Field Instrumentation For Soil

Copyright by ASTM Int'l (all rights reserved); Mon Jun 26 17:14:37 EDT 2017
Downloaded/printed by
University Of Western Ontario (University Of Western Ontario) pursuant to License Agreement. No further reproductions authorized.
STP 1358
Field Instrumentation for Soil

and Rock
Gary N. Durham and W. Allen Marr, editors
ASTM Stock Number: STP 1358
ASTM
100 Barr Harbor Drive
West Conshohocken, PA 19428-2959
Field instrumentation for soil and rock / Gary N. Durham and W. Allen Marr, editors.
(STP ; 1358)
Papers from a symposium held in Atlanta, Ga., June 18-19, 1998.
Includes bibliographical references and index.
ISBN 0-8031-2604-2
1. Engineering geology--Instruments. 2. Soil-structure interaction. 3. Geophysical
instruments. I, Durham, Gary N., 1942- II. Marr, W. Allen. III. ASTM special technical
publication ; 1358.
TAT05 .F48 1999

624.1'51'028--dc21
99-046576
Copyright 9 1999 AMERICAN SOCIETY FOR TESTING AND MATERIALS, West

Conshohocken, PA. All rights reserved. This material may not be reproduced or copied, in whole or in
part, in any printed, mechanical, electronic, film, or other distribution and storage media; without the
written consent of the publisher.
Photocopy Rights
Authorization to photocopy items for internal, personal, or educational classroom use, or the
internal, personal, or educational classroom use of specific clients, is granted by the
American Society for Testing and Materials (ASTM) provided that the appropriate fee is paid to
the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923; Tel: 508-750-8400;
online: http://www.copyright.com/.
Peer Review Policy
Each paper published in this volume was evaluated by two peer reviewers and at least one
editor. The authors addressed all of the reviewers' comments to the satisfaction of both the technical
editor(s) and the ASTM Committee on Publications.
To make technical information available as quickly as possible, the peer-reviewed papers in this
publication were prepared "camera-ready" as submitted by the authors.
The quality of the papers in this publication reflects not only the obvious efforts of the authors
and the technical editor(s), but also the work of the peer reviewers. In keeping with long standing
publication practices, ASTM maintains the anonymity of the peer reviewers. The ASTM Committee on
Publications acknowledges with appreciation their dedication and contribution of time and effort on
behalf of ASTM.
Printed in Chelsea, MI
October 1999
Foreword
This publication, Field Instrumentation for Soil and Rock, contains papers presented at the sym-
posium of the same name held in Atlanta, Georgia, on June 18-19, 1998. The symposium was held
in conjunction with the June 14-17, 1998 standards development meetings of Committee D-18 on
Soil and Rock and its Subcommittee D18.23 on Field Instrumentation, the symposium sponsor. The
symposium was chaired by Gary N. Durham, Durham Geo Enterprises Inc., GA; W. Allen Man',
Geocomp Corporation, MA, was the co-chairman. They also both served as STP editors of this pub-
lication.
Contents
Overview ix
INSTRUMENTATION ASSOCIATED WITH SOIL STRUCTURE INTERACTION
Comparison of Horizontal Load Transfer Curves for Laterally Loaded Piles from Strain
Gages and Slope Inclinometer: A Case Study--PAULOL. PINTO,BRIANANDERSON,AND
FRANK C. TOWNSEND 3
An Instrumented Geogrid Reinforced Slope in Central Italy: Field Measurements

and FEM Analysis Results--FILIPPOMONTANELLI,PIERGIORGIORECALCATI,
AND PIETRO RIMOLDI 16
Uses of Automated Geotechnical Instrumentation Systems---w. ALLENMARR 36
Instrumentation and Long Term Monitoring of a Soil Nailed Slope at Madeira Walk,
Exmouth, UK--ANTHONY D. BARLEY, MICHAEL C. R. DAVIES, AND ALUN M. JONES 51
Instrumentation for Monitoring Large-Span Culverts---MARK C. WEBB, ERNEST T. SELIG, AND

TIMOTHY J. MCGRATH 66
Measurement of Earth Pressures on Concrete Box Culverts under Highway

E m b a n k m e n t s - - M i C H A e L ZHIQIANG YANG, ERIC C. DRUMM, RICHARD M. BENNETT, AND
MATTHEW MAULDON 87
Instrumentation for Monitoring Buried Pipe Behavior During Backfilling--

TIMOTHY J. MCGRATH, ERNEST T. SELIG, AND MARK C. WEBB 101
INSTRUMENTATION SUPPORT CONSTRUCTION ACTIVITIES
Monitoring Pavement Response and Performance Using In-Situ Instrumentation--

DAR-HAO CHEN, JOHN BILYEU, AND FRED HUGO 121
Apparatus and Method for Field Calibration of Nuclear Surface Density Gauges---
ALI REGIMAND AND ALAN B. GILBERT 135
Two Weight System for Measuring Depth and Sediment in Slurry-Supported Excavations--
PETERDEM1NG AND DAVID GOOD 148
INSTRUMENTATION TO MONITOR LANDFILLS
A Comparison of Portable and Permanent Landfill Liner Leak Detection Systems---

STEVEB. TAYLOR,CHRISC. WHITE, AND RON D. BARKER 157
Instrumentation of Dredge Spoil for Landfill Construction--MICHAEL J. BYLE,

MONICAL. MCCULLOUGH,RODNEYALEXANDER,N.C. VASUKI,AND JAMESA. LANGER 175
Field M e a s u r e m e n t s of F r o s t Penetration into a Landfill Cover That Uses a P a p e r

Sludge Barrier--HORACE K. MOO-YOUNG,CHRIS LAPLANTE,THOMASF. ZIMMIE,
AND JUANQUIROZ 191
Stability M o n i t o r i n g System for the F r e s h Kills Landfill in New Y o r k C i t y - -

THOMAS G. THOMANN,MAJEDA. KHOURY,JACKL. ROSENFARB,AND
RICHARD A. NAPOLITANO 206
INSTRUMENTATION FOR MONITORING SETTLEMENTAND STABILITY
Field M e a s u r e m e n t of W a t e r Contents and Densities by Nuclear M e t h o d s

in Clay Deposits---VINCENT SILVESTRI 223
M o n i t o r i n g P r e l o a d Performance---JAMES B. RUSSELL,CHRISTOPHERA. LAWRENCE,

MICHAELW. PADDOCK,AND DOUGLASROSS 236
Instrumentation for Slope S t a b i l i t y - - E x p e r i e n c e from an U r b a n Area--PHIL FLENTJE

AND ROBIN CHOWDHURY 250
DATA ACQUISITIONAND DATA MANAGEMENT
R e m o t e Field M e t h o d s to M e a s u r e F r o s t Depth----cRAi6 H. BENSONAND PETER J. BOSSCHER 267
Integrated Automation of the New Waddell Dam Performance Data Acquisition S y s t e m - -

L. RANDALLWELCHAND PAULE. FIELDS 285
Using Geomatics in the Acquisition and Management of Field Data--WILLIAM S. BROKAW 298
Documentation of Data Collected from Field Instrumentation for Spatial Applications---

DAVIDT. HANSEN 307
Data Acquisition and Management for Geotechnical Instrumentation on the Central

A r t e r y / T u n n e l P r o j e c t - - M . r~wI~S AND W. A. MARI~ 319
Data Management System for Organic Soil--PAMELA STINNETTE 333
Review of GIS Developments in Geotechnical Instrumentation on the C e n t r a l A r t e r y /

Tunnel Project, Boston Massachusetts--D.J. BOBROWAND S. VAGHAR 360
INSTRUMENTATION FOR MEASURING PHYSICAL PROPERTIES IN THE FIELD
Use of Tensiometer for In Situ Measurement of Nitrate Leaching--

KANGLINLI ANDM.R. REDDY 371
Compaction Comparison Testing Using a Modified Impact Soil Tester and Nuclear Density
Gauge---a.A. ERCHUL 382
Selecting Surface Geophysical Methods for Geological, Hydrological, Geotechnical, and

Environmental Investigations: The Rationale for the ASTM Provisional Guide---
WAYNE SAUNDERS, RICHARD BENSON, FRANK SNELGROVE, AND SUSAN SOLOYANIS 392
The Key Role of Monitoring in Controlling the Construction of the New Jubilee
Line Between Waterloo and Westminster, London--N. H. OSBORNE 404
Indexes 411
Overview
This overview summarizes the results of the Symposium on Field Instrumentation for Soil and
Rock held in Atlanta, Georgia June 18-t9, 1998 as part of Committee D18 scheduled meeting
week. ASTM's subcommittee D 18.23 on Field Instrumentation sponsored the symposium. Major
session topics include:
9 Field Data Acquisition and Data Management

9 Geotechnical Instrumentation of Landfills
9 Instrumentation for Project Cost Management
The following is an overview of the information presented in this volume.
Instrumentation Associated with Soil Structure Interaction and Construction Monitoring

Three papers address field instrumentation associated with culverts and buried pipe. McGrath et
al. presented the results of backfilling on pipe of various material and diameter. Unique methods
were presented for instrumenting corrugated pipe. Webb et al. discussed instrumentation of concrete
and corrugated steel arch culverts with spans of about 9 m. Various instruments were installed to
monitor structural strain and deformation and culvert-soil interface pressure. The arches were back-
filled and tested under live loads. Deformation measurements of the walls of the culverts were made
with a custom-built laser device developed for this research. Yang et al. presented convincing evi-
dence on earth pressure instrumentation of concrete box culverts beneath highway backfill that cur-
rent ASSHTO guidelines underestimate earth pressures perhaps as must as 30%.
Barley et al. described instrumentation of strain gauged steel bars used in a soil nailed slope some
10-m high. Montannelli describes instrumentation of geogrid used in a reinforced slope. In situ
pullout tests that were performed were in good agreement with scaled laboratory pullout tests. The
study found good agreement between observed stress and deformations with those predicted from
FEM modeling.
The paper by Chen et al. discussed the effectiveness of in situ instrumentation of diagnosing the
pavement layer conditions under full scale accelerated traffic loading. Multi-depth deflectometers
were used to measure both permanent deformations and transient deflections resulting from traffic
loading and falling weight defleetometer tests.
Russel et al. presented a case history of monitoring preload performance for a highway alignment
traversing deposits of organic silt and peat. Flentje and Chowdhury used a case history to point out
the unique problems associated with instrumentation programs installed in urban areas.
Boston's Central Artery/Tunnel Project

Two papers were presented concerning the geotechnical instrumentation aspects of the Central
Artery/Tunnel Project, Boston, Massachusetts. Bobrow and Vaghar discussed the Geographic Infor-
mation System (GIS) that were developed for rapid analysis and reporting of instrumentation and
Copyright by ASTM Int'l (all rights reserved); Mon Jun 26 17:14:37 EDT 2017 ix
X FIELDINSTRUMENTATION FOR SOIL AND ROCK
survey data on the Central Artery/Tunnel Project. The paper by Hawkes and Marr discussed the role
of automation and reviewed the data management system developed for processing large quantities
of field data on a daily basis.
Instrumentation to Monitor Landfills

Four papers were specific to geotechnical instrumentation of landfills. The paper by Taylor et al.
discussed landfill liner lead detection systems. Their work indicated that a below liner monitoring
grid combined with above line surveys to pinpoint leaks accurately offer a successful approach to
leak detection of lined waste sites. The paper by Thomann et al. presented a case history of the
unique problems connected with installing geotechnical instruments within refuse fill and the under-
lying foundation to monitor pore pressures, lateral and vertical movements and temperature. They
reported that the attrition rate of instruments in a refuse environment is about 10 to 15% per year;
therefore, redundancy of instruments is mandatory to insure a successful program. Byle et al. dis-
cussed instrumentation of a landfill liner constructed of dredge spoil. Moo-Young et al. presented a
method using thermistor and conductivity probes for measuring frost penetration that might deterio-
rate the permeability of the landfill's liner or cap. Another paper by Benson and Bosscher discussed
results of their study indicating frost depth can be more reliably measured using electrical resistivity
or dielectric constant rather than with thermistors.
Data Acquisition and Data Management

A paper by Marr described the application of recent developments in electronics and instrumenta-
tion to three geotechnical cases to illustrate some of the benefits of good automated field instrumen-
tation systems. He also reviewed the considerations involved in deciding whether to automate a ge-
otechnical instrumentation sytem. Welch and Fields described the automation of the field
instrumentation associated with a large dam. Fiber optic cable was used extensively from the vari-
ous sensors to eliminate voltage surges. The automation included satellite as well as land line re-
mote transmission of the data. A related paper by Brokaw discussed the use of geographic informa-
tion systems (GIS) coupled with global positioning satellite technology for field data acquisition
and data management. Hansen furthered the overall review with a discussion of relating field instru-
mentation data that has been incorporated into GIS framework and the implications to other geospa-
tial data.
Acknowledgments
The editors express grateful appreciation to the authors and the many engineers and scientists
who provided peer review for all papers. The ASTM editorial staff deserves special credit for their
support and encouragement throughout.
Gary N. Durham
Durham Geo-Enterprises Inc.,
Stone Mountain, GA 30087;
Symposium chairman and editor
W. Allen Marr
Geocomp Corporation, Boxborough,
MA 01719;
Symposium cochairman and editor
Instrumentation Associated with
Soil Structure Interaction
Paulo L. Pinto l, Brian Anderson2, and Frank C. Townsend3
Comparison of Horizontal Load Transfer Curves for Laterally Loaded Piles from
Strain Gages and Slope Inclinometer: A Case Study
REFERENCE: Pinto, P. L., Anderson, B., and Townsend, F. C., "Comparison of

Horizontal Load Transfer Curves From Strain Gages and Slope Inclinometer: A
Case Study," Field Instrumentation for Soil and Rock, ASTM STP 1358, G.N. Durham
and W.A. Mart, Eds., American Society for Testing and Materials, West Conshohocken,
PA, 1999.
ABSTRACT: Laterally loaded deep foundations are commonly analyzed using the
Winkler model with the soil-pile interaction modeled through nonlinear springs in the
form ofp-y curves. Computer programs such as FloridaPier and COM624P use default
p-y curves when performing lateral analyses. These curves are based on input soil
properties such as subgrade modulus, friction angle, undrained shear strength, etc. Soil
properties must be deduced by laboratory testing or correlation to in situ test results. This
is a source of uncertainty. In a few cases, lateral load tests are performed on
instrumented piles, and the validity of such assumptions can be assessed. Test piles are
commonly instrumented with strain gages and/or inclinometers. P-y curves can be back
computed from these data, and the curves obtained with the two methods should agree
closely. Results from a field test on a concrete pile are presented and the critical factors
for the analysis are discussed. One added difficulty with concrete is its nonlinear
behavior particularly near structural failure. The curves obtained are also compared with
those developed from the Dilatometer/Cone Pressuremeter Test and the Standard
Penetration Test.
KEYWORDS: piles, deep foundations, instrumentation, data reduction, lateral load test,
P-y curves, strain gages, slope inclinometers
1Assistant Professor, Department Engenharia Civil, Universidade de Coimbra, 3004

Coimbra, Portugal.
2phO candidate, Department of Civil Engineering, University of Florida, Gainesville, FL

32611-6580.
3professor, Department of Civil Engineering, University of Florida, Gainesville, FL

32611-6580.
Copyright by ASTM Int'l (all rights reserved); Mon Jun 26 17:14:37 EDT 2017 3
Copyright 9 1999 by ASTM International www.astm.org
4 FIELD INSTRUMENTATION FOR SOIL AND ROCK
Introduction
The analysis of laterally loaded piles usually utilizes the concept of a beam on an
elastic foundation represented by nonlinear curves of soil resistance versus horizontal
deflection (p-y curves). It is a problem of soil-structure interaction, such that deflection
of the pile depends on the soil response and the soil response depends on the pile
deflection. Ap-y curve represents the total resistance at a particular depth to the lateral
displacement of a horizontally loaded pile.
In current practice, the load transfer curves are generated based on soil type,
strength and deformability properties, and geometry of the pile as well as loading
conditions. Several authors (Reese et al. 1974, 1975; O'Neill and Murchison 1983) have
proposed curves, for different types of soil and loading conditions, based on a limited
number of field tests. In these tests, the piles have been instrumented with strain gages or
with inclinometers. It is of great interest to compare thep-y curves from lateral load tests
on piles with both strain gages and inclinometers. The inclinometer has become the
instrumentation of choice for concrete piles, for reasons that will be presented later.
Theory
P-y Curves From Strain Gage Data
A laterally loaded pile can be analyzed as a beam under composed bending (pure
bending plus axial load). The governing equation is,
E1~4
dx +pxd2y +E , y -- o (1)
where Elis the flexural rigidity of the pile, Px is the axial load, Es is the secant modulus
of soil reaction (slope ofp-y curve), y is the lateral deflection and x the depth. On most
lateral load tests, the effects of the axial load may be neglected and the second term is
dropped.
During a lateral load test, strains are measured at given depths, for each load
increment. With the strains, the curvature, ~, and the bending moment at the section are
determined.
M = EI ~ = Elcb (2)
dx
The horizontal displacements are obtained by double integration of the moment

diagram, and the lateral earth pressures are computed by double differentiation. In order
to obtain the horizontal displacements some boundary conditions, such as the deflection
and slope at the top of the pile, have to be known. For a pile with linear elastic behavior
the solution is straightforward. The integration and derivation &the moment distribution
can be performed using several numerical techniques. Due to the scarcity and spacing of
PINTO ET AL. ON A CASE STUDY 5
the data points, they are commonly fit using a polynomial or a spline. This continuous
function is differentiated and integrated. The procedure is repeated for each load
increment. At specific depths, the pressure from each load increment is plotted versus the
corresponding deflection resulting in ap-y curve.
For a nonlinear analysis, the flexural rigidity of the pile has to be adjusted,
depending on the strain level. The procedure is more elaborated and a correct
characterization of the nonlinear behavior of the pile is crucial. Table 1 and Figure 1
show the nonlinear variation of stiffness with curvature for a prestressed concrete pile. It
was computed by applying increasing loads at the top of a pile fixed at the tip and
determining the corresponding curvature.
TABLE 1--Nonlinear variation offlexural rigidity with curvature for test pile T1
Moment ~smax 8srni n Curvature El
(kN-rn) (tract) (comp) (l/m x 10"6) (kN-rn2 xl03)
0 0.000 9i2.6
113 3.70E-05 -3.70E-05 0.124 911.6
226 7.40E-05 -7.40E-05 0.248 911.6
339 1.11E-04 -1.11E-04 0.372 910.1
452 1.50E-04 -1.53E-04 0.506 893.2
565 1.89E-04 -1.97E-04 0.647 873.8
678 2.28E-04 -2.42E-04 0.787 861.3
791 2.70E-04 -2.90E-04 0.940 841.8
904 3.34E-04 -3.50E-04 1.145 789.2
1017 5.40E-04 -4.61E-04 1.678 606.1
1073 8.78E-04 -5.62E-04 2.412 444,9
1130 1.03E-03 -6.24E-04 2.777 406.8
1186 1.20E-03 -6.90E-04 3.158 375.7
1243 1.36E-03 -7.56E-04 3.538 351.3
1356 4.23E-03 -1.22E-03 9.122 148.6
1000
800_~
600
O
"E
i
Z 400 ~. . . . .
[]
20O
0
2 4 6 8 10
Curvature, ~(l/m x 10 ~)
FIG. 1--Graphical depiction of nonlinear variation offlexural rigidity with

curvature for test pile T1
6 F I E L DINSTRUMENTATIONFOR SOIL AND ROCK
The main advantage of using strain gages is that stresses and bending moments
can be computed directly from the measured data. The p-y curves can be obtained
without estimating soil properties. The major drawback is that the gages are often
damaged during the installation of the piles and drilled shafts. Strain gages are sensitive
equipment that can easily be damaged by improper handling.
P-y Curves From Slope Inclinometer Data
The use of slope inclinometers to measure lateral deflection of the pile has
substantial advantages compared to the traditional strain gages. A grooved PVC pipe is
inserted in the pile or drilled shaft and the variation from the original vertical position is
measured using a slope inclinometer. It is quite economical since the only material not
recovered is the PVC pipe. When used in concrete piles/shafts it avoids the difficult and
expensive task of installing strain gages. The main disadvantage is that data reduction
requires more judgement. A direct computation of soil pressure is hindered by numerical
errors, because the slope distribution would have to be differentiated three times.
A method based on the assumption o f a predefinedp-y curve and on the
optimization of the soil parameters that match the deflected shape has been proposed
(Brown et al 1994), with good results. The disadvantage of such an approach on the
calibration ofp-y curves is the assumption of the shape of the curve. Additional effort is
also required for the iteration over two or more soil parameters (angle of internal friction
and modulus of lateral soil reaction for sands, or undrained shear strength and strain at
50% of strength for clays).
The authors have used a simpler approach, by iterating on a single parameter, the
modulus of lateral soil reaction, k (F/L3). For uniform sands, the initial slope of the p-y
curves increases linearly with depth, z, and is equal to
Es(z) = k z (3)
Once the pile is loaded and deformation occurs, the secant modulus of the p-y curve will
no longer have a linear variation as shown in figure 2. It is reasonable to consider that if
the pile is flexible, the horizontal displacements at a depth close to the pile tip will be
very small. In this case the secant modulus will have a small variation, for each load
increment.
The variation of the secant modulus with depth (z), for a given horizontal toad, is
approximated to a parabola with equation where a and b are arbitrary coefficients:
Es(z) = az ~ + bz (4)
Note that this corresponds to a linear variation of the modulus of lateral soil reaction, k,
with depth given by:
dEs
k= = 2az + b (5)
dz
/"
H t
$ s
I/ s
( l Y
I if
It I
Y
FIG. 2--Evolution of secant modulus with depth and deformation
If the secant modulus at the pile tip Esmax,is assumed constant for all loads, the parameter
a is defined and k will depend solely on the value of the modulus of lateral soil reaction at
the surface, b (figure 3).
a= (Em~x-bZmax) (6)
2
Zmax
Es max
Es (secant)
Load 3
Depth Zmax
FIG. 3--Assumed evolution of secant modulus with depth and load level
The parameter b is iterated upon, until good agreement of the deflected shapes is found.
The measure of the error is given by
8 FIELDINSTRUMENTATIONFOR SOIL AND ROCK
~d = ElYpredtcted(i)--Y ........ d(i)l (7)

i=1
where n is the number of points where the deflection Was evaluated. The prediction is
obtained using a finite-difference technique where the governing equation (Eq 1) is
solved for particular points along the depth. This is similar to the method used in the
program COM624 (Wang and Reese 1991) but was performed using a simple
spreadsheet. The same spreadsheet is capable of carrying out the iterative analysis.
Curves from DMT and Cone Pressuremeter
As a comparison to the curves obtained from instrumentation, curves based on in

situ tests were computed. A composed curved with the initial slope, Kl, estimated from
the dilatometer modulus (Ed) and the ultimate pressure (Plim) based on the cone
pressuremeter was chosen. The dilatometer modulus is obtained using the Marchetti
dilatometer and is related to the initial slope of the p-y curve by an empirical factor F~,
varying from 2 to 3.
K l = F, E~ (8)
The combined curve is linear up to a horizontal displacement of 1.1-mm (DMT B

reading) and hyperbolic from then on. The ultimate p is estimated using the cone
pressuremeter limit pressure multiplied by the pile width, D, and an empirical factor a
(Robertson et al, 1985) and is given by:
Putt = (X P l i m D (9)
Robertson and co-authors proposed a linear variation ofc~ to a depth of 4 diameters. At

the surface a is zero and at 4D it is equal to 1.5. Thereafter it is constant and equal to
1.5.
Results
As an example of the application of both methods, the authors selected a pile

tested as part of a large (4x4) group tested near Roosevelt Bridge, in Stuart, Florida
(Townsend et al. 1997). This was a well instrumented test, with extensive in situ testing
performed to characterize the soil layers.
The prestressed concrete pile was 16.2 m long (2.4 m of free length, above the
mud line) and had a square section of 0.76 x 0.76-m. (fc'=41 MPa, Ec --- 30.4 GPa). It
was located at the right comer of the trail row (pile 1 in figure 4). The pile was
instrumented with a series of nine full bridges of strain gages. A slope inclinometer
guide pipe was installed in the center of the pile, to a depth of 9.0 m from the ground
surface. The reduced data from both the strain gages and slope inclinometer are shown in
figures 5and 6, respectively.
FIG. 4--Location of piles in test group
Bending Moment (kN*m)
0 200 400 600 800 1000 1200

i i i i i
4, 9 9 9 o [] a o
..... ~ t -- ~A- - - ~ - ~ - - - - Z ~ - 0 . . . . . . . . . . . . . . . . . . . . . . . .
Lateral
9 9 9 9 0 [] A 0
Load in
Kilonewtons
9 9 9 9 0 rl A 0
E
9 62
=~3 ~ - II 9 --Q o D ~, ---- u 9 93 [
r 9 9 9 9 0 [] A 0 9 121 L
01531
[J 182
a211
o 258
9 9 .... ~ 9 o [] A 0 1
...........................................................................................................................................................................................................................
9 9 9 9 o [] zk 0 .,,..I
FIG. 5--Bending Moments from Strain Gage Readings
Horizontal Displacement (m)
-0.01 0.00 0.01 0.02 0,03 0.04 0.05 0.06 0.07

-2
Lateral
Load in
Kilonewtons
36
A
.-a-62 I
9- k - g r J I
o --e- 122 I
--e- 153 I
___%m --B- 183 I
--a- 208 I
--e-- 243 I
10
FIG. 6--Displacements from the Slope Inclinometer
The local geology consists of a surface layer of slightly silty fine sand (d~' =35 ~
~,'=10.4 kN/m3), 4.0 m thick overlaying a layer of cemented sand (d~' = 44 ~ ~/'=12.0
kN/m 3), 10.0 m thick. The ground water table was located above the mud line.
Following the steps outlined for strain gage data and accounting for the nonlinear
properties of the concrete section, the p-y curves for the first four meters were determined
as those shown in Fig. 7.
The shape is hyperbolic, as suggested (O'Neill and Murchison 1983) for sands,
and the limiting force per unit length increases with depth. The curves can only be traced
to the maximum horizontal displacement observed during the test. Proper extrapolation
is required for larger displacements.
As shown in Fig. 8, the method used with the inclinometer data yielded curves in
good agreement with the strain gages. A key factor to obtain similar and meaningful
curves is the correct assessment of the flexural stiffness. It has contrary effects on the
results, depending on which data are analyzed. A pile stiffness higher than correct will
result in higher bending moments, shifting up the p-y curves obtained from strain gage
data. The inverse will be observed for inclinometer results. If the pile stiffness is
overestimated, the soil will have to be softer to provide the same horizontal
displacements. Therefore, the curves will be shifted downward. On the particular case of
concrete piles, loaded past the moment that causes the concrete to crack, the proper
modeling of the non-linear moment curvature relation is critical. The inclusion of
nonlinear pile behavior marginally increases the complexity in the data reduction from
the strain gages. That is not the case for the inclinometer, where the iterative loop would
have to consider the curvature of each section and the bending stiffness would be
PINTO ET AL. ON A C A S E STUDY 11
140
Distance
120 to Ground
Surface
100 (m)
A 80 f ,
9--~- 3..o
40
A
20
0
0 5 10 15 20 25 30 35
Displacement,y (mm)
FIG. 7--Curvesfrom Strain Gages
140
Distance
120 to Ground
Surface
100 (m)
" 8O
Q. 60
40
20
0 iI--- ....... ~ - - - - ~ ~- ........ +-
0 5 10 15 20 25 30 35
Displacement, y (mm)
FIG. 8--Curves from Inclinometer
computed for each iteration. Although not extraordinarily difficult, that would limit the
benefit of the simple procedure outlined and implemented on a spreadsheet.
Curves Based on In situ Testing
The fact that the pile was included in a group complicates the scope of the
problem. Ideally, a single pile would have been chosen, but the data available for
concrete piles is very scarce and few tests have been conducted with both strain gages
and inclinometers. For the purpose of comparison of the curves with the curves based on
the in situ tests the situation is not ideal. The fact that the pile was included in a group
greatly influences the pile-soil interaction. A p - y multiplier (Brown et al. 1988) was
required to allow comparison of curves. The concept of the multiplier forces the curve to
shift down, reducing capacity, to account for the pile-soil interaction. P-y multipliers are
dependent on position within the group, relative to the point of application of the load,
pile spacing and soil type. (McVay et al. 1995)
The curves calculated using DMT and pressuremeter data measured at a depth of
2 meters are displayed in Fig. 9. The DMT curves seem to have slightly lower soil
resistance but are still in reasonable agreement with the curves obtained from both
inclinometers and strain gages. A multiplier of 0.3 (70% of reduction in the curve) had
been back computed based on the overall group performance.
140
120
100
--==-2.0 m
60
40 --e- 2.0 m with P-Y

Multiplier of 0.3
20
0 10 20 30
Displacement, y (mm)
FIG. 9--Curves from Dilatometer and Pressuremeter
If Oneill Sand P-y Curves (O'Neill, 1983) are used with ~'=35 ~ and ~,'=10.2
kN/m 3 estimated from SPT tests and a multiplier of 0.3, the resulting curves are plotted in
Fig. 10. The resulting p-y relationship is stiffer than the prior and would have resulted in
smaller displacements than observed in the test.
250
- - l - 2.0 m
200
k
I - * - 2.0 rn with P-Y
~- 150 Multiplier of 0.3
a. 100
50
0 5 10 15 20 25 30 35
Displacement, y (ram)
FIG. 1O--Theoretical curves based on SPT
In order to evaluate the p-y curves obtained from instrumentation, a program with
full nonlinear analysis capability was required. The program FloridaPier (McVay et al.
1998) was selected for that purpose. User defined curves, based on strain gage and
inclinometer data, was used to a depth of 4.5 m. From that point to the pile tip, the soil
springs were modeled using O'Neill curve for sands with d~'=l 5 ~ k=16,290 kN/m 3.
These low values correspond to the application of a multiplier of 0.3. The predicted pile
head displacements match well with the values observed during the load test (Fig. 11).
The pile behaves almost linear until the moment caused by a horizontal load of 200 kN
cracks the concrete section.
Conclusions
Lateral load tests in deep foundations are important to access the validity ofp-y
curves used in design. Two methods used to determine the p-y curves using strain gage
and inclinometer test data were presented and discussed. A simple method to reduce the
inclinometer data, for cohesionless uniform soils, was proposed and the resulting curves
compared well with the curves deduced from strain gage data. The scarcity of lateral
load tests on concrete piles/shafts, instrumented with strain gages and slope inclinometers
prevented the extension of the current comparison to a larger database.
Reinforced concrete piles loaded close to structural failure exhibit nonlinear
behavior. The consideration of this nonlinear behavior is important to interpret the test
14 F I E L DINSTRUMENTATION FOR SOIL AND ROCK
300
250
200
z
-e-Wire L i n e - A c t u a l Deflection
"o 150-
F L P I E R Prediction -
Strain Gages
100 -
- - B - F L P I E R Prediction -
Inclinometer
50-
0 il
0 10 20 30 40 50 60 70 80 90
Top Displacement (mm)
FIG. 11 -- Pile Head Deflection, comparison between test results and predictions
results correctly. The poor evaluation of the bending stiffness will cause the p-y curves
from strain gages and inclinometer to move in opposite directions.
For the current case, where the pile was included in a group, comparison of the
curves based on instrumentation with curves based on in situ testing required the use of
p-y multipliers. I f a value of 0.3 is used, the curves agree reasonably. When the
instrumentation curves are used in a full nonlinear analysis, Figure 11 shows there is an
excellent match of pile head deflection with the behavior observed in the load test.
References
Brown, D. A., Morrison, C. and Reese, L., 1988, "Lateral Load Behavior of Pile Groups
in Sand," ASCE, Journal of Geotechnical Engineering, Vol. 114, No. 11, pp.
1261-1276.
Brown, D. A., Hidden, S. A. and Zhang, S., June 1994, "Determination of P-Y Curves
Using Inclinometer Data," Geotechnical Testing Journal, Vol. 17, No. 2, pp. 150
-158.
McVay, M. C., Casper, R. and Shang, T., 1995, "Lateral Response of Three- Row Groups
in Loose to Dense Sands at 3D and 5D Pile Spacing," ASCE, Journal of
Geotechnical Engineering, Vol. 121, No. 5, pp. 436-441.
McVay, M., Hays, C. and Hoit, H., 1998, "User's Manual for FloridaPier for NT Version
1.22," Department of Civil Engineering, University of Florida, Gainesville.
O'Neill, M. W. and Murchison, J. M., 1983, "An Evaluation of P-y Relationships in

Sands," Research Report No. GT-DF02-83, Department of Civil Engineering,
University of Houston, Houston, Texas.
Reese, L.C., Cox, W.R. and Koop, F.D., 1974, "Analysis of Laterally Loaded Piles in
Sand," Paper No. OTC 2080, Proceedings, Sixth Annual Offshore Technology
Conference, Houston, TX.
Reese, L.C., Cox, W.R. and Koop, F.D., 1975, "Field Testing and Analysis of Laterally
Loaded Piles in Stiff Clay," Paper No. OTC 2312, Proceedings, Seventh Annual
Offshore Technology Conference, Houston, TX.
Robertson, P. K., Campanella, R. G., Brown, P. T., Grof, I. and Hughes, J. M., 1985,
"Design of Axially and Laterally Loaded Piles Using In situ Tests: A Case
History," Canadian Geotechnical Journal, Vol. 22, No. 4, pp. 518-527.
Townsend, F. C., McVay, M. C., Ruesta, P. and Hoyt, L., 1997, "Prediction and
Evaluation of a Laterally Loaded Pile Group at Roosevelt Bridge," Report No.
WPI 0510663, Department of Civil Engineering, University of Florida,
Gainesville, FL.
Wang, S. T. and Reese, L. C., 1991, "Analysis of Piles Under Lateral Load - Computer
Program COM624P for the Microcomputer," Report No. FHWA-SA-91-002,
U.S. Department of Transportation, Federal Highway Administration,
Washington, DC.
Filippo Montanelli, 1 Piergiorgio Recalcati, 1 and Pietro Rimotdi 1
An instrumented geogrid reinforced slope in central Italy: field measurements and

FEM analysis results
Reference: Montanelli, F., Recalcati, P., Rimoldi, P., "An instrumented geogrid
reinforced slope in central Italy: field measurements and FEM analysis results,"
Field Instrumentation for Soil and Rock, ASTM STP 1358, G.N. Durham and W.A. Marr,
Eds., American Society for Testing and Materials, West Conshohocken, PA, 1999.
Abstract: A 15 m high green faced reinforced slope was built in 1996 to stabilise a
landslide situated on the Montone hill in the province of Perugia (Italy). The reintorced
slope was built using locally available soils as fill material and HDPE mono-oriented
extruded geogrids as reinforcement. The reinforced slope has been instrumented with
strain gages connected to the reinforcing geogrids and with total pressure cells; full scale
in situ pull-out tests have been performed as well.
The preliminary results demonstrate the good performance of the geogrids and define
possible failure mechanisms for slopes with stepped geometry. The in situ pull-out test
results validate laboratory tests performed in the past. To evaluate the field stress and
deformation behavior of the reinforced slope a finite element analysis has been carried
out. The paper describes the model technique developed to evaluate the field stress and
deformation of steep reinforced slope using special interface elements. The results of
finite element analysis are in agreement with the field measured results.
Keywords: slope reinforcement, field instrumentation, geogrid, FEM analysis
Introduction
The Montone hill, located in the province of Perugia (Central Italy), shows
important instability phenomena along its slopes, resulting from a complex evolutive
situation. Many factors have an important role, including the lithology, the structural
1TENAX SpA, Vigan6, LC, Italy.
MONTANELLI ET AL. ON GEOGRID REINFORCED SLOPE 17
arrangement of the geological units, the still active neotectonic phenomena. The hills also
suffered from hydrological and hydrogeological conditions, and the anthropic
modification suffered by the territory during the last decades.
During the last decades few landslides concerned the North-East slope of Montone
hill, particularly along the Montone-Pietralunga main road. These phenomena involved
an area about 15 m high and 200 m long; in the proximity of the Fosso Fornaci a road
failure occurred (Fig. 1).
In one of the sub-projects for the stabilisation of the hill, along the Montone-Perugia
main road, the use of geosynthetics was foreseen for different functions, and more
precisely:
9 soil reinforcement to build reinforced slopes (using mono-oriented HDPE
geogrids);
9 erosion control to protect the slopes affected by erosive phenomena (using
geocells, polypropylene geomats and biomats composed of coconut and straw
fibres);
9 separation and filtration for the construction of drainage trenches (using
nonwoven geotextiles).
FIG.1 - Reinforced slope location.
The reinforced slope presents a maximum height of 15.5 m and a horizontal length of
about 53 m (see Figure 1). The geogrid reinforced steep slope was divided in 3 steps of
about 5 m each. The instrumented section was equipped with 40 strain gages (electrical
wire resistance type) bonded on the geogrid ribs of the first and the second reinforced
block; 3 total vertical pressure cells positioned at the bottom of the first reinforced block;
2 vertical inclinometers installed on the first block and on the top of the third block; 1
pore pressure cell and 1 acquisition system installed on top of the first block.
Field pull-out tests have been performed up to failure on dummy geogrid layers
purposely embedded into the bottom step of the slope.
Geotechnical characteristics
Due to the extension of the area and the complexity of the problems, a large scale
geophysical and hydrological test campaign has been carried out, with boreholes,
piezometric and penetration tests, plus extensive geotechnical laboratory testing (Coluzzi
et al., 1997). The area involved in the works is characterised by the presence of a silty
soil of lacustrine origin, locally emerging in correspondence of the instability phenomena.
Along the main landslide surfaces, yellow silty sands with layers of gravel, up to 1 m
thick, are present (see Figure 2).
F I G . 2 - G e o l o g i c a l cross section.
The water table in this area is located at 2 m depth from the ground level. A
geological cross section has been prepared for the main landslide as shown in Figure 2.
MONTANELLI ET AL. ON GEOGRID REINFORCEDSLOPE 19
Through traditional slope stability analysis, a sliding-rotational failure mechanism was

identified in this section very close to the ground surface (approximately at 2.5 m depth).
This failure was due both to the slope cutting operations for the construction of the road
and to the loss of mechanical characteristics in the above mentioned soils due to the
saturation under the water table level. Old landslide bodies have been found in the area
by site investigation. The geotechnical parameters for the surface colluvial layer of soil
and for the deeper subgrade are reported in Table 1.
TABLE 1 - Geotechnical characteristics of the soils involved in the landslide.
Soil Surface coUuvial layer Deeper subgrade
Unit weight [kN/rrd] 21.2 20.3 - 21.7
Friction angle [o] 19.7 22.4 - 27.5
Undrained cohesion [kPa] 0 8.3 - 95
Cohesion [kPa] 37 - 69
Water content [%] 15.7- 19.9
A back-analysis of the global stability using Janbu method has been performed on
the failed slope: considering that Montone is in a highly seismic area and therefore
applying a seismic load corresponding to an acceleration ratio a/g equal to 0.07, and
considering the in-situ soil with a unit weight of 19.5 kN/m ~ and a cohesion of 0 kPa, an
overall friction angle of 26 ~ was found. This value was in very good agreement with the
geotechnical data of the tested soils. For both economical and practical reason (access to
the site, easiness and speed of construction), it was decided to stabilise the slope by
rebuilding it with the local soil reinforced with HDPE geogrids.
Design and construction of the reinforced slope
The problems to be faced were the presence of a soil with poor geotechnical
characteristics and the presence of a very high water table.
Therefore, for a running length of 50 m, the body of the landslide has been
completely removed (Figure 3). The existing ground surface has been modelled to a
maximum height Of about 15 m, in three 5 m high blocks, at 60 ~ slope, separated by two
berms, 3.7 and 4.5 m wide (Figure 4). After excavation and before reconstructing the
slope, the exposed soil was protected from rainfall with a light polyethylene membrane.
A drainage layer made up of about 0.5 m gravel, separated from the existing
surface by a 800 g/m z nonwoven geotextile, has been placed beneath the reinforced
blocks. This drainage layer was connected to a water removal system consisting in steel
pipes embedded in a draining trench (as shown in Figure 4).
FIG.3 - The landslide body is removed, and the ground surface modelled.
Geocells and geomats - -

Comeacted soil
-~,<:
Geogrid reinforcement
- - Fill soil
- - Drainage system
o
tt~ 0 2m
t.~
0)
LEGEND
:! 9 Strain gage
m Total cell pressure
J Inclinometer
Remodeled ground surface
i Pore pressure cell
[~ Acquisition system
FIG.4 - Cross section of the instrumented geogrid reinforced slope.
The steep reinforced slope has been designed according to the Jewell method;
(Jewell, 1991) from calculations, Tenax TT060 SAMP HDPE geogrids, having 60 kN/m
ultimate (peak) tensile strength and 25 kN/m Long Term Design Strength have been used.
(ICITE N. 508/98, 1998) The main geogrid properties are reported in Table 2.
A global stability analysis on the reinforced section (with the geotechnical data
previously reported), using the STABGM code, has been performed. This software,
based on Bishop's modified method (Duncan et al. 1985), determines the circular
surfaces, passing through a specified point or tangent to horizontal surfaces, whose
Factor of Safety (FS) is minimum.
The program takes into account the presence of reinforcement layers in the
determination of the resisting moment, by multiplying the tensile strengths of the
reinforcements by their arms. The complexity of the geometry verifies not only the
circular surfaces passing through the toe of the first step, but also through the toe of the
upper steps and through other points along the face of the first step of the slope.
The same analysis was performed with surfaces tangent to deeper horizontal planes.
TABLE 2 - Main properties of the geogrids used in the project.
Test
Properties Value Unit
method
Polymer type HDPE
Extruded
Structure
i mono-oriented
ISO
Unit weight 400
g/~ 9684
Peak emile strength 60.0 kN/m GRI-GG1
Strain at peak 13.0 % GRI-GG1
Strength @ 2% strata 17.0 kN/m GRI-GG1
Strength @ 5% strak~ 32.0 kN/m GRI-GG1
Long TermDesign ASTM

25.0 kN/m
Strength D5262
The results of the analysis are presented in Figure 5: the failure mechanism with the
lowest Factor of Safety seems to encompass several failure surfaces; in fact, low Factors
of Safety are found both for the circular surface tangent to a horizontal layer (FS=1.289),
and for circular surfaces passing through the toe (FS=1.264) and through the external end
of the third geogrid (FS=1.355). These considerations are supported by the
instrumentation results, as shown later.
FS= 1.28.c
FS=1.264- J
FS=1.355
\
t ,.,
FIG. 5 - Global stability analysis
The steep reinforced slope has been built using the Tenax RIVEL System (Rimoldi
and Jaecklin, 1996) (Figure 6). It consists in the use of sacrificial steel mesh formworks
that help in the construction of the face slope and obtaining an uniform geometry of the
slope; moreover the time necessary for the construction and hence the costs of it are quite
low. It is possible to create steep slopes, completely vegetated thanks to the use of
biomats, which provided a very good medium for preventing the washout of the soil and
for the support of the growing plants. The vegetation of the face was further enhanced by
hydroseeding the face at the end of the construction of the reinforced soil structure.
FIG. 6 - Tenax R1VEL System
To build the reinforced slope it was necessary first of all to prepare the geogrid
layers cut at the required length, and to bend a steel mesh sheet at the required angle.
These sacrificial "formworks" were lined at the face of the slope, overlapped for about
50 nun and jointed by steel wires.
The reinforcing geogrids were placed and anchored by means of "U" shaped staples
with an overlapping of two ribs. A biomat was placed on the internal side of the geogrid,
in the wrapping-around of the slope face. Hooked bars were fixed to the formworks (one
every 500 mm) to avoid outward bending during soil compaction (Fig.7).
Following the common practice for geogrid reinforced slopes, the soil was laid and
spread on the geogrid in 300 mm lifts and compacted by using a hand roller near the face
of the slope (within 1 m), while the rest of the soil (more than 1 m away from the face)
was compacted with a heavy roller compactor (Fig.8). Finally the geogrid was wrapped
around the face, stretched and fixed with "U" shaped steel bars.
When the work was completed, the face was hydroseeded to encourage fast
vegetation growth.
The Tenax RIVEL System (Fig.9) allowed more than 50 m 2 of face area per
working day to be built, with a typical crew of 4 people equipped with 1 excavator, 1
roller and 1 hand compactor.
FIG. 7 - Steel meshformworks, FIG. 8 - Compaction with

hooked bars and biomats. a heavy duty roller.
FIG. 9 - View of the reinforced slope before hydroseeding.
Instrumentation
Due to the importance of the project, the slope has been instrumented, in order to verify
the long term behaviour of the structure. In particular, strain gages and total pressure
cells data are hereby reported.
Strain gages
The dimension of the geogrid ribs required the use of strain gages with very small dimensions
(3.18 mm x 2.54 mm). Figure 10 shows a typical Tenax TI" 060 SAMP geogrid rib with a
glued strain gage. Similar specimens have been used for calibration of the stress strain
curve in the laboratory by either static or dynamic tensile testing.
The strain gages were completely encapsulated in polyamide resin for protecting them
and to allow for their use in soils. The strain gages are self-temperature compensated within
a range of -20~ up to 60~ and with a strain limit up to 5% of the gage length.
The installation of this kind of gages on the geogrid ribs was particularly difficult
due to their dimension and their fragility. The strain gages have been electrically
connected with a 3 wires quarter bridge system to an automated data acquisition system.
FIG. 10 - Tenax I T 060 SAMP geogrid rib with a glued strain gage.
First of all, the geogrid to be instrumented was prepared in the laboratory by

making it perfectly flat; it was cleaned with alcohol, then the gage area was abraded with
180 and 400 silicon-carbide paper to increase the roughness. A light flame was quickly
passed over the rib to eliminate any residual powder from the abrasion and to increase
the number of radicals. A neutraliser was applied to the gage area, keeping the surface
wet by scrubbing it with a cotton tipped applicator not to allow evaporation. Immediately
after applying the neutraliser, the residual was removed with a gauze sponge. Then, the
gages were glued using a cyanoacrylate adhesive to the fibs, and cured in an oven at 40 ~
for at least 8 hours. The strain gages have been coated with a layer of silicon rubber to
protect and waterproof them.
In particular, it was very difficult to apply a thin and uniform adhesive layer (due to
the non planarity of the surface); then, taking into account the fact that the strain gages
had to be glued on a 10 m long geogrid layer, it was difficult to handle it in all phases
(from preparation to transportation on site and installation) with the necessary care. The
number of "dead" strain gages (9 out of 37) can therefore be considered a success.
The instrumented geogrids were (from the bottom) the first, third and fifth layers of
the first step, with gages at 0.5, 1.0, 1.5, 2.0, 2.5, 3.0 and 3.5 m from the face; the 7th
geogrid with gages at 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.5, 5.5 and 6.5 m from the face; the
first and fifth geogrids of the second step, with gages at 0.5, 1.0 and 1.5 m from face
(Fig.ll).
The geogrid layers have been installed in tension and the strain gages and cables covered
with 150 mm of fine sand.
fa
//
"Vt:
i ..""
. + 9
FIG. 11 - Gages outputs, 90 days after the installation.
The outputs of the strain gages are shown in the same Figure 11, 90 days after the
installation.
From the analysis of these preliminary data, it is possible to see the presence of
strain peaks in the strain envelopes, corresponding to a stress concentration in the
geogrids. Analysing the 7th geogrid, it is possible to see an important peak (about 1.7%
strain) at 2.0 m distance from the face which, together with the peak at the face point of
the third geogrid, seems to be justified by the circular surface found during the global
stability analysis and passing through the toe of the third geogrid layer (see Figure 5). The
second peak of the 7th geogrid, at 4.5 m from the face, together with the peak of the first
geogrid (at 3.5 m from the face) could belong to the deep seated surface (interesting also
the upper step), which provides the lowest FS in Figure 5. The loss of 9 strain gages
unfortunately does not allow a more accurate interpretation, but the result of this
instrumentation is considered much more than satisfying. Some of the relatively high
strains recorded near the face may be a result of the contractor using the heavy duty roller
too close to the slope face and the consequent face bulging and geogrid tension.
Total pressure cells
Three total pressure cells where placed at the bottom of the first reinforced step,
respectively at 1.0, 2.0 and 3.0 m from the face (Fig.4), in order to verify the actual
distribution of the vertical load on the base. The pressure cells made available were the
vibrating wire type with 300 mm diameter, with a standard working range from 0 to 2070 kPa.
Installation of the cells consist simply in laying down horizontally the cell, and
covering it with a thin sand layer before spreading the fill soil.
100
~ 80 eeoo..o..e'l ,eeoo-''''o'oeoea'~ "''eooo~176176176
I---i
= 60
Q.
4t1
- - 1.00 m from toe
20 9-- 2.00 m from toe
-3.00 m from toe
0
500 600 700 800 900 1000
time [hours]
FIG. 12 - Total pressure cells outputs (after end o f construction).
MONTANELLIETAL.ONGEOGRIDREINFORCEDSLOPE 27
The total pressure cells output signal, after an initial phase with a lot of discontinuity due
to the construction of the whole slope, stabilised itself as shown in Figure 12.
In Figure 13 the base pressure envelope has been compared with the "static"
pressure due the weight of the embankment. The larger output of the cell 1, closest to the
face, compared to the static pressure, seems to be due to the presence, very close to the
face itself, of the road embankment (constructed while the reinforced slope was built) and
to the eccentricity of the vertical load on the base, due to the thrust of the backfill.
The values of the inner cell C3, instead, are not so far from the expected static pressure.
surcharge ~ 1 2 3
FIG.13 - Envelope of pressure at the base, from pressure cells readings, after 500 hours
In situ pull-out tests
In situ pull-out tests have been performed on sacrificial geogrid layers, about 500 mm
wide, with different anchorage lengths and positions along the slope. In particular, three
tests, with an anchorage length of 1000, 1500 and 2000 mm have been performed. About
1000 rnm length of geogrid has been left outside the slope, in order to allow for the
connection of the geogrid with the tensile apparatus.
The installation procedure for the slope area interested by these tests has been
modified as follows:
28 FIELDINSTRUMENTATIONFORSOILAND ROCK
1. lay down the steel formwork, the geogrid and the biomat for the layer inside
which the geogrid specimens for pull-out test will be installed;
2. cut a rectangular slot, 550 mm wide x 50 mm high, in the steel mesh at half
height of the steel formwork itself. Then both the geogrid and the biomat at the
face have been cut;
3. lay down and compact a 300 mm lift of soil;
4. lay down, in correspondence to the previously prepared slot, the required
sacrificial geogrid layer. As said before, a 1000 mrn length of geogrid has been
left outside the steel formwork; this geogrid specimen was placed exactly in the
middle of the slot, in order to avoid any friction between the geogrid and the
steel mesh during the pull-out tests;
5. lay down and compact the soil on the special geogrid specimen;
6. wrap around the main geogrid layer;
7. continue the construction as usual.
A reaction structure, built with timber boards and steel pipes, was placed against
the slope face in order to avoid any damage to the whole steep slopes during pull-out
tests (see Figure 14).
1000 mm steel pipe e = 42 mm
actionC~ - ; ~ ~
pull-out geogrid
width 500 mm /./~ pull-out geogrid
main geogrid
timber board
FIG. 14- Set-up for in situ pull-out tests.
The operation was repeated in three different locations (one for each pull-out tests),
keeping the geogrid specimens always at the same depth from the top of the berm. It is
important to notice that these pull-out tests will not affect the behaviour of the whole
slope; in fact, the geogrid specimens were not calculated in the internal stability analysis
(so, no loss of performance is suffered due to the extraction of these geogrids).
Furthermore, the dimension of the slot allowed only a very small loss of soil during the
pull-out tests.
Finally, the total pull-out was limited to about 560 mm, hence the disturbance to the
soil in the slope was minimal.
In fact no further movements occurred in the adjacent soil or at the face.
Figure 15 provides an overall front view of the geogrids during the pull-out tests.
Tests were performed using a hand tensioning apparatus, connected to a special
clamp attached to the geogrid. A load cell was inserted between the clamp and the
tensioning apparatus. Reaction was made with the backhoe of the excavator. The pull-out
rate was about 10 mrn/min. Load (measured at the load cell) and position of the clamp
were measured every 15 seconds.
The results were converted in kN/m by dividing the value of the measured load by
the number of geogrid ribs tested and multiplying by the number of ribs in one metre
width. The use of the hand tensioning apparatus allowed a nearly constant test speed.
From the analysis of results, it appears that the anchorage length of 1.5 m and 2.0 m
were far larger than required to allow the pull-out of the geogrid sample. In both cases, in
fact, the failure was due to tensile failure of the geogrid, and not to pull-out. In other
words, the minimum required anchorage length was lower than 1.5 m. The results for the
three tests are presented in Figure 16.
geogrid width
500 mm,
L 1000, 1500
e 2000 mm
I mesh formwork
;ogrid
FIG. 15 - Overall front view for the pull-out tests.
FIG. 16 - Results of the pull-out tests.
The two upper curves can be considered as a sort of wide width confined tensile tests,
and not pull-out tests. The failure, in fact, was due to tensile failure of the geogrid, and
not to pull-out. The failure load was very close to the peak tensile strength of the geogrid
used (60 kN/m); the difference observed (a few kN/m) is due to the different test
conditions in respect to laboratory wide width tests, both in term of speed (not strictly
constant) and of application of the force (not perfectly distributed).
The only valid curve for the pull-out evaluation is the lower one in Figure 16.
From this curve, considering the vertical load transmitted by the total weight of the soil
on the geogrid specimen, it is possible to find a pull-out coefficient for Tenax TT
geogrids with this soil, as follows.
The resistant shear stress to pull-out of the geogrid is given by
zb=o~' fpo tan 9' (1)

where
fo = factor of pull-out;
G'.= normal stress = 21.16 kPa;
9' = soil friction angle = 26 ~
The maximum pull-out resistant force that the reinforcement can develop is given by
Tb= 2 L ~ b (2)
where
L = length of the reinforcement in the anchorage zone.
MONTANELLI ET AL. ON GEOGRID REINFORCEDSLOPE 31
Since it is
L = 1.00 m,
T b = 24.53 kN/m
we have
xb = TJ2 L = 12.26 kPa
and, finally,
fo = xb/a n' tan t)' = 1.16
This value seems to be in good agreement with the results obtained in laboratory
pull-out tests on the same kind of soil and geogrid (Cancelli et al., 1992). The value
obtained in situ is higher in respect to the laboratory one (1.00); this is probably due to
soil arching effect caused by the boundary conditions provided by the reaction structure
during the test.
Finite element analysis
To evaluate the field stress and deformation behavior of the Montone reinforced
slope, a finite element model of the entire structure has been developed (Ghinelli and
Sacchetti, 1998). The numerical simulation has been performed using the CRISP90
computer code (Britto et AI., 1990).
The model uses the following elements:
9 "BAR" elements, with linear-elastic or linear elasto-plastic behavior, for
reinforcements.
9 "LSQ" (Linear Strain Quadrilateral) and "LTS" (Linear Strain Triangle) elements, with
linear elastic-plastic behavior, for fill soil.
9 "SLIP" elements, with behavior based on the Goodman & Taylor (1968), for soil
reinforcement interface.
9 "LSQ" and "LST" elements, with linear-elastic behavior, for foundation soil.
Material properties
Field tests on the reinforced slope and laboratory tests on materials taken from t h e
reinforced block were carried out for assessing the most important mechanical
parameters of the numerical model.
The tests undertaken are:
9 Field density tests, carried out on all three reinforced blocks, that supplied the unit
weight of the fill soil.
9 Field pull-out tests, previously explained.
9 Rigid plate load tests, carried out on the first reinforced block using circular plates with
300 mm and 600 mm diameter, that supplied the principal deformation parameters of the
fill soil.
9 Classification tests, performed on several samples of fill soil taken from the reinforced
slope, that supplied the limit indexes and the particle size distributions.
9 Direct shear tests with 60x60 mm standard box, performed on samples of fill taken
from the reinforced slope, that provided the shear resistance of fill with particle size finer
than ASTM number 10 sieve.
The most important test results achieved are summarized in Table 2.
TABLE 2 - Geotechnical properties offill soil and mechanical properties of

reinforcement geogrids adopted for finite element analysis.
U~ [ ~s~V~e
FILL SOIL
Unit weight [kN/n-?] 18.4
Friction angle [o] 33
Effective cohesion [kPa] 0

Initial elastic modulus [MPa] 35.2
: Poisson ratio [-] 0.2
FOUNDATION SOIL
Unit weight [kN/m3] 20.0
Friction angle [o] 23
Effective cohesion ,[kPa] 100
It
EJastic modulus [MPa] 10
Poisson ratio [-] 0.4

INTERFACE
PuU-out coefficient I [-] I 1.00
Figure 17 shows the comparison between the distribution of reinforcement strains

obtained with the numerical analysis (light dots and dashes) and from field measurements
(dark dots) at two months after the end of construction, relating to the first reinforced
block (layers A to H in Fig.4). Figure 18 shows the same comparison relating to the
second reinforced block (layers I to R in Fig.4).
~+ : : : t ~ ........~...........,...........,......................,...........I
,.o .......... i ........... i ........... i ........... i ........... i ........... i........... t ...........
I+ ::I~ ........,...........~...........,...........+...........+...........I # ........ 'i...........
,o .......... ~........... +i........... i!........... +i........... i...............
4 ..........i...........'...........+...........++~-o++ + ...........I . O ~~ ~
o+[. :
~.......... i .......... i~::~::o::i
, '~ ....
....... i ........... :...........
ooI~-i---: ......... ~.... ! -+.~ I

2.0 ; . +
|z_."'
"~
+
~
+ ........ ,:...........
: .+...........~'...........!,...........i,...........
' +
~.o ..........
i...........
........
i...........!...........
........... + ............................................
!...........i...........i...................
I
i-p, o,,r
'.+, .........++-+:
- .:........... :~...........
+ 9 ~::.,
. ............... +++
........... ........... ,+...........
o"++:~ +:;::;.+.
........... ~ .......
o.0~o.].o
o.~
i ~,.i.,.,;i...........
.......... i ...........
~i........... ~~............................................
o....i.....o...i..o..~..,<,..~, ~ i I
0 2 4 6 0 2 4 6 $
D i s t a ~ from the dope (m) Distanoe from the dope (m)
2.0f
Lo ...................... +....................... ~....................... ! .......................

"~ ...................... i....................... !........................ !......................
o.5I.-..-..-+
...<,--i,,-.--~-,,,-.--r.......................
!.......................
o.,i..+......~
.........
++,,,,,,,.+-.......;
......................
:-.........
:.::,,~
o.o~ i , i .............. -0-'"~" o.ot,'<' i ~ .........
i ......... i~
I~~" ,::i~~ .........!~.......................

~.o ......................
..........i+.......................
i.......................
.............~'i l+ :::f~ "~ .......................
LO~-.....................
i.,.~','-I.......... +!.......................
....................+.......................
i..............................................
............~
o.~ .-.--:.<>,.,o-..i.-o-.-,,iij.-:.+,.:::~:::.--::::::::::::::-.i.:::::::::............. o.sl::+o.....~ .......... ~ ...................... u:.::,-.,,. ......... i~+, ...................

o.ol'<' i +<, l o or i i
2"0 f ' -
~.o~-.............. ~-....i .....................................................................

I .~ ....... T '~ .........0...4...+. i | ,o
o++,~
...................... ! ....................... i .......................
...........+ ......... i.......................

i .......................
i+ +.................................
o.+I.o ................
',::I~ <,.,+: ...;.........
::..........
....................+.......................+.......................
o+, i i " o.d" i i ........ i~
2.0 ~
i o 'i ....................... i........ ::::::: ..........................

j+ 'i ZZ:::ZZZ:Z:ZZ
o.y.'i
....................
:+.......................
!....................
::t'::: ............... ]| o+~:+:,+:i:,~,,~:++!::: ; ................................
o.o+ i i ""~ o of' : i ............ i-o
20 , , . 2,0 i :
'+ ............
. -+:.............................................
.+.+ ........., . ............ ................
+++ ...................
9 -~ ................. i ....................... !'"*::::: ........... '~.....................
:[.....+,: ..........i.:....................i ..............................................
o.+~:: .................... i ....................... i............ .::::,,..~ ...................... 0.5~"

.o i .....,
..................... i ...................
i ......... ~.......................
00~t i ""'""-,. o.o-~' i i "]'o
2.0 2.0
~+,,
,.o~......................
i...................... +....................... i....................... 1.0
~ o+ ~ =..it ........~...-! i ~io,
9 "';i: ....... ~ ....................... ~- "+0::':: ........... :.......................
od., i ......... ,6 ..... 0.0
I 2 3 I 2 3 4
Distance ~ m i ~ slope (m) l~lsncc fi'om tlw slope (m)
FIG. 17 - Strain distributions on the geogrids FIG. 18 - Strain distributions on the geogrids
of the first reinforced block (dark dots: field of the second reinforced block (dark dots."
data; light dots: FEM analysis). field data; light dots: FEM analysis).
With reference to numerical analysis results, all layers present more or less the
same pattern of strain distribution. In particular, the second layer of each block
(respectively B and L) has the largest strain whereas the layers closer to the top
(respectively F and P) has the smallest. The peak reinforcement strains in the bottom
layers occurred closer to the slope surface than in the top layers. The very particular strain
profiles of layers G and H in the bottom block and Q and R in the second block are
mainly related to their specific boundary conditions (stepped embankment).
From Figures 17 and 18 it is also possible to highlight a good affinity between
measured values and calculated values.
Conclusions
The following conclusions can be drawn from the field measurements and the FEM analysis:
1. In a stepped slope several potential failure mechanisms exist and therefore the
envelope of the tensile strength developed along the geogrid layers is complex;
2. The strain gages readings provide, so far, qualitative results which can be well
explained with the global stability analsyis;
3. The pull-out behaviour of the geogrids, observed in full scale in situ tests, are in
very good agreement with the results of laboratory pull-out tests for a similar type of soil;
4. For a complex geometry like a 3 steps slope the pressure at the base is the result
of a rather complex superposition of the effect of the self weight of the fill soil and the
thrust from the backfill. It is anyway the Authors' opinion that in a simpler geometrical
situation the self weight of the fill soil would prevail and the envelope of the base
pressure would be very close to the static pressure one.
5. When performing a FEM analysis the use of "SLIP" elements with frictional and
adhesive behavior, which permits relative displacements, allows to simulate the complex
interaction mechanisms between the soil fill and the reinforcement geogrids.
6. The strain distribution results of finite element analysis are in good agreement
with field measurement results. Nevertheless the boundary conditions have a remarkable
influence on the base reinforcement layer and on the geogrid layers between each
reinforced block. This influence is negligible in the other reinforcement layers.
7. To the Authors' opinion, the good agreement achieved confirm the validity of
modeling technique developed and highlights the possibility to use FEM to obtain
detailed information about stress and deformations under field condition.
Aknowledgments
The field measurements here presented were part of a project in association with the
Department of Civil Engineering of Florence University (Florence, Italy), TENAX SpA
(Vigan6, Italy) and R.P.A. SpA consultants (Perugia, Italy). The contributions of these
organizations is gratefully aknowledged.
References
Britto, A., Gunn, M., 1990, Critical State Soil Mechanics via Finite Elements, Ellis
Horwood Limited, Chichester, England.
Cancelli, A., Rimoldi, R, Togni, A., 1992, "Frictional characteristic of geogrids by means
of direct shear and pull-out tests" Proceedings of lS Kyushu '92 International Symposium
on Earth Reinforcement Practice, Fukuoka, Japan.
Coluzzi, E., Montanelli, E, Recalcati, R, Rimoldi, R, Zinesi, M., 1997, "Preliminary

results from an instrumented Geogrid Reinforced Slope for the Stabilization of the
Montone Hill in Central Italy" International Symposium on Mechanically Stabilized
Backfill (MSB), Colorado, USA.
Duncan, M., et Al., 1985, STABGM." a computer Program for slope stability analysis of
reinforced embankments and slopes. Virginia Polytechnic Institute and State University.
Ghinelli, A., and Sacchetti, M., 1998, "Finite Element Analysis of instrumented geogrid
reinforced slope" Proceedings of the 6 th International Conference on Geosynthetics,
Atlanta, USA.
ICITE, 1998, Sistema di rinforzo e stabilizzazione di terreni mediante geogriglie Tenax

TTSAMP. Certificato di idoneit?~tecnica n. 508196, ICITE - CNR, Milano, Italy.
Rimoldi, E, Jaecklin, F., 1996, "Green faced reinforced soil walls and steep slopes: the
state-of-the-art" Geosynthetics: Applications Design and Construction, De Groot, Den
Hoedt & Termaat Editors, Balkema, Rotterdam, pp.361-380.
W. Allen Marr ~
Uses of Automated Geotechnieal Instrumentation Systems
REFERENCE: Marr, W.A., "Uses of Automated Geotechnical Instrumentation

Systems," Field Instrumentation for Soil and Rock, ASTM STP 1358, G. N. Durham and
W.A. Marr, Eds., American Society for Testing and Materials, West Conshohocken, PA
1999.
ABSTRACT: Developments occurring in electronics and instrumentation promise to

lower the cost and improve the reliability of electronic systems to monitor geotechnical
instrumentation. The miniaturization of electronics, reduction of power consumption,
reduced component cost, and improved component reliability all help to make new
instruments possible and geotechnical instrumentation more cost effective.
This paper describes the application of some of these developments to three

problems and summarizes the potential benefits to the engineer from their use. The
increased capabilities of instrumentation and data acquisition equipment combined with
their improved reliability and lower cost will make future applications of geotechnical
instrumentation more cost effective.
KEYWORDS: field instrumentation, tilt meters, piezometers, automated data

acquisition
Introduction
The geotechnical practitioner faces many unknowns in design. In many cases, the
effort required to remove these unknowns during design is too costly. Construction must
proceed anyway. As a result, geotechnical engineers rely on field instrumentation to
monitor the constructed facility and forewarn them of adverse performance resulting from
these unknowns. Where required, the design or construction method is adjusted to avoid
unacceptable performance. On most major geotechnical projects, field instrumentation
constitutes an important element of the engineer's overall design. The results from field
instrumentation programs have contributed in a major way to many of the advances in our
profession.
~Chief Executive Officer, GEOCOMP Corporation, 1145 Massachusetts Ave.,

Boxborough, MA, 01719, USA, e-mail: wam@geocomp.com,
MARR ON INSTRUMENTATION SYSTEMS 37
The potential benefits of good field instrumentation programs are many and
varied. Evaluated data from field monitoring programs can reduce risks, avoid failures,
avoid litigation, reduce delays and reduce costs. While our profession learns more each
day about geotechnical behavior and pushes the envelope of knowledge further out, we
are also faced with increasingly complex site conditions, tighter restrictions, new
construction materials and processes, pressure to cut costs, and concerns about litigation.
Field instrumentation can many times help us deal with these requirements.
With all these potential benefits, field instrumentation systems are not as widely
used as one might expect. Cost has always been a hindrance. The costs to procure the
equipment, install the instrumentation, collect the data, and interpret the results can
become substantial. On the other hand the benefits cannot always be quantified. In a
strict cost-benefit picture, it becomes hard to justify many of these systems, especially the
automation component. A second hindrance to more widespread use of field
instrumentation has resulted from problems with reliability. Many clients have
experience where they spent a lot of money only to obtain confusing and contradictory
data. These clients become reluctant to spend money for field instrumentation on their
next project. Thirdly, inexperience is a major hindrance to more use of instrumentation.
I am speaking of the inexperience of some who try to do instrumentation projects without
the knowledge and experience to do the job properly, the inexperience of those procuring
instrumentation systems who do so solely on the basis of price, and the inexperience of
those who don't recognize the benefits of a well conceived and executed instrumentation
program. Some of the new technology discussed in this paper help us overcome some of
these hindrances by lowering the cost and increasing the reliability of field
instrumentation systems.
Characteristics of Typical Applications of Automated Systems
Cost considerations have previously limited the use of data acquisition systems for
field instrumentation. Typically a data acquisition system can only be justified in one or
more of the following circumstances:
large number of sensors -~ more than 50
need for frequent readings = more than once per day
need for simultaneous readings -- take all readings within seconds of each other
difficult access -- safety issues or long access time
The given numbers are rough guidelines from my experience.
Most automated field data acquisition systems being installed in the US employ
variations on two systems. The CR10 unit from Campbell Scientific is highly adaptable
to geotechnical requirements and forms the backbone for systems provided by GEO-
KON, RockTest and Slope Indicator. Geomation has released a new product that reads
up to 20 channels of voltage, current, or frequency input and provides a variety of ways to
transmit the data to a central location. These units have been adapted to readily accept
input from the commonly available geotechnical instruments.
The typical costs of an automated instrumentation system are revealing. Table 1
summarizes these in a general way for a typical installation on a per sensor basis. These
costs are just those costs for automation. The costs of support services to get the sensor
in place, such as drilling costs, are not included as these costs more-or-less apply equally
to automated and non-automated instruments.
Table 1: Cost Components for Automatin a Field Instrument

Component Cost Pe~entofTotal
Sensor $100-$600 13-20
Wire $100-$300 10-13
Data acquisition, including communications $300-$500 17-40
Install sensor, excluding drilling $50-$500 6-17
Install wire $100-$500 13-17
Install data acquisition system $50-$200 6-7
Maintain system $50-$300 6-10
TOTAL $750-2,900
The costs in Table 1 are only approximate and typical. There are plenty of exceptions.
Nevertheless in a general way, Table 1 shows typical cost components and some
conclusions are possible. The cost of components makes up 50-70% of the cost of
automating a field instrument. Installing the system makes up about 25-35% and
maintenance of the system during operation consumes 5-10% of the total cost. New
technology can potentially lower each of these cost elements by lowering the cost for the
electronic components, reducing the cost of installation and reducing the need for
maintenance by providing more reliable systems.
The total cost in Table 1 is significant. If we consider that a typical field
technician charged with manually reading sensors might cost $50 per hour, we would
have to save 15 to 60 person hours per instrument to recover the cost of automation.
Clearly a large number of readings per instrument, a large amount of time to get a reading
per instrument, or factors that preclude manual readings push the decision to automation.
Merely committing the money to automate an instrumentation system does not
insure success of the field monitoring program. I am aware of recent situations where
large sums of money were spent to automate the instrumentation in embankment test
sections placed on soft ground with drains added to accelerate consolidation. While these
projects used current products, they failed to produce the data required for a meaningful
assessment of the test sections. Problems with installation and system operation
produced too many obstacles to interpreting the data. Perhaps even more than with
conventional systems, automated field monitoring systems must follow the 25 links to
success set forth by Dunnicliff (1988) for implementing a field instrumentation system.
MARR ON INSTRUMENTATIONSYSTEMS 39
Current Developments
Recent electronics has considerably lowered the cost of data loggers, increased
their capabilities, and improved reliability. Availability of compact and low-cost
computers with high reliability to log and process data provide additional opportunities.
New types of piezoresistive pressure transducers provide highly stable readings but cost
considerably less than vibrating wire sensors. These transducers may help reduce the
sensor cost for piezometers, total stress cells and settlement monitoring gages. Strain
gage technology has improved considerably as has the data acquisition equipment to read
strain gage output. Advances in electronics for wireless communications are dropping the
cost and improving the reliability to send data through the air rather than through a wire.
This makes installation easier, troubleshooting faster, and lowers costs.
Whole new classes of low-cost sensors are being developed to measure
groundwater chemistry, flow rates, changes in position, pressures, and forces to name a
few. I fully expect the next five years to provide the geotechnical community with an
entirely new set of sensors and readout equipment that costs considerably less than
today's equipment, is easier to install and has higher reliability. We have been working
with some of these new approaches. The next three cases describe some of our
experiences.
Case 1 - Concrete F o r m Carrier for Chicago Tunnel
The 73-060-2H Contract on the North Branch Chicago River of Chicago's TARP
project consisted of 50,000 ft of machine bored tunnel lined with concrete. For
concreting, the Contractor assembled a set of five cars which carried collapsible forms.
Each cal" was 35 ft long and rode on four wheels running over a curved concrete invert.
The wheels consisted of a 17-inch diameter steel drum covered with a 2 inch thick, 10
inch wide solid polyurethane tire.
Within the first 2,000 ft of concreting, several of the tires experienced catastrophic
failure. The polyurethane tire was separating from the steel hub. Each failure completely
stopped concreting operations for several hours until the wheel could be changed. This
problem posed a major threat to the project schedule and cost.
Our evaluation of the problem indicated three unknowns. What loads were being
exerted on the tires? What stresses did those loads cause in the tires? What was the
strength of the polyurethane and the bond between the polyurethane and steel hub?
Visual observations of the cars during a move cycle indicated a widely varying load
distribution on the wheels. The start of the cycle involved collapsing the forms onto the
car. Sometimes the form would be frozen to the cured concrete. By manipulating the
hydraulic system, the operator could set the form into a dynamic oscillation that would
produce more force to jar the form loose. While moving, the car would tend to weave
from side to side over the curved invert. Since the wheels had no suspension, this would
cause a wheel to lose contact with the invert. At times during a move, a wheel was
observed to experience sudden slip sideways of as much as one inch. Substantial lateral
loads were developing that were causing unknown shear stresses in the polyurethane.
Clearly the load in each wheel was far different than the weight of the car divided by the
number of wheels.
W e decided to instrument the wheel support system to measure the force in each
wheel during a complete move cycle. Due to the dynamic nature of the move cycle, the
measuring system needed to be capable of collecting data quite rapidly. The Contractor
was anxious that this work not disrupt his normal production. We decided to use strain
gages welded onto the struts for each wheel and read them with a laptop-based data
acquisition system. The entire system would sit on the car and operate while the car was
moved from one setup to the next. One strain gage was welded to each side of each steel
strut, W e intended to use the average strain in each of the four gages on a strut to
compute axial force and the difference in strain on two opposite sides to compute a
moment from which we could determine the lateral force acting on the tire. This made a
total of 16 strain gages. In addition each wheel was supported by a hydraulic cylinder
contained within the strut that transfers the weight of the forms to the wheel. Each
cylinder was instrumented with a pressure transducer.
The sensors were connected into a signal conditioning box which provided
excitation to the sensor and gained the sensor output to the voltage level required by the
A/D. W e used a standard signal conditioning box normally used for laboratory test
equipment. The strain gages were 120 ohm 1/4" single element weldable strain gages
coupled with a microchip Wheatstone bridge from all from Micro Measurements. The
A/D was a Computer Boards PPIO-AI08 external card with a 32 channel add-on
multiplexor card. The A/D connected to the parallel printer port of a laptop. We
modified some in-house data acquisition software to read the 20 channels of input and
save them to a data file. The entire system was assembled, calibrated and debugged in
our facilities prior to shipping them to the site. All wires running from the data
acquisition system to the individual sensors were precut, fitted with connectors and
labeled prior to delivery to the site.
The complete pour cycle was taking
24 hours if everything went well. It would
take about eight hours to make the pour,
eight hours for the pour to cure, and eight
hours to move and reset the forms. W e had a
window of about 16 hours during which we
could install the instrumentation. By
preparing as much of the system in advance
as possible, our main task in the tunnel was
mounting the strain gages. This required
grinding the struts at each gage location to
fresh steel, cleaning the steel, then welding
the gage. Figure 1 shows a typical
installation. For scale the leg is FIG. 1 - Strain gages on two sides of strut.
approximately 18 inches by 10 inches (460 by 250 mm). The installation was greatly
hampered by the working conditions. Water, grease and dirt prevail in a tunnel. These
elements ruin strain gaging operations quickly.
W e spread the work over two move cycles. During one cycle, we installed the
gages and tested their output with a standard readout box. During the next cycle we
added the wires from the gages to the data acquisition system and monitored the readings
during the move. We were frequently barraged with water from every direction. This
made it very difficult to protect the wiring connections and data acquisition components.
We kept the data acquisition components boxed as long as possible, then covered them
with plastic to try to keep water away.
We decided to read each channel 10 times per second. With 20 channels of data
and a move time of up to 30 minutes, the data file would become unwieldily with more
than 300,000 individual readings. However, the move was not one continuous operation.
The operator usually halted several times to deal with unexpected problems or
difficulties. We adapted the data collection to these starts and stops. Just before the
operator would begin an activity, we would start the data logging. When move operations
stopped we would stop the data logging. In the interim we would keep a written log of
what was happening to the car so that we could later compare the data we were collecting
to the operations of the car. During the stopped periods we could check that the data
logging operation was working correctly and that the sensors were continuing to give
output.
Figure 2 gives a sample output obtained from the strain gages. It shows the axial
force in the strut determined by taking the average strain measured in the four strain gages
on that strut times the cross sectional area of steel on the strut times the modulus of
elasticity of steel. From examining the data for the entire move, we were able to
determine two important conclusions. First, whenever the operator manipulated the
200,000-
Strut 3 - Move 1
150,000-
r
.D
uZ
(D
r~ 100,000-
O
ii
50,000-
0 I I I I I I
0 100 200 300 400 5OO 600
TIME, s e c o n d s
FIG. 2 - Measured Force in Strut (1 lb = 4.45 N)
42 FI E L DINSTRUMENTATIONFOR SOIL AND ROCK
hydraulics to try to jockey a stuck form loose, he would create a large force in one or
more wheels. The hydraulic jacks were pushing the forms against the obstruction with
more force than the weight of the car. Second, on several occasions during the move, the
full weight of the car would be carried mostly by only two wheels. This happened when
the axis of the car did not align with the axis of the curved invert. Since the car had no
suspension, the entire weight would be transferred to one front tire and the rear tire on the
opposite side.
Contrary to our expectations, measured lateral forces were not substantial. While
lateral forces of several tons would develop, the side slip that we observed occurred when
the vertical weight was suddenly reduced on the tire which then allowed it to easily slip
over the invert to relieve any lateral forces.
The end result of this effort was a determination that the tires were being loaded
to as much as 150,000 lb., a value which far exceeded the apparent design load of
approximately 80,000 lb. The maximum force was developed during the jockeying of the
car to free it from an obstruction. We recommended that the hydraulic system be altered
to limit the force that could develop in the tires from the hydraulics. The measured forces
were used to assess the stresses in the polyurethane with which a revise wheel geometry
and polyurethane molding process was developed. New wheels were placed on each car.
The remaining 46,000 ft of tunnel were completed with only a few additional tire failures.
Those failures were attributed to the cars running over sharp steel objects and not to
overstressing.
This case illustrates several important points. Field measurements can be
essential to determining the cause of failure and what corrective actions to take. Fairly
elaborate measuring systems can be deployed relatively quickly and successfully operated
in severe conditions. Portable electronics now makes it possible to instrument many
construction operations without hampering the Contractor's activities. These capabilities
give us intelligent ways to help Contractors identify the cause and formulate the cure for
unexpected behavior instead of relying solely on the traditional tool box containing a
bigger hammer, a cutting torch and welding rod.
Case 2 - W D A S for tilt of b u i l d i n g facade
This opportunity arose from a need to determine whether the brick facing of
historical buildings will be deformed by underground construction activities in the
vicinity of the building. Previous monitoring methods such as crack gages and visual
inspections were not sufficiently sensitive and reliable to prevent construction related
cracks. From a safety perspective, the damage of concern is primarily architectural in
nature, i.e., minor cracks and limited differential settlement. The structural integrity of
the buildings and their foundations is not expected to be impacted by construction
activities. However cracks and differential settlement are things owners can see. They
cause high levels of concern to the people involved. Hence the need to find some means
to measure impact of construction activities on adjacent buildings before the visual
damage appears.
Other instrumentation methods such as plumb lines dropped from the roof,
inclinometer casings attached to the building face, and tiltmeters attached to the face of
each building, were considered. Each of these had serious problems, such as: having to
bother building residents to gain access to read the instrument, vandalism, overall
expense for the equipment or the time to take readings, and the limited number of
readings that could be obtained considering the necessity to arrange for readings in
advance.
Of the options available, we concluded that tilt meters connected to data loggers
using radio telemetry of the data offered a possible solution. Multiple units would be
required on each side of a building to increase the chances of measuring localized
movement of the building face. However, putting multiple units of commercially
available data loggers onto each of the large number of buildings involved would be
expensive. We located a less expensive data logger with built in wireless capability but
its resolution was only one part in 256. This was no where near the required resolution
for the project of one part in 10,000.
After considering the scope of the project, we decided to build a system to meet
the job requirements. Our goal was to have a system with the following capabilities:
-read data to at least one part in 10,000
-read temperature at the sensor location
-transmit the data to a handheld data logger in the street upon command
-operate for several months without servicing
Figure 3 shows basic components. The Sensor Connectors

design concept is an integrated CPU - A/D - RF
(central processing unit - analog to digital
converter - radio frequency transceiver ) on a 22bit 1
analog to ~ Power
single board. All electronic components are digital ]~ ~u_p~
standard, off-the-shelf items. The CPU controls
all components and can operate in a sleep mode
...... c~ IJ
to conserve power. The A/D is a 22-bit unit with cpu
an input range of five volts. That means that an
optimally configured unit can read the sensor to an Radio
Frequency ....................
accuracy of one part in 4,000,000. However it is Transceiver
almost impossible to remove sources of small Power Supply
levels of noise so we use the 22 bit A/D but ignore
the lowest 6 bits of the reading which contain the
noise. This gives us an accuracy of one part in To antennae
65,500, which more than meets the requirements FIG. 3 - WDAS components.
for the project. An electronic temperature sensor
mounted on the board provides temperature
readings directly to the digital inputs of the CPU. An off-the-shelf radio transceiver
placed directly onto the board receives incoming signals requesting readings and
transmits outbound data. To avoid having to obtain an FCC broadcast license, the
transmitting power is limited to 10 micro watts. This limits the broadcast range to a
theoretical line-of-sight range of 1,000 ft (300 m) and a practical working range of about
300 ft (100 m). The system is referred to as the W D A S unit, which stands for Wireless
Data Acquisition System.
The electronics are normally shut down to conserve power. The CPU looks at the
RF transceiver every few seconds to see if a request to read has been received. If a
request has been received, the CPU checks to see if the request is for it. If the request is
for this cell, it wakes up all components, powers
the sensor and takes a reading. The reading is
transmitted out the RF transceiver to be picked
up by the host data logger.
Figure 4 shows a typical cell. The board
is mounted inside a 4 in. by 6 in. by 3 in. (100 by
150 by 75 mm) NEMA enclosure. An Applied
Geomechanics Model 800 tilt meter sensor is
also placed inside along with batteries to supply
power. Figure 5 shows the unit installed on the
side of a brick-faced building. Figure 6 shows a
readout unit, in this case a hand-held Psion
Workabout. This ruggedized computer has a FIG. 4 - Components of WDAS.
serial port that connects to a host
communications unit. The host
communications unit contains the same
radio transceiver as those on the individual
cells. The software on the Psion sends a
request to read a particular cell ID. The
cell receives the request, wakes up and
powers the sensor, takes the reading, and
broadcasts the reading. The Psion receives
the reading and adds it to its data base.
The original concept was that a
technician would drive by the instrumented
buildings every few days and read each
cell. Figure 7 shows a sample of the type
of data received with this approach from
two units mounted on the sides of a brick
faced building. The data show
considerable variation with a range of over
300 arc seconds. Data on a similar unit FIG. 5 - WDAS installed on brick wall.
located on a concrete slab and kept at
constant temperature varied less than five
arc seconds of tilt. We suspected that the
sensors were picking up movements of
the building face caused by temperature
changes and wind blowing against the
building. These effects were probably
being amplified by the type of mounting
brackets used for the trial installation.
Cracking of panel walls shows up
FIG. 6 - Handheld readout unit for WDAS.
MARR ON INSTRUMENTATION SYSTEMS 45
at tilts of 1/300 which equates to 700 arc seconds. Standard practice to avoid cracking of
panel walls is to keep tilt below 1/500, or 400 arc seconds. Since the trial installation
showed changes of over 300 arc seconds, from environmental effects, i.e., temperature
and wind, the measurement system and approach of taking a reading every few days
would not work. It would create a situation where separating construction effects from
normal environmental effects at tilt levels below those which cause cracking would be
impossible. Construction effects should be less time dependent than temperature and
wind effects. We also expected that temperature and wind effects should be recoverable,
i.e., there should be a response signature for the building that reflects temperature and
wind effects. Any measured response outside this signature, excepting that from an
unusual condition, would more likely be related to construction. These considerations
lead us to conclude that the tilt readings had to be taken at close enough intervals to
measure the environmental response envelope for the building. Since the building goes
through a thermal cycle each day, we concluded that we needed to take readings hourly
every day. These readings should be taken for some time prior to the start of construction
in the vicinity of the building to establish the response signature for the building to
normal environmental effects.
With readings taken each hour of every day it was no longer cost effective to have
a technician collect data with a hand-held data logger. Fortunately the design of the
system permitted this new requirement to be easily met. By making use of a rebroadcast
feature of each cell, we can use the cells to transmit reading requests and data over
considerable distances. As long as each cell is within 300 ft of another cell and at least
one cell is within 300 ft of the host unit, each cell in a group can be automatically read by
the group host as frequently as we desire. Figure 8 shows a typical set of data from the
revised system. These data come from a unit rigidly mounted to the face of a brick wall
using a three point mounting system with anchor bolts inserted into the mortar. The data
clearly show a cyclic response of up to 50 arc seconds that correlates with temperature.
Analysis of the data shows another 10-20 arc seconds of variation that does not correlate
with temperature. We think it results from wind effects and extraneous vibrations of the
building. By correlating the tilt reading with temperature to obtain a temperature
correction, removing the temperature effect and averaging the resulting data for each 24
hour period, we obtain one reading per day which is called the corrected daily average.
For the case shown in Figure 8, the corrected daily average fluctuates less than 10 arc
seconds over the eight day reading period. This is a major improvement over the 300-400
arc second variation obtained on the original system and the 70 arc second range in the
raw data show in Figure 8. It is important to recall that these variations are coming from
the building itself and its response to environmental conditions, and not from the
instrumentation. In separate tests, we measured the variations due to instrumentation to
be less than 10 arc seconds.
This project illustrates several important points. First, electronics are available
today which permit us to inexpensively collect data from inaccessible locations. I expect
wireless technology to permit us to greatly expand the situations where monitoring is cost
effective. Second, we face increasingly strict limits on how construction activities can
affect adjacent facilities. Some of these restrictions approach levels of normal
environmental change. Monitoring these situations requires round-the-clock readings to
define the response signature to normal environmental events. Taking a reading every
few days does not provide sufficient data to make meaningful evaluations.
Case 3 - Groundwater monitoring
We frequently encounter situations which involve tens of observation wells and/or

piezometers to monitor effects of construction on groundwater levels. For extensive
dewatering applications, each sensor may have to be read daily. This is a potential
application for automation. However, time and again we have found that the cost of
automating the job is not competitive with dedicating a person full-time to data
collection.
Some recent developments have changed the cost equation. One development is
the wireless system described in the last case. The wireless cells remove the cost for wire
and multiplexers to get the signal or readings back to the host unit. These are significant
costs. Wire connections and multiplexers are prone to fail with time. Problem resolution
requires a site visit by a trained technician, further increasing the cost. The wires always
seem to "attract" construction equipment and become damaged. The wireless units can
be prefabricated and tested. Field installation consists of installing the sensor, mounting
the cell and plugging the two together. These tasks do not require a data acquisition
expert. A defective unit can be quickly replaced by site personnel. By using a hand-held
readout unit, each cell can be checked when it is installed. This greatly simplifies
problems and system troubleshooting.
The other development is the availability of new low cost piezoresistive pressure
sensors made for industrial applications. These sensors use silicon circuitry etched
directly onto the stainless steel sensing diaphragm. The result is a sensor with
exceptional accuracy, stability, repeatability and temperature response. Several
manufacturers are offering sensors with combined linearity and hysteresis less than 0.1%,
long term stability of span and offset over one year less than 0.1% and repeatability less
than 0.01%. These values are an error in the reading as a percent of the full scale reading.
Table 2 provides the approximate error in measured water level using various ranges of
these transducers. We have used some of these sensors in laboratory applications for
about one year and can confirm their excellent stability. We are just beginning to deploy
them in field applications.
Table 2: Measurement Error from Modem Piezoresistive Pressure Transducers

Nominal Pressure Range Range in Water Head Expected Maximum Error
psi (kPA) ft (m) ft (mm)
10 (70) 23 (7) 0.02 (7)
25 (170) 57.7 (18) 0.06 (18)
50 (350) 115 (35) 0.12 (35)
100 (700) 230 (70) 0.23 (70)
Figure 9 shows a model that has a water proof cable. It has a maximum outside
dimension of 1 inch. The wire can be fitted with a connector to plug it directly into a data
acquisition system such as the WDAS described above. This combination can be
calibrated as a complete system, then sent to the field for installation. The complete
assembly, including pressure transducer, cable and wireless cell costs less than just the
sensor for some piezometers.
A low air entry stone can be
fitted to the sensor end to conveniently
obtain a closed deaired system for long-
term monitoring of pore water pressure.
This particular model has a thin plastic
tube embedded in the wire that normally
serves to vent the back side of the
transducer diaphragm to atmospheric
pressure. Readings with the transducer
are relative to atmospheric pressure.
This tube can be used to check the
sensor. By applying gas pressure to the FIG. 9 - Piezoresistive pressure transducer.
tube, we can return the diaphragm to its zero condition, (i.e., increase the pressure until
the reading of the transducer equals the original reading at zero pressure). The required
gas pressure should equal the reading of the transducer.
It is useful at this point to review some historical information. Piezoresistive
pressure transducers were first used in geotechnical applications about 30 years ago. That
initial experience was not always successful, particularly in situations requiring
measurements over several months or years. Readings from these sensors tended to drift
with time. For field applications, voltage losses in cable extensions altered the sensor
calibration which produced errors in the measurement. Many installations experienced
serious electrical noise problems. These problems led geotechnical engineers to exploit
the benefits of vibrating wire strain gage sensors. Vibrating wire sensors transmit a
frequency that does not change with cable length. Vibrating wire sensors available in the
1970s gave much better stability than the piezoresistive sensors. With time the
geotechnical community has come to rely heavily on vibrating wire type sensors.
Essentially all of the electrical piezometers, strain gages, and load cells on the Boston
Central Artery Tunnel are vibrating wire sensors.
The problem with the piezoresistive sensors lay in their method of manufacture.
A strain gage was bonded to a thin metal diaphragm with epoxy resin. Pressure on one
side of the diaphragm caused the diaphragm to bend. The sensitive strain gage responded
to the bending with a change in resistance which could be read as a change in voltage.
The problem was that the epoxies tended to creep with time. This creep led to drift in the
output of the transducer. This problem has been solved by etching the strain gage circuits
in silicon directly onto the back of the thin metal diaphragm. There is no longer any
epoxy to creep. The new process was not only better but less costly. Manufacturing
quantities have also increased dramatically to supply sensors for the automotive and
industrial control markets. Consequently the cost of piezoresistive pressure sensors has
dropped more than 50% over the past 30 years.
I believe we will see an increase in the use of these sensors in geotechnical field
applications in the coming years. It is much easier, with off-the-shelf equipment, to
excite and read a voltage-based device than it is for a frequency-based device.

Additionally, the problem of line losses in voltage type devices can be solved by using
current loop versions of these sensors. In this case the sensor gives a current output that
is proportional to pressure. At the reading end, the current can be passed through a
resistor and the voltage drop across the resistor input to off-the-shelf readout equipment.
Electrical noise problems have been greatly reduced by using shielded cable and noise
canceling readout techniques. By using the vent tube to pressurize the back of the
diaphragm, it is possible to obtain a check on the reading of a suspect sensor. Finally,
today's piezoresistive type sensors cost less than the alternatives.
Summary and Conclusions
The cases described in this paper illustrate some of the benefits of a good
automated field instrumentation system. There is no substitute for the right data from a
field instrumentation system to give the geotechnical engineer control over field
performance. Recent developments in electronics, sensors and software provide the tools
to help us implement automated field instrumentation systems that are reliable,
responsive and affordable. However, automation systems have previously suffered from
problems of cost, reliability and ignorance.
Many new low-cost sensors are available and more are coming. Cost of sensors,
readout equipment and related materials are dropping. Off-the-shelf software, sometimes
as simple as using a spreadsheet, provides low cost but powerful ways of reducing and
reporting data to meet project needs. Off-the-shelf hardware components are making it
easier to automate data collection from field instrumentation. These components are
becoming more reliable and easier to use.
For projects with a lot of instruments and/or lot of readings, it may be cost
effective to design a project specific data collection/management system tailored to the
specific needs of the project (Hawkes and Marr, 1999). Such systems can save many
man-hours and more importantly can get accurate data to the place where it is needed
when needed. These systems involve knowledge beyond the domain of the typical
geotechnical engineer. To effectively use today's technology in automated field
instrumentation systems requires individuals with current knowledge and experience in
software, instrumentation, electronics, signal processing as well as geotechnical
engineering. A team with a weakness in any of these areas can create the opportunity for
failure.
Once an instrumentation system is automated, the cost of collecting data more
frequently is relatively low. Frequent data collection, i.e., several times a day, can help us
establish a response signature for the facility that includes the normal fluctuations from
temporal environmental effects. With this response signature, it becomes much easier to
separate the true effects of our activities from the normal response of the facility. This
more complete data set helps us avoid the difficult situation of what to do when we have
a single reading that has suddenly changed and there is no explanation for the change.
Acknowledgments
The dedicated efforts of Martin Hawkes, Salim Werden and Chafik Hankour,
engineers at GEOCOMP, helped develop the described systems and put them into
practice. Mr. Louis Brais of Perini Corporation provided the opportunity to put our
capabilities to work in the Chicago Tunnel case. Dr. Thorn Neff of Bechtel/Parsons
Brinkerhoff encouraged the development of the tilt monitoring system. The cooperation
and assistance of these gentlemen are acknowledged and appreciated.
References
Dunnicliff, J. (1988), Geotechnical Instrumentation for Monitoring Field Performance,

John Wiley & Sons, New York.
Hawkes, M. and Marr, W. A. (1999), "Data Acquisition and Management for Boston's
Central Artery/Tunnel Project," Field Instrumentation for Soil and Rock, ASTM
STP 1358, G. Durham and W.A. Marr, Eds., American Society for Testing and
Materials, 1999.
Peck, R.B. (1969), "Advantages and Limitations of the Observational Method in Applied
Soil Mechanics," Geotechnique, Vol. 19, No. 2, pp. 171-187.
Anthony D. Barley, i Michael C.R. Davies,2 and Alun M. Jones 3
Instrumentation and Long Term Monitoring of a Soil Nailed Slope at Madeira

Walk, Exmouth, UK
REFERENCE: Barley, A.D., Davies, M.C.R., and Jones, A.M., "Instrumentation and
Long Term Monitoring of a Soil Nailed Slope at Madeira Walk, Exmouth, UK,"
FieM Instrumentation for Soil and Rock, ASTM STP 1358, G.N. Durham and W.A. Marr,
ABSTRACT: This paper describes the field instrumentation of a soil nailed slope
undertaken at Madeira Walk, Exmouth, UK. The instrumentation consists of strain
gauged steel bars that are employed as soil nails to maintain stability of a slope. The
nature of the stabilisation works and the instrumentation used to monitor the axial load
induced in the soil nails are described. Initial results are presented showing the changes in
axial load induced in the nails.
KEYWORDS: field instrumentation, soil nails, slope stabilisation, strain gauges
Introduction
Exmouth is a small seaside resort located on the United Kingdom's South Devon
coastline at the mouth of the River Exe. Madeira Walk is a footpath that stretches the
length of this section of coastline, between 400 and 500 metres from the current high tide
line. The walk runs along the base of a slope known as the 'Maer'. Historically the Maer
was the cliffline above the beach. Subsequently the sea has regressed some 500m.
Director of Engineering, Keller Ground Engineering, Thorp Arch Trading Estate,

Wetherby, West Yorkshire, LS23 4SJ, United Kingdom.
2 Professor of Geotechnics, Civil Engineering Department, University of Dundee,

Dundee, DD1 4HN, United Kingdom.
3 Research Student, Division of Civil Engineering, University of Wales Cardiff, P.O. Box
917, Cardiff, CF2 1XH, United Kingdom.
Tourism is one of the main industries of Exmouth. Madeira Walk provides a focal
point for walks along the coastline and is also regularly used by locals. The safety aspects
of the cliff and slope above the path were of major concern to the local authority when
localised failures began to occur.
Background
The cliff is formed from New Red Sandstone, which has weathered heavily
leaving a residual soil mantle several metres thick. Slope angles on the weathered face
are between 45 ~ and 60 ~ standing up to 12 metres high. Bent trees point to long term
movement of the steep slope, and tension cracks are evident along the footpath running
along the crest.
Generally, movements of this weathered zone have been by creep, but in the
winter of 1994 a larger slip blocked Madeira Walk. Following this, East Devon District
Council commissioned an assessment of the stability of the slope. The result of the
ground investigation (Geotechnics Limited 1995) was a proposal to improve the stability
of the slope using soil nails.
Objectives
The objective of the programme described here was to investigate the forces
induced on the soil nails by further movements of the slope over the period of a number
of years. The project was considered suitable for monitoring since it was a relatively
small scheme where soil nailing was the main remedial works being undertaken. Access
to the site was easy, and the cooperation of the consulting engineer and the contractor
was significant if the instrumentation was to be successfully installed. In addition, with
only a few soil nailing projects currently being undertaken in the UK, this project was
within suitable distance of the research group to allow attendance during the installation
of the instrumented nails and regular monitoring.
Slope Stabilisation Design
Site Investigation
Three boreholes were drilled at the crest of the slope to determine ground
conditions. The boreholes were sunk until bedrock was reached between 4.5 and 5.5
metres below the crest. Drilling was continued 10.5 meters below ground level to confirm
the bedrock.
The borehole near the section to be instrumented indicated made ground for the
first 0.5 metres, consisting of tarmacadam over gravel sub base. This was underlain by a
1.3 metre thick layer of firm red brown clayey silt with some gravel and occasional
decomposed organic matter. A head of stiff red brown silty clay with frequent gravel
sized blocks of mudstone proceeded the rock head. The rock was encountered at a depth
of five metres. This consisted of Exmouth Mudstone, a red brown, closely fissured, very
silty weak Mudstone. No groundwater was encountered whilst cable percussive drilling.
BARLEY ET AL. ON SOIL NAILED SLOPE 53
Table 1 summarises the soil tests conducted in the field, and the laboratory tests
results from samples taken are given in Table 2. These results were used for the basis of
the geotechnical design.
TABLE 1 -- Soil properties derived from field tests.
Depth (m) Standard penetration Rock Quality Fracture

Test (blow counts) Designation, % Index
1.0 10 . . . . . .
2.4 20 . . . . . .
5.4 ... 98 6
6.5 ... 92 3
TABLE 2 -- Soil properties derived from laboratory tests.
Sample Moisture Plastic Liquid Plasticity Bulk unit Undrained

content, limit, limit, index, weight, shear
% We WL WL-Wp kN/m 3 strength,
kN/m 2
Firm, dense
weathered red
brown clayey silt 15 23 43 20 20.9 99
with occasional
clay pockets
Firm, dense friable

red brown clayey
silt, with 21 . . . . . . . . . 20.2 111
calcareous silt
pockets
Design of Slope Stabilisation Works
Stabilisation of the slope was required to prevent any further localised slippage
and the possibility of a major slide. Access to the slope was relative easy on foot, but to
mobilise plant and machinery in a built up area would be difficult. The solution proposed
was a regrading of the slope together with stabilisation using soil nails.
The environmental sensitivity of the project meant that there were two major
components to the stabilisation work, The design of a soil nailing system to prevent
54 FIELDINSTRUMENTATION FOR SOIL AND ROCK
further creep of the slope, and the design of a facing to minimise localised surface
movement between the nails and to facilitate the re-establishment of vegetation.
Nail Design
The design of the nail layout was undertaken according to the Department of
Transport design guide for the reinforcement of highway slopes by reinforced soil and
soil nailing techniques (HA68/94 1994). The nails were designed to a working load of
40kN and were inclined at an angle of 15~ to assist with the grouting process.
The final design section layout is given in Figure 1 (Rust Environmental 1996).
The nails were spaced at 1.5 metres centres and installed in a diamond grid pattern. The
nails consisted of 20mm diameter bars grouted into 100mm diameter holes (Tables 3 and
4), varying from 4 to 9.5 metres in length depending on location. Each nail was protected
from corrosion using a single 70mm diameter corrugated duct.
TABLE 3 -- Nailproperties.
Nail Bar diameter, Young's Ultimate

mm modulus, Tensile
kN/mm 2 Strength, kN
MacAlloy 500, galvanised steel 20 205 173
TABLE 4 -- Grout properties.
Grout Compressive Strength, kN/m2

14 day 28 day
Ordinary portland cement 53 - 62 67 - 87
Surface Protection
Removal of the existing vegetation was required prior to nail installation.

Therefore, a system of surface protection was designed, with resistance to localised
failure and facilitated the growth of new vegetation.
Geogrid netting (Table 5) was used to prevent localised movement between the
nails and was fixed to each nail head using a 250mm by 250mm steel plate. The growth
of the vegetation was facilitated by using a biodegradable hessian mat ('geojute'
impregnated with seed laid over the geogrid). In addition to the vegetation, trees and
shrubs were planted to improve the root matrix with the aim of reducing surface soil
erosion.
TABLE 5 - Geogridproperties.
Geogrid Thickness, Grid size, Mass per Strength,
mrn mm 2 square metre, kN/m
kg/m2
Tensar GM4 1.7 62 x'62 0.34 15
.... C o n n e c t i o n s o f Bars
FIG. 1-- Nail layout: Section view.
Instrumentation
The objective of the instrumentation was to provide a monitoring system to assess

the contribution of the nail system to the stability of the slope, within the budget
limitations. This allowed for installation, and monitoring for 18 months, of three strain
gauged soil nails.
Strain Gauged Nails
The strain gauges are installed to provide measurements of strains induced by the
soil mass loading the nail. The gauges are bonded to the prepared surface of the bar.
Protection against damage and waterproofing are made by potting with Flexonoss
polyurethane (Figure 2). The lead wires from the gauges were three strand waterproof
Teflon. This system was both effective and suitable for the harsh conditions, with
minimal change to the surface properties of the nail.
The gauges used were of 350-ohm resistance, and temperature compensated for
steel. The active ,~auge length was 3.2mm, with a range of microstrain and a cyclic
loading life of 10 cycles. Details of the gauge set-up are given in Figure 2.
Protective t a p e - - ' , ' ~- Polyurethane coat
/ t/~ ~\, ' ~ - - Strain gauge

'--- Solder tabs
Three w i r e s ~
teflon coated _ Centre line
- - Cable tie
FIG. 2 -- Gauge mounting and protection detail.
The gauges were positioned in pairs at one metre spacing along the bars as shown
in Figure 3. At each measuring point, one gauge was mounted on the top of the bar, the
other on the underside. This accounted for any strain variation between top and the
bottom of the nail to be monitored. The cables were then fed along the nail to the surface.
Three bars were instrumented: 7.5m, 5.5m and 4m length with 6, 5 and 4 pairs of gauges
respectively. To allow transport to site, the 7.5m and 5.5m nails were in two sections and
couplers were used on site.
5550 mm
I 750 1000 750 250 1000 1000 750 !
BTI BT2 ' BT3 BT4 BT5 I

4'
f +
BB1
+
BB2
'i BB3
I
+ +,
BB4
T
BB5 2oI
Nail i~f Two bars joined
head together with a coupler
FIG. 3 -- Strain gauge layout along the nails (Example: nail B).
Wiring and Bridge Completion System
The lead wire length of the nails (up to 10 metres) meant that using a standard two
wire 88 bridge arrangement would lead to temperature fluctuations having a significant
effect on the stability of the bridge measuring system. This is countered by implementing
a 'three wire' configuration. The resistance change in the wires now appears on both
arms of the bridge, and measurement is made at the gauge.
The wires are terminated at the head of each nail using a multi-way plug with
gold plated pins. The plug is then connected to a switching box that completes half of the
measuring bridge with another 350-ohm resistor, and allows up to l0 gauges to read.
DataLogger
The strain changes are measured using a HBM DMD 20A digital strain meter.
The meter completes the second half of the bridge with two 120-ohm resistors, and has a
gauge factor setting that allows the direct digital readout in microstrain. The supply
voltage to the completed bridge is one volt and has a measuring range of+19 999
microstrain.
The HBM allowed 10 strain gauges to be monitored in sequence using a
switching mechanism. Measured values are recorded manually during each visit and
input into a spreadsheet for processing.
Construction Works
The nails were delivered to site at Exmouth prior to installation for testing to
ensure all gauges were working correctly prior to installation.
Location in the Slope
The section the nails were installed in the slope was chosen for a number of
reasons:
9 the section was fairly constant in height and slope inclination

9 from the borehole logs, rock head appeared to be furthest from the surface
9 the section was adjacent to a borehole
9 good access from both the toe and the crest of the slope in this area
9 minimal trees to influence the soil nailing system.
The instrumented nails were to be in the second, third and fourth rows of the five
rows of nails installed in this section (Figure 1). Due to financial constraints, not all the
nails in the section could be instrumented. The three selected were chosen since they
were in the central body of the slope.
Slope Preparation
The slope surface before nailing began comprised of large areas of brambles and
small trees. A number of these had to be removed to provide access to the slope surface.
In areas, a certain amount of regrading was required to maintain the stability of the slope.
To allow operation of the drilling equipment to be undertaken safely, a scaffold

arrangement of temporary works was erected before nail installation could begin. The
temporary work was erected to provide a platform to give access for drilling for the top
row first. As nailing was completed, a level of scaffolding was removed and the next row
installed.
Nail Installation Procedure
Once the platform had been erected the nail positions were marked out and the
drilling rigs moved into placed. A Twin Tech 4 rig was used to drill 100mm diameter
holes, using 1.5 metre extension rods to reach the required lengths. The method of
drilling involved a short length of casing then open hole using air flush.
The tensile element of the nails were 20ram diameter MacAlloy 500 hot dipped
galvanised ribbed steel bar (ultimate tensile strength of 173kN). These were provided to
site in up to four metre lengths. Where necessary lengths were connected together using
threaded couplers.
The design of the nail included a 70mm diameter Drossbach plastic corrugated
duct, encapsulating the nail for corrosion protection. The duct was installed in the drilled
hole first, using 'Chinese lantern' centralisers to position them in the hole. A grout hose
was then coupled to the head of the duct and grout pumped down to the base until it
returned up the outside of the duct to the surface. The grout hose was then removed and
the steel bar with centralisers at 1.5m spacing pushed down the centre of the duct.
The instrumented nails were carefully installed in this way, ensuring the strain
gauges were in the correct orientation and that the lead wires were safely fed along the
nail to the surface.
As nailing proceeded, drainage pipes were also installed in the slope at ten metre
intervals between the nails and of similar length. These were intended reduce pore water
pressure within the slope.
Surface Replacement
Once a row of nails had been installed, the scaffolding platform was lowered and
the next row of nails installed. In parallel with this activity, the surface of the previous
levels was protected. The 'soft, flexible' facing applied was a geogrid. Lengths of the
geogrid were rolled down the slope and fixed to the surface at the nail heads using facing
plates bolted to the nail heads, and in between the nails using 0.5 metre long, 10mm
diameter pins.
The face plates providing connection between the facing and the nails consisted of
a 250mm by 250mm galvanised steel plate embedded on a concrete base around the nail
protruding from the slope. A nut was tightened down so the plate was firmly bedded onto
the concrete. The remaining length of nail was then cut off at the nut. A grease filled cap
was then applied to the nail head to protect it from corrosion,
Once the nailing had been completed, hessian 'geojute' was rolled down the
slope, and fixed to the geogrid. Small shrubs and trees were then planted through the
surface protection.
Monitoring and Results to Date
Once the nailing work was complete, a base reading was taken from the strain
gauges. The initial monitoring period was from August 1996 to January 1998. This
involved a number of visits to the slope, approximately every eight weeks to take a
measurement from the gauges (Jones 1998).
In conjunction with the strain measurements, a visual survey was made of the
slope during this period to observe whether any further cracks had developed, or if there
were any other signs of localised failure.
The strain gauge results presented here are for the period October 1996 to January
1998. There was initially a difficulty in acquiring a stabilised zero reading for the strain
gauges. This was diagnosed as a capacitance problem that required an earth of the system
once the nails were installed. Once this had been resolved, a base reading was taken. All
changes calculated were relative to this.
Pullout Tests
Pullout tests were undertaken during nail installation to verify the design pullout
capacity of 10kN per metre length. The tests were undertaken using a hollow ram jack
with hand pump, a pressure gauge, and a 50mm dial gauge to measure displacement. The
dial gauge was mounted remotely from the stressing frame.
Nine tests were carried out in accordance with the contract specification.
Incremental loading and unloading took place over two cycles with movement being
recorded at 1,2,3,4,5,6 and 10 minutes for each load increment/decrement. During this
10 minute period, the nails showed a small creep movement. This is shown by an
increased displacement at a constant load, Figures 4 and 5 present the loading and
45
I /A
40 A
35
9 30
~ 20 I
O ~ 7
15
10
,///
5 ;4
/
0 i | i i i i | I
0 0.5 1 1.5 2 2.5

Extension (mm)
FIG. 4 -- Force versus extension plot for pullout test PT6.
45
i
40
35
9 30
~2s E /
20 =
0
15
10
7
5 Y ./
i i | ! !
0
0 0.5 1 1.5 2 2.5
Extension ( r a m )
FIG. 5 -- Force versus extension plot for pullout test PT9.
unloading curves for pullout test nails PT6 and PT9, each installed at 3m and 9m below
the crest level respectively. The bonded length of test nail PT6 was 4.0m and was loaded
to 40kN. The bonded length of test nail PT9 was 2.5m and was loaded to 30kN.
Nail Loading
Figures 6-8 present the changes in axial force measured at each gauge point over
the initial period. This was calculated by averaging the strain measurements at each point
shown in Figure 3. This axial strain was multiplied by the modulus of elasticity of the bar
and the cross sectional area to give the resulting axial force. The corresponding axial
force profiles developed along the nails are given in Figures 9 to 11, for nails A, B and C,
respectively. The axial loads induced are quite small; hence they are currently
contributing only a small amount to the stability of the slope. The peak load in the bottom
level nail is about 6kN, but loading is typically 3kN. The loading pattern along the nail,
consistent with expectation, increases towards the middle and tails off towards the distal
end of the nail, albeit at such low loads the trend is not strongly marked.
The figures of axial force versus time for each gauge position does show a general
trend of increased loading. This appears to oscillate slightly, probably due to the varying
moisture content of the slope. This is highlighted more clearly in Figures 12-14 where
gauge point results are compared for each instrumented nail for particular distances from
the slope face. Though generally they show an increase in load, some drops in load
(shrinkage in slope) occurring in the summer are consistent.
Distance of gauge
from nail head
" * - 1.Sm
~2.5m
~3,5m
--'><--4.5 m
~5.5m
-4- 6.5m
Sep-97 Dec-97
FIG. 6 -- Nail A, Axial force at each gauge point versus time.
8
Distance of gauge
7 from nail head
6 -*-" 0 . 7 5 m
--m-- 1 . 7 5 m
5 --~- 2 . 7 5 m
g4 --~- 3 . 7 5 m
-4~4.75m
2
1
0
_~o
-2 -
FIG. 7 -- Nail B, Axial force at each gauge point versus time.
8 --~ Distance of gauge

7 -~ from nailhead
6-[ ~0.75m ~
5-~ ~t--2.75m f ~ _ ~ . . " ~ / / / /
.~o~96 Feb'-97 May-97 S e p '- 9 7 D e e '- 9 7

-2 -~ Time (months)
FIG. 8 -- Nail C, Axial force at each gauge point versus time.
8 Dates of readings
-*--12111196
7 --4t- 19/12/96
6 ~17/01/97
--><--05/03/97
5 ~12/06/97
4 ~ ~K )K ~04/08/97
o
.o
321 ~ ~ ' ~ ~ i
,r
-1 ~ 1 2 3 4 5~ ' ~ ~ 7 8
Distance from face (m) - - ' - C
-2
FIG. 9 -- Nail A, Axial force developed along the nail.
Dates of readings
--0--12111196
--w- 19/12/96
~17/01/97
--r 05/03/97
~12/06/97
3 --0--04108197
~,~,,,~ --+--18109197
2 -.-w--13/11/97
,r 1 ~20101/98
, . . . . . . . . .
0 ~ ' ~ Z '
1 2 " 3 ~ 5
-1
-2 Distance from the face (m)
FIG. 10 -- Nail B, Axial force developed along the nail.
7 Dates of readings
-r 12/11/96
---o-19/12/96
5 ~17/01/97
.-.)<--05103197
4 ~12/06/97
~04/08/97
8t. 3 -+--18/09/97
& -...-o--13/11/97
2 ---w--20/01/98
< 1
0 . , v ,
1 2 3 4
-1
Distance from the face (m)
-2
FIG. 11 -- Nail C, Axial force developed along the nail.
7 --.*--Nail A
6 -~ ~Nail B
5~ ~Nail C
.~o~9~ Feb-97 M ay-97 ~ Se;-97 Dec-97

2 -~ Time ( m o n t h s )
FIG. 12 -- Axial force in each nail at 1.75m from the slope face.
8
7
--*--Nail A
6
-'r Nail B / /
9 5
~ 4
~3
0
;-96 Feb-97 May-97 Sep-97 Dee-97
Time (months)
-2
FIG. 13 -- Axial force in each nail at 2.75m from the slope face,
7 --*--Nail A
6 ~Nail B
5 ~ ~Nail C / "~
.~ o ~ 9 6 ~ May-97 Sep-97 Dec.97

Time (months)
-2 _It.
FIG. 14 -- Axial force in each nail at 3.75m from the slope face.
Most gauges have performed well and show steady changes in results, though there
appears to be some problems with nail A (hence the absence of recent data points). The
gauges have recently been providing rather erratic results. Also, contraction and
expansion of the soil due to climatic conditions may induce the negative axial forces
measured at some points along the nails, but these loads tend to be very small.
Conclusions
Over the eighteen month period of monitoring no visually detectable movement

of the slope occurred which is an indication of the soil nails fulfilling their intended
purpose of stabilising the slope. However, slope movement is necessary to cause loading
of the soil nail. The developments of only nominal loads (3 to 6 kN) in the nails during
the monitoring period is consistent with the occurrence of only small undetectable
amounts of movement. It is noted that these changes are measured from post construction
in an attempt to distinguish the load induced by the soil mass as opposed to the effects of
the curing of the grout.
It may well be that the installation of drainage holes within the slope itself made a
major contribution in reducing loading and enhancing soil strength. However, the
monitoring period of eighteen months is a very short spell of the intended life span of the
soil nailing system. Over a longer period of say 15 to 20 years the monitoring would
detect greater nail forces and a more significant contribution.
The project has shown the successful application of strain gauges to monitor nail
loading in geotechnical works with a small budget that can provide valuable information
from long term monitoring. It is noted that success is only achieved as a result of careful
preparation, installation and protection of the instrumentation used. Hence the experience
of this project has shown, with two of the three sets of instruments on the nails still
providing data.
ACKNOWLDGEMENTS
The authors would like to express thanks to Keller Ground Engineering for
financing the research project and EPSRC for supporting the research studentship. Also
East Devon District Council, Rust Consulting and Geotechnics Ltd for the data provided
and Measurements Group UK for instrumentation of the nails.
References
Geotechnics Limited, "Madeira Walk, Exmouth Ground lnvestigation, '" Site

investigation report, Geotechnics Limited, Exeter, UK. March 1995.
Jones, A.M., "The performance of soil nailed slopes, "PhD thesis, Division of Civil
Engineering, University of Wales Cardiff, UK, 1998,
HA68/94, "Design methodsfor the reinforcement of highway slopes by reinforced soil

and soil nailing techniques, "Design manual for Roads and Bridges, Volume 4.1, HMSO,
London, 1994.
Rust Environmental, "Design of cliff stabilisation works, Madeira Walk, Exmouth,"

Design Report, Rust Environmental, Bristol, UK, February 1996.
Mark C. Webb, 1Ernest T. Selig, 2 and Timothy J. McGrath3
Instrumentation for Monitoring Large-Span Culverts
Reference: Webb, M.C., Selig, E.T., and McGrath, T.J., "Instrumentation for
Monitoring Large-Span Culverts," Field Instrumentation for Soil and Rock, ASTM STP
1358, G.N. Durham and W.A. Marr, Eds., American Society for Testing and Materials,
1999.
ABSTRACT: An extensive instrumentation plan was devised to monitor large-span

culvert behavior, soil behavior and culvert-soil interaction during backfilling and during
live load testing. The goals of the instrumentation plan were to help control construction
deformations, including insuring safe working conditions, and to gather sufficient data to
advance the state of the art in the design and construction of large-span culverts. The
instrumentation plan and sample results are discussed.
A 9.1 m (30 ft) span reinforced concrete arch and a 9.5 m (31 ft) span structural
plate corrugated steel arch large-span culvert were tested. Both had shallow cover and
live loads. Two tests were conducted, one with dense backfill and one with loose backfill.
Measurements were made of culvert shape, culvert strains, culvert-soil interface

pressures, soil density, soil stresses, soil strains, foundation movements, temperature,
cracking of the concrete culvert and shear movements between precast concrete segments.
The most unique measurement was made by a custom built laser device for measuring
movements of specific points on the culvert wall. The accuracy of this device was
verified by level sur~,ey and tape measurements.
The instnmaentation performed well and measured results were within the range
expected. Culvert deformations were efficiently measured without impeding construction
operations. Culvert-soil interaction effects during backfilling and during live load testing
were reasonably defined with the instruments selected.
KEYWORDS: instrumentation, large-span culverts, culvert-soil interaction
ISenior Engineer, Department of Civil and Environmental Engineering, University of Massachusetts,

Amherst, MA 01003.
2professor of Civil Engineering, Department of Civil and Environmental Engineering, University of
Massachusetts, Amherst, MA 01003,
3principal, Simpson Gnmpertz & Heger, Inc., 297 Broadway, Arlington, MA 02174-5310.
WEBB ET AL. ON LARGE-SPAN CULVERTS 67
INTRODUCTION
The paper describes the experience gained from instrumenting two large-span
culverts for monitoring their behavior during backfilling and during live loading. Two
purposes of the instrumentation were to insure that excessive culvert distress did not
occur and to provide safety for the workers during construction and during live load
testing. In addition, detailed measurements of soil-structure interaction were collected for
advancing the state-of-the-art in designing and constructing these large-span culverts.
Sample results from the test program are included in this paper.
TEST DESCRIPTION
Two full-scale tests were performed as part of a project called Recommended

Specifications for Large-Span Culverts, with the objective of developing reliability-
based design specifications and construction criteria, for flexible and rigid large-span
culverts. Details about the testing program and measured results are given in Webb
(1998) and Webb et al. (1998 and 1999).
The tested large-span reinforced concrete arch and large-span structural plate
metal arch culvert installed end-to-end in a pre-excavated wide trench are illustrated in
Figures 1 and 2. The depth of the trench was 3.0 m (10 ft) so that the culverts were
buried with tops about 0.6 m (2 ft) above the existing ground surface (Figure 2). Each
culvert was founded on separate reinforced concrete strip footings. The trench was
back_filled and a shallow earth embankment was constructed over both culverts in layers.
Existing site material was taken from the excavated trench and used for backfilling the
culverts and for building the embankment. The in-situ material is a well-graded sand
with gravel, per ASTM D 2487, with 1% fines. The AASHTO M 145 classification is
A-l-b, while the Unified Soil Classification is SW. Live load testing with a heavily
loaded truck was carried out at various stages of shallow cover.
The installation was done twice. In the first construction sequence, Test 1,
structural backfill was compacted and a depth of embankment over the culvert was
limited to 0.9 m (3 ft) of soil cover. After the final embankment height was reached a
live load test was conducted, which was repeated twice more while the embankment was
removed in 0.3 m (1 ft) lifts. Then the culverts were uncovered by removing the
remaining embankment and structural backfill material. In the second construction
sequence, Test 2, the structural backfill material was placed without compaction up to the
top of the arches and then compacted above the top. Live load testing was conducted as
in Test 1.
After the completion of Test 2, embankment material was added over the top of
the culverts to bring the height of cover to about 1.4 m (4.5 ft). This cover height
remained in place for one year to permit monitoring culvert performance over a period of
time with constant dead load,
The concrete arch culvert is a 9.1 m (30 ft) span by 3.5 m (11 fi 4 in.) rise (inside
dimensions) structure with a constant wall thickness of 254 mm (10 in.). The 12.8 m (42
ft) long structure is assembled from precast segments with a width of 1.82 m (5 ft 11-1/2
in.).
FIG. 1--End-to-End Installation of Concrete and Metal Culverts
FIG. 2--Cross-Sectional Hew of Culvert Installation
The metal arch culvert is a non-galvanized corrugated steel low-profile arch

structure with a 9.50 m (31 ft 2 in.) bottom span and a 3.7 m (12 ft 1 in.) total rise. The
culvert is manufactured from structural plates with 152 mm by 51 m m (6 in. x 2 in.)
corrugations. The plate thickness is 5.5 m m (0.218 in.) designated as 5 Gauge. The
assembled length was 12.2 m (40 ft) long.
The measured parameters and their purpose are summarized in Table 1. A

description of each instrument is presented in the following sections.
CULVERT DEFORMATION
Deformation Categories
Culvert deformation measurements were grouped into the following categories:

1) remote readout measurements taken during construction operations,
2) detailed measurements taken when no construction operations were active, and
3) measurements during live load testing.
WEBB ET AL. ON LARGE-SPANCULVERTS 69
TABLE 1--Purpose of Measurements
Measurement Purpose
Metal Culvert
Deformation Culvert shape and distortion
Strain Culvert deformations
Stresses, moments, and thrusts in culvert wall
Interface Pressure Culvert-soil interaction
Distribution and effects of vehicular loads
Wall Temperature Correlate temperature change with culvert behavior
Concrete Culvert
Deformation Culvert shape and distortion
Interface Pressure Culvert-soil interaction
Distribution and affects of vehicular loads
Crack Length and Width Culvert performance and distress
Relative Segment Movement Distribution and effects of vehicular loads
Wall Temperature Correlate temperature change with culvert behavior
Foundation
Settlement Culvert distortion and foundation stability
Transverse Spread Culvert distortion and foundation stability
Rotation Settlement and transverse spread
Soil
Soil Stress Culvert-soil interaction and live load modeling
Soil Strain Culvert-soil interaction
Soil Moisture Correlation with standard field practice
Soil Unit Weight Weight of soil over culvert
Soil Stiffness Settlement predictions and culvert-soil interaction
Surface Elevations Depth of cover and weight of soil over culvert
Other
Photographs Visual documentation
Live Load Magnitude Document loading
Tempera~tre and Rainfall Document environmental conditions
Category 1 - Construction Monitoring--Culvert movements during construction

operations were monitored to evaluate structural stability and warn of excessive
deformations. For this purpose it was necessary to select instruments with remote reading
capabilities for safety of workers. Conventional level survey was selected for monitoring
vertical deformations while structural extensometers (with electrical displacement
transducers) were selected for monitoring horizontal deformations.
Category 2 - Detailed Measurements--Detailed deformation measurements were

used for structural analysis and also to expand on, as well as check, readings taken during
construction operations. Electrical readings with an infrared laser device at three stations
along the length of each culvert were used for this category supplemented by tape
extensometer readings. Digital level survey and tilt sensors were used to monitor vertical
footing movements at the same stations.
Category 3 - Live Load Monitoring--Evaluating live load effects on the culverts

required monitoring of the culvert shape at multiple points quickly. The infrared laser
device was selected for this purpose and readings were taken at the three longitudinal
stations in each culvert. Level survey was used to obtain the deflection basin of the metal
culvert crown during live load testing.
Level Survey
For the digital level readings, photocopies were made of the bar code staff and
attached to the inside walls of the culverts to monitor vertical movements at various
locations. The paper photocopies were bonded to strips of plastic, about 450 mm (18 in.)
long and covered with a transparent polyurethane coating for a moisture seal. Bar codes
covered with less transparent coatings were found to be more difficult to read with the
digital level even in very bright surroundings. Since the bar codes were installed on the
insides of the culverts (in the shade) less transparent coatings would definitely not have
worked. To further expedite the reading process all bar codes were extended up or down
from the permanent point to the same level to prevent needing to reset the digital level
every time that a higher or lower bar code had to be read.
The reported reading accuracy of the digital level was 1.0 mm (0.04 in.) when a
foldable bar code staff is used. This was sufficient for monitoring the construction
deformations. The digital level was a Model DiNi 10 manufactured by Zeiss.
The digital level together with photocopied bar codes installed on the culverts
worked well to monitor culvert behavior without impeding construction operations or
risking injury to personnel. One problem that was encountered with this technique,
however, was the inability to read the bar codes during vibratory compaction when both
the culvert and the foundation soil experienced large vibrations. This problem was most
pronounced when a large ride-on vibratory roller compactor was operated close to or
immediately above the culverts. However, for smaller walk-behind vibratory plate
compactors vibrations caused less problems. This problem was further alleviated by the
fact that in most cases it was acceptable to obtain culvert deformation only during
placement and after compaction of backfill material layers.
Three reading locations were selected at each of two cross-sections along the
length of each culvert as shown in Figures 3 and 4 to monitor construction deformations.
For the metal culvert bar codes were attached at the crown and points of curvature change
where the small radius side plates meet with the large radius top plates. For the concrete
culvert bar codes were attached at the crown and shoulders. The two selected cross-
sections for metal culvert were Stations M-P1 and M-S3 (Figure 3) and Stations C-P1 and
C-$2 for the concrete culvert (Figure 4). The digital level was set up at an angle to the
centerline of the culverts to permit seeing all the bar codes along the length of both
culverts. Results of level survey measurements with this arrangement will be discussed
in a later section.
The deflection basin under the live load vehicle in the metal culvert was measured
with the digital level by holding the inverted bar code staff at 12 longitudinal crown
locations.
Structural Extensometers
The structural extensometers consisted of a displacement sensor attached to the

structure with tensioned cables using inductance coils as the displacement transducer
(Figure 5). These gages have been extensively used in the past and are described by Selig
(1975a and b), Locations of the structural extensometers are shown in Figures 3 and 4 for
the metal and concrete culverts, respectively,
Inductance Coils--Each transducer consists of a pair of wire coils wrapped around

non-conductive disks. An alternating voltage applied to one coil induces a voltage in the
second coil which is related to the distance between the coils. The readout device,
automatically balances the output and converts it to a direct current signal that could be
collected with a data acquisition system.
Gage Description,-The coils were mounted on two transfer rods which were each
connected to steel cables which were attached to eye bolts inserted in the culvert walls
(Figure 5). Therefore, the structural extensometers could be installed at any position on
the culvert and spanning any distance, Springs were used on the outside sleeve tubing of
the sensor to maintain an adequate tension in the cables. To increase the working range
of the inductance coils, steel rings were added or removed to adjust the length of the
device when the sensor approached the ends of its range. This aided the monitoring
process since, during the stage where backfill is placed at the sides of a flexible culvert,
causing the sides of the culvert to move inward, the rings could be removed without
cutting or adjusting the steel cables. Conversely, during the stage where backfill material
is placed over the top of these flexible culverts the inverse is true as the sides of the
structure move outward thus causing the sensor to extend.
FIG 3--Monitoring Deformation in Corrugated Metal Culvert
FIG 4--Monitoring Deformation in Reinforced Concrete Culvert
WEBB ET A L ON LARGE-SPAN CULVERTS 73
FIG. 5--View of Structural Extensometer
The sensitivity and range of the coils is a function of the size and number of turns
of wire. For this project 25 mm (1 in.) diameter coils were used which had a range of
approximately 75 mm (3 in.), bias of 1.3 mm (0.05 in.) and a resolution of 0.13 mm
(0.005 in.) which was controlled by the characteristics of the data acquisition system.
Readout Instrumentation--The readout unit, called the e-mu and developed at Nottingham
University, had only one channel. Thus, manual switch boxes were used to direct the
signal from the appropriate sensor into the unit. The processed signal coming out of the
readout was collected by a pc-based data acquisition system. The data acquisition
program (LabTech Notebook by Laboratory Technologies of Wilmington, MA) for these
gages had a manual trigger to allow the operator time to change the switch boxes.
Infrared Laser Device
InfraredLaser--The laser device, an AccuRange Model 4000-LIR manufactured

by Acuity Research Inc., Menlo Park, CA emits near infrared light (780 nm wavelength)
and can be used for measuring distances up to 15 m (50 ft). Measurement accuracy is
affected by three types of noise in different ways and as a function of sample rate. These
are detector thermal noise, laser diode noise and a resolution limitation imposed by the
maximum sample range. In addition, the reflectivity of the target and therefore the
amplitude of the return signal also affects measurement accuracy. Since most of these
issues could be controlled by selecting the optimum sample rate, maximum range
specification and even improve the reflectivity of targets by painting them broken white
or grey colors, the measurement accuracy could be controlled. For the un-galvanized
metal culvert (erected from black structural steel plates), white stripes were painted on the
inside surface at each measurement station to enhance the reflectivity of the culvert
surface. The reflectivity of the concrete culvert surface was sufficient and did not require
any painting. Therefore, the laser configuration as used was capable of reading distances
up to 6 m (20 ft) with a maximum bias (inaccuracy) of about 2 rmrn (0.08 in,). Since the
laser beam is near infrared, the beam could not always be visually detected to check target
alignment. Thus special filter paper had to be used to find the beam in bright light
conditions. However, in darker conditions the beam could easily be spotted with the
naked eye. Laboratory evaluation showed that the laser device has a warm-up period of
about 25 to 50 minutes. Therefore, during field testing the laser device was turned on
about 1 hour before any measurements were made and left on during testing.
Pan-Tilt Unit--The laser was mounted on a pan-tilt unit as shown in Figure 6.

The pan-tilt unit, Model PTU 46-17.5 manufactured by Directed Perception, Inc., has a
tilt range of approximately 78 ~ while it could simultaneously pan through 340 ~
Specifications of the pan-tilt unit include precise control of position, speed and
acceleration with a resolution of 3.086 arc minute (0.051428 ~ and speeds up to 300 ~ per
second. However, too fast acceleration speeds were observed to cause the pan-tilt unit to
skip steps causing misalignment of the laser beam. This was a result of a combination of
the weight of the laser and restrictions of movement by electrical cables (it was important
to ensure that no cables were caught underneath the support fixture). The pan-tilt unit
connected to a RS-232 interface on the computer.
FIG. 6--Laser Device Mounted on Pan-Tilt Unit
Mounting and Operation--The laser and pan-tilt unit were mounted on a sliding
support fixture as shown in Figure 6. In the center of each culvert a concrete wall was
constructed extending down the length of the culvert on which steel channels were
installed to guide the support fixture. The support fixture could be moved down the
length of the culverts to any longitudinal station. The base of the support fixture was
designed so that it could be accurately repositioned at the same station location and
locked in place during a laser scan.
Using the pan-tilt unit the laser could be pointed at any location on the inside
circumference of the culverts fairly quickly and very accurately. This enabled tracking
the movements (x, y and z directions) of unique targets around the culvert circumference
as the construction and testing progressed. To achieve this, software was developed to
control the pan-tilt unit while simultaneously allowing measurements to be made with the
laser. Each unit had to be controlled separately through two serial ports on a computer.
The basic procedure was to first aim the laser beam to a specific point on the structure by
advancing the pan-tilt unit through software and then recording the exact coordinates.
This procedure was repeated for each location around the culvert circumference of a
specific longitudinal station, therefore creating an input data file for that station. Next, a
second software program would read the input data file just created and instruct the pan-
tilt unit to advance to each location and allow the laser device to make measurements.
Typically, this scanning of the inside circumference was repeated twice to improve the
measurement precision which took about two minutes for 26 reading locations around the
circumference. On subsequent scans a third software program would read the input data
file created during Step 1 (or any other increment) and stop at each location allowing the
user to fine adjust the laser beam if any movements occurred. The coordinates were then
automatically recorded for use in Step 2. Finally, the measured results were written to
hard disk and processed for almost immediate graphical display.
Measurement Locations--A total of 26 locations around the circumference of the

metal culvert and 23 locations for the concrete culvert were selected for the detailed laser
measurements at each of 3 monitoring stations. All measurements for the metal culvert
were made on the inside crest of the corrugations. Example results of laser measurements
are discussed in a following section.
Tape Extensometer
Triangulation with distance measurement using a tape extensometer provided a

check on the laser measurements. This procedure is highly reliable and accurate, but is
time consuming and thus was used only a few times during each test (including the
beginning and end of each test) at a relatively few locations. A total of 5 points (Figures
3 and 4) on the culvert wall at each cross-section were measured relative to a fixed
reference. Since footings may move and rotate under the applied loads, instrumentation
reference walls were constructed as the fixed reference (Figures 3 and 4). Eye bolts
installed on the culvert walls and on the reference walls were used to connect the tape.
Measurements were recorded as a combination of a tape measure reading in feet

and inches to the nearest 2 inches (25 ram) and a dial gage reading from 0 to 2 inches in
0.001 inch (0.025 mm) increments. All measurements were corrected to account for
temperature changes in the steel tape length. For the distances measured the precision
was about 0.13 mm (0.005 in.). Distances of up to about 60 m (200 ft) can be measured.
METAL CULVERT STRAIN
Metal culvert wall strains were measured with weldable foil electrical resistance
strain gages (Measurements Group, Inc. Type CEA-06-W250A-350). Weldable gages
were selected because of their suitability for field installation and long-term stability.
Gages with pre-attached leads were not used because of their size and difficulty in
waterproofing the Teflon coated leads. However, thin jumper wires were pre-soldered to
the gages in the laboratory to speed up actual field installation. Weldable gages are made
up of polyimide-encapsulated foil grids pre-bonded in the laboratory to thin (0.13 mm)
metal carriers with a high-performance adhesive. However, the effective gage factor of
the foil gage on its metal carrier is less than that of the basic strain gage because of the
stiffness of the metal shim. Therefore, a steel test beam was instrumented with both
conventional and weldable strain gages to obtain the correct properties of the weldable
gages.
These encapsulated strain gages had large rugged soldering tabs for easy
soldering. Furthermore, the gages could be contoured to culvert corrugation radii as
small as 13 mm (0.5 in.). The gages are fully encapsulated for optimum environmental
protection however, additional waterproofing was added. The strain gages are self-
temperature-compensated for steel, with a resistance of 350 ohms. The gages were wired
into a full Wheatstone bridge circuit using the three leadwire system and bridge
completion modules supplied by Campbell Scientific, Inc.
Installation
The gages were mounted to the metal surface by spotwelding. Since a reasonably
clean surface is required for spotwelding, the un-galvanized and slightly rusted metal
surface was de-greased and then cleaned with a small grinder and a cloth. Then grit and
grease were removed with a lint-free cloth soaked in a solvent. Capacitor-discharge
resistance welding is controlled by the welding time and the probe contact force.
Therefore, trials were made with sample gages first before attempting to install actual
gages. Furthermore, skill is required to master the spotwelding pattern recommended by
the gage manufacturer.
Gage Locations
At each of the 25 locations on the inside wall of the metal culvert (Figure 7), four
strain gages were installed (circumferential and longitudinal directions on inside crest and
valley locations) in order to compute wall thrust and bending moment. Unfortunately, the
technique used to compute hoop strain (and thus wall thrust) from the measured outer
fiber strains (assuming a linear strain profile across the wall section) produced erroneous
results (Webb et al. 1998). A better technique would have been to measure strain at the
centroidal axis of the wall section directly without the need to infer it from the outer fiber
strain measurements. Additional strain gages were subsequently added to the centroid of
the corrugation profile which produced much better results.
FIG. 7--Strain Gagesfor Corrugated Metal Culvert
INTERFACE PRESSURES
Backfill soil pressures acting on the culvert as a result of culvert-soil interaction

were measured with vibrating wire earth pressure cells at locations shown in Figures 8
and 9. A total of 16 earth pressure cells were installed around each culvert. Because of
the difficulty in measuring soil stress accurately, cells were installed at duplicate
locations.
Gage Description
The pressure cells were 230 mm (9 in.) diameter, fluid filled cells with vibrating
wire pressure transducers, manufactured by Geokon, Inc. Each cell had two pressure
sensitive faces and also had a built in temperature sensor. The cells were installed
radially around each culvert wall 150 to 300 mm (6 to 12 in.) away from the wall.
Pressure cells with ranges of 345 kPa (50 psi) and 690 kPa (100 psi), over-range capacity
of 150 % full scale, resolution of0.1% full scale and reported transducer accuracy to
within 0.25 % of full scale were selected. However, accuracy of the earth pressure
measurement with the pressure cell is considerably less because of the effects of
installation and conformance of the installed cell. The 690 kPa (100 psi) pressure cells
were installed above the top of the culverts for measurement of live load pressures while
the 345 kPa (50 psi) pressure cells were used at the remaining locations around the
culvert wall where lower pressures were expected. In addition, similar cells with a range
of 70 kPa (10 psi) were used in regions around the culverts where the lowest pressures
were expected. Some of the latter cells only had one pressure sensitive face.
FIG. 8--Interface Pressure Cellsfor Corrugated Metal Culvert
Gage Calibration
The earth pressure ceils are all supplied with factory calibration sheets giving
pressure and temperature coefficients; however, these coefficients are determined for the
vibrating wire transducer itself, and not the complete soil stress cell. Experience with the
230 mm (9 in.) diameter gages shows that the calibration of the complete cell is similar to
that of the transducer alone (McGrath and Selig, 1996). This was confirmed by an
extensive calibration program.
W A L L TEMPERATURE
Temperature of the metal and concrete culvert inside walls were measured using
OMEGA Type T copper-constantan thermocouples. Eleven thermocouples were installed
on the two culverts at the crown and springline locations. Some of the thermocouples
were read automatically by data logger while the rest were read manually.
FIG. 9--Interface Pressure Cells for Reinforced Concrete Culvert
CONCRETE CULVERT INSTRUMENTATION
Crack Length and Width
Cracks in the concrete culvert were monitored visually and crack length and width
dimensions were determined manually using a magnifying crack gage to evaluate culvert
performance and distress. Cracks were marked by chalk for identification. Most of the
cracks noted in the precast concrete segments occurred during formwork removal,
transport and/or installation at the test site.
Relative Segment Movement
Vibrating wire crack gages manufactured by Geokon, Inc. were attached to the
concrete culvert wall to determine relative joint movement. Nine gages were installed on
the concrete culvert between segments at the shoulder and crown locations to measure
relative shear movement. Three gages were installed one each at the crown and both
shoulders at the joint between Segments 2-3 (see Figure 4 for segment numbering). Two
gages were installed, one each at the South shoulder and crown between Segments 3-4, 4-
5, and 5-6. These gages worked well even though very small shear movements (typically
less than 2 ram) occurred during testing.
FOOTING MOVEMENT
Settlement, Rotation and Spread
Level survey using a digital level with photocopied bar codes (as discussed in a
previous section) were used to measure settlement of the footings under both culverts.
Bar code targets were installed near the outer edge of the footing to permit extending
them to the height required by the bar code targets hanging from the culvert. Footing
settlements were very small (about 4.5 mm maximum) and fairly tmiform. Footing
rotation was needed to compute settlement of footings at the base of the culvert leg and
was measured using a portable digital level device. The device was able to read rotations
to the nearest 0.1 o. No rotations of the footings occurred. Footing spread was measured
with a tape extensometer which indicated less than about 2 mm (0.08 in.) of outward
movement.
SOIL MEASUREMENTS
Soil Extensometers
Six soil extensometers were installed in the backfill next to the metal culvert to
measure displacement of the structural backfill relative to the culvert wall during
backfilling and live load testing. Four of these gages were installed at the springline
elevation on both sides of the metal culvert while the other two were installed at the
change of curvature on the one side only (Figure 7).
Gage Description-Inductance coils, similar to those used for the structural

extensometers, were used for soil extensometers. The coils were mounted on two transfer
rods which were each connected to end plate anchors that moved with the soil within
which they were embedded causing the coil spacing to change. The portion of the
transfer rod near the coil was non-conductive to minimize interference with the
inductance field. A schematic drawing o f the soil extensometer is shown in Figure 10.
The overall length of the installed gage was 1.8 m (6 ft). The guide tubes served two
functions. One function was to keep soil away from the transfer rods to minimize any
friction that could affect the measurement. The second purpose was to provide a tube for
the guide blocks that keep the inductance coils in alignment. The guide tubes were non-
conductive because of their proximity to the inductance coils.
Gage Readout Instrumentation--The soil extensometers were read the same way
as the structural extensometers.
Other Soil Measurements
In-place soil density and soil moisture content measurements were taken with a
direct aansmission nuclear density gage (Troxler Model 3411B ). A soil compaction
modulus gage Model H-4140 manufactured by Humboldt Manufacturing Company was

used to measure backfill stiffness for analysis of culvert-soil interaction. Conventional
level survey was used to measure backfill layer surface elevations and wheel rutting of
the live load vehicle.
FIG. 1O--Schematic of Soil Extensometer
DATA ACQUISITION SYSTEMS
Due to the variety and number of sensors three different data acquisition systems
were used to acquire and process data. The structural and soil extensometers were read
with a pc-based data acquisition system. The vibrating wire earth pressure cells and
thermocouples were read with a Campbell Scientific, Inc. CR10X data logger, while the
strain gages were read with a Campbell Scientific, Inc. CR7 data logger. In order to
support the large number of sensors each of the Campbell Scientific data loggers was
expanded with three AM416 Campbell Scientific, Inc. multiplexers.
SAMPLE DATA
Sample data are included for illustration purposes. Detailed results can be found
in Webb et al. (1998).
Culvert Deformation
Vertical deformations of the metal culvert during backfilling operations, measured

with the laser device and with the digital level at the crown and curvature locations, are
shown in Figure 11 for Test 2 (loosely placed backfill). The crown continued to rise until
earth was placed over the top of the structure at which point the movement was reversed.
The relative rate at which the crown rose was about the same as that during downward
movement of the crown because of the loose state in which the soil was placed.
Maximum measured peaking during Test 2 was about 72 mm (2.8 in.). The South
curvature points showed more downward movement than the North points. This trend
was also observed by more flattening o f the side plates on the South side than on the
North side. The metal culvert was top loaded in both tests to prevent excessive peaking
during backfilling operations. Concrete blocks were added to the metal culvert with
about 2.4 m (8 fi) of backfill in place. The effect of top loading the metal structure can be
seen in Figure 11.
The deformed metal culvert shapes due to backfilling operations are shown in
Figure 12 for Test 2 as measured with the laser device. This figure shows the maximum
peaking of the structure which occurred at 0.3 m (1 ft) of soil cover and the final shape
after backfilling. Culvert displacements have been magnified 7.5 times. Flattening of the
South plates can be seen. The deformed metal culvert shapes for Test 1 were similar with
slightly more peaking (maximum peaking of 80 mm or 3.1 in.) and slightly less flattening
of the South plates.
Backfill Depth, ft
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
' I ' I ' I ' I ~ I ' I ' I ' I ' I J I ' I ' ' I ' I ' I ' I ' I
80
Test 2 - Uncompacted Backfill .~ Too of - 3.0
70 Top loading ( ~ ' ~ N structure
60 / added ,~ - 2.5
50 2.0 .~
~r 40 1.5
20 f 1.0 "~
0.5 ~
k
9 0 0.0 ~
~ -10 aftei- uiir Test 1 \ N
-0.5
-20 >r
: Notes : - Solidlinesare laser measurements ~
- Dashedlines are digital level measuren-ents
D~o:,._ a~r^--.,.:^- :o pm,~.g -1.0
-30
- All measurementshavebeen correctedfor foot
--40 ' ' ' ' ' . . . .
, i ,
' . . . . ' . . . . ' . . . . , . . . , i , , , , ,
. . . . . . . . - -1.5
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Backfill Depth, m
FIG. 11--Structural Deformations o f Metal Culvert During Backfilling
Maximum peaking (Cover = 0.3 m) ~ - r - " ~ .

/,,,~,'~-. _ ~ Final cover = 0.9 m
~.~jkr ~ fl NH~"~I~ / Measurement
Initial shape / point
SS " NS
.,;o.too.;o
Displacement, mm
Test 2 - Culvert displacements magnified x 7.5
FIG. 12--Average Metal Culvert Displacements During Backfilling - Laser Device
Bending Moments During Live Load Testing
The distribution of the change in bending moment due to the tandem axles of the
live load vehicle positioned over the culvert crown is plotted in Figure 13 for Test 1.
Moment is plotted on the tension side of the structure. Three depths of soil cover and
average measurements of the two stations underneath the wheel paths are included. This
figure indicates that the live load bending moments at the shoulders at the stations under
the wheel paths are inversely related to the depth of soil cover, i.e. decreasing cover
results in increasing moment. No significant moments developed at the springline and
curvature changes from the live load vehicle. The maximum live load moments did not
develop directly underneath the tandem axles but, instead, developed at locations outside
the tandem axles (shoulder locations). Bending moments for both tests were similar
(Test 1 results are not shown here).
Radial Pressures During Long Term Monitoring
Radial pressures were measured over a 9 month period under constant earth load
and the results are plotted in Figure 14 for both culverts. The first set of long term
measurements was taken about 3 months after the additional soil cover had been placed
and indicates a drop in radial pressure for both culverts and all gages compared to the
final cover pressures, However, after the initial drop the pressures generally increased to
almost the same values measured at the final cover condition. Temperatures were
measured during the long term observation period using thermocouples installed on the
inside of the culverts and thermistors located in the earth pressure cells and are plotted in
Figure 14a. A comparison between these temperatures and corresponding radial
pressures generally suggests a lowering of radial pressures when the culverts contract
during the colder months and subsequent increase in radial pressures when the culverts
expand again during warmer periods.
CR
,-lO NH Cover = 0.3 m
SH
-7.S N ~ , , / C o v e r = 0.6 m
-s.o ~'~Cover = 0.9 m
-2.5
2.s
5.0
7.5 J ~ NC
SS= : IC
1~ - _ _
~,
/
- = - NS
Note:BendingmomentunitsarekN.m/m
FIG. 13--Test 2 Bending Moments - Live Load Vehicle Over Culvert Crown
SUMMARY AND CONCLUSIONS
The instrumentation plan for two large-span culvert tests on a rigid and a flexible
culvert has been presented. Experience gained from instrumenting these culverts to
monitor construction operations and to obtain detailed measurements for soil-structure
interaction analysis has been discussed. The successful procedures followed for
monitoring construction deformations using level survey with photocopied bar codes for
vertical deformations and inductance coil structural extensometers for horizontal
deformations have been documented. The successful use of a laser device mounted on an
automated pan-tilt unit for detailed culvert deformation measurements has been
described. Problems with the strain gage measurement technique in computing wall
thrust from outer fiber strain measurements have been described and a better
measurement technique has been given.
Measured vertical deflection results showed uplift of the metal culvert crown until
about 0.3 m (1 ft) of soil cover was in place after which movement was reversed during
placement of additional soil cover. With the tandem axles of the live load vehicle
positioned over the culvert crown measured bending moments at the shoulders are
inversely related to the depth of soil cover while moments at the springline and curvature
change are small. Long term monitoring of radial pressures around both culverts indicate
a lowering of pressures when the culverts contract during the colder months and
subsequent increase in pressures when the culverts expand again during wanner periods.
30
80 --
20' 70
E3 60 Ci
insiC~vec~ulverts /..,./'~
10 50
(thermocouples) ~-c~ ~ .o_ / / ---
g Backfillsoil(thermistors)~ 40 ~
~ o 30 E
I-- a) AverageTemperature 2o ~
- 1 0 , , I , , I , , I , , I , , I , , L , , i , , I , ,
0 30 60 90 120 150 180 210 240 270

. . , . . , 9 . , . . , . . , . . , . . , . . , . .
70 - 10
b) Concrete Culvert ~- 9
60
- s ~
~ 50 _--:7 u;
40
o. 30
~F-mal Shoulders ~ ---- 4
cover . . ^ A , ~ ' - ~ , ~ ~ 3
~ 20
Curvatures ~ 2
10 , u p
Crown 1
, I , , I i , I i , I i , I i , I i i l i i i l i 0
30 60 90 120 150 180 210 240 270
7 0 1 ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' - 10
L c) Metal Culvert / ~ ' ~ t " - ~ . . ~ v ~ . ~ _~
9
60~ Springlines " z(~--/""- i
~.
"~ 50 _-" 7 ~
9" kFinal .~ ~ - 0)
~ "covQr Curvatures =,=~ _'-

4O m
ft.
4
3~tr
2
1
0
30 60 90 120 150 180 210 240 270
Days After Backfilling
FIG. 14--Average Long Term Radial Pressures on Culverts
ACKNOWLEDGMENTS
The full-scale tests were conducted at the University of Massachusetts in

Amherst, MA. The project was funded by the American Association of State Highway
and Transportation Officials (AASHTO), in cooperation with the Federal Highway
Administration (FHWA), and conducted through the National Cooperative Highway
Research Program (NCHRP) which is administered by the Transportation Research
Board. The concrete culvert was provided by BEBO of America~ Inc. and was
manufactured by a local licensee, Rotondo Precast of Avon, CT. The metal culvert was
provided by CONTECH Construction Products, Middletown, OH. The culvert tests were
performed in a gravel pit owned by Delta Materials Corp., Sunderland, MA. The readout
device for the inductance coils was developed at Nottingham University. The soil
compaction modulus gage was provided by CNA Consulting Engineers on a trial basis,
Val Moser of Campbell Scientific, Inc. provided invaluable assistance with their data
acquisition systems. The Massachusetts Highway Department provided the Troxler
nuclear density gage. Jeanne Sussmann, Raymond Frenkel, Dan Lovett, Troy Thiele,
David Bamett, Zack Barnett, Donna Carver and David Glazier provided substantial
assistance with various aspects of the instrumentation and test programs.
REFERENCES
McGrath, T.J., and E.T. Selig., 1996, "Instrumentationfor Investigating the Behavior of
Pipe and Soil During Backfilling," Geotechnical Report No. NSF96-442P, Department of
Civil and Environmental Engineering, University of Massachusetts, Amherst, MA.
Selig, E.T., 1975a, "Soil Strain Measurement Using Inductance Coil Method,"
Performance Monitoringfor Geotechnical Construction, ASTM STP 584, American
Society for Testing and Materials, pp, 141-158.
Selig, E.T,, 1975b, "Instrumentationof Large Buried Culverts," Performance Monitoring

for Geotechnical Construction, ASTM STP 584, American Society for Testing and
Materials, pp. 159-181.
Webb, M,C., 1998, "Improved Design and Construction of Large Span Culverts,"
Geotechnical Report No. NCH98-458D, Department of Civil and Environmental
Engineering, University of Massachusetts, Amherst, MA.
Webb, M,C., Sussmann, J.A., and Selig. E.T., 1998, "Large Span Culvert Field Test
Results," Geotechnical Report No. NCH98-456F, Department of Civil and Environmental
Engineering, University of Massachusetts, Amherst, MA.
Webb, M. C., Selig. E. T., Sussmarm, J, A. and McGrath, T. J., 1999, "Field Tests of a
Large-Span Metal Culvert," Submitted to Committee A2CO6 for Consideration for
Transportation Research Board's 1999 Annual Meeting, National Research Council,
Washington, D.C.
Michael Zhiqiang yang,1 Eric C. D r t l m m , 2 Richard M. Bennett 2, and Matthew Mauldon 3
Measurement of Earth Pressures on Concrete Box Culverts under Highway

Embankments
REFERENCE: Yang, M. Z., Drumm, E. C., Bennett, R. M., and Mauldon, M.,
"Measurement of Earth Pressures on Concrete Box Culverts under Highway
Embankments," Field Instrumentation for Soil and Rock, ASTM STP 1358, G.N.
Durham and W.A. Marr, Eds., American Society for Testing and Materials, West
Conshohocken, PA, 1999.
ABSTRACT: To obtain a better understanding of the stresses acting on cast-in-place

concrete box culverts, and to investigate the conditions which resulted in a culvert failure
under about 12 meters of backfill, two sections of a new culvert were instrumented. The
measured earth pressure distribution was found to depend upon the height of the
embankment over the culvert. For low embankment heights (less than one-half the culvert
width), the average measured vertical earth pressures, weighted by tributary length, were
about 30% greater than the recommended AASHTO pressures. The measured lateral
pressures were slightly greater than the AASHTO pressures. As the embankment height
increased, the measured weighted average vertical stress exceeded the AASHTO
pressures by about 20%. Lateral pressures which exceeded the vertical pressures were
recorded at the bottom of the culvert walls, and small lateral pressures were recorded on
the upper locations of the wall. The high lateral pressures at the base of the wall are
consistent with the results from finite element analyses with high density (modulus)
backfill material placed around the culvert.
KEYWORDS: box culvert, field instrumentation, pressure cell, vibrating wire, highway
embankment, earth pressure
Introduction
Cast-in-place concrete box culverts are commonly incorporated into highway
embankments.To simplify the design of these structures, the earth pressures are usually
taken as a function of the equivalent fluid stress due to the overburden. Although the
tGrad. Res. Asst., Civil & Environ. Engrg. Dept., Univ. of Tenn., Knoxville, TN 37996.
2prof., Civil & Environ. Engrg. Dept., Univ. of Tenn., Knoxville, TN 37996.
3Assoc. Prof., Civil & Environ. Engrg. Dept., Univ. of Tenn., Knoxville, TN 37996.
Copyright9 1999 by ASTM International www.astm.org
structural response may depend upon the stress level, there is no distinction made
between low overburden heights and high overburden heights. The actual loadings
experienced by the structures can be complex, and may change during construction and
the subsequent service life. Soil-structure interaction effects result in a state of stress
around the structure that is dependent upon the stiffness of both the backfill materials and
the structure.
Small size circular culverts have been instrumented and studied over the past 60
years (Davis and Bather, 1968; Corotis and Krizek, 1977; Davis and Semans, 1982).
Because circular culverts have equal rigidity and strength in the horizontal and vertical
directions, culvert installation methods were developed to reduce the vertical pressures
acting on the culvert. These methods divert the vertical stresses from the culvert to the
adjacent soil and result in an increase of the lateral pressures on the culvert sides
(Spangler and Handy, 1982; Vaslestad, et al., 1994). For culverts built on level ground,
the "imperfect trench" (Spangler and Handy; 1982) condition installation method may be
used. This method involves spreading a specified thickness of compressible material such
as baled straw or plastic foam immediately above the culvert followed by compaction
with normal backfill reducing the vertical earth pressures acting on the culverts. A
comparison between this method and normal compaction on two instrumented (2.0 m
height and 2.55 m width) concrete box culverts under a 10 m silty clay embankment
height showed that the imperfect trench method resulted in a significant vertical load
reduction (Vaslestad et al. 1994). The earth pressure immediately above the box culvert
installed with the imperfect trench condition was 62% of the pressure due to the weight of
the soil column above the culvert. The earth pressure on the culvert roof under normally
compacted backfill was 125% of pressure due to the soil weight. However, field
measurements (Yang et al., 1997) of pressures on an instrumented double cell concrete
box culvert (3.66 m in height and 9.91 m in width) under about 12 m backfill height,
indicated that the vertical earth pressure was not reduced by placing 2 m loose fill soil
around the culvert roof. The average measured vertical pressure was 124% of the soil
prism pressure (actual backfill unit weight of 22 kN/m 3) above the culvert and 145% of
the current AASHTO recommended pressure(recommended unit weight of 18 kN/m3).
Furthermore, the induced differential settlement due to the loose fill may cause damage to
the pavement at the top of the embankment.
Field test data on box culverts with normally compacted backfill are limited. The
reported instrumented culverts have been either small size culverts with a width less than
3 m (Russ, 1975; James et al., 1986; Vaslestad et al., 1994) or large size culverts with
relatively low backfill heights (Tadros et al., 1989). The suitability of the current
AASHTO recommended design pressures for these culverts was examined by monitoring
the pressures on an instrumented culvert throughout the backfilling process.
Instrumentation Description
The instrumented box culvert was 99 m long and 3.9 m high by 7.0 m wide. It
was a double cell culvert, constructed of cast-in-place reinforced concrete. Typically, the
top and bottom slabs were 0.78 m thick and the side walls were 0.41 m thick, although
YANG ET AL. ON MEASUREMENT OF EARTH PRESSURES 89
Unit: Meter
i i 7.01 ,~
0.80
ir J J f J i Jx~'~
(~1
0.80
= Vibrating Wire Pressure Cell
Figure 1 Illustration of Culvert Dimensions

and Pressure Cell Location
the dimensions varied along the length of the culvert with the change of the overburden
embankment height. The culvert's typical dimensions under the full embankment height
are illustrated in Figure 1. The culvert was constructed on relatively level ground at the
bottom of a broad valley. The site soil was a soft silty residual clay with shallow outcrops
of limestone rock. About 0.6 m of well graded crushed gravel was spread immediately
below the culvert in order to level the foundation and adjust for the variable thickness of
the bottom slab along the length. Well graded crushed stone was backfilled to a height of
0.6 m above the culvert roof to provide drained conditions for the culvert, then the culvert
was backfilled with silty clay, and high plasticity clay with occasional limestone boulders
of 0.35m or smaller diameter.
To monitor the earth pressures during the construction, 12 vibrating wire
hydraulic type soil contact pressure cells were installed in two separate sections. Upon
completion of the embankment, section A would be under about 19 meters of fill, and
section B would be under 11.7 meters. The relative locations of instrumented sections are
shown in Figure 2. The slope of the embankment was 1:2. Each instrumented section
consisted of 6 pressure cells with 3 cells mounted on the surface of the culvert wall and 3
on the roof. The location and pressure gage numbering scheme are shown in Figure 1.
The pressure cell (Geokon model 4810) consists of two 230 mm diameter circular plates
welded together around their periphery. One of the plates is thicker and designed to bear
against the external surface of the structure in order to prevent flexure of the cell. The
total thickness of the cell is 6 mm, and the aspect ratio (cell diameter over plate thickness)
is 38. This "intermediate" cell size is appropriate for the measurement of soil pressure as
suggested by Weller and Kulhawy (1982). Two different cell capacities were chosen: 345
kPa on the roof and 172 kPa on the culvert wall. The cells are capable of operating at up
to twice the rated capacity, but the accuracy decreases.
The cell was first fixed to the culvert wall and roof with concrete anchors through
4 mounting lugs around the edge of the plate, and a quick setting high strength grout pad
Figure 2 Schematic of Embankment and Culvert Cross-Section

with Typical Sections A and B
was used to assure uniform contact between the plate and concrete. Medium sand was
used to cover the cell and transducer housing to protect the cell from possible point loads
or other stress distortions induced by the large size particles in the crushed gravel. A
geosynthetic cover was attached to the concrete with adhesive to separate the gravel and
the sand. This installation is illustrated in Figure 3. The backfill was placed with
conventional compaction control criteria, with the dry density greater than 95%
maximum dry density determined by standard Proctor compaction. The unit weight of the
backfill gravel was measured in-situ by the sand replacement method (ASTM D 4914-89)
and the average was found to be 22.0 kN/m 3. The unit weights of the silty clay and high
plasticity clay were determined by the drive tube method and the average value for both
materials was determined to be about 18.0 kN/m 3.
Measured Vertical Stresses on the R o o f
The pressure changes during placement of the first 13.6 meters of backfill on the
culvert roof were recorded with respect to the backfill height. The recorded vertical
pressures and the surveyed backfill height above the cell are shown in Figure 4. The data
were collected over about 600 days (20 months), during which there were two periods of
about 3 months each when no construction took place. Figure 4 indicates that the
Figure 3 Installation of Contact Pressure Cell on the Culvert

variation in backfill height in both instrumentation sections A and B are similar until
about day 580, when section B (under the slope) reached the final embankment height.
Up to about 2 m backfill depth, the recorded pressures are similar in sections A and B,
and the measured pressures were nearly uniform across the roof. When the backfill height
reached about 6 meters, the pressures were found to vary significantly across the roof.
The highest recorded pressures were on the culvert comer (gage 4), which corresponds to
the location of greatest structural stiffness. This is consistent with results from previous
instrumented culverts (Tadros et al., 1989; Yang et al., 1997). The large deviation of
vertical pressures may reflect the influence of soil-structure interaction effects, which
become more significant as the structural deflections increase with increasing
embankment height. At backfill heights greater than about 6 m, the vertical pressures
measured at section B are less than those at section A. This is likely a result of the
position of section B close to the embankment slope.
Measured Horizontal Stresses on the Walls
The recorded lateral pressures on the culvert wall (Figure 5 and Figure 6) show an
increase with backfill height similar to the vertical pressures on the roof. However, the
Figure 4 Recorded Vertical Pressures on the Culvert Roof

with Backfill Height above Roof
92 FIELD INSTRUMENTATIONFOR SOIL AND ROCK
cells at the bottom of the wall (Figure 6) registered pressures exceeding the vertical
pressures, whereas the upper cells recorded relatively low lateral pressure. The influence
of compaction equipment on the lateral pressure can be identified as the recorded stress
peaks in Figure 5 during the 155-170th day, corresponding to the time when the gravel
was placed around the culvert. This is followed by a period of stress relaxation with
residual stresses differing at sections A and B. This relaxation does not appear to occur in
the lower portion of the wall (Figure 6), where the overburden stress is greater. The added
lateral pressure contributed by the additional thickness of overburden is negligible with
respect to the short term increase caused by compaction. As the backfill height was
increased above the culvert roof(about day 155), the upper cells in both sections recorded
a decrease in lateral pressure, whereas the pressure cell reading at the bottom of the wall
kept increasing and exceeded the vertical pressure. The high horizontal pressures were
observed at both section A and section B, and were larger than the manufacturer's rated
capacity for the pressure cells.
Figure 5 Recorded Lateral Pressures on the Culvert Wall

with Backfill Height above Cells
Figure 6 Recorded Lateral Pressures at Bottom of WaU

with Backfill above the Cells
Discussion
1. Comparison of Measurements with AASHTO Recommended Vertical Pressure
The instrumentation results were compared with the AASHTO (1996)
recommended design pressures. The AASHTO recommended pressure is an equivalent
uniform fluid pressure that will give approximately the same internal forces (moments
and shears) that are generated by the actual pressure distribution. The AASHTO vertical
pressure for the embankment condition is the embankment height times an assumed unit
weight of 18.8 kN/m3. To account for soil-structure interaction effects, AASHTO
suggests that this pressure be increased by a dimensionless correction factor or soil-
structure interaction coefficient, Fel
H = ~1.15 for well - compacted side soil

F~,=l+020 <Fo,.m~x where Fo,.m~ ~1.40 for uncompacted sidesoil (I)
and H is the embankment height above the box culvert and B is the width of the box
culvert. For well-compacted fills at the sides of the culvert, Fel should not be taken
greater than 1.15, and for uncompacted fills at the sides of the culvert, it should not be
exceed 1.40.
The AASHTO recommended vertical pressure ov used in design is then:
0"~ = F~I)'H (2)
The AASHTO soil-structure interaction coefficient (Equation 1) increases linearly with

backfill height, H. For the current instrumented culvert (B = 7 m) with well-compacted
fill, the limiting value of 1.15 is reached at a backfill height of 5.3 m.
The results presented in Figure 4 suggest that the vertical pressure distribution on
the roof is not uniform and thus cannot directly be compared to AASHTO design
pressures. An equivalent uniform pressure can be calculated by assigning a tributary
length to each cell and obtaining a weighted average. Based on the predicted distribution
of vertical pressure from a finite element analysis, a tributary length was assigned to each
pressure cell and the equivalent uniform pressure determined. The tributary width for
cells 4 and 6 at the culvert comer and at the centerline was taken as 0.2b and tributary
width at cell 5 in the middle of the span was 0.6b, where b is the span of one cell. Figure
7 compares the weighted average vertical pressures with both the uncorrected (Fr
and the corrected AASHTO pressures. The weighted average vertical earth pressures
were generally greater than the AASHTO design pressures, even at the lower backfill
heights (H/B<0.5). Figure 7 shows that the soil pressure recommended by AASHTO is
less than the tributary-weighted average pressure observed at the site.
The measured soil-structure interaction coefficient (the measured ratio between
the recorded vertical pressure and the calculated overburden pressure, 7H), can be
compared to the coefficient Fel recommended by AASHTO (Eq. 1). Although the
AASHTO coefficient is intended to produce an equivalent uniform pressure, it is
instructive to compare the coefficient Fet determined from individual pressure cells with
the AASHTO values. Figure 8 compares the coefficients from the instrumented culvert
with the AASHTO coefficients. Of 96 recorded pressure data from the instnLmented
culvert with backfill height less than 3.5 m (H/B < 0.5), the range of the recorded soil-
structure interaction coefficient Fel is from 0.73 to 1.85. The average value is 1.30 with a
standard deviation of 0.34. The highest recorded Fel values were readings taken
immediately after compaction of the backfill.
Also shown in Figure 8 are the soil-structure interaction coefficients calculated
from an instrumented culvert in Nebraska reported by Tadros (1989). This was a double
cell concrete box structure 8.1 m wide and about 4.3 m high. The permanent backfill
height was 2.6 m, but it was temporarily backfilled to a height of 3.7 m for several days.
Compacted silty clay was used as fill adjacent to the culvert. This culvert was under a low
embankment height (H/B<0.5), with a maximum H/B value of 0.47, and a corresponding
Fel = 1.06. As noted in Figure 8, the majority of the measured coefficients from the
Nebraska culvert also exceed the AASHTO design coefficient. Of 30 recorded vertical
pressures from 6 locations on the roof at backfill heights of 1.1, 2.4, 2.6, and 3.7 m, the
recorded F~I ranged from 0.94 to 2.07. The mean value was 1.36 with a standard
deviation of 0.26. Some of the cells recorded slightly less than the soil column pressure
3'H, when the fill height reached the maximum height of 3.7 m.
Although the materials adjacent to the two culverts were different, the vertical
pressures on the roofs under low backfill height were similar. When the embankment
height was less than 0.5B, the average recorded vertical pressures were greater than the
AASHTO recommended vertical pressure by approximately 30%.
In the current study, when the embankment height exceeded 4 m (H/B > 0.5),
some of the recorded soil-interaction coefficients are lower than the AASHTO Fez values.
Of 90 recorded pressure data points, the range of the recorded soil-structure interaction
Figure 7 Weighted Average of Recorded Pressures

Compared with AASHTO Pressures
2.5
9 O,.e~ Study(S~m A)
o Curre'tStudy(Se:~me)
o ~ D e a
'E
9~ 2.0 9 9
Aa,$1-ffOC0eflfdenl~ RII) oo
O
r ~o
c-
oo o ~ o 9
o
., o
\ O O0 9 9~
1.5 o @o o~ t o\O| 9 o o
o 9 @ ~ ~ o~ o 9 oO
.=.
o
"'~ .............................
, ..o....... .,~.- ..............
0 . . . 0 .... .
if)
~ /
O O i ~ ]
1.0 ..... ~ I~ o Oo o
o A~HTO~ ( ~ RnD
0.5 I I r I I I I I I I L I I I [ I I L I I I I I I I
0.o 0.4 0.8 1.2 1.6 20
t~-,-,~izedHacdil0-1~)
Figure 8 Comparison of the Normalized Vertical Pressures

with Normalized AASHTO Pressures
coefficients was from 0.8 to 2.0, with a mean value of 1.31 and a standard deviation of
0.27. This suggests that 90% of the measured pressure data were greater than the soil
column pressure. The average recorded soil-structure interaction coefficient was still
greater than the recommended AASHTO soil-structural interaction coefficient by about
33%.
For any given backfill height, the mean of the weighted average pressures (using
the assumed tributary widths of 0.2b for the pressure recorded at cells 4 and 6, and 0.6b
for the pressure recorded at cell 5) is greater than AASHTO design pressure. For low
embankment heights (H/B<0.5), the mean value is greater than the current AASHTO
design pressure by 31%. For high embankment heights (H/B>0.5), the mean value is
greater by 19%. The AASHTO upper limit ofFe#l.15 does not seem justified based on
these measurements.
In Figure 8, all of the recorded soil-structure interaction coefficient values less
than 1.0 from the current study are at locations in section B, which are close to the edge
of the slope (Figure 2). Lower vertical pressures recorded in this section may be
attributed to the lower lateral constraint compared with that in section A.
2. Comparison of Measurements with AASHTO Recommended Lateral Pressure
The current AASHTO (1996) recommended lateral pressure is obtained by the
YANG ET AL. ON MEASUREMENTOF EARTH PRESSURES 97
Figure 9 Recorded Lateral Pressures Compared

with AASHTO Pressures
embankment height multiplied by an equivalent liquid unit weight. For the maximum
lateral pressure, this equivalent unit weight is 9.4 kN/m 3, and for the minimum pressure,
the equivalent unit weight is 4.7 kN/m 3. Based on the unit weight of 18.8 kN/m3 used for
the vertical pressure, this corresponds to lateral earth pressure coefficients of 0.5 and
0.25, respectively. Figure 9 compares the recorded lateral pressures at different fill
heights with the AASHTO recommended pressures. The bottom cells (Number 1)
recorded larger than vertical pressure and the intermediate and upper cells (Numbers 2
and 3) recorded pressures below the AASHTO minimum pressure. This was observed at
both instrumented sections A and B.
Very high lateral earth pressures were observed for some combinations of backfill
modulus during a parametric study using the finite element method (Yang et al., 1997). It
was assumed that the backfill modulus was proportional to compaction energy, and
lateral pressures against a 9.9 m high, 3.7 m wide double cell culvert under 11.7 m fill
L a t ~ Pr,~urr (kP.)
200 160 120 80 40
0.5
1.0
1.5 o
o
E
ZO
25
3.0
3.5
4.0
Figure 10 Parametric Study of Earth Pressures under Different

Compaction Conditions (Yang et al., 1997)
were investigated. For a range of modulus values, both the distribution and the magnitude
of the lateral pressure were found to depend strongly on the modulus of backfill. As
shown in Figure 10 (Yang et al., 1997), very large lateral pressures at the bottom of
culvert walls were obtained with the very dense gravel (high modulus) surrounding the
culvert. The elastic modulus of the very dense gravel was 74 MPa, or 5 times the
modulus of gravel in the "standard analysis" which used a value of 14.8 MPa (Penman et
al., 1975). These analytical results suggest that high lateral pressures may be induced at
the bottom of the culvert wall when the backfill is well compacted. These high lateral
pressures may result in significant shear forces in the bottom of the culvert wall.
Pressure measurements on an instrtunented culvert during backfill construction
provided a record of the change in vertical and lateral pressure with increasing
embankment heights. The pressure acting on two different sections along the culvert
changed in a similar manner as the height of the embankment increased. At lower backfill
height (H/B < 0.5), the vertical pressures acting on the roof are consistent with the results
of other instrumented culverts of similar size. Based on the results from a numerical
analysis on a similar box culvert, the tributary width of pressure cells at different
locations can be determined and then a weighted average vertical pressure can be
obtained. The recorded weighted average vertical pressures are about 30% greater than
the recommended AASHTO pressure at low backfill heights. At high backfill heights
(H/B>0.5) the weighted average vertical pressures exceed the AASHTO recommendation
by about 20%. Although there are no other reported test results for box culverts at high
values of I-I/B, consistent measurements were obtained at two different instrumented
sections along the box culvert.
The lateral pressures acting on the culvert wall vary in a complex manner with the
increase of backfill height. At low values of H/B, the AASHTO recommended pressures
were found to slightly underestimate the actual lateral pressure on the culvert. At higher
H/B ratios, or H/B>0.5, the recorded lateral pressure became very large at the base of the
wall, exceeding the vertical pressure. The pressure at the top and mid-height of the wall
experienced some relaxation after compaction, and thereafter remained below the
AASHTO recommendation. The high lateral pressure at the base of the wall was
consistent with results from finite element analyses reported previously, in which high
modulus, a well compacted gravel backfill was found to result in large lateral pressures at
the base of the culvert wall.
The reported field measurements of earth pressures on large concrete culverts
suggest that the current AASHTO recommended pressure does not reflect field
observations. Furthermore, the current soil-structure interaction coefficient may not
account for important features such as the effect of backfill compaction effort.
Acknowledgment
This investigation was supported by the Tennessee Department of Transportation,
contract #CUT123RES1085. This support, and the input from William D. Trolinger,
Division of Materials and Tests, and Billy R. Burke, Structures Division, are appreciated.
References
Corotis, R. B. and Krizek, R. J., "Analysis and Measurement of Soil Behavior Around
Buried Concrete Pipe," Concrete Pipe and the Soil-Structure System, ASTM STP
630, 1977, pp. 91-104.
Davis, R. E. and Bather, A. E., "California's Culvert Research Program-Description,
Current Status, and Observed Peripheral Pressures," Highway Research Record
249, 1968, pp. 14-23.
Davis, R. E. and Semans, F. M., "Rigid Pipe Proof Testing under Excess Overfills with
Varying Backfill Parameters," Transportation Research Record 878, 1982, pp.
60-82.
James, R. W., Brown, D. E., Bartoskewitz, R. E. and Cole, H. M. "Earth Pressures on
Reinforced Concrete Box Culvert" Research Report 294-2F, Texas Trans. Ins.,
The Texas A&M Univ. Sys. College Station, 1986.
Penman, A. D. M., Charles, J. A., Nash, J. K. and Hnmphreys, J. D. "Performance of
Culvert under Winscar Dam," Geotechnique, Vol. 25, No. 4, 1975, pp. 713-730.
Russ, R. L., "Loads on Box Culverts under High Embankments: Positive Projection,
without Imperfect Trench," Research Report No. 431, Division of Research,
Department of Transportation, Lexington, KY., 1975.
Spangler, M. G. and Handy, R. L., "Soil Engineering" 4th edition, Harper & Row, New
York, 1982.
Tadros, M. K., Benak, J. V., Abdel-Karim, A. M. and Bexten, K. A., "Field Testing of a
Concrete Box Culvert," Transportation Research Record 1231, 1989, pp. 49-55.
Vaslestad, J., Johansen, T. H. and Holm, W., "Load Reduction on Rigid Culverts Beneath
High Fills: Long-Term Behavior," Transportation Research Record 1415, 1994,
pp. 58-68.
Weller, W. A., Jr., and Kulhawy, F. H., "Factors Affecting Stress Cell Measurements,"
Journal of Geotechnical Engineering Division, ASCE, Vol. 108, No. 12, 1982,
pp. 442-449.
Yang, M. Z., Drumm, E. C., Bennett, R. M. and Mauldon, M., "Influence of Compactive
Effort on Earth Pressure on a Box Culvert," Proceedings, 9th International
Conference of the Association for Computer Methods and Advances in
Geomechanics, Wuhan, China, 1997, pp. 2021-2026.
Standard Specificationsfor Highway Bridges. 16th Ed., The American Association of
State Highway and Transp. Officials (AASHTO), Washington, D. C. 1996.
Timothy J. McGrath, ~Ernest T. Selig,2 and Mark C. Webb 3
Instrumentation For Monitoring Buried Pipe Behavior During Backfilling
Reference: McGrath, T.J., Selig, E.T., and Webb, M.C., "Instrumentation for
Monitoring Buried Pipe Behavior During Backfilling," Field Instrumentation for Soil
and Rock, ASTMSTP 1358, G.N. Durham and W.A. Marr, Eds., American Society for
Testing and Materials, 1999.
ABSTRACT: An extensive instrumentation plan was devised to monitor buried pipe

behavior, soil behavior and pipe-soil interaction during backfilling. The emphasis of the
instrumentation plan was to monitor these parameters under different installation
techniques without impeding construction operations.
Different types and sizes of pipe were selected for installation in trenches
excavated in undisturbed in situ soil conditions. Installation variables included in situ
soil conditions, trench widths, backfill material (including controlled low strength
material), haunching effort, and compaction methods. A total of fourteen tests, each
including reinforced concrete, corrugated steel, and corrugated HDPE, were conducted.
Eleven of the installations were conducted with 900 mm inside diameter pipe and three
with 1500 mm inside diameter pipe. The pipes were buried to a cover depth of 1.2 m.
Measurements of pipe shape, pipe strains, pipe-soil interface pressures, soil

density, soil stresses, and soil strains were collected. Pipe shape changes were measured
by a custom built profilometer. Custom designed bending beam pressure transducers
were used in the steel pipe to measure interface pressures.
Most of the instrumentation performed well and measured results were within the
range expected. Pipe-soil interaction effects were effectively measured with the
instruments selected. Pipe shape changes were a very valuable parameter for
investigating pipe-soil interaction.
KEYWORDS: instrumentation, buried pipe, pipe-soil interaction, measurements
1principal, Simpson Gumpertz & Heger, Inc., 297 Broadway, Arlington, MA 02174-5310.
2professor of Civil Engineering, Department of Civil and Environmental Engineering, University of
Massachusetts, Amherst, MA 01003.
3Senior Engineer, Department of Civil and Environmental Engineering, University of Massachusetts,
Amherst, MA 01003.
INTRODUCTION
The interactions between a buried pipe and the soil surrounding it are complex
and challenging to monitor. Difficulties arise from the nature of soil, which is widely
variable in terms of particle size, and stiffness, and due to the presence of the pipe which
causes significant redistribution of soil stresses. The interactions between pipe and soil
during backfilling when soil is placed in a very loose state and then compacted,
sometimes under high impact loads, can be even more complex and have not been well
documented. This paper documents the instrumentation and data acquisition hardware
and soft-ware used to collect data during extensive field tests conducted to study the
behavior of buried pipe during the placement and compaction of backfill. It provides
details of instrument development, describes the performance of the instruments used in
these tests, and provides suggestions and guidance for similar measurements.
Scope
Instrumentation used in the field studies included both new devices developed
specifically for this project and commercially available instruments. This paper provides
construction details of the new devices as well as information on accuracy and precision.
Four groups of equipment are discussed:
computer hardware and software to perform the tasks of data collection,

instrttrnents for monitoring pipe behavior,
instruments for monitoring soil behavior, and
instruments to monitor pipe-soil interface pressures.
Background
The project formed part of Fundamentals of Buried Pipe Installation (McGrath,

1998) evaluating current and proposed construction practices for achieving quality
installations. Part of this task was accomplished by conducting laboratory tests (Zoladz et
al. 1996) and field tests (Webb et al. 1996, 1998) on different types of pipe using a variety
of soil types, equipment and techniques. However, only the field test instrumentation is
discussed in this paper. Detailed information on the instrumentation for both laboratory
and field tests is given in McGrath and Selig (1996).
The pipes used in the field tests were corrugated steel (metal), reinforced concrete
(concrete), corrugated high density polyethylene pipe with 900 mm (36 in.) inside
diameter, and profile wall high density polyethylene pipe with 1500 mm (60 in.) inside
diameter. The two polyethylene pipes are referred to as the plastic pipe in this paper.
Each test included all three types of pipe (Figure 1). Field tests were conducted by
installing an access manhole and then conducting four tests, two in each direction (narrow
trench condition, wide trench condition) from the manhole. Thus every test was
conducted in undisturbed in situ soil conditions. After the four tests were conducted the
McGRATH ET AL. ON BURIED PIPE BEHAVIOR 103
manhole was retrieved and m o v e d to the location o f the next series o f tests. A total o f
fourteen field tests was completed.
Instruments used to make measurements, and the purpose o f the measurements,

are summarized in Table 1. Figure 2 is a typical instrument layout for the field tests. The
figure shows the instruments for the metal pipe; however, the instruments for the concrete
and plastic pipe were similar, except as noted in the subsequent sections.
Access manhole \
Wide trench 3.0 m 2.4 m 2.2 m ,2.2 m , 2.4 m 3.0 m
(approx. 1.7 m) I~ ~ ~-~ ~I ,~ ,~,~ ~I~ ~I
rr_izlzz_-_-_-zz:_-zz_-_-iizz_-:_-z_-:iz_-]_-_-_-z_-_-s]z_-zz~
~:_-_-zz_-_-zz:_-z:_-]iz2zz_-zz_-zz--:--zzz_-zz]:z:]
I i
f ~ = , i ~-Metal IPlastic IConcr.~ )/ IConcr. )Plastic I Metal I I! I,
Narrow trench Trench 1 li~ Trench 2

(approx. 2.gm) ~ _~
Total length approximately 24 m'
lff=0.3m
FIG. 1--Typical Field Test Trench Layout
T A B L E 1--Summary of Measurements
Measurement Instrument Purpose

Pipe profiles Profllometer Pipe deflections and distortions
Metal and plastic - resistance strain Structural deformations,
Pipe wall strains gages moments and axial forces in
RCP none
- pipe wall
Correlation with standard field
In-place density & Nuclear gage practice, relationship to soil
moisture
stiffness
Relative motions of pipe and
Soil strains Inductance coil soil strain gages
trench wall
Horizontal soil Trench width affects, in situ-
stresses at backfill-in 230 mm (9 in.) earth pressure cells
backfill soil interaction
situ soil interface
Vertical soil stresses 230 mm (9 in.) earth pressure cells Soil loads on pipe (arching)
Soil strength & Quality of haunching technique,
Penetrometers
stiffness uniformity of haunch support
Concrete - 100 mm (4 in.) earth
Pipe-soil interface pressure cells embedded in pipe wall Pipe-soil interaction, quality of
pressures Metal - bending beam cells haunching
Plastic - none
104 FIELD I N S T R U M E N T A T I O N FOR SOIL A N D R O C K
Notes:
1. Instruments occurring at the same locations in this figure were offset longitudinally
in the ground.
2. The 900 mm metal pipe is shown in this figure. Details of instrumentation for other
pipe and diameters are similar.
1 in. = 25.4 mm
1 ft = .305 m
Soil strain gage

Earth pressure cell
I I
Interface pressure cell (concrete and metal pipe)
Circumferential and longitudinal resistance strain gage
on inside and outside surface (plastic and metal pipe)
FIG. 2--Typical Field Test Trench Cross-Section with Instrumentation
DATA ACQUISITION SYSTEM
A computerized data acquisition system was assembled to acquire and process

data for the project. The measurement instruments were connected by cabling to a signal
conditioning board which amplified the signal as required. From the signal conditioner
the signal was sent to an analog to digital (A/D) conversion board installed in the
computer. After conversion the signal was processed into engineering units, displayed,
and stored for future analysis. Several software programs were used to read the different
instruments. An external power supply was used to power the instruments and part of the
data acquisition system. The system was installed in a van.
A/D Board
A Model CIO-DAS08 A/D board manufactured by Computer Boards, Inc. was

used. Conversion resolution was 12 bits, thus the selected a:5 volt input range could be
divided into 4,096 increments of 2.44 millivolts. Readings were all static in nature,
therefore actual sampling rates were not an issue.
Signal Conditioner
Two 32-channel signal conditioning boards were used. On-board multiplexer

units coordinated the data from sixteen signal conditioner channels through a single
channel of the A/D board. Thus the 64 channels available on the signal conditioner board
used four channels on the A/D board. The remaining four channels on the A/D board
were not used. The signal conditioners, Model CIO-EXP32, were manufactured by
Computer Boards, Inc. On board amplification (gain) was available in increments of 1,
10, 100, 200, 500, or any sum of these increments, Strain gage based instruments were
read with a gain of 300 and the other instruments were read with a gain of 1.
Software
Software used for data acquisition was LabTech Notebook (Notebook) by

Laboratory Technologies of Wilmington, MA. Notebook had capabilities including
complex programs, channel triggers, and direct conversion of voltage signals to
engineering units. The last feature allowed real-time assessment of the data.
Power Supply
Three external voltage supplies were required to operate the system. The signal
conditioner boards were operated with supply voltages of +12 v and -12 v, while the
profilometer, pipe strain gages, and soil strain gages were powered with +4.76 v. This
voltage level was constantly monitored using one of the data acquisition channels.
PIPE INSTRUMENTATION
Profilometer
The changes in the pipe shape during installation and backfilling are indicative of
the pressure distribution on the pipe and of the stresses in the pipe wall. A device called a
profilometer was developed to measure the distance from a point near the center of the
pipe to the pipe surface at one degree intervals around the entire pipe circumference. This
data was processed to compute the pipe shape and change in vertical and horizontal
diameter and to estimate the radius of curvature of the pipe wall at all stages o f the
installation process.
Design and Construction--Several criteria were developed for the performance of the
profilometer. These are:
9 Mobility - The profilometer was designed to be easily installed and then removed
to avoid causing pipe restraint and to allow measurements at several locations in a
single test.
Centering in pipe not required - The mobility requirement introduced a second

requirement that the center of rotation of the profilometer would not need to be
located in exactly the same position for every measurement o f a single pipe, nor
would it need to be located at the center o f the pipe.
Measurement of Corrugated Pipe - The profilometer was designed with a roller to

bridge the corrugations o f the metal pipe and provide a smooth curve.
Range, Precision and Accuracy - It was estimated that the pipe deflections during
a test could be plus or minus 2.5 percent o f the inside diameter and that the center
of measurement might be up to 25 mm (1 in.) away from the center of the pipe.
Thus, for the 1500 mm diameter pipe, the range of the measurement arm had to be
greater than 100 mm. Radial measurements accurate to the nearest 0.05 percent of
the inside diameter were desired. Precision similar to the accuracy was also
required.
The major elements of the profilometer design, based on the above criteria are
shown in Figure 3 and include:
a three legged support arrangement to hold the device in a pipe during

measurements with adjustability of the legs achieved through threaded feet on the
ends,
9 a bearing to allow free rotation o f the measurement arm,
9 a rotary optical encoder to monitor the orientation of the measurement arm, and
the measurement arm, which consisted principally of a LVDT, to measure the

distance from the center of rotation of the profilometer to the inner surface of the
pipe wall.
McGRATH
ETAL.ONBURIED
PIPEBEHAVIOR 107
Note: not drawn to scale
[Topleg
. ~ ~ RotatingarmwithLVDT--
,jf~ / / " - ~ \ ~-~ Bearingwithrotary ]=~

(~//,~ \ L ~ e:~iZilin')"~
B o / t t ~ o el gsm 100R~ in., ~',
Flat metal foot
a) Front View b) Side View
FIG. 3--Schematic of Pipe Profilometer
Electronic Components--The principal measuring elements of the profilometer

were the optical encoder for rotational position and the LVDT for radial measurements.
The optical encoder was Model No. K15DM-360-5-SE-4A purchased from Lucas
Ledex Inc. The profilometer used two channels on the encoder. Channel M provided a 5
volt signal once per revolution. This channel was used to start the series ofprofilometer
readings. Channel A provided a separate 5 volt signal once for every degree of rotation.
This channel was used to signal when to read the LVDT.
The LVDT was actually a DC-LVDT because it incorporated the necessary

circuitry to allow DC input and output. The LVDT was a model LDC/3000/A purchased
from RDP-Electrosense Inc. and had a range o f + 75 mm (:t:3.0 in.). The manufacturer's
specified calibration coefficient was 33.9 mm/v (1.33 in./v); however, when assembled
and operated by the data acquisition system the coefficient was 39.9 mm/v (1.57 in./v).
Resolution of the LVDT is infinite, thus the A/D card resolution of 2.44 mv limited the
profilometer resolution to 0.10 mm (0.004 in.).
Operation--A data acquisition program, written in Notebook, monitored the

optical encoder, read the LVDT at the appropriate times, and stored the data on a
computer disk. Both instruments were read with a gain of 1.0.
Calibration--The calibration of the assembled profilometer was completed using a

750 mm inside diameter concrete pipe. The profilometer was installed in the pipe and
two complete 360 degree profile measurements were taken. The profilometer was then
removed from the pipe, reinstalled and two more complete profiles were taken. This
procedure was repeated until a total of four installations, and thus eight profiles, were
taken. The results of these tests showed that for a second set of readings without
removing the profilometer from the pipe the average change in reading is less than 0.1
mm (0.004 in.), the resolution of the A/D board. When the profilometer is removed and
reinstalled between tests the average change increases to 0.5 mm (0.02 in.) with a
standard deviation of 0.9 mm (0.04 in.).
Sample Data--The profilometer was used to determine the change in vertical and
horizontal diameter, and the complete pipe shape at several stages of a test. Sample data
is presented in Figure 4.
Backfill level:
-- Before backfilling
- - - - - Top of pipe
- ........ End of test
_-____.,..~. ." 9 ,.
f
I , I, I.i"~
_.-........... "~
~'.1,1,1
'~"'""?"~'"'
1,1,1,
"'"'~'~:l"""'+""'"';If"'"
,I,I,
I'1'1'1'1
"..
% Deflection" ..... % Deflection
I'l'l~lJI
J
01234 01234
a) Metal pipe - Field Test 5 b) Metal pipe - F i e l d T e s t 6
I,I,I, ' ~ ' . , I, I , I
''.i .' ..'"

% Deflection" ~ ..... %DefleXion
I'1'1'1'1 I'1'1'1'1
01234 _.~ 01234
c) Plastic pipe - F i e l d T e s t 5 d) Plastic pipe - F i e l d T e s t 6
FIG 4--Sample Profilometer Data, Magnified 10 Times
McGRATH ET AL, ON BURIED PIPE BEHAVIOR 109
Strain Gages
Strain gages were used to monitor the plastic and metal test pipe. The concrete
pipe was not expected to deform sufficiently under 1.2 m of cover to warrant gaging.
Strain gages were mounted on the inside and outside surface at the crown, invert, and
springlines of each pipe. Two gages, one each oriented in the circumferential and
longitudinal directions, were mounted at each location, thus each pipe had a total of 16
gages (Figure 5).
Strain gages were also used to build the interface pressure cells for the metal pipe.
Details presented here on gage selection, installation and waterproofing are also
applicable to that instrument.
Gage Selection and Installation--All strain gages, adhesives, and waterproofing materials
were purchased from Micro-Measurements Division of Measurements Group, Inc. (MM).
The gages selected were standard constantan foil faced, polyimide backed,
resistance gages. All gages had a gage length of 6.4 mm (0.25 in.). Selection of specific
features of the gages was based on the characteristics of the test pipe materials. Steel is a
good heat sink and gage heating is not a problem during operation; therefore standard 350
ohm gages were used (Type EA-06-250BF-350). Steel is also a common material and
self-temperature-compensating gages can be purchased directly from gage manufacturers.
Polyethylene is not a good heat sink, thus 1,000 ohm gages were used (Type EA-06-
250BK-10C) to reduce the power consumption and associated heat output. Thermal
properties in polyethylene can vary widely; thus no attempt was made to purchase self-
temperature-compensating gages.
All strain gages were bonded using M-M M-Bond 200, a cyanoacrylate adhesive.
M-Bond 200 is an easy to use, fast setting, room-temperature curing material. This
adhesive is somewhat sensitive to moisture; however, the short, two month, life of the
project, and the extensive waterproofing system planned allowed its use rather than
requiring slower curing epoxy adhesives. All gages were waterproofed with M-M's M-
Coat F.
Procedures for bonding the strain gages generally followed M-M

recommendations; however, the steel and plastic pipe each created some problems.
Polyethylene is difficult to bond to with any adhesive because of the waxy nature of the
surface of the material, and the galvanized surface of the metal pipe is difficult to bond to
because of the acidic nature of the surface after the galvanizing process. For both types
of pipe high humidity interfered with successful bonding. All the gaging was completed
in air conditioned laboratories.
For bonding gages to the plastic pipe the surface was treated with several cycles of
scrubbing with a household cleanser (Ajax with bleach), rinsing with distilled water, and
abrading with sandpaper, and then degreasing with isopropyl alcohol.
FIG. 5--Strain Gage Locations for Plastic and Metal Pipe
Procedures for bonding the strain gages to the metal pipe involved the following
steps:
The surface was degreased with isopropyl alcohol.
Cleaning with M-Prep Conditioner A was omitted.
The surface was warmed slightly with a heat gun prior to bonding. This step was
added to protect against humidity forming on the surface of the pipe.
Wheatstone Bridge Circuits--All strain gages were read using standard

Wheatstone bridge circuits. To minimize lead wire effects, all bridges were completed in
the test pipe, as close to the active strain gage as possible. After the test pipe were set in
place, the lead wires from the data acquisition system were brought into the pipe and the
lines were connected.
The gages for the plastic pipe were not self temperature compensated for
polyethylene, therefore dummy gages had to be provided. This was accomplished by
cutting out sections of pipe (from spare pieces) and bonding gages to the same portion of
the pipe profile as the actual sensing gage. The remaining half of the bridge was
completed using precision resistors. Wiring for the metal pipe was similar to that of the
plastic pipe except that temperature compensating strain gages were not required.
There was some random noise in the system when the data acquisition system was
set to a gain of 300. This was resolved by reading each gage ten times and averaging the
ten readings.
SOIL INSTRUMENTATION
Nuclear Soil Density Gage
In-place soil density and soil moisture content measurements were taken with a
Troxler, Model 3411B nuclear density gage on loan from the Massachusetts Highway
Department.
Soil Strain Gages
The soil strain gages (Selig, 1975) were used in the field tests to monitor the
change in the sidefill width, which is the dimension between the trench wall and the pipe
(see Figure 2). Six gages were used, one on either side of each of the three test pipe.
Gage Description--The soil strain gages operate on the principle of

electromagnetic inductance and are sometimes referred to as inductance coil strain gages.
A schematic drawing of the soil strain gage as used is shown in Figure 6.
FIG. 6--Schematic of Soil Strain Gage
The measuring element of the soil strain gage is the pair of inductance coils, each
consisting of a non-conductive disk wrapped with wire conductor. Driving one with an
alternating current induces a voltage in the second coil that is a function of the distance
between the two coils. The output of the gages is non-linear and is an alternating current
signal that must be balanced for phase and amplitude to accurately determine the distance
between the two coils. A signal processing unit, called the e-mu, developed at
Nottingham University, was used to automatically balance the phase and amplitude, and
to convert the output to a direct current signal. This is a redesigned version of the
original Bison Instruments, Inc. device. This allowed measurements to be taken through
the data acquisition system.
The sensitivity and range of the coils is a function of the size and number of turns
of wire. For this project 25 mm (1 in.) diameter coils were used which had a range of
approximately 75 mm (3 in.). As with many of the instruments, the sensitivity of the
gages was limited by the data acquisition system.
End plates were sized to insure movement with the soil within which they are
embedded. The transfer rods translate any motion of the end plates to the inductance
coils. Thus the measured change in coil spacing is the change in the end plate spacing.
The transfer rods were adjustable to allow for the different trench widths that were being
tested. The portion of the transfer rod near the coil was non-conductive to minimize
interference with the inductance field.
The guide tubes served two functions. One function was to keep soil away from
the transfer rods to minimize any friction that could affect the measurement. The second
purpose was to provide a tube for the guide blocks that keep the inductance coils in
alignment. The guide blocks control the lateral movement of the transfer rods to keep the
inductance coils in alignment. The guide tubes and guide blocks were non-conductive
because of their proximity to the inductance coils.
The soil strain gages were installed at the springline level of the pipe. A shallow
cross-trench was dug into the loose backfill and the gage was placed with one end plate
against the springline of the pipe and the other against the trench wall. The trench wall
was smoothed as much as possible and any spaces that were left were filled with a
medium sand.
Earth Pressure Cells
Soil stresses were measured at twelve locations, four for each pipe. As shown in
Figure 2, two earth pressure cells were set at an elevation 150 mm (6 in.) above the top of
the pipe to measure the vertical soil stresses directly above the pipe and over the trench
sidefill, while the other two gages measured horizontal soil stresses at the interface
between the trench backfill and the in situ material at the elevation of the pipe springline.
McGRATH ET AL. ON BURIED PIPE BEHAVIOR 1 13
The earth pressure cells were 230 mm (9 in.) diameter, fluid filled cells with
vibrating wire pressure transducers, manufactured by Geokon, Inc. Each cell also had a
built in temperature sensor. The cells for measuring vertical soil stress were built to read
pressure on both faces. The cells to measure horizontal stresses were designed with only
one pressure sensitive face. The other side o f these gages was a 6.25 mm ( 0.25 in.) thick
heavy plate. The heavy plate was set against the in situ material in the trench wall to
protect against erroneous readings if the trench wall was not smoothed adequately. The
Geokon cells have a range of 70 kPa (10 psi), accuracy of 0.25% of full scale, precision
of 0.2% of full scale, and a resolution of 0.1% of full scale. Resolution of the
temperature sensors is 0. I~
The gages were monitored with a GK-403 vibrating wire readout by Geokon, Inc.
The GK-403 reads both the vibrating wire transducer and the temperature sensor.
PIPE-SOIL INTERFACE INSTRUMENTATION
Interface pressures between the pipe wall and the soil were measured for the
concrete and metal pipe. Different approaches were used for the two pipe because of the
difference in flexural stiffness and wall profile of the pipe.
Concrete Pipe Interface Pressure Cells
Pipe-soil interface pressures for the concrete pipe were measured with 100 mm (4
in.) diameter fluid filled cells with vibrating wire transducers. The cells were custom
made for the project by Geokon, Inc. The pressure cell consists of two diaphragms
welded at the edges, with the fluid tube to the vibrating wire pressure transducer coming
off of the back diaphragm, as shown in Figure 7a. The transducers had a range o f 170
kPa (25 psi), accuracy of 0.25% of full scale, precision of 0.2% of full scale and a
resolution of 0.1% of full scale.
Design and Construction--The needs of the project imposed several criteria on the
design of the cells:
9 Cell diameter - The size of the measuring face was restricted because of the
curvature of the outside diameters of the test pipe. A cell diameter of 100 mm (4
in.) was considered an acceptable compromise as being large enough to minimize
the effects of local nonuniform stresses on the face of the cell, yet small enough to
minimize deviation from the curved surface of the pipe.
Transportability - The test program was designed to use two sets of test pipe. This
approach allowed backfilling of one test trench while the next trench was being
prepared for a subsequent test (Webb, 1996). This sequence required that the
interface pressure cells be removable so that they could be transported from one
set of test pipe to the next.
FIG. 7--Interface Pressure Cell for Concrete Pipe
A mounting ring and clamp system were developed to achieve transportability.

The complete interface pressure cell assembly, shown in Figure 7b, includes the pressure
cell, a mounting ring, and two expansion rings with bolts. The mounting ring and
pressure cell were cast together with non-shrink gout. The mounting ring was
manufactured with circumferential and longitudinal slots to create flexible tabs that
would push out against the concrete pipe wall when expanded with the expansion clamp.
Calibration--After being cast into the ring clamp, each gage was calibrated under
fluid pressure and for changes due to temperature. The fluid pressure calibrations showed
that the special construction did not change the response of the cell to externally applied
pressure. The calibration factors were within 3% of the transducer calibrations supplied
by Geokon. The small size of the cells, however, did result in a significant increase in
sensitivity to temperature. The average transducer temperature correction factor supplied

by Geokon was 0.04 kPa/~ (0.006 psi/~ while the average temperature correction
factor for the complete interface pressure cell was 1.24 kPa/~ (0.18 psi/~
Installation in Pipe--Ten interface pressure cells were used, two at both the crown
and invert, and the remaining cells were located at 30 ~, 60 ~ and 90 ~ from the invert on
each side of the pipe (Figure 2). This provided redundancy for all measurements.
Corrugated Metal Pipe Interface Pressure Cells
Measuring the soil pressure on the surface of the corrugated metal pipe is
complicated by the profile surface. Standard pressure cells have flat surfaces which, if
installed in the corrugated profile, create edges and sudden geometry changes that result
in conformance problems, causing arching of load onto or off of the cell. To minimize
this problem, custom interface pressure cells were designed to preserve the exterior
profile of the pipe. This was accomplished by using a close fitting cutout of the pipe wall
as a sensing element, supported by a beam instrumented with strain gages to serve as a
transducer.
Cell Configuration--The sensing element of the metal pipe interface pressure

transducer was a cutout of the pipe wall as shown in Figure 8.
The cutout was supported on a 140 mm long by 25 mm wide by 6.35 mm thick

(5.5 in. by 1 in. by 0.25 in.) aluminum transducer beam and secured by three screws, one
at the center of the cutout into the transducer beam and two at the edges of the cutout into
a transverse cross brace (see Figures 8b and 9). The ends of the transducer beam were
secured to the pipe by screws spaced at 127 mm (5 in.).
Calibration--The cells were calibrated by dead weights and through soil. The
dead weight calibrations were in general agreement with design calculations (correlation
coefficients r2 of about 0.99 and the reproducibility was typically 0.6 kPa).
Subsequent testing showed that the cells were less reliable in measuring actual
interface pressures than indicated by the dead weight tests because: 1) the screws securing
the cells to the pipe wall restrained the ends from free rotation and introduced an apparent
pressure as a result of pipe deflection, and 2) as the pipe deflected under soil load the
change in curvature of the pipe would cause the cutout to project out from the face of the
pipe or withdraw back from the face of the pipe, causing arching of load onto or off of the
cutout. These two effects were demonstrated in calibration tests.
Overall the two issues discussed above raise questions about the accuracy of the
metal pipe interface pressure cells. While the cells were used in the tests and the data is
reported, the results should be considered qualitative rather than quantitative in
demonstrating interface pressures.
Transducerbea~ipewall
76.mm
(3 m.)/
S
I.Onecorrugatior]]
I~ width vl
a) Outside of pipe
Pipewall -- __ 1;Orr~ubc~y25emmm
B r a c e ~
W e l ~ Corugatiocnutout
b) Inside of pipe
FIG. 8--Corrugated Metal Pipe with Interface Pressure Cell
-- Strain gages
~- Corrugation \ Mounting h o l e ~
cutout ~ \
I ~ Brace Strain gages~~,o,,,Transducer \

beam
L = 127 mm 9
End view Section A-A
FIG. 9--Metal Pipe Interface Pressure Cell
CONCLUSIONS
This paper documents the selection and application of instruments used to monitor
soil-culvert interaction during backfill placement. A wide array of instruments were used
to monitor the pipe strains and deformations, the soil stresses and strains and the
pressures at the soil-culvert interface. Most instruments were read using a computer data
acquisition system that is also documented.
Most of the instrumentation functioned as desired; however, some of the

instruments and sensors required special care and pertinent issues were discussed. In
addition, the interface pressure cells that were developed for the corrugated metal pipe
provided good qualitative information but the results may not represent actual values.
ACKNOWLEDGMENTS
The work reported in this paper was supported by funding from the National
Science Foundation, the Federal Highway Administration, the states of California, Iowa,
Kansas, Louisiana, Massachusetts, Minnesota, New York, Ohio, Oklahoma,
Pennsylvania, and Wisconsin, and the Eastern Federal Lands Highway Division of the
Federal Highway Administration. The pipes used in the tests were donated by Contech
Construction Products, CSR/New England, Hancor Inc., and Plexco/Spirolite Inc. UMass
graduate students Glen Zoladz and George Costa provided substantial assistance with the
tests. The Massachusetts Highway Department provided a nuclear density gauge for use
during the field tests.
REFERENCES
McGrath, T. J., 1998, "Pipe-Soil Interactions During Backfill Placement," Geotechnical

Report NSF98-444D, Department of Civil and Environmental Engineering, University of
Massachusetts, Amherst, MA.
McGrath, T. J. and Selig, E. T., 1996, "Instrumentation for Investigating Behavior of Pipe
and Soil During Backfilling," Geotechnical Report NSF96-443P, Department of Civil and
Environmental Engineering, University of Massachusetts, Amherst, MA.
Selig, E. T., 1975, "Soil Strain Measurement Using Inductance Coil Method",
Performance Monitoringfor Geotechnical Construction,ASTM STP 584, American
Society for Testing and Materials, pp. 141-158.
Webb, M. C., McGrath, T. J., and Selig, E. T., 1998, "Field Test of Buried Pipe with
CLSM Backfill," "The Design and Application of Controlled Low-Strength Materials
(Flowable Fill)," ASTMSTP 1331, A.K. Howard and J.L Hitch, Eds., American Society
for Testing and Materials.
Webb, M. C., McGrath, T. J., and Selig. E. T., 1996, "Field Tests of Buried Pipe
Installation Procedures," Transportation Research Record 1541, TRB, National Research
Council, Washington, D.C., pp. 97-106.
Zoladz, G. V., McGrath, T. J., and Selig. E. T., 1996, "Laboratory Tests of Buried Pipe
Installation Procedures," Transportation Research Record 1541, TRB, National Research
Council, Washington, D.C., pp. 86-96.
Instrumentation Support
Construction Activities
Dar-Hao Chen, 1 John Bilyeu,1 and Fred Hugo2
Monitoring Pavement Response and Performance Using In-Situ Instrumentation
REFERENCE: Chen, D., Bilyeu, J., and Hugo, F., "Monitoring Pavement Response
and Performance Using In-Situ Instrumentation," Field Instrumentation for Soil and
Rock, ASTM STP 1358, G. N. Durham and W. A. Mart, Eds., American Society for
Testing and Materials, West Conshohocken, PA, 1999.
ABSTRACT: The purpose of this paper is to present the effectiveness of in-situ

instrumentation on diagnosing the pavement layer conditions under full-scale accelerated
traffic loading. The test section is an in-service pavement (US281) in Jacksboro, Texas.
Multi-Depth Deflectometers (MDDs) are used to measure both permanent deformations
and transient deflections, caused by accelerated traffic loading and Falling Weight
Deflectometer (FWD) tests. Four different FWD loads of 25, 40, 52, and 67 kN were
applied in close proximity to the MDDs at various traffic loading intervals to determine
pavement conditions. It was found that the majority of rutting occurred in the newly
recycled asphalt mix. The aged (> 40 years) underlying base and subgrade layers
contributed less than 30% to overall rutting. Only the top recycled asphalt layer
underwent notable deterioration due to traffic loading. Up to 1.5 million axle repetitions,
the test pad responded to FWD load almost linearly, not only over the whole pavement
system but also within individual layers. However, under higher FWD loads, the
percentage of total deflection contributed by the subgrade increased.
KEYWORDS: accelerated pavement testing, deflectometer, remix, instrumentation,

load simulator
1Data Analysis Engineer and Assistant, respectively, Texas Department of

Transportation, Design Pavement Section, 4203 Bull Creek Road, #37, Austin, Texas
78731, Tel (512) 467-3963, Fax (512) 465-3681.
2Research Fellow, Center for Transportation Research, University of Texas at Austin,
3208 Red River, Austin, Texas and the Institute for Transport Technology, University
of Stellenbsoch, South Africa, 7600.
Accelerated Pavement Testing (APT) provides an avenue for authorities to find

fast and reliable solutions to pavement problems. The Texas Mobile Load Simulator
(MLS), an accelerated pavement testing device, was employed in this investigation. The
MLS is a machine that rolls tandem axles with standard truck fires over a 3m by 12m test
pad. The load on each tandem axle is set to carry a legal load of 151 kN at a speed of
approximately 19 kph. More details of the MLS can be found in references [Metcalf,
1996; Chen et al., 1997a].
Since the testing was conducted under closely watched conditions, the specific
pavement response and performance information could be obtained. As with all APT, a
link must be established between the data obtained from the pavement responses and the
pavement performance models being investigated. This link is generally established via
the pavement instrumentation and elastic-layer modeling. The effects of accelerated
traffic on the pavement was measured using in-situ instrumentation installed at different
depths and locations within the pavement structure.
In this investigation, the FWD was used to determine the response of the
pavement structure to four different loads, which are approximately 25 kN, 40 kN, 52 kN,
and 67 kN. The FWD testing was done at regular traffic loading intervals. Other
analyses included the comparison of these responses to those from the loading applied by
the MLS. Performance was measured in terms of rutting relative to the load applications.
Also the rutting was also analyzed in terms of permanent deformation in the respective
pavement layers.
Test Section
The test section is an in-service pavement located on US281 near Jacksboro in the
Fort Worth District. US281 is a two-lane highway in each direction. The Fort Worth
District Pavement Engineer indicated that there was an average of 3,100 vehicles per day
(1550 per direction) in 1994. The percent of trucks is approximately 17.4%.
Approximately 10% of traffic falls on the inside lane. Since the pavement was
rehabilitated in 1995, an estimated 9,850 trucks had traveled on the inside lane (or
approximately 10,000 to 19,700 ESALs of traffic, depending on the conversion factor
used) before the MLS was moved onto the test site.
The first asphalt layer of the test section was constructed in 1957. There were
four major upgrades/rehabilitations that were completed in 1971, 1976, 1986, and 1995.
Figure 1 shows the complete pavement history. TxDOT forecasts that the pavement
section in the outside lane will be subjected to 2 million ESALs over a twenty-year
period. The last major rehabilitation was done in 1995 with 50mm of recycled ACP
using the Remix process. Prior to that there was a major rehabilitation, using 76 mm of
lightweight aggregate ACP, in 1986. The inside southbound lane of US281 was closed
to traffic in April of 1997 for testing. The MLS was then moved onto the test section in
May. The outside lane remained open for public use. At the time of this analysis, 1.5
million axle repetitions had been applied to the test pavement by the MLS. Each axle had
been set to carry a load of 75.6 kN.
CHEN ET AL. ON PAVEMENT RESPONSE 123
FIG. 1 -- Pavement Section o f Southbound US281

The wander pattern was set such that the maximum percent coverage would be
83% over a cumulative width of 100 mm, centrally located, while 405 mm of the section
would be covered with 67% of the MLS "traffic", as shown in Figure 2. This calculation
was based on a 230 mm wide tire and 100 mm spacing between dual tires, which were
measured from the MLS. The total width of the wander pattern was approximately 860
mm.
................................................................................................................... 1.00
0.83
0.67
o.so
g
0.33
0.17
0.00
Distance (ram)
FIG. 2 --Wander Pattern Applied by the MLS
Objective
The main objective of the current testing was to determine the effectiveness of the
rehabilitation process performed in 1995. It was important to make sure that the
pavement failure would not be caused by a subsurface layer(s) within the 40-year-old
pavement. Several measures were taken to prevent incorrect adjudication of the
rehabilitation process, since failure of an overlay is sometimes caused by a subsurface
layer and not the overlay itself. Thus, it was imperative to have in-situ instrumentation
with the capacity to identify the condition of the various layers under traffic loading. The
Multi-Depth Deflectometer (MDD) was found to be a very suitable tool for this purpose.
Multi-Depth Deflectometer (MDD)
The Texas Transportation Institute (Trl) was awarded the contract to install
MDDs. Two MDDs were installed, one in the center of each wheel path. MDDs reveal
not only the transient responses under load but also the accumulated permanent depth
deformation. Each MDD hole contains three LVDTs to measure deflections at three
different depths, as shown in Figure 3. Ideally, an LVDTshould be placed at each layer
interface, so that the deflection contribution by each layer could be directly measured.
Mechanical limitations made this impossible, so the MDD sensors were installed as close
as possible to the ideal depths. MDD-T, MDD-M, and MDD-B has been used to refer to
the LVDTs at the depths of 90 mm, 320 mm and 570 mm, respectively. Figure 3 shows
the MDD depths and their relation to the pavement layers.
L R
0 ~ if
50 mm E
recycled ACP
MDD -T
133 mm
old A CP
381 mm
flexible
MDD-M t31 n 131
base
MDD -B
subgrade
anchors
FIG. 3 --MDD and Pavement Layer Depths
FWD-MDD Interrelationship
Deflection profiles from FWD tests provide valuable information to assess

pavement layer moduli through a back-calculation process. The back-calculated moduli
are often obtained from a best fit to the measured deflection profile. The verification of
back-calculated moduli has been done in two ways: (1) engineering judgment; and (2)
comparison with other test results such as laboratory and field seismic testing. Kim et al.
[1992] and Uzan and Scullion [1990] applied FWD tests on top of an MDD and
measured the resulting surface deflections and depth deflections simultaneously using
both the FWD geophones and the MDDs, respectively. They found that the depth
deflections measured by the MDD can be a powerful tool in evaluating the accuracy and
dependability of back-calculated moduli values from FWD data. The authors also used
the MDD data for analyzing the two-dimensional pavement structure of the pavements
tested in Victoria. In addition, FWD drops on MDDs provide valuable information to
determine if different pavement layers respond to FWD loads in a linear or nonlinear
fashion, thus validating the appropriateness of using linear elastic theory in the
calculation.
Four different loads, 25, 40, 52, and 67 kN, were applied in close proximity to the
MDDs with each load repeated three times to examine the repeatability of the results. In
the previous study [Chen et al. 1997b] applied four different loads with each load
repeated 10 times and found that the repeatability was high for the MDD readings and for
most of the FWD readings, The results indicated that errors were observed in the FWD
W1 measurement because the first sensor was in direct contact with the MDD cap, which
generated slippage that caused erratic readings. Therefore, two changes were made in this
study.
1. The center of the FWD loading plate was offset approximately 220 mm from
the center of the MDD cap, as shown in Figure 4. This was done so that the
FWD loading plate would not be in direct contact with the MDD cap.
2. The number of repeated tests was reduced from 10 to 3 because of the

excellent repeatability shown in the previous study. The results show
insignificant variation among the 3 tests.
F'WD
load 2213
p l a t ~ m
e~9 wl _J ~_ i
305 mm 183mm
(typical) ACP
818 rnm
381 mm
BASE
572 mm
MDD
FIG. 4--MDD-FWD Schematic
FWD-MDD data was collected prior to testing and after 2500, 5000, 20k, 40k,
80k, 160k, 300k, 450k, 600k, 750k, 900k, 1050k, 1200k, 1350k and 1.5 million axle
repetitions. Analyses were performed on the data to determine pavement degradation due
to axle loading. Figure 5 shows the history of depth deflections under 40 kN load. It is
encouraging to see that the MDD registered clear signals under FWD load at all three
depths. All collected FWD-MDD signals were high in quality. Figure 5 shows the
duration of the stress pulse to be approximately 28 milliseconds; and that the subgrade
contributed up to approximately 50% of total deflection.
FIG. 5 --Deflection Response History on Left Wheel Path for 42 kN FWD Load
Figure 6 illustrates the MDD depth deflections vs. axle repetitions for the left
wheel path. Only the averaged results from 40 kN FWD loads are presented in Figure 6.
Figure 7 shows the top MDD deflection (MDD-T) measurements under four different
load levels. Note that the pavement temperature was not controlled but was recorded.
Asphalt pavement temperatures were collected from three different depths: top (13 mm
from surface), middle (90 mm from surface), and bottom (165mm from surface). The
middle pavement temperature data were also incorporated into figures 6 and 7. The
deflections increased not only due to the traffic loading but were also affected by
pavement temperature. The MDD deflections increased up to 1 million axles due to
increased temperatures but decreased after 1 million because of decreased temperature.
This trend for deflections is the same at all three depths.
FIG. 6 --Depth Deflections at Various Loading Stages on Left Wheel Path
FIG. 7 --Pavement Responses Under Four Levels of FWD Loads on Left Wheel Path
The depth deflections at MDD-M and MDD-B at the end of 1.5 million axles
were approximately of the same magnitudes as at zero axles. However, there was some
deflection increase at MDD-T. No surface cracking that was not already present at the
beginning of the test (two thermal cracks), was observed at the end of 1.5 million axle
repetitions. It was thus concluded that no significant damage had occurred in the base
and the subgrade at the end of the test. A prime reason for this was the fact that very little
water (if any) could have filtered into the pavement layers, because the pavement test
section was covered by the MLS. Also, the base and subgrade had been in service and
consolidated for 40 years.
One of the advantages of the MDD is its ability to measure the permanent depth
deformation within the pavement structure. The surface rutting (S) or total rutting minus
the rutting registered at the top (T) MDD yielded the net rutting for the top 90 mm of
asphalt. Similarly, the top minus the mid MDD (M) reading was the rutting for the
bottom 100 mm of asphalt and top 127 mm of granular base. Also, the mid minus the
bottom (B) reading represent the bottom 254 mm of base material rutting. The B reading
in Figure 8 denotes the rutting from the subgrade layer. It was no surprise to see from
Figure 8 that the major rutting (> 60%) was contributed by the top 90 mm of asphalt. The
top 50 mm of asphalt is the newly overlaid (1995) layer. The aged underlying pavement
layers had reduced their rate of consolidation. In view of Figure 8, the rate of rutting in
the newly overlaid ACP layer increased after 300,000 reps partly because of the high
temperature (mid June) during that testing period. The underlying pavement layers also
began to contribute some rutting after 300,000 axles, because the soft surface layer
resulted in higher stress penetrating into the subsurface. After 1 million reps, the rate of
rutting for underlying pavement layers decreased, because the temperature was lower
during this testing period. The test passed 1 million repetitions in mid September.
LWP Permanent Deformation

180
160 I
A140
B m
"~120
.~ 100
.'AA
60
" . AAA
:4o
2O
0 _--
1000 10000 100000 1000000 10000000
Axles
FIG. 8 --Layer Permanent Deformation For Left Wheel Path
CHEN ETAL. ON PAVEMENTRESPONSE 129
At the end of 1.5 million repetitions, the surface rut depth, on average, had
reached 3.4 mm. No surface cracking was observed. Both the MDD transient responses
and permanent deformations indicated that the majority of material property changes
occurred in the top asphalt layer. Since the subsurface layers have been in service since
1957, their material characteristics were assumed to have remained fairly constant during
the test.
The test pavement had carried the equivalent of 8 year's traffic in accelerated form
without severe distress. Of course, the pavement had not been subjected to the full
environmental impact. Nevertheless, the district pavement engineer was satisfied with
the performance of the rehabilitation process.
The test pad is currently being subjected to a 25% overload to study the effect of
overload and to investigate whether it would bring about more visible distress.
Subsequently, it is planned to continue the test with the surfacing subjected to water
infiltration during testing.
BISAR Analysis and Anchor Movement Monitoring
The measured peak surface deflections were used to back-calculate the moduli of
the pavement layers using the MODULUS software. For this step, loads of
approximately 40 kN (from the 2"d FWD drop height) were used. The moduli from
MODULUS were then used in BISAR to forward-predict deflections both at the surface
and at depth. The moduli obtained from MODULUS were used only as a starting point.
Several iterations within BISAR were needed to obtain a set of moduli that would
estimate both the depth and surface deflections accurately at any load.
The accuracy of pavement analysis is greatly affected by the estimation of the

bedrock depth. Particularly, when shallow bedrock is encountered, the back-calculated
subgrade modulus can often be overestimated. Before the test site was selected, ground-
coupled ground penetrating radar was employed to assess the subsurface structure. From
the drilling and ground penetrating radar, bedrock was found in the range of 0.76m to
3.5m from the surface, for the 182 m long section which includes the MLS site. During
the MDD anchor installation, no bedrock was encountered up to a 2 m depth, where the
anchors were placed.
It is believed that the bedrock depth under the MLS test section is between 2.5 m
and 3.5 m. Since the anchors were not located in the fixed bedrock, any anchor
movement will affect the true MDD deflection readings. Therefore, anchor movement
monitoring becomes important to assure the integrity of the test data. In addition, anchor
movement information could be used to estimate the subgrade stiffness and depth to
bedrock. The lesser anchor movements mean that there may be shallow bedrock and/or
stiffer subgrade. Anchor movement data under FWD load was collected by attaching the
7th FWD geophone to the MDD core rod, which is connected directly to the anchor.
Tables 1 and 2 present the FWD deflections W1 (at the center), W2 (300 mm)
from the center of the load, and MDD deflections MDD-T, MDD-M, MDD-B. The FWD
load plate was offset approximately 220 mm from the MDDs. The magnitudes of anchor
movement are also given in tables 1 and 2. As shown in tables 1 and 2 and Figure 6, the
subgrade deflection constitutes slightly less than 50% of the total deflection. It was found
that at higher loads, a slightly higher percent of deflection was contributed by the
subgrade.
TABLE 1 --MDD Deflections Including the Anchor Movement for the Left Wheel Path at
the End ofl.2 Million Axle Repetitions
Load (kN) Wl(mm) W2 (mm) MDD-T (mm) MDD-M (ram) MDD-B (mm) Anchor (mm)
25 0.201 0.150 0.189 0.145 0.113 0.044
40 0.343 0.259 0.319 0.251 0.195 0.076
52 0.454 0.345 0.421 0.337 0.264 0.101
67 0.600 0.460 0.558 0.456 0.359 0.136
TABLE 2 --MDD Deflections Including the Anchor Movement for the Right Wheel Path
at the End of 1.2 Million Axle Repetitions
Load(kN) Wl(mm) W2 (mm) MDD-T (mm) MDD-M (mm) MDD-B (mm) Anchor(mm)
25 0.203 0.139 0.185 0.139 0.112 0.015
40 0.338 0.239 0.310 0.240 0.192 0.025
52 0.444 0.318 0.408 0.323 0.258 0.031
67 0.588 0.428 0.543 0.437 0.351 0.040
In view of tables 1 and 2, the anchor movements were three times higher in the
left wheel path than in the right wheel path. However, the deflections recorded by the
MDD-T, MDD-M, and MDD-B were only 2% higher in the left than the right wheel path.
Also, the surface FWD deflections are approximately the same between the left and right
wheel paths, as shown in Figure 9. It is the authors' opinion that the left wheel path has
deeper bedrock and higher moduli values for asphalt, base and subgrade. Bedrock in the
left and right wheel path was estimated to be at the depth of 3.2 m and 2.7 m,
respectively. The variation in bedrock depth and subgrade moduli is possible (even
within 1.8 m) in this case. Since the test section was built in a fill section, variation of the
subsurface properties is not uncommon. The confirmation of bedrock depth through
drilling will be performed once the test is complete. Figure 9 also shows that both
surface and depth deflections responded linearly to increased load. This observation was
supported by the fact that the data fit linear trendlines, with R-squared values exceeding
0.99.
IeFWD(WI)I
0.600 II InMDO-T (
i }I9
0.400ItXMDO-B I 9 !o.,o1 ~
~-' ~176176II o.o 1 m I I
R o,oo0/ "
N 0.00 /
20 40 60 80 20 30 40 50 60 70
Load (kN) Load (kN)
(A) (B)
4FFWD(W1)
0.60 mlMDD-T
9
~ 0.40 XMDD-B I
. i~ 9
ii
9 (k I 0.40 1 ~ 9
x9 ,,, o.2o 1 t x
t0 ~176
o.oo
l N 0.00 1
20 40 60 80 20 40 60 80
Load (kN) Load (kN)
(c) (D)
FIG. 9 --FWD-MDD Response Data (A) at 0 Axle Repetition Left Wheel Path (69 F mid-
Depth Pavement Temperature) (B) at 0 Axle Repetition Right Wheel Path (69 F
temperature) (C) at 600 000 Axle Repetitions Lej~ Wheel Path (94 F) (D) at 600 000 Axle
Repetitions Right Wheel Path (94 F)
Additional efforts were made to place the MDD between the second and third
sensors of the FWD with a spacing of 480 mm (center of FWD load plate to center of
MDD hole). Graphical displays of the measured surface and depth deflections at a single
load level (but two different offsets) are shown in Figure 10. At an offset between 220
mm and 480 mm from the load, the deflections measured within the asphalt layer (MDD-
T) are equal to or greater than those measured on the surface.
0.4 ..................................................................................................................................
A
E
03o3
E
qJ
E 0.2
O Suroooo,.eotion(FVVO)
O
m
Q.
o
0 l I I
0 400 800 1200 1600
Distance from Load (mm)
FIG. 10 --Measured Surface and Depth Deflections on Left Wheel Path for 40 KN FWD
Load
There are four major layers shown in the cores taken from the MLS test site.
Thus, the asphalt was divided into 4 different layers in the BISAR analysis. The total
thickness of ACP is 183 mm, as determined from the cores. The moduli and layer
thicknesses used in the analyses are presented in Table 3. The analyses included two
different load offsets (220 and 480mm) and four levels (25, 40, 52 and 67 kN) of loads.
After running MODULUS to obtain surface deflection-based layer moduli, these values
were adjusted (using BISAR) to better fit the actual measurements taken at various
offsets, depths, and loads. The comparison between measurements and predictions for
surface deflection from FWD and depth deflections from MDD are given in Table 4. As
shown in Table 3, the moduli are higher in left than right wheel path but the bedrock is
deeper, as explained in the previous section. It should be noted that the analyses in tables
3 and 4 represent the pavement condition at the end of 1.2 million axle repetitions. The
moduli variation in relation to the axle repetitions will be discussed in subsequent
publications.
TABLE 3 --Layer Properties Used in the BISAR Program

LEFT WP RIGHT WP
Modulus (Mpa) Thickness Poisson's Modulus Thickness PoJsson's
(mm) Ratio (Mpa) (ram) Ratio
3101 (AC) 51 0.35 2963 (AC) 51 0.35
2412 (AC) 25 0.35 2274 (AC) 25 0.35
1723 (AC) 25 0.35 1723 (AC) 25 0.35
1860 (AC) 81 0.35 1585 (AC) 81 0.35
289 (BASE) 381 0.38 276 (BASE) 381 0.38
76 (SUBG) 508 0.4 69 (SUBG) 508 0.4
48 (SUBG) 2159 0.4 55 (SUBG) 1651 0.4
1723 0.38 1723 0.38
(BEDROCK) (BEDROCK)
TABLE 4 --Left Wheel Path Measured Vs. Predicted Deflections, mm

(Predicted in Parentheses)
Left WP Offset = 220mm
Load Wl W2 MDD-T MDD-M MDD-B
25 KN 0.224 0.166 0.211 0.163 0.127
(.253) (. 177) (.201) (.181) (. 156)
40KN 0.860 0.272 0.335 0.268 0.207
(.381) (.267) (.302) (.272) (.234)
52KN 0.467 0.355 0.437 0.353 0.276
(.48) (.335) (.381) (.343) (.295)
67 KN 0.625 0.479 0.586 0.484 0.378
(.63) (.439) (.5) (.452) (.386)
Offset = 480mm
Load Wl W2 MDD-T MDD-M MDD-B
25 KN 0.213 0.162 0.126 0.119 0.099
(.254) (.177) (.143) (.14) (.129)
40KN 0.342 0.263 0.204 0.182 0.160
(.384) (.267) (.216) (.211) (. 194)
52KN 0.444 0.343 0.264 0.241 0.208
(.48) r (.272) (.267) (.244)
67 KN 0.602 0.465 0.363 0.329 0.288
(.63) (.439) (.356) (.348) (.32)
Conclusions
The purpose of this paper is to present the effectiveness of the in situ

instrumentation on diagnosing the pavement layer conditions under accelerated traffic
loading. It was found from this study that the MDD is a valuable tool in assisting
pavement engineers in characterizing the pavement material properties. Pavement
response and performance information of underlying layers under traffic loading were
evident from the MDD data. Several observations were derived from the analyses:
1. The majority (> 60%) of rutting occurred in the top new recycled asphalt mix.
The aged (40 years) underlying base and subgrade layers contributed less than
30% of overall rutting.
. Only the top recycled asphalt layer suffered notable deterioration due to
accelerated traffic loading. Deflections, at any given load, increased in this
layer with increased axle repetitions.
3. At intervals up to 1.5 million axle repetitions, the test pad responded to FWD
load linearly. Increased FWD loads caused proportionately larger deflections
not only within the whole pavement system, but also within individual layers,
4. High loads caused a slightly higher percent of total deflection to be

contributed by the subgrade. The subgrade exhibited this non-linear behavior
while the asphalt layers were still in their linear range.
5. The test pavement carried 8 years of equivalent traffic in accelerated form

without severe distress. The pavement engineer was satisfied with the
performance of the rehabilitation process.
Acknowledgments
The authors thank Mr. Ken Fults, Dr. Andrew Wimsatt and Mr. Mike Murphy for
their input and suggestions. In addition, the authors also thank Mr. Tom Scullion and Mr.
John Ragsdale for installing the MDDs. Thanks and appreciation are extended to Mr.
Sherwood Helms and the MLS field operation personnel for conducting the field tests and
collecting data.
References
Chen, D., Fults, K., and Murphy, M. (1997a) "The Primary Results for the First TxMLS
Test Pad," Transportation Research Board. Record No.1570. p. 30-38.
Chen, D., Meilahn, N., Murphy, M. and Fults, K. (1997b) "Evaluation of the Change of
Material Properties Under Accelerated Trafficking of Texas Mobile Load Simulator,"
Proceeding CD-ROM, 13th International Road Federation (1RF) Worm Meeting.
Toronto 1997.
Kim, Y. R., Khosla, N.P., Satish, S., and Scullion, T. (1992) "Validation of Moduli
Backcalculation Procedures Using Multi-depth Deflectometers Installed in Various
Flexible Pavement Structures," Transportation Research Board.
Record 1377. p. 128-142.
Metcalf, J. B., (1996) "Application of Full-Scale Accelerated Pavement Testing,"
Synthesis of Highway Practice 235. National Cooperative Highway Research
Program.
Uzan, J., and Scullion, T. "Verification of Backcalculation Procedures, "Proceeding
Third International Conference on Bearing Capacity of Roads and Airfields. 3-5 July,
1990. Trondheim, Norway, p. 447-458.
Ali Regimand 1 and Alan B. Gilbert2
Apparatus and Method for Field Calibration of Nuclear Surface Density Gauges
REFERENCE: Regimand, A. and Gilbert, A.B., "Apparatus and Method For Field
Calibration of Nuclear Surface Density Gauges," Field Instrumentation for ,Soiland
Rock, ASTMSTP 1358, G.N. Durham and W.A. Marc, Eds., American Society for
Testing and Materials, West Conshohocken, PA. 1999.
ABSTRACT: Nuclear gauge density measurements are routinely used for compliance
verification with specifications for road and construction projects. The density of
construction materials is an important indicator of structural performance and quality.
Due to speed of measurement, flexibility and accuracy, nuclear gauge density
measurement methods are becoming the preferred standard around the world.
Requirements dictate that gauges be verified or calibrated once every 12 to 18 months.
Presently, there are no field portable devices available for verification of the gauge
calibration. Also, the density references used for calibration of gauges, are large and not
designed for field portability. Therefore, to meet the present standards, users are required
to ship gauges back to a service facility for calibration.
This paper presents results obtained by a newly developed device for field verification
and calibration of nuclear density gauges from three different manufacturers. The
calibrations obtained by this device are compared to the factory calibration methods and
accuracies are reported for each gauge model.
KEYWORDS: verification, calibration, density, moisture, nuclear density gauge, density

gauge, backscatter, direct transmission
Nuclear density gauges are used for measurement of wet density and moisture in the
construction industry. Due to the large errors in the measurements of the density gauges,
a conference was held in Virginia to discuss and examine the variations in gauges (Hughes
and Anday 1967). Later in 1969, a symposium was held at North Carolina State
University to address improvements for nuclear gauges (McDougall et al. 1969).
Since their initial development in the 1950's, nuclear gauges have undergone many
improvements. Several variations of gauges are now available in the market to address
different needs in the industry (Regimand 1987). In the last twenty years, the major focus
has been on the accuracy of the calibration references and the methods used. In the last
version of ASTM Test Method on Density of Soil and Soil-Aggregate in Place by Nuclear
1Chief Nuclear Engineer, InstroTek, Inc., 3201 Wellington Ct. Ste. 101 Raleigh, NC 27615.
2product Manager, InstroTek, Inc., 3201 Wellington Ct. Ste. 101 Raleigh, NC 27615.
Methods(D 2922) and Test Method for Density o f Bituminous Concrete in Place by
Nuclear Methods(D 2950), the five block calibration method was specified as the
preferred calibration method. However, in the new revision of these standards the
calibration requirements were changed to allow flexibility in gauge calibration methods
used by the manufacturers. This was done to encourage research and development
innovations in the gauge calibration area.
A weakness that has always existed in the nuclear gauge application has been the field
verification of gauges to determine the amount o f drift in calibration. For years, gauge
standard counts were used to determine the gauge calibration accuracy. However, for
density measurements, standard counts can not be relied on for calibration accuracy check.
Standard counts are primarily utilized in the calculations to correct for radioactive source
decay. The geometrical relationship of source to detectors in the standard count position is
completely different as compared to the measurement positions used on the gauge. For
this reason, changes in counts at a particular position is not necessarily reflected in the
standard count position taken on a reference block provided with each gauge.
ASTM standards D 2922 and D 2950 require that gauges be verified or calibrated
once every 12 to 18 months. But due to the expense and difficulties related to shipment o f
nuclear products, gauge owners, in many cases, continue to use these devices for years
without a current verification or calibration.
For best results, a field device with a known density is needed to allow the users an
opportunity to collect density readings at each source position and determine the status o f
the gauge calibration.
Nuclear Gauge Theory
Nuclear gauges are used for measurement of density and moisture of soil, asphalt and
concrete. The material density is measured by employing a small radioactive source such
as Cesium-137 and one or more photon detectors. There are two modes of operation
routinely referred to as "backscatter" and "direct transmission". In the backscatter mode,
the source and the detector are placed on the test nnaterial in the same plane. Enough
shielding is provided in the base to stop the majority of direct photons from the source to
the detector. The photons from the source penetrate the test material and a fraction will
scatter back to the detectors. In the construction density range o f interest, the number o f
photons scattered back to the detector is inversely proportional to the material density.
Therefore, the higher the number of detected counts, the lower the density.
In the direct transmission mode, a hole is formed in the test material and the source is
placed into the material to a predetermined depth. The photons detected in this case are
representative & t h e density of the material in the path between the source and the
detectors. Again, in direct transmission mode, the number o f photons detected is inversely
proportional to the material density. Both backscatter and direct transmission densities
are calculated from the relationship in (Eq 1).
CR = A e x p ( - B D ) - C (1)
REGIMAND AND GILBERT ON APPARATUS AND METHOD 137
where A,B and C are gauge parameters determined at the time o f calibration, D is the
material density and CR is the ratio of counts on the test material to standard counts taken
on a reference block with the radioactive source in the shielded position.
Moisture is generally measured by employing a fast neutron source such as an
americium-241 :Beryllium and a thermal neutron detector such as a helium-3 tube. Both
the source and the detector in most gauges are placed in the base o f the gauge. The fast
neutrons from the source penetrate the material and interact with the hydrogen nucleus in
water (HzO). Multiple interactions reduce the energy of the incident neutron to a level
known as thermal energy. The detector in the gauge base is primarily sensitive to thermal
neutrons and there is a direct proportionality between the number o f counts detected by
the tube and the amount o f water in the material. The relationship used for calculating
moisture from the gauge counts is given by
MCR = a + bM (2)
where a is the intercept and b is the slope of the line determined at the time of calibration,
M is the moisture content and M C R is the ratio o f counts on the test material to the
standard count taken on the reference block.
In most manufactured gauges, the source and detector position for moisture
measurement is fixed inside the gauge base and for both the standard count and the
measured count the source to detector relationship is the same. The advantage of this
method is that once the gauge is calibrated at the factory, the moisture count ratio (MCR),
if taken at relatively the same time as the measured count will correct for all variations in
the gauge. However, in the density measurement geometry, the measurement position is
different from the standard count position and the density count ratio (DCR) does not
correct for variation in gauge readings.
Nuclear Gauge Calibration
Gauge density calibration is performed by using large blocks with known densities.
The gauge is placed on these blocks and a count is collected. A minimum o f three blocks
are needed to find the constant parameters in (Eq 1).
There are two methods o f choice for gauge calibration. Namely the "Three" and
"Five" block method.
In both these methods, magnesium (mg), magnesium/aluminum laminate (m/a) and
aluminum (al) blocks are used. The gauge is placed on each o f these blocks and a count is
collected. The known block densities are normalized by using the ratio of the metal mass
attenuation coefficient to that of an average soil. The detectors used in the manufactured
gauges are Geiger Mueller tubes and not capable of energy discrimination. The
normalization factors are calculated by assuming an average detected energy by the gauge.
The counts along with the normalized metal densities are used with an appropriate
curve fitting routine to determine the A,B, C constants in (Eq 1). This method is referred
to as three block calibration. It is important to note that in some cases, other combinations
o f materials are used as the three calibration blocks to determine the A, B and C constants.
In the five block method, the A, B and C constants are calculated using the metallic
blocks used for the three block method. In addition, counts on limestone and granite
blocks are used to adjust the B value. This forms the basis for the five block calibration
and the fundamental argument for this approach is that soil composition is halfway
between granite and limestone. Whereas, three block calibration is based on historical
five block calibration data to determine the position of the soil curve and to adjust the
metal block densities accordingly. Again, this is done because accurate detected photon
energy information is not available for these gauges.
Moisture calibration is performed by using at least two reference blocks, with different
hydrogen densities. Counts on the two blocks with the known moisture content are used
to calculate the slope and intercept of(Eq 2).
All gauges are calibrated prior to shipment from the factory. However, due to
electronic and mechanical variations with time, calibrations have to be performed
periodically. The main problem is the determination of calibration frequency. While some
gauges can continue operation with accurate calibration for many years, others might need
adjustments right after shipment from the factory. To effectively address this problem, a
portable device is required to determine the status of the calibration and to allow accurate
on site calibration.
The details on moisture calibration and density verification is outside the scope of this
paper. The main emphasis of this paper is comparisons of factory to the newly developed
field calibration method for the manufactured gauges.
Gauge Errors
There are three errors inherent in gauges manufactured. Namely, nuclear precision,
surface roughness and composition error. Gauge precision is the function of source to
detector geometry and the source intensity. Composition error is the effect of elemental
composition of the material to the density measured by the gauge. Surface roughness
error is the effect of photon streaming from the source to the detectors between the base
of the gauge and the material surface. Since test material surface varies for each test
location, surface roughness is the most difficult error parameter to define.
Composition error in most gauges can be minimized by appropriate filtering of the
detection system or adjustment of the source to detector distance. Even though this error
can be controlled, the adjustments to reduce this error usually causes an increase in the
surface roughness error. Manufacturers optimize the gauge geometry to balance these
errors and to hold these effects within a tolerable limit.
There are other external factors that significantly influence the gauge readings. Walls
or structures in the proximity of gauges can change the gauge measurements.
Manufacturers recommend gauge measurements to be taken at least one meter from any
walls. However, due to space restrictions and production schedules, most calibration
facilities, including some of the manufacturers, ignore this critical restriction. Also,
radiation background from other sources will reduce the gauge measured density. The
effects from the wall and background are easy to control and should be avoided during
calibration and field measurements. Significant errors can be introduced in calibration and
field results between gauges can vary, if these effects are not controlled.
Field Verification and Calibration Method
Recently a new development in the area of field verification and calibration was
introduced to the market. The portable engineered block (PEB) is a portable device that
can be used to determine the gauge calibration status and to generate new calibration
constants for the gauge. This device weighs approximately 18 kg and works with a
number of different manufactured gauges.
A combination of materials is used within a same block to create densities within the
construction range of interest. Composition error, surface roughness and boundary
conditions for each manufactured gauge is considered when determining the PEB density.
The block densities are traceable to NIST by standard procedures used by the
manufacturer. The gauge can be placed on the block and gauge density readings can be
compared to the densities provided with the PEB. The comparison may serve as the basis
for acceptance of the gauge calibration accuracy.
If calibration verification tails, the added benefit of this system is its capability of
field calibration. Counts taken on the PEB along with a PC program can be used to
simulate counts for blocks employed during the gauge factory calibration. Since the
density and the composition of the calibration blocks are well established, one can
accurately predict the counts based on the present status of the gauge. This routine can be
used to generate counts for a new three block or a five block calibration.
Test Method
The purpose of this test was to compare the results obtained on construction materials
using calibrations provided by the conventional factory methods and the one performed on
the PEB.
Test Procedure
Three different manufactured gauges, Gauge 1 to Gauge 3, were used for this study.
All gauges were calibrated at InstroTek according to the standard methods used by the
manufacturer of each gauge. Wall and background effects were eliminated in the
calibration facility by isolating the calibration area. Twenty minute calibration counts were
taken at each depth on known NIST traceable and ASTM required standard blocks.
Gauges 1 and 3 were calibrated with the three block method. Blocks ofmg, m/a and
granite were used for Gauge 1 and rag, m/a and al for Gauge 3. Gauge 2 was calibrated
with the five block method using rag, m/a, al, limestone and granite blocks. For a direct
comparison, the PEB calibration method performed on each gauge used the same
calibration method as used by the factory. The PEB calibration involves taking counts on
one block, and simulating the counts for the blocks used in the conventional factory
method.
Measurements were taken in the field on six different materials with the density range
of approximately 1922 to 2563 kg/m3(120 to 160 Ib/fl3). The field counts were used to
calculate densities by using the constants generated from the following:
1- Factory Method- Calibration normally provided with the gauge, when shipped
from the factory.
2- PEB Method- Calibration performed using the PEB device.
Test Materials
The test material consisted of three different types of asphalt mixtures and three soils.
The asphalt mixture readings were performed in the backscatter mode on a base, binder
and a surface mixture. Two sites with North Carolina red clay, a pure limestone block and
a pure granite block were used for backscatter and direct transmission measurements.
The asphalt projects were located on Interstate 40 between Raleigh and Research
Triangle Park, North Carolina. The clay site was located on Highway 401 and U.S. 1,
north of Raleigh, North Carolina. The limestone and granite block measurements were
performed in the laboratory. These two blocks along with the other blocks used during
this study are traceable to National Institute o f Standards and Technology (NIST).
Since gauge calibrations are optimized for construction materials, the best comparison
between the calibration methods for different manufactured gauges should be based on
construction materials similar to the ones selected in this study.
a- Asphalt Surface Mixture (ASM)- This material is generally used for riding surface
material on asphalt construction projects. The gauge density results presented in
the tables are averages often 4 minute readings, for each manufactured gauge.
The actual density (AD) for this material is taken as the average often core density
tests performed by the contractor.
b- Asphalt Intermediate Mixture (AIM), This material is used as an intermediate
layer on asphalt construction projects. The gauge density results presented in the
tables are averages often 4 minute readings, for each manufactured gauge. AD for
this material is taken as the average of eight core density tests performed by the
contractor.
c- Asphalt Base Mixture (ABM)- This mixture is used as the base layer for new
construction sites with larger proportion of large size aggregates. The gauge
density results presented in the tables are averages often 4 minute readings, for
each manufactured gauge. AD for this material is an average of twenty five core
density tests performed by the contractor.
d- Clay sitel and Clay site 2- This material was compacted in preparation for a new
road. The gauge density results presented are averages of two 4 minute readings,
one from each site, for depths backscatter (BS), 2,4,6,8, 10 and 12 of each
manufactured gauge.
e- Limestone Block- These counts were collected in the laboratory at depths BS, 2,
4, 6, 8, 10 and 12 for each manufactured gauge. The limestone block density used
in this study is NIST traceable.
f- Granite Block- These counts were collected in the laboratory at BS, 2, 4, 6, 8, 10
and 12 for each manufactured gauge. The granite block density used in this study
is NIST traceable.
Results
Four minute count time was utilized throughout this study to minimize the statistical
variations in each gauge.
The average count for all test sites for each manufactured gauge, 1, 2 and 3, is
reported in (Table 1) through (Table 3).
Table 1--Accumulated counts for all projects for Gauge 1
Depth ASM AIM ABM Clay Limestone Granite

Inches
DS 3 35065 35065 34798 35114 34520 34526
BS 15998 15618 14352 20027 15201 12572
2 . . . . . . . . . 86336 68950 51777
4 . . . . . . . . . 71216 56565 40398
6 . . . . . . . . . 48734 38963 25493
8 . . . . . . . . . 28854 23002 14164
10 . . . . . . . . . 16291 12514 7652
12 . . . . . . . . . 9408 7131 4759
To calculate the density of the test material, it is necessary to use the corresponding gauge
constant parameters A, B and C in (Table 4) to (Table 6) calculated by utilizing the
constants provided by the Factory Method and the ones performed by the PEB method.
Table 2-- Accumulated counts for all projects for Gauge 2

Inches
DS 3469 3469 3496 3508 3460 3449
BS 993 934 881 1293 951 761
2 . . . . . . . . . 3908 3113 2276
4 . . . . . . . . . 3220 2544 1671
6 . . . . . . . . . 2365 1866 1123
8 . . . . . . . . . 1575 1224 703
10 . . . . . . . . . 1000 747 414
12 . . . . . . . . . 606 436 256
Table 3-- Accumulated counts for all p r ~ e c t s f o r G a u g e 3

Inches
DS 2901 2901 2909 2902 2930 2874
3 DS is the daily density standard count.
BS 813 783 705 1134 797 596

2 ... 3590 2728 1862
4 ... 3342 2575 1673
6 ... 2456 1889 1140
8 ... 1526 1189 662
10 ... 892 661 356
12 ... 497 342 180
Table 4 - Gauge constant parameters for Gauge 1, for each method
Depth Factory Method PEB Method

A B C A B C
BS 2.9734 57.40371 0.19588 2.3828 67.80975 0.15349
2 11.4374 74.89223 0.22950 9.6361 97.31010 -0.28891
4 14.0418 63.03570 0.12731 11.6818 73.46961 -0.07307
6 17.3612 48.11310 0.16631 12.8271 57.27014 -0.00009
8 19.8939 38.10837 0.14181 13.4902 44.74362 0.05611
10 20.3064 31.94563 0.10585 12.8428 37.20080 0.06355
12 20.6020 26.98492 0.09367 13.4682 30.25024 0.08053
Table 5-- Gauge constant parameters for Gauge 2, for each method

A B C A B C
BS 2.8789 1.24874 -0.10566 2.9389 1.25607 -0.10771
2 6.9728 0.92231 -0.01640 7.1775 0.93334 -0.02671
4 9.0004 1.07577 0~07625 9.3703 1.08689 0.070326
6 10.5446 1.27141 0.06881 11.0273 1.28290 0.065651
8 11.9348 1.53628 0.02167 12.4651 1.54835 0.019517
10 13.1583 1.84295 -0.01086 13.5364 1.85564 -0.01216
12 15.1417 2.20183 -0.02693 14.9782 2.21519 -0.02651
Table 6-- Gauge constant parameters for Gauge 3, for each method

A B C A B C
BS 5.6871 1.59553 -0.12106 4.8485 1.49027 -0.10700
2 16.9252 1.42566 -0.25673 14.0234 1.30207 -0.18828
4 21.6530 1.53094 -0.20321 17.9546 1.40740 -0.13323
6 24.8264 1.71560 -0.13758 21.0438 1.61328 -0.09670
8 26.8041 1.95809 -0.08181 23.5309 1.88292 -0.06487
10 28.4525 2.25320 -0.04735 24.6028 2.17730 -0.04061
12 26.1565 2.50666 -0.02743 27.5310 2.54279 -0.02888
Densities calculated for each gauge uses the constants and the counts at each depth.
The density equation (Eq 3) is used for calculating the densities for Gauge 1 and (Eq 4) is
used for Gauge 2 and Gauge 3.
D = Bin (3)
D = --ln
B
'I--)
CR + C
(4)
where CR is the count ratio.

The densities for each gauge for all test sites are given in (Table 7) to (Table 9).
Analysis
The objective of this study is to determine if the factory and the PEB methods of
calibration yield similar density results. The data in (Table 7) shows that in all but one
case, the percent difference between the factory and the PEB methods were less than
+1.0%.
For Gauge 2, at depth 12, the percent difference for clay was -1.19 %. The depth 12
reading for limestone using the PEB method is in line with the measurements with the
other gauges. Also, assuming the granite block is homogenous, the reading at depth 12 is
consistent with the other depths when using the PEB method. Therefore, the factory
method readings at depth 12, in this case can be considered to be slightly off. Since
counts at depth 12 are lower than the other depths, it is possible that the higher statistical
variation in the counting rate resulted in a slight deviation in factory calibration.
To determine if the calibration performed by the PEB method is accurate, it is
necessary to compare the densities measured by this method to an acceptable value for
each material. If one accepts the fact that the gauges similar to the ones used in this study
are used daily for control of density on job sites, then the average of the density
measurements from the three gauges on each material at each depth should serve as a
good estimate for the density of the tested materials. Therefore, the overall standard error
o f estimate for each gauge at all depths will provide accuracy values for the two methods
o f calibration.
The data in (Table 8) is a comparison of the individual measurements taken with the
factory calibration method for each gauge, at all depths for each material and (Table 9)
represents the same data with the PEB method.
Table 7--Comparison of densities (kg/m 3) for all test sites, Factory Method (FM) and
PEB Method (PEBM)
Gauge 1 Gauge 2 Gauge 3

Depth FM PEBM %diff FM PEBM %diff FM PEBM %diff
A s p h a l t surface Mix: Avg. core density 2273 kg/m 3
BS 2239 2241 0.07 2216 2229 0.56 2240 2235 -0.25
Asphalt I n t e r m e d i a t e Mix: Avg. core density 2419 kg/m a
BS 2279 2281 0.09 2296 2309 0.57 2282 2276 -0.28
Asphalt Base Mix: Avg. core density 2513 kg/m 3
BS 2409 2424 0.63 2385 2388 0.12 2410 2397 -0.57
Clay: N/A
BS 1905 1892 -0.72 1916 1915 -0.06 19t0 1903 -0.34
2 1962 1954 -0.40 2004 2009 0.26 1997 1990 -0.35
4 2019 2017 -0.09 2047 2059 0.58 2042 2038 -0.22
6 2045 2039 -0.33 2085 2097 0.55 2072 2066 -0.29
8 2061 2055 -0.28 2104 2111 0.37 2093 2087 -0.27
10 2066 2066 0.00 2099 2097 -0.13 2083 2077 -0.29
12 2063 2070 0.36 2108 2083 -1.19 2075 2069 -0.27
Limestone: N I S T traceable density 2225 kg/m 3
BS 2297 2300 0.10 2271 2284 0.56 2273 2267 -0.27
2 2240 2242 0.11 2248 2265 0.75 2260 2255 -0.18
4 2251 2260 0.42 2236 2257 0.93 2264 2260 -0.18
6 2229 2230 0.01 2244 2263 0.83 2267 2260 -0.30
8 2219 2219 0.00 2243 2257 0.63 2254 2248 -0.28
10 2237 2241 0.18 2264 2268 0.14 2250 2245 -0.24
12 2251 2264 0.57 2283 2262 -0.91 2265 2259 -0.25
Granite: N I S T traceable density 2629 kg/m 3
BS 2641 2635 -0.22 2556 2571 0.60 2624 2601 -0.88
2 2637 2625 -0.44 2585 2603 0.68 2641 2624 -0.67
4 2625 2637 0.42 2584 2604 0.78 2641 2620 -0.82
6 2630 2619 -0.43 2587 2605 0.69 2658 2634 -0.93
8 2628 2609 -0.74 2593 2607 0.54 2652 2632 -0.78
10 2644 2620 -0.90 2597 2601 0.15 2625 2612 -0.53
12 2656 2645 -0.41 2626 2601 -0.95 2636 2635 -0.03
REGIMAND AND GILBERT ON APPARATUSAND METHOD 145
Table 8--Comparison of densities (kg/m 3) of the Factory Method densities and the
average density for each test site.
Material Avg. Dens. Gauge 1 Gauge 2 Gauge 3 %diff %diff %diff

ASM 2232 2239 2216 2240 0.33 -0.70 0.37
AIM 2286 2279 2296 2282 -0.29 0.45 -0.16
ABM 2401 2409 2385 2410 0.32 -0.68 0.36
Clay BS 1910 1905 1916 1910 -0.28 0.30 -0.02
Clay 2 1988 1962 2004 1997 -1.29 0.82 0.47
Clay 4 2036 2019 2047 2042 -0.83 0.54 0.29
Clay 6 2067 2045 2085 2072 -1.08 0.85 0.23
Clay 8 2086 2061 2104 2093 -1.20 0.86 0.34
Clay 10 2083 2066 2099 2083 -0.80 0.78 0.02
Clay 12 2082 2063 2108 2075 -0.91 1.25 -0.34
Lime BS 2280 2297 2271 2273 0.73 -0.41 -0.32
Lime 2 2249 2240 2248 2260 -0.41 -0.06 0.47
Lime4 2277 2251 2236 2264 -1.13 -1.79 -0.57
Lime 6 2247 2229 2244 2267 -0.79 -0.12 0.91
Lime 8 2239 2219 2243 2254 -0.88 0.19 0.68
lime 10 2250 2237 2264 2250 -0.59 0.61 -0.01
lime 12 2266 2251 2283 2265 -0.68 0.74 -0.06
Granite BS 2607 2641 2556 2624 1.30 -1.96 0.65
Granite 2 2621 2637 2585 2641 0.61 - 1.37 0.76-
Granite 4 2617 2625 2584 2641 0.32 -1.25 0.93
Granite 6 2625 2630 2587 2658 0.19 -1.45 1.26
Granite 8 2624 2628 2593 2652 0.14 -1.19 1.05
Granite 10 2622 2644 2597 2625 0.84 -0.95 0.11
Granite 12 2639 2656 2626 2636 0.63 -0.51 -0.13
Accuracy + 0.79 + 0.98 + 0.57
The standard error o f estimate, defined here as accuracy, indicates that for all gauges the
two calibration methods will approximately produce the same results. For Gauge 1, the
accuracy of the factory calibration is 0.04% better than the PEB method. However, for
Gauge 2 and Gauge 3, the PEB method is 0.12% and 0.21% better than the factory
method.
Table 9--Comparison of densities (kg/m 3) of the PEB method densities and the
average density for each test site.
Material Avg. Dens. Gauge 1 Gauge 2 Gauge 3 %diffl %diff2 %dill3

ASM 2232 2241 2229 2235 0.42 -0.12 0.15
AIM 2286 2281 2309 2276 -0.20 1.02 -0.42
ABM 2401 2424 2388 2397 0.94 -0.56 -0.18
Clay BS 1910 1892 1915 1903 -0.96 0.24 -0.38
Clay 2 1988 1954 2009 1990 -1.69 1.07 0.12
Clay4 2036 2017 2059 2038 -0.93 1.13 0.10
Clay 6 2067 2039 2097 2066 - 1.37 1.44 -0.06
Clay 8 2086 2055 2111 2087 -1.49 1.20 0.05
Clay 10 2083 2066 2097 2077 -0.80 0.69 -0.27
Clay 12 2082 2070 2083 2069 -0.58 0.05 -0.62
Lime BS 2280 2300 2284 2267 0.86 0.16 -0.58
Lime 2 2249 2242 2265 2255 -0.33 0.70 0.25
Lime 4 2277 2260 2257 2260 -0.73 -0.86 -0.75
Lime 6 2247 2230 2263 2260 -0.74 0.73 0.59
Lime 8 2239 2219 2257 2248 -0.88 0.82 0.42
Lime 10 2250 2241 2268 2245 -0.41 0.79 -0.24
Lime 12 2266 2264 2262 2259 -0.10 -0.19 -0.32
Granite BS 2607 2635 2571 2601 1.07 -1.38 -0.23
Granite 2 2620 2625 2603 2624 0.15 -0.69 0.11
Granite 4 2617 2637 2604 2620 0.78 -0.48 0.13
Granite 6 2625 2619 2605 2634 -0.23 -0.76 0.34
Granite 8 2624 2609 2607 2632 -0.58 -0.66 0.29
Granite 10 2622 2620 2601 2612 -0.08 -0.80 -0.38
Granite 12 2639 2645 2601 2635 0.21 -1.45 -0.16
Accuracy + 0.83 + 0.86 + 0.36
Conclusion
Results o f this study indicates that the PEB calibration method produces density
results comparable in accuracy to the factory calibration method. Tests conducted on six
different construction materials with three different nuclear density gauges calibrated with
the factory calibration and PEB calibration methods produced accuracy values of
+ 0.79%, _+0.98% and _+0.57% for the factory method and + 0.83%, + 0.86% and
+ 0.36% for the PEB method. Direct density comparisons between the two methods
showed that 71 out o f 72 densities had percent differences o f less than + 1%. Based on
this study, the PEB calibration method can be used as a reliable and accurate alternative to
the factory calibration methods used for nuclear density gauges.
References
[1] Hughes, C. S., and Anday, M.C. (1967), Correlation and Conference of Portable
Nuclear Density and Moisture Systems. Highway Research 177,.
[2] Gardner, R. P., and Ely, R. L. (1967), Radioisotope Measurement Applications in

Engineering, Reinhold Publishing Corporation, New York,.
[3] McDougail, F. H., Dunn, W. L., and Gardner, R. P. (1969), Report on Nuclear
Soil Gauge Calibration Workshop-Symposium, Nuclear Engineering Department, North
Carolina State University, Raleigh,.
[4] Regimand, A. (1987), "A Nuclear Density Gauge for Thin Overlays of Asphalt
Concrete," Transportation Research Board, Record 1126, Washington, D. C.,.
Peter Deming,1 and David Good2
Two Weight System for Measuring Depth and Sediment in

Slurry-Supported Excavations
REFERENCE: Deming, P. W. and Good, D. R., " T w o Weight System for

Measuring Depth and Sediment in Slurry-Supported Excavations," Field
Instrumentation for Soil and Rock, ASTM STP 1358, G.N. Durham and W. A. M a r l
ABSTRACT: This paper describes a two weight system using bar and flat shaped
weights for measuring depth and detecting sediment at the bottom of slurry-supported
excavations. Currently there are no standard depth measurement weights or methods for
reliably identifying bottom sediment. Two weights and a procedural system for using the
weights is described. Details suitable for manufacture are provided.
K E Y W O R D S : Slurry, Sediment, Field Inspection, Bentonite, Slurry Trench, Slurry

Wall, Drilled Pier
Introduction
This paper describes two sounding weights and procedures for measuring depth
and thickness of sediment at the bottom of slurry-supported excavations. At present, there
is no standard weight or procedure for depth measurement. There is also no standard
means for measuring sediment, although "bobbing" a weight to discern bottom drag is
often performed to infer the presence of sediment. The current typical practice of bottom
sounding, particularly for sediment control, is inconsistent, inaccurate and subjective.
Sounding with a heavy weight provides little quality assurance information to the field
engineer even though proper depth control and sediment removal are crucial to the
performance of foundations and hydraulic barriers constructed using the slurry method.
We use a two weight system for quality control inspection of slurry-supported
trench, panel and drilled shaft excavations. One weight is a solid "bar" which readily
penetrates loose bottom sediment. The other weight is a "flat" weight which rests on the
sediment surface. The thickness of sediment at the bottom of the excavation at any one
location is determined by comparing depths carefully measured by each weight, from a
uniform reference.
With the ability to reliably measure bottom sediment, the engineer can determine the
suitability of the excavation bottom for backfill or concrete placement. Contract
specifications should include a provision for using the two weight system to determine
sediment thickness and should define criteria for bottom acceptance.
In addition to the shape and mass of the sounding weights, the ability to detect
sediment depends on other factors. These include: the time readings are taken relative to
trench events, slurry viscosity, character of sediment, excavation depth, reliability of the
fixed reference and most notably inspection technique.
1 Partner, Mueser Rutledge Consulting Engineers, 708 Third Ave., New York, NY 10017
2 Associate, Mueser Rutledge Consulting Engineers, 708 Third Ave., New York,
NY 10017.
DEMING AND GOOD ON TWO WEIGHT SYSTEM 149
Description of Weights
The two weights which we have used are detailed for manufacture in Figures 1 and
2. The weights are attached to the end of fiberglass reinforced graduated survey tapes.
Use of a light cable or rope to support the weight has been shown to allow the measuring
tape to wrinkle, giving falsely deep and non-repeatable readings. The weight should be
raised and lowered only by the survey tape. In our excavation specifications, the
foundation contractor is required to provide the weights and graduated tapes. Several
weights should be provided, as weights are lost when the graduated tape breaks.
The "flat" weight consists of a 345 m m (14 in.) long piece of 19 mm (3/4 in.)
square or round bar stock welded to the center of a 160 mm (6 in.) diameter steel disk cut
from 12 m m (1/2 in.) thick steel plate. A 32 m m (1-1/4 in.) wide band is welded around
the edge of the disk to create a 20 mm (3/4 in.) high lip. The lip retains sediment when the
weight is removed from the slurry column, giving the operator the ability to recover a
specimen of sediment or thick slurry. A slot in the bottom of the weight receives the
survey tape grip. The survey tape is attached to the shaft of the weight with duct tape. The
flat weight weighs 3.1 kg (6.9 lbs) and has a bearing intensity of 1.5 kPa (0.015 tsf), less
than 1/30th that of the bar weight.
The "bar" weight is a 610 mm (24 in.) long piece of 25 m m (1 in.) square or 29
m m (1-1/8 in.) diameter round bar stock. It weights approximately 3.1 kg (6.8 lbs) and
has a tip bearing pressure of 47 kPa (0.5 tons per square foot). We have used bar weights
ranging from approximately 3 to 5 kg (7 to 12 lbs), and find equivalent performance of
these for sinking the measurement tape and penetrating through sediment-laden slurry.
The grip at the end of the survey tape is extended and placed over the bottom of the bar,
and duct tape is wrapped around the tip and shaft of the weight to attach it to the graduated
measuring tape.
Causes of Sediment
Slurry is used to resist earth pressure and groundwater, enabling excavations to
progress safely to great depths without structural bracing. When initially prepared, slurry
is a mixture of water, bentonite clay and additives which enhance the colloidal suspension.
Freshly prepared slurry is only slightly heavier than water, because it contains only a small
amount of solids. As the excavation progresses, clay, silt and sand particles become
suspended in the slurry. Suspended soil particles increase the slurry weight, which
enhances the stability of the excavation.
The ability of slurry to suspend particles depends primarily on its viscosity and gel
strength. Most construction slurries will support fine sand sized particles in suspension for
long periods of time. However, viscous slurry is required to suspend coarse sand and fine
gravel particles. Solids which cannot be suspended by the slurry fall to the bottom.
Bottom sediment is a deposition of these solids within a semi-fluid slurry. The sediment
can include sand and gravel particles, clay balls, debris, and silt or clay-laden slurry.
Theoretical Bearing Capacity

The pointed weight fails the sediment in end bearing as it passes through the soft
sediment to reach the bottom of the excavation or the top of tightly packed granular
sediment. Because of its larger bearing area, the flat weight rests near the top of the
sediment when it attains a factor of safety of 1.0 against bearing failure. The bearing
intensity of the tip of the bar weight is approximately 30 times that imposed by the flat
weight. If the fiat weight is considered to be a simple circular footing statically bearing on
the surface of a soft clay, a shear strength greater than 0.24 kPa (5 psf) should support the
flat weight. However, the flat weight may be supported by weaker sediments because of
the following:
Weight:
-- 3130-J:225 g
= (6.9-~0.5 lb)
Bearing Area:
A =214 em 2
= (33.2 in) 2
i Bearing Intensity:
= 1.5 kPa
19ram - = (0.21 lb/in)2
-~- /1/8 V
.... J
II I~ i II ~,'/ // /A I'/// // // // /AII
IN""" ~ L.Slot I
__ 165---ram(6 1/2") _
II_ 1 5 ~ m (61/4")_ Ib~_____.~
I-- --I
I--
165ram (6 1/2") 2'1 - --I
Elevation Section A - A
Materials:
A. 3/4" round bar or No. 6 deformed
bar.
B. Rolled steel plate 1/2" thick.
C. Rolled steel plate 1 1/4" wide x
1/8" thick.
Notes:
1. Steel shall be ASTM A615 Grade
40 Billet Steel.
2. All edges shall be de-burred.
---,.-.___~25mm (1") 3. Bottom shal be smooth (grind, ete).
4. End of measuring tape shall be
A,P referenced to bottom of fiat
weight.
Plan
5. Slot provides drainage to plate and
reeieves end grip of measttring tape.
Figure 1 - Flat Weight
/
77 Weight:
// = 3100 g
// = (6.8 lb)
// Bearing Area:
# //
= 6.45 cm 2
= (1 in) 2
Bearing Intensity:
/ / f E = 47 kPa
//
// = (6.8 lb/in) 2
# //
9" //
//
//
# //
# //
# //
Materials:
# //
D. 1" square bar.
E. 1 1/8" round bar or No. 9 deformed
# //
bar.
Notes:
# // 1. Steel shall be ASTM A615 Grade
# //
40 Billet Steel.
2. All edges shall be de-burred.
// 3. Either round or square bar may
// be used.
4. End of measuring tape shall be
referenced to end (bottom) of
Section Section bar weight.
D (3
Plan Plan
Figure 2 - Bar Weight
9 Sediments are suspended in a thixotropic fluid of varying gel strength.

9 Sediments vary in composition and size (clay balls, granular particles, etc.) The
particle packing density and angularity of the granular sediments also varies.
9 Frictional drag on the falling weights, particularly the flat weight, may play a
significant role in the depth within the sediment at which the weight is supported.
9 Support provided by the underlying hard bottom of the excavation which influences
the location of the bearing failure plane within the sediments.
The flat weight has been observed to slowly sink into the bottom sediment, even when
suspended under slight tension, suggesting sediment may behave as a viscous fluid.
Depth Measurement Procedures
1. Use normal caution when working adjacent to slurry supported excavations. A

man-bridge spanning over the excavation may be required to provide access for use
of the weights in some cases, especially where the excavation walls have caved and
the center of the excavation cannot be reached from the side wall with an
outstretched arm. Use work gloves when performing soundings.
2. Specifications often require the contractor tO take depth measurement readings.
This is satisfactory for the contractor's quality control, but the inspecting engineer
should operate and feel the weight performance during all quality assurance
measurements for bottom approval. Proper technique is needed to take repeatable,
reliable depth soundings and such functions are normally best performed by the
engineer with contractor support.
. Bar and flat weight depth readings are compared in order to define sediment
thickness. These depth measurements should be taken at the same location and
time, in immediate sequence. Bar and flat weight readings both need to be taken to
document bottom depth and sediment thickness following a cleaning attempt for
final bottom approval.
4. Lower the sounding weights under tension to maintain a continuously plumb line.
When the weight is carelessly thrown in the slurry, it may not be plumb below the
assumed sounding station because the slurry viscosity slows its return to vertical.
Allow the graduated tape of the descending weight to feed through the fingers. Do
not allow the weights to "free fall." Do not permit the weight to slide down the
excavation walls. This can happen when sounding near the steep end face of an
excavation or along side walls that have sloughed. Sliding weights on end faces
can mislead the completion of panel construction, at comers of excavations and at
the leading edge of a trench. The operator probably will not realize the weight is
sliding, except for feeling a slight drag. If this is suspected, pull the weight up out
of the trench and perform the sounding from the center of the excavation, As an
altemative raise the weight at least one meter (3 feet) and bob it slowly to attempt to
"catch" the excavation face, defining it for consideration.
5. The weights, particularly the flat weight, should have a slow, controlled descent
near the bottom so that they land without impact. The intention is that the flat
weight will rest on top of sediment, not penetrate it. Conversely, bar weight
soundings must be coaxed to penetrate through the slurry-supported sediment. A
drag on the bar weight or a "soft bottom" feel indicates sediments are present.
Progressively work the bar weight into the bottom sediment by repeated up and
down motion, raising the bar weight about 300 mm (one foot) and allowing it to fall
through the sediment as required. Confnm that the weights are not "laying down."
6. Pull the tape taut and read the depth of the weight below the reference elevation.
Flat weight readings should be recorded next to corresponding bar weight readings.
7. Provide a description of bottom sediment recovered with the flat weight. Be aware
that material recovered could be "filter cake" if the weight drags the side wall during
removal.
Project Experience
The two weight system has been used successfully to identify bottom sediments on
four projects we have been directly involved with. The amount of sediment detected has
varied considerably, in a manner consistent with our understanding of sediment generation
discussed above. Our experience with the two weight system has been similar on each of
the four projects. Typically, at the start of a project, the two weights reveal some amount
of sediment, undetected by the "drag feel" of the bar weight. If the sediment problem
became sufficiently troublesome, the contractor was required to ref'me the bottom cleaning
and/or excavation method in order to achieve sediment criteria given in the specifications.
Where extensive sediment removal was required, various modifications were attempted to
mitigate sediment generation. Modifications to the excavation system which met with some
degree of success included: increasing slurry viscosity, dragging the bucket parallel to the
trench bottom when cleaning, increasing the number bucket teeth and using short teeth,
welding plates between long bucket teeth, and separating the toe of the backfill from the
excavation area. On two trench projects backfill placement was held at one location until no
additional backfill could be accepted. We believe this practice reduced the quantity of
sediment by minimizing mixing of backfill with the active slurry column. Using the two
weight system allowed observation of this influence.
Once a proper cleaning and excavation system was developed, the two weight
system sediment check became confn'matory, typically showing little sediment. However,
it is the opinion of the authors that the two weights should still be used routinely. Using
the weights can detect small side wall collapses below the slurry level as observed and
documented on one project, can detect unusually thick slurry as occurred on two projects,
and in general promotes good trench quality, giving information about the excavation
bottom to the inspection staff.
Notes For Field Inspection

1. The precision and repeatability of depth measurements is dependent on the reference
used and the care exercised when sounding with the weights. The writers
recommend using survey stakes with elevation reference marks about every 3 m (10
feet) along trench and panel wall alignments.
2. The excavation equipment used and its operation will influence the bottom
roughness and therefore the repeatability of bottom soundings. Holding the bucket
teeth parallel to the trench bottom while the backhoe tracks backwards often
improves sediment retrieval and bottom cleaning, and enhances sounding
repeatability. This method should typically be employed during the last few feet of
excavation advance and for final bottom cleaning.
3. A sufficiently viscous slurry will prevent excessive sedimentation; conversely, thin
slurry will increase sedimentdevelopment. Adding water, chemicals, or thin slurry
to trench slurry must be done with caution, as this procedure may generate large
amounts of sediment along the entire trench alignment in areas where backhoe
excavation equipment cannot reach for sediment removal.
4. In our experience, mixing fresh slurry at a 6% bentonite content to a viscosity of
about 48 Marsh funnel seconds provides a workable slurry with good particle
holding capacity. Fresh slurry pumped to excavations should be hydrated, well
blended and have consistent viscosity. In a long and deep trench slurry will stratify
by density and viscosity. Heavy and viscous slurry will fall to the lower portion of
the trench.
5. Sediment thickness will increase with time as large and heavy particles fall from
suspension. Soundings for sediment thickness measurement should therefore be
performed immediately prior to allowing placement of backfill over the excavation
bottom.
. Overly viscous slurry is capable of suspending coarse sand and gravel particles or
soil clods and may build up at the bottom of the slurry column. Chemical attack can
both thin and thicken slurry. Slowed weight penetration rates and poor
reproducibility of flat weight soundings are indicators of a thickened slurry
condition. In cases of extreme slurry thickening the fiat weight may rest on the
surface of the submerged slurry. Thick slurry can be sampled with the use of the
fiat weight.
7. Effective removal of sediment from the bottom of a slurry filled trench using a
backhoe is bucket dependent. Bucket shapes, teeth length, teeth spacing and the
use of side cutters, ripper teeth, etc. can influence the efficiency of material removal
and opportunity for sediment generation. Typically, buckets with long and few
teeth cannot retrieve sediment, but instead create sediment. Short teeth and flat
cutting edges promote sediment retention for removal. Holes cut into the back of
the buckets to drain slurry increase slurry sand content and can generate large
amounts of sediment. These observations were made by repeated use of the flat
weight under various conditions.
8. Overfilled backhoe buckets can drop debris as they are raised through the slurry
column. This can be particularly worrisome when the bucket is operated above the
backfill slope when raised. Consider use of an end stop pipe to maintain a rigid
separation between the excavation and backf'lll ends of trench excavations.
Instrumentation to Monitor Landfills
Steve B. Taylor, 1 Chris C. White, 2 and Ron D. Barker ~
A Comparison of Portable and Permanent Landfill Liner Leak Detection

Systems
REFERENCE: Taylor, S. B., White, C. C., and Barker, R. D., "A Comparison of
Portable and Permanent Landfill Liner Leak Detection Systems," Field
Instrumentation for Soil and Rock, ASTM STP 1358, G. N. Durham and W. A. Marr,
ABSTRACT: Monitoring of the integrity of electrically non-conductive geomembrane

liners installed at waste sites using electrical geophysical techniques has been carried out for
a number of years using above-liner leak location surveys and, more recently, below-liner
monitoring systems. We compare the theoretical response of both types of survey to a hole
in a liner and then compare with measurements made in the field. The theoretical leak
response indicates that above-liner surveys are sensitive to leaks over a greater area, though
both responses result in comparable leak detectability. However, field data suggest that in
practice, measurements made on a sparse grid below the liner have the greater sensitivity to
certain leaks. This may be due to the differing leak geometries and background conditions
present above and below the liner.
The results indicate that a sparse below-liner monitoring grid, with its long-term monitoring
capabilities, combined with above-liner surveys to pinpoint leaks accurately offer a
successful approach to ensuring liner integrity throughout the lifetime of a lined waste site.
KEYWORDS: leak detection, geomembrane liners, landfill monitoring, geophysics
Introduction
Plastic lined impoundments are frequently used for the containment of waste, with the aim
of reducing the possibility of environmentaldamage due to escape of contaminants into the
~Researcher and SeniorLecturerrespectively,School of Earth Sciences,Universityof Birmingham,

Edgbaston, Birmingham,B15 2TT, UK.
2TechnicalDirector, TechnosLtd., WalfordManor, Baschurch, Shrewsbury,Shropshire, SY4 2ttH, UK.
157
surrounding land. Monitoring of the integrity of these liners is essential for the early
detection of leaks from lined waste sites, in order to reduce the extent of pollution of the
groundwater and the financial costs of remediation.
To this end, electrical geophysical techniques have been employed in the detection and
location of leaks in electrically non-conductive geomembrane liners installed in waste
impoundments, for a number of years. This monitoring normally takes the form of a
single Geomembrane Leak Location Survey (GLLS) carried out above the liner on a
dense half to two metre grid using portable equipment. Such surveys are regularly
employed in the UK as the final part of a Construction Quality Assurance (CQA) plan to
ensure the integrity of the liner prior to waste disposal. A theoretical analysis of the
above-liner electrical leak location method has been presented by Parra (1988) and Parra
and Owen (1988), where the response of a gradient array to a leak in a lined, liquid filled
impoundment is assessed. Adopting this array type, Darilek et al. (1989), reported on
electrical leak location surveys conducted at a number of lined sites in the United States
where an average of 26 leaks per hectare were detected. The first GLLS to be conducted
in the UK was in February 1993 (Laine and Moseley 1995) at Craigrnore Landfill Site
near Belfast, Northern Ireland. Ten holes were located in a 1.4 ha site with 0.3 m of soil
cover above the geomembrane liner.
More recently, permanently installed monitoring systems, comprising a grid of electrodes

located beneath the geomembrane liner, have been developed offering the possibility of
long term monitoring of the liner's integrity. Frangos (1994) presents a system
comprising electrodes on an 8 x 10 m grid located beneath a lined waste-disposal pond.
The system proved successful in the location of six holes in the liner.
This paper compares the theoretical responses to leaks for both above and below-liner
survey methodologies. Field data collected using a permanent below-liner monitoring
system (White and Barker 1997) are discussed with data using an above-liner GLLS also
conducted at the same site, to offer direct comparison of the capabilities and limitations
of each survey type.
In some cases, leaks detected with the permanent below-liner system, which can be
nominally located to within a 20 m grid square, have proved difficult to locate easily with
a GLLS. Knowledge of the presence of such holes provided by the permanently installed
system has required that the holes be located with a GLLS, but it has only been through
perseverance that all holes have eventually been located and repaired. This experience
suggests that some of these holes would have gone undetected with an above-liner GLLS
alone, and appears to indicate that the permanent below-liner system does provide a
greater sensitivity in practice.
The Methodology o f Leak Detection and Location

The method of leak location in lined impoundments relies on the electrically insulating
properties of the geomembrane liner onto which the waste is deposited. These liners,
constructed of high density polyethylene (HDPE) or polyvinyl chloride (PVC) have
TAYLOR ET AL. ON LEAK DETECTION SYSTEM 159
typical resistivities of between 109 ~ m and 1014 ~m. A current source located inside
the lined impoundment and a current sink located outside the impoundment set up a
potential difference across the geomembrane liner. The resulting distribution of potential
is sampled on a regular grid either within the impoundment, for manual GLLSs, or below
the liner in the case of permanent monitoring systems. In the ideal case where the site is
electrically isolated, i.e. where there are no paths for current to flow over the edges of the
liner and there are no holes in the liner, current flow is impeded by the high resistivity of
the liner material and the resulting potential distribution is relatively uniform (White and
Barker 1997).
If holes are present in the liner, conductive liquids passing through these holes will form
conducting paths along which current may flow. These conductive paths offer the route
of least resistance to current flow and a high current density will be present in the region
of the leak. The highest potential values from the resulting distribution of electrical
potential are therefore located in the vicinity of the leak. As the dimensions of the hole
are normally small compared with all other dimensions of the impoundment, the leak
effectively behaves as a point source of current.
Parra (1988) determined the above-liner electrical response of a leak in a lined

impoundment. He characterised the impoundment using a three layer model comprising
an above-liner layer of material with resistivity pw and thickness h, a thin highly resistive
liner of resistivity pl and thickness t, and a semi-infinite half-space beneath the liner of
resistivity ps (Figure 1). The formulation combines the response of this three layer
model with the response of a leak acting as a point source of current located in the thin
highly resistive liner layer.
To current Current Measurement Above-liner reference

sink - Is source +Is point potential electrode
Liner resistivity, Pl h Above-liner material resistivity, Pw

.......
Below-liner electrode ~....._~ D
~.."..... o.. ---o O o o O--- . . . . . . . c ~o.----- T
To below-liner remote Below-liner material resistivity, p.

potential electrode
Figure 1 - Schematic representation of impoundment indicating electrode

arrangement both above and below the liner.
Extending this to measurements made beneath the liner, the contribution to the below-
liner potential distribution due to the current flowing in the material above the liner is
negligible due to the presence of the highly resistive liner. The resulting distribution of
potential, d~, below the liner is therefore dominated by the leak response which
approximates to the response of a point current source, Is, located at the position of the
hole on the surface of a semi-infinite half-space of resistivity ps, simply
/~p~
~b- 2z~ r (1)
where r is the distance of the measurement point from the leak.
Parametric Study
Introduction
Most above-liner leak location surveys conducted in lined impoundments use the
gradient array configuration (Figure 2a). The response of such an array to a leak is
bipolar with the leak located at the amplitude midpoint of the resulting anomaly. One of
the problems associated with sampling the electrical potential using this type of electrode
array is that the potential difference measurements tend to be very small. This requires
accurate equipment and stable electrodes if these small potential differences are to be
accurately recorded. A second more significant problem is the fact that the gradient
tends to decrease to background levels within a very short distance of the leak source.
Consequently if the thickness of material covering the liner is too great, or the grid on
which measurements are made is too sparse, there is a likelihood that leaks will not be
detected at all. Indeed, the parametric study with such an array conducted by Parra and
Owen (1988), using the model described in Parra (1988), concludes that practical leak
detectability occurs only within 0.1 m of the leak with the dipole (P1 and P2 in Figure 2a)
close to the liner.
a) J
cO
P1 P2 + I., - Ix
~ ..........00
b)
< d ........ > l
P2 P1 + Is - Is
Figure 2 - Electrode array configurations, with - I s, the current sink, located

outside the impoundment, and the current source, +1s, located inside but distant
from the potential electrodes, P1 and P2. a ) Gradient array, with moving
potential electrode dipole, P1 and P2, with separation, d, o f typically I - 2 m.
b ) Pole-Pole array where only P1 moves and the separation, d, o f P1 and P2 is
maximised
Parra and Owen's study was directed specifically at liquid containment sites where the
distance at which measurements are made above the liner could be minimised. In lined
solid waste sites the distance to the liner is set by the thickness of the drainage layer
which is normally found to lie between be 0.3 m and 1.0 m. This automatically reduces
the level of detectability possible using above-liner leak detection methods.
The Pole-Pole Configuration

An alternative electrode arrangement to the gradient array is the pole-pole configuration
of electrodes (Figure 2b). With the potential distribution in the impoundment set up in
the same way as for the gradient array case, this configuration samples the electrical
potential distribution using a single moving electrode and compares the value to that
measured at a distant stationary electrode also in the impoundment. The potential
difference measurements made with this configuration are therefore much larger than
those measured with the gradient array. Secondly, the position of the leak is coincident
with the maximum amplitude of the anomaly curve associated with it and not at the
amplitude mid-point of the anomaly as with the gradient array. This makes identification
of leaks and their location during a survey more straight forward and does not necessarily
require the plotting of the data to determine the location of any leaks.
Using the equations presented in Parra (1988) and equation 1 above, a theoretical
comparison of the above-liner and below-liner responses using the pole-pole array can be
made. In this comparative study, an impoundment of 4 ha with a hole of radius 0.4 mm
is chosen. From Parra (1988), the amount of current passing through this hole in an
impoundment of this size is approximately 75 % of the current injected above the liner.
The current source electrode is located on the surface of the above-liner material of
resistivity pw =30 f2m, forty metres from the start of the profile line. For the data
calculated above the liner, the reference potential electrode is located forty metres along,
and offset forty metres from the profile line. For the below-liner data the reference
potential electrode is located two-hundred metres along the profile line. The liner is
characterised by a resistivity, pl = 1014 f~m and thickness, t = 2.5 mm, with the material
beneath the liner characterised by a resistivity, ps = 80 f~m. The above and below-liner
resistivities are chosen as they are representative of the site discussed in the case study.
Comparison o f Profile Shapes

Figure 3 shows a comparison of profiles over a leak located ten metres along the profile
line where the thickness of material above the liner is h = 0.5 m. The data are displayed
as normalised potential difference, that is the potential difference divided by the source
current, I s. Note that for a current source, I s, above the liner and return current beneath
the liner, potentials measured in the vicinity of a leak lead to a relative low above the
liner and a relative high below the liner. The data are calculated at the same distance
from the leak both above and below the liner. It is clear that the response to a
given leak and impoundment conditions is greater when measured above the liner
than below. It also appears from the shape o f the curves that the above-liner
response to the leak is laterally more extensive than the response below the liner,
with a larger gradient present in the below-liner response than in the above-liner
response. This evidence suggests that any given leak should be easier to detect
with an above-liner survey than with a below-liner survey.
35 9 -35
.-30 ~ [
/A
30.
-~:,:,~oL,.~ 9 -2s g ~"
- - A b o v e Liner
20
9 -10 ~ ~
9-5 e~
0 o
5 10 15 20
Distance along Profile ( m )
Figure 3 - Comparison of above and below-liner leak profile shapes calculated at

the same distance from a leak located 10 m along the profile.
Effect of Distance to Liner and Material Resistivity

Figure 4 shows the variation of the maximum normalised potential difference against
logarithmic resistivity calculated across a leak for varying material resistivity and
distance from the liner. The plot displays data calculated on the surface of the above-
liner material with varying thickness and resistivity, with the resistivity beneath the liner
kept at ps = 80 s Also displayed are data calculated at different depths below the liner
for differing below-liner resistivity values, with the above-liner material resistivity kept
at pw = 30 f2m. Due to the electrical properties of waste material inside the
impoundment it will in general have a lower resistivity than the ground beneath the liner,
and the different ranges over which the data have been collected reflect this.
Both sets of data indicate that the nearer to the liner a survey is conducted, the greater the
maximum leak response and hence the better the likelihood of leak detection. As the
distance from the liner increases, the maximum leak response decreases, suggesting a
limit on the distance from the liner that a survey can be conducted for successful leak
detection.
A greater maximum leak response is also observed for measurements made in more
resistive materials. This highlights the potential problem of very small leak anomaly
responses when conducting surveys over highly conductive waste or drainage materials.
Indeed comparison of the two data sets presented in Figure 4 indicates that for below-
liner materials that are more than five times as resistive as above-liner materials, the
maximum normalised potential difference response to a leak will be greater below the
liner than above.
60 -60
H=0.5 m D=0.5 m f
-50
40
~ A b o v e Liner D=l.0m -30 ~ ~

=-~
Response H = 1.0 m
S ~ 20. D~ 1.5 m -20 ~ ~[ ~ <
~=.~.
~ z ~ lo.
0
1'0 1'00 1000
Log [Material Resistivity (Ohm-m)]
Figure 4 - Variation of maximum normalised potential difference over a leak with

above and below-liner material resistivity and distance from the liner.
Variation of Profile with Lateral Offset

The range of profiles presented in Figures 5a and 5b are calculated across a single leak at
a distance of 0.5 m from the liner. The figures show how the theoretical leak response
varies with offset distance from the leak centre both above and below the liner. As
expected, the smaller the offset from the leak, the greater the maximum leak response
and the greater the likelihood of detection. For the data collected above the liner, it is
clear that for offsets of 0.25 m or less, there is little variation (less than 5 % from that at
zero offset) in the observed maximum leak response. This suggests a minimum line
separation for practical leak location surveys of approximately 0.5 m with no extra useful
information being gained from having survey lines any closer together. Similarly, at
greater distances from the liner, the minimum offset increases, with less than 5 %
variation in the maximum leak response observed at offsets of 0.45 m and 0.65 m for
distances above the liner of 1.0 m and 1.5 m respectively. At offset distances of
approximately 2 m or greater, the peak of the anomaly becomes much broader and less
well defined in both above and below liner cases.
a) b)
_ _ OItset = 0 . ~ m 20 __C~et=O.~m
. . . . . . O~fset = 0.1 m ...... Ol~t--0.1 m
-30. - - G(Tset ~ 0.25 m
"~ ~ 1 5 ,
~ .. Off.=,I~m
...
n
Gffs~=0,$m
- - O i ~ t =2 m
~,~ I~ " " " Ofl~=0.5m
I;,~ . . . . O l ~ r m
~;~ ~ --GIf~t=2m
' --- =
"~ ~., 10
.~
9 ~ -I0.
zN zN
-5.
r
0
0
.... ; .... 10 . . . . ; .... ~0'
.... ; .... ;o . . . . ; .... ;'
Distance along im~me (m) Distance a l o n g profile (In)
Figure 5 - Variation o f leak profile with offset from the centre o f a leak,
calculated at the same distance, O.5 m, a) above, and b) below the liner.
Variation o f Fractional Current Conducted Through Leaks

The most important parameter in determining the ability of a leak detection survey to
detect a specific leak is the amount of current conducted by that leak. The profiles
displayed in Figures 6a are calculated on the surface of a 0.5 m thick above-liner
drainage layer across a single leak conducting different fractions of the source current.
Figure 6b shows profiles calculated at a depth of 0.5 m beneath the liner across the same
leak. As the fractional current flowing through the leak is reduced, so is the maximum
normalised potential difference response. As the fraction of current flowing through the
leak drops below approximately 0.1 Is, the above-liner leak anomaly becomes
indistinguishable from the background potential. Below the liner the peak anomaly
amplitude also reduces sharply, though the anomaly shape is still observable even at very
small fractional currents.
a) b)
~]~IlIs=O'l
-45
-35
....
...... IIIs=OA5 i
I / I s =0.05
20
--Ills=O'l
. .
......
.
111s=0.25
. 1/Is = 0.05
. . . I/Is=0.5 . l/Is=03
. -
"= 25 --_ Ills=0.75 ~ --- l/Is=0.7S

, ~ --I/11=0.95
I I I s = 0.95 "~ t
-5 ~ .~176176 / ',.. "'" ........ "~
9 ,'":---- . . . . . . . . . . . . -:-:,',~" .
0 5 10 15 20 0 5 10 15 20
D i s t a n c e a l o n g Profile ( m ) Distance along Profile ( m )
Figure 6 - Variation of profile shape with fraction of current conducted through a

hole in the liner calculated at the same distance (0.5 m) a) above and b) below
the liner.
Variation of Below-liner Response with Grid Size

It is impractical to have a grid of electrodes located beneath the liner sampling the
potential distribution with the spatial frequency achievable with above-liner surveys.
It is therefore important to assess the limits of detectability of leaks using as sparse a
sampling grid as possible. The plots in Figure 7 show contoured below-liner data
calculated 0.5 m
o 50
biooi ii lc
.o_s0~
9 , ~ . .
/
\
0-2 _
, ~s'o . . . . . . :ii0 0 .... ~0 . . . . 1~0

X Position (m) X Position (m) X Position (m)
d ) lo0-
Contour Interval - 0.2 V o W A m p
i X Locationof Leak
/
vE /
.~ 50- SamplingPoint
O Figure 7 - Contoured below-liner response

ft.
>- 0.2 84184 to a leak located at (46,46)i Data calculated
0
- ~ 0~70~ - at a depth of O.5 m for different sample grids
X Position (m)
ata) 5m, b) 1Om, c) 2 0 m a n d d ) 30m
beneath the liner for a leak located at (46,46) and sampled on a 5, 10, 20 and 30 m grid.
The concentric pattern of contours indicate the presence of a leak, though the location of
the leak becomes less clear with the increasing size of the sampling grid. The presence
of a leak is most clearly seen in the data sampled on 5 m and 10 m grids and is still
apparent, though the maximum amplitude is smaller, in the data sampled on a 20 m grid.
At larger grid sizes, the anomaly is still apparent, though less identifiable, and the
location of the leak is less well defined.
Anticipated response implied by the theoretical analysis

The results of the parametric study indicate that for the given set of impoundment
characteristics, the above-liner leak response produces a leak anomaly that is larger than
the leak response produced below the liner. This holds for below-liner materials that
have resistivity values up to five times those of the material above the liner. For ratios
larger than this, the maximum leak response below the liner is larger than that above.
The largest anomaly responses are produced closest to the leak and therefore to achieve
the greatest level of leak detection, measurements should be made as near to the liner as
possible, echoing the findings of Parra and Owen (1988). Similarly, the further laterally
from the centre of a leak, the smaller the leak response. This is important in determining
the line separation for GLLS surveys. The results show that there is a minimum practical
line separation of 0.5 m for an above-liner material thickness of 0.5 m, increasing to
0.9 m and 1.3 m for thicknesses of 1.0 m and 1.5 m respectively. At these offsets, the
difference between the maximum leak anomaly and that observed at zero offset is less
than 5 %, with smaller line separations offering little, or no improvement on leak
detectability. With data collected on a sparse grid beneath the liner, it is still possible to
detect a leak even with a grid spacing as large as 30 m, though at this spacing the
maximum anomaly amplitude is small and the location of the leak is poorly defined.
By varying the amount of current conducted by a leak, the maximum anomaly response
is observed to change, becoming smaller for smaller conducted currents. To achieve the
greatest level of leak detectability in any leak location survey, the highest possible
current should flow through any leaks in the liner.
Comparison of Field Data
Introduction
The examples presented below compare data collected above the liner during a GLLS
with that collected using a below-liner monitoring grid of electrodes at a landfill site in
the UK. The site, Sandy Lane Landfill Site, operated by Cleanaway Ltd., is the first fully
operational landfill site in the UK to have a below-liner geophysical monitoring system
installed.
Brief Description of the Below-liner Monitoring System

The data presented in the following section were collected during evaluation of the
integrity of the geomembrane liner in Phase 1 of the Sandy Lane Landfill Site, (White
and Barker 1997). The present extent of the landfill covers an area of approximately
5.5 ha and is located in the base of an old sandstone quarry. The liner construction
comprises a 2.5 mm thick HDPE liner underlain by a 0.3 m thick layer of bentonite
enriched sand (BES). Directly beneath the BES layer are located monitoring electrodes
on a 20 m grid, with alternate east-west lines having electrodes at 10 m separation.
Presently, the grid comprises 242 below-liner electrodes, all of which are connected to an
electronically controlled switchbox installed within the landfill weighbridge building,
enabling any of the electrodes to be selected for current conduction or potential
measurement.
Data are collected using the pole-pole array with return current and reference potential
electrodes located over 200 m away from the electrode grid. A constant current supply
injects a 20 mA current through a stainless steel plate located in the 0.5 m thick sand
drainage layer at the base of the landfill. Potential differences are recorded between each
of the grid electrodes and the reference potential electrode.
To date three cells have been completed within the landfill with the geophysical
monitoring system installed during each construction phase. Its use as an additional
CQA survey has revealed 3 holes in Cell 1, 19 holes in Cell 2 and 6 holes in Cell 3. This
averages 5 holes per hectare, compared to 26 cited by Darilek et al. (1989). Giroud and
Peggs (1990) relate the number of holes per hectare to the effectiveness of the CQA
programme. The number of holes located in Cells 1 and 3 relate to good CQA whilst
Cell 2 relates to average CQA. However, this needs to be put in the context that only
four of the holes detected could be easily seen with the naked eye. In both Cells 2 and 3,
the holes were concentrated in distinct areas; Cell 2 along a berm mid way up the landfill
slope where the upper and lower halves of the liner were welded together; and Cell 3
where construction was most intense in the sump area.
Real Data Examples

Example 1 - - Figure 8a shows a typical example of contoured data collected using the
below-liner system when there is a single hole present in the liner. Displayed on the plot
are the locations of the electrodes in the monitoring grid and overlaid on the data is an
outline map of Phase 1 of the landfill. Towards the centre of the plot there is a single
sharp positive peak anomaly which causes a departure from the otherwise smooth
distribution of normalised potential difference. This peak is indicative of the presence of
a leak in this area.
With the data collected using the below-liner grid of electrodes, it is possible to estimate
the location of any leak to within a 20 m grid square. To improve on this location
accuracy it is possible to model the data to produce a best-fit estimate of the leak
location. This can normally improve the location accuracy to a 1 to 2 square metre area,
though multiple leaks can reduce this accuracy. To locate the source of the leak precisely
it is still necessary to conduct a GLLS over this localised area. The results of the GLLS
conducted over the estimated position of the leak is shown in Figure 8b.
a) Scale , ii -. b) 15
Survey Line : 1 2 3 4
1o-
[
;" 5
0
0 5
X position (m)
Contours o f Potential D i f f e r e n c e
395100 395200 E a s t i n g 395300
C o n t o u r Interval = 0.02 Volts
C o n t o u r s o f N o r m a l i s e d Potential D i f f e r e n c e
C o n t o u r I n t e r v a l = 0.05 V o l t s / A m p
c)
-~9
"~ ~ -07
-0.65
g~ -- Model Potential I~ffetence - Line 4
-0.6 , , ., . . . . . . . . . . . .
0 5 10 15
Distance along Profile (m)
Figure 8 - a) Data collected using the below-liner monitoring grid of electrodes.

Data indicate the presence of a single leak. b) Above-liner survey results over
the leak clearly show its location, c) Comparison of model fit to the data
measured along survey line 4.
The survey was conducted on the surface of the 0.5 m thick drainage layer o f sand
present in the base of the landfill on a 1 m grid using a constant 250 V supply. The data
clearly show a single localised sharp peak anomaly indicating the location of the leak.
Also observed is a sharp drop in potential between survey Line 1 and Line 2. This is
caused by the presence of a gravel drain running parallel to, and located between these
survey lines. The leak was found to be two adjacent 5 mm knife cuts in the basal liner.
Using the survey parameters and resistivity values measured at the site of 30 f2m for the
above-liner sand and 80 f)m for the sandstone below the liner, the model of Parra (1988)
was used to reproduce the observed data and a comparison is shown in Figure 8c. This
example shows that both above and below-liner methods are sensitive to very small leaks
and that the model is able to reproduce the measured data with a high degree of accuracy.
Example 2 - Figure 9a shows another example o f data collected using the below-
liner system. Located in the north o f the site are a number of leak anomalies.
Concentrating on the largest anomaly, anomaly A (Figure 9b) this anomaly is
observed to have a maximum amplitude o f 0.32 Volt/Amp (V/A).
a)
el.
Electrode Location 395180 Easting 395200 395220

3951100~ ~3951200 Easting 395300 Contours of Normalised Potential Difference
Contour Interval = 0.01 Volt/Amp
A
c) 20
15
! 84
d)
g
t- e-
O
10 0
O 0
a. O.
>- >.
0 X Position (m) 10 0 X Position (m) 10

Contours of Potential Difference g Location of Leak
Contour Interval = 0.01 Volt Measurement Point
Figure 9 - a) Data collected using the below-liner monitoring grid of electrodes

indicating the presence of a number of leaks, b) Expanded view of the Northwest
corner of data in Figure 9a showing the area of the above-liner survey, c)
Above-liner survey data before and d) after background removal
The results of the GLLS conducted over the estimated position of the leak is
shown in Figure 9c. Once again, the survey was conducted on the surface of the
0.5 m thick drainage layer of sand, In this case the data set shows no obvious sharp peak
anomalies and is dominated by a strong gradient running approximately east-west. Even
with a best fit estimate to this background gradient removed (Figure 9d), the data do not
clearly indicate the presence of a sharp anomaly of the kind predicted in theory or of the
type observed with the data collected using the below-liner grid of electrodes.
Knowledge gained by the below-liner system that a hole was present necessitated
locating the other holes first, in order to create a more identifiable anomaly at this precise
location.
Further GLLSs were conducted to locate other leaks observed to be present in the data
shown in Figure 9a. These surveys did establish clearly identifiable anomalies similar to
the anomaly presented in Example 1. All the holes associated with these anomalies were
repaired, the above-liner survey repeated over the remaining leak and the hole, a pin-hole
sized split on a weld joint, was finally located.
Discussion
Though most above-liner GLLSs have proved successful in the swift and accurate
location of leaks detected with the below-liner system, there have been a number of
occasions, similar to the one presented in Example 2, where leak location, although
always successful, has proved difficult and time consuming. These results highlight the
fact that above-liner GLLSs may not be sensitive enough to detect leaks in certain
situations and there is a possibility when conducting a GLLS that some leaks may go
undetected.
Example 1 has shown that the model fit to certain leak data is very accurate. However,
other findings (Example 2) also indicate a disparity between the results of some above-
liner surveys and the theoretical response, which may be due to the assumptions made in
the development of the theory. Primarily, the observed maximum amplitude response
over a leak is often much less than that predicted in theory for a hole and impoundment
of given size. The model predicts the percentage of current flow through a circular hole
of given diameter. However, in reality holes are very rarely circular and tend to be very
thin knife cuts, pin holes or splits on welds. In such circumstances, the level of
connectivity between the material inside and outside the impoundment may result in less
current being able to flow through the hole than predicted in theory.
As the amount of current available to flow through any hole is such an important
parameter, it is important that the lined impoundment be electrically isolated when leak
location surveys are performed. If the material inside the impoundment is not adequately
electrically isolated from the material outside the impoundment, current will flow out of
the impoundment at the places of inadequate isolation. This may be caused by sand or
soil lying over the edge of the liner, or due to conduction along water filled pipes, metal
guide ropes, or earth wires leading out of the impoundment. Any low resistance
pathways such as these will significantly reduce the current densities in the vicinity of
leaks and as a consequence reduce the sensitivity of any leak location survey. A GLLS
conducted in these circumstances is likely to produce a false negative.
Masking of holes can occur if there are a number of leaks present in the liner, as the
current passing through each hole will be reduced and, moreover, current will
preferentially flow through the leaks offering the least electrical resistance. Hence as a
consequence, this can preclude detection of smaller leaks in the presence of much larger
ones. A further reason for the departure of the measured response from that predicted in
theory is the assumption that the impoundment can be approximated by a model
consisting of three infinite homogeneous horizontal layers. In the base of a landfill when
no waste is present, this model is fairly representative. However, on the slopes of the
landfill, or when waste is present, these assumptions are less valid due to the presence of
topography and heterogeneity in the waste complicating the potential distribution.
The above comments apply equally well to below-liner surveys as they do to above-liner
surveys. However, there are still important differences in the results of the two types of
surveys which should be considered.
With the below-liner survey it is possible to locate the remote current and potential
electrodes at a significant distance from the measuring grid. This is advantageous as it
removes the strong background gradient that is observed to be present in the above-liner
pole-pole data. Location of the potential reference electrode well away from the grid also
precludes the random chance that it is located near to a leak and therefore in the region of
a steep gradient. This is a possibility with the above-liner survey method, though this can
be negated with good field practice. A further advantage of the below-liner system is that
the current flows through a larger volume of ground than in the above-liner case and the
resulting potential distribution is less affected by localised heterogeneities in the ground.
Though prior to landfilling the drainage layer above the liner should be relatively
homogeneous and the potential distribution relatively uniform, once landfilling has
begun, the heterogeneity of the waste can result in significant localised variation in the
potential. Also, as has been shown in example 1, the presence of gravel drains, gravelled
sump areas and bunds 3 may complicate normally uninterrupted current flow and lead to
complicated gradients.
Another factor that may explain why the results of below-liner surveys are more sensitive
to certain leaks than above-liner surveys is the leak geometry. During construction, the
external side of the geomembrane liner is laid directly in contact with the secondary BES
liner or engineered base, with little overlap and no repair taking place on this side of the
liner. All flaps, secondary cuts, welds, patches and repairs are located on the upper side
of the liner, onto which the drainage cover is laid. In some circumstances, such as T-
weld junctions or overlapping patches, a very complicated and tortuous path for current
flow may be set up. A small hole present in the basal liner underneath a weld flap, for
example, will allow current to flow from above to below the liner. When 'observed'
from below the liner, the current flows into the material below the liner through this
small hole which acts as a point source of current. However, above the liner the hole is
not in direct contact with the drainage material, but located beneath an electrically
insulating flap. Current is dispersed between the flap and the basal liner and only flows
into the above-liner material at the edge of this flap. The area over which the current is
dispersed into the above-liner material is now much greater than for the original hole, and
the current density is reduced resulting in a much smaller anomaly. When an above-liner
survey is conducted over such a leak, the leak may not be detected, or if detected, the
broad, small amplitude anomaly may not be recognised as a that due to a leak.
Other benefits afforded by below-liner systems include their ability to detect leaks in
regions that would normally not be surveyed using GLLS's. For example, partially
constructed or removed bunds. If current can enter the bund and holes are located
beneath the bund they may be detected using a below-liner monitoring system.
By far the biggest benefit of below-liner systems is their ability to continue to monitor the
integrity of the geomembrane liner throughout the lifetime of the landfill, and to continue
monitoring post closure. Though repair of the liner may become impractical and
prohibitively expensive for leaks that develop late on in the life of the landfill, such
systems can be used to monitor the development of leachate migration in the subsurface.
3 Bund - interceU boundaries usually formed by sand enclosed by a sheet of membrane above the
basal membrane and should by inference be electrically isolated
This extra knowledge offers advantages in the implementation of remediation systems as

these can be targeted more cost effectively.
Conclusion
A comparison of the theoretical leak response calculated both above and below the liner
suggest both offer comparable leak detectability, though the theoretical above-liner leak
response produces a laterally more extensive anomaly than the below-liner response and
an anomaly that is n ot always observed in practice. The reason for the departure of the
observed above-liner response from that predicted in theory is thought to be due to a
smaller than predicted current flowing through the leaks in the liner. Field data collected
over a number of leaks both above and below the liner have indicated that the below-liner
system is more sensitive to certain leaks. Possible explanations, including the differing
leak geometries and background conditions present above and below the liner, have been
discussed.
The use of the GLLS for leak detection is a successful and proved method. However,
from our experience, small leaks that have been detected with the below-liner monitoring
system have sometimes proved difficult to locate with a GLLS and without the below-
liner system may have gone undetected. The use of a sparse below-liner monitoring grid,
with its long-term monitoring capabilities, combined with above-liner surveys to pin-
point leaks accurately offers a successful approach to leak detection and long term
monitoring of the liner's integrity.
Acknowledgements
The authors would like to thank Cleanaway Ltd. for allowing us to present the data in this
paper and Professor D. Griffiths for his comments and deliberations.
References
Anderson, W. L., 1979, "Numerical Integration of Related Hankel Transforms of Orders
0 and 1 by Adaptive Digital Filtering," Geophysics, Vol. 44, pp. 1287-1305.
Darilek, G. T., Laine, D. L., Parra, J. O., 1989, "The Electrical Leak Location Method for
Geomembrane Liners : Development and Applications," Industrial Fabrics Association
InternationalGeosynthetics '89 Conference, San Diego, CA, pp. 456-466.
Frangos, W., 1994, "Electrical Detection and Monitoring of Leaks in Lined Waste
Disposal Ponds," Proceeds of the Symposium on the Application of Geophysics to
Engineering and Environment, 1994, Englefield, USA, Vol. 2, pp. 1073-1082.
Giroud, J. P., and Peggs, I. D., 1990, "Geomembrane Construction Quality Assurance,"
In : Waste Management Systems: Construction, Regulation and Performance.
Geotechnical Special Publication No 26, ASCE, pp. 190-225.
Laine, D. L., Mosley, N. G., 1995, "Leak Location Survey of a Soil Covered
Geomembrane at a Landfill Site in the UK," In : Waste Disposal by Landfill, A. A.
Balkema, Rotterdam, pp, 151-156.
Parra, J. O., 1988, "Electrical Response of a Leak in a Geomembrane Liner,"
Geophysics, Vol. 53, pp. 1445-1452.
Parra, J. O., Owen, T. E., 1988, "Model Studies of Electrical Leak Detection Surveys In
Geomembrane-LinedImpoundments," Geophysics, Vo 1.53, pp. 1453- 1458.
White, C. C., Barker, R. D., 1997, "Electrical Leak Detection System for Landfill Liners:
A Case History," Ground Water Monitoring & Remediation, Vol. 27, No 3, pp 153-159.
Michael J. Byle, ~Monica L. McCullough, 2 Rodney Alexander, 3 N.C. Vasuki 4 and James
A. Langers
Instrumentation of Dredge Spoil for Landfill Construction
REFERENCE: Byle, M.J., McCullough, M.L. Alexander, R.,Vasuki, N.C., and Langer,
J.A., "Instrumentation of Dredge Spoil for Landfill Construction," Field
Instrumentation for Soil andRock, ASTMSTP 1358, G.N. Durhana and W. A. Mart, Eds.,
American Society for Testing and Materials, West Conshohocken, PA, 1999.
ABSTRACT: The Delaware Solid Waste Authority's Northern Solid Waste

Management Center is located outside of Wilmington Delaware at Cherry Island, a
former dredge disposal site. Dredge spoils, of very low permeability, range in depths up
to 30 m (100 feet) which form a natural "liner" and the foundation for the 140 ha (350-
acre) municipal solid waste landfill. The soils beneath the landfill have been extensively
instrumented to measure pore pressure, settlement and deflections, using inclinometer
casings, standpipe piezometers, vibrating wire piezometers, pneumatic piezometers,
settlement plates, liquid settlement gages, total pressure cells and thermistors. The nature
of the existing waste and anticipated settlements (up to 6 m (19 feet)) have required some
unique installation details. The instrumentation data has been integral in planning the
landfilling sequence to maintain perimeter slope stability and has provided key
geotechnical parameters needed for operation and construction of the landfill. The
performance of the instrumentation and monitoring results are discussed.
KEYWORDS: instrumentation, piezometers, total pressure cells, landfills,

inclinometers, thermistors, settlement, municipal solid waste
Background
In 1984 the Delaware Solid Waste Authority (DSWA) acquired 140+ ha (350+ acres)
of land at the U.S. Army Corps of Engineers (USACE) Cherry Island dredge disposal site
from private landowners and the City of Wilmington. DSWA also initiated the development
1Region 3 Geotechnical Practice Manager, Gannett Fleming, Inc., 650 Park Avenue, King
of Prussia, PA 19406.
2project Engineer, Gannett Fleming, Inc., 650 Park Avenue, King of Prussia, PA 19406.
3Senior Engineer, Delaware Solid Waste Authority, 1128 S. Bradford Street, Dover, DE
19903.
4Chief Executive Officer, Delaware Solid Waste Authority, 1128 S. Bradford Street,
Dover, DE 19903.
5Vice President, Gannett Fleming, Inc., 207 Senate Avenue, Camp Hill, PA 17011.
175
of a complex memorandum of understanding between the USACE, Delaware Department

of Natural Resources and Environmental Control (DNREC), and DSWA under which
USACE agreed to release portions of the dredge disposal site for DSWA to use as the new
landfill site for New Castle County.
The Cherry Island Landfill (CIL) site is bordered by 1-495 on the west side, the City
of Wilmington regional wastewater treatment plant on the north side, the Delaware River on
the east side, and the Christina River on the south side. The site offers a suitable location
because of the ready access, available land area and lack of nearby sensitive neighbors.
However, the site is very challenging from a geotechnical viewpoint because of the depth and
nature of existing unconsolidated sediments.
The CIL design was based on five phases of construction starting with Phase I in
1984. The location for Phase I was selected on reasonable access from 1-495 and the least
depth of unconsolidated sediments. DSWA planned a sequential construction program
assuming normal growth rate of solid waste generation. However, the closure of several
landfills previously operating in New Jersey and Pennsylvania resulted in a large increase
in the commercial and industrial waste being delivered to DSWA landfills. Consequently,
the planned phasing construction of the ceils had to be accelerated to accommodate the
substantial increase in solid waste delivery. This had a significant effect on the performance
and design requirements for the latter phases of the CIL.
Site Conditions
The CIL is situated at the confluence of the Delaware and Christina Rivers and
comprises approximately 138 ha (342 acres) of land divided into five filling sections, or
Phases. A wastewater treatment plant is located to the north and an active dredge spoil
disposal area is located to the west. The active landfill areas consist of Phases III, IV and V
which are approximately 18, 20, and 24 ha (45, 50, and 60 acres), respectively. All three
active phases are being filled concurrently to accommodate required consolidation time
between lifts in each phase. Phases I, IA, and II are no longer receiving municipal solid
waste. They are used for stockpiles, asbestos disposal, and haul roads. The five phases are
located on dredge spoils recently deposited by the USACE.
During the pre-design investigations, the dredge material was so soft that tripod drill
rigs on plywood rafts and low contact pressure all terrain vehicle drill rigs were necessary
for the geotechnical investigations for each phase. The subsurface profile consisted of three
primary horizons: dredge spoils (beneath the landfill interior and in the dikes), dense sands
of the Pleistocene Columbia formation and very stiff silts and clays of the cretaceous
Potomac Formation.
The dredge spoils typically classify as silt (ML), and elastic silt (MH). Standard
penetration test (SPT) N-values range from weight of rods to 4 within the interior dredge
spoils and from 2 to 7 blows per foot (bpf) in the dikes. Moisture content of dredge spoil
samples is typically greater than 50%. Undrained shear strengths from unconfined
compression tests and unconsolidated undrained triaxial tests were typically very low,
generally 0.4 to 1.9 kPa (100 to 400 psi). The dredge spoil thickness generally increases
from a low of 8 m (25 It) in the northwest comer of the site to a maximum of 35 m (114 ft)
near the confluence of the Christina and Delaware Rivers at the southeast comer of Phase V.
BYLEETAL. ON DREDGESPOIL 177
The soils
used in the
construction of the
perimeter dikes
consisted
materials excavated
of
WILMINGTON WWTP
[ ]
from the desiccated [ ]
crust material
removed from the n-
LU
i n t e r i o r dredge _>
n,
spoils. This crust ILl
material, having
been compacted at a
'i .....-i 'Ui
~l~illlllilllllllllllllllfl~]lllllnlllllllll~Phase
IV~ ~ I~
L ~ ~ nf
lower m o i s t u r e
5
LU
U S ARMY ~ / = El
content, is typically coiS.PS OF ~_ / I=-
of higher strength ENGINEERS ~ , I V ~ _==
DREDGE DISPOSAL-~ Phase V f f J ==
than the AREA _~ / / ~
hydraulically placed =E-I DREDGEDISCiC~GE / / . ~ !
interior materials. /
_J
The dike soils have
CHRISTINA RIVER
gradations and
Atterberg limits
identical to the
i n t e r i o r dredge Figure 1 Dredge Placement Pattern in Phases III, IV and V
spoils.
The particle sizes of the dredged materials vary across the site, since the dredge
disposal area was designed to allow the solids to settle out of the slurry as it flowed from the
dredge discharge point at one end of the disposal sitel around a baffle dike to an outlet weir
at the other. Then, after the solids settled out, the water was discharged through a weir
structure to the Delaware River. The dredge spoils typically have more than 85% passing
the No. 200 sieve and permeability in the range of 1 x 10.6 to 5 x 10-s cm/sec. The area of
Phases I and II consisted of upland dredge disposal cells with reduced thickness of dredge
spoil: The dredge spoil was normally to slightly under consolidated at the time of landfill
construction. Some natural alluvial deposits underlie the dredge spoil, but are so similar as
to be indistinguishable from them.
The Columbia Formation which underlies the dredge spoil consists of dense coastal
plain sands and gravels with interbeds of silty sand, silty clay, and clayey silt from 3 to 20
m (10 to 66 ft) thick. SPT values typically range from 35 to 100 bpf. The Columbia
Formation is an aquifer of limited quality with local hydraulic gradients toward the southeast
confluence of the Christina and Delaware Rivers. The deeper Potomac Formation is a lower
confining layer for the Columbia Formation. The soils consist of very stiff silts and clays
interbedded with silty sand and gravel.
The hydrogeology of the site is somewhat complex. A water table, perched in the
dredge spoil, actually has an upward gradient due to the under-consolidated nature of the
dredge spoils. The hydraulic head increases with depth through the near surface dredge spoil
to the middle of the dredge material and then decreases toward the bottom where the dredge
materials/recent deposits meet the Columbia Formation. There are granular layers within the
recent deposits, isolated from the Columbia Formation, that locally distort the hydraulic
head.
Landfill Design
The landfill design and operation are controlled by geotechnical issues. The great
thickness of low permeability dredge spoils creates an excellent barrier to protect the
groundwater from landfill leachate. DNREC granted an exception to the Subtitle D
requirements for a composite bottom liner based on a demonstration of equivalent hydraulic
performance of the existing dredge spoils to a manufactured liner system. An additional
factor in the approval was the upward groundwater gradient induced by consolidation of the
dredge spoil that prevents infiltration ofleachate into the subsurface. The upward gradient
is expected to remain until consolidation is completed, approximately 50 to 70 years. At that
time the soils will have an average permeability less than 1 x 10-7 cm/sec.
Settlement of the dredge spoils under the weight of the municipal waste was
estimated to be up to 6 m (19 ft). The leachate collection system beneath the landfill has
been designed to accommodate this settlement while maintaining adequate slopes for
positive drainage.
The low strength of the dredge spoils controls the height and steepness of slopes that
can be safely constructed. With a high drained angle of internal friction, on the order of 38 ~
the dredge materials will be capable of supporting the proposed 3H: 1V slopes when fully
drained. However, consolidation of the dredge materials must take place for this drained
strength to be realized. This consolidation is time dependent and cannot be accelerated,
because methods that could be used to improve drainage of the dredge spoils would
compromise its effectiveness as a liner.
Therefore, the landfill filling operation requires staged filling among the phases so
that the dredge spoils can consolidate under each lift before the next lift is placed. The
original designs of each phase projected the expected rate of consolidation and magnitude
of settlement. Because of the very low strength of the dredge spoils, it is not possible to
construct slopes with high factors of safety. Slope stability analyses were completed using
the Modified Bishop Method of Slices which resulted in factors of safety on the order of 1.2
for the design slopes under each successive lift of waste.
Because of the relatively small factor of safety, consequences of failure, very slow
rate of drainage, uncertainties associated with predicting time rate of consolidation and
strength gain in these materials, and the need to verify soil conditions prior to each
successive lift of waste, the design included instrumentation to verify the design assumptions
and improve the time rate correlations. This was considered essential to control the rate of
filling to maximize the available air space without over stressing the weak foundation soils.
The proposed final landfill height for all phases is 24 m (80 ft). Because of the low
dredge spoil strength, geogrids were included at the bottom of the waste in Phases II and III
to reinforce the base soils against basal failure. The geogrids were embedded in the sand
drainage layer at the top of the dredge spoils. Phases IV and V incorporated vertical strip
drains to accelerate consolidation and the dissipation of pore pressures in the perimeter dikes
for improved global stability.
BYLE ET AL. ON DREDGE SPOIL 179
Instrumentation Systems Design
The purpose of the instrumentation systems at Cherry Island is to monitor the

behavior of the perimeter and the interior soils to determine if they are behaving as required
and as expected. The primary instrumentation focus areas for each phase of the landfill are
slope stability of the berms and consolidation of the underlying interior dredge spoils. At
present the total instrumentation consists of 109 standpipe piezometers, 5 liquid settlement
gages, 33 pneumatic piezometers, 5 vibrating wire piezometers, 43 settlement plates, 45
inclinometers, 32 total pressure cells, and 28 thermistors (Table 1). The manufacturers and
model numbers of the discussed instrumentation are included in the following sections,
where known. A typical cross-section of the landfill is shown in Figure 2.
Table 1
I n s t r u m e n t a t i o n Summa1T
Instrument Phase I Phase II I Phase III I Phase IV Phase V
Type
Casagrande 14(Perimeter) 17(Perimeter) 24(Perimeter)
Piezometers
Well-Screen 17 (Perimeter) 17(Perimeter)
Piezometers 3 (Interior) 6 (Interior)
Pneumatic 15 (Interior) 18 (Interior)
Piezometers
Vibrating Wire 5 (Interior)
Piezometers
Inclinometer 6 (Perimeter) 6 (Perimeter) 12 (Perimeter) 8 (Perimeter) 13 (Perimeter)
Stick-Up 7 (Interior) 9 (Interior) 9 (Interior) 6 (Interior) 10 (Interior)
Settlement Plate
Settlement 0 (Interior)
Transducer 5 abandoned]
Total Pressure 10 (Interior) 10 (Interior) 12 (Interior)
Cell
Thermistor 8 (Interior) 8 (Interior) 12 (Interior)
Pneumatic Piezometers
Pneumatic piezometers were included in the interior of Phases III and V to verify the
direction of groundwater gradients for permit compliance and to measure consolidation
related pore pressures. The piezometers were planned at elevations of the upper, middle and
lower sections of dredge spoil. The pneumatic piezometers specified are two-tube normally
closed piezometers read under no flow conditions with 50 micron stainless steel filters and
synthetic rubber diaphragms (Slope Indicator Model 514178 in Phase III and RST Model P-
102 in Phase V). The piezometer tips are placed in 0.9 to 1.2 m (3 to 4 It) of well sand, with
a bentonite seal and bentonite grout. Small sand layers were used to keep the monitoring
zones small.
The Phase III piezometers were installed using hollow stem augers. The Phase V
design used the same type of piezometers but used the direct push method of installation.
The piezometers are enclosed in a perforated stainless steel housing that is pushed into the
subsoil without a borehole or below the bottom of a casing.
The Phase III design routed the tubing from the boreholes, through the leachate
collection sand layer to the remote readout locations on the perimeter berms. In Phase III,
the p n e u m a t i c
tubing consists of total pressure cells
two 5 mm (3/16 settlement plate ~
inch) outer- liquid settlement \ \ Surface
diameter Nylon 11 transducer . - ' ,~' ,~~t - ~ - Readout Station
tubes inside a inclinometer~,~ / / " ~ T - ~ - - - ] JL. . . . . . . . . . . . . . . . . . . .
polyethylene sheath ;tandpipe ~'~ .,,,.'/ "[ ~ / solid waste
with fittings of 316 9 i e z o m ete r s ~ ~']1-~<,_
"_. . . . .= .[. . . . ~ . . . . . . . . . . . . . . . . . . . . . .
stainless steel. --~'-~-"-~--1 ~I[ dredge spoils
These materials
were selected for
resistance crushing
/
/J sands
and chemical Vibrating wire //
degradation. The and pnuematic
distance between piezometers
the instruments and Figure 2 Typical Instrumentation Cross Section
monitoring stations
range from
approximately 60 to 200 m (200 to 700 feet).
The Phase V design called for high density polyethylene (HDPE) pneumatic tubing
routed through a 152 mm (6-inch) diameter HDPE conduit, buried in the leachate collection
zone. Flexible rubber hose was used to protect the tubing where sharp bends occurred.
, Several pneumatic piezometers were installed after the landfill was in operation to
augment the original instruments (Rocktest Model FPC-2). These piezometers were installed
through about 6 m ( 20 ft) of municipal solid waste (MSW) via boreholes drilled through the
MSW in the interior of Phase III. The tubing for these instruments could not be trenched
through leachate collection zone so these piezometers are read from temporary movable
monitoring stations in the landfill interior. The tubing is routed vertically through the waste
within telescoping 76 and 102 mm (3 and 4-inch) Schedule 40 galvanized steel casing. A
bentonite slurry is placed inside the telescoping casing to provide lubrication and extra
protection for the pneumatic tubing.
Vibrating Wire Piezometers
Vibrating wire piezometers were used in the interior of Phase IV. These piezometers
include a 50 micron sintered stainless steel filter and a stainless steel body (Irad Gage Model
PWS). The vibrating wire piezometers were installed in a 50 mm (2 in) PVC casing with a
1.5 m (5 ft) slotted well screen. The well screen was set in a 1.8 m (6 ft) filter sand pack.
The piezometers cable was routed vertically from the dredge layer to the landfill
surface through the MSW inside a steel casing. Additional same diameter steel casing is
installed as the MSW height increases. These instruments are monitored from the landfill
surface. The frequency and temperature of the instruments are monitored with an IRAD
Gage Vibrating wire readout unit.
Standpipe Piezometers
Conventional standpipe piezometers are used in the perimeter berms surrounding all
five phases of the landfill. The Phase designer's preferences resulted in two general types
o f standpipe piezometers. In Phases I, II and III, the piezometers are Casagrande
piezometers (from various manufacturers) containing porous tips connected to 19 mm (0.75-
inch) PVC riser pipes (Casagrande 1949). The tips were installed in 900 to 1500 mm (three
to five feet) of sand and with a bentonite seal and bentonite grout to the surface.
In Phases IV and V, the standpipe piezometers are well-screen type piezometers
containing 3 m (10-foot) lengths of 51 mm (2-inch) diameter PVC well screen with a 0.2 mm
(0.006-inch) slot size and 51 mm (2 in) PVC casing. These well screens were backfilled with
4.6 m (15 feet) of filter sand and hence have a longer monitoring zone. Phases IV and V also
contain interior standpipe piezometers constructed in the same manner. However these
interior piezometers include a 400 mm (16 in) HDPE outer protective casing through the
MSW. This interior placement was intended to verify initial vertical gradients for the first
two lifts of waste only.
Stick- Up Settlement Plates
Each phase o f the landfill contains stick-up settlement plates installed during cell
construction. The plates in Phase I, II and III are 610 x 610 x 6 mm (24" x 24" x 1/4")
structural steel plate with a 19 mm (3/4 inch) threaded steel riser pipe attached to the plate
center. The plates were placed on a prepared subgrade of sand constructed on the dredge
spoil surface, prior to placement of the sand drainage layer. As additional lifts of MSW are
placed, extensions are added to the riser pipes. In Phase I and II, the riser pipe is protected
from the MSW with 51 mm (2-inch) outer-diameter threaded steel casing extended as MSW
heights increase. In Phase III, the riser pipe is protected by telescoping 51 and 76 mm (2
and 3 inch) steel pipe that is grouted with a bentonite slurry.
In Phases IV and V the plates consist of 1219 x 1219 x 6 mm (48 x 48 x 0.25 inch)
structural steel plates with a 19 mm (3/4 inch) steel riser pipe attached to the plate center.
The plates are supported on a prepared sand base similar to the other phases. The riser pipes
are protected from the MSW with 70 mm (2.75 inch) outer-diameter ABS plastic
telescoping inclinometer casing.
Liquid Settlement Transducers
Five settlement plates in Phase III are attached to vertically aligned liquid settlement
transducers (LSTs) (Sinco Model 51483). Like the stick-up plates, these plates are supported
on prepared sand base placed directly on the dredge spoil surface during cell construction.
A transducer was attached to each plate and backfilled with sand. The transducer tubing was
routed through the leachate collection zone to readout stations on the perimeter berms. The
pneumatic instruments (piezometers and LSTs ) were constructed in clusters to minimize
trenching and share common readout stations. The tubing and fittings for these instruments
are similar to those described for the Phase III pneumatic piezometers. Unplasticized nylon
11 tubing with a polyethylene sheath was selected for low permeability to gas and water
(Dunnicliff 1993).
One pair of the transducer tubing for the LSTs contains a deaired glycol solution that
is used to provide a fluid head for the measurement. The glycol solution is used to prevent
freezing. The remote transducer consists of a pneumatic piezometer, in a liquid filled cell,
connected by tubing to a reservoir outside the area of filling. The reservoir of glycol solution
is situated at a higher elevation than the transducer, on the perimeter berm. The elevation
of the glycol reservoir is surveyed and the elevation of the plate is calculated from the
differential head measured by the piezometers. The reservoirs are protected within metal
housings but must be refilled regularly because of evaporation. The piezometers in the
transducers are three tube, normally closed piezometers which are read under two-tube
normally closed conditions like the other pneumatic piezometers (per manufacturer's
instructions).
lnclinometers
Inclinometer casing was installed along the perimeter berms of all phases of the
landfill to monitor berm stability. Each inclinometer casing was installed to the bottom of
the dredge layer, approximately 15 to 34 m (50 to 110 feet), so that a complete profile of the
dredge movement could be developed.
In Phases I, II and IV, 70 mm (2.75 inch) O.D. ABS plastic standard inclinometer
casing was installed. In Phases III and V, 85 mm (3.34 inch) O.D. ABS plastic casing with
telescoping couplings was installed. The telescoping couplings allow for up to 150 mm (6
in) of settlement per coupling (Slope Indicator 1996). Couplings are spaced every 1.5 to 3.0
m (5 to 10 ft) allowing for a minimum shortening of approximately 5% to accommodate
consolidation settlement of the berms. The casings in all phases were grouted in place with
cement-bentonite grout, except for Phase IV where the casings were backfilled with gravel.
Surveys were conducted with standard inclinometer probes.
Total Pressure Cells and Thermistors
Total pressure cells (TPCs) were installed in Phases III, IV and V of the landfill to
measure the load applied by the MSW on the dredge spoils. Thermistors were included to
monitor temperature variations that could affect the TPC readings (Felio and Bauer 1986).
Retrofit Installation Phases 11land IV-- During the TPC design process for Phases
III and IV, MSW was already in place, to depths of approximately 3 to 6 m (10 to 20 feet).
The installation required excavation through the existing MSW. Four clusters of five TPCs
and four thermistors were installed to minimize excavation and provide redundancy.
Corrosion resistance was a key factor in the TPC design, since the equipment would be
exposed to landfill leachate.
In order to minimize the effect of point loads from the mixed materials in MSW,
large instrument bearing areas were preferred. The largest off-the-shelf TPC size was
approximately 305 mm (12-inches) diameter. This diameter was considered too small to
accurately monitor the pressure because of the potential for nearby interference from large
point loads and/or load arching. The

concept of clustering multiple TPCs
beneath a structural steel bearing plate TPC Thermistor
was considered to spread the load over
several TPCs. This option was ruled out
because of the corrosive environment of I "--/ I+
the steel, high cost, and installation
difficulty.
t i _ li
In lieu of the steel plate, a \+ . \ J /" ,/
geomembrane sandwich was developed. 7
This approach relies on load distribution
through two layers of sand and tensile
reinforcement by the membranes. The
reduced stiffness of this system results in
less concentration of landfill loadings than
HDPE Mernbranl ?
for a steel plate would induce, while
providing adequate protection of the 1 _j~onduit
instruments.
Five 305 mm (12 in) TPCs and
four thermistors were placed in a square
pattern in a 3.4 x 3.4 m (11 x 11 It) area. Figure 3 Phase III and IV Typical Total
The area was then covered with two layers Pressure Cell and Thermistor
of 60 mil HDPE geomembrane and sand. Layout
The instrumentation was aligned so that at
least 1305 mm (1-foot) spacing was
maintained between instruments with at least 305 mm (1-foot) geomembrane extends beyond
the outermost instruments. This confguration is illustrated in Figure 3.
A typical cross-section of Phase III and IV TPC installation is included as Figure 4.
This illustrates the designed system o f alternating layers of compacted sand backfill with two
3.4 x 3.4 m (11 x 11 ft) HDPE sheets. HDPE was the chosen geomembrane type because
of its likely resistance to chemical degradation. The geometry of the system was intended to
promote even beating across the 3.4 x 3.4 m (11 x 11 ft) area and to increase protection for
the instruments against puncture from objects in the MSW. Areas larger than 3.4 x 3.4 m
(11 x 11 ft) were considered and rejected in order to minimize the amount of excavation
through existing MSW.
Resistance to chemical degradation was an important factor when ordering the TPCs
and thermistors from the manufacturer. The chosen 305 mm (12-inch) diameter cells are
normally supplied with stainless steel bodies (RST Model TP-101-P-12). The cells and
thermistors were specially ordered with epoxy coating.
A unique protective conduit system was designed to carry the pneumatic tubing and
thermistor cable to readout locations. Trenching through the waste to the periphery of the
landfill was not considered practical. Instead, tubing was routed vertically to temporary
readout stations on the landfill surface. The protective conduit is designed to maximize
protection against settlement-induced stresses on the TPC/thermistor tubing system.
The conduit system is shown in Figures 4 and 5. The system was constructed of
Schedule 80 PVC for robustness and durability. The system consists of a horizontal PVC
J
/
/
Protective
I
\
Conduit //"'/"
x " \xGe~ ~
11
Figure 4 Phase III and IV Total Pressure Cells
conduit leading from the instrument locations to an

Expansion Assembly. The Expansion Assembly
isolates the instruments from settlement of the
dredge spoils below and compression of MSW
above. The Expansion Assembly is similar to a sink
1
trap in configuration, with the center U-shaped
section free to slide up or down relative to the
instrument side and waste side conduits. The
vertical sections consist of telescoping lengths of
127 mm and 152 mm (5-inch and 6-inch) diameter
PVC pipe, to the landfill surface. The telescoping Figure 5 Expansion Assembly
design reduces stress on the conduit from MSW
settlement. The PVC conduit was filled with a bentonite slurry for extra tubing protection.
The entire excavation was backfilled with MSW. The tubing is stored in protective
containers which serve as moveable readout stations on the landfill surface. As the MSW
height increases additional telescoping sections will be installed.
New Installation - P h a s e V - - In Phase V, the TPCs and then-nistors were installed
during cell construction before the first lift of MSW was placed. Again redundancy being
an important issue, six sets of two TPCs and two thermistors each were installed. The
pneumatic tubing and thermistor cable were routed in HDPE pipe embedded in the leachate
collection zone to remote readouts on the perimeter berms.
Two total pressure cells and two thermistors were placed at each interior cluster. The
TPCs were embedded in 300 mm (12 in) of sand with 900 mm (1.5 ft) of aggregate placed
above the sand. The tubing utilized the same HDPE piping as the pneumatic piezometers
to travel to the remote terminal. Twelve total pressure cells (RST Model TP- 10 l-P- 12) were
installed. The thermistors were also provided by RST. The tubing supplied for the project
was a twin 5 mm (3/16") OD HDPE tubing in a polyethylene sheath. Thermistor cables were
22 AWG in a 6.3 mm (0.25") diameter heavy duty, direct burial cable. Six terminal posts
were constructed of 100 mm (4") steel box tubing, 3 mm (1/8") wall and zinc plated.
Instrument Installation and Performance
In general, all of the instruments were installed as describe above. Several items of
interest were noted. The protective casings around instruments tend to settle; often the
instrument cap falls below the top of the riser pipe which prevents closing and locking the
cap. The protective casing around the Phase V interior standpipe piezometers consisted of
HDPE pipe that tapers so that each additional section of pipe reduces in diameter. The pipe
sections were so heavy that wood forms were required prior to landfilling around them.
Five pneumatic piezometers were recently placed in Phase III through existing MSW
to depths ranging from 1.5 to 18 m (5 to 60 ft) below the bottom of waste (top of dredge
spoil). This installation occurred after two lifts, about 6 m + (20' 40, o f MSW had been
placed. This required drilling through existing MSW to reach the dredge layer. Nested
installations were used to minimize the number ofboreholes required through MSW. The
pneumatic tubing for these installations was extended to the landfill surface inside the steel
casing backfilled with a bentonite slurry. The installation had to be suspended several times
to allow for venting of natural gas in excess of the lower explosive limit. The gas was from
decomposition of solid waste and organic matter in the dredge spoil. Continuous monitoring
with a combustible gas monitor was required throughout the installation.
Phase V interior pneumatic piezometers are push-in pneumatic piezometers. The
tubing was conducted through a flexible rubber hose that turned 90 degrees and inserted into
the minimum 150ram (6 in) SDR 15.5 HDPE protective conduit. The conduit was placed in
the leachate collection zone one foot above the subgrade. A protective layer rock gravel was
placed over the HDPE protective conduit before waste placement. This surface installation
required no trenching and was simple to construct, but left ridges of gravel across the landfill
floor that became obstacles to construction equipment.
Generally standpipe piezometers, both the porous tip and well screen types, are
performing satisfactorily. Sediment accumulates in the well screen piezometers and has to
be removed by flushing periodically. Several piezometers in Phase IV have experienced
such high pore pressures after waste placement, that they overflowed and required long
extensions to reach the hydrostatic level. Some of these require a lift to read and others have
been read with a pressure gage attached to the top of the casing. In the winter, the water can
freeze in the riser making the piezometers unreadable until a thaw occurs.
Some of the newer standpipe piezometers encountered gas pockets during their
installation. These piezometers overflow with effervescent water and cannot be read except
with a pressure gage. The gas pressure typically dissipates over time and the water level in
the casings stabilize.
Settlement stick-up plates are generally performing well. Any instrument that sticks
up through the landfill can be damaged by the waste placement process. Settlement plates
were being regularly damaged and required frequent repair until the landfill operator's
contract required him to pay for the replacement of instruments damaged by his operations.
Prior to waste being placed around the settlement plates, several of the riser pipes snapped
at the base and fell over due to wind load.
Some of the inclinometer casings in the early phases have had to be replaced because
of excessive deflection. Many of these inclinometers have sustained more than 1 m of
horizontal deflection. In general, the larger diameter casings have a longer life and tolerate
larger deflections.
The liquid settlement transducers installed in Phase III have failed to provide
consistent data. All five LSTs have failed despite repeated purging and de-airing. All of the
piezometers in the LSTs became unstable and would not give stabilized readings. The
problem appears to be with the instruments rather than the tubing, since the tubing is of the
same material and shares a common trench with the pneumatic piezometers which are
functioning quite well.
Of the ten total pressure cells installed in Phase IV, one may have been damaged by
the installation process and has not shown a stabilized reading. Another one initially
indicated a steady decrease in pressure inconsistent with the unchanged loading conditions,
but has since recovered with the placement of additional waste. This behavior appears to be
the result of load redistribution and demonstrates the need for multiple instruments when
measuring pressures under waste with relatively small instruments.
A few of the Phase IV interior standpipe piezometers had to be abandoned due to
tilting. They were placed in a triangle to prevent damage by landfill equipment working
between the riser pipes. The lack of compaction on the waste between the piezometers
resulted in the riser pipes settling inward towards each other to a point where the instruments
were not readable.
The thermistors indicate generally rising temperatures under the waste. The
temperature measurements appear to be affected by both seasonal changes and
biodegradation of the waste.
In Phase V, several standpipe piezometers were located adjacent to the pneumatic
piezometers and screened at similar elevations to correlate the two types of instruments. The
standpipe piezometer data correlated well with the pneumatic piezometer data. However,
potentiometric groundwater gradients consistently differ by approximately 3 m (10 ft)
between the standpipes and the pneumatics. This is likely due to the differences in the size
of the screened zone between the standpipe and pneumatic piezometers.
In order to preserve the pneumatic tubing in Phases III and V, the first lift of waste
material excluded bulky waste with large, heavy, or pointed objects. Extending the interior
pneumatic tubing, vibrating wire settlement plates and standpipe riser pipes through the
MSW interferes with the normal waste filling operation. Waste tends to settle around interior
instruments due to poor compaction creating areas where water ponds around the instrument
making monitoring difficult. Compacting around the interior instruments is a delicate task.
Findings
Landfill operations have had a substantial and profound influence over the
interpretation of instrumentation data. Early readings from the instruments did not agree
with the known waste surcharge on the dredge spoils. Record keeping by the facility
manager had to be improved to include information about the area of filling, location and
size of stockpiles, tonnage of waste received, tonnages of cover material used, etc. Only with
good filling and operations records can variations in the instrumentation data be attributed
to real life conditions.
The pore pressure data support the design assumption that strength gain via
consolidation due to the applied waste loadings will occur within the dredge spoil stratum.
The measured piezometric heads indicate that the excess pore water pressures develop
consistent with this consolidation process and dissipate as vertical flow in two directions
(Figure 6). The flow is upward into the landfill's leachate collection system within the upper
half of the dredge spoil stratum and downward into the underlying Columbia Formation
within the lower half of the dredge spoil stratum thickness. The instrumentation generally
indicates that excess pore pressures dissipate more quickly in the lower one quarter of the
dredge spoils than in the upper portions.
Pore pressure measurements, while accurate, have proven to be difficult to interpret.
Uncertainties regarding the surcharge loading, initial degree of consolidation in the dredge
spoils prior to waste placement and determination of the initial hydrostatic water table in the
dredge spoils made the estimation of future behavior from pore pressures alone impossible.
Pore pressure readings in Phase III after placement of the first lift of waste indicate that pore
pressures continued to increase when no additional waste was being placed.
It was essential to measure the load imposed by waste placement in order to gain an
understanding of the pore pressure data. The TPCs were installed for this reason. TPCs
cannot directly measure the overburden pressure because they are subject to the "inclusion
effect". That is, they, being stiffer than the surrounding soil attract loading. Reduction of the
TPC data required estimating this effect. This was accomplished by embedding the TPCs
in sand of known density and computing the adjustment by taking a reading after the three
feet of protective sand cover was in place. Weiler and Kulhawy (1982) recommend such in-
soil calibration of load cells. Based on this calibration the measured TPC pressures are 10%
greater than the overburden pressure. Temperature corrections are small and calibration
curves between TPCs show large scatter. Temperature con'ections were based on a statistical
mean of many calibrations. The TPC data indicate that the average waste density is 84 pcf.
This value represents the total unit weight of the waste layer including cover soil and
absorbed moisture at the time of the most recent measurement. This value is increasing with
time (Figure 7). A typical thermistor record is included in Figure 8.
Settlements measured from the settlement plates (Figure 9) are consistent with
calculated values for the first 3 to 9 m (10 to 30 ft) of waste. The data indicates wider scatter
and increasing settlements
with additional lifts. This
appears to be the result of Pneumatic Piezometer Pair at P3-104
distortion of the pipe 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
extensions as the waste
settles. Though the m
extensions are protected
from downdrag forces by
protective casings, it is a .
100 0
possible that s o m e
downdrag loading is
transmitted to the base
- 4 - P3-1 04-A!
50 0
plate, either by the ac.
- l - P3-104B
distorted extensions
coming into contact with O0 "~ H 4
the protective casing or 8/1~3 8/1/94 8/1/95 7/31/96 7~1/97 7~1N8
from the casing being
Date
forced down into contact
with the base plate. Figure 6
The inclinometers Total P r e s s u r e C e l l s

have shown that the perimeter Phase IV - W e s t C l u s t e r
dikes are deflecting outward 120.0
under the waste loading. The
100O
pattern o f movement is
indicative of general shear 80.0
with no defined shear plane.
60.0
The maximum inclinometer
deflections have been plotted 40.0
over time and coincide with
20.0
expected rates of
consolidation for the dredge O.0
11/1/96 1/30/97 4130197 7129197 10127197 1/25/98 4/25198
spoils. This data is consistent
with horizontal consolidation ~ Date
within the dikes. Cumulative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
inclinometer deflections up to Figure 7
I m (3 fl) o f horizontal
movement have been Thermistors
measured without evidence of Phase IV - W e s t Cluster
failure.
20
Implications for Design
The landfill has been 15 _. _- ;+TH4- [

in operation since the
opening o f the 33-acre Phase E tit --A- T H 4 - 7 ~ ! ~
I in October 1985. Filling in /
/ , i ~T TH4-8J
Phase I was completed in 10
11/1/96 1/30/97 4/30/97 7/29/97Date10/27/97 1/25/98 4/25/98
1987 without incident. The
Phase II filling began in 1987.
Routine monitoring o f the Figure 8
instrumentation in 1990,
showed accelerating S e t t l e m e n t Plate S P - 8 ( P h a s e III)
movement in two 1000.0
inclinometers in Phase II after 3 m waste
800.0
placement of the first four
lifts o f waste. Based on these g
600.0
measurements, three ==
additional inclinometers were 400.0
installed and monitored r
200.0
weekly. The monitoring data
from these instruments 0.0
confirmed that a shear plane 12/1/93 12/1/94 12/1/95 11/30/96 11/30/97
was forming in the dredge Date
spoils. A topographic survey
o f the waste revealed that the Figure 9
waste had been placed about 3 m (10 ft) higher than the design waste height. The actual
slope configuration was analyzed and found to have a safety factor near unity. The slope
was quickly stabilized by constructing a toe berm from amended sludge imported from the
neighboring wastewater treatment plant. The movement abated and filling was resumed.
This failure and analysis confirmed that soil parameters used in the design of Phase II were
generally accurate and showed the value of instrumentation monitoring to the safe operation
of the facility. Phase II filling topped out without further incident in 1991.
Inclinometer casings in Phases I and II have deflected sufficiently to prevent insertion
of the inclinometer. The early installations in these phases used 70 mm (2.75 inch) casings
with rigid couplings. The later phases using 85 mm (3.34 inch) casings with telescoping
couplings have undergone similar deflections and remained readable. The telescoping
couplings can be problematic for taking repeatable readings and care must be used to avoid
positioning the inclinometer instrument across a coupling. This problem has been overcome
by logging the installation and carefully positioning the inclinometer during readings. All
replacements are now being made with the larger diameter casing with telescoping couplings.
Phase III was constructed in 1991 and continues to receive waste. The
instrumentation has confirmed that pore pressures caused by the placement of waste are
dissipating very slowly. In fact, more slowly than would be expected from the design data.
This has resulted in modifying operation plans of the facility to alternate waste filling among
the three active phases to permit a longer period of time for pore pressures to dissipate and
dredge spoil strengths to increase in the critical slope areas. The second lift of waste to be
placed in Phase III was set back 45 meters (150 feet) from the perimeter slope to maintain
stability since the excess pore pressure had not dissipated sufficiently under the first lift of
waste. Continued monitoring and evaluation indicated that pore pressures had dissipated
sufficiently to fill this bench and place a third lift with a setback of 30 m from the slope face.
Fill placement in Phase IV began in 1995. Placement of the first lift o f waste was
completed in October of 1995. Based on the designer's review, the second lift was placed
in Phase IV with a set back of 135 m (450 ft). On-going analysis is required to determine
when the setback area can be filled. Phase V received its first lift of waste later in 1996. An
evaluation of this phase is also in progress to assess the stability of the second lift to be
placed.
Conclusions
We have made the following conclusions based on the instrumentation data:
Interpretation of pore pressures is closely tied to estimated waste loading.

Pore pressures were often higher than expected considering the amount of waste
placed. This was initially assumed to be a lack of consolidation under the waste, but
has been determined to be a result of higher than expected waste density.
Total pressure cells indicate that the average waste density is 25 to 100 percent
higher than estimated in the design.
The waste load is not a constant for a given lift. The waste density increases as it
absorbs moisture from storm water infiltration and as soil cover materials are worked
into the waste pore spaces.
Consolidation is proceeding. Evaluation of consolidation requires periodic
reassessment of the waste loads.

Liquid settlement transducers have not proven to be durable or reliable. All LSTs
have been abandoned.
Total pressure cells do incur an inclusion effect. This has been adjusted by
measurements under known loadings during installation of the instruments in final
placement.
Measured settlements generally agree with expected settlement. The accuracy of
settlement measurements is variable because of distortion of the vertical extensions
and potential for downdrag on the protective casings.
Larger diameter inclinometer casings with telescoping couplings remain readable to
larger deformations than smaller casings with rigid couplings.
As the site continues to be filled, instruments are regularly monitored to validate the design
assumptions. The instruments have proven to be valuable tools in assessing the performance
of the subsoils and dredge spoil materials. Periodic instrument monitoring combined with
evaluation of slope stability is used to maximize both operational efficiency and to maintain
stable slopes. DSWA is planning to install additional instrumentation within the existing
cells and Phase V to better define dredge spoil behavior and the loadings from the waste.
Filling is expected to continue for the next 15 to 20 years and this instrumentation will play
a vital role in the successful operation and performance of the Cherry Island Landfill.
References
Casagrande, A. (1949), "Soil Mechanics in the Design and Construction of the Logan
Airport," Journal of the Boston Society of Civil Engineers, Vol. 36, No. 2, pp. 192-
221.
Dunnicliff, J., (1993), Geotechnical Instrumentationfor Monitoring Field Performance,

John Wiley & Sons, New York, NY.
Felio, G.Y., and G.E. Bauer (1986), "Factors Affecting the Performance of a Pneumatic
Earth Pressure Cell," Geotechnical Testing Journal ASTM, Vol. 9, No. 2, June, pp.
102-106.
Slope Indicator (1996), Geotechnieal and Structural Instrumentation, Slope Indicator

Company, Bothel, WA.
Weiler, W.A., and F.H. Kulhawy (1982), "Factors Affecting Stress Cell Measurements in
Soil," Journal of the Geotechnical Engineering Division of ASCE, Vol. 108, No.
GT12, Dec. Pp. 1529-1548.
Horace K. Moo-Young 1, Chris LaPlante 2, Thomas F. Zimmie 3, and Juan Quiroz 4
FIELD MEASUREMENTS OF FROST PENETRATION INTO A LANDFILL

COVER THAT USES A PAPER SLUDGE BARRIER
REFERENCE: Moo-Young, H. K., LaPlante, C., Zimmie, T. F., and Quiroz, J. "Field
Measurements of Frost Penetration into a Landfill Cover that uses a Paper Sludge
Barrier," Field Instrumentation for Soil and Rock, ASTM STP 1358, G. N. Durham and
W. A. Marr, Eds., American Society for Testing and Materials, West Conshohocken, PA,
1999.
Frost penetration is a major environmental concern in landfill design. Freezing

and thawing cycles may deteriorate the permeability of the liner or cap. In this study, the
depth of frost penetration into a landfill cover that uses paper sludge as the impermeable
barrier (the Hubbardston landfill in Massachusetts) was measured using a frost
measurement system. A thermistor probe measured the temperature at various depths.
Although temperature measurements are important, soil resistivity measurements are
required to accurately predict the freezing level, since soil resistivity increases greatly
upon freezing. A conductivity probe measured the half bridge voltage between
conductivity rings and a ground rod. Data were collected in data loggers.
The data collected from 1992-1996 showed that the frost level did not penetrate
the paper sludge capping layer. Heavy snow cover throughout the winters decreased the
depth of frost penetration by insulating the landfill. The high water content in the sludge
also contributed to the lack of freezing.
KEYWORDS: instrumentation, freezing, thawing, permeability
1Assistant Professor, Department of Civil and Environmental Engineering, Lehigh University,

Bethlehem, PA 18015.
2Assistant Professor, Department of Civil and Environmental Engineering, Union College, Schenectady,
NY.
3professor, Department of Civil Engineering, Rensselaer Polytechnic Institute, Troy, NY 12180.
4 Graduate Research Assistant, Department of Civil Engineering, Rensselaer Polytechnic Institute, Troy,
NY 12180.
INTRODUCTION
A major environmental problem facing the world today is waste disposal.

Although several methods of solid waste disposal are available, landfilling is by far the
most popular method of disposal. Low permeability hydraulic barriers are utilized to
reduce the movement of leachate from the landfill. Although regulatory requirements for
the permeability of the hydraulic barrier vary, the most common maximum allowable
value is 10-7 crn/sec, and compacted clay soils are commonly used as the cover material.
When clay is not locally available, the cost of landfilling is significantly

increased. This has sparked interest in the use of unconventional material such as paper
sludge to substitute as the barrier protection layer. Paper sludges have been successfully
used to cap landfills in Maine, Wisconsin and Massachusetts (Moo-Young and Zimmie,
1997; Moo-Young et al., 1997; Zimmie et al., 1997; Moo-Young and Zimmie, 1996a;
Moo-Young, 1995; Zimmie et al., 1995; Moo-Young, 1992).
A major concern in landfill design is the effect of freezing and thawing on the
hydraulic conductivity of the hydraulic barrier. Numerous researchers have studied the
effects of freezing and thawing on the hydraulic conductivity of compacted clays
(Othman et al., (1994) summarizes these works). In general, freeze and thaw cycles can
cause an increase in the hydraulic conductivity of compacted clays of one to two orders
of magnitude. In studying the effects of freeze/thaw cycles on paper sludge, Moo-Young
and Zimmie (1996a, 1996b) obtained a similar increase in hydraulic conductivity to
compacted clay. However, unlike compacted clays that show a greater increase in
hydraulic conductivity at low effective stresses (two to three orders of magnitude), the
paper sludge's hydraulic conductivity increased about one order of magnitude, regardless
of the effective stress. The better performance of the paper sludge may be due to the
fibers in the sludge and the high compressibility of the sludge. Fibers give the sludge
some ability to resist tension. There are a wide variety of paper sludges, and the
observation on the role of fibers may be unique to the type of sludge tested (Moo-Young
and Zimmie, 1996a).
Although there is a significant amount of data that supports the detrimental effect
of freezing and thawing cycles on clay and paper sludge low permeability layers, there is
little information available on the amount of frost penetration at landfills. In this study,
instrumentation to measure frost penetration was installed in the paper sludge low
permeability layer at the Hubbardston landfill in Massachusetts and at the Corinth landfill
in Corinth, NY and into the clay cap at the Greenwich landfill in Greenwich, NY and at
the Corinth landfill.
Frost Action in Soils
Most of the literature on frost heave action in soils pertains to highways,

foundations, construction in permafrost, and chilled pipelines buried in unfrozen soils.
Frost action in soils has received considerable attention in the literature (Jumikis, 1966,
Tsytovich, 1975; Nixon, 1991; Anderson and Morgenstern, 1973). Frost heaves are
MOO-YOUNG ET AL. ON FIELD MEASUREMENTS 193
caused by the freezing of in situ pore water and by the flow of water to the freezing front.
Pressure develops in the direction of crystal growth that is determined by the direction of
cooling. The frost fringe transports water to the active ice lens from the unfrozen soil
under a suction gradient. Pore water in fine grained soils does not necessarily freeze at 0 ~
C. In some clays, as much as 50% of the moisture may remain as a liquid at -2 ~ C
(Tsytovich, 1975; Anderson and Morgenstem, 1973). The freezing temperature depends
on pore size, water content, applied pressure, and solute concentration.
Extensive research has been conducted on the effects of applied pressure on frost
heaves. Applied pressure inhibits frost heaves and affects the freezing temperature and
permeability of the soil. When the rate of cooling is near zero, the freezing front attracts
water as long as the applied load is less than the actual shutoff pressure in which no water
flows to the ice lens (Konrad and Morgenstem, 1982). For all practical purposes, the
shutoffpressure in fine grained soils is quite high and would never be exceeded in
landfill covers, which typically are subjected to low values of effective stress.
Some major factors controlling frost penetration in soils are the soil thermal
conductivity (k), volumetric sensible heat (C), and latent heat (L). Thermal conductivity
is the ratio of heat flow through a unit area under a unit gradient. Volumetric sensible
heat is the change in thermal energy in a unit volume of soil per unit change in
temperature. It is derived from the specific heat, which is the change in thermal energy
per unit mass per unit temperature change. Latent heat is the change in thermal energy in
a unit volume of soil to freeze and thaw the soil moisture at its melting point without
change in temperature. Sludges require more energy loss to freeze pore water than
typical soils, due to the high water contents. The latent heat of soil moisture controls
frosts penetration.
Soil Resistivity
In fine-grained material, the pore water does not necessarily freeze at 0 ~ C
(Konrad and Duquennoi, 1993). Liquid water well below 0 ~ C, in a super cooled
condition, moves towards the frozen fringe to form ice lenses. Although temperature
measurements are very important, an accurate determination of freezing requires
additional measurements, one possibility being soil resistivity. Soil resistivity increases
greatly upon freezing. The resistance, R, of a material is defined as the voltage, V,
divided by the current, I, and is proportional to resistivity, p. Soil Resistivity, p, is equal
to the resistance,R, multiplied by the length, 1, and divided by the cross sectional area, A
(where p = R1/A = V1/AI). Electrical conductors have very low resistivity, and insulators
have very high resistivity. Water has a lower resistivity than ice. Before the pore water
in the soil forms the active ice lens, the resistivity of the soil is low. As ice lenses form in
the soil, resistivity increases.
INSTRUMENTATION EQUIPMENT
Conductivity Probe
Geonor Inc. built the frost measurement system utilized in this study. A
conductivity probe measures the half bridge voltage between the given conductivity ring
and a reference electrode, when a 60 ms pulse is fed to the ring from a 2.5 V DC source
through a 10,000 ohm series resistor. The probe has a length of 79.6-cm and a diameter
of 2.22-cm. The stainless steel reference electrode has a length of 94.5-cm and a
diameter of 2.22-em. Each conductivity probe has eight conductivity rings that are
spaced 7.62-cm apart and are located on the outside of a schedule 40 PVC pipe. The
sensors on the conductivity probe are assigned a number in a top-down direction. The
top of the probe connects to the data acquisition cable, which relays the conductivity
reading to the data acquisition system. For thawed paper sludge, a voltage reading of 0.3
to 0.5 V is typical; while for completely frozen material (zero liquid pore water), a
voltage of about 2.5 V is common. The formation of ice lenses and subsequent freezing
is indicated by a voltage increase (Geonor, 1990). Figure 1 illustrates the frost
measurement system.
In a DC excitation situation, galvanic potential is a common problem. Two

techniques are used in this probe to minimize this effect: the rings and reference electrode
are made of the same grade and lot of stainless steel; and the excitation pulse width is
short, minimizing electromigration and other undesirable effects.
Figure 1 Instrumentation Equipment
Thermistor Probe
Thermistor probes (as shown in Figure 1) measure the temperature at various

depths in the soil and consist of eight thermistors spaced 7.62-cm apart. The probe has a
length of 79.6-cm. Thermistors are inside a schedule 40 PVC with an outer diameter of
2.22-cm. The thermistor temperature range is between -75 ~ C to 150~ C. At the top of
the thermistor, a data acquisition cable transmits the temperature readings to a data
logger. Correlation between temperature and conductivity readings at the same depth
occurs when the thermistors and conductivity rings are installed parallel to each other.
Data Loggers
R.S. Technical Instruments Ltd. (RST) built the data loggers and created the
LE8200 Data logger Software. The data logger (station) has three external ports:
conductivity, thermistor, and data retrieval. Interface cables from the thermistor and
conductivity probes connect to the data logger ports. Microprocessors collect and store
the data. A 10,000-ohm series resistor located in the data logger feeds a 60 ms pulse to
the conductivity rings.
Each data logger is capable of storing 32,000 bits. Each probe requires 16 bits per
reading, hence the data logger can store 1,000 readings overall. The data logger can be
programmed to take readings in various modes. Typically, hourly readings are obtained,
and thus 42 days of data can be stored.
LANDFILL INSTRUMENTATION
The Hubbardston Landfill, located in Hubbardston, Massachusetts, is a four-acre

municipal waste disposal facility that utilized Erving paper mill sludge as the low
permeability cover layer. Erving paper mill is conducting studies to measure the
performance of their sludge serving as the low permeability barrier in landfill covers.
Two sets of the frost measurement system were installed into the low permeability
layer in the Northwest portion of the Hubbardston Landfill in December 1992. The
impervious layer is overlain by a 30.5-cm thick top soil layer and by a 15.5-cm thick
drainage layer. The drainage layer has a permeability equal to or greater than 1 x 10-3
era/see. A frost chisel was used to remove the top 7.6-cm of frozen soil and to outline the
site. The remaining soil was excavated, and the site was prepared for installation of the
probes. An elevated concrete platform was constructed to place and store the data
loggers. The frost measurement systems were placed 3.81-m to the north (Hole #1) and
3.86-m to the west (Hole #2) of the concrete pad.
Two guide holes were drilled into the sludge layer perpendicular to the surface
and were spaced 7.62-cm apart on center (the 7.62-cm spacing was suggested by the
manufacturer). Electric ground rods were used to puncture the holes; two 1.9-cm
diameter water pipes were welded together to provide a guide for the two parallel holes.
A = THERMISTOR PROBE
B = GROUND ROD
C = CONDUCTIVITY PROBE
Figure 2 Vertical Profile of the Hubbardston Landfill
The conductivity and ground probes were placed into these holes. The holes were slightly
smaller than the probe's diameter to allow a good soil-probe contact. A third hole for the
thermistor was punctured perpendicular to the surface.
The probes were pushed by hand into the two holes made in the sludge. Figure 2
shows the vertical profile for the instrumentation equipment into the Hubbardston
landfill. In Hole #1, the probes were placed 61-cm into the sludge. The vegetative
support layer was 30.5-cm thick, the sand drainage layer was 20.3-cm thick, and the
sludge layer was 76.2-cm thick (Figure 2). The top bands of the conductivity and
thermistor probes were in the sand layer. In Hole #2, the probes were placed 74-cm into
the sludge which allowed all the conductivity bands and thermistors to be contained in
the sludge (Figure 2). For Hole #2, the topsoil layer was 30.5-cm thick, the sand drainage
layer was 15.2-cm thick, and the sludge layer was 76.2-cm thick.
Two trenches 7.6-cm deep were excavated from the concrete pad to the holes
containing the frost measurement system. The thermistor and conductivity data
acquisition cables were placed in the trenches and covered with top soil. The data
loggers were placed on top of the concrete pad, and the instrumentation cables were
connected to the data loggers through a hole in the middle of the concrete pad.
To prevent vandalism and to protect the data loggers from the environment, half
of a fifty-five gallon drum was placed over the data loggers and secured to mounts
constructed into the concrete pad with a lock. The sand and top soil layers were carefully
backfilled into both holes so that the instrumentation equipment was not disturbed
(Myers, 1996).
INTERPRETATION OF FROST MEASUREMENT DATA
Although there were two sets of the frost measurement system, Hole #1 will be
used to depict the depth of frost penetration into the Hubbardston Landfill, since Hole #2
replicates the data of Hole #1. Figure 3 displays the minimum air temperature at the
Hubbardston Landfill from December 1992 to April 1996, The mean air temperature
during the four freezing periods, Ts, and the average duration of freezing, tavg,were -5~
and 163 days respectively. The average minimum (Tomin)and maximum (Tomax)surface
temperatures are 2.5~ and 12~ C, respectively. In Figure 3, the four periods of freezing
(winters) are represented by 0 to 159 days, 348 to 543 days, 746 to 903 days, and 1096 to
1239 days.
Figure 4a displays a three dimensional plot of temperature, time, and voltage for
Hole #1 at Level #1, which represents the lower portion of the sand drainage layer (see
Figure 2). During the four winters from 1992 to 1996, Figure 4a shows a decrease in
temperature below 0~ and an increase in soil voltage in the sand drainage layer. The
increase in voltage (i.e. this infers an increase in soil resistivity) indicates that freezing
(the formation of ice lenses) occurred in the sand drainage layer during each winter.
Figure 4b displays a three-dimensional plot of temperature, time and voltage for

Hole #1 at Level #2, which represents the upper layer of the sludge barrier. Figure 4b
shows a decrease in temperature slightly below 0~ during each winter. Since there was
no increase in the voltage plane during the four period of freezing, it can be inferred that
the sludge layer did not freeze. Levels #3-8 for Hole #1 and Levels #1-8 for Hole #2
follow a similar trend of no freezing to Level #1 for Hole #1 (Moo-Young et al., 1997;
Zimmie et al., 1997; Moo-Young, 1995).
ANALYSIS OF FROST PENETRATION DATA
Frost measurement data from the landfill was compared to two conventional
thermal analysis tools to determine the applicability of these methods to predicting frost
penetration into landfills: Corp of Engineers design charts and Stefan formula (Brown,
1964; Linell et al., 1963; Mitchell, 1993).
10 . . . . . . . . . = |,)0 c
0 t~u--~ - ~ _ ~ - -Nltt-~,~ - - ~,-~i-~ r T~in = Zs~

i ' i ' r __ ,111 ,LT --5.5~
Ltavz~ 163days
-30 i i i i f i i
0 200 400 600 800 1000 1200 1400
TIME (DAYS)
Figure 3 Air Temperature Profile for the Hubbardston Landfill
Estimating Properties
Frost penetration depths in soils depend on numerous factors including weather
conditions, ground conditions, and soil conditions. Weather conditions for predicting
frost penetration include the freezing index, the mean annual temperature, Tavg,and the
duration of freezing, Gtvg. The freezing index, F, is equal to the product of the mean
surface temperature during freezing, Ts, and the duration of freezing, tavg. Weather data
for the Hubbardston landfill is given in Figures 3.
Frost penetration also depends on the temperature of the ground surface. It should
be noted that the ground surface is generally warmer than the air temperature, and the
31
3~. ~, ! !LlP-z_ltl I I1! 11 I I

-" ~, ~, ~,~,~1 T-AII--I-t~z-JI_ I I
~ i ~1 6"
TI~ (~AYs)
a)
31 ' ' ~ I I I I I~-----1-~__ 1 I I
~ A- I I I I I I i !
/// I I I I/---~-- I f t I
2 II,/4 ; ---m~-f--l~_~_
I
I }/I I I L I F I , I I
II I _~. ~ so ,~%
TIME (DAys)
b)
Figure 4a Temperature, Voltage, and Time Profile for Hole #1 at Level #1

4b Temperature, Voltage, and Time Profile for Hole #1 at Level #2
difference between the ground temperature and air temperature is enhanced by snow
cover.
Frost penetration depths are controlled by the soils' thermal conductivity,

volumetric sensible heat, and latent heat. The thermal conductivity for frozen and
unfrozen paper sludge can be estimated from the thermal properties of its constituents.
The thermal conductivity for unfrozen paper sludge at 0~ is equal to 0.98 J/m-s~
assuming that paper sludge has an water content of 180% and is composed of 50% clay
and 50% organic (paper). Typical values for the specific heat and volumetric sensible
heat for paper sludge are 3.1 J/kg~ and 1,630 J/m 3 ~
Latent heat, L, depends on the amount of water in a unit volume of soil. One
kilogram of water gives off 151.3 kJ as it freezes. The latent heat of soil, L (kJ/m3), can
be estimated as follows: L = 0.80 W~'dry (Jumikis, 1966). For paper sludge (~'dry= 530
kg/m3), the latent heat is estimated to be 49,900 kJ/m3.
Thermal Analysis
The Corp of Engineers design chart was utilized to predict frost penetration
depths into the landfill covers (Brown, 1964). From Figure 3, the freezing index was
predicted for the landfill. Table 1 summarizes the measured depth of frost penetration
and compares it to the predicted results from the Corp of Engineers. Table 2 shows that
the Corp of Engineers design chart over predicts the depth of frost penetration into the
landfill cover. It should be noted that the Corp of Engineers design curve was developed
for a well drained granular non-susceptible base coarse material. This method may be
applicable to the sand drainage layer in the landfill cover.
Utilizing the thermal and physical properties of paper sludge and sand, the Stefan
formula was utilized to estimate the depth of frost penetration into the landfill. This
solution assumes that the latent heat is the only heat removed during freezing and is
shown in equation 1:
Z = (2kF/L) x/z (1)
Where Z is the depth of frost penetration, k is the thermal conductivity, F is the freezing
index, and L is the latent heat.
Table 1 displays the depth of frost penetration predicted by the Stefan formula for
the barrier layer and for sand. It should be noted that these estimated depth of frost
penetration into the paper sludge layer represent freezing into an exposed landfill cover
(i.e. no frost protection, vegetation, or drainage layer.). To predict the effects of frost
penetration into the drainage and vegetation layers at The Hubbardston Landfill, the
Stefan formula was utilized to estimate the depth of frost penetration in sand. For The
Hubbardston Landfill, the Stefan formula predicts that frost penetration will be a
problem, since the depth of frost penetration predicted for sand in Table 1 is greater than
the depths of the vegetation and drainage layers (Figures 2). For the paper sludge barrier
Table 1: Summary of Measured to Predicted Frost Penetration Depths
LANDFILL YEAR MEASURED STEFAN CORP OF

FROST FORMULA ENGINEERS
PENETRATION (cm) (cm)
(cm)
Sludge Sand
Hole #1 92-96 _<50.8 175 554 140
at the Hubbardston Landfill, the Stefan formula over predicted the depth of frost
penetration in comparison to the measured frost penetration. The Stefan formula may
have over predicted the depth of freezing, because the removal of the volumetric sensible
heats of frozen and unfrozen soil are neglected (Mitchell, 1993).
FACTORS CONTRIBUTING TO T H E LACK OF FROST PENETRATION
Four factors may have contributed to the lack of frost penetration in the protected
landfill cover at the Hubbardston Landfill: high water content of paper sludge, applied
stress created by the drainage layer and vegetation layer, snow cover, and the heat of
decomposition in the landfill. The high water content of the paper sludge may have
attributed to the lack of freezing. After one year of in situ consolidation, the sludge water
content at the Hubbardston Landfill ranged from 100% to 110% (Moo-Young, 1995).
For typical sands, the in situ water content is about 10 to 15%, and for clay, the in situ
water content typically ranges from 20 to 30%. Moo-Young (1995) showed that the pore
water from paper sludge has a freezing point depression of-5~ Significantly more
energy loss is required to freeze the pore water in sludge than to freeze the pore water in a
typical sand or clay. The Stefan formula, and Corp of Engineers design chart do not
account for the flow of water into or out of the soil during freezing. This may have
attributed to the over prediction of frost penetration depth by these methods.
Another factor, which may contribute to the lack of freezing, is the applied stress
on the barrier layer. Applied pressure affects the freezing temperature and the
permeability of clays (Othman et al., 1994, Konrad and Morgenstern, 1982). For paper
sludges, Moo-Young and Zimmie (1996a and 1996b) showed a decrease in hydraulic
conductivity with an increase in applied stresses. Since paper sludge consolidates
considerably under a low effective stress and since paper sludge is composed of 50%
kaolinite, applied stresses may affect the freezing temperature of paper sludge in a similar
fashion to clays.
The third factor, which may have reduced the depth of frost penetration, is snow
cover at the landfill. Although the Corp of Engineers design chart, and Stefan formula,
utilize climatic data, these solutions do not account for the insulation of the soil by snow
cover. Figure 5 shows the observed snow cover depth at the Hubbardston Landfill from
25-
20.
15
1
10
II
5
0 I I i
I I i I I
~ 0 200 400 600 800 1000 1200 1400
alME (DAYS)
Figure 5 Observed Snow Depth
December 1992 to April 1996. Snow cover throughout the winters of 1992-1993 and
1993-1994 at the Hubbardston Landfill may have insulated the landfill cover. In
analyzing the data for Hole #1 at Level #2, although the temperature dropped below 0~
freezing can not be inferred from the voltage data. During the winters of 1994-1995 and
1995-1996, the observed snow cover depth at the Hubbardston Landfill was less than the
previous two winters.
The fourth factor, which can be postulated to reduce the m o u n t of frost

penetration into the landfill cover, is the heat of decomposition in the landfill. The heat
of decomposition is created by the decomposition of the waste caused by biological
activity of the waste mass. The contribution of the heat of decomposition to reducing the
frost penetration into the Hubbardston landfill cover was unknown. However, this is an
area for future research.
SUMMARY AND CONCLUSION
A frost measurement system was installed into the Hubbardston landfill to

measure the depth of frost penetration into clay and paper sludge low permeability barrier
layers. A thermistor probe measured the temperature at various depths, and a
conductivity probe measured the half bridge voltage between a given conductivity ring
and the ground probe.
Data from the landfill were collected from 1992-1996. For the Hubbardston
Landfill, there was no frost penetration into the barrier layer. Thus, the paper sludge
acted as an effective hydraulic barrier. It should be noted that even though no frost
penetrated into barrier layers of the unexposed landfills, future landfill designers should
design for maximum frost penetration. Freezing and thawing in the barrier layer has been
shown to increase the hydraulic conductivity of paper sludge or clay (Moo-Young and
,1996b and Othman et al., 1994).
The measured frost depths were compared to the two theoretical solutions for
predicting the depth of frost penetration: Corp of Engineers design chart and Stefan
formula. These solutions over predicted the depth of frost penetration into the unexposed
landfill covers.
Three factors may have contributed to the reduction in the depth of frost
penetration: high water content, applied stress, and snow cover. Heavy snowfall
throughout the winter covered the landfill may have acted as an insulation blanket, which
reduced the depth of frost penetration. The high water content of paper sludge may have
also contributed to the lack of freezing. Lower temperatures and more energy loss are
required to freeze the pore water in paper sludge than to freeze the pore water in a typical
sand or clay. This indicates that low temperatures are required to freeze paper sludge.
Moreover, the applied stress on the sludge layer that is created by the drainage layer and
vegetation layer may have affected the freezing temperature of the paper sludge
ACKNOWLEDGEMENT
The Erving Paper Company, in Erving, MA, the United States Environmental Protection
Agency, and the National Science Foundation provided funds for this research. Their
generous support and cooperation is greatly appreciated. However, the opinions
expressed herein are solely those of the authors. The authors would also like to thank
Timothy Myers of the Hydro Group in Bridgewater, New Jersey and Warren Harris and
Dr. Carsten Floess of Clough Harbor and Associates in Albany, New York for their
assistance in conducting this project.
REFERENCES
Aldrich, H.P., 1956, "Frost Penetration Below Highway Pavements," Highway

Research Board, Bulletin 135.
Anderson, D.M. and Morgenstem, N.R., 1973, "Physics, Chemistry and Mechanics of
Frozen Ground," Proceedings, 2nd International Conference on Permafrost,
Yakutsk, U.S.S.R., pp. 257-288.
Brown, W.G., 1964, "Difficulties Associated with Predicting Depth of Freeze or Thaw,"
Canadian Geotechnical Journal. Vol. 1, No. 4, pp. 150-160.
Geonor, Inc., 1990, Frost Measurement System, Geonor, Inc., Milford.
Jumikis, A.J., 1966, Thermal Soil Mechanics~ Rutgers University Press, New
Brunswick.
Konrad, J.M. and Morgenstem, N.R., 1982, "Effects of Applied Pressure on

Freezing Soils," Canadian Geotechnical Journal, Vol. 19, pp. 494-505.
Linell, K.A., 1963, "Corp of Engineers Design in Areas of Seasonal Frost,"

Highway Research Record, No. 33, pp. 77-128.
Mitchell, J.K., 1993, Fundamentals of Soil Behavior. John Wiley and Sons, New York.
Moo-Young., 1992, Evaluation of the Geotechnical Properties of a Paper Mill Sludge

for Use in Landfill Covers, Master of Science Thesis, Rensselaer Polytechnic
Institute, Troy.
Moo-Young, H.K., 1995, Evaluation of Paper Mill Sludgesfor Use as Landfill

Covers, Ph.D. Thesis, Rensselaer Polytechnic Institute, Troy.
Moo-Young, H.K. and Zimmie, T.F., 1996a, "Geotechnical Properties of Paper of

Mill Sludges for Use in Landfill Covers," Journal of Geotechnical and
GeoEnvironmental Engineering. Vol. 122, No. 10, pp. 768-775.
Moo-Young, H.K. and Zimmie, T.F., 1996b, "Effects of Freezing and Thawing on
the Hydraulic Conductivity of Paper Mill Sludges Used as Landfill Covers,"
Canadian Geotechnical Journal, Vol. 33, pp. 783-792.
Moo-Young, H.K. and Zimmie, T.F., 1997, "Utilizing a Paper Sludge Barrier Layer in a
Municipal Landfill Cover in New York," Testing Soils Mixed with Waste or
Recycled Materials, ASTM STP 1275, M.A. Wasemiller and K. Hoddinott, Eds.,
American Society of Testing and Materials, West Conshohocken, pp. 125-140.
Moo-Young, H.K., Zimmie, T.F., and Myers, T.J., 1997, "Environmental

Monitoring of the Depth of Frost Penetration into Landfill Covers," ISCORD, H.
Zubeck, C. Wollard, D. White, and T.S. Vinson, (Eds.), Fairbanks, AK.,
American Society of Civil Engineering, Reston, pp. 403-406.
Myers, T.J., 1996, Landfill Cover Instrumentation to Determine Depth of Frost

Penetration, M.S. Thesis, Rensselaer Polytechnic Institute, Troy.
Nixon. J.F., 1991, "Discrete Ice Lens Theory for Frost Heave Beneath Pipelines,"
Canadian Geotechnical Journal Vol. 29, pp. 487-497.
Othman, M., Benson, C.H., Chamberlain, E.J., and Zimmie T.F., 1994, "Laboratory
Testing to Evaluate Changes in Hydraulic Conductivity of Compacted Clays
Caused by Freeze-Thaw: State of the Art," _Hydraulic Conductivity and Waste
Contaminant Transport in Soils. ASTM STP 1142, D.E. Daniel and S.J.
Trautwein. American Society for Testing and Materials, West Conshohocken, pp.
227-254.
Tsytovich, N.A., 1975, The Mechanism of Frozen Ground. McGraw-Hill.

Washington, D.C.
Zimmie, T.F., Moo-Young, H.K. and LaPlante, K., 1995, "The Use of Waste Paper
Sludge for Landfill Cover Material," Green 93--Waste Disposal by Landfill,
R.W. Sarsby, Ed., A.A. Balkema, Rotterdam, pp. 487-495.
Zimmie, T.F., LaPlante, C.M., and Quiroz, J.D., 1997, "Evaluation of Frost Penetration
in Landfill Cover Systems," 1SCORD, H. Zubeck, C. Wollard, D. White, and T.S.
Vinson, (Eds.), Fairbanks, AK., American Society of Civil Engineering, Reston,
pp. 407-410.
Thomas G. Thomann, 1 Majed Ao Khoury, 1 Jack L. Rosenfarb, ~ and Richard A.
Napolitano2
Stability Monitoring System for the Fresh Kills Landfill in New York City
REFERENCE: Thomann, T.G., Khoury, M.A., Rosenfarb, J.L., and Napolitano, R.A.,
"Stability Monitoring System for the Fresh Kills Landfill in New York City," Field
Instrumentation for Soil and Rock, ASTM STP 1358, G.N. Durham and W.A. Marr, Eds.,
ABSTRACT: The Fresh Kills Landfill, located in Staten Island, New York, serves as the
repository of all municipal solid waste from the five boroughs of New York City. Because
of the existence of compressible soils under most of the filling areas and the urban
environment surrounding the landfill, considerable importance is being placed on the
relationship between filling operations and the stability of the landfill. As a result of this
concern and to address Order on Consent requirements, a program of geotechnical site
characterizations, stability analyses, and design and implementation of a geotechnical
instrumentation program was undertaken. Geotechnical instruments have been installed
within the refuse fill and foundation soils to monitor both the magnitude and rate of
change of pore pressure, lateral and vertical movements, and temperature. This paper
presents an overview of the subsurface conditions, the overall instrumentation plan for
assessing the landfill stability, a description of the various instruments, the performance of
these instruments to date, an overview of the collected measurements, and a description of
how these measurements are used to monitor the stability.
KEYWORDS: geotechnical instrumentation, slope stability, landfill, lateral deformations,

settlement, refuse fill
Introduction
The Fresh Kills Landfill, located in Staten Island, a borough of the City of New York (see
Figure 1), serves as the repository of all municipal solid waste from the five boroughs of New
York City. The landfill is operated by the New York City Department of Sanitation
(NYCDOS) and has received as much as approximately 115,650 kN (13,000 tons) of
municipal refuse per day. Construction of the landftU began in 1948 and now encompasses
SeniorProjectEngineer,SeniorPrincipal, and SeniorProjectEngineer,respectively,URS Greiner

Woodward Clyde, 363 Seventh Ave., New York, New York.
2 Senior Project Manager, New York City Department of Sanitation, New York, New York.
206
THOMANN ET AL. ON STABILITY MONITORING SYSTEM 207
NEW
JERSEY
Bay
BROOKLYN
FRESH KILLS STATEN

LANDFILL
ISLAND
Lower New
York Bay
Atlantic
Ocean
Not to Scale
Figure 1 - General Site Location Plan
more than 9.7 sq. km. (2,400 acres) of former marshland. The landfill is comprised of four
ections (see Figure 2) identified as the Fresh Kills Bargefill (Sections 1/9 and 6/7), the Victory
Boulevard TruckfiU (Section 3/4) and the Muldoon Avenue Trucklill (Section 2/8). Sections
2/8 and 3/4 are no longer receiving refuse fill and have undergone final closure. Sections 1/9
and 6/7 are currently receiving refuse fill. Originally, the landfill operations were to continue
into the year 2015 to reach a maximum height of about 152 m (500 ft). However, present
plans are to close the landfill by the year 2002.
Considerable importance is being placed on the relationship between filling operations and
the safety and stability of the landfill because of 1) the planned size of the landfill (Section 1/9
when completed will be about 76 m (250 ti) high based on the most recent plan); 2) the
existence of compressible soils under most filling areas, and 3) the presence of an urban
environment in the vicinity of the landfill. As a result of these concerns, the NYCDOS has
undertaken a program ofgeotechnical site characterizations, stability and deformation analyses,
design and installation of an automated monitoring instrumentation system, and development of

an operations and maintenance manual for this system. This program was also designed to
fulfill a portion of an agreement between the NYCDOS and the New York State Department
of Environmental Conservation (NYSDEC).
Because of the complex stratigraphy at the site and the uncertainties regarding the short-
term and long-term behavior of the refuse and foundation soils (in terms of their field behavior
versus laboratory behavior), an observational methodology has been adopted (Peck, 1969).
This entails installing a monitoring system to obtain measurements of key parameters (e.g, pore
pressure and deformation) during landfillingto ascertain the performance of the landfill, and
then to modify, if necessary, landfilling operations (e.g., location of filling rate of filling
schedule of filling) to maintain the satisfactory performance of the landfill. Similar approaches
have been adopted for other projects involving earthen embankments and landfills(Handfelt, et.
al. 1987, Withiana,et. al. 1995, Oweis, et. al~ 1985, Duplanic 1990)
Figure 2 - Landfill Section Plan
This paper presents an overview of the subsurface conditions, the overall instrumentation
plan for assessing the landfill stability, a description of the various instruments, and the
performance of these instruments to date. Summaries of the data, collected to date, are also
presented and discussed in the context of their use in conjunction with on-going stability
assessment of the landfill.
Site Characterization
Fresh Kills Landfill is located in a very complex lithology due to the past existence of two
and, perhaps, three known glacial end-moraines found south of the site. As a result, the soil
stratigraphy under the landfill is extremely variable in layer thickness, extent, and material
properties. The strata of interest to the evaluation of the stability of the landfall are generally in
the following sequence from the ground surface:
9 Refuse Fill
9 Recent Silt and Clay, Q~
9 Recent Sand, Q~
9 Glacial Sand, Qp
9 Glacio-lacustrine Clay, Q~
9 Glacial Till, Q~
9 Cretaceous sequence of Clays and Sands, I~ and K,
Not all strata are present throughout the landfill complex. For example, the Q~ deposits
are absent from historic upland areas (prior to the construction of the landfill) which are
generally located in the eastern portions of Sections 1/9 and 6/7, the southern portion of
Section 2/8, and the northeast portion of Section 3/4.
Monitoring System Approach
Considerable importance has been placed on providing adequate instrumentation coverage

for the complex and diverse stratigraphy at a reasonable cost. The results from pre'lnninary
slope stability analyses indicated that the landfill stability is primarily dependent on the
presence, consolidation, and strength behavior of the Qr~ and the Q~. The results from these
analyses and geotechnical site characterization studies were then used to select representative
cross-section,s, or profiles, and locations for installation of instrumentation. The criteria used
for selecting representative profiles included the presence and characteristics of the Qrc and
Qgl layers and the planned height and side slopes of the landfall.
Along a profile, the instruments are typically located in groups referred to as dusters.
In general, the cluster locations were successfully selected so that the instruments would
be positioned within a potential failure surface zone. Consideration was also given to
locating the clusters near an existing or future bench so as to minimize the amount of
lateral trenching across the landfill (to protect instrument signal cables) and to facilitate
installation, manual readings, and future maintenance. Typically, three to four instrument
clusters are installed within a profile (see Figure 3). One cluster is generally located near
the toe, one near the mid-point of the existing slope, and one or two dusters are located
inward of the present landfill crest, so that they will be below the edge of the crest of the
future landfill stages.
Leaend
I Open Standpipe Push-Type Piezometsr Cluster 3
U~ Instrument Cables I . . . . . . . . . . .
9 Inclinometer with Magnetic Extsnsorneter Placed in Trench . "
[] Closed Push-Type and Drop-In ~ 2 \ . . . " Planned Refuse Fill
Type Piezometsr
Cluster I ~ "~
-ndosures for R e m o ~ - - - - - ~ J JL
~onitofing Units (RMU)IB v J "i~" I Existing Refuse Fill
I I Qrc (Max. Thickness = 15 ft)[

.L
9 ii" I Qul (Max. Thickness = 40 R)I
II B S~II
Figure 3 - Typical Profile Instrumentation
In general, the distribution of instrument types within the clusters at each profile is designed
to measure the following:
9Pore pressures in the cohesive foundations softs (Q= and QgO;

9 Pore pressures in the sand layer(s) (Q, and 04,);
9 Pore pressures in the refuse fall;
9 Vertical deformation (settlement or heave) and horizontal deformation of the refuse fall
and the foundation soils; and,
9 Temperature within the refuse fall.
The instruments were designed and installed in a phased approach to follow the progress
and location of refuse filling. This phased approach also spread the cost over several years and
allowed for the refinement of instrument designs and installations based on experience gained
during previous phases. A total of three (3) instrument installation phases were performed
between 1992 and 1997 resulting in a total of sixteen (16) instrument profiles. The majority
(13) of these profiles have been located in Sections 1/9 and 6/7 since these sections continue to
receive refuse fall.
THOMANN ET AL. ON STABILITY MONITORINGSYSTEM 211
Instrument Types
To date, over 260 geotechnical instruments have been installed within the refuse fall and
foundation soils to monitor both the magnitude and rate of change of pore pressure, lateral and
vertical movements, and temperature. These instruments and their associated components
(e.g., electrical cables, inclinometer casing readout locations, etc.) had to be designed to
survive on-going landfilling and closure activities, and environmental hazards at and below the
landfill surface. These hazards include temperatures within the refuse fall in excess of 65~
(150~ temperatures within the foundation soils near the refuse fall in excess of 32~ (90~
lateral and vertical movements within the refuse fall on the order of tens of centimeters and
meters, respectively, and the harsh chemistry of the landfill leachate. An Automated Data
Acquisition System (ADAS) was also installed because of the large number of instruments to
be installed, difficult access conditions, large areal coverage, and the necessity to make timely
changes in landfill operations in case of instability concems.
Piezometers
The following six types of piezometers were selected for making pore pressure
measurements (Dunnicliff 1988):
9 Closed push-type vibrating wire (VW) piezometers;
9 Dual closed push-type VW piezometers;
9 Closed drop-in type VW piezometers;
9 Closed drop-in type vibrating strip (VS) piezometers;
9 Open standpipe push-type piezometers.
The closed push-type VW piezometer refers to a single pressure sensor installed at the end
of a 1.5 rn (5 It) long rod with special couplings and attachments. The dual version of this
piezometer contains two VW sensors along a maximum 4.6 m (15 ft) long rod. The spacing of
the transducers along the rod is either 1.5 m (5 ft), 2.1 m (7 It), or 3.0 m (10 It). The purpose
of this version is to measure pore pressure at two different elevations within the same borehole.
This arrangement eliminates the need for an additional boring and results in reduced installation
costs.
The closed drop-in type VW piezometer and VS piezometer refer to a pressure sensor that
is placed in a small sand-filled burlap bag and lowered into the borehole to a predetermined
location. These types of piezometers are only installed in the upper portions of the refuse fill.
The body of each sensor used in the VW and VS piezometers is made of stainless steel and
contains a thermistor (temperature sensor). The falter stone in front of the pressure-sensitive
diaphragm is made from sintered stainless steel having a maximum opening size of 50 microns
and is therefore considered a low air entry falter.
The open standpipe push-type piezometer consists of a porous tip attached to a

polypropylene riser pipe. The porous tip consists of sintered bronze manganese with a
maximum opening size of 50 microns (i.e., low air entry) and is designed to be pushed from
within a borehole to the desired location. The riser pipe consists of Schedule 80 polypropylene
pipe. To protect this pipe from lateral and vertical deformations within the refuse fill, a
Schedule 40 steel pipe is installed around the polypropylene pipe. To allow for automated
measurements, a VW pressure sensor is installed within the riser pipe. One of the primary
advantages of an open standpipe piezometer is that it allows for the retrieval and replacement,
if necessary, of the sensor and/or permits the water level to be read manually.
Inclinometers
Measurement of horizontal deformations are made in inclinometer casings that are installed
in boreholes (Dunniclitf 1988). These casings are constructed of self-aligning pultruded
fiberglass. Measurement of the angular distortions (i.e., tilt) along the length of the casing is by
means of a portable inclinometer probe. The probe is manually lowered and raised in the
casing and the readings are taken at 0.61 m (2 ft) intervals. The readings are recorded in a
portable computer and then transferred to a desktop computer for further data analysis.
The refuse fill generally undergoes large vertical deformations relative to the foundation
soils. To accommodate this, telescoping coupling sections are provided in 1.5 m (5 ft) lengths
in the refuse fall in order to minimize the development of "down drag" forces on the
inclinometer casing.
Magnetic Extensometers
Vertical deformation measurements within the refuse fill and foundation soils are made
using a series of magnetic extensometers installed at pre-determined locations along the outside
of each inclinometer casing (Dunnicliff 1988). Two types of magnetic extensometers were
used for this project: datum ring magnets and spider magnets. Datum ring magnets are fixed to
the bottom of the inclinometer casing and provide a reference point. Each spider magnet is
fitted with six leaf springs that are extended by remote means fi'om the ground surface during
installation and serve to anchor the magnet into the surrounding soil or refuse fill. The spider
magnets are not fixed to the casing; therefore, they move with the soil or refuse fill as it
compresses. The location of each spider magnet relative to the datum magnet (and hence the
vertical movement of the surrounding refuse fill or soil) is determined by using a magnetic reed
switch probe which is manually lowered into the casing.
Temperature Probes
Vibrating wire temperature probes are used to obtain temperature measurements at varying
depths within the refuse fill (Dunnicliff 1988). These measurements were useful in designing
instruments within the refuse fill. In addition to the probes, thermistors are included within
each vibrating wire and vibrating strip piezometer sensor. These provide additional
temperature data and are also used in applying temperature correction factors to VW
piezometer readings. The vibrating wire probes and the thermistors generally provide similar
temperatures.
Automated Data Acquisition System (ADAS)
The ADAS system is used to remotely obtain data from the piezometers and the
temperature probes. The ADAS system allows for nearly real-ftme readings of the instruments,
if needed. The ADAS system also reduces the potential for human input errors and the time
needed for manual monitoring and data entry.
The ADAS consists of two main components: a set of individual Remote Monitoring Units
(RMUs) and a Central Network Monitor (CNM). Typically, for a given profile, one RMU is
installed to collect data from the instrument cluster installed at or near the toe, while another
RMU collects data from the dusters installed at the edge of the working crest or at an
intermediate bench. Each RMU consists of a Measurement and Control Unit (MCU), an
antenna for transmitting data and receiving instructions from the CNM, and a solar panel to
provide power. The MCU is a microprocessor-controlled data acquisition unit for taking
instrument measurements. The CNM consists of a personal computer, a Network Repeater
Unit (NRU) and an omni-directional radio antenna. The NRU provides the interface between
the computer and the RMUs.
Instrument Performance
The first phase of instrument installations was completed in 1992; therefore, there is over 5
years of instrument performance that can be analyzed to determine the survivability of the
instruments within the landfill and the foundation soils. This information was used to develop a
basis for estimating future instrument replacement costs.
Piezometers
The open standpipe piezometers have been very durable to date. Although failures of this
type of instrument have been rare, many of these piezometers are located at the toe of the
slopes where the vertical and horizontal deformations are relatively small. In addition, the
piezometers installed within the landfill were generally located in areas where refuse filling
operations were located relatively far away from these piezometers.
Out of a total of 139 dosed vibrating wire piezometers installed in landfill Sections 1/9 and
6/7, 104 piezometers of these type are currently fimctioning properly. A statistical analysis was
performed to determine an Annual Percent Failure (APF) rate for these piezometers. The APF
is defined as the ratio of the number of properly functioning instruments existing at the
beginning of a one year period to the number of instruments not functioning at the end of the
same one year period. The APF can be calculated for any time interval; one month intervals
were chosen for this statistical analysis. The frequency distribution of the APF for the closed
vibrating wire piezometers is shown in Figure 4. The relatively high APF rates of 50%/yr and
more were encountered after the first phase of instrument installations when the total number
of instruments was relatively small. Therefore, a small change in the number of non-
functioning instruments would result in a large change to the APF rate. More recent APF rates
Figure 4 - APF Rate for Closed Piezometers
range from about 1O~ to 15*/dyr. A value of 14%/yr was selected for estimating the
number ofpiezometers that may need to be replaced on a yearly basis.
Inclinometers
In general, inclinometers are installed at the toe of the landfall and within the landfall. Due
to the relatively small vertical and horizontal deformations occurring at the toe of the landfill,
no failures of the inelinometers at these locations have occurred to date. Failure is defined as
not being able to advance the inclinometer probe to the bottom of the casing. A total of
eighteen (18) inclinometer casings have been installed within the landfiU of Sections 1/9 and
6/7. The number of inclinometers is too small to perform meaningful statistical analyses;
therefore, an Annual Number of Failures (ANF) is calculated. In general, the ANF has
increased over time because of the greater number of inclinometer casings installed. Based on
the more recent years of data and the current number of inclinometer casings, it is estimated
that approximately four inc'lmometers located within the landfill will need to be replaced
annually.
Magnetic Extensometers
Since all the vertical deformation measurements are referenced to the datum ring magnets
near the bottom of the casing, it is critical that the magnetic reed switch probe be
advanced to these magnets located near the bottom &the inclinometer casing. In
approximately 20% of the inclinometer casings located within the landfill, it is no longer
possible to advance the probe to the datum ring magnets. Therefore, vertical deformation
measurements in these casings can no longer be obtained.
Temperature Probes
A relatively few number of temperature probes have been installed within the refuse fill of
the four landfill sections. Out of a total of 17 probes installed, only 4 are currently
functioning. Approximately 50% of these probes stopped functioning or indicated large
temperature variations after approximately two years. As indicated previously, these
probes have been primarily used to confirm that the other instruments were properly
designed to resist the high temperatures within the refuse fill. They have not been
replaced with the same frequency as the other instruments. The reason for the relatively
large failure rate of these instruments is not known.
Instrument Measurements
Pore Pressures
The pore pressures witlfm the refuse fill typically increase when additional refuse fill is
placed above them and then typically decrease with time. In addition, there appears to be some
correlation between the pore pressure and the refuse fill thickness, as shown in Figure 5. The
four piezometers within the box of the figure have high pore pressures when compared to pore
pressures from other piezometers. At two of these piezometers, placement of approximately
180
160
140 9 9
120
O o
== 100
u)
.= 80
13.
0 6O O o
Q.
4O 9
. 2 gee
20
0 n I I I ~ I I ~ n i'84 n i ~ I I I I I t I I n I n
0 10 20 30 40 50
Refuse Fill Thickness Above Sensor (m)
Figure 5 - Pore Pressure within Refuse Fill
18.3 m (60 it) of refuse fill in the vicinity of these piezometers resulted in pore pressure
increases on the order of 9.1 m (30 it) to 12.2 (40 it), with the greatest increase occurring in
the piezometer located closest to the area of refuse fill placement. Approximately 6 months
after stopping refuse filling, the pore pressures have dissipated only slightly. The other two
piezometers have had relatively high pore pressures since their installation and the pore
pressures from these piezometers may be measuring gas pressures since these piezometers may
be above the phreatic surface.
The pore pressure within the recent silt and clay (Q~) stratum of Section 1D generally
increases with increasing refuse fill thickness, as shown in Figure 6. Also shown on this figure
are the pore pressures from those piezometers installed in the strata above (i.e., refuse fall) and
below (i.e., recent sand, Q,,) the Q,~. The relatively high pore pressures within the Q~ indicate
that this stratum is continuing to consolidate.
450 -
9 RecentSilt and Clay - Qrc 1
400
[] RecentSand beneath Qrc cl S 9 9
i1.
350 ~/~ RefuseFill within 15 tt of top of Qr
~- 300 e9
o
250
9 |e
o0
200 09 9 9
r
150 e9 9 &
.=
9 [] ~a o
1oo
o - _e'. , D ~ ~ D
Q. oe
50
0
0 10 20 30 40 50
R e f u s e Fill T h i c k n e s s ( m )
Figure 6 - Pore Pressures within Section 1/9
The pore pressures within the overconsolidatcd Glaciolacustrine Clay (Q~) strata have
remained relatively low when compared to the Q~ stratum and show very little increase with
increasing refuse fill thickness. Similar trends have been observed from piezometers installed in
other strata.
THOMANN ET AL. ON STABILITYMONITORING SYSTEM 217
Horizontal Deformations
The inclinometers indicate the horizontal deformations that have taken place since their
installation. In all cases, the inclinometer casings were installed after the placement of refiise
fill. As a result, the measured horizontal deformations only represent a small portion of the
total horizontal deformations that have taken place since landfilling began. In addition, since
the inclinometer casings were not all installed at the same time, the total horizontal
deformations at some inclinometer locations may be greater than others only because they were
installed earlier.
The horizontal deformation rate at each inclinometer casing location is shown in Figure 7.
The rate is calculated at the location where the greatest horizontal deformations have occurred,
which is always within the refuse fall. There are many factors that influence the horizontal
deformation rate and not all of these factors are reflected in Figure 7; however, this figure is
intended to provide information that may be useful for designing instruments within landfills or
other compressible soils. The results indicate that for refuse fill thicknesses less than
approximately 18.3 m (60 ft), the horizontal deformations are relatively small. However, for
refuse fill thicknesses greater than approximately 30.5 m (100 ft), there is a wide range in rates
with some inclinometers having relatively constant deformation rates for almost three years.
The higher deformation rates at greater refuse fill thicknesses may be due to greater
deformations occurring in the more recently placed refuse fill.
0.6
~ Maximum Yearso ~ 0.5
A
Measureme~
0.5 (~ Minimum r 2.7
E
n, 0.4
t-- 1.5
o
02 9
2.9
Q
0.2 0.5
E
L,.
o 0.1
0.0 . . . . . . . I . . . . I . . . . . . . . I . . . . I . . . .
0 5 10 15 20 25 30 35 40 45
Refuse Fill Thickness (m)
Figure 7 - Horizontal Deformation Rate
Vertical Deformations
The vertical deformation rate at each inclinometer location is shown in Figure 8. The rate is
calculated using the measurements fi'om the magnet located closest to the landfill surface.
Almost all of the vertical deformations occur within the refuse fill. As expected, there is a good
correlation between the vertical deformation rate and the refuse fill thickness. As with the
horizontal deformation rates, there are many factors that influence the vertical deformation rate
and not all of these factors are represented in Figure 8; however, this figure is intended to
provide information that may be useful in the design of instruments within landfills or other
highly compressible materials.
2.6 18
2.4
~- 2.2 Maximum
E 2.0 C) M n mum
II)
,-, 1.8
n, 1.6 Years of
c Measureme~
o
~= 1.4
.e 1.2
~ 1.0
D .7
0.8
o
0.6
> 0.4 ~3 O.4
0.2 o=oTs
0.0 . . . . . . . . . , . . . . , . . . . . , . . . . . . . . . . .
0 5 10 15 20 25 30 35 40 45
Refuse Fill Thickness (m)
Figure 8 - Vertical Deformation Rate
Use of Instrumentation Measurements for Monitoring Stability
Documents and procedures have been developed for this project to aid in the
interpretation of the instrument measurements and to assess the overall stability of the
landfill side slopes. One key document that is produced is a monthly monitoring system
report. This document presents the measurements in graphical form and is useful in
assessing overall spatial and temporal trends in the data. However, due to the changes in
the configuration of the landfill side slopes due to active refuse filling and the large volume
of data collected, it is difficult to rapidly assess the stability condition of the instrumented
profiles. Therefore, a unique computer program has been developed to perform
preliminary analyses based on a portion of the key instrument measurements. The input to
this program is data from the piezometers and inclinometers. The primary output is an
action level for each profile, which is determined based on the factor of safety against
slope instability, the maximum shear strain rates within the foundation soils, and the
changes in these parameters over time. Associated with each action level are contingency
measures that range from maintaining normal operating procedures to stopping refuse fill
placement and moving operations to a contingency area. The action levels provide a
relatively straightforward and frequent method of informing the client so that grading
plans can be developed that take into consideration stability concerns.
Two other key documents include the Slope Stability Monitoring System Operation &
Maintenance Manual and the Slope Stability Monitoring System Report. The O&M
manual is updated on a yearly basis and includes detailed information concerning the entire
monitoring system, information regarding the identification of action levels and
contingency measures, the preparation of monitoring system reports, and general
guidelines for the evaluation of measurements. The slope stability report is updated on a
yearly basis and includes detailed information regarding the performance of the
instruments, detailed interpretation of the measurements, and detailed slope stability
analyses.
Conclusions
A geotechnical instrumentation system has been designed and implemented to monitor the
stability of the Fresh Kills Landfill in New York City. To date, over 260 geotechnical
instruments have been installed within the refuse fall and foundation soils for the primary
purpose of monitoring both the magnitude and rate of change of pore pressure, and the lateral
and vertical deformations. The conclusions regarding the monitoring system are as follows:
9 The instruments, which include closed vibrating wire piezometers, open standpipe
piezometers, inclinometers, magnetic extensometers, and temperature probes, have been
successfully designed and installed to acammmodate on-going landfilling and closure
activities, high temperatures, large movements, and the landfill leachate.
9 Based on the collection of over 5 years of instrument performance data from 139 vibrating
wire piezometers, the failure rate ranges from about 10%/yr to 15%/yr.
9 Based on the past performance and the current number ofinclinometers within the landfill,
it is estimated that approximately 4 inclinometer casing will need to be replaced annually.
9 The horizontal and vertical deformation rates presented herein may be useful in the design
of instruments within landfdls or other highly compressible materials.
9 Action levels and associated contingency levels have been successfully used to monitor the
stability of the landfall side slopes and to provide information to the client for use in the
development of grading plans.
Acknowledgments
The design, installation, and monitoring of the instrumentation system for the Fresh Kills
Landfill started in 1989 and is still on-going URS Greiner Woodward Clyde has performed all
aspects of this program under contract with the New York City Department of Sanitation. The
authors wish to acknowledge the contribution to this work by their colleagues: Melvin Esrig,
~fm Fan, Drina Ferreira, Aaron Goldberg, Harry Horn, and Niels Jensen. Also, Charles Ladd
of the Massachusetts Institute of Technology and John Dunnicli~ Consultant, have provided
valuable discussions as members of the Technical Review Board. Finally, the authors want to
express their appreciation to Mr. Phillip Gleason, NYCDOS Director of Landfill Engineering,
for his support and guidance on this program~ His continued interest has been instrumental to
the overall success of the project.
References
Duplanic, N., 1990, "Landfill Deformation Monitoring and Stability Analysis," Geotechnics of
Waste Fill, ASTMSTP 1070, Landva and Knowles, Eds., pp. 303-239.
Handfelt, L.D., Koutsoftas, D.C., Foott, R., 1987, "Instrumentation for Test Fill in Hong
Kong," Journal ofGeotechnicalEngineering, Vol. 103, No. 2, pp. 127-146.
Oweis, I.S., gftlls, W. T., Leung, A., Scarino, J., 1985, "Stability of Sanitary Landfills",
Geotechnical Aspects of Waste Management, American Society of Civil Engineers
Metropolitan New York Section, pp. 1-30.
Peck, R. B., 1969, "Advantages and Limitations of the Observational Method in Applied Soil
Mechanics" 9th Rankine Lecture," Geotechnique, Vol. 19, No. 2, pp.171-187.
Withiana, J.L., Tarvin, P.A., Bushell, T.D., Snow, R.E., G e n n a ~ H.W., 1995, "Prediction and
Performance of Municipal Landfill Slope", Geoenvironment 2000, ASCE Special
Pubfication No. 46, pp. 1005-1019.
Instrumentation for Monitoring
Settlement and Stability
Vincent Silvestri ~
Field Measurement of Water Contents and Densities by Nuclear Methods in Clay

Deposits
Reference: Silvestri, V., "Field Measurement of Water Contents and Densities by

Nuclear Methods in Clay Deposits", Field Instrumentation for Soil and Rock, ASTM
STP 1358, G. N. Durham and W. A. Mart, Eds., American Society for Testing and
Materials, West Conshohocken, PA, 1999.
Abstract : In recent years natural water content and density determinations have been
carried out in the stiff clay deposits on Montreal's Island by means of nuclear depth
probes. The probes are lowered down aluminum access tubes that are permanently
inserted into the ground, thus providing a means of measuring the volumetric water
content and the density of the soils surrounding each tube, at any convenient time and
depth. Access tubes have an external diameter of 50.9 to 57 mm and a wall thickness of
1.3 to 1.9 mm, and are sealed at the bottom to prevent entry of water. The tubes are
pushed into holes of the same diameter which have been reamed out of the soil by a
mechanical auger. Tubes have been inserted to a maximum depth of 6 m. Readings in
each access hole are taken at 100-200 mm depth intervals along the length of the tube.
Values of volumetric water contents and densities are determined by using field
calibration curves. The paper gives the results obtained on several projects, reports on
the precision of the predicted values and presents typical problems which were
encountered during installation of the access tubes and the long-term (up to 8 years)
monitoring programs. It is also shown that factory calibration curves furnished with the
nuclear probes cannot be used in high-water content clay soils. It is further indicated that
even though the water content calibration curves show a high degree of correlation, the
neutron scattering technique will not effectively detect changes in soil water content for
periods ranging from a few days to about one week. Finally, it is shown that voids
around the access tubes adversely affect the nuclear gage readings.
KEYWORDS: water content, density, field, nuclear depth gages, long-term program,
precision, clay
Introduction
The use of subsurface gamma and neutron probes for the measurement of soil
density and water content, which originated nearly five decades ago (Belcher and
Cuyckendall 1950, Gardner and Kirkland 1952), has become common practice in
geotechnical engineering. However, despite the progress in understanding the physical
phenomena involved and the improvements in equipment design, there is still much
apprehension pertaining to the accuracy in the use of the equipment (Tan and Fwa 1991).
Reasons commonly invoked for uncertainties associated with gage accuracy are soil
1 Professor, Department of civil, geological and mining engineering, Ecole Polytechnique, P.O.
Box 6079, Station Centre-ville, Montreal (Quebec) Canada H3C 3A7.
composition and porosity effects, and adequate methods and procedures for relevant
calibration (Tan and Fwa 1991, Ruygrok 1988).
This paper focuses on the use of nuclear density and water content depth gages in
the stiff clay deposits on Montreal's Island (QC). It reports on the accuracy of the field-
measured values, and presents problems encountered during the installation of the access
tubes and the subsequent monitoring programs.
The use of nuclear measurements was found necessary because the clay deposits
had to be monitored for several years in order to determine the relationship between soil
shrinkage and evapotranspiration. The study was triggered by the severe summer
drought of 1983 in the Province of Quebec which affected surficial clay deposits in urban
areas and caused substantial structural damage to a large number of buildings founded on
these clays. Damage resulted from excessive differential foundation settlement.
Materials and Methods
Nuclear Probes
Two standard Campbell Pacific Nuclear Corp. (CPN) water content/density and
water content gages, models 501DR and 503D1L respectively, were used in several stiff
clay deposits on Montreal's Island. Each gage was supplied with a 10 m long cable and
movable cable stop. The cable was marked at 100 mm intervals. The bottom of the gage
contains an oversize hole to allow inserting an adapter ring with a diameter to match the
type of access tube being used. With the gage sitting on top of the access tube, the probe
is lowered to the depth of measurement and the cable is clamped. Upon retraction of the
probe into the shield inside the gage, the probe locks automatically in place for transport.
The probes were lowered in thin aluminum access tubes, which had been permanently
inserted into the ground. Two sizes of aluminum tubes were employed: a) 50.9 mm
outside diameter with 1.33 mm thick wall, and b) 57 mm outside diameter with 1.9 mm
thick wall. The tubes were pushed into pre-drilled boreholes of essentially the same
diameter. Soil samples were recovered in these boreholes by means of either split spoons
or shelby tubes. While split spoon samples were used only for identification and water
content determinations, shelby tubes samples served for water content and bulk density
determinations.
The maximum length of the access tubes used in these projects was 6.1 m. The
upper ends of the access tubes were protected with vandal-proof caps; their lower ends
were sealed with fiat plates welded to the tubes.
Field Calibrations
The calibration equations for density and water content content, respectively, take
the form (Morris and Williams 1990):
Ra = F (water content, density, dry composition) (I)
and
SILVESTRI ON WATER CONTENTSAND DENSITIES 225
Rm = G (water content, density, dry composition) (2)

where Ra is gage response for density, and Rm is gage response for water content.
Each gage parameter, Ra or R~ in Eqs. 1 and 2, represents the count ratio (CR),
that is, the ratio of the field count to the equivalent standard count (SC). The SC is the
mean value of a series of counts of equal duration, taken with the gage in a fixed position
in a medium whose composition and dimensions are stable (Morris and Williams 1990).
The use of ratios rather than experimental counts alone minimizes the effects of
electronic drift, source and detector decay, and aging.
Following the manufacturer's recommendations the two probes were recalibrated
in the field because the soil characteristics were much different from those of the sand
used for the original factory calibration of the equipment. Field counts were taken in the
access tubes at the exact positions of the soil samples which had been previously
recovered in the boreholes.
Density and Water Content Determinations
Nuclear probe readings in the access holes were usually taken at 100 mm intervals
along the length of the aluminum tubes. This interval was increased to 200 mm in the
deeper soil layers, and also when there was no evidence of any seasonal variation.
For the water content content determination, a 16s count period was used for all
measurements. For density determinations, 64s readings were employed. The volumetric
water content, 0(%), and the bulk density, T(kN/m3), of the soils were determined using
the aforementioned field calibration curves.
Soils and Monitoring Programs
As the vast majority of foundation settlement problems caused by the drought of

1983 occurred in the eastern part of Montreal, field studies were undertaken in that area
with the aim of measuring the variation of soil density and water content, especially
during the subsequent summer months. Fortunately, in the period of study from 1988 to
1995, Montreal's region experienced moderate droughts in 1988 and 1991, during which
significant decreases in volumetric water contents occurred in the surficial clay deposits,
leading to excessive foundation settlements (Silvestri et al. 1992, Silvestri and Tabib
1994).
The soils found in the area at study consist of a superficial layer of fill of 0.3 to
1.5 m in thickness, followed by a 2 to 3 m thick layer of stiff brown clay in which the
natural gravimetric water content varies between 35 and 45 %. This silty clay is part of
the undisturbed sensitive Champlain Sea clay found below, in which the natural
gravimetric water content ranges from 60 to 80 % and whose thickness varies between 3
and 15 m on the various sites studied in the program. Below the clay lies a layer of till of
variable thickness in contact with fractured bedrock.
Results and Discussion
Gage Precision and Probe Calibration
First, regarding the precision of the gages, several sets of repeated readings o f
equal duration were taken on different occasions, at fixed positions in the access tubes.
For both the 503DR water content and the 501DR density/water content gages, 16s
readings showed that for an average volumetric water content 0 (%), 95 % confidence
limits on O were computed to be 0 _+ 0.7 %. In addition, for 64s readings with the
501DR gage, it is expected that for an average bulk volumetric weight ~(kN/m3), 95 %
confidence limits are ~(kN/m 3) + 0.09 kN/m a. Using the chi-square distribution for the
95 % confidence limits, the two gages were found to work properly with an acceptable
level of variation in the devices' electronics.
The precision of the predicted field values of the two gages was determined using
the field techniques previously discussed. Results are reported in Figs. 1 to 3. These
figures also present the original factory calibration curves. The observed discrepancy
between the factory and the field values shown immediately points out to the need of
recalibrating the gages whenever in situ soil properties are different from those of the
standard materials used by the manufacturer. Note that in Figs. 1 to 3 are also given the
straight-line relationships obtains by means of linear regression analyses, and as well as
the corresponding coefficients of determination, R2's, for the field calibrations of the two
gages.
Concerning the precision of the predicted field values of the mean volumetric
water content, 0 (%), the calibration curve of the 503 DR gage shown in Fig. 1 gives, for
example, the following 95 % confidence limits: a) 0 = 50.29 % _ 1.16 % for CR = 2.0,
and b) 0 = 67.55 % -+ 1.43 % for CR = 2.5. For the data presented in Fig. 2 obtained by
means of the 501DR gage, 95 % confidence limits are: a) O = 52.86 % _ 1.21% for CR
= 0.65, and b) ) O = 64.19 % _ 1.42 % for CR = 0.8. Comparison of these results shows
that the precision of the water content contents predicted by both gages is quite similar.
Finally, concerning the precision given by the 501DR gage for the predicted
values of the mean bulk volumetric weight, y(kN/m3), the data of Fig. 3 show, for
example, the following 95 % confidence limits: a) y = 19.00 kN/m 3 _+ 0.50 kN/m 3 for
CR = 2.6, and b) y = 15.96 kN/m 3 _ 0.27 kN/m 3 for CR = 3.1.
Even though the water content calibration curves shown in the preceding figures
indicate a high degree of correlation, as shown by the relatively high values of the
coefficient of determination, R 2, some remarks are necessary. Consider, again, the
calibration curve for gage 503DR reported in Fig. 1. As just mentioned, the 95 %
confidence limits on 0 were computed to be equal, for example, to 50.29 % _ 1.16 % for
CR = 2.0, using just one neutron-probe reading. For a 100 mm thick layer of saturated
clay, the quantity of stored water is thus equal to 100 O or 50.29 mm + 1.16 mm. The _
1.16 mm deviation from the mean is + 2 . 3 1 % which at first glance is not very large.
However, for the field sites tested, for which the daily evapotranspiration was calculated
SILVESTRI ON WATER CONTENTS AND DENSITIES 227
75 - :-- : 9 : :. : .: :. : : : : 9 :- :
I i i i i i i i i i i i i i i
7o ........... i ........... i............ ~............ i............ i ............ i ............ ~............ i ............ i ........... i .......... ~............ i ........... i ............ ~...........
.-. i i i i i i i i ;o i i i i i
65 ........... i ........... i ............ ! ............;~........... i ............ i ............i............ i - ~ / ~ ~ - - i .......... i ............ i ...........
:: i :: :~ i :: :: ~ "~" ; \ ' o : " . . . .
" " 6 0 [ ...........;:............i............:,............i............i............; - . - ~ - . - . . . i . . ~ . ~ ' ~F'ield C a l i b r a t i o n : [...
~.' i i i i i o i ~ 10~%)=~,,.6~(0R)-~8751
o55 ............i ...........i............!............i ...........i ~ - - ~ R 2 = 0 . ~ 7 L1/
O i :: :: :: i ~ / ::. i :: "
~ i ~ i i OQI i 0 I ! i ~ i i :
5 0 ............i ...........~............i'-'-"-o"i . . . . . . . . . . . .
~ ~5 ...........i............i. . . . . . . . . . .
~ 40 .........i . . ~ F ' a c t o r y Calibration: ......

~ [ e(O/o)=~e(cR) I
~35 / ' ~ ~ ' " " "" - '
30 i ..........
25 .... ~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1,5 1,6 1,7 1,8 1,9 2,0 2,1 2,2 2,3 2,4 2,5 2,6 2,7 2,8 2,9 3,0
Count Ratio, CR
Figure 1 - Water Content Calibration of 503DR Gage.
70
65
60 ................... i .................... ~i .................... io

~ ................
...........~. i.............
. ~ ~o ~ ~'~
~...~ ...........
"~ ......',i '~.o._._...._~
/..............
/ i i .................
........................
;....................
t-
55 ...................}....................}..................~ i e l d Calibration: ......i....................

O
i i o / - ~ ' ~ R)+3.78, i
~ 50 ................... i ............. ~ J ~ - . . I R'=0.80 ]......i ....................
o
"E
45
E
_~ 4 0 .......... ~ ~ ! 0 ( % ~ - - ~ 9 . 8 6 < c R ) J......... i .................... i .................... i ....................
>O
3 0 . . . . r , , , i . . . . . . . i . . . . i . . . . i . . . . i . . . . i . . . .
0,50 0,55 0,60 0,65 0,70 0,75 0,80 0,85 0,90 0,95
Count Ratio, CR
Figure 2 - Water Content Calibration of 501DR Gage.
using a simple hydrologic balance model with eight soil layers (Silvestri and Tabib
1994), the cumulative resulting error in millimeters of stored water could be very large, in
many cases larger than the daily amount of evapotranspiration. Under operating
conditions of one 16s reading per 100 mm depth layer and a total layer thickness of 1 to 2
m, the neutron scattering technique will not effectively detect changes in soil water
content for periods ranging from less than a few days to about one week, depending on
total root depth and the prevailing evapotranspiration rate. This limitation is primarily
due to the expectation of large variances in estimates of soil water over the total soil
profile affected by evapotranspiration. In order to decrease the instrument component of
the standard error, a/d'n, where a = standard deviation, and n = number of readings, the
procedure is both to increase the number of readings per soil layer and to decrease the
soil layer thickness, that is, the vertical distance between successive neutron-probe
readings. Note also that increasing the count time is equivalent to increasing the number
of readings per soil layer, e.g., the expected variance of four 16s counts is equivalent to
the variance of one 64s count (Haverkamp et al. 1984, Carrijo and Cuenca 1992).
Overboring and Soil Shrinking away from Tubes
In one of the first field studies undertaken in the program, 3 to 4.5 m long access
tubes were installed in July 1988, at various distances from a selected number of trees
growing in Montreal's Maisonneuve Park. The trees retained were a 22 m high elm, a 20
m high silver maple, a 30 m high cottonwood, and a row of 18.5 m high Lombardy
poplars. The investigation aimed at determining the maximum spreading of the drying
front caused by water absorption by tree roots.
Due to lack of adequate drilling equipment, some of the access tubes installed at
the beginning of the study were inserted in boreholes that were slightly too large. This
resulted in two problems. First, due to the fact that the surrounding soil was not in
contact with the tube, initial experimental counts predicted rather low water contents
compared to the values obtained by means of laboratory measurements on the soil
samples which had been retrieved during boring of the access holes. Such a phenomenon
is illustrated in Fig. 4 for the depth interval ranging between 1.6 and 3.0 m. Second, as
the monitoring program went on, it was found that when count readings were taken in the
hours following heavy rain storms, volumetric water content values increased
considerably. Such a phenomenon was caused by rain water filling the cavity between
the soil and the tube. A similar response occurred in late spring and fall when the water
table rose near the surface. Such a behavior is also shown in Fig. 4, for the readings
taken, for example, on May 16, 1989.
In addition, because the access tubes had been installed in the summer, an annular
cavity began to progressively form around them as the soil continued to dry and slowly
moved away from the metallic surface. The diameter of these cavities was maximum at
ground surface and decreased almost linearly with depth. At the end of the summer, the
maximum thickness of these cavities varied between 20 and 30 mm, and their depth
reached approximately 1 m. These cavities were found to form particularly around tubes
inserted quite close to the trees, typically at distances of less than 1 H, where H represents
the height of the tree.
In order to minimize the formation of such cavities, the access tubes should be
installed at the end of the summer, when soil shrinkage caused by evapotranspiration has
already attained its maximum and the soil is at its lowest water content content.
21
2O I ........................................
!o ...........................................
~.............................................
~...............................................
i................................................
I Field Calibration: I.
.......................
i..........
o ................................
17( kNIm3)=-6.07(CR)+34.80, I........
~18 .......................~ 1 7 6............... , I...........
e.- ............................................
i. . . . . . . . . . . . . ~"T .... ;;0% ~ o o~o [i o
m15
...............
[ 7(k_N/m3)=-6.2(CR1+ 33.5 .......................................................
iI................
~ .....................
14
13 i-"'~
2,4 2,6 2,8 3,0 3,2 3,4
Count Ratio, CR
Figure 3 - Density Cafibration of 501DR Gage.
Volumetric Water Content, 0 (%)

20 25 30 35 40 45 50 55 60 65 70 75 80
0,0
Lobora,o ,pec, en, i
0,5 ..1 (08-30-88) ........
~ ~ ~ . . , ............
...............
,...............
~-,, .... 09-29-88 ~ ~ t~" i i
1,0 .j 9 08-30-88 ...............
i......~.~l ............~ .............
i.............
i...............
/ ~, 11-28-88 i "~Tr-~ - ~ i i
1 9 05-16-89 i ~ d ~ i i
1,5 -| --=-- 09-11-89 ...........~ i i ~ : ; ~ ............i..............i ...............
E |'~ 06-11-90 i ~ "~--"x~ i i
..E
~. 2,0
O
2,5
3,0
3,5
4,0
Figure 4 -Effect of Cavity on Water Contents.
Water Content Variations
For civil engineering purposes, water content is usually expressed in terms of a

gravimetric ratio w (%), that is, the ratio between the mass of water and that of the dry
soil. Such an approach is easily implemented in the laboratory as masses can be directly
determined. However, when long-term, monitoring programs are carried out in the field,
it becomes impractical to recover soil specimens and to repetitively determine
gravimetric water contents. As a consequence, the volumetric water content 0 (%),
which is equal to the volume & w a t e r divided by the total volume of the soil, is preferred
over the gravimetric ratio w, because it can be directly determined using a nuclear gage.
In addition, variations in 0 can be related to soil settlements for as long as the material
remains saturated. For example, in the stiff clays of Montreal, Silvestri et al. (1992)
reported that, as the soils remained saturated for volumetric water contents a's low as 40
%, the decrease of one percentage point in 0 corresponded to a vertical settlement of 1/4
to 1/3 (%). This finding implies that for a 10 % decrease in 0 in a 1 m thick soil layer,
the vertical settlement would probably range between 25 and 33 mm.
Typical field water contents obtained using the two gages are presented in Figs. 5
and 6. The data shown in Fig. 5 correspond to water contents measured in access holes
drilled on the unwatered lawn around a residential building. One access hole, that is,
TA2-1, is located on the front yard, at a distance of 3 m from a 13.7 m high red ash. The
other, that is, TA2-3, is situated on the back yard, far away from any tree. The profiles
shown for each access holes were obtained when the soil deposit was in its driest and
wettest states during the year 1991. Examination of the two profiles for tube TA2-1
indicates that evapotranspiration proceeded to a maximum depth of 4.2 m. As for the soil
conditions in the back yard, the two profiles obtained for tube TA2-3 show almost no
climatic effect in the soil layers studied. The large variation in the water contents around
tube TA2-1 resulted in a surface settlement of approximately 85 mm.
Figure 6 reports water contents measured on a watered site adjacent to the one for
which the data is shown in the preceding figure. In the present case, tube TAI-1 is
located on the front yard o f a residentiaibuilding, at a distance of 3.4 m from a 13 m high
red ash. As for TA1-3, it is situated on the back yard, far away from any tree. A
comparison between the driest and wettest profiles measured around tube TAI-1
indicates that evapotranspiration proceeded to a maximum depth of 2.2 m, in spite of
continuous irrigation by means of a trickle system. As for the back yard, the two profiles
show that the maximum depth affected by evapotranspiration does not exceed 1 m.
In order to compare the results obtained by the two gages, readings were taken
with both of them on almost the same days, in the same access holes and at the same
depths. The results are shown in Fig. 7. Examination of the data indicates an almost
perfect correlation. The straight line relationship and the coefficient of determination R 2
are also given in the figure.
Volumetric Weights
Data obtained with the 501DR gage are shown in Fig. 8. Readings were taken in
access hole TA2-1 of Fig. 5. Comparison of the data given in these two figures indicates
V o l u m e t r i c W a t e r Content, e (%)
38 44 50 56 62 68 74
0
~3
~4
s ....... _-'............Tui~eT~:los;is:9~ ..........................
i.......... ~ ............................
" ii......................................
,-~
6 ' ~ TubeTA2-3, 05-15-91 .........................................................................

i..........................................................................................
I --o- TubeTA2-3, 09-18-91 I i
F i g u r e 5 - Water Content Profiles around Tubes TA2-1 and TA2-3.
V o l u m e t r i c W a t e r C o n t e n t , e (%)
38 44 50 56 62 68 74
0
~3
B4
5
............. i ...............................
Tube TAI-1,04-25-91 l~'5 ~ ~ -- i --.
6 9 T u b e T A 1 -3,' 0 9 - 1 8 - 9 1 [ / ...................:~...............................~................................~:................................
l
,--o
,~ T u b e T A 1 - 3 , 04-25-91 ]
7 . . . . . i . . . . . i . . . . . i . . . . . i . . . . r . . . . .
F i g u r e 6 - Water Content Profiles around Tubes TAI-1 and TA1-3.
68
[+01~-.~--1.005m03D.~-0.683,
I i .~i
~64
t~
....................... :~....................... : ' 1 R 2 = ~
i :: ~ ~ ! o
I
........ ~ ' " .......... Z
el
.................
+
i i i ': :: oe ~ oo i +
"~ 60
....................... i~.......................':i....................... ':i.......................~
::....................... !i............
~ 9 ee":: ~ ........................
+ + +
~" 56
t-
+
i
+
i
+
i
~+ 9 +,,,,. i
+i :
~ 52
0
~ 48 ....................i.......................::..................Ih ................. ~.......................i.......................i.......................~........................

i 9 i,. " ~ :: i :: :: i
.~ 44 ...................... +................ ~ - - - - - + + +
E
_~ 40
0
p+ + .......... + + p+
..................... ~- J'-+"o+-'""i ....................... i ....................... i ....................... ":....................... +:....................... i ........................
36
36 40 44 48 52 56 60 64 68
Volumetric Water Content, 0 (%), by 503DR
Figure 7 - Comparison between Water Contents Given by Gages 501DR and 503D1~
Bulk Density, ~, (kN/m 3)

15,0 15,5 16,0 16,5 17,0 17,5 18,0
0
v
E
s
a
5 --o-- 08-06-91 .................i ...............................i .............................................................
J ~ 02-05-91
6 . . . . i . . . . n . . . . i . . . . n . . . . + . . . .
Figure 8 - Volumetric Weight Profiles around Tube TA2-1.
that, as expected, whenever the volumetric water contents increase, the bulk volumetric
weights decrease, and vice versa.
In addition, to appreciate the effect of the existence of annular cavities around

access tubes, measurements of volumetric weights were carded out in the tube for which
water contents are reported in Fig. 4. The data are shown in Fig. 9. Analysis of the two
relationships given in this figure indicates that in the region affected by the cavity, that is,
in the depth interval between 2 and 3.4 m, the bulk unit weight is unrealistic low, i.e.,
around 14.4 kN/m3, for an unrealistic high average volumetric water content of 76 %
when the probable value ofe was determined to be in the range from 55 to 65 %.
Finally, using both the field density and water content values determined by
means of the 501DR gage, Fig. 10 presents the bulk volumetric weight as a function of
the corresponding value of 0. In order to better appreciate the results, theoretical
relationships given by 7=Tw[p,0-0/lOO)/pw +0/100] are also indicated for solid
densities p, of 2.6, 2.7 and 2.8 g/cm3, assuming the soil to be saturated, with 7w = 9.81
kN/ma and Pw = lg/cm 3, Please note that the solid density p, was also determined in the
laboratory on twenty nine soil specimens and was found to have a mean value of
2.722 g/cm3 and a standard deviation of 0.029 g/cm3. The field data shown in Fig. 10
indicate that most of the values correspond to an average solid density of about 2.7 g/cm3.
Volumetric Water Content, 0 (%) Bulk Density, 7 (kN/m 3)

54 60 66 72 78 12 14 16 18 20
0,6
1,0
1,4
1,8
e-
~.. 2,2
|
Q
2,6
3,0
3,4
3,8
Figure 9 - Effect of Cavity on Volumetric Weight and Water Content Profiles.
i::iI ............................. .......... p..2.8o,cm3 ..................... , .............................. :.............................
>o.....o.~.i ...?.......................~.............................
o o o :
' )o 9 : o
16,0 ........................................................................................... ~...o..~ ~%....o. ~,~,,, ........ ~.............................
o'~ 0 % o '%
15,5 ......................................................................
Ps=2,6 g/cm3 . . ~ ~ ' ~ , ~ , .........................
~ o
15,0 ...........................................................................................................................................................
~.............................
14,5 . . . . . . . . . . . . . . . . . . . .
46 50 54 58 62 66 70
Volumetric Water Content, O (%)
Figure 10 -FieM Volumetric Weight- Water Content Relationship

Given by Gage 501DIL
Conclusion
On the basis of the contents of this paper, the following conclusions are drawn:
i) Nuclear gages can be to effectively useful for long-term monitoring programs of

water contents and densities of clay soils subject to evapotranspiration. The two
nuclear gages used in this investigation were found to perform adequately. The
precisions of the two gages were quite similar.
ii) Nuclear gages cannot effectively measure water content variations over short
periods of time in clay deposits. It is estimated that minimum time intervals
required to detect meaningful changes of water contents in clays are of the order of
one week.
iii) The installation method of access tubes is of prime importance. Methods that result
in a void space between the access tube's outside diameter and the wall of the
borehole will adversely impact readings made with the nuclear gage. Possible
formation of cavities around access tubes may be minimized by installing them
when the soil is at its driest point.
iv) Water content and density profiles obtained by means of nuclear gages give a
detailed picture of the soil state as function of time.
Acknowledgment
The author expresses his gratitude to NSERC of Canada for the financial support
received in the course of this study.
Rerefences
Belcher, D.J., and Cuyckendall, T.R., 1950, "The Measurement of Soil Moisture and
Density by Neutron and Gamma Ray Scattering", Civil Aeronautics Administration
(USA), Technical Development and Evaluation Center, Washington, Technical
Development Report 127, 1950.
Carrijo, O.A., and Cuenca, R.H., 1992, "Precision of Evapotranspiration Estimates Using
Neutron Probe", Journal of Irrigation and Drainage Engineering, ASCE, Vol. 118,
No. 2, pp. 943-953.
Tan, S.-W., and Fwa, T.-F., 1991, "Influence of Voids on Density Measurements of
Granular Materials Using Gamma Radiation Techniques", Geotechnical Testing
Journal, Vol. 14, No. 3, pp. 257-265.
Haverkamp, R., Vauclin, M., and Vachaud, G., 1984, "Error Analysis in Estimations of
Soil Water Content from Neutron Probe Measurements: 1. Local Standpoint", Soil
Science, Vol. 13, No. 2, pp. 78-90.
Morris, P.H., and Williams, D.J., 1990, "Generalized Calibration of a Nuclear

Moisture/Density Depth Gauge", Geotechnical Testing Journal, Vol. 13, No. 1, pp.
24-35.
Ruygrok, P.A., 1988, "Evaluation of the Gamma and Neutron Radiation Scattering and
Transmission Methods for Soil Density and Moisture Determination", Geotechical
Testing Journal, Vol. 11, No. 1, pp. 3-19.
Silvestri, V., Soulid, M., Lafleur, J., Sarkis, G., and Bekkouche, N., 1992, "Foundation
Problems in Champlain Clays During Droughts. II. Case Histories", Canadian
Geotechnical Journal, Vol. 29, No. 2, pp. 169-187.
Silvestri, V., and Tabib, C., 1994, "Settlement of Building Foundations on Clay Soils
Caused by Evapotranspiration", Proceedings, ASCE Conference on Vertical and
Horizontal Deformations of Foundations and Embankments (Settlement '94),
College Station, Texas, Vol. 2, pp. 1494-1504.
James B. Russell, l Christopher A. Lawrence,2 Michael W. Paddock, 3 and Douglas R o s s 4
Monitoring Preload Performance
Reference: Russell, J. B., Lawrence, C. A., Paddock, M. W., Ross, D., "Monitoring
Preload Performance," FieM Instrumentation for Soil and Rock, ASTM STP 1358, G. N.
Durham and W. A. Marr, Eds., American Society for Testing and Materials, West
Abstract: Reconstruction of State Highway 29 (STH 29) in north central Wisconsin was
complicated by two separate deposits of compressible peat and organic silt within the
highway right-of-way. The deposits ranged from about 3 to 11 meters in thickness over
about 900 meters of the reconstruction alignment. Excavation and removal of the peat
and organic silt was evaluated, but was determined to be cost prohibitive. Preloading was
determined to be a cost effective and environmentally friendly solution, and was
implemented at both deposits to induce the expected consolidation and long-term
secondary compression of the final highway embankment and pavement sections. This
paper presents the design, instrumentation, and rheological modeling procedures used to
complete the project.
Keywords: marsh, peat, organic silt, highway, settlement, vertical drains, preload
monitoring, rheological modeling
In 1997, the Wisconsin Department of Transportation completed the

reconstruction of State Highway 29 (STH 29) between the cities of Hatley and
Wittenberg, Wisconsin. The reconstruction was part of the overall reconstruction of STH
29 between the cities of Green Bay and Wausau, Wisconsin, and was undertaken to
convert the existing and heavily traveled 2-lane highway (l-lane each way) to a 4-lane
highway (2-lanes each way separated by a median). Within Marathon County, the 2
existing lanes became the new westbound lanes, and a new highway embankment was
constructed to carry 2 new eastbound lanes.
1Managing Engineer, Pare Engineering Corporation, 8 Blackstone Valley Place,

Lincoln, R102865
2Senior Project Manager, CH2M HILL, 1700 Market Street, Suite 1600, Philadelphia,
PA 19103-3916
3Senior Project Manager, CH2M HILL, 411 East Wisconsin Avenue, Suite 1600,
Milwaukee, WI 53211
4project Manager, Wisconsin Department of Transportation, 2811 8th Street, Wisconsin

Rapids, WI 54495
RUSSELL ET AL. ON PRELOAD PERFORMANCE 237
Two separate marsh deposits of peat and organic silt adjacent to the existing
highway complicated construction of the new highway embankment. At the first deposit
(Marsh A), there was about 3 to 4.5 meters of peat underlain by up to 4.5 meters of
organic silt. At the second deposit (Marsh B), there was about 3 to 6 meters of peat
underlain by up to 6 meters of organic silt. Typical sections at Marsh A and B are
presented in Figures 1 and 2, respectively.
FIG. 1-Typical Section Marsh A
Excavation and removal is the Wisconsin Department of Transportation's

preferred alternative when constructing new highway embankments through marsh.
However, at Marsh A, there was a Class A trout stream immediately south of the
proposed highway embankment. Excavation and removal of the peat and organic silt in
this area would have required a braced sheet pile wall to allow the peat and organic silt to
be removed without damaging the trout stream. The braced sheet pile wall and depth of
the marsh (up to 9 meters) made excavation and removal cost prohibitive in this area. At
Marsh B, the depth of the marsh (up to 12 meters) and height of the existing embankment
adjacent to the marsh (Figure 2) also discouraged an excavation and removal alternative
since the existing roadway would need to be braced during removal.
FIG. 2-Typical Section Marsh B
Based on the above, preloading of the marshes was evaluated and determined to
be a cost effective and environmentally friendly solution. Preloading began at Marsh A
in the fall of 1995 and at Marsh B in the spring of 1996. After inducing the expected
immediate, consolidation, and long-term secondary compression settlements, the preloads
were stripped to the pavement subgrade elevation and the pavement sections installed in
the spring of 1997. This paper presents the design, instrumentation, and rheological
modeling procedures used to complete the project.
Soils Information
Geotechnical investigations were performed at Marsh A and B to determine the

depth and stratigraphy of the marshes, and geotechnical parameters for design. The
geotechnical investigations revealed 3 main strata at each marsh, as summarized in Table
1.
TABLE 1-Generalized Soil Conditions at Marsh A and B
Average Estimated
Average Undrained Angle of
Moist Shear Internal
Type of Thickness Unit Weight Strength, Cu Friction (0)
Stratum Material (m) (kg/m 3) (kN/m2)(1) (2)
1 Peat 3-6 1,073 14.4 28
2 Organic Silt 0-4.5 1,602 9.6 20
3 Sand/Gravel 3-9 1,922 ... 35
Footnotes:
(1) Undrained shear strength, Cu, based on vane shear test results.
(2) Total stress angle of internal friction based on CU triaxial test results.
Stratum 1 is a brown to black fiberous peat. Stratum 2 is a black to grey spongy

organic silt with traces of animal shells. Stratum 3 is a well-graded sand and gravel
underlain by bedrock. The groundwater table ranged from about 0.3 to 1 meter below
existing grade.
Consolidation tests were performed on undisturbed Shelby tube samples collected

from the peat and organic silt. The consolidation tests revealed that the peat and
organic silt are normally consolidated and highly compressible, as summarized in Table
2.
TABLE 2-Typical Consolidation Parameters at Marsh A and B
Estimated Laboratory
Laboratory Secondary Coefficient of
Type of Compression Compression Consolidation, Cv
Stratum Material Index, Cc Index, C~ (mZ/s) (1)
1 Peat 1.15 0.06 1.9 x 10 -o
2 Organic Silt 0.65 0.03 2.8 x 10-7
Footnote:
(1) Coefficient of Consolidation, Cv, determined using Cassagrande's Logarithm of
Time Fitting Method.
Preload Design
The elevation difference between the existing pavement surface, and the existing
ground surface at the proposed eastbound embankment ranged from about 1 to 4 meters
(Figures 1 and 2). The finished pavement surface at the proposed eastbound
embankment would be at approximately the same elevation as the pavement along the
existing westbound embankment.
One-dimensional settlement analyses were performed using the laboratory

consolidation data to determine the preload thickness required to induce the expected
immediate, consolidation, and long-term secondary compression under the proposed
highway embankment and pavement section service loads. Based on the settlement
analyses, a 3.5 to 4-meter-thick preload embankment would be required at Marsh A,
and a 6 to 6.5-meter-thick preload embankment would be required at Marsh B.
It was estimated that about 1 meter of the preload embankment would be

stripped at the completion ofpreloading for construction of the pavement sections. The
initial preload design parameters are presented in Table 3.
TABLE 3-Initial Preload Design Parameters
Elevation Maximum Estimated

Difference Thickness of Thickness to
Between Existing Peat and Estimated Total Estimated be Stripped at
and Finish Grade Organic Silt Preload Settlement (m) Completion
Marsh (m) (m) Thickness (m) (1) (m)
A 1 9 3.5-4 1-1.5 1
B 3 12 6-6.5 1.5-2 1
Footnote:
(1) Total of immediate, consolidation, and long-term secondary over 30-years.
Embankment stability analyses were performed with the UTEXAS computer

program using undrained strength parameters (i.e., O = 0). The analyses indicated that
the preload embankments would need to be constructed in stages to minimize the
chances of a general bearing capacity or slope stability failure. Based on the analyses,
the initial preload stage thicknesses were limited to 0.6 meters, and subsequent stages
were limited to 1 to 1.2 meters. A total of 5 preload stages were planned to construct
the preload embankment at Marsh A, and 6 to 7 preload stages were planned to
construct the preload embankment at Marsh B.
Time rates of settlement analyses were performed using the laboratory

coefficients of consolidation and assuming double drainage would occur in the field.
These analyses indicated that it would take between 60 and 90 days for each preload
stage to undergo its full immediate and majority (i.e., 90 percent) of consolidation
settlement. Therefore, it was determined that up to 15 months would be required to
construct the Marsh A preload embankment (5 preload stages x 3 months/stage) and
that up to 21 months would be required to construct the Marsh B preload embankment
(7 preload stages x 3 months/stage). Paving was scheduled for the Spring of 1997. The
project schedule allowed the Marsh A preload to be constructed within this time period,
however, vertical drains were needed at Marsh B to accelerate the consolidation process
and allow preloading at this location to be completed before the spring of 1997.
Site Preparation
Prior to field instrumentation and vertical drain installation, brush and vegetation within
the limits ofpreloading were cleared and removed. The sod and grass root mat were
not removed because the root mat helped to provide a stable surface. Holes and
depressions were filled with granular fill and a woven geotextile fabric was placed over
the entire preload area. The fabric was placed perpendicular to the roadway and
anchored in the slopes of the existing highway embankment. The fabric was overlapped
a minimum of 0.6 meters at the joints. After placing the fabric, a 0.15 to 0.30 meter
thick granular fill working surface layer was placed to allow field instrumentation and
vertical drains to be installed.
The vertical drains were installed at Marsh B on a triangular grid pattern at 1.8
meters on center. The drains were installed before the field instrumentation was
installed to help minimize the chances of damaging the instrument cables. Vertical
drains (AMERIDRAIN 407) were pushed about 0.6 to 0.9 meters into Stratum 3 and cut
offabout 15 cm above the granular fill working surface layer to allow double drainage
to occur. The Barron and Kjellman (1992) formula was used to estimate the
acceleration of pore pressure dissipation. The analyses indicated that the drains would
induce the majority of consolidation settlement within about 2 to 3 weeks for each
preload stage, about 5 times faster than without the drains.
Field Instrumentation
Field instrumentation was installed at Marsh A and B and included settlement

plates, a liquid settlement gauge, and vibrating wire piezometers, as summarized in
Table 4. Settlement plates were generally installed about every 30 meters along the
preload embankment. The plates were used to track the actual preload settlement and a
schematic of the settlement plates used is presented in Figure 3.
A liquid settlement gauge (GEOKON Model 4600) was installed at Marsh A to

serve as a back-up instrument in case the settlement plates were damaged, and to
corroborate the settlement plate data. Vibrating wire piezometers (GEOKON Model
4500AL) were installed in the center of the consolidating layers to monitor pore
pressure dissipation and help determine when consolidation settlement for each stage
was complete and if the next preload stage could be placed.
TABLE 4-Field Instrumentation Summary

Vibrating
Preload Length Settlement Liquid Settlement Wire
Marsh (m) Plates Gauge Piezometers
A 700 21 1 7
B 200 6 ... 2
242 FIELD I N S T R U M E N T A T I O N FOR SOIL A N D ROCK
E 100mm Sch. 40 PVC pipe,

o extend as required
25mm Steel pipe, threaded both

ends, extend as required
100ram PVC pipe

not connected Thread 25mm Steel pipe to
pipe flange or weld to base
i r*-~
1.3m x 1.3m x 19ram
Elevation Pressure Treated Base
FIG. 3-Settlement Plate Schematic
The vibrating wire piezometers and liquid settlement gauge at Marsh A were
connected to a central control box located outside the preload area. The control box
allowed up to 10 separate instruments to be read from one location. At Marsh B, the
instrument cables were connected to wood posts outside the preload limits to obtain
readings.
Preload Construction and Monitoring
Marsh A preload construction began in November 1995 and Marsh B

construction began in April 1996. For Marsh A and B, a total of 4 and 5 preload stages
were constructed, respectively. The first stages were about 0.6 meters thick and
subsequent stages were about 0.9 meters thick. The final stages for Marsh A and B
were constructed in April 1996 and September 1996, respectively.
The preload embankments were constructed using granular fill from a nearby
cut section of the highway alignment and the fill was hauled to the preload areas in 8 m 3
dump trucks. The trucks were able to drive directly on the preload areas and the fill was
placed in 0.3 meter lifts and compacted using tracked equipment.
RUSSELLETAL.ONPRELOADPERFORMANCE 243
Settlement and pore pressure readings were taken prior to and after placing each
preload stage. After each stage was placed, readings were taken at 1 week intervals for
the first month and at 2 week intervals for the following months. In general, subsequent
preload stages were not placed until about 65 to 90 percent of the excess pore pressures
had dissipated. A typical settlement and pore pressure dissipation curve from an
adjacent settlement plate and vibrating wire piezometer is presented in Figure 4.
2.5- O-
2.0 - 0.2 -
i l.s- E~ Settlement
1.0- ~ 0.6-
0.5 - 0.8 -
O- 1.0 I I I I I I
20 40 60 80 100 120
gays
FIG.4-Typical Settlement and Pore Pressure Dissipation Curves
Rheological Modeling
The initial preload design parameters presented in Table 3 were estimated using
conventional one-dimensional settlement analyses and laboratory consolidation data.
The conventional settlement analyses performed, developed for inorganic silts and
clays, was considered approximate. To determine when preloading was completed, the
Gibson and Lo Rheological model, as further developed by Edil, was used to generate
field settlement curves and predict settlements under the expected STH 29 service loads
(Gibson and Lo 1961; Edil 1982). The predicted settlements were compared to actual
field settlements to determine when preloading was completed (i.e., when the
anticipated immediate, consolidation, and long-term secondary compression was
induced). Input parameters for the model were developed using the field settlement
data at the Marsh A and B preloads.
The following steps were followed to estimate the rheological parameters for
each preload:
(1) Settlement plate data was used to develop vertical strain versus time plots and
determine when primary consolidation appears to end and secondary compression
begins (Figure 5).
(2) The logarithm of strain rate versus time was plotted for data obtained after primary
consolidation had occurred A best fit straight line was obtained through the data
points (Figure 6) and the rheological parameters "b" and "X" were calculated using
the slope and intercept of the best fit straight line (Figure 7).
(3) The rheological parameter "a" was determined from "b" and "~." per the procedure
outlined by Edil (Figure 7).
Using the above parameters in Equation 1, the predicted and actual strains were
compared. The parameters that produced the predicted strains closest to the actual
strains were used to predict the service load (i.e., embankment and pavement section)
settlements.
6(t) = A6[a+b(1-e" 2/b(t))] (1)
Where
8 (t) = Strain at time, t

Ac = Applied Load, F/L2,
a = Rheological Parameter,
b = Rheological Parameter,
2 = Rheological Parameter,
t = Time, t
Using Equation 1 and the marsh deposit thicknesses, settlement prediction curves
were developed along different sections of each preload (Figure 8). The preload was
intended to eliminate the settlements under the service load over a 30-year design life.
The predicted settlement at 30 years was compared to the actual field settlements. If the
actual settlement was less than the predicted settlement, the preload was left in place or
additional fill added. If the actual settlement was greater than the predicted settlement,
the preload was stripped down to the pavement subgrate elevations for construction of
the pavement sections.
STRAIN vs. TIME
0"000T
0"010l
0,020-~
0,030-
0,040-
0.050
0.060 I I
0.00 50.00 100.00 150.00
Time In Days
FIG. 5-Typical Vertical Strain Versus Time Plot
LOGARITHIM OF STRAIN RATE vs.TIME
0.000
-1.000 -
=.
rr
e-
-2.000-
u}
E
-3,000-
r
/o
-4.000 -
-5.000
0.00 20.00 40,00 60.00 80.00 100.00

Time in Days
FIG. 6-Typical Logarithm of Strain Versus Time Plot
LOGARITHIM OF STRAIN RATE vs.TIME

0.000--
m = "9~ = -0.009 -0.434 ( ~ ) = - 0 . 0 0 9
-0.500-- Intercept = ~ ~ = 0.021
-1.000-
+3.2 - Ioglo Ao'~,
n- -1.500- ~.: 0.00063/345 PSF
_= ~,= 1.83 x 104 PSF "1
r162
-2.000- b = 1.83 x 10 "s PSF "1 = 0.00009

0.021
6
-2.500- a = E (tk). b + be " ( ~
r- &o-
= 0.058,. 0.00009 + O.O0009e (.o.ezl)(12s)
-3.000- 345
= 0.00009
-3,500-
-4.000-
-4.500 I 1 I I
0.00 20.00 40.00 60.00 80.00 100.00
Days
FIG.7 - E s t i m a t e d Rheological P a r a m e t e r "'b . . . . ,~ " a n d "'a '"
PREDICTED SETTLEMENTS AT SPRANGER CREEK

BETWEEN ABOUT 1036+00 AND 1041+00 AT VARYING SERVICE LOADS
0.0
0.3 t
~ o.6
~ 1.0-
.; 1.3-
i 1.6--i
~ 2.0-
1.100 PSF
2.3- [] 1.1soI ti
1,250 /
2.6-
.r
x
1#oo
3.0- I I I I I I 1
50 100 150 200 250 300 350 400
Days
FIG.8 - T y p i c a l Settlement Prediction Curve
Results
Typical site-specific rheological parameters for Mash A and B are presented in

Table 5. Using site-specific parameters and Equation 1, Settlement Prediction Curves
were developed and it was determined when preloading was completed.
TABLE 5-Typical Site-Specific Rheological Parameters
Marsh 2 b a
A 1.8x10 ~ 0.00009 0.00009
B 9x10 "7 0.00006 0.00006
When the preloading was determined to be completed, the preloads were

stripped down to the pavement subgrade elevation and the pavement sections installed.
The paving work was completed in the Spring of 1997, and to date there has been no
measurable settlement of the pavement surface in the preloaded areas.
During pore pressure monitoring, the coefficient of consolidation, Cv, was back
calculated, based on the amount of pore pressure dissipation, in accordance with
Equation 2 and 3. The estimated apparent field coefficients of consolidation are
presented in Table 6.
U
U~ = 1- -- (2)
Ui
Where
UZ Percent Consolidation and Mid-Point of Consolidating Layers

U = Pore Pressure Reading at time t,
U/ = Initial Pore Pressure Reading at t = o.
CV ~-- THat2 (3)

t
Where
CV Coefficient of Consolidation, m2/s,
T = Time Factor,
Hdr = Length of the Drainage Path, m,
t = Time, see.
TABLE 6-Estimated Apparent FieM Coefficients of Consolidation
Apparent Field Apparent Field

Coefficient of Coefficient of
Stratum Type of Material Consolidation, Cv Consolidation, Cv
(mZ/s)(1) (m2/s)(2)
1 Peat 1.4 x 10.4 4.5x10 "~
2 Organic Silt ...(3) ...(3)
Footnotes:
(1) At Marsh A without vertical drains.
(2) At Marsh B with vertical drains at 1.8 meters on center.
(3) Vibrating wire piezometers were not installed in the organic silt layer.
Based on the results presented in Table 6, the estimated apparent field

coefficient of consolidation for the peat is approximately the same as the laboratory
coefficient of consolidation (1.4 x 10.6 m2/s versus 1.9 x 10"6m 2/s). It was not possible
to compare the coefficient of consolidation for the organic silt because vibrating
piezometers installed in the organic silt deposit malfunctioned. The apparent field
coefficient of consolidation for the peat at Marsh B was about 3 times greater due to the
installation of vertical drains.
Conclusions
Preloading was a cost-effective solution for constructing about 900 linear meters
of new 2-lane highway for the reconstruction of STH 29 between the cities of Hatley
and Wittenberg, Wisconsin. To date, there has been no measurable settlement of the
pavement surface along the areas preloaded. Therefore, settlement prediction curves,
developed using the Gibson and Lo rheological model, where reliable for predicting
when preloading was complete and when the preload surcharge could be stripped.
Settlement plates spaced about ever 30 meters along the preloads were adequate
for obtaining the settlement data necessary for generating the rheological Settlement
Prediction Curves. Vibrating wire piezometers spaced about every 100 meters along
the preload were useful for estimating the time rate of settlement, and for determining
when the next preload could be placed. However, this information could also be
determined by interpreting the settlement data strain versus time plots. Based on this, a
wider vibrating wire piezometer spacing, say every 130 to 200 meters, is considered
reasonable for providing back-up information for projects of similar size and type.
References
Canadian Foundation Engineering Manual, 3 rd Edition, Canadian Geotechnical

Society, pp. 261-263, 1992.
Gibson, R. E. and Lo, K.Y., "A Theory of Consolidation of Soils Exhibiting

Secondary Compression", Acta Polytechnica Scandinavica, Ci10296, 1961, pp.
1-15.
Edil, T. B., 'q_lse of Preloading For Construction on Peat", 13th Minnesota

Conference on Soil Mechanics and Foundation Engineering, St. Paul,
Minnesota, 1982.
Phil Flentje' and Robin Chowdhury 2
Instrumentation for Slope Stability - Experience from an Urban Area
REFERENCE: Flentje, P. N., and Chowdhury, R. N., "Instrumentation for Slope

Stability--Experience from an Urban Area," Field Instrumentationfor Soil and Rock,
ASTMSTP 1358, G. N. Durham and W. A. Man-, Eds., American Society for Testing and
ABSTRACT: This paper describes the monitoring of several existing landslides in an

urban area near Wollongong in the state of New South Wales, Australia. A brief
overview of topography and geology is given and reference is made to the types of slope
movement, processes and causal factors. Often the slope movements are extremely slow
and imperceptible to the eye, and catastrophic failures are quite infrequent. However,
cumulative movements at these slower rates do, over time, cause considerable distress to
structures and disrupt residential areas and transport routes. Inclinometers and
piezometers have been installed at a number of locations and monitoring of these has
been very useful. The performance of instrumentation at different sites is discussed in
relation to the monitoring of slope movements and pore pressures. Interval rates of
inclinometer shear displacement have been compared with various periods of cumulative
rainfall to assess the relationships.
KEYWORDS: landslides, subsurface shear movement, monitoring, inclinometer,

piezometer, velocity of movement, pore water pressure
Introduction
Monitoring and observation of slopes is an essential part of the assessment of

slope stability. Observational approaches facilitate the assessment of hazard and risk
associated with slope movement and landsliding processes. The most important
measurements associated with foundation investigations, natural slopes and earth dams
aReseareh Fellow, Department of Civil, Mining and Environmental Engineering, University of

Wollongong, Wollongong, NSW, 2522 Phone +61 2 42213056 Fax +61 2 42213238 email:
pflentje@uow.edu.au.
2Head of Department, Department of Civil, Mining and Environmental Engineering, University of

Wollongong, Wollongong, NSW, 2522 Phone +61 2 42213037 Fax +61 2 42213238 email:
robin_chowdhury@uow.edu.au.
FLENTJE AND CHOWDHURY ON SLOPE STABILITY 251
concern pore water pressures, settlements, lateral movements and shear movements.
Amongst these, the two most important for slope stability and landslide monitoring are
subsurface shear movement and pore water pressures.
In geotechnical engineering applications, instrumentation, observation and

monitoring may serve several purposes. For some projects it is necessary to check
important assumptions concerning the basic geotechnical/geological models used for
analysis and interpretation. In slope stability problems it may be necessary to verify the
shape and location of the slip surfaces and the character of the movement, e.g., rotational
or translational. Furthermore, it may be necessary to check the assumptions made in
analysis and design about the magnitude of subsurface pore water pressures and their
spatial and temporal variability.
Apart from checking design assumptions, continued monitoring can be a useful

tool for updating of assessments concerning stability, hazard and risk. In addition,
continued monitoring is essential for assessing the effectiveness of remedial measures
after their installation and/or construction.
Figure 1 - Study Area Location Plan showing the extent of Triassic and Permian
sediments in the Southern half of the Sydney Basin
This paper is concerned with monitoring of geotechnical instrumentation in

sloping urban terrains. Subsurface shear movements were monitored at a number of
existing inclinometer installations near the city of Wollongong, on the south coast of
New South Wales, Australia. Subsurface shear movement was compared to antecedent
rainfall patterns in a number of instances. Water levels in open standpipes were measured
at several sites. In some instances subsurface pore pressures were also measured more
accurately using vibrating wire piezometer installations. However, the emphasis during
the research was on subsurface shear movements and their relationship to rainfall.
The Wollongong Area
The City of Wollongong, including the northem suburbs, is a major urban centre
on the south coast of NSW with a population of approximately 200,000 people. A total
of 328 past and present sites of land instability are known in an 87 km 2 area within the
Wollongong City Council local government area. At least 60 houses have been damaged
and a further 29 houses have been destroyed since the turn of the century by land
instability. As well as affecting suburban areas, landslides affect the major road and rail
transport routes which link Wollongong to Sydney and other regional centres (Shellshear
1890, Hanlon 1958, Bowman 1972 and Pitsis 1992).
Wollongong is situated on the coast approximately 80 km south of Sydney (Fig.

1). The local government area includes a highland plateau area to the west of the
escarpment topped by a near vertical sandstone cliff line which grades down to the east
into terraced lower slopes. The eastern and southern margins of the district comprise a
narrow coastal plain which is flanked on the east by the Pacific Ocean.
Geological Setting
The Sydney Basin comprises a basin sequence of dominantly sedimentary rocks

extending in age from approximately Middle Permian to at least Middle Triassic. The
study area lies on the south eastern margin of the Sydney Basin (Fig. 1).
The geological bedrock sequence of the Illawarra district is essentially fiat lying
with a low angle dip, generally less than five degrees, towards the northwest. This gentle
northwesterly dip is a result of the relative position of the district on the southeastern
flanks of the Sydney Basin. The northwesterly dip is superimposed with relatively minor
syn-depositional and post-depositional structuring (folding and faulting). Normal
faulting within the Illawarra area is common, although the fault throws infrequently
exceed 5 metres. The structural geology of Wollongong, Kiama and Robertson, mapped
on 1:50000 sheets, has been discussed in detail in Bowman (1974).
The geological units encountered within the district, in ascending order, include
the Illawarra Coal Measures (locally including intrusive/extrusive bodies collectively
known as the Gerringong Volcanic facies), the Narrabeen Group and the Hawkesbury
Sandstone. The geology of these units has been discussed at length in several
publications such as Bowman (1974) and Herbert and Helby (1980).
Extending down from the base of the upper cliff line, the ground surface is often
covered by debris of colluvial origin. This material comprises variably weathered
bedrock fragments supported in a matrix of finer material dominantly weathered to sand,
sandy clay and clay and brought downslope under the influence of gravity.
Instrumented Sites
Inclinometer monitoring records of forty two boreholes from 19 landslide sites

have been examined and analysed. Records of movements prior to 1993 have been
acquired from government authorities and from consulting engineers who were
responsible for the initial drilling of boreholes and installation of the inclinometers.
Monitoring at several sites was continued during the period 1993-1997 as part of a
Landslide Hazard Assessment project. The inclinometer monitoring records comprise the
recorded profiles of the borehole inclinometers and the date of each monitoring visit.
A sample of 10 boreholes each with a brief summary of monitoring coverage and

movement data are listed in Table 1. This table also shows the borehole identification
number and the site reference code of each landslide.
Table 1 - A Sample of lO Boreholes out of a total of 42 in the borehole inclinometer

dataset with site reference code (SRC), monitoring coverage and rates of movement.
Borehole 1st last numher I depth to peak rate average rate

SRC
Name readin[~ readin~ I readin[~sl shear (m) mm/da~, mm/da~,
2 026 28/03/93 9/12/96 12 15-16 0.03 0.02
4 026 28/03/93 9/12/96 12 8.5-9.5 0.03 0.02
6 026 28/03/93 9/12/96 12 11.5-12.5 0.04 0.02
3 064 9103/89 8/08/96 30* 7.0-8.0 1.93 0.07
7 064 19110189 8/08/96 27* 1.5-2.5 2.09 0.06
WH2 134 14/12/90! 16/05/96 16" 11.0-12.0 1.1 0.01
WH10 134 12/11/91 15/05/96' 1 I* 13-13.5 0.03 0.004
WH13 134 12/11/91 16/05/96 11" 16.5-17.5 0.11 0.01
WH14 134 12/11/91 16/05/96 10* 3.4-3.9 0.08 0.01
WH15 134 18/02/92 16/05/96 9* creep 0.03 0.002
* denotes that some readings were collected as part of this research ~roject
It was considered necessary to adopt a standard numbering system for the 328
sites of instability incorporated into a landslide inventory of this area (Chowdhury and
Flentje, 1997, 1998a). This landslide inventory was compiled during the research project
mentioned above. Each site is identified with a 3 character Site Reference Code (SRC).
This site reference code is unique to each location, is plotted for each site on a map
record and constitutes the "primary" or "key" field in a land instability database.
In general, landslides in the Wollongong area can be summarised as follows:

9 predominantly rainfall triggered but often a consequence of urbanisation
9 80 % o f the total number of landslides have volumes in the range 500 m 3 to
50,000 m 3, but the maximum landslide volume is 600,000 m
9 velocities are typically very slow to extremely slow according to the
international scale (WP/WLI, 1993)
9 depths to the slip surfaces range from 0.5 m to 20 m
9 movements are rotational and/or translational
9 landslide material generally comprises of clay and sandy-clay colluvium

9 more than 50 % of the landslides can be classified as complex debris slump -
debris flows (Varnes 1978)
In addition to inclinometer monitoring, open standpipe piezometers have been

used to record standing water levels at some sites. Two vibrating wire piezometers
(VWP) are installed in the head area of one landslide. These VWP have been monitored
as part of this research project.
Inclinometer Profiles
Before giving details of three significant landslide sites in subsequent sections,

some general comments on the installation of inclinometers and generated inclinometer
profiles are provided here.
The inclinometers have been installed by various government organisations and

engineering consulting firms as part of geotechnical investigations of landslide sites.
These installations generally comprise 70 mm outside diameter plastic probe
inclinometer casing installed within the full length of the borehole. The space between
the casing and the wall of the borehole is backfilled with sand.
Inclinometer records used during this research project include the displacement at
0.5 m intervals along the length of the borehole, relative to the position of the bottom of
the borehole. Cumulative displacements are summed from the base of the borehole to the
top. Recording displacement at 0.5 m intervals provides optimum accuracy, such that
probe inclinometers can measure changes in inclination of the order of 1.3 mm to 2.5
mm over 33 m lengths of inclinometer casing (Mikkelsen 1996), although one
manufacturer claims a system accuracy of_+ 6 mm per 25 m.
Analysis of each record enables the identification of time intervals between each
subsequent monitoring visit and the magnitude of the movement during each such time
interval. This enables the cumulative displacement in subsurface shear movement to be
compared over time and, therefore, allows the interval rates of shear to be determined.
There are numerous different methods of reporting and discussing inclinometer

profiles. For example, the New South Wales Government Railway Services Authority
Geotechnical Services (RSA) employs an ATA- and B axis nomenclature system
with the A axis orientated perpendicular to the railway line and pointing downslope,
with the B* axis clockwise 90 ~ from the A axis. The RSA typically reports the Aaxis
data only. The New South Wales Government Roads and Traffic Authority (RTA)
employs an A/B and C/D axis nomenclature system with the A axis orientated along the
direction of maximum slope and the C axis orientated 90 ~ in a clockwise direction from
the A axis. The RTA typically reports the profiles of both axis. Coffey Partners
International (CPI) uses the same axis labelling system as that employed by the RSA,
except the A axis is usually orientated along the steepest inclination downslope. CPI
reports both axes profiles, and then calculates and reports the maximum resultant
displacement vector (magnitude and azimuth).
Composite 20 Year Rainfall Histogram and Antecedent Rainfall Curves
Rainfall is recognised as an important causal and triggering factor for slope instability.
To assess the relationship between slope instability and rainfall within the study area
daily rainfall totals have been used to determine daily rolling antecedent rainfall curves,
for which Antecedent Rainfall Percentage Exceedance Time (ARPET) values have been
determined (Chowdhury and Flentje 1998b). These antecedent rainfall curves and
percentage exceedance values have been compared with monitored landslide shear
displacement curves to examine upper and lower bound antecedent rainfall thresholds
associated with landslide movement.
Daily rainfall totals have been collected from several rainfall stations to compile
an unbroken 20 year composite record, which is currently being extended to cover a 100
year period. The existing 20 year record extends from 1 January 1977 to 31 December
1996. This period is significant not only for its duration, but also for some exceptional
rains that have fallen during this period and several periods of significant land instability
which it encompasses. The data were entered into a computer spreadsheet whereby
antecedent rainfall for daily rolling periods of 7 days, 30 days, 60 days, 90 days and 120
days have been computed.
In the following sections three significant landslide sites are discussed in some
detail.
The Coalcliff Terrace Landslide, Site 026
This landslide site covers an area of approximately 66,000 m e, and has a volume
of approximately 600,000 m 3 making it the largest volume landslide in the Wollongong
study area. According to Cruden and Varnes (1996) this landslide can be described as an
active, advancing, composite, extremely slow, moist to wet debris slump in the head area
and a debris slide-debris flow lower down. The landslide is rotational in the head area of
the site, but predominantly translational for most of the site below the head area. The
depth of the colluvium material at this site is up to 16 m. Shear displacement is typically
occurring along the bedrock/colluvium interface.
This site is traversed by the South Coast Railway Line (a dual electric freight and
passenger railway line) and a two lane main road. Ground movements have been
reported at least every decade at this site since 1942. The site has been the subject of
several detailed geotechnical investigations. During the most recent of these, 13
boreholes were drilled, 6 of which had inclinometers installed, and another 6 had a total
of eleven pneumatic piezometers installed.
In late 1997 and early 1998 this site has had an elaborate series of remedial
measures installed, and as a result may have been stabilised. These remedial engineering
construction works include a 200 m long row of vertical drainage wells with an
interconnecting series of subhorizontal, gravity-fed, drainage relief drives.
Inclinometer profiles (A axis only) recorded by the Railway Services Authority,
from three borehole inclinometers at this site have been examined for rates of movement.
One of these inclinometer profiles is shown in Figure 2. This profile displays a block
style of movement with a depth to the slip surface of 9.25 m. Progressive movement at a
depth of 4.5 m in borehole 4 is shown in Table 2.
Borehole Log
~ Legend
9/12/99
~, 4 19/5/99
7/2/96
: ,, ,/11/.
IO/2RS
30/11/94
3t/3/94
i 13/5/94
911 - . ___ <, 19/2/94
.^i, I tede,,h 1.3t/9/93

iu ~I I ~-'~ I ~ of shear 9.25m l I 2013/93
1't I L_
I o:"-~
m I r A Axis
~ cumulative
,5 =; displacement
r ~ (ram)
r ,.,.;
End of Borehole
15,58m
Figure 2 - Summary Borehole Log and lnclinometer Profiles for Borehole 4, Site 26 at
Coalcliff, near Wollongong in New South Wales, Australia
Table 2 - Borehole 4, Site 026, monitoring period and displacement (A axis data only)
recorded at 4.5m depth
Date of Days Interval Displacement (mm) Rate of shear {ram/day)

reading (since start) (days) Resultant Interval Interval mean
28/03/93 0 0 0 0.00 0.000 0.000
31/08/93 156 156 2.81 2.81 0.018 0.018
15/02/94 324 168 7.61 4.80 0.029 0.023
13/05/94 411 87 9.67 2.06 0.024 0.024
30/08194 520 109 11.86 2.19 0.020 0.023
10/11/94 592 72 12.04 0.18 0.003 0.020
10/02/95 684 92 13.8 1.76 0.019 0.020
2/05/95 765 81 15.49 1.69 0.021 0.020
7/11/95 954 189 t9.61 4.12 0.022 0.021
7/02/96 1046 92 22.37 2.76 0.030 0.021
19/09/96 1271 225 27.74 5.37 0.024 0.022
9/12/96 1352 81 29.42 1.68 0.021 0.022
Total displacements indicated for boreholes 2, 4 and 6 over the period 28th March
1993 to 9th December 1996, are 27.6 mm, 29.4 mm and 29.0 mm respectively, all at
average rates of approximately 0.02 mm per day, or, if extrapolated, 8 mm per year. On
the WP/WLI (1993) velocity scale, such a rate of displacement is classified as extremely
slow. Yet, RSA engineers maintaining the dual electric railway line that traverses this
site have reported an annual maintenance cost exceeding tens of thousands of dollars.
Cumulative displacement and rate of shear for three boreholes at this site,
including borehole 4, are shown in Figure 3 together with the 90 day antecedent rainfall
curve. The cumulative displacement curves for the three boreholes show the continual
movement at the site over the period of monitoring, 28th March 1993 to 9th December
1996. Periods where movement is lower than 0.015 mm per day (the average rate of
movement is approximately 0.02 mm per day) correspond to periods of low antecedent
rainfall, except Borehole 2 for the period November 95 to February 1996. Movement has
continued at this site for the duration of the whole monitoring period.
~" 0.05
0.04
0.03
~,
Q.
o,o2
Q 0,01
~:o.oo
30
iW
,0
10
0
~800
E
...4oo
g o
200
E
E
"~ lOO
0
1993 1994 1995 1996 I
Figure 3 - Site 026, daily and antecedent rainfall and shear displacement for 3
boreholes.
Periods of accelerated movement, leading either to peaks or steep positive slopes

on the rate per day curves, generally correspond to high values in the antecedent rainfall
curves over the monitoring period.
The effects of the recently completed remedial stabilisation works on shear

displacement at depth and ground movement at this site is yet to be assessed. This
assessment will be largely based on data obtained from the instrumentation installed at
this site, and on other ground surface survey data. The results of this assessment will be
reported elsewhere.
A Landslide in Goodrich Street, Scarborough, Site 064
Site 064 covers an area of approximately 5000 m 2 and a volume of almost 18,000
m 3, making it the 77th largest landslide in the study area. The head area of this landslide
is also traversed by the South Coast Railway Line. The area below and downslope of the
railway side-fill embankment is residential land.
The most recent geotechnical investigation of this site was carried out by the then
State Rail Authority of New South Wales, now a subsidiary of that organisation known
as the Railway Services Authority Geotechnical Services, and by their consultants in the
period from 1989 to 1991. These investigations followed various land instability
"events" during the preceding 40 years. These events include scour wash outs, general
track "subsidence", development of tension cracks between the two tracks, a train
derailment and inundation of the site with debris resulting from ground movements
upslope.
These investigations included the excavation of test pits, the drilling of numerous
boreholes with installation of two inclinometers and two vibrating wire piezometers.
Three additional open standpipes were also installed.
The monitoring results from all the instrumentation installed at this site, combined
with daily rainfall and a daily rolling 90 day antecedent rainfall curve (rainfall station
7km away from the site) are shown in Figure 4. The inclinometer profiles (not included
here) clearly define the depth to and style of movement at this site, which, combined
with surface surveys of ground movement, allowed the approximate plan area and
volume of this landslide to be calculated. The cumulative shear displacement and rates of
shear displacement curves in Figure 4 clearly show the response to rainfall, and, in
particular, the 90 day antecedent rainfall in the periods March - June 1989 and February
- April 1990. A series of longitudinal subsurface drainage trenches were designed and
installed at this site, at a cost of almost $400,000.
Monitoring of the site following the installation of these remedial works has been
continued. This "post construction" monitoring has revealed a significant decrease in
slope movement but the site has not been stabilised fully.
Interval rates of displacement following the first monitoring visit since

completion of the remedial works have ranged from a maximum of 0.03 mrn/day down
to less than 0.01 mm/day. However, in the head area of the landslide, as indicated by
inclinometer 3, 21 mm of displacement has occurred since the remedial works, whereas
only 2.5 mm of displacement has been indicated in the toe area of the landslide, as
indicated by the monitoring data from inclinometer 7. Monitoring data from the shallow
vibrating wire piezometer, installed at a depth of 4.2 m in the head area of the landslide,
in the vicinity of borehole inclinometer 3, has shown a significant increase in pore water
pressure during 1994 and 1995. This build up of pore pressure may be part of the reason
for continued movement in this area of the landslide.
Figure 4 - Site 064, MonitoringData for all of the Instrumentation Installed at this site
plotted with Daily Rainfall Totals and 90 day Antecedent Rainfall Curve.
It must also be noted that in other research work carried out by the writers and
reported elsewhere (Flentje 1998, and Chowdhury and Flentje 1998b), the 90 day
antecedent rainfall curve has been considered to show a close relationship to landslide
movement at this site. Since the remedial works were completed in June 1992 the
landslide site has not experienced 90 day antecedent rainfall magnitudes of similar scale
to those which have been correlated with "failures" in 1989 and 1990. Hence, there is
concern about the effectiveness of the remedial measures and monitoring of this site is
continuing.
The Woonona Heights Landslide, Site 134
The Wonoona Heights Landslide has a plan area of approximately 20,500 m 2, a

volume of approximately 225,000 m 3 and is ranked in the land instability database as the
9th largest landslide within the subject area. The landslide site is located within a densely
developed urban area, which contains up to 100 residential houses, 29 of which are
situated within the landslide. Another 10 houses are located immediately adjacent to the
margins of the landslide. One house has been destroyed by ground movement, whilst at
least 19 have required major repairs.
Borehole Log
0 -1
1"5-i
2-1"4
M o n i t o r i n g Date
3-1
[ 12/11/91
1 =S
4-t _= (~ 20/12/91
"4
>, [] 23/I 2/I)1
5-1
..J (~ 10/3/OS
.J
O
u ,A, 10/S/OS
[] 15/5/08
[~ 1710/95
20/10/U
==8 1 w
1ol r
w
[]
20/1R$
23/4R6
;111 as
m
|
:13-1
"15-4
"o
lS-j e
ol
=
17-1 <
18-t ate depth

r 17.2m
19 -t ~,,,z
~go
20 -1
End of Borehole o in o ~ o
21.0 m A Axis cumulative displacement (turn)
Figure 5 - Summary borehole log and inclinometer profiles for borehole 13, Site 134 at
Woonona Heights, near Wollongong in New South Wales, Australia.
This landslide has a long history but is a subtle feature, very little damage being
visible from the streets. According to a previous geotechnical report, disturbed ground
can be observed in a 1948 black and white aerial photograph. Accelerated movement
phases have been documented at this site in 1977, April 1990, March - June 1991 and
November 1991 - March 1992, all of which have resulted in road and residential damage.
The site has recently been investigated on behalf of the local council by an
international geotechnical consulting firm. The comprehensive investigation included the
drilling of 22 boreholes, and included the installation and periodic monitoring of 7
inclinometers and numerous open standpipe piezometers. Following this investigation
remedial works have been proposed. These works have not yet been implemented.
Consequently, the site remains episodically active as an extremely to very slow moving
landslide (WP/WLI, 1993).
This landslide is located within the area of subcrop of the Illawarra Coal
Measures. The slide material comprises of a sequence of some fill and a gravelly and
sandy-clay colluvium sliding over residual bedrock, which includes several coal seams.
Monitoring of three inclinometers and piezometers has been carried out at this site. All
the inclinometer profiles within the landslide show similar block styles of movement,
with an approximate average depth of 11 m to the slip surface, the maximum depth being
almost 18 m, as shown in Figure 5.
This monitoring has confirmed that movement at this site, under residual strength
conditions, continues to occur when the piezometric levels associated with seepage reach
about l m below the ground surface (Fig. 6). The results of laboratory shear strength
testing and stability analyses are outside the scope of this paper.
Further instrumentation and monitoring is required to investigate the lower half of

the area of this landslide. In this area, the basal slip surface of this landslide is essentially
horizontal and immediately overlies a highly fractured coal seam with numerous low
strength claystone bands of tuffaceous origin. Additional inclinometers and piezometers
would provide extremely useful information for the geotechnical model of this landslide
and may facilitate the design of more economical remedial measures than those which
have already been proposed.
Discussion
With the types of landslides common in this study area, an acceptable minimum
inclinometer monitoring interval is 30 days. Additional monitoring visits are required in
response to rainfall "events", with additional follow up visits one day to several weeks
later, dependent upon the rainfall and the associated movement response. During periods
of dry weather, the period between monitoring visits can be extended. The preferred 30
day period allows accurate interval rates of displacement to be calculated. More frequent
monitoring does incur significant costs, and it is accepted that monitoring schedules will
often be controlled by available financial resources,
Use of instrumentation for subsurface monitoring of slopes along the Illawarra
escarpment has been very useful for establishing the character, range and velocity of
slope movements. The episodic nature of landslide movements has been established and
more research is required to determine the conditions under which catastrophic failures
may take place. This is likely to be a very difficult area of research because catastrophic
failures are rare and because instrumentation may not be installed at the particular site
which undergoes catastrophic failure. Furthermore, if instrumentation existed at the site,
it may not survive the large movements which occur during a catastrophic landslide.
Figure 6 - Site 134, daily and 90 day Antecedent Rainfall, Cumulative and Rate of Shear
and Standpipe Water Levels.
Some success has been achieved in correlating landslide movements to

cumulative rainfall magnitudes and different levels of antecedent rainfall have been
considered in this research. These relationships have been quantified on a site by site
basis, as well as for the landslide "population" as a whole. It appears from this research
that 90 day and 120 day antecedent periods correlate fairly well to peak or accelerating
movement phases at many of the sites.
Open standpipes (slotted pipes, with the annulus back filled with sand over the
entire borehole depth) have been widely used in the Illawarra region because of their low
cost. Yet, it is recognised that the associated pore pressure data may not be reliable or
accurate. Vibrating wire or pneumatic piezometers, whilst more expensive on a per unit
basis, provide more reliable and accurate data.
Concluding Remarks
The use of inclinometers has proved to be immensely valuable for monitoring the
performance of slow moving landslides in the Northern Illawarra Region of New South
Wales, Australia. In fact this monitoring is considered to be the single most important
factor in detailed site assessments of all such landslides regardless of their size and
previous history. Experience has shown that simple and robust instrumentation is the best
in these situations.
The inclinometers used at all sites have performed well and have proved to be
reliable. Based on the experience gained during this research, movement rates of up to
17.6 mm per day and cumulative movements of up to 200 mm at slower rates of
displacement can be measured with confidence. Rates of displacement as slow as 0.002
mm per day have been recorded over 90 day monitoring intervals. This level of accuracy
is only relevant when considering the longer monitoring periods, e.g, more than a month.
Measurements concerning extremely slow movements are not expected to be accurate or
reproducible over very short monitoring periods, e.g, a few days. Of course, the use of
inclinometers in faster moving landslides will require shorter intervals bewtween
monitoring visits. In such cases, weekly or daily monitoring may be necessary.
References
Bowman, H. N., 1972. "Natural Slope Stability in the City of Greater Wollongong,"
Records of the Geological Survey of New South Wales, Vol. 14, Part 2, pp. 159 -
242.
Bowman, H. N., 1974. Geology of the Wollongong, Kiama and Robertson 1:50,000
Sheets 9029-H 9028-1&IV Sheets, Department of Mines, Geological Survey of
New South Wales. Government Printer, New South Wales.
Chowdhury, R. N. and Flentje, P. N., 1997. "Relevance of Mapping for Slope Stability
in the Greater Wollongong area, New South Wales, Australia," Proceedings of
the International Symposium on Engineering Geology of the Environment,
Marinos, P. G., Koukis, G. C., Tsiambaos, G. C. and Stournaras, G. C., (editors),
Athens, Greece. June 23 - 27, pp. 569 - 574, Balkema, Rotterdam.
Chowdhury, R. N. & Flentje, P.N. 1998a. "A Landslide Database for Landslide Hazard
Assessment." Proceedings of the Second International Conference on
Environmental Management. Wollongong Australia. Sivakumar, M. &
Chowdhury, R. N., (editors), February 10 -13, Vol. 2, pp 1229 - 1237. Elsevier:

Amsterdam.
Chowdhury, R. N. & Flentje, P. N. 1998b. "Effective Urban Landslide Hazard

Assessment," Proceedings of the 8th Congress of the International Association
of Engineering Geology and the Environment (in press), 7p.
Cruden, D. M, and Varnes, D. J., 1996. "Landslide Types and Processes," Chapter 3,
Turner and Schuster, 1996, pp 36 - 75.
Flentje, P. N., 1998. Computer Based Landslide Hazard and Risk Assessment (Northern
lllawarra Region of New South Wales, Australia), unpublished, University of
Wollongong, Department of Civil, Mining and Environmental Engineering
Doctor of Philosophy thesis.
Hanlon, F. N., 1958. "Geology and Transport, with Special Reference to Landslides on
the near South Coast of NSW," Presidential Address, Part II, Journal of the
Proceedings Royal Society of NSW, Vol. 92, pp. 2-15.
Herbert, C., and Helby, R., (editors) 1980. A Guide to the Sydney Basin, Bulletin No. 26.
Department of Mineral Resources, Geological Survey of New South Wales.
Mikkelsen, P., E., 1996. "Field Instrumentation," Chapter 2, in Turner and Schuster.
Pitsis, S. E., 1992. Slope Instability along the lllawarra Escarpment, University of New
South Wales, Master of Engineering Science thesis, unpublished.
Shellshear, W., 1890. "On the Treatment of Slips on the Illawarra Railway at Stanwell
Park," Journal of the Proceedings of the Royal Society NSW, Vol. 24 (1), pp. 58 -
62.
Turner, A. K. and Schuster, R. L., (editors), 1996. Landslides, Investigation and

Mitigation, Special Report 247, Transportation Research Board, National
Research Council. National Academy Press, Washington DC.
Varnes, D. J., 1978. "Slope Movement Types and Processes", Chapter 2, Landslides,
analysis and control. Schuster, R. L. and Krizek, R. J., (editors).
WP/WLI (Working Party on World Landslide Inventory), 1993. "A Suggested Method
for Describing the activity of a Landslide," Bulletin of the International
Association of Engineering Geology, No. 47, pp 53 - 57.
Data Acquisition and Data Management
Craig H. Benson 1 and Peter J. Bosscherz
Remote Field Methods to Measure Frost Depth
REFERENCE: Benson, C. H. and Bosscher, P. J., "Remote Field Methods to Measure

Frost Depth," Field Instrumentation for Soil and Rock, ASTM STP 1358, G. N. Durham
and W. A. Marl Eds., American Society for Testing and Materials, West Conshohocken,
PA, 1999.
ABSTRACT: This paper reviews the three common methods used to remotely measure
frost depth: (1) soil temperature measurements, (2) soil electrical resistivity
measurements, and (3) soil dielectric constant measurements. Soil temperature
measurements are often made using thermocouples or thermistors. Specially designed
probes are used to measure frost depth based on electrical resistivity. Inferences based on
dielectric constant are usually made using time domain reflectometry. Frost depths based
on soil temperatures are made equally well using thermistors or thermocouples.
However, frost depths inferred from soil temperature measurements can be erroneous if
the freezing point is depressed (due to unsaturated conditions or the presence of salts) or
if the temperature profile is essentially isothermal and near 0~ More reliable
measurements of frost depth are made using electrical resistivity or dielectric constant
measurements because the electrical properties of soil are affected more by phase change
of the pore water than temperature.
KEYWORDS: frost, frost depth, freezing and thawing, thermocouples, thermistors,

electrical resistivity, dielectric constant, time domain reflectometry, geophysics
In cold regions, seasonal freezing and thawing often have a significant impact on
engineering behavior of earthen materials. For example, the strength and stiffness of
pavements increase during freezing and decrease during thawing (Mahoney et al. 1985,
Nordal and Hanson 1987, Janoo and Berg 1996, Jong et al. 1998). Also, the hydraulic
conductivity of hydraulic barriers changes as pore water freezes and thaws (Chamberlain
et al. 1990, 1995; Benson and Othman 1993). In other cases, soil is intentionally frozen
for earth retention or to prevent contaminant migration (Tumeo and Davidson 1993,
Andersland et al. 1996). In each of these situations, the depth or thickness of the frozen
IAssociate Professor, Dept. of Civil & Environmental Engineering, University of Wisconsin-Madison,

Madison, W153706, chbenson@facstafs
2Professor, Dept. of Civil & Environmental Engineering, University of Wisconsin-Madison, Madison, WI
53706, bosscher@engr.wisc.edu.
and/or thawed zones is important to characterize engineering behavior of the soil and the
response of engineered systems.
Since the onset and rate of freezing are usually controlled by nature, frost depths
are often measured remotely to reduce the number of site visits while maintaining a
nearly continuous data record that includes critical events such as rapid thaws (e.g.,
Chamberlain et al. 1995, Jong et al. 1998). Three techniques are commonly used to
remotely measure frost depths: (1) soil temperature measurements, (2) soil electrical
resistivity measurements, and (3) soil dielectric constant measurements. Sensors and
hardware used for each of these techniques can readily be deployed in the field, and
connected to a datalogger and telecommunications system for automated data storage and
subsequent downloading for analysis at the office.
A schematic of a typical setup is shown in Fig. 1. The setup consists of a

datalogger equipped with a multiplexer, a cellular phone with a modem, and a probe (or a
set of probes) equipped with a series of sensors used to measure soil properties related to
frost depth. Additional sensors can be added to monitor meteorological conditions or
other parameters relevant to the project. This simple system can be constructed with off-
the-shelf hardware and software and can be installed and operated remotely. Solar panels
and deep cycle batteries can provide the power needed for operation, and data
observation and retrieval can be conducted from the office using a personal computer
with a modem. Details regarding how to construct such a setup are described in Benson et
al. (1994).
Ground
[ ,~Surface
Cellular Phone ~1~ ///\\\ - ///\\\
I Bundle of --2
Datalogger .,=----Sensor ___ Temperature

I Wires --- = Sensor
Multiplexer
F, n
Z
L' II
FIG. 1- Schematic of typical field setup showing datalogger, cellular phone,

multiplexer, and string of sensors buried in soil.
This paper reviews the three primary measurement techniques and provides
recommendations for their application. The paper is based primarily on the authors'
BENSON AND BOSSCHER ON REMOTE FIELD METHODS 269
experience in remotely measuring frost depths, and its intent is to share this experience
with others practicing cold regions engineering.
Soil Temperature Measurements
Frost depths are often inferred from soil temperature measurements because the
free pore water in moist soils without significant salt concentration freezes at about 0~
Soil temperatures are usually measured using thermocouples or thermistors.
Thermocouples
A therrnocouple consists of two wires of dissimilar metal that are joined at the
point where the temperature is to be measured. This point is called the "measuring
junction." The junction of two dissimilar metals results in an electromotive force (e.g.,
voltage), which is known as the Seebeck effect. The voltage across the open end of the
thermocouple (i.e., the "reference point") is measured, and is proportional to the
difference between the soil temperature at the "measuring junction" and the air
temperature at the "reference point." Thus, when using thermocouples, the temperature at
the reference point must be known or measured using another technique. The
relationship between temperature and voltage is non-linear, and special polynomials are
required to convert voltage to temperature (Bums et al. 1993). These polynomials are
standardized by the National Institute for Standards and Technology (NIST) and are
normally pre-progranuned in modem dataloggers. Thus, the user only sees the measured
temperature as data.
Type-T (copper-constantan) thermocouples are commonly used in practice

because of their suitability for temperatures encountered in most natural environments
and their affordability. Typically 20 AWG wire insulated with a PVC jacket is robust
enough for field applications, although the authors have successfully used wire as thin as
24 AWG. Stranded wire is preferred because it is less likely to crack during installation
or use. However, stranded thermocouple wire is more costly than solid wire.
Thermocouples are normally attached to a multiplexer so that measurements can

be made at various depths using one channel on a datalogger. In this case, the reference
temperature is the air temperature at the multiplexer connection. This temperature is
usually made using a single precision thermistor or a resistance-temperature detector
located adjacent to the multiplexer connections. To ensure that the reference temperature
is representative of all the multiplexer connections, a metallic heat sink is placed over the
connections and the enclosure in which the multiplexer is stored is heavily insulated and
sealed against drafts.
A key advantage of using thermocouples is that they can be easily constructed and
installed in the field. Thermocouple wire can be purchased in long spools and wire can
be laid out from the measuring point to the multiplexer. The measuring junction is then
formed by twisting the two wires together at the measurement point and soldering the
junction using good quality electronics solder with flux. This seemingly minor advantage
becomes significant during installation, because the wire length required between the
measuring point and the multiplexer is rarely known a priori. Relatively long lengths are
acceptable. The authors have used thermocouples with leads 80 m long with no loss in
accuracy (Benson et al. 1994). Another advantage is that the measuring junction can be
moist (or submerged) without affecting the measurement.
Thermistors
A thermistor is a temperature-sensitive resistor. Temperature measurements are
made by inducing a small current across the resistor, and measuring the corresponding
voltage drop. A calibration curve is then used to relate the voltage drop to temperature.
The resistance-temperature characteristics of thermistors are highly dependent on the
manufacturing process, yielding a unique calibration for each type of thermistor.
Thermistor calibrations are normally well-behaved, which makes the conversion from
voltage to temperature simple.
As with thermocouples, thermistors are normally connected to a multiplexer so

that multiple measurements can be made from a single datalogger channel. However, a
temperature reference is not necessary at the multiplexer connection and no special effort
must be made to ensure all multiplexer connections have the same temperature.
The key advantage of using thermistors is that no temperature reference is

required, which makes multiplexing easier. Thermistors are also more accurate than
thermocouples, although the differences in the accuracy become important only when
temperature variations in the milli-degree range become essential (Andersland and
Ladanyi 1994). The primary disadvantage of thermistors is that the calibration is
sensitive to the length of the lead wires, and thus thermistors are normally assembled and
calibrated in the laboratory or factory with a fixed length of wire. In addition, the
temperature sensitive resistor must be carefully sealed to prevent water entry with a
sealant that permits easy heat transfer between the soil and resistor. Leaky thermistors
almost always fail in geotechnical applications. In some cases, high freezing pressures
can alter the calibration of thermistors, resulting in substantial error.
Comparison
In principle, thermocouples and thermistors should yield equally reliable
temperature data for most geotechnical applications. However, engineers usually specify
thermistors based on anecdotal reports suggesting that thermistors are more accurate than
thermocouples. In a recent study the authors installed thermocouples and thermistors at
various depths beneath three secondary highways in Wisconsin to assess whether the two
devices yield similar temperatures (Benson et al. 1997, Jong et al. 1998). Temperatures
were measured using both devices for 18 mos. and then compared.
Data are shown in Fig. 2 that were collected from a thermistor and a
thermocouple installed 45 cm below the base of a pavement near Westby, Wisconsin.
The subgrade was silty clay. Nearly identical temperatures were measured with both
devices for soil temperatures ranging from -6~ to 25~ In this case, type T
thermocouples were used with a solid state multiplexer equipped with a platinum
resistance-temperature detector. Nearly identical results were obtained at other depths

and at the other sites (Jong 1998).
The data shown in Fig. 2 illustrate that thermocouples or thermistors can be used
reliably to measure soil temperatures. Thus, selection of the appropriate device depends
on cost, availability, and ease of installation. In general, thermocouples are less costly
(even when thermocouple-specific multiplexers are used), are more easily installed, and
are more likely to perform over extended periods even in very moist conditions. For
example, Benson et al. (1994) made hourly measurements of soil temperatures at 30
locations with Type-T thermocouples for five years without a single failure.
30 .,,,,i,,,,i,,,,i,,,,I,,,,i,,,,i,,,,l~r,,
25
20
I-
-5
-101( . . .
-5.
. . .
0.
. . .
5.
. . . .
10
. . .
15
. . .
20.
. . . .
25 30
Thermocouple (~
FIG. 2 - Comparison of temperatures measured with a thermistor and a thermocouple

below the base of a pavement near Westby, Wisconsin.
Problems with Frost Depths Inferred from Soil Temperatures
Freezing Point Depression -. Freezing point depression is the reduction in the

freezing temperature of pore water that occurs as soil becomes unsaturated or the pore
water becomes contaminated with salts. In most cases where the soil is moist and the salt
concentration is not unusual, the freezing point is depressed very little and the freezing
temperature is 0~ for practical purposes. For example, Abichou (1993) found that the
freezing point of a low plasticity glacial clay compacted near optimum water content was
essentially 0~ Kraus et al. (1997) report similar results for hydrated bentonite in a
geosynthetic clay liner. In cases where the freezing point is depressed significantly,
laboratory tests can be conducted to determine the freezing temperature and this
temperature can then be used to infer frost depths from soil temperature data.
Freezing point depression is problematic when the salt concentration changes

during the monitoring period, resulting in a change in the freezing temperature. For
example, Leonards and Andersland (1960) show that contamination with a 1M solution
of LiC1 depressed the freezing point of a clay to -8~ This situation can occur in
roadway applications where deicing salts are employed. As deicing salts melt overlying
ice, the salt-laden melt water infiltrates the subsurface. The frozen pore water then melts
with essentially no change in the soil temperature. Consequently, the soil is believed to
be frozen when it is actually thawed. Atkins (1979) describes an illustrative case history
where this type of event occurred.
I s o t h e r m a l C o n d i t i o n s -- Another significant problem associated with inferring

frost depths from soil temperature data is that nearly isothermal conditions can make
locating the frost depth difficult. For example, Fig. 3 shows temperatures beneath a
pavement near Westby, Wisconsin. Throughout February and until mid-March, the
temperature at 152 cm is essentially 0~ Similarly, the temperature is about 0~ at a
depth of 107 cm from mid-March until the end of April. During these periods the exact
location of the frost depth is difficult to determine because it is determined indirectly via
temperature rather than directly by monitoring for phase change.
20 i , i i
Depth:
15
9 107 cm
O 152 cm
0~ to
~ s ~ K ~ Datalogger Down
E 0
-5
I , , I I I
-10
11/1/95 12/17/95 2/1/96 3/18/96 5/4/96
Date
FIG. 3 - Temperatures at depths of 107 cm and 152 cm below the base of a pavement
near Westby, Wisconsin (adapted from Jong 1998).
A more complicated condition can exist during spring thaws, when nearly
isothermal conditions can exist throughout the near-surface soil profile at a temperature
near 0~ (Atkins 1979). When this condition occurs, the frost depth is undefined because
freezing conditions cannot be identified at any depth.
B E N S O N A N D B O S S C H E R ON R E M O T E F I E L D M E T H O D S 273
Electrical Resistivity Measurements
Another technique to measure frost depths is by measuring the vertical profile of

electrical resistivity (Pe), or analogously electrical conductivity (t~e = 1/p~). This method
is superior to temperature measurements because it provides a direct index of phase
change. In particular, when water changes to ice, its electrical resistivity increases
dramatically even though the temperature of the water may change only slightly. A
similar effect occurs in most soils when they freeze or thaw.
Electrical resistivities of three compacted clays varying in plasticity are shown in

Fig. 4 (Abu-Hassanein et al. 1995). The electrical resistivity (60 Hz) of these clays varies
approximately one order of magnitude as the freezing point is crossed. Similar data are
reported by Hoekstra and McNeill (1973). They report electrical resistivities for biotite
granite, saturated sandy gravel, silt, and clay and in each case the electrical resistivity
jumps at least one order of magnitude as the phase change occurs.
10 8 ' i ' ' I
9 Soil A
A Soil B
E o Soil D
?E 1 05
J~
o
% i
10 4
& ~ ~F~DO0O0 GSDo00O0 0 0 0
tr"
9o.-" 10 3
m
I
102 , I , I , , I I
-20 0 20 40 60 80
Soil Temperature (~
FIG. 4 - Electrical resistivity as a function of temperature for three compacted clays

(adapted from Abu-Hassanein et al. 1995).
Probes
Probes for electrical resistivity measurements can be easily made from readily
available construction materials or purchased from geotechnical equipment vendors. A
detailed description of probe construction can be found in Atkins (1990). A brief
summary is described here.
In its simplest form, an electrical resistivity probe consists of an electrical supply

source, two electrodes buried in the soil, a measurement resistor, and a voltmeter (Fig. 5).
9~ AVm
I Measurement
AC
Source - -
Vo
FIG. 5 - Circuit for Electrical Resistivity Probe.
Soil resistivity, Rs, is determined using Ohm's law:
V=IR (1)
where V is the voltage across a resistor having resistance R with a current I. When
Ohm's law is applied to the circuit shown in Fig. 5, Rs can be computed from:
R, ~AV~
( Vo -1) (2)
where Rm is the resistance of the measurement resistor, Vo is the applied voltage from the
source, and AVm is the measured voltage drop across the measurement resistor.
Resistivity (pe) is then computed from the resistance, Rs, via:
where A and L are the cross-sectional area and length of the soil through which the
current flows. Often times A and L are ill-defined because current travels through the
soil in a complex three-dimensional field. This problem is circumvented by ensuring that
each set of electrodes has the same A and L, which results in the same constant of
proportionality between Rs and Pe for each adjacent pair of electrodes. Also, inspection
of Eq. 2 shows that the resistivity can be inferred simply by the voltage drop across the
measurement resistor. Thus, voltage data can be used as a surrogate for resistivity
without loss of generality when determining frost depths.
A suitable arrangement for the electrodes is described by Atkins (1990) and is

illustrated in Fig. 6. A solid PVC rod is used to retain the electrodes in position. Any
practical length rod can be used. The rod is solid so that it does not fill with water during
thawing, which could short out the electrodes. PVC pipe or conduit should not be used
unless it is filled with material that will not leak. A narrow slot is milled along the length
of the rod to carry the wire (24 AWG stranded) for the electrodes. The electrodes are
constructed by wrapping 14 AWG copper wire (i.e., ground wire for household electrical
lines) around the PVC rod three times, and then twisting and soldering the two ends. One
lead wire is also soldered to this junction. The spacing between each adjacent pair of
electrodes is held constant so that L and A in Eq. 3 remain the same.
The probe is inserted in a pre-drilled hole having slightly smaller diameter than
the rod. A rubber mallet can be used to tap the rod in place. A small hole must be used
to ensure good contact exists between the electrodes and the soil and to ensure that an
annulus is not present that can collect melt water. The top of the rod should be below
ground surface to prevent water from intruding along the soil-rod interface. A small
layer of bentonite can be placed above the rod to ensure water does not funnel down
along the leads. The leads can daylight directly above the rod, or be routed through a
shallow narrow trench to a multiplexer.
Electrode Leads to
Datalogger
- - M i l l Slot 6 mm wide
& deep for electrode leads
Seal
14 AWG Copper
Wire Wrap
Electrode
Solid 2 cm
T10 cm
_L
FIG. 6 - Schematic of frost depth probe with electrical resistivity electrodes (adapted
from Atldns 1990).
Data Acquisition
Most modem dataloggers can be used to generate the source and to measure the
voltage drop at the measurement resistor. The source needs to have a capacity of several
rnilliamps, and must supply an AC signal to prevent polarization of the electrodes.
Atldns (1979) suggests an AC frequency < 60 Hz and an operating peak to peak voltage
276 FIELDINSTRUMENTATION
FORSOILANDROCK
of 3 V. Modem dataloggers generally can output this type of signal at frequencies less
than 30 Hz.
A multiplexer is required to switch the electrode leads in and out of the circuit so
that electrical resistivity can be measured at various depths along the rod. One single-
ended channel is required for each electrode. Voltage across the measurement resistor is
measured using a differential input channel on the datalogger each time the multiplexer
connects to a pair of electrodes.
Sample Data
A profile of voltage measurements (i.e., AVm) is shown in Fig. 7 that was
obtained from a resistivity probe placed beneath a pavement near Rhinelander,
Wisconsin. The data were collected on February 8, 1998. Two frozen zones exist. The
deeper zone is from a deep freeze in early January where the frost depth had reached a
depth of 160 cm. Top-down thawing followed during a warm period in late January,
which thawed the upper 70 cm of soil. This thaw was then followed by a short freeze and
a subsequent light thaw, which resulted in the thinner frozen zone near the surface.
ACVolts
0.0 0.2 0.4 0.6 0.8 1.0
I I I I I
Q
"E 50 ~" Frozen

E
100
ft.
~ o I Frozen
_o 150
I:l:l
D
r
~. 200
a
250
FIG. 7 - Voltage profile from frost depth resistivity probe beneath pavement near
Rhinelander, Wisconsin.
A case history is described by Moo-Young et al. (1999) where resistivity probes

were used to determine if frost penetrated the barrier layer in a landfill cover. The barrier
layer was constructed with paper mill sludge. In this application, the resistivity probe
was used to show that frost did not penetrate the paper mill sludge, whereas temperature
measurements showed that a portion of the sludge layer dropped below 0~ during the
winter months.
Advantages and Disadvantages

The primary advantage of the electrical resistivity technique is that the response
being monitored is directly affected by the phase change and thus the frost depth can be
accurately determined. Additionally, the probes are relatively simple to construct and can
be connected to conventional multiplexers used with dataloggers. No special
instrumentation is required and the cost of construction and installation is only slightly
greater than the cost associated with thermocouples.
A disadvantage of the technique is that different soils have different electrical

resistivity at the same degree of saturation due to differences in the electrolyte in the pore
water and surface conductance (Andersland and Ladanyi 1994, Abu-Hassanein and
Benson 1994). Thus, the voltage drop and electrical resistivity are not uniquely related to
phase change. A clear understanding of the subsurface layering is necessary to employ
the method. Another disadvantage is that gaps between the probe and soil invalidate the
method.
Dielectric Methods
Like electrical resistivity, the dielectric constant of moist soil is significantly

affected by freezing and thawing of the pore water. Thus, frost depth can be determined
by monitoring the soil dielectric constant, Ka. The soil dielectric constant is the square-
root volume-weighted average of the dielectric constants of the soil phases (Whalley
1993): air (Ko), solid (Ks), water (Kw), and ice (Ki):
~K'~-~= fo ~ o + fw K~"~ + f~~'~'~ + fi ~ i (4)
where fo, fw, fs, and fi are the volume fractions of air, water, solids and ice (i.e., fo + fw + fs
+ fi =1). Because soil solids, pore water, air, and ice have different dielectric constants
(3-10, 80, 1, and 3, respectively) (Topp et al. 1980, Patterson and Smith 1981), the
apparent dielectric constant of the soil changes significantly as the unfrozen volume
fraction of water changes.
The most common technique to measure Ka is time domain reflectometry (TDR),

which is an electromagnetic technique similar to RADAR. A pulse generator sends
electrical pulses down a coaxial cable. When the pulse encounters a change in the
electrical properties of the cable, a portion of the pulse is reflected back towards the
generator. The elapsed time between emission of the pulse and reception of the reflection
is used to determine the speed of the pulse (Vp), which is a function of Ka:
c
Vp ~ ~ (5)
For soil measurement, Vp is determined by measuring the travel time of the

electromagnetic pulse in a set of un-insulated conductors buried horizontally in the soil.
The conductors are generally placed at the end of a coaxial cable and are referred to as a
"waveguide." A three-rod wavegulde is most common (Fig. 8), although two-rod
waveguides are used. The rods are typically made from stainless steel and are 2 mm in
diameter, 30 cm long, and separated by 2 cm (Benson and Bosscher 1998).
FIG. 8 - Three-rod TDR waveguide.
A typical TDR waveform from a waveguide is shown in Fig. 9 as a graph of

voltage versus apparent distance. The waveform is the response recorded by the device
that generates and receives the pulses. Such devices are commonly called "cable testers,"
because they are employed for testing utility cables. The apparent distance is determined
based on the travel time of pulses measured by the cable tester and assuming Vp equals
the speed of light (i.e., Ka = 1 in Eq. 5). The apparent length of the probe, La, is
determined from the difference in the apparent distances corresponding to the inflection
points in the waveform created by the beginning and ending points of the probe (Fig. 9).
FIG. 9 - TDR Waveguide and Waveform (adapted from Suwansawat and Benson 1998).
B E N S O N A N D B O S S C H E R ON R E M O T E F I E L D M E T H O D S 279
The true length of the probe, Lp, is known. Thus, from Eq. 5,
K~ = " (6)
because La is determined assuming Ka =1 or effectively Vp = c.
KaResponse During Freezing

The change in Ka that occurs during freezing is illustrated in Fig. 10. The data
were collected beneath a pavement near Unity, Wisconsin in a layer of clayey sand
subgrade. A distinct drop in Ka occurs when the soil freezes because the dielectric
constant of water is 80 and that of ice is 3. Similarly, a distinct jump in I ~ occurs when
the soil thaws.
,v. *~
35
," 30 9 ,aw, n .
tl:l
t-
O 25
Q
=" 20
i~ 15
< 5 i i i i I I i ~ i I i i i i I i ~ i I llllllllJllll~lllll
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0
Soil Temperature (~
FIG. 10 - Apparent dielectric constant versus soil temperature beneath a pavement near
Unity, Wisconsin (adapted from Jong et al. 1998).
Figure 11 shows how dielectric constant measurements are superior to

temperature measurements for assessing phase change. The data were obtained 60 cm
beneath a pavement in a silty clay subgrade at a site near Spirit, Wisconsin. The
temperature slowly fell below 0~ over a three-week period in early December 1995 as
the soil began to freeze. Based on the temperature data, frost could have arrived at a
depth of 60 cm any day during this period. In contrast, the dielectric constant dropped
dramatically, as the frost depth reached and passed 60 cm within two days in early
December. Similar statements can be made regarding the thaw period, which began in
early April 1996.
351- ' ' ' ' ' ' ' ~ ' ' ' ~.~qa~ -20
L.
30 I- Freeze -15
-10 -I
~ 25
3
-5 "O
(1)
o 20
0
o {..
"- 15
5
"-'o
N 10 c)
10
E) 5 I~" O Dielectric Constant

15
A Temperature ~l
OF , ~ ~ I ~ , , I , = t I , 1 20
111111995 111/1996 312/1996 5/2/1996
Date
Fig. 11 - Dielectric constants and temperatures measured beneath a pavement near Spirit,
Wisconsin (adapted from Benson et al. 1997).

The primary advantage of measuring frost depth by monitoring dielectric constant
is that the phase change is clearly indicated by an abrupt change in Ka. And, unlike
electrical resistivity measurements, similar Ka are obtained for most soils having similar
water content. That is, for moist soils a nearly unique relationship exists between similar
Ka and the phase of the water (Ka ~ 5-10 for frozen soil and - 15-25 for moist unfrozen
soil). Another advantage of measuring Ka is that unfrozen water contents can be
determined simultaneously with frost depth (Stein and Kane 1983, Smith and Tice 1988,
Spaans and Baker 1995).
The primary disadvantage of dielectric constant measurements is the high cost of

the equipment. A cable tester and its electronic interface typically costs about
US$10,000. In addition, special coaxial multiplexers must be used that have a limited
number of channels (8 - 10). These coaxial multiplexers are typically twice as expensive
as traditional multiplexers used with dataloggers. The probes are also more difficult to
construct, and require a large diameter hole (> 30 cm) for installation.
The three primary methods to measure frost depth have been reviewed in this
paper: (1) temperature measurements made with thermocouples or thermistors, (2)
electrical resistivity measurements and (3) dielectric constant measurements. The
advantages and disadvantages of each measurement technique have been described, and
their relative accuracy in predicting frost depths has been discussed using field data.
Techniques to incorporate the three methods in conjunction with field data acquisition
equipment have also been described.
Comparison of the advantages and disadvantages of the three methods indicates

that the best method to measure frost depth is to monitor the dielectric constant using
time domain reflectometry (TDR). This method is extremely reliable, because the phase
change from water to ice (and vice versa) is associated with an abrupt change in the
dielectric constant. Moreover, essentially the same dielectric constants are obtained,
regardless of soil type. The TDR method is also easy to implement, but the
instrumentation is expensive. Electrical resistivity methods are the next most desirable
because a large change in electrical resistivity of soil occurs when the pore water changes
phase. The electrical resistivity method is also far cheaper to implement than the TDR
method, but electrical resistivity is sensitive to soil type and installation quality.
Measuring temperatures with thermocouples or thermistors is the least desirable method,
but is also the least expensive and the simplest to implement. Either thermocouples or
thermistors can be used, although installations using thermocouples typically are less
expensive. The main drawback associated with the temperature method is that freezing
and thawing temperatures in soil do not equal those of free water, and can change as the
solute concentration of the pore water changes. Consequently, soil temperature can be an
unreliable index of frost depth.
Acknowledgement
A variety of sponsors have supported the authors' research on field

instrumentation including the National Science Foundation, the United States
Environmental Protection Agency, Envotech Limited Partnership, NCASI, the Wisconsin
Dept. of Transportation (WisDOT), and WMX, Inc. This support is gratefully
acknowledged. The findings and opinions reported in this paper are solely those of the
authors, and endorsement by the sponsors should not be implied. Brian Gaber of
WisDOT provided the data from the resistivity probe in Rhinelander, Wisconsin.
References
Abichou, T. (1993), Field evaluation of geosynthetic insulation for protection of clay

liners, MS Thesis, University of Wisconsin-Madison.
Abu-Hassanein, Z. and Benson, C., and Blotz, L. (1996), Electrical resistivity of

compacted days, J. of Geotechnical Engineering, 122(5), 397-407.
Abu-Hassanein, Z. and Benson, C. (1994), Using electrical resistivity for compaction

control of compacted soil liners, Proc., Tailings and Mine Waste '94, Jan. 19-21, Fort
Collins, CO, 177-189.
Andersland, O. and Ladanyi, B. (1994), An Introduction to Frozen Ground Engineering,

Chapman and Hall, New York.
Andersland, O., Wiggert, D., and Davies, S. (1996), Hydraulic conductivity of frozen
granular soils, J. of Environmental Engineering, ASCE, 122(3), 212-216.
Atldns, R. (1979), Determination of frost penetration by soil resistivity measurements,

Special Report 79-22, US Army Cold Regions Research and Engineering Laboratory,
Hanover, NH.
Atldns, R. (1990), Frost resistivity gages: assembly and installation instructions,

Technical Note, US Army Cold Regions Research and Engineering Laboratory,
Hanover, NH.
Benson, C. and Bosscher, P. (1998), Metallic time domain reflectometry (TDR) in

geotechnics: a review, Nondestructive and Automated Testing for Soil and Rock
Properties, ASTM STP 1350, W. A. Marr and C. E. Fairhurst, Eds., American Society
for Testing and Materials, West Conshohocken, PA, 1998, in press.
Benson, C., Bosscher, P., and Jong, D. (1997), Predicting seasonal changes in pavement
stiffness and capacity caused by freezing and thawing, Geotechnical Engineering
Report 97-9, University of Wisconsin-Madison.
Benson, C., Bosscher, P., Lane, D., and Pliska, R. (1994), Monitoring system for
hydrologic evaluation of landfill final covers, Geotech. Testing J., ASTM, 17(2), 138-
149.
Benson, C. and Othman, M, (1993), Hydraulic conductivity of compacted clay frozen and
thawed in situ, J. of Geotech. Eng., ASCE, 119(2), 276-294.
Bums, G., Scroger, M., and Strouse, G. (1993), Temperature-electromotive force

reference functions and tables for the letter-designated thermocouple types based on
the IPTS-90, NIST Monograph 175, Washington, DC.
Chamberlain, E., Erickson, A. and Benson, C. (1995), Effects of frost action on

compacted clay barriers, Geoenvironment 2000, ASCE, GSP No. 46, 702-717.
Chamberlain, E., Iskander, I., and S. Hunsicker (1990), Effect of freeze-thaw cycles on
the permeability and macrostructure of soils, Proc. Frozen Soil Impacts on
Agricultural, Range, and Forest Lands, K. Cooley, ed., US Army Cold Regions
Research and Engineering Laboratory, Hanover, NH.
Hoekstra, P. and McNeill, D. (1973), Electromagnetic probing in permafrost, Proc. 2nd

Intl. Conf. on Permafrost, Yakutsk, USSR, National Academy of Sciences,
Washington, DC, 517-526.
Janoo, V.C., and Berg, R.L. (1996), PCC airfield pavement response during thaw-
weakening periods: a field study, CRREL Report 96-12, U.S. Army Corps of
Engineers, Cold Regions Research and Engineering Laboratory, Hanover, NH.
BENSON AND BOSSCHERON REMOTE FIELD METHODS 283
Jong, D., Bosscher, P., and Benson, C. (1998), Field assessment of changes in pavement
rnoduli caused by freezing and thawing, Transportation Research Record, in press.
Jong, D. (1998), Freeze-thaw effects on the stiffness of flexible pavement systems, MS

Thesis, Dept. of Civil and Environmental Engineering, University of Wisconsin-
Madison.
Krans, J., Benson, C., Erickson, A., and Chamberlain, E. (1997), Freeze-thaw and
hydraulic conductivity of bentonitic barriers, J. of Geotech. and Geoenvironmental
Eng., ASCE, 123(3) 229-238.
Leonards, G. and Andersland, O. (1960), The clay-water system and the shearing
resistance of clays, Proc. Research Conference of Shear Strength of Cohesive Soils,
ASCE, 793-818.
Mahoney, J., Lary, J., Sharma, J., and Jackson, N. (1985), Investigation of seasonal load
restrictions in Washington State," Transp. Research Record 1043, Transportation
Research Board, National Research Council, Washington, D.C., 58-67.
Moo-Young, H., LaPlante, C , Zirnmie, T., and Quiroz, J. (1999), Field measurements of
frost penetration into a landfill cover that uses a paper sludge barrier, Field
Instrumentation for Soil and Rock, ASTM STP 1358, G. N. Durham and W. A. Marr,
Nordal, R. and Hansen, E. (1987), The Vormsund Test Road, Part 4: Summary Report,
Norwegian Road Research Laboratory.
Patterson, D. and Smith, M. (1981), The measurement of unfrozen water content by time
domain reflectometry: results from laboratory tests, Canadian Geotech. J., 18, 131-
144.
Smith, M. and Tice, A. (1988), Measurement of the unfrozen water content of soils;
comparison of NMR and TDR methods, Report 88-18, U.S. Army Cold Regions
Research and Engineering Laboratory, Hanover, NH.
Spaans, E. and Baker, J. (1995), Examining the use of time domain reflectometry for
measuring liquid water content in frozen soil, Water Resources Research, 31(12),
2917-2925.
Stein, J. and Kane, D. (1983), Monitoring the unfrozen water content of soil and snow
using time domain reflectometry, Water Resources Research, 19(6), 1573-1584.
Suwansawat, S. and Benson, C. (1998), Cell for water content calibration using time
domain reflectometry, Geotech. Testing J., ASTM, in press.
Topp, G., Davis, J., and Annan, A. (1980), Electromagnetic determination of soil water
content, Water Resources Research, 16(3), 574-582.
Tumeo, M. and Davidson, B. (1993), Hydrocarbon exclusion from groundwater during

freezing, J. of Environmental Engineering, ASCE, 119(4), 715-724.
Whalley, W. (1993), Considerations on the use of time-domain reflectometry (TDR) for

measuring soil water content, J. of Soil Science, 44(1), 1-9.
L. Randall Welch and Paul E. Fields
Integrated Automation of the New Waddell Dam

Performance Data Acquisition System
Reference: Welch, L. R., and Fields, P. E., "Integrated Automation of the New
Waddeil Dam Performance Data Acquisition System," Field Instrumentationfor Soil
andRock, ASTMSTP 1358, G.N. Durham and W.A. Man', Eds., American Society for
Testing and Materials, West Conshohocken, PA, 1999.
Abstract: New Waddell Dam, a key feature of the U.S. Bureau of Reclamation's Central
Arizona Project, had elements of its dam safety data acquisition system incorporated into
the design and construction. The instrumentation array is a reflection of the dam's large
size and foundation complexity. Much of the instrumentation is automated. This
automation was accomplished while maintaining independent communication
connections to major divisions of the instrument array. Fiber optic cables are used to
provide high quality data, free from voltage surges that could originate in a nearby
powerplant switchyard or from lightning. The system has been working well but there
are concerns with a lack of continued equipment manufacturer support.
Keywords: dam performance, automated dam monitoring, New Waddell Dam, fiber
optic cable
Introduction
New Waddell Dam is 104 m high with a length of 1 433 m. It is located on the south
end of the 1 360 x 106 m 3 Lake Pleasant reservoir about 40 km northwest of Phoenix,
Arizona (Designers' Operating Criteria 1993). It is a zoned embankment structure having
a thin central impervious core, internal chimney and blanket drainage zones and gravel
cobble fill shells. Slopes are 2:1 downstream and 2 1/4:1 upstream. Construction was
completed in 1992.
The dam is founded on bedrock except in the narrow valley center where the
foundation excavation extends only to the top of a dense, permeable, older alluvium.
Crest elevation, 527 m, provides 1.5 m of freeboard above the maximum water surface
Technical Specialist and Automation Specialist, respectively,

Structural Behavior and Instrumentation Group,
U. S. Bureau of Reclamation, P.O. Box 25007, Denver, CO 80225.
285
elevation. Crest width is 12 m and paved to enhance the ability to detect deformation
cracks and to increase erosion resistance. A 183 m-long grouting gallery is beneath the
embankment core on the right abutment just downstream of embankment centerline and
upstream of a pump/generating (P/G) plant.
Hydraulic structures include two 4.3 m-diameter Central Arizona Project (CAP) outlet
works tunnels (connect to the P/G Plant), a 4.9 m-diameter river outlet works tunnel, a
198 m-wide ogee crest overflow service spillway, and a 107 m-wide fuseplug dike
auxiliary spillway. Outlet capacities are 122 m3/s and 158 m3/s, respectively. The service
spillway, capacity 5 235 m3/s, is on the reservoir rim and approximately 2 286 m west of
the right end (looking downstream) of the dam. The auxiliary spillway, capacity 3 540
m3/s, is on the fight reservoir rim about 1 220 m west of the service spillway.
New Waddell Dam provides the primary storage reservoir for the CAP. It also stores
and passes the flows of the Agua Fria River, part of which is allocated to the Maricopa
County Municipal Water Conservation District.
FIG. 1-New Waddell Dam
CAP water is delivered to the dam, from the Colorado River at Lake Havasu, through
nearly 260 km of the Hayden-Rhodes Aqueduct system and about 10 km of the reversible
Waddell Canal. Power is generated from the releases of reservoir water through the
Waddell Pumping-Generating Plant at the base of the dam on the right abutment.
Figure 1 is a photo of the dam.
WELCH AND FIELDS ON THE DAM PERFORMANCE DATA 287
Foundation
The site location for New Waddell was not a first choice. It was selected after
environmental considerations ruled out more favorable locations. Geology of the
foundation has a complex history and structure. This includes a fault zone, three buried
stream channels, layers of weak material and layers of high permeability. Foundation
rock consists largely of volcanic material such as tuff and andesite.
Instrumentation
New Waddell Dam was provided with instruments to allow the measurement of
embankment and foundation deformation, appurtenant structures deformation, collected
seepage flow, and piezometric pressures. Specific portions of the embankment and
foundation were selected for internal deformation measurement and provisions for
monitoring general surface deformation of the embankment were also made. Structural
measurement points were installed in both CAP outlet works tunnels, intake towers, and
service spillway walls. In addition, tiltmeters were installed on each intake tower and
each of the three associated access-bridge piers. Weirs were installed in the toe drain
system to allow monitoring of the seepage trends of individual segments of the
fonndation. Arrays ofpiezometers were installed at five cross sections and along the
downstream embankment toe. Data from those piezometers show the effect of cutoffs
and drainage features upon the piezometric pressure distribution.
Automation of a portion of the instrumentation was desirable for two reasons. 1)
Due to the difficult nature of the foundation, many instruments were installed to check
design assumptions. (Monitoring during first filling was necessarily frequent and the
number of instruments involved required a large manpower commitmer~i evenwith
automation.) 2) The dam is upstream of a large and rapidly growing suburban area.
Consequently, future dam safety considerations are likely to continue requiring the
highest level of instrument monitoring and dam surveillance.
Generally, nine areas of the embankment are instrumented (Prudhom 1995). Other
than the embankment surface, downstream toe, and appurtenant structures, these areas
include the following six instrument sections:
Section A, a single zone of highly fractured rock (an inactive fault zone)
beneath the far right embankment and downstream of a secant pile cutoff
wall.
Section B, a shallow but steep-sided ancient stream channel backfilled

beneath the embankment in the vicinity of the outlet works tunnel.
Section C, downstream of a shallow cutoff wall beneath the central right

portion of the embankment.
Section D, downstream of a deep cutoff wall beneath the central left

portion of the embankment.
Section E, a zone of possible weak tuff material beneath the downstream

embankment shell at the foot of the left abutment.
Section F, a large mass of core material overlying a steep left abutment

slope.
Below is a table of the phenomena measured and instruments used to obtain the
measurements:
TABLE--New Waddell Dam instrumentation
Measurement Instrument
Embankment Surface Measurement Points
Reference Piers
Embankment Internal Inclinometers
Deformations Extensometers*
Foundation Internal Inclinometers

Shear Indicators*
Appurtenant Structures Measurement Points
Tiltmeters*
Cutoff Wall and Drain Vibrating-wire
Efficiency Piezometers*
Foundation Water Vibrating-wire

Piezometric Level Pressure Piezometers*
Seepage Exit Gradient Porous-Tube

Piezometers*
Local Groundwater Level Observation Wells
Toe Drain Flow V-notch Weirs*
Seepage Quantity
Seeps and Springs Various Devices
Automated Readings
Design provided ready accommodation for an automated data acquisition system.

The layout and components are such that most of the piezometric level monitoring, much
of the seepage quantity monitoring, and a significant portion of the internal and structural
deformation monitoring were easily automated. Five instrument terminals service the six
instrtnnent sections listed above. A fiber optic cable system connects these instrument
terminals to a central monitoring station in the switchyard control house. Fiber optic cable
was chosen for its indifference to lightning and radio interference and superior data
transmission. Automated installations are indicated by an asterisk in the table above and
are the only instruments discussed below.
Figure 2 shows general locations of the embankment performance instruments with an
emphasis on the automated installations. Measurement points, observation wells and
other non automated instruments are not shown.
Internal Embankment Deformations
Measurements of internal embankment deformations are obtained from inclinometers

and embedded extensometer installations. Only the extensometers are automated. They
allow measurements of lateral compression and extension along lines parallel to the
embankment centerline, at 6 m intervals, at the elevations of the installations. Locations
of the extensometers are in instrument section F.
The possibility of embankment cracking, caused by an abrupt change in steepness of
the left abutment profile, is monitored with an embedded extensometer installation. The
installation is a series of extensometers placed end-to-end to span the entire mass of the
core material that may be subjected to tensile stress. It is high in the core and along the
embankment centerline.
Each individual interval of the extensometer installation (assembly) is 6 m long and
can measure up to +/- 75 mm of horizontal displacement parallel to the embankment
centerline. The left end of the installation is anchored into a stable rock portion of the
foundation creating a reference point. Horizontal displacements of the flanges on both
ends of each extensometer interval are measured relative to the stationary reference point
to provide absolute movements.
Readings of all the extensometer intervals are obtained at a central location,
instrument terminal F. A switch box within the terminal allows selecting individual
intervals. Each interval is identified by the number of the flange on the right end.
Readings are displayed as a percent of the 150 mm full range. Measurement resolution is
0.15 ram.
Internal Foundation Deformation
Measurements of internal foundation deformations are obtained from three shear

indicators and four inclinometer casing installations. The shear-indicators show the
elevation of horizontal displacement beneath the downstream shell. Use of them reduces
the needed frequency of inclinometer readings. The shear-indicators consist of a parallel
electrical circuit of resistors. Readings are simply a count of the number of resistors in
the circuit from each end. A shear (break) in the circuit results in lower reading values.
IX.)
r
0
"13
FiberOpticCable m
Instruments .................... r-
tD
'~.,,~ FiberOpticCable
)("~./~ \ Toswitchyard N
|~'x'~"JJ ~ controlhouse Of)
"-I
StabilityBerm~. ~ / I
",,,.a",c~.\~I / TerminalB / C
~ TermJnaL. C - D ~ , ~ / / // E
m
~-~erminalE~ ~ ~y~ ~ / ~T~-I,/ z
.-I
\ / ~ /\ ~ \\ I /Jl fl >
-I
// \ \
~ ~'q,,-~.--~\\.Pu,
\x. Box~ ~1~
! ./'~" III ! lii ~ A 5
9 \,q, r " .... Z
Term,na,F\ Y L i ~ instrument . . . ~ ~ x \ . . . . J ~ l ~ "33
0
:g
O0
o
\\ II I C -I
~ I ~ e e t l o n E "~ ~ . Instrument x,, II | i t / \ Instrument
\ SectionF ~ SectionC \~..%~'~'~l~ SectionA Z
\Instrument , ~ j j~-~-------- AccessBridge U
SectionD ~ II j L J [ J ~ T u n n e l s , CAP 0
~ outletworks o
7~
TerminalG
2010.......0 2~}0 4~0 ~0 IntakeTowers
SCALEOFFEET
FIG. 2 - Instrumentation Cable & Terminal Location
Shear Indicators
All shear indicators extend 30 m into possibly weak foundation material, at

instrumentation section E. These shear indicators, along with nearby piezometers to
show foundation pore pressure, make up instrument section E. Like the inclinometers,
these installations provide data on horizontal deflections, however, no measurement of
the amount of shear is given. They are capable of indicating that shear of 3 mm or more
has occurred and approximately at what depth it occurred.
Shear indicators are read quickly and thus more frequently than inclinometer casings.
If shear is shown, the manual reading frequency of nearby inclinometers would be
increased. Polyvinyl chloride (PVC) casing was installed within the shear-indicator
borehole to protect the installation from minor displacements that could result from
caving or other readjustment of stress that occurs only near the borehole. Because of the
specified spacing of electrical resistance elements (0.3 m) the shear zone elevation
bounds can be determined within 0.3 m.
Appurtenant Structures Deformation
Tiltmeters are installed atop each of three piers supporting the CAP intake tower
access bridge and atop the two intake towers. The bridge pier installations are placed
horizontally on the upper surface of the bridge pier cap, between the bridge girders.
Tower tiltmeters are on the floors inside each respective control house.
Each tiltmeter contains two orthogonally fixed electrolytic transducers to allow
measurements of tilt in two planes. The resolution is 0.2 arc second with a repeatability
of 0.4 arc second. The range is :k4.6 ~ A power supply unit and switch box that services
all the tiltmeters are in the right CAP tower. This unit operates on 105-205 VAC with
internal backup batteries. A portable digital readout unit can be connected to the switch
box to obtain readings. These readings, even when not automated, are obtained easily
and much more frequently than surveys of the associated structural measurement points.
Use of the tiltmeters reduces the number of costly surveys that would otherwise be
required to monitor the stability of the tall structures associated with the CAP tunnels.
Seepage Quantity
Seepage is collected and measured by a drain system that runs within and parallel to
the downstream toe of the embankment. The surface of the embankment foundation
downstream of the core is divided into sections by concrete baffle walls contained within
the blanket drain. Each such section drains through a single V-notch weir placed within
a toe drain inspection well located at the lowest comer of the area defined by the baffle
walls. Seven weirs are on the right abutment and two on the left.
Accuracy of the measurement depends upon flow, however, the measurements are
expected to be accurate to the nearest 10 L/s for flows between 75 and 1 500 L/s. Besides
a staff gauge, an electrical (vibrating-wire) pressure transducer is installed upstream of
each weir. Cables from the weir transducers are routed to the nearest instrument terminal.
This allows the weirs to be read remotely at the instrument terminals and be automated.
The weir transducers are placed deep into the weir pool so that a zero reading on the
weir crest would still result in a significant amount of pressure on the transducer. That
was done to avoid the problems of obtaining accurate readings for near zero flows: The
depth range of these transducers is zero to 3.4 m, however, their calibration is based only
upon the transducer behavior in the zero to 0.6 m range.
Cutoff Wall and Drain Efficiency
The primary performance monitoring at instrumented sections, A through D, is for the

efficiency of foundation-seepage cutoff features.
Efficiency of the cutoff walls is determined by the hydrostatic pressure distribution
within the foundation and lower portions of the embankment core. At instrument section
A, it is the efficiency of a secant pile cutoff wall that is of interest. At section B, it is the
efficiency of a grout curtain and backfill of an excavated buried channel that is of interest.
An absolute determination of efficiencies will not be available because, to preserve core
integrity by avoiding the use of cable trenches upstream of centerline, no instruments are
installed upstream of the embankment centerline. Four piezometers distributed between
instrument sections B, C and D and embedded in trenches in the lower core are the only
piezometers embedded in the core. They show the piezometric effect resulting from
seepage passing over the top of cutoff features.
Only vibrating-wire piezometers (VWs) are used. Closed piezometers are necessary to
provide a practical means to monitor piezometric levels beneath such a large
embankment. Cables run downstream from the transducers, within the sandy portion of
the drainage blanket, to the nearest instrument terminal.
All vibrating-wire piezometers and transducers at the site contain thermistors. This
allows the application of temperature corrections. For piezometers, temperature
corrections are insignificant due to the low in situ temperature variation and instrument
insensitivity to such variations. However, using seasonal variations of temperature data
does provide a sense of speed and/or quantity of seepage flow. Only larger flows moving
relatively fast can significantly affect the surrounding ground temperatures.
At sections A through D, the piezometers closer than 15 m to the embankment
centerline and deep, below elevation 488 m, have a range of zero to 168 m of pressure
head (17 of these high range transducers are in use). All others beneath the embankment
have a lesser range of zero to 67 m. Specified accuracy of all vibrating-wire piezometers
was 0.5 percent of full range.
Instrunaent section A has three piezometers in each of five drill holes beneath the
downstream shell and downstream of a secant pile cutoff wall. In a plan view the array
forms a general "T"-shape. The cap of the "T" is a plane ofpiezometers located near the
wall to detect wall leakage. The stem is placed to map the energy dissipation of seepage
as it moves downstream of the wall. (This pattern of piezometers is used downstream of
each of the three foundation cutoff walls). One other piezometer is placed within the base
of the chimney drain above each "T" section. At depth, these piezometers show the
upward gradient. The shallow piezometers, including the chimney drain installation,
show the capacity of the drainage system to convey collected seepage.
Instrument section B has two shallow foundation piezometers and four embankment
piezometers. The embankment piezometers provide data from a horizontal slice of the
piezometric pressure distribution and with the foundation installations provide an
indication of flow direction. Effectiveness of the drainage features is shown by the three
downstream piezometers in an upper row. Foundation installations are high range
transducers beneath the core.
At instrument section C two columns of two piezometers each are installed in drill
holes just downstream of the cutoff wall beneath the downstream portion of the core and
beneath the drainage blanket. Also, two columns of two piezometers are installed at two
separate offsets downstream of the cutoff wall. This foundation array allows a mapping
of the piezometric pressure distribution in the region of the foundation and core just
downstream of the cutoff wall and reveals any foundation piezometric levels greater than
the tailwater elevation. One piezometer is embedded in the lower core downstream of the
cutoff wall. Six piezometers of this group of 10 are 15 m or more downstream and have a
range of zero to 67 m of pressure head. The other four have a zero to 167 m range.
Instrument section D is similar to sections A and C. However, the array is larger and
contains 18 piezometers. The downstream side of the cutoff wall is monitored with three
elevations of piezometers. Each piezometer is installed in a drill hole and has a high
range. There is also a line of two embedded in the lower core (monitoring wall
overflows) and one column of three under the downstream core edge and one column of
three under the drainage blanket. As in section C, the array shows where pressure is
greater than tailwater, the direction of flow, and the ability of the blanket drain to convey
collected seepage.
Foundation Pore Pressure
Instrument section E contains three columns of three piezometers each beneath the
downstream shell. These vibrating-wire piezometers, installed in drill holes, are for
monitoring pore water pressure within possibly weak tuffaceous foundation materials.
Each column of 0- to 60 m range piezometers is near a shear indicator. The drill holes
extend to a depth of 30 m.
Seepage Exit Gradients
Seventeen porous-tube piezometers (PTs) placed approximately at a 30 to 60 m

spacing are along the downstream embankment toe. These locations are on the abutments
above tailwater. All are used to detect the presence of water in zones of higher
permeability and to allow an estimate of the exit gradients if seepage does occur. In some
cases these toe installations are complimentary to the instrumented sections. All behave
as standpipe installations, however, each is outfitted with a vented vibrating-wire
transducer having a range of zero to 34 m of head. Cables from these transducers are
routed to the nearest instrument terminal. The transducers allow the remote standpipes to
be read more quickly and at a central location. If a transducer malftmctions, manual

readings can be taken and the transducers easily replaced.
Data Acquisition System
As shown on Figure 2, there are six instrument terminals. Each contains a

conventional 120-volt power supply with a battery backup. These terminals allow
centralizing the reading of 177 instruments. Switch boxes that allow manual as well as
automated readings for all of the following instrument installations are found in the
terminals:
9 83 vibrating-wire piezometers
9 9 weir vibrating-wire transducers
9 69 thermistors
9 3 shear indicators
9 8 extensometers intervals
9 5 tiltmeters
These terminals are designated as A through G to correspond to the instrument

sections. Two instrumentation sections, C and D, terminate at a single terminal
designated as C-D. Terminals A and B are bolted on the walls within the right abutment
grouting gallery; A is in the west end near the junction with a west access adit; B is near
the east entrance. Terminals C-D, E and F are within concrete manhole structures as
shown on Figure 3. Terminal G is the tiltmeter wiring panel in the right CAP intake
tower control house. All terminals are supplied with electrical power, grounding, and
protection from transient voltage and electromagnetic interference (EMI). EMI
protection is circuitry within the switch boxes that contains gas discharge tubes. This is
only limited protection but helps insure that an electrical potential that arises at the
terminal will not be transmitted to the transducers.
Each terminal other than G is connected by a separate heavy-duty fiber optic(FO)cable
to a point under a raised floor in the P/G plant switchyard control house. On each FO
cable end is a bidirectional signal conversion interface device called an optical-
transceiver. They are AC powered and convert electrical signals to optical signals and
vice versa. Each is provided with a water-resistant enclosure.
The optical fiber is multi-mode having a 62.5 micron core specified to have no signal
point loss greater than 0.2 decibels. Signal attenuation is no more than 3.5 decibels per
kilometer with a bandwidth of at least 160 megahertz. No splices were allowed. Each
FO cable was field tested twice, once on the reel after delivery and again immediately
after installation. Optical time domain reflectometry was used on both cable ends in the
testing to verify cable length, signal attenuation and that there were no discontinuities or
breaks. The optical fiber is embedded in a loose tube baffle and has a minimum bend
radius of at most 150 mm. "ST" connectors were used in the coupling with the
transceivers.
FIG. 3-Typical terminal well equipment layout.
FO cable routing is through a 75 mm diameter buried flexible polyethylene electrical

duct. No elbows or other joints were needed to make duct bends. The duct lays flat when
uncoiled, was lubricated internally and contained a 6 mm rope for pulling the FO cable.
Terminal G is connected to the central network monitor by UHF radio through
terminal F.
During construction, automation equipment was installed at each terminal location as
instrumentation installations were completed. This provided an opportunity to evaluate
and test the automation equipment prior to first filling. Reclamation personnel performed
all system installation, programming and testing.
A measurement and control unit (MCU) capable of reading up to 70 instruments is
located at each terminal. The MCUs are capable of reading a variety of instruments
depending upon which multiplexer modules are installed. Each is programmable and
independently performs all physical measurements and computations based on user
specified parameters according to specified scheduling and sequencing information.
During construction the MCUs were linked to the network monitoring station (NMS)
located in the switchyard control house by means of UHF radios. This was done to
reduce the potential for fiber optic cable damage while heavy construction was underway.
As previously noted, all but one of the MCUs are networked to the NMS with the fiber
optic cable. Each wired MCU is connected separately to the NMS with a duplex cable
(two optical fibers). The duplex feature assured independent two-way operation of each
MCU while allowing the use of fewer and simpler fiber optic transceivers.
The NMS consist of a microcomputer with telephone modem, WWV time-code
receiver, and Geostationary Orbiting Environmental Satellite (GOES) transmitter.
GOES
Geoatationary Operational
Environmental Satellite ~
~--~.._.~ ~ ~ ~ . . ~ ~.-- ~ Bureauof Reclamation
" ~ ~ BuroXXau
ofRealamatlo.~
New Waddell Dam ~ . _ ~
MCU- Measurement and Control Units ~1~ Central ArizonaWatar

NMS9Network Monitor Station Conservation District
FIG. 4-Communication links to a u t o m a t e d

instruments.
Communications software is installed on the NMS to allow downloading by the Project

Office in Phoenix, Arizona and Reclamation's Technical Service Center in Denver via
telephone modems. The NMS is the logging destination for all measurement data from
the MCUs. Data can be logged to the database, screen printer, and the GOES transmitter.
Remote users dialing into the system may request real time measurements, reprogram,
transfer files and trouble shoot. Figure 4 illustrates the communication links.
It is noteworthy that the MCU and NMS equipment is no longer supported by the
manufacturer. Although the equipment continues to perform well, there will come a time
when failures will begin to be a problem. Exact replacements are no longer available.
Conclusion
The performance instrumentation at New Waddell Dam represents a state-of-the-art

system for monitoring the behavior of a large embankment dam on a complex foundation.
It represents the first case for Reclamation where full advantage was taken of the
opportunity to design an infrastructure specifically for the automation equipment.
Unique aspects, in Reclamation experience, included the embedment of embankment
extensometers, the use of shear indicators beneath the embankment and the use of fiber
optic signal cable for performance data acquisition. All of those components are
integrated into the automation system. Automation has worked well and has been a
necessary feature in this age of"downsizing" and growing populations exposed to dam
hazards. A specialized technician was needed on site two for two days to check fiber
optic cable quality. However, the cable proved to be sturdy and there were no other
complications. Even though this automation has been working well, designers of long
term monitoring systems for dams are cautioned to consider the ability of automation
equipment manufactures to provide continued service.
References
"Designers' Operating Criteria, New Waddell Dam," U.S. Bureau of Reclamation,

Engineering and Research Center, Denver, Colorado, March 1993
Prudhom, B.D., July 1995, "Final Instrumentation Report (L-16) - New Waddell Dam -
Stage 2," U.S. Bureau of Reclamation, Phoenix Area Office, Phoenix, AZ.
William S. Brokaw l
Using Geomatics in the Acquisition and Management of Field Data
REFERENCE: Brokaw, W. S., "Using Geomatics in the Acquisition and Management

of Field Data," Field Instrumentation for Soil and Rock, ASTM STP 1358, G.N. Durham
and W. A. Mart, Eds., American Society for Testing and Materials, West Conshohocken,
PA, 1999.
ABSTRACT: The term Geomatics refers to the integration of geographic information

systems (GIS) with global positioning satellite (GPS) technology, remote sensing and
communication technology, (i.e., the Intemet). A project involving the collection of
environmental data at a site follows a series of consecutive events. The initial step is to
identify project parameters and determine data needs. Data capture methods (i.e. collection
and analyses) are selected and then implemented in the field. Once the data are generated,
organization and communication are essential in making the data useful. The objective of
using Geomatics is to facilitate the acquisition and management of data and to provide a
mechanism for effectively communicating this information.
Before a field work is initiated, valuable information about the site and surrounding area
can be obtained via the Intemet. Homepages maintained by the US Geographic Survey, US
Census Bureau and environmental organizations provide maps and other regional data in
digital format, enabling a project manager to assess a wide range of regional conditions.
Aerial and satellite photographs can be obtained to further understand the site and
surrounding area. As site-specific data are generated, the geographic coordinates of each
sampling point can be accurately and rapidly determined with GPS receivers. Using GIS
sottware, a project manager can efficiently integrate the GPS geographic coordinate data
with analytical data and the information obtained from the Intemet-aerial photographs. GIS
serves as the primary data management engine, organizing and providing access to all
information generated as part of the project. The Intemet can be utilized to disseminate the
information to concerned parties.
KEYWORDS: geomatics, GIS, GPS, intemet
Geomatics, a term more commonly used in Canada than in the US, refers to a set of
technologies used to manage spatially related data. The homepage for the Ministry of
1Scientific Associate, LFR, Raritan, New Jersey 08869.
BROKAW ON MANAGEMENT OF FIELD DATA 299
Natural Resources of Canada defines Geomatics as "...the science and technology of

gathering, analyzing, interpreting, distributing and using geographic information.."
(www.geocan.NRCCan.gc.ca). The University of Florida homepage adds "...it is a new
profession that includes activities previously known as surveying, mapping,
photogrammetry and geographic information systems (GIS)." (www./surv.ufl.edu:443).
Geomatics technologies are efficient and powerful tools that can be used to acquire,
manage and communicate environmental information. Over the last few decades, advances
in hardware and soRware took these computer tools from the mainframe to the desktop
computer. The natural progression is to take these tools from the desktop into the field,
where the data originate.
By nature, environmental projects revolve around data and progress in three general
phases.
1. Desk-based collection and processing of existing data to plan the field work,
2. Field-based data collection, verification, updating, and
3. Use of data in decision making. (Carver et al. 1996).
An environmental investigation is often an iterative process. The above phases may

be conducted repeatedly for a single site, especially when monitoring is involved.
Geomatics integrates well with iterative activities due to its many components and
flexibility.
This paper follows the flow path of a typical field project involving the assessment
of a known or potentially contaminated site. Opporttmities are identified within each phase
where Geomatics can be incorporated into the process.
Planning
Sampling projects are undertaken when there is a need to make a decision about a
site. It may be a purchase decision (i.e. a due diligence investigation), to comply with a
regulatory directive or to determine if a remedial action is needed. The data acquisition
objectives are identified and then a scope of work is developed and agreed upon by the
parties involved.
Implementation of a project usually begins in the office, on the desktop. Data needs
are identified and procedures formulated to acquire the data. It is also important to consider
the ultimate use of the data. Will they be used to develop fate and transport models or to
assess risk? Is there likely to be a long-term investigation, or is the project simply a brief,
short-term investigation?
During the initial planning phase, existing data are acquired and evaluated such as:
1. Surface and subsurface hydrologic data,
2. Topography and planemetrics,
3. Historic aerials and orthophotos,
4. Historic maps and surveys,
5. Soil logs and geotechnical data, and

6. Previous laboratory result reports (Douglas 1995).
To take advantage of Geomatics, pertinent, existing information may need to be

converted into digital format. Hardcopy data, from preliminary assessments or prior
investigation should be organized into a database management system (DBMS), 1995. The
DBMS allows a user to sort, retrieve, analyze, delete and update the data as needed. Paper
maps, drawings and photographs can be digitized or scanned. When possible, original
computer files such as computer-aided design (CAD) drawings or orthophotos should be
obtained (if they exist). The scale, orientation and projection of the maps and drawings all
need to be known so that the maps and images can be properly integrated.
The source and date of information should be carefully recorded. In addition, users
should be aware of the accuracy and precision of the maps and images that are utilized.
Accuracy is how correct, or free from error, the data are. Precision is a measure of how
fmely one is able to measure differences in either position or value (Madry 1996). These
attributes of existing data (source, age, accuracy and precision) allow a user to assess the
quality and reliability of the data.
When the data and other information are in digital format they can be incorporated
into a geographic information system (GIS) which couples the information in the DBMS
with the maps, images and other graphics. In addition to providing the organizational and
analytical power of a DBMS, GIS allows a user to visualize the information. GIS provides
an easy method to render numeric data, allowing a user to better understand and interpret
the information. Over half the neurons in a human brain are devoted to processing optical
information. People have a natural ability for recognizing and interpreting visual patterns
and visualization comes naturally to users (Buttenfield 1996).
Another advantage of GIS is that users can incorporate existing regional data with
site-specific data. GIS projects are layered in a manner similar to CAD drawings. The
layers, often referred to as coverages, contain the graphic images (points, lines, polygons,
pixels) that are mathematically linked to attributes. Many state and federal environmental
agencies maintain Internet sites containing public-domain coverages, satellite images, aerial
photos and other data that can be incorporated into a GIS project. These data may cover
large regions but often contain details that are useful on a site-by-site basis. Information
such as soil type, geology, elevation, topography, surrounding land use and surface
hydrology is useful when planning a sampling event.
As opposed to hardcopy, site-specific maps and drawings, GIS coverages are
already in digital format and usually projected in a geographic coordinate system, either
latitude/longitude or a Cartesian system (e.g. state plane feet or Universe Transverse
Mercador). Digitized site-specific maps and data can be projected into the same coordinate
system and become an integral part of these coverages.
In addition to allowing individuals to obtain valuable data from a multitude of
sources, including governments, universities and private vendors, the Intemet also serves as
a conduit for information between users. E-mail allows direct transfer of digital data
between private parties. Files containing images and drawings are often too large to be
BROKAW ON MANAGEMENTOF FIELD DATA 301
copied on a disk and Intemet transfers have become the most efficient way to transfer the
files.
Once sufficient information has been collected and organized, a virtual concept of a
site can be generated and then used for planning purposes. This may be as simple as
selecting the location for a few soil samples or as complex as generating extensive regional
data for a comprehensive risk assessment.
Acquisition
What ultimately happens in the field is the least predictable element of a data
acquisition project. This lack of predictability is the strongest argument for advanced
planning. During the planning phase, when the data requirements are identified, the data
acquisition methodologies are selected for use in the field.
Field instruments associated with Geomatics include dataloggers, portable
computers, GPS receivers and cellular phones. Measurement and sample collection tools
vary widely but all share the attribute of generating data. Geomatics provides the link
between the measurement and management of data and thus can reduce the lag time
between the generation and use of these data.
During the planning phase, simple considerations should be taken into account, such
as creating ways to make written field notes compatible with the records and fields of the
project's DBMS. Forms or checklists can be used to assure that all pertinent data are
properly recorded.
Dataloggers and portable computers can be used in the field to add information
directly into the DBMS as it is generated. Numerous database tables are often generated for
various aspects of a project and can be integrated using the simple relational database
techniques such as keeping at least one field in common between various database sets. For
example, a table (database set) used to record sample locations and descriptions can share
the sample name field that is used for the analytical results report database.
Global positioning systems (GPS) receive satellite-generated signals and record this
information in a database. This database is usually formatted prior to collecting the satellite
signals. Attributes (i.e. labels or descriptions) such as site name and the name of the GPS
point can be entered into the database prior to using the GPS receivers in the field. The
database fields for the GPS coordinate points can be kept consistent with the fields used in
other database sets (e.g. well sample results). When the positional coordinates are then
established, the GPS database can be linked to other field databases allowing use of data as
they are generated.
GIS further integrates the positional data with the site data. GIS can create graphic
images from the numeric values in the site-specific databases (i.e. model the data), merge
this information with data generated during the planning phase and better illustrate site
conditions for field crews. This can give field crews the ability to redesign sampling
strategies in the field as new information is collected (Carver et al. 1996). Trends and
patterns may emerge that were not anticipated or were not evident until the data could be
visualized.
There are two basic forms of GIS, raster and vector. Both can be used by field
crews for on-site analysis of data. In raster GIS, data and other attributes are assigned to
cells, or pixels. Each point on the computer screen can have a host of information attached
to it. Raster is especially useful for cell-based modeling, such as contouring contaminant
isopleths and can be used to take advantage of photogrammetry information such as satellite
or orthophoto images. In vector GIS the data and other attributes are assigned to points,
lines and polygons.
Much of the GIS data available from government agencies are in vector GIS format.
Vector-based GIS allows a user to integrate site CAD drawings and regional GIS coverages
with the data that are generated in the field. GIS software programs are increasingly
combining both vector and raster capabilities into the same program, giving users the
advantages provided by both systems.
Although having database, GPS and GIS in the field can add new dimensions to the
data gathering capabilities, there may simply not be enough time in a typical field day to
both collect and analyze the data as they are generated. However, with the use of cellular
phones and the Interact, field capabilities can be extended through an electronic link, a two-
way link. Data and images can be transmitted via E-mail as they are generated and then
analyzed and interpreted by others. The results of the analyses can be E-mailed back or
even posted on a homepage and accessed for use by the field crews.
Geomatics can dramatically shorten the time interval between the generation and the
use of data and create a new paradigm (Huxhold and Levensohn 1995), a conceptual
framework around which a project is implemented. It adds a new element and requires new
skills. It requires field crews and project managers to think differently about the way they
implement their projects.
Even if field personnel do not use Geomatics tools themselves, they should be
trained in the fundamentals of Geomatics for two reasons. One, they will understand why
they are being required to implement new and perhaps unfamiliar procedures in the field
and two, they can contribute their expertise to the process. What may seem logical and
simple in the office may actually be ridiculously impractical to implement in the real world.
One of the objectives of using Geomatics in the field is to collect and store the data (as
generated) for the benefit of all users. The more field personnel understand the process, the
better they will be in making the effort successful.
Example Project
A recent ground water assessment project conducted in New Jersey illustrates how
even a simple project can be accelerated by incorporating GPS, GIS and the Intemet into the
data acquisition process.
Buyers were considering the purchase of a four-acre light industrial site. However a
previous Phase I assessment indicated potential environmental impacts. A contract closing
was scheduled for the next week, but there were no data on the ground water conditions at
the site. Samples needed to be collected, but there was not sufficient time to install and
sample monitoring wells. Furthermore, the site was in full operation and access was
BROKAW ON MANAGEMENTOF FIELD DATA 303
limited. The parties involved agreed to use temporary wells to obtain ground water samples
for analysis.
While waiting for the utility markout, available GIS data were gathered for the site.
Soil, geology, surface hydrology and land use coverages were obtained through the Interact
and from NJDEP. Digital topographic maps and aerial photos were obtained through the
USGS homepage and commercial vendors.
These layers of information, plus a CAD-generated outline of the site, were
compiled in a GIS using commercially available software and a virtual model of the site and
surrounding area was created. Seven temporary well point locations were selected and a
simple DBMS created. Using GPS, the precise state plane coordinates and elevations were
determined for each of the seven points on the site, establishing an x, y, z coordinate system
for the site. These coordinates were added to the DBMS.
When the utility markout was complete, a direct-push soil boring rig was used to
insert seven temporary well points during the morning. The depths to water and field
parameters (pH, conductivity, and headspace) were measured in each well point and the
data entered into a datalogger. Then, ground water samples were collected for analysis by
an offsite laboratory. All of these attributes were added to the DBMS and a profile of the
site's ground water was established. Ground water elevation contours and direction of flow
were determined and correlated with headspace and conductivity data.
Based on these findings, arrangements were made with the laboratory to expedite
certain samples. The following morning, while the lab worked on the samples, a second
round of headspace analyses, field parameters and water level measurements was taken and
logged. Because the data from each round of measurements were compiled into a DBMS
and linked to GIS maps, the data from the two rounds of measurements could be readily
compared, both in tabular form and visually.
By that afternoon, when the temporary wells were being abandoned, the preliminary
laboratory data were available. Instead of faxing analytical sheets to the office, the lab
simply downloaded the pertinent data from their laboratory information management
system (LIMS) into a database file and E-mailed them to the project manager.
Using a cellular phone and a modem, the project manager (who was still in the field)
accessed his mailbox, downloaded the file and added the data to the existing DBMS.
During the road trip back to the office, various maps were generated using GIS that
illustrated ground water conditions at the site.
The next day (one day before contract signing), the maps were plotted and tables
were generated using pertinent data from the DBMS. This visualization and tabulation of
the data allowed for a rapid peer review process. A senior hydro-geo engineer was able to
examine all of the project data, including the GIS coverages and aerials obtained before the
fieldwork, within a single software program. The maps were converted to bitmaps and
transmitted to the clients and their attorneys for review and discussion.
Upon viewing the information, the clients were able to quickly see site conditions
and gained a better understanding of the risks they were facing. The site was not pristine
but it also was not severely contaminated. The contract was signed.
If improperly implemented, a Geomatics program can be a nightmare and a waste of

energy. The wrong sottware, inadequate hardware or lack of training can result in time-
consuming mistakes. Moving data from paper into digital format can potentially be a
daunting task involving a considerable cost (DeFina et al. 1997).
Conversely, a properly implemented Geomatics program can provide a significant
competitive advantage. It can accelerate the organization, review and statistical
manipulation of spatially related information and enhance the decision making process
(DeFina et al. 1997).
Training is a critical element of a successful program. Personnel need to become
familiar with and understand the fundamentals of Geomatics in order to properly integrate it
with existing methodologies.
A USEPA study (1994) concluded that typical Geomatics software programs
become more cost effective with use. The magnitude and number of projects the software
was applied to greatly influenced the cost and effectiveness.
When field-based Geomatics was used for a remote mapping of a proposed national
park in Siberia, a team of researchers gave Geomatics a tree acid test. They identified a
series of advantages and disadvantages which are itemized below (Carver et al.. 1996).
Advantages:
1. Interactive development of sampling strategy through visualization and
feedback.
2. On-the-spot environmental modeling and feedback through field-based
verification.
3. Greater appreciation of the problem in the context of the processes operating.
4. Improved confidence in the data and greater user awareness of its limitations.
5. Reduction on the number of field visits and associated project costs.
6. Ability to integrate local ideas and knowledge at the start of the analyses.
7. Increased ability to convince local decision makers of the approach adopted.
Disadvantages:
1. Logistics problems associated with power sources and transportation/protection
of equipment in the field.
2. Lack of technical backup facilities/services.
3. Problems of data availability, sensitivity, and security in some countries/regions.
4. Education and training of field teams in GIS concepts and techniques.
5. Long fieldwork preparation lead times.
6. It is very hard work.
While Siberia is more remote than most areas, the advantages and disadvantages
itemized above are applicable to many field situations. Ideally, the disadvantages are offset
by the advantages.
Geomatics has the potential of producing an enormous amount of additional data for
BROKAW ON MANAGEMENT OF FIELD DATA 305
a particular project and at the same time provides an organizational structure for this
information. Information can be reeled in from a wide range of sources but these sources
will have varying degrees of accuracy and relevance. As data are acquired, the metadata
(the data about the data used in a Geomatics project) need to be kept current. Adequately
recording and maintaining a metadata requires a higher level of discipline and additional
effort above and beyond the scope of the project itself.
Geomatics allows the user to take further advantage of the abilities of computers for
data acquisition, data management and analyses. Field and laboratory data can be rapidly
queried and visually displayed to test hypotheses or to illustrate site conditions. What used
to take hours or days for an individual to do with separate software programs can be done in
seconds with an integrated Geomatics system, assuming proper protocols are established
and followed.
Summary
Most environmental scientists and engineers generating data in the field already use
many components of Geomatics such as databases, digital maps and E-mail. These
components often are merged into GIS in the office where the data are ultimately stored and
used. Implementing Geomatics in the field simply extends these capabilities and is an
example of the evolution of information management.
Geomatics integrates a myriad of tools.
1. Intemet transfer protocols such as http (the world wide web) and file transfer
protocol (tip) for acquiring and communicating data.
2. Raster and vector generated images such as orthophotos and CAD drawings.
3. Relational databases, spatial database engines and structured query language
(SQL).
4. Satellite photogrammetry and global positioning systems.
5. GIS topology, which creates the mathematical relationships between graphic
spatial elements and associated attributes.
6. Printers and plotters that produce visual representations that can be intuitively
understood.
The most basic Geomatics tools are a desktop computer, a modem with Interact
access, plus database and GIS software. CAD software, GPS equipment, cellular phones,
dataloggers and portable computers help to complete the picture but are not essential,
especially in the initial stages of implementing Geomatics.
However, computer skills are essential. One needs to possess a basic understanding
of computers and software systems in order to apply these tools. The learning curve may be
steep in certain areas but can result in a higher level of attainment.
As these technologies develop, they become more seamless and easier to use. Word
processing software can now produce Intemet compatible documents with a click of a
mouse. CAD programs are increasingly adding GIS capabilities and object oriented
programming, such as Visual Basics is greatly simplifying the use of database programs.
By adopting a global, inclusive approach with respect to personnel, these tools can
be utilized by individuals at all levels of expertise. While GIS and database experts are
useful in helping to set up systems, they may not at all understand how the data are
generated and how it is ultimately used by decision makers. Whatever systems are set up,
they should be designed with maximum flexibility to incorporate improvements. Those
using the systems should be encouraged to provide input improve them and those who run
and maintain these systems should constantly seek out this advise and offer solutions and
new innovations.
The overall objective is to increase efficiency. Increasing efficiency means
accomplishing more with less energy.
References
Buttenfield, B., "Scientific Visualization for Environmental Modeling: Interactive and

Proactive Graphics". GIS and Environmental Modeling: Progress and Research
Issues, Fort Collins, CO, GIS World Books, 1996, pp. 463-467.
Carver, S., Heywood, I., Cornelius, S., and Sear, D., "Evaluating Field-Based GIS for
Environmental Characterization, Modeling and Decision Support". GIS and
Environmental Modeling: Progress and Research Issues, Fort Collins, CO, GIS
World Books, 1996, pp. 43-47.
Date, C.J., An Introduction to Database Systems. USA, Addison-Wesley Publishing

Company, 1995, p. 2.
DeFina, J., Maitin, I., and Gray, A., "Site-wide collection of remediation data in support of
environmental quality objectives", http.'//www.earthsofl.com/articles, 1997.
Douglas, W. J., Environmental GIS: Applications to Industrial Facilities, Boca Raton FL,
CRC Press Inc. 1995, p. 41.
Huxhold, W.E. and A. Levinsohn, Managing Geographic Information System Projects.

New York. Oxford University Press, 1995, pp. 5, 31.
Madry, S., The Fundamentals of Geomatics, unpublished draft, 1996, p. 81.
Ministry of Natural Resources of Canada, Intemet Homepage, http.'//www.geo.NRCan.gc/

geomatics/htmle/gen-gOl.html, 1997.
United States Environmental Protection Agency, GIS/Key TM Environmental Data

Management System: Innovative Technology Evaluation Report, Cincinnati OH,
USEPA, 1994, p. 107.
University of Florida Intemet Homepage, http://www.surv.ufl.edu:443, 1997.
David T. Hansen 1
Documentation of Data Collected from Field Instrumentation for Spatial

Applications
REFERENCE: Hansen, D. T., "Documentation of Data Collected from Field

Instrumentation for Spatial Applications," Field Instrumentation for Soil and Rock,
ASTM STP 1358, G.N. Durham and W.A. Marr, Eds., American Society for Testing and
ABSTRACT: Data from field instruments are increasingly being incorporated into
Geographic Information Systems (GIS). This adds a geospatial component to time series
data and permits analysis and display with other geospatial data. From the perspective of
GIS analysis, two issues are of importance. The first is the identification and description
of the instrument location. The second is the description of the data stream from the
sensor package and the access to that data for analysis and display. They are the basis for
effective and proper use of the data from the instrumented site.
The spatial location of the sensor or sensor array in latitude and longitude or in a map
projection system is a critical parameter for the use of the data in GIS. The accuracy of
the spatial locations of sensors is important in identifying the spatial analysis that is
appropriate for the data. Besides geographic coordinate information, local coordinate
information available for the site is valuable information for locating the sites and
instnmaents. Local coordinate information often is more accurately known than the
geographic position. Local coordinate information can be used in computer modeling of
1Soil Scientist/GIS Specialist, U.S. Bureau of Reclamation, Mid Pacific Region, 2800
Cottage Way, Sacramento, CA 95825-1898.
307
the sensor data independently and can be used to refine the initial values of geographic
coordinates.
To effectively use data captured from the instruments, information on the structure and
contents of the data base is required. Sensor data may then be carried into the GIS data
base management system (DBMS) or accessed from the DBMS set up for the time series
data. For effective use and application of the data, all require information on the
parameters being captured. For GIS data, this information is described as elements in
ASTM D 5714-95 Content Specifications for Geospatial Metadata.
KEYWORDS: geospatial, metadata, site investigation, data base management system,

datum, field instrumentation, time series data
Geographic information systems (GIS) are valuable tools for displaying and
analyzing time series data. This includes data collected via field instrumentation. In GIS,
the location of the field instruments is part of a data base that also has access to the
parameters being capture by the instrument sensors. This location or geographic position
permits display, manipulation and spatial analysis of the data. When site location is
carried into the same coordinate system as other data sets, information from these other
data sets can be incorporated into the analysis. This includes construction of cross
sections, computer visualization, and surface analysis of the data values.
Spatial location includes horizontal position and vertical elevation (or depth) in a
defined spatial reference system. The spatial reference system may either be a local
system or a global system. In a local system, the instrument location is identified based
on measurements relative to other features at the site. Usually, horizontal location is
identified globally as the instrumented site's latitude and longitude. These spherical
coordinates are converted into a planar map projection system for analysis with other
data. The vertical position (elevation or depth) of the instrtmaent is identified with respect
to a local datum for the site or to an established vertical datum.
Instrument position is one set of parameters from what is often a complex data set
captured from field instruments. Data from an instrumented site may represent several
physical and chemical parameters which are collected continuously or at intervals. The
data base management system (DBMS) must be able to manage this data stream and
usually has tools for the manipulation, analysis and reporting these data. GIS is a data
base system for the management, analysis, and display of geographic information. It
operates on spatial entities representing location and attributes associated with those
HANSEN ON DOCUMENTATIONOF DATA 309
entities. Often in practice, the DBMS of the GIS system is linked to a separate DBMS
storing data captured from the instrumented site. Recognizing the structure and
definitions of fields in each DBMS improves the ability to access data for query,
analysis, and display.
Information on the data captured and stored from the field instruments and
coordinate systems identifying the instrument location is metadata. Several standards are
in effect which identify metadata elements for describing these characteristics. Metadata
describing GIS data sets have been identified in ASTM Specifications for Content of
Digital Geospatial Metadata (D 5714). The elements identified in that guide provide
detailed descriptions of the spatial representation of the instrument location and data
captured from the instruments. Other guides identify elements that should be considered
in site selection for instrumentation, description of the site, or description of the
parameters being measured. These include the following:
9 ASTM Guide for Site Characterization for Environmental Purposes with
Emphasis on Soil, Rock, the Vadose Zone and Ground Water (D 5730),
9 ASTM Practice for a Minimum Set of Data Elements to Describe a Soil Sampling
Site (D 5911),
9 ASTM Practice for the Minimum Set of Data Elements to Identify a Ground
Water Site (D 5254),
9 ASTM Guide for Selection of Data Elements for Ground-Water Investigations (D
5474),
9 ASTM Guide for the Set of Data Elements to Describe a Ground-Water Site, Part
1 - Additional Identification Descriptors (D 5408),
2 - Physical Descriptors (D 5409),
3 - Usage Descriptors (D 5410).
Many elements in these guides are complimentary with D 5714. In addition, ASTM
Guide for Continual On-Line Monitoring Systems for Water Quality Analysis (D 3864)
identifies items that should be considered in the installation of field instnmaents. While
the focus of this paper is on information needed for the management and application of
time series data within a GIS environment, these standards need to be reviewed with
respect to field instrumentation for soil and rock.
The Spatial Dimension
In GIS, the instrumented site is represented as a spatial feature. The feature may
be represented either in vector or raster format. The format for representing the site or
instrument is dependent on the types of spatial analysis that are to be performed. Spatial
location including horizontal and vertical position are attributes stored and maintained by
GIS. Figure 1 shows sampling locations represented as points in vector format for a
portion of the San Francisco Bay and the Sacramento - San Joaquin Delta Estuary.
Monitoring at these locations are coordinated through the Interagency Ecological
Program (IEP). They represent a variety of sampling conditions from trawling surveys to
instrumented sites for a variety of chemical, physical, and biological parameters. IEP
provides summaries of this data at Web site, wwwiep.ca.gov. Representation of the sites
as points is the simplest implementation of time series data in GIS where site location
alone is shown. These locations are displayed with other spatial data showing San
Francisco Bay and the Sacramento - San Joaquin Delta along with the global coordinate
system of latitude and longitude.
FIG 1--Site Locations in San Francisco Bay and the Sacramento - San Joaquin Delta
Estuary
Coordinate Systems
Location in GIS is carried in defined coordinate systems with horizontal and

vertical components. A common coordinate system for the spatial representation of
features in GIS permits display and spatial analysis of different data sets as shown in
Figure 1. The parameters of this coordinate system and the accuracy in the placement of
the sample sites in the coordinate system are of major importance for successful operation
of GIS analysis. These parameters include the definition of the datums or reference points
for the coordinate system and other characteristics of the coordinate system. Control for
these coordinate systems is based on networks of control points at the local, regional, and
national level. Control systems at the national level are coordinated and reported by the
National Geodetic Survey (NGS).
Horizontal Position-- For horizontal position, latitude and longitude are the
recommended coordinates to report the position of the instrumented site location on the
surface of the earth. It is a universally recognized system for determining positions on the
globe. This is the recommended practice in D 5911 for identifying a soil sampling site (D
and D 5254 for identifying a ground-water site. This global coordinate system is
dependent on the model of the ellipsoid that is used. For the United States, there are two
different ellipsoid models in active use. These are either the Clarke 1866 model or the
WGS 1984 model. These ellipsoid models are the basis for the North American Datum of
1927 and the North American Datum of 1984. When coordinates of latitude and longitude
are reported, the datum on which those values are based is a key parameter to be
identified.
Many maps in active use in the United States such as the topographic series
produced by the United States Geological Survey (USGS) until very recently were
produced with map projections based on the 1927 datum. Positions captured with Global
Positioning Systems (GPS) are frequently based on the 1984 datum. The datums are
based on different spheroids of the earth and affect the values reported for latitude and
longitude across the United States. In Delta area of California, identification of the
incorrect datum for values of latitude and longitude will displace the site by several
hundred meters. GIS data for an area frequently includes horizontal coordinates based on
both systems. For proper display and analysis of data in GIS, the underlying datums for
each of the data sets must be identified and converted to a common datum. Peter Dana of
the University of Texas has a Web page describing geodetic datums.
In practice, the locations of instrumented sites are usually carried in a planar map
coordinate system rather than latitude and longitude. This permits the direct measurement
of distance, area, or direction. However, when representing a spherical surface in a planar
system not all properties of distance, direction, area, or shape of features can be preserved
throughout the geographic extent of a project. For a GIS project, a map projection system
selected for a project area is determined by which properties are of most importance for
that project area. Data to be used in that project is then transformed into the common map
projection. Peter Dana has Web pages describing map projection systems. Snyder and
Snyder and Voxland provide comprehensive descriptions of the parameters and effects of
a variety of map projections.
Vertical Position-- The vertical position (depth, height, altitude, or elevation) of

the instrument location is generally a major characteristic for the application and use of
312 FIELD INSTRUMENTATIONFOR SOIL AND ROCK
the data. As identified in D5911, D5254, and D5714, the reported vertical position is
dependent on the vertical datum on which the measurement is based. The reported value
is often assumed to be reported as elevation with respect to mean sea level. Mean sea
level is based on measurements at regional tide gage stations and in this sense it is a local
or regional datum. Besides local and regional vertical datums, two vertical datums are in
common use in North America and the Delta area of California. In 1929, the National
American Vertical Datum (NAVD29) set a fixed datum for the United States. This is the
datum used for most USGS 7.5 minute topographic sheets. Since mean sea level actually
varies over the Earth, the North American Vertical Datum (NAVD88) was developed
based on adjustments back to mean sea level at Father Point, Canada in 1988. For the
Delta area of central California, there is about a one meter difference in elevations based
on these two datums. In addition to differences in elevations based on these two datums,
the Delta area is subject to active subsidence. Subsidence in the Delta region is primarily
due to compaction, erosion, and decomposition of organic soils from agricultural
development over the past 100 years. Other areas of California as well as the United
States are subject to active vertical shifts from seismic events or ground water
withdrawal.
In the fall of 1997, several agencies cooperated in improving the network of
stations set for NAVD88 by the National Geodetic Survey (NGS). This project set
stations identifying elevation based on NAVD88. Global positioning systems (GPS) were
used to adjust and refine the network for vertical elevations with less than a 2 cm
difference in elevation across the network. At the same time, it improved the network of
horizontal control within the region. Unpublished presentations by Don D'Onofrio of
NGS, Marti Ikehara of USGS and Monte Lorenz of the United States Bureau of
Reclamation (USBR), indicate a difference of less than 0.5 cm to more than 30 cm in
height form elevations shown on the USGS 7.5 minute map series after they had been
converted to NAVD88.
These differences in elevation (or depth) due to local or national datums and
changes in surface elevations from subsidence have major implications for surface
modeling and water modeling in the Delta area. Existing elevations and measurements
from stream gages and tidal gages, as well as ground elevations need to corrected to a
common datum. Development of this control network by NGS, USGS, and USBR
provide the control for adjusting existing GIS data based on other datums to the NAD83
and NAVD88 datums.
GIS Coordinate Accuracy--Where site positions can be clearly located on the

USGS 7.5 minute map series, locations have been expected to be within about 10 to 20
meters in the horizontal position and 1 to 10 meters in the vertical position. This map
series is constructed to meet National Map Accuracy Standards and these are the errors
expected for well defined locations at that map scale. Increasingly, global positioning
systems (GPS) or similar devices are being used to directly record the position of sites.
Currently with GPS units, the horizontal position of the site can be captured more
accurately in a shorter period of time than the vertical position. Often software in the GPS
unit permit the reporting of the location either in latitude and longitude or a map
projection. Presently, GPS can rapidly report horizontal positions within 100 meters and
with survey grade equipment and proper procedures within the centimeter level. At
present, the vertical position reported by a GPS unit will not be as accurate as the
horizontal position unless it can be tied into a NAVD88 station. The National Geodetic
Survey (NGS) provides access to geodetic control information for both horizontal and
vertical control at the Web site www.ngs.noaa.gov and provides data sheets on known
NGS control at http://sinbad.ngs.noaa.gov/FORMS/ds_area.html.
The horizontal and vertical coordinate systems o f a GIS data set are reported as
part of the metadata in the spatial reference information section of D5714. This
information includes parameters defining these coordinate systems such as map
projections and datums. The resolution of the coordinate values and the accuracy of the
coordinate values are also to be reported.
Local Measurement Systems-- At the time of instrument installation the relative

positions of instrument have often been located by survey or measurement relative to
local features. This relative position is often more accurate than the coordinates that are
carried in GIS and may be very important for the analysis of the data collected by the
instruments. D 5730 provides guidance on locating site locations. D5474 identifies
elements for further definition of the site location by a map and description on the
position of the ground-water monitoring. These relative positions for the local site
configuration should be recognized for the site. This is particularly true for horizontal
position where instruments are placed for sampling particular depths in the water column
or particular stratums. These positions can be carried into GIS for spatial modeling and
visualization independently of other data that is in a map projection. These local
measurements can also be used to adjust GIS stored coordinates where the local
coordinate system can be tied to a survey control station with known geographic
coordinates. Elements for identifying and describing local coordinate systems for the
position of GIS spatial entities are identified in D 5714.
The Data Base Dimension
Attributes associated with spatial features in GIS provide access to data that will
be used in any spatial analysis. Depending on the structure defined for the GIS data base,
the spatial feature may identify individual instruments or clusters of instruments that are
reporting a variety of physical and chemical parameters. Basic attributes will identify the
instrumented location and instruments or clusters of instruments being represented.
Attributes will include a unique identifier for linking to the data base structure storing the
parameters or summary values captured from the instruments. Unique identifiers provide
the link or key fields for accessing the data for individual parameters captured from the
instruments. Data captured from the instruments are usually stored in separate tables in a
data base structure designed specifically for time series data. Typically, this is in date
stamp format.
To use GIS as a tool in the display and development of models utilizing time
series data, an understanding of the data structures and DBMS is required. This is
particularly true where GIS is being used as a tool to visualize or model values from the
data base. Of particular importance are:
9 Table Names - Identification of tables carrying time series data or data summaries
from the instrumented sites.
9 Key Fields - Field that carry the unique identifiers providing access to data for the
site, instrument, and parameter values. These fields serve as the basis for relates
or joins between tables.
9 Date and Time Representation - Time stamp field and other information on:
Time period represented by sensor data.
Frequency of measurement.
9 Attributes - Additional fields containing date and time, measured values or
summaries. This includes attributes that will be used for queries, analysis and
display. The attribute description includes:
Data type
Field width
Domain of valid values
Units of measurement
Measurement resolution or precision of reported values
Special values or flags for:
No data conditions,
Censored data values,
Calculated values,
Other special conditions
This information provides the basis for query, analysis, and display of the data captured
from the instrumented site.
With access to the time series data, GIS can be used to query and display the
results of analysis. Figure 2 shows a portion of IEP data for the Chipps Island site in the
Delta. Chipps Island has been a key point for monitoring water quality conditions in the
Delta. The graphs have both average monthly and average daily salinity values for the
months of February and June between 1958 and 1994. Winter months are typically
periods of high river flows as the result of seasonal storms from the Pacific Ocean and
salinity levels generally remain low. The winters of 1976, 1977, 1989, 1990, and 1991
represent drought periods with low volumes of run off in the Sacramento and San Joaquin
River systems. Low river flows and corresponding higher average salinity levels are
typical during the summer months. The variation in average daily values are due to tidal
flows from San Francisco Bay.
FIG 2 -- Salinity Values at the Chipps Island Site in the Delta.
Development of this display is based on information from IEP describing the data
files, fields containing the recorded data, and special values. For GIS data, D5714
identifies this information as elements of the entity and attribute section of the metadata
reported for a GIS data set. Date and time information is expressed in the format
YYYYMMDD and HHMMSSSS as identified in ANSI standards for time dependent
information. The four character requirement for year avoids potential confusion between
the year 2000 and 1900 for computer processing. The time element includes
specifications to identify local time and its relationship to Universal Time (Greenwich
Mean Time). In addition, D5714 identifies a data quality assessment for the attributes
that are carried into a GIS system. For ground-water monitoring sites, D5714
complement elements in D5474 for identifying the types of data, frequency of sampling,
and data quality. Similar elements are also identified in D3864 for on-line monitoring of
water treatment systems.
Elements in D5714 represent only information for documenting and managing
data actually used and accessed in GIS. It can identify the purpose and limitations for the
digital representation of data from the instrumented site. Other guides such as D5474,
D5730, D3864, D5409, and D5410 identify considerations for the design and installation
of instrument systems. ASTM Guide for Continual On-Line Monitoring Systems for
Water Quality Analysis (D 3864) is focused on monitoring systems installed in a
laboratory or for water system control rather than field installation. It identifies items that
should be considered such as instrument calibration, validation of instrument reported
values, and review of outlier or exceptional values reported by the instrument. It also
includes guidance on reporting instrument precision. D5474 identifies the elements that
should be reported for ground-water monitoring sites. It complements D5409 which
identifies the elements for describing the measurement of physical parameters at a site
and D5410 which identifies the elements for describing the measurement of chemical and
other parameters at a site. In addition to these guides, Gordon and Katzenbach (1983)
provide guidance on the installation of instruments for water quality assessment.
Summary
The data base is the central focus for query and display of time series data from
instrumented sites. The location of the instrument is one set of characteristics which may
be represented in GIS in vector or raster format for spatial display, query, and analysis.
When utilizing GIS as a tool, it is important to recognize that a specialized DBMS may
be required with special tools for maintaining and for analysis of time series data. The
GIS component of any overall system is used to store and maintain the spatial location for
the site. As part of the GIS data set representing the site, GIS attribute tables provide
access to the records of the time series data for query, analysis, and display in GIS.
The universal system used to identify the location of the instrumented site are
latitude and longitude. In actual GIS display and analysis, a Cartesian coordinate system
is used in a defined map projection. This permits the display and analysis of data from the
instrumented site with other GIS data layers. The altitude or elevation of the site is an
additional attribute locating the position of the site. For effective use of these coordinates
for the site, the datums used for latitude and longitude and for elevation need to be
identified. Where the instrument location is carried into a map projection system, the
parameters which define the map projection system need to be identified. These are
required elements in D5714 which identifies metadata for describing GIS data sets. The
local coordinate systems for site instrument position or local measurements of instrument
location are important to maintain. As national and regional survey control networks
improve, the local coordinates can be tied into map projection systems with other GIS
data. The local coordinate values can be carried into GIS and used in modeling the data
independently of other GIS data.
For effective use in GIS additional information is required on the data structure
and format of the time series data. This metadata is identified in the entity and attribute
section of D5714. This information identifies the tables and fields representing the time
series data. It includes information on the period of record of the data, the frequency of
data values, units of measure, and resolution of the reported data values. This metadata
may have application in describing the DBMS independent of the GIS data set. It should
be reviewed for application to time series data. D5714 complements other standard
guides for site characterization, describing a soil sampling site, and for describing a
ground-water site. These guides and the standard guide on continual on-line monitoring
systems for water analysis contain additional information that should be considered in
field instrumentation.
Gordon, A. Brice, and Max Katzenbach, 1983, Guidelines for Use of Water Quality
Monitors, U.S. Geological Survey, Open File Report 83-081
Snyder, John P., 1983, Map Projections used by the U.S. Geological Survey, Bulletin
1532. Second Edition, Washington D.C.: U.S. Government Printing Office.
Snyder, John P., 1987, Map Projections - A Working Manual, U.S. Geological Survey
Professional Paper 1453, Washington D.C.: U.S. Government Printing Office.
Snyder, John P.,and Philip M. Voxland, 1989, An Album of Map Projections, U.S.
Geological Survey Professional Paper bulletin 1453, Washington D.C.: U.S.
Government Printing Office.
On Line:
Dana, Peter H., "Geodetic Datum Overview", University of Texas Department of

Geography, http ://www.utexas, edu/depts/grg/gcrafi/notes/datum/datum, html
(October 14, 1997) - Description of datums.
Dana, Peter H., "Global Positioning Systems Overview", University of Texas,

http://www.utexas.edu/depts/grg/gcraft/notes/gps/gps.html
(February 15, 1998) - Description of GPS.
Dana, Peter H., "Map Projections Overview", University of Texas,

http ://www. utexas, edu/depts/gr g/gcr afi/notes/mapproj_ftoc, html (October 14,
1997) - Description of map projections.
IEP "Interagency Ecological Program Home Page", http://wwwiep.water.ca.gov!

(Accessed: February 2, 1998) - Interagency Ecological Program for the Delta
Estuary in Calfomia. Provides access to data files on reported a variety of
physical, chemical, and biological parameters collected by the agencies.
NGS, "Data Sheet by Area", http.'//sinbad.ngs.noaa.gov/FORMS/ds_area.html

(Accessed: February 2, 1998) - National Geodetic Survey data sheets of control
point information for the United States.
M. Hawkes, 1 and W. Allen Marr2
Data Acquisition and Management for Geotechnical Instrumentation on the Central

Artery/Tunnel Project
Reference: Hawkes, M. and Marr, W. Allen "Data Acquisition and Management for
Geotechnical Instrumentation on the Central Artery/Tunnel Project," Field
Instrumentation for Soil and Rock, ASTM STP 1358, G. Durham and W.A. Marr, Eds.,
Abstract: Boston's Central Artery/Tunnel Project has implemented a unique method of

subcontracting for monitoring geotechnical instrumentation. The M026J contract includes
a provision for the collection of 1.3 million readings over a six-year period. The contract
uses a fixed unit cost per reading. This paper describes some aspects of the data
acquisition and management processes that handle the flow of data for the project.
Keywords: instrumentation, data management, databases
Introduction
The Central Artery/Tunnel Project in Boston is one of the largest, most complex
and technologically challenging highway projects in American history. The project
design reduces traffic congestion and improves mobility in one of America's oldest and
most congested major cities, improve the environment and support economic growth for
New Englanders. The project has two major components:
An eight-to-ten-lane underground expressway replaces an existing six-lane

elevated highway. The new expressway is directly beneath the existing road,
culminating at its northern limit in a 14-lane, two-bridge crossing of the Charles
River. When the underground highway is finished, the crumbling elevated road
will be demolished.
1-90 (the Massachusetts Turnpike) will be extended from its current terminus
south of downtown Boston through a tunnel beneath South Boston and Boston
Harbor to Logan Airport. The first link in this new connection - the four-lane Ted
Williams Tunnel under the harbor - was finished in December 1995.
t Manager and Geotechnical Instrumentation Engineer, GEOCOMP-Brown, IV - CA/T Instrumentation

Monitoring Contract M026J - 421 Dorchester Ave., South Boston, MA 02127,
2 Chief Engineer, GEOCOMP Corporation, 1145 Massachusetts Ave., Boxborough, MA 01719.
To put this highway improvement in the ground in a city like Boston amounts to
one of the largest, most technically difficult and environmentally challenging
infrastructure projects ever undertaken in the United States. The Project spans 7.5 miles
of highway, 160 lane miles in all, about half in tunnels. The Project will place 3.8 million
cubic yards of concrete - the equivalent of 2,350 acres, one foot thick - and excavate 13
million cubic yards of soil. The larger of the two Charles River bridges, a ten-lane cable-
stayed bridge, will be the widest ever built and the first to use an asymmetrical design.
The Project also includes four major highway interchanges to connect the new
roadways with the existing regional highway system. At Logan Airport, a new
interchange will carry traffic between 1-90 and Route 1A as well as onto the airport road
system. In South Boston, a mostly underground interchange will carry traffic between 1-
90 and the fast-developing waterfront and convention center area. At the northern limit of
the Project, a new interchange will connect 1-93 north of the Charles River to the Tobin
Bridge, Storrow Drive, and the new underground highway.
At the southern end of the underground highway, the interchange between 1-90
and 1-93 will be completely rebuilt on six levels, two subterranean to connect with the
underground Central Artery and the Turnpike extension through South Boston. The
interchange will carry a total of 28 routes, including High Occupancy Vehicle lanes, and
channel traffic to and from Logan Airport to the east. A fifth interchange, at
Massachusetts Avenue on 1-93, will function as a part of the larger 1-90/I-93 Interchange
when the Project is finished but today is already helping improve Southeast Expressway
traffic flow following early phases of reconstruction.
The Project has been under construction since late 1991. As of November 1998,
final design is 98 percent complete, construction 46 percent complete. The next
construction milestone, a bridge across the Charles River connecting 1-93 in Charlestown
with I.everett Circle and Storrow Drive, will be reached in1999. The 1-90 extension
through South Boston to the Ted Williams Tunnel and Logan Airport will open in 2001.
The northbound lanes of the underground highway replacing the elevated Central Artery
open
in 2002, the southbound lanes shortly thereafter in 2003. The entire project will be
finished in 2004, including demolition of the elevated highway and restoration of the
surface. The final cost of the project is $10.8 billion.
The 7.5-mile alignment is generally constructed using cut-and -cover excavations.

The excavations are supported by variations of cast-in-place concrete diaphragm walls
that are tied back or braced. The alignment passes directly through the downtown area of
Boston in the same location as an existing elevated highway. Construction must occur
while maintaining existing traffic flow and without affecting a large number of adjoining
structures.
A joint venture of Bechtel and Parsons Brinkerhoff (B/PB) is the Management

Consultant (MC) to the Massachusetts Transportation Authority, the owner. B/PB is also
the Construction Manger.
HAWKES AND MARR ON ARTERY/rUNNEL PROJECT 321
The subsurface conditions along the alignment consist of glacial stratigraphy, the
most prominent strata being organic silt deposits underlain by Boston blue clay. A
variety of jet grouting, soil mixing, ground freezing and dewatering methods are being
used to improve the subsurface conditions of these strata prior to construction.
A substantial geotechnical instrumentation program was developed by the

designers to monitor the performance of constructed components and adjacent facilities.
Currently, the instrumentation system includes monitoring groundwater pressures,
deformations, loads, strains and vibrations at approximately 3000 locations. Movements
of a large number of reference points are also monitored by surveying; however these are
done by a separate contractor and are not considered in this paper.
A computerized application manages the large amount of data collected from

these instruments and supports the project's need for rapid turnover of reduced data. The
following sections present an overview of the system developed to collect, validate, store
and report the geotechnical instrumentation data for the CA/T project.
Project Requirements and Objectives

The project consists of over 40 separate design and construction contracts. To
obtain consistency in instrumentation and data management procedures across the entire
project, the Management Consultant defined the primary objectives of the
instrumentation, developed a project wide specification, and defined a means to collect
and process the data [1]. Of major concern to the project is to put in place a system that
produces reliable data from the geotechnical instrumentation, as it is needed, in the form
that it is needed.
The large size of the excavations for the project, the proximity to numerous
existing structures that are sensitive to movements, and the variable subsurface conditions
along the alignment were factors pushing the need for rapid turn around of data. To meet
the goals of the MC, readings taken on one day need to be reduced and reported by
08:00AM the next day. This is to permit all of the involved parties (engineers,
contractors, and abutters) to have access to the data within 24 hours. The MC decided to
award a single contract to monitor all geotechnical instrumentation for a six-year period.
This was to provide a consistent data management process over the entire contract for the
principal earthwork portions of the project. That contract was structured as a unit price
contract, i.e. the Contractor would be paid a set price for each successful reading on each
type of instrument. The work would require an estimated 1.3 million readings. The work
was put out for bid.
In April 1997, the joint venture of GEOCOMP Corporation (Boxborough, MA.)

and TLB Associates (Millersville, Maryland), was awarded the geotechnical
instrumentation contract [2]. The total contract amount was approximately $10,000,000
for the six-year period. This represented approximately 60% of the engineer's estimate
of the contract value.
Table 1 summarizes the types and quantities of instruments that are currently
monitored. Typical monitoring frequencies are weekly, except for dally seismograph.
Table 1 - Geotechnical Instruments
Instrument Type Quantity ~

Seismographs 45 units, 149 locations
Convergence gages 135
Crack gages 992
Dial gages 4
Inclinometers 290
Multi-position borehole extensometers 6
Multi-point heave gages 23
Observation Wells 278
Probe extensometers 165
Shear displacement gages 6
Single-point borehole extensometers 9
Tilt meters 43
Vibrating wire load cells 28
Vibrating wire piezometers 725
Vibrating wire strain gages 12
Total 2910
1
as of 06/30/98
Project Concept
GCB's approach to the job is to automate all aspects of the work that make
financial sense. The overwhelming cost for the work is labor. Consequently every aspect
of the data collection and reporting process was reviewed to determine what labor saving
measures could be applied. Since we are only paid for readings when we provide a
correct reading, errors must be kept to a minimum. On time readings are required to
avoid penalties. Therefore, we desired approaches that would help us perform our work
quickly. The project deliverables are tabulations of reduced data for each portion of the
project, tabulations and graphs of instrument data which have readings that exceed pre-
established threshold values, and placement of the raw data into the MC's project
database.
Since the data must be provided to the Client in electronic form, it is immediately
logical that data should be put into electronic form as early in the process as possible.
After reviewing the possible approaches, we decided to employ handheld industrial
computers to collect the data from each instrument. These devices could be used to
collect data from a wide variety of instruments and readout devices. They reduce the
opportunities for human errors and decrease the time required to get the data into the
Client's database.
HAWKES AND MARR ON ARTERY/TUNNEL PROJECT 323
Contractors
Designers
Abutters
Distribute Data
Project Management
Bechtel/Parsons Brinckerhoff
1
Oracle Database
Consolidate data into daily reports

and load into Project Oracle Database
using network link over a phone line
GECOMP-Brown J.V.
Field Database
[1
Transfer data to Field Database
using serial communication (RS232)
GECOMP-Brown J.V.
Personnel Data Recorder
[1
Collect and validate data
using custom sottware
Measured Parameters
strain, pressure
1
displacement, load
Figure 1 - Data Flow

DATA F L O W
Figure 1 shows a schematic of the structure of the data collection and reporting
process. The ultimate destination for the data is on the MC's computer network, and
mostly in an ORACLE database system. Seismograph and Inclinometer data files are
stored in series of electronic folders (directories) and files corresponding to construction
contracts, instrument identifications and dates.
All data from the previous 24 hours must be reported to the Client by 08:00AM.
Most data collection occurs in the daytime because of limits on access times to some
locations and difficulties getting technicians to work night shifts. The majority of the data
are collected during the construction work hours, typically from 06:00 to 16:00. As the
work got underway, it became clear that seismograph data needed to be available the
same day it was collected to permit the project staff to address complaints faster.
Seismograph data are collected every morning and reported in the early afternoon of the
same day.
Collecting Readings
A field data acquisition device (Personnel Data Recorder or PDR) collects
instrumentation readings in electronic format. The PSION Workabout [3] is the primary
data acquisition device. Figure 2 gives the specification for the Workabout.
Figure 2 - PSION Workabout Personnel Data Recorder
We chose this device as our basic data logging unit because of its ruggedness, its
small size, its ability to run for weeks on a single battery, its ability to be programmed in
a language similar to VisualBasic, and its large data storage capacity. We wrote software
for the unit to collect and validate data for all the instruments read on the project. The
entire database of instrument readings is kept on each PDR and used to check new
readings immediately when they are obtained.
Data are obtained from the instruments manually, electronically, or automatically.

Regardless of the method of collecting, the data ultimately resides in the computer data
management system. Electronic collection of data in a format that interfaces with the data
management system in the office is the most efficient method. The PDR provides a
HAWKES AND MARR ON ARTERY/TUNNELPROJECT 325
common gateway for all data into the data management'system. The PDR provides the
following features:
9 determines the anticipated data range for the current instrument, using several
methods.
9 validates the current reading against the anticipated range.
9 flags the reading to indicate which data range method was used.
9 tags the reading with the technicians id.
9 tags the date and time for the reading.
9 stores reading.
9 transfers reading to the data management system
M a n u a l l y C o l l e c t e d D a t a - Manually collected readings are entered into the PDR

using an alphanumeric-numeric keypad. Manually collected readings include readings
from tape measures, dial gages, observation wells, probe extensometers, convergence
gages, and crack gages. Manually collected readings also include a category of data that
describe the status of instruments that may not be readable. The PDR provides a selection
of typical remarks the technician can choose from without having to type the remark.
E l e c t r o n i c a l l y C o l l e c t e d D a t a - For electronically collected data, the PDR is

connected directly to the instrument readout unit. This method is used to read vibrating
wire strain gages and piezometers, inclinometers, and tiltmeters. The PDR is connected
to the readout unit using a standard 9-pin RS232 serial and programmed to communicate
with the readout unit. Electronically collected data are not subject to data entry errors,
thereby eliminating one possible source of error.
A u t o m a t i c a l l y C o l l e c t e d D a t a - Automatically collected data are collected and

stored electronically. The readout-data storage device is typically referred to as a
datalogger. The datalogger is connected to several instruments. It collects data at
specified intervals. The data can be transferred to another location by radio or telephone
links or by connecting a computer directly to the datalogger. Dataloggers are used in the
following special situations on this project:
9 collecting multiple readings from individual piezometers during pumping

tests.
9 collecting multiple readings from multiple instruments in difficult to access
locations, for example in the subway tunnels.
9 collecting "real-time" readings.
Surprisingly with the exception of Seismograph data, there are very few occasions
where dataloggers are used on the CA/T project. We examined the potential to use data
loggers on a large scale for this project. However, since we on average collect data from
a sensor once per week, the cost savings provided by using dataloggers do not cover the
cost of the added electronics.
EvaluatingReadings
An immediate requirement for the PDR is to help determine whether a reading is
good or not and whether a reading exceeds established threshold values. If the reading
for a critical instrument exceeds established threshold values, project personnel must be
notified within 15 minutes. However, an erroneous reading could exceed the threshold
value and cause an invalid alarm, It becomes important that ways be found to identify
erroneous readings as quickly as possible to minimize the number of false alarms and to
minimize the cost associated with the bad reading. Minimizing the cost is especially
important since we are not paid for an erroneous reading.
There are four possible categories for readings collected in the field:
9 Good readings that fit previous trends
9 Good readings that do not fit previous trends
9 Erroneous readings that fit previous trends
9 Erroneous readings that do not fit previous trends
The PDR can be programmed to compare the current reading with previous
trends. A reading may fit previous trends and still be in error; however this will not
trigger a false alarm. Conversely, a reading that doesn't fit previous trends may be good.
These are the important readings because they are the ones that indicate change in
performance which may be the warning of an impending failure.
Questionable or erroneous readings are due to:

9 Misread or inaccurate instrumentation reading.
9 Incorrectly keying numbers into the data acquisition device, for example
mistakenly keying a 6 as using and adjacent key such as 5, 9 or 3.
9 Misreading written data, for example reading a 5 as a 3, or a 1 as a 7.
9 Erratic or unanticipated instrumentation responses resulting from construction
activity.
Questionable readings due to construction activity can often be identified from

field observations (jet grouting, pile driving, pumping test) or discussion with the
engineers. Usually the readings are confirmed by taking an additional verification
reading. All erroneous readings, except for reproducible equipment malfunctions can be
eliminated by additional verification of the reading.
The PDR is programmed to set an acceptable data range for the current reading. If
the reading is within the established range the reading is considered as valid. Valid
readings are processed with no further quantitative checking.
Anticipated reading ranges are determined using one of the following methods:
1) plus-minus two standard deviations from the mean using the most recent eight
readings to compute mean and standard deviation.
2) maximum and minimum of the most recent eight readings.

3) predetermined anticipated value plus-minus an acceptable error.
4) some other value determined by the specific conditions for the instrument.
The data window for a reading is computed "on-the-fly" by the PDR using the
database of previous readings. The technician can switch between the different methods
for setting the range. However, the ranging method is stored with the reading so that an
engineer can further evaluate the reading. If the reading is determined to fall outside the
acceptable range, the technician is instructed to take a second reading. If the second
reading is also verified to fall outside the anticipated range, the PDR instructs the
technician what action to take. Possible actions include telephoning the resident
engineer immediately, telephoning the instrumentation superintendent, or tagging the
instrument for further evaluation.
Data range computations using the previous readings (mean 2sd, and max/min)
have proven to be not useful for instruments with a greater potential for erroneous
readings. For example probe extensometer instruments using an electric reed switch
which trigger a beeper in the presence of the magnetic reference anchor, are prone to
erroneous readings. If the data range is set using previous readings, there is a tendency
for the range to grow to an unacceptable size. Windows based on mean 2sd have
grown more than 1 foot for these instruments. This is clearly unacceptable and therefor a
data range using a predetermined reading with a range of 2/100 foot is used
exclusively for probe extensometers. An engineer further checks these data before
loading them into the MCs database.
Pore pressure instruments, such as vibrating wire piezometers, located in the zone
of influence of pumping wells, jet grouting, and pile driving can display large changes in
readings that may or may not fit the previous history. These instruments may change
readings by a 20 or 30-foot pressure head in less than twenty-four hours. Data entry
ranges for these instruments are typically 1 or 2 feet. Clearly this is a problem from the
perspective of the field technician who is usually not aware of the cause of such
apparently erratic readings. These readings are usually entered using data range based on
the specific conditions of the instrument. The reading is flagged, and the flag results in a
recheck of the reading.
ProcessingReadings
The PDR program dumps all data for the current date and selected instrument
type into a daily data file. At the end of the day each technician creates the daily dump
files and places the PDR and form listing the quantity and types of data dumps in an
inbox for further processing. The daily dump files are transferred and loaded into the
M026J Database Management system (the field database) using standard serial
communications.
During uploading from the PDR, the database automatically assigns a unique
identification integer to the reading. The identifying integer is referred to as the
ReadingId. The ReadingId's are manipulated as supersets in the database using another
integer called RepNo (report number). All data loaded or entered into the database are
assigned Readingld's and RepNo's.
The reason that report number and reading identification number are used instead
of reading dates, is that database queries using integers execute several times faster than
the same query using dates. Only for the database administrator can see or use the
Readingld's. RepNo's are used to select and process sets of readings collected on a
specified date corresponding to the RepNo.
Transmitting Data
Data is transmitted to the Project Oracle system using standard telephone lines
and modems. The data are transmitted in bulk grouped by RepNo and instrument type.
All the data for each report and instrument type are downloaded from the Field Database
into a single data file for transmitting to the Project Management. The main data
transmission is usually preformed in one session late in the afternoon. A second session is
performed the next morning to transfer all data collected during the night.
Once the data files are transmitted, they are loaded into the Project Oracle
database system, using utility programs. It is the responsibility of the Project
Management to further distribute data to Resident Engineers, Construction Contractors,
Designers and Abutters.
Documentation
Documentation is accomplished using a series of hardcopy reports generated at
each data processing stage. The first report is generated directly from the PDR daily
dump file when it is loaded into the Field Database. The second report is generated when
the Field Database downloads the data into a single file (for each instrument type) for
transmission to the Project Management's network. A third report is generated from the
Project Oracle Database detailing any problems during transfer. The final report is
generated from the Project Oracle Database containing the reduced readings. The final
report is transmitted to the Project Management with a formal transmittal letter listing the
instrument types and construction sites that are reported. The transmittal letter also
references the RepNo from which the data are processed.
All reports are placed in a single hardcopy file containing all documentation and
field operations, comments associated with the report, and the day's activities.
Scheduling and Tracking

The Field Database system contains an element to schedule and track readings.
This element was built to better respond to the Client's requests for a constantly changing
set of instruments to be read. We receive work orders each Thursday for readings to take
in the upcoming week. These work orders are then modified on a daily basis depending
on what may occur in the project. Readings are scheduled by flagging an instrument as
unread in the work order section of the Field Database system. A report number (date)
and a technician is assigned to each unread instrument. The work order system creates
reports of unread instruments that are distributed as daily assignments to the field crew.
In a typical week, 1200 instruments are assigned to 10 field technicians.
When the readings are uploaded from the PDRs, the work order status is updated.
The Report number, technician initials, and "unread" status are updated. At any time
reports of read and unread instruments can be generated. If a technician was unable to
complete his day's assignment, the unread instruments are reassigned to a new Report
number.
Reporting
Most data reports are generated from the readings after they have been loaded into
the Project Oracle Database system. Reports are generated in this manner because they
can be recreated by anyone who is granted access to the Project Oracle system. The
Oracle system is the final database destination for most of the readings. Once the
readings are loaded in Oracle, they are secure from further modification.
Problems
We continue to modify our methods of identifying data that are in error. Primary
sources of errors result from taking inappropriate actions when the data fall out side of an
acceptable range, or not accurately determining the acceptable range. The final reports
are printed in draft. To determine if erroneous readings are loaded into the Project Oracle
database, and engineer checks the draft reports. Any erroneous readings are deleted from
the Project Oracle and Field Database. By using a more reliable validation of data when it
is collected, we hope to eliminate this laborious task.
At any particular time up to 18% of the instruments cannot be read. Reasons for
not reading instruments are a result of access problems or instrument damage. We track
reading problems and provide reports summarizing problem instruments. Because we are
paid only for instruments that are read, there is an incentive to work with the Project
Management to reduce the percentage of unread instruments.
The fast turnaround from data collection to final reporting and the large quantity
of data put severe demands on the data collections and processing system. We have
experienced failures in most of the electronic hardware, including network server, printer,
copier, instrument readout units and PDR's. We have developed backup and redundant
systems for all critical components of the data processing system. We have identified and
worked around problems with instrumentation readout units.
A large amount of communication is required to design, install, initialize, collect,

process, report and interpret instrumentation data. Add to this process a large number of
contractors, designers, abutters and other interested parties and potential for
communications problems increase. Communication problems arise because the projects
data collection requirements change rapidly (daily, sometimes hourly). We use a system
of pagers and cell telephones to communicate with the field personnel and to allow the
field personnel to communicate with the Clients resident engineers. Project Management
distributes verified data to the different parties using the Project Oracle database.
Future Trend
We expect to continue to look for more cost-effective ways to complete our work.
Since just about all aspects of processing the raw data and providing it to the client have
been automated to the extent possible, future efforts have to focus on ways to obtain valid
readings. We continue to focus on how to reduce the number of invalid or failed
readings. Recently, we reassigned a technician to focus on this task. He is to review
each instrument that we have had problems obtaining valid readings for and seek out
ways to repair, restore or retire the instrument.
Better communications technology may help us improve data flow. Instead of

waiting until the end of the day, the technician might transmit readings several times a
day. This would allow more time to prepare the reports and potentially make it possible
to provide the reports on a shorter schedule. Options now available include; direct cabling
from the instrument to the office, telephone links, or radio links. These types of
connections are only viable in special situations requiring multiple readings from
multiple instruments.
Low cost instruments with built-in low power radio communications using
cellular technology may provide some assistance. These instruments communicate
through one another to transmit readings over distances that could encompass the entire
CA/T project. Devices of this type would be especially useful in locations where access
is difficult or where frequent readings are required.
To reduce the possibilities for error, we are considering adding unique electronic
identifications to each instrument. A device such as the iButton [4] (Figure 3) gives a
unique number to the instrument that can be read with the PDR. This eliminates the
possibility for the technician to mix up instrument numbers and also verifies that the
technician actually got to the instrument. With the PDR, it can also help track the time at
which the instrument was reached, permitting us to more closely consider the time
requirements for each activity
Dallas Semiconductor iButton DS 1990A (DS 1420) 64-Bit ROM

iButton: Contains a unique, unalterable, 64-bit unique registration
number engraved both on the silicon chip and on the steel lid of
the button
Figure 3 - D a l l a s iButton for Instrumentation Identification
Although we are providing all data in electronic format, we also provide a

hardcopy report each day. This report varies between 2 and 6" of paper per copy per day.
As computers become more pervasive on the project, we anticipate being able to dispense
with the hardcopy report at some point.
HAWKES AND MARR ON ARTERY/TUNNELPROJECT 331

The Central Artery/Tunnel project involves one of the most extensive
geotechnical instrumentation systems ever deployed. The size of the excavations, the
proximity of movement sensitive structures, and the sensitize nature of the abutters place
great demands for a comprehensive, accurate and timely data management system. The
project's management team must have timely, accurate information at all times to help
maintain the safety of the project, address issues with the contractors, and answer the
questions of concerned abutters.
Computers have significantly eased the data tracking and reporting tasks
associated with the instrumentation monitoring. Data entry in the office is eliminated.
The PDRs are a valuable tool for us to obtain accurate data in a cost-effective manner.
However, there is still a considerable amount of labor required to routinely collect the
data and validate questionable readings.
Automatically collected data are desirable for reducing manpower, collecting

large quantities of data, and supporting a rapid reporting system. Technically, it is
feasible. However, on the Central Artery/Tunnel an automated data collection system is
not financially feasible because of the large cost of installing and maintaining such a
system in a large congested construction environment.
After one year of work, we are continuing to refine our data collection and
management techniques. We currently track the number of unsuccessful reading attempts
so we can isolate the cause of the problems and work to increase the successes. Our
database is being expanded to track other items that affect our cost and performance. For
example, we have added provisions to track the maintenance activities on each piece of
readout equipment. This permits us to identify trends in equipment performance so that
we can look for ways to decrease our maintenance costs.
We believe that our efforts on this project are showing that geotechnical
instrumentation can be monitored in a timely and cost effective manner. We consistently
place data into the project team's possession within 24 hours of data collection. This is
done with a delicate balance of dedicated technicians and electronic equipment. That
balance is constantly altered to adjust to the changing needs of the project. We think the
approach being used on this project can serve as a model for other large geotechnical
instrumentation projects.
Acknowledgements
The authors would like to acknowledge and thank the Massachusetts
Transportation Authority for permission to publish this paper and Bechtel/Parsons
Brinkerhoff for their cooperation. In particular we recognize the cooperation of Dr.
Thorn Neff, Mr. Charles Daugherty and Mr. Dave Druss, all of the B/PB management
team. We also acknowledge the dedicated efforts of our instrumentation team lead by
Mr. Terry Knox of TLB Associates. This paper was prepared with the support of
GEOCOMP Corporation without public funds. It contains the views and opinions of the
authors and does not represent the views or opinion by the CA/T project.
References
[ 1] Bechtel/Parsons Brinckerhoff, "Geotechnical Instrumentation", Concept Report No.

2AB24, prepared for the Massachusetts Department of Public Works, 1991.
[2] Massachusetts Department of Public Works, "Central Artery/Tunnel Project,

Geotechnical Instrumentation Monitoring (M026J)", Section 7210.505,
Addendum 18, 1996.
[3] Psion Inc., "Psion Workabout: user guide", 150 Baker Ave., Concord, MA 01742,
1998.
[4] Dallas Semiconductor, Inc. "Data Sheet : DS 1990A Serial Number iButton ",
www.ibutton.com, 1998.
Pamela Stinnette, Ph.D., P.E)
Data Management System For Organic Soil
REFERENCE: Stinnette, Pamela, "Data Management System for Organic Soil,"

Field Instrumentation for Soil and Rock, ASTM STP 1358, G. N. Durham and W. A.
Marc, Eds., American Society for Testing and Materials, West Conshohocken, PA, 1999.
ABSTRACT: A Data Management System for Organic Soil (DMSOS) has been
developed that enables the acquisition, management and analysis of organic soil data as
well as the presentation of results to be conducted effectively through a common
interface. This development was in response to the data management needs of research
investigating the engineering properties of organic soil and its extension to the
stabilization of organic soil through dynamic replacement (DR). It is shown how the
above functions are implemented efficiently using Windows-based software to perform
comprehensive data management and analysis of data gathered from both laboratory and
field tests. When the engineering properties of a given organic soil deposit are needed, a
built-in Computer Advisor for Organic Soil Projects (CAOSP) predicts the properties
from DMSOS based correlations. A unique and useful feature of the CAOSP is its
ability to estimate the anticipated ultimate settlement of an organic soil deposit given the
loading conditions and the moisture or organic content. Also incorporated in the DMSOS
is a quality control system that utilizes computerized data acquisition/data management
techniques in order to evaluate the degree of improvement of an organic soil layer at a
given stage of treatment using DR.
KEYWORDS: database management systems, data acquisition, organic soil, consolidation,

spreadsheets, ultimate settlement
Generally, geotechnical projects require site investigation, laboratory testing,

model studies and field monitoring. Each of these phases produces extensive data which
must be acquired, stored, analyzed and interpreted. In order to achieve an accurate
measurement of test data with high resolution, state-of-the-art data acquisition techniques
must be utilized. Once collected, the data should be stored in a database using a standard
format. Analysis of the acquired data can then be performed using appropriate sottware.
' Assistant Professor, Hillsborough Community College, 1206 N. Park Road, Plant City,
Florida 33566.
Currently there are database management systems (DBMS) being used in many
commercial and business enterprises. These systems have found application in the area
ofgeotechnical engineering as well. Geotechnical DBMS have been reported by
researchers such as, Adams et al. (1993), Benoit et al. (1993), Ishii et al. (1992), Lee et al.
(1993) and Sykora and Koester (1991).
Organic soil is considered an undesirable foundation material due to its low
strength and high compressibility. Therefore, the Florida Department of Transportation
(FDOT) and researchers at the University o f South Florida (USF) cooperatively
investigated: (1) the geotechnical properties of Florida organic soil; and (2) the
feasibility of stabilizing it through the use of DR with sand. Since the quantity of data
collected for the above research was voluminous, the development of a Data
Management System for Organic Soil (DMSOS) was essential for effective
implementation of the current and subsequent research projects on organic soil.
Although, the work presented in this paper addresses the setup of the basic DMSOS
structure from laboratory and field tests including DR, conducted on Florida organic
soils, its scope could easily be extended to include all soil types.
In order to develop the data management system, a common interface was created
using Microsoft Visual Basic programming. The Microsoft Visual Basic user interface of
the DMSOS operates in a Windows 3.1 integrated environment which supports Microsoft
Office Applications. These applications include: Access, a relational database, Excel, a
spreadsheet program and Query, a structured query language (SQL) that enables access
and retrieval of data stored in the Access database to other applications such as Excel
The exclusive use of Microsoft products provides consistency and automation among the
applications which results in their seamless integration in the DMSOS.
The above facilities were effectively utilized to store all measured properties of
organic soil samples collected from test sites and formulate any possible correlations. A
unique feature of the system is the inclusion of a Computer Advisor for Organic Soil
Projects (CAOSP). This program, also written in Visual Basic, assists the user by
providing the ability to estimate index properties and the ultimate settlement of any
organic soil site from DMSOS established correlations. Additionally, the optimum
energy required for effective field implementation of the DR technique can be evaluated
through the use of the developed DR quality control system which is also incorporated in
the DMSOS.
In order to illustrate the functionality of the DMSOS, the paper will be divided
into three parts. Part I gives an overview of the DMSOS structure and details the
acquisition of test data, data reduction, data entry in the Access database, data analysis
and presentation. Part II describes the capabilities of the CAOSP such as prediction of
index properties and ultimate settlement of an organic soil deposit under given loading
conditions, based on DMSOS data and correlations. Finally, Part III outlines the DR
quality control system integrated into the DMSOS.
STINNETTE ON ORGANIC SOIL 335
PART I DMSOS STRUCTURE

VisualBasic 3.0 was selected to develop the common user interface for the
DMSOS. This software offers database communication protocol to many of the
commercially available database management systems. Additionally, as it is a Windows-
based product, the resulting user interface can be graphical and it will be able to take
advantage of the Windows 3.1 operating environment. "Buttons" are commonly used
graphical objects found in this operating environment. The on-screen buttons can be
clicked or double-clicked with a mouse causing any number of actions to be performed.
They have been incorporated in the DMSOS to facilitate an intuitive graphical user
interface (GUI).
Figure 1 shows the opening screen of the DMSOS created using Visual Basic 3.0
programming. An overview of the structural blocks and their functionality in the
DMSOS is shown in Figure 2.
Data Acquisition
Due to the advancement in microcomputer technology, the use of data acquisition
has become commonplace in both geotechnical laboratory testing and field monitoring.
Many of the laboratory and field investigations on organic soil have utilized automated
data acquisition. The data storage and simultaneous screen presentation during testing
can be completely automated by a microcomputer equipped with a multifunction
input/output interface board and appropriate software. The core of the described
laboratory data acquisition system includes a 486-33 MHZ microcomputer with 20
megabytes of random access memory (RAM). The computer interface with the
transducers is accomplished with the AT-MIO-16F-5 multifunction interface board. It
contains a 12-bit analog to digital converter (ADC) which samples up to 16 separate
single-ended (SE) analog inputs or 8 channels of differential input (DI) at 200 kilo-
samples per second.
Visual Basic 3.0 was also selected for interfacing the data acquisition board to the
microcomputer~ Visual Basic communicates with the selected data acquisition board
through the use of a dynamic link library (DLL) that includes functions for controlling
the data acquisition board for the AT microcomputer. Through the use of this DLL,
executable programs have been developed using Visual Basic for both laboratory and
field tests. The laboratory tests include (but are not limited to): 1) the consolidation test
which utilizes a linear variable differential transformer (LVDT) for displacement
measurement and a pore pressure transducer, and 2) the triaxial compression test which
utilizes a LVDT, a load cell and a pore pressure transducer. Data acquisition during field
DR records data from an accelerometer, laser photoelectric ceils and dynamic pore
pressure transducers. Figure 3 shows a list of the available test options for data
acquisition once the DATAACQUISITIONbutton has been clicked on the MAIN MENU. As
an example, a double-click on DYNAMICREPLACEMENTwill bring up the user interface
shown in Figure 4. It is on this screen that the user provides pertinent test information
such as an output filename, if and how long the residual pore pressure and stress are to be
monitored, and the data recording time interval.
Figure 1 Opening screen of the DMSOS.
i 84~ ~ s ~ , ~ !
........................ j
........................................ [ ................................................................. I. . . . . . . . . . . . . . . . . . . . . . . . . . ] .....

DaaEmy Data~ly~s Data , ~ i o n O~mter ~'visw
...... ii ~IIIII!/IIII/III iii 9 i iiiiii ...... i/~ i/i .........Iiiii ii ....i i~ ....... i ii 1
IMa En~y/l~eval Fonm Data Amly~ Forms lMa A~lJsition ~Formto :
DataFormsAvailaHe DataAmlysesAvailalzle lnteffac~for~ ard Hdd ProvideesltmaledValues
~ b y t e s t type IrdexedbyTestType TesdngSch~nes aFlndexPmpc~cs
&Fern DescaStlion &TqzcDescflaion
iiiiiii Ii iiiiiiii i iiii iii i i 1 ...........iii i .....i....... i .........

~ecific DataEntry ~oecificI~taAnalyses Data Beducfim lnteracliveFormsto
Forns for mchtest Av~ahle for eachTest ExcelWod~e~ Prov~ ~
~tYI~aeq~ew Q.~ryDa~ese forSite Wnecrcs to Settlenm Preticeon u ~
or~iaData SpecificG ~ t a ~ ~ b s e s O~lihat~Jze dsta Z ~ e d ~ and ~ i r i e ~ M ~ s
Figure 2 DMSOS block functionality flowchart.
Figure 3 Available data acquisition tests screen.
Figure 4 Dynamic replacement user interface.
Figure 5 Data reduction screen of the DMSOS.
Data Reduction
After data (i.e., consolidation, triaxial, DR) have been acquired using data
acquisition techniques and stored in the computer (*.prn file), the data may then need to
be reduced depending on the specific test conducted. This is accomplished using Excel
worksheets specific to each test type which include macros written in Visual Basic for
Applications. These macros can be used to calibrate data that have been acquired and
also plot and extract important parameters specific to each test. For instance, a macro
may prompt the user to select information associated with the test such as the specific
transducer(s), sample height, diameter, etc. from a list box. Once this information has
been selected, calibration and analysis are then initiated. Specific results of the analysis
can then be extracted for input into the DMSOS. Figure 5 shows the DATA REDUCTION
SCREEN of the DMSOS.
Data Entry into the Access Database

The structure of the Access database is founded on a relational model (RM).
Although other data models, such as object oriented and feature-based geographic
information system (GIS) data models are available which have both spatial and non-
spatial components, the relational model proved to be sufficient for the DMSOS, as the
spatial component for the present study could be adequately represented by the relational
model.
The RM is presently the most commonly used type of DBMS since it allows
sufficient flexibility in data representation and provides access to its data for different
STINNE'I-IE ON ORGANIC SOIL 339
Figure 6 Incremental consolidation results table.

applications in a program-independent fashion (Rasdorf and Spainhour 1993). The
relational DBMS is also well suited for representing tabular data. Although many
relational DBMS are available, such as dBase 5.0 and Approach 3.0, Microsoft's Access
2,0 was selected as the Relational Database Management System (RDMS) for the
DMSOS. This selection was based on the flexibility of the program and the ease of
integration with the other Microsoft Office products. By using the Visual Basic 3.0
programming language, access can be made directly to the Microsoft Access database.
Additionally, coupled with the use of appropriate programming techniques, a high degree
of control is afforded over how data is stored in the Access database.
The National Geotechnical Experimentation's Site (NGES) (Benoit et al. 1993)
has developed a data dictionary (standard format for data entry) for its Central Data
Repository (CDR) database, This data dictionary has been incorporated in the DMSOS
to the extent that a smooth data exchange between the DMSOS and the CDR database
will be facilitated. Data may be entered in the database in a number of formats.
Inaddition to data being in the common numerical or text form, a data field may represent
an embedded object such as a static picture (i.e, *.pcx) or a motion video (i.e., *.avi).
Data entry into the DMSOS can be achieved by using one of two approaches: (1) direct
data input or (2) data entry forms. Data may also be entered directly into the database
through the use of Access 2.0, Figure 6 shows a portion of the Incremental
Consolidation Results Table. Data are stored tabularly with each row representing a new
Figure 7 Design mode for the consolidation table.
Figure 8 Available data entry / retrieval forms.
STINNE'I-I-E ON ORGANIC SOIL 341
Figure 9 Consolidation data entry form.
record. For instance, records 1-6 correspond to oedometer specimen "SR 951B" which
underwent six loading stages of testing. Restrictions can also be made on the data type,
data format and on the input of a particular data field in the form of a "validation rule"
incorporated in the Access database. In order to illustrate this concept the "Saturation"
data field (CONS_SATI) is selected from the design mode for the Consolidation Table
(Figure 7). The data type is seen to be "number" with a "fixed" format. The validation
rule displayed reads "0 <= 100." This rule ensures that the data entered in this field stay
within the prescribed limits. If the user enters data outside of these limits, the validation
text "Maximum saturation should be less than 100% (0<=100)" is displayed. Although
this mode of direct data entry is available, it is not envisioned as the primary method due
to the development of data entry forms which provide a GUI for ease of data entry from
the DMSOS. Data that already reside in the database may also be retrieved, reviewed
and printed using these forms. As the different tests require separate tables for storage,
there will also be separate forms for data entry and/or review. To access each of the
various data entry forms, a list of all available data forms is provided to the user on a
separate form with a description of the fields associated with that form. Figure 8 shows
the data entry/review forms that are available in the DMSOS after the VB DATAFORMS
button has been depressed on the MAINMENU. The FIELDDR PILOTSTUDYRESULTS
FORM is highlighted and hence a list of included data fields are shown in the FORM
DESCRIPTIONbox. As an example, a double-click on the MAINCONSOLIDATIONDATA
FORM reveals the CONSOLIDATIONDATAENTRYFORM (Figure 9). This form may then be
used for new data entry, or the review of pertinent consolidation results. As seen in
Figure 10 Data analysis available.

Figure 9, a static picture of the test apparatus used is displayed on the form. The
reference to this picture and all of the pertinent data fields displayed are stored in the
Access database. A separate data form is required for entry of data associated with each
incremental loading stage and may be accessed by clicking on the INCREMENTALbutton.
The data entry forms, developed using Visual Basic 3.0, provide additional
guidance for data entered into the database. As an example, the "water content" data field
is highlighted on the CONSOLIDATIONDATAENTRYFORM (Figure 9). The field
comments read "Enter water content as a percentage," thus ensuring that the user is aware
of the input data format. Additionally, when the CHECKDATA ENTRIESbutton is clicked,
pertinent data fields are checked against previously established correlations to assure data
integrity. The user is prompted of the results in the adjacent box. Coupled with Access,
this will help maintain data integrity and ensure that data stored is in the appropriate
units/form etc. and follows standard testing convention.
Data Analysis and Presentation

Data stored in the Access database is retrieved by using Query. Query enables dynamic
communication between Excel and the Access database by directly querying the Access
database, extracting the information sought, and then importing the data into the Excel
spreadsheet. The data can then be manipulated and analyzed using macros programmed
with Excel "s Visual Basic for Applications. Hence, with little noticeable delay and
Figure 11 Consolidation Data Analysis using Microsoft Query and Excel.

without requiring any special input from the user other than a mouse click, Excel can
directly query the database using Query and statistically analyze and plot the results.
Moreover, as new data becomes available the relevant correlations can also be updated.
Figure 10 shows a list of the available test types and their relevant analyses after
the DATAANALYSISbutton has been depressed on the MAIN MENU. In order to illustrate
the role of DATA ANALYSISin the DMSOS, once again the consolidation test will be
used. Hence, a double-click on CONSOLIDATIONANALYSISshown in Figure 10 reveals
the form shown in Figure 11. A box in the upper left hand comer of this screen indicates
the data plotting options that are available for the consolidation test. The first three
options in this list represent cumulative test results based on data currently available in
the database. Results of the statistical analysis, if appropriate, are included on the graph.
Double-clicking on one of the graphical analyses listed initiates object linking and
embedding (OLE) automation. OLE can be used to start applications, view and/or edit
specific information, allow users to send and/or retrieve data from other applications and
control the actions of other applications. Since, a complete description of OLE
automation is beyond the scope of this paper, details outlining its use and application can
be found in a related Microsoft publication (Microsoft Corp. 1994). Through OLE
automation, the appropriate Excel spreadsheet is activated which then reveals the specific
graph of interest. Double-clicking the REFRESHDATA button on the graph will then
automatically cause the database to be queried and reflect, graphically and statistically,
any new additional data that has been recently entered into the database. Additionally,
the graph may be printed by clicking on the PRINTbutton.
The remaining graphical analyses listed in the box are site specific. In order to
view a specific site of interest, Excel and Query must be utilized together. As an
example, if one were to double-click on the VOIDRATIOVS. PRESSUREgraphical analysis,
Excel and Query would be activated. Excel would then display the Void Ratio vs.
Pressure graph for the last site selected (Figure 12). A click on the SELECTNEW SITE
button enables the user to select a specific site of interest. This screen is shown in Figure
13. At this point, the site of interest denoted by project identification (example: SR 951
B) is highlighted under each of the graphical analyses list boxes. This form gives the
user the option of performing any desired graphical analyses for any selected site. Once
a new site is selected from the list, Query would then extract the pertinent information
from the appropriate fields of the consolidation table associated with that site, import the
data into Excel, and then plot the results on the appropriate graph. A click on the OK
button will return focus to the originally selected graphical analysis. Movement between
the remaining graphical analyses is facilitated by clicking the appropriate button on the
graph. The graphical analyses for site $951B are shown in Figure 14. At this point, the
graphs may be printed, the process repeated for a new site, or one may exit and return to
the previous DATAANALYSISscreen of the DMSOS (Figure 10).
PART II COMPUTERADVISORFOR ORGANICSOIL PROJECTS(CAOSP)

Index Property Correlations
A comprehensive laboratory study was previously undertaken to determine the
engineering properties of Florida organic soils (Stinnette 1992). This evaluation included
the determination of index properties such as the moisture content, organic content,
specific gravity and density. Correlations which have been determined consequent to the
above study will be maintained and continuously updated in the Computer Advisor for
Organic Soils Projects (CAOSP). The following linear correlations with the indicated
coefficients of correlation (R 2) have been found to exist among index properties gathered
from test sites in Central and South Florida (Stinnette 1992):
Figure 12 Void ratio vs. Pressure graphical analysis.
Figure 13 Site Selection Form.
Figure 14 Graphical analysis for site $95 lB.
Figure 15 CAOSP.
Dry Density and Water Content; R 2 = 0.98

Wet Density and Water Content; I{2=0.87
Organic Content and Water Content; R 2 = 0.86
Specific Gravity (1/Gs) and Organic Content; R 2 = 0.93
Initial Void Ratio (eo) and Water Content; R 2 = 0.95
Initial Void Ratio (%) and Organic Content; R 2 = 0.96
As more data becomes available from experimental research, efforts will be made
to further verify these correlations. In addition to incorporating correlations determined
by the USF researchers, organic soil correlations established by other researchers will
also be included in CAOSP. Currently this includes investigations by Yaromko and
Ses'kov (1979) in which the shear wave velocity of peat and organic mud are
characterized by the moisture content. Clicking on the CAOSP button on the MAIN
MENU reveals the screen shown in Figure 15. This screen shows the CAOSP user
interface for estimating a number of index properties of any other organic soil site based
on the moisture content and, if available, either the organic content or density. As an
example, if the water content and organic content were determined to be 600% and 85%,
respectively, then through the use of the CAOSP (Figure 15), values of the shear wave
velocity (Vs), wet density (~/w), dry density (ya), initial void ratio (e) and specific gravity
(Gs) are estimated as:
Vs = 20.2 m/s
Yw = 1.00 g/cm 3
Ya = 0.156 g/cm 3
e = 9.46
Gs = 1.50
Ultimate Settlement Prediction

Rheologieal Model
A critical construction problem associated with organic soil deposits is the
excessive long-term settlement. In order to address this issue, extensive consolidation
tests were performed on numerous organic samples from across the state. The author's
previous work (Stinnette 1992) also showed that the Gibson and Lo (1961) simplified
expression for large time intervals accurately predicts the ultimate settlement for these
samples. The ultimate settlement of each sample at all loading stages was then
determined using appropriate rheological models and curve fitting techniques due to
Gibson and Lo (1961) and Tan (1971), respectively. Since descriptions of these models
can be found in previous work (Stinnette 1992), the basic equations are given in
Appendix.
Based on the above laboratory data a new technique was then formulated to
predict the ultimate settlement of normally consolidated (NC) organic soils based on the
organic content and the consolidation pressure (Gunaratne et al. 1998). This technique
enables quick and accurate estimation of the settlement of an organic deposit anticipated
under a given loading condition and hence, it is invaluable for construction operations. It
involves the development of separate expressions for primary and secondary
compressibility parameters (a and b) of the Gibson and Lo (1961) model in terms of
organic content and consolidation pressure. While a complete description of this
rheological model is found in Gunaratne et al. (1998), the resulting expressions have been
incorporated in the CAOSP.
However, the previously described model is valid in the field only for one-
dimensional consolidation under extensively placed loads such as landfills. Stinnette
(1996) extended the previous expressions to suit the field case of three-dimensional
consolidation. This extension becomes important when, for instance, a circular surcharge
is applied to preload the organic soil to induce settlement. The extended model predicts
the ultimate field strain due to a pressure increase from o'~ to 0'2 as (Stinnette, 1996):
% ) , . , : f o,o.(o') ao' + f b,~ ao' + : do (1)

(Ill 0/1 0/1
(2o')
69 kPa (10 psi) ]
- r
y = 11 kN/m 3
(70 pef)
u = .25; Ko =.33 +(r,z)
E = 150 kPa
Figure 16 Example of imposed stress conditions in

the field.
where
a and b are defined in Appendix A
u = Poisson's Ratio
E = Elastic Modulus, kN/m 2
Ko = Coefficient of lateral earth pressure at rest
K = Coefficient of lateral earth pressure under loading
The following example illustrates the estimation of the anticipated ultimate

settlement using the extended model (Eq. 1) incorporated in the CAOSP. The
formulation of this model and the comparison of its prediction with field results is found
in Stinnette (1996).
Figure 16 shows a circular load of radius of 6. l m (20') exerting a pressure of 69
kPa (10 psi) on the ground surface of a homogeneous organic soil. The soil has a density
of 11 kN/m ~ (70 pcf), an organic content of 50%, a coefficient o f at-rest lateral earth
pressure (Ko) of 0.33 and an elastic modulus, E, o f 150 kPa. It is desired to estimate the
ultimate settlement due to the single stage loading in the upper 6 m of the deposit. A
click on the ULTIMATE SETTLEMENT - LANDFILL AND UNIFORM CIRCULAR LOAD (GIBSON
ArID Lo MODEL) button on the form shown in Figure 15 leads to the screen shown in
Figure 17. The pertinent information required for the analysis is then entered in the
appropriate boxes on the upper portion of the screen. If necessary, estimates of the
engineering properties (E, Ko, OC, etc.) can be obtained by clicking on the TYPICAL
PROPERTIESbutton. A mouse click on the CALCULATEbutton initiates the mathematical
analysis. Once the analysis is performed utilizing Visual Basic programming coupled
Figure 17 Ultimate settlement analysis - Landfill and uniform circular

surcharge.
with Mathcad, the results are then displayed on the lower half of the screen (Figure 17).
As seen from Figure 17, the ultimate settlement predicted from the extended model for
the circular surcharge is 1.303 m. The 1-D "landfill" analysis prediction of 1.248 m is
also given for comparison. The above analysis assumes that the soil is homogeneous.
On the other hand, if the soil is layered, a multi-layer analysis should be conducted using
an appropriate El/E2 ratio (Figure 15).
Empirical Model
A simple curve-fitting technique was also found to give reasonable results for the
estimation of the ultimate settlement of an organic soil. The method is also based on
laboratory consolidation tests conducted on saturated samples of Florida organic soils
with organic contents which varied from 25% to 90%. These soils were found to exhibit
a hyperbolic strain versus time relationship and a hyperbolic ultimate strain versus
pressure relationship. Utilizing these relationships, an empirical equation was developed
to predict the ultimate strain of an organic soil based on the pressure increase and organic
content. A complete formulation of this model and the comparison of its predictions to
field results is found in Stinnette (1996). This model is also incorporated in the CAOSP
and can be accessed by clicking on the ULTIMATESETTLEMENTANALYSIS- LANDFILL
(HYPERBOLICMODEL) button on the form shown in Figure 15.
PART III - DR QUALITYCONTROL SYSTEM

Dynamic replacement (DR) is a relatively new ground modification technique that
has been used successfully to stabilize organic soil deposits by replacing the organic soil
with sand columns. A quality control system that evaluates the degree of improvement
of the organic soil layer at a given stage of treatment has been developed (Stinnette et al.
1997). In order to activate the quality control system, an accelerometer must be attached
to the drop weight and monitored during impact. The quality control system evaluates the
effects of treatment by progressively computing the dynamic settlement modulus (DSM)
from the impact acceleration record during the entire treatment period. The role of DSM
in evaluation of field DR is illustrated by using results from a field DR study on organic
soils as follows.
The DR field pilot study site is located in Plant City, FL. A 0.2 ha area of the site
was cleared and filled with approximately !-1.5 m of sandy fill material (approx. 7%
fines). This material acted as a work platform for the crane and Cone Penetration Test
(CPT) rig and as initial replacement material for the dynamic replacement process. The
thickness of the surficial organic deposit at the pilot study site was 1.7 m. The peat was
characterized as primarily amorphous in nature having an organic content of 95%. A
complete description of the field study and its results are given in Stinnette et al. (1996c).
During the pilot studies, the drop weight (3.64 tonne) was instrumented with a
piezoelectric accelerometer and connected to the data acquisition system described
previously. When the photoelectric cell beam (2.13 m above the ground surface) sensed
the falling drop weight, data acquisition was initiated. Measurements are recorded during
the impact for a duration of 1 second and were then stored in the computer (*.prn). The
acceleration vs. time graph is automatically plotted on the Visual Basic data acquisition
screen during testing (Figure 14). The file is then automatically analyzed by the DMSOS
by clicking on DYNAMICREPLACEMENTMACRO in the drop down menu on the DATA
REDUCTIONSCREEN (Figure 15). This action launches a specially written Excel
spreadsheet which uses interactive macro programming. The values of acceleration,
velocity and displacement vs. time are automatically calculated. Once the data has been
reduced, both the continuous data such as the accelerometer record and discrete data such
as the drop height, weight etc. are then stored in the Access database. Figure 18 shows
the discrete data entry/retrieval form for the DR Pilot Study Results. In addition to
pertinent test results being displayed, a double-click on the photo shown in the lower
right portion of the screen plays the video of the field DR during that drop. A click on
the GRAPHICALANALYSISbutton reveals the DYNAMICREPLACEMENTGRAPHICAL
ANALYSISSCREEN (Figure 19). As an example, a double-click on the ACCELERATION,
VELOCITYANDDISPLACEMENTGRAPHICALANALYSIS initiates OLE automation and
opens the appropriate Excel spreadsheet which then reveals the specific graph of interest
(Figure 20). At this point, the DR data stored in the Access database may be retrieved for
any impact number by clicking on the SELECTNEW TESTbutton which reveals the DR
PILOT STUDYLOCATIONFORM (Figure 21).
Figure 18 Dynamic replacement pilot study results form.
Figure 19 Dynamic replacement graphical analyses screen.
Figure 20 Acceleration, velocity and displacement graphical analyses

screen.
The DSM is evaluated from the measured data as follows. First, the record of
impact stress, o, is determined from the acceleration record, a, by using Eq. (2)
m a
O - (2)
A
where:
m mass of the drop weight
A = base area of the drop weight
Figure 21 DR pilot study location selection form.
Figure 22 Impact stress vs. Relative displacement analysis.
Figure 23 DSM vs. Blow Number analysis.
Then, the ratio of the instant displacement to the equivalent diameter, Do (relative
displacement), is then plotted against impact stress, o. The dynamic settlement modulus,
DSM, is then calculated as the slope of the tangent of the loading portion of the impact
stress-relative displacement curve given by:
Ao
DSM -
6(~o ) (3)
A click on the SHOW GRAPHbutton next to Impact Stress vs. Relative Displacement
(Figure 21) reveals the graphical analysis of the DSM (Figure 22), Once the DSM is
calculated for each blow, it is then stored in the Access database. The plot o f D S M vs.
Blow Number is displayed once the SHOWGRAPHbutton has been clicked (Figure 23).
Poran et al. (1992) outlined the use of this technique for the quality control of
dynamic compaction on dry sand and report that the DSM values obtained based on
experimental tests conducted in the laboratory have shown good correlation to soil
densities. F u ~ e r the relative change in DSM values was found to be proportional to the
rate ofdensification (density increase as a function of number of drops). As seen in
Figure 21, during the illustrated pilot study DR seems to be more effective in the first 15
drops, while after 15 drops there is an insignificant increase in the DSM, showing little
additional improvement. This indicates that the optimum number of drops should be
approximately 15 for this particular site.
CONCLUSION
An in-house data management system was developed in response to the data
management needs of an ongoing organic soils research project that produces extensive
laboratory and field data. The management system developed is based on the Windows
3.1 operating system and integrates many advanced features afforded by this operating
system. The main features of the DMSOS are: (I) automated data acquisition during
consolidation, triaxial and field dynamic replacement tests, (2) analysis of automatically
acquired or manually entered data, and (3) easy retrieval of general correlation results for
organic soils or test parameters specific to a user selected organic soil site. In addition, a
Computer Advisor for Organic Soils Projects (CAOSP) built into the DMSOS is capable
of furnishing many useful engineering properties of an organic soil deposit once easily
measurable parameters such as organic content or moisture content are input. Of these,
the most critical properties from a construction perspective is the compressibility of the
organic soil deposit. Also incorporated in the DMSOS is a quality control system
developed for the evaluation of dynamic replacement (DR) of organic soils. The system
facilitates rapid and efficient analysis and graphical presentation of the DR test results.
These include the plots of acceleration, velocity and displacement of the drop weight for
each impact. Additionally, the DSM parameter is computed for each blow and
graphically presented such that its trend may be observed as treatment proceeds. From
this plot, the optimum energy (number of blows) can be predicted when the change in the
DSM becomes relatively insignificant. The system can easily be implemented in any
laboratory with existing microcomputer facilities and it will accommodate any new
Windows-based software developments in the foreseeable future.
ACKNOWLEDGMENTS
The author would like to thank research colleagues Dr, A. Mullins, Dr. S.
Thilakasari, and Professor M. Gunaratne of the University of South Florida.
Additionally, the author would like to thank Mr, Brian Jory and Ms. Terry Puckett of the
Florida Department of Transportation, District 7 for their support during the study.
Finally, the FDOT grant B-8452 WPI 0510665 is gratefully acknowledged.
APPENDIX ULTIMATE SETTLEMENT DETERMINATION OF ORGANIC SOIL
Gibson and Lo Rheological Model (1961)

The Gibson and Lo equation for large values o f time is expressed as:
(--.F) (AO
e(t; = Ao[ a + b ( 1 - e )]
where
Ao = effective stress increment
t = time
a = primary compressibility
b = secondary compressibility
L/b = secondary compression rate factor
Hence, the ultimate settlement ( t - oo) may then be expressed as:
8u = AO H [a + b] (A2)
where
H height o f organic soil layer
Tan's Empirical Model (1971)

Tan's empirical relationship for large values o f time is given as:
t
-Mt+C (A3)
6
where
t = time
6 = total settlement at any time t
M&C = empirical constants
It can be seen that as t becomes very large ( t ~ oo) the ultimate settlement can be
expressed as:
~ult = 1 (A4)
M
REFERENCES
Adams, T.M., Tang, Agatha Y.S. and Wiegard, N. (1993). "Spatial data models for
managing subsurface data." J.. of Computing in Civil Engineering, 7(3), 260-277.
Benoit, J., P.A. de Alba and Sawyer, S.M. (1993). "National geotechnical experimentation
sites central data repository." In Geographic Information Systems and their
Application in Geotechnical Earthquake Engineering, 17-20.
Gibson, R.E. and Lo, K.Y. (1961). "A theory of consolidation of soils exhibiting secondary
compression." Acta Polytechnical Scandinavia. Ci 10 296: Scandinavian Academy
of Science.
Gunaratne, M., P. Stinnette, G. Mullins, C. Kuo and W.F Echelberger. (1998).
"Compressibility relations for natural organic soil." ASTMJ. of Testing and
Evaluation, January.
Ishii, M., K. Ishimura and T. Nakayama. (1992). "Management and application of
geotechnical data: the geotechnical data information system of the Tokyo
metropolitangovemment." Environ. Geol. Water Sci., 19(3),169-178.
Lee, F.H., T.S. Tan, G.P. Karunaratne and Lee, S. (1990). "Geotechnical data management
system." J. of Computing in Civil Engineering, 4 (3), 239-254.
Microsoft Corporation. (1994). Microsoft Office Developer's Kit, Microsoft Corporation.
Poran, C.J, Heh, K.S. and Rodriquez, J.A. (1992) "A new technique for quality control of
dynamic compaction," in Grouting, Soil Improvement and Geosynthetics,
Proceedings of conference sponsored by ASCE, February 25-28, 1992, Geotechnical
Special Publication No. 30, ASCE, New York, pp. 915-926.
Rasdorf, W.J. and Spainhour, L.K. (1993). "Developing and implementing a conceptual
computing materials design database." In Proc. of ASME Computers in Engrg.
Conf., San Diego, CA.
Stinnette, P. (1992). Engineering Properties of Organic Soils. Masters Thesis, University
of South Florida, Tampa, Florida.
Stirmette, P. (1996). GeotechnicalData Management and Analysis Systemfor Organic Soil.
Doctoral Thesis, University of South Florida, Tampa, Florida.
Stinnette, P, Gunaratne, M., Mullins, G. And Thilakasiri, S. (1997). "A quality control
program for performance evaluation of dynamic replacement of organic soil
deposits." Geotechnical and Geological Engineering,Vol. 15, No. 4, 283-302.
Sykora, D.W. and Koester, J.P. (1991). "Database of seismic body wave velocities and
geotechnical properties." In Geotechnical Engineering Congress 1991 ASCE

Geotechnical Special Publication No. 2 7, 690-700.
Tan, Swan Beng. (July 26 - August 1, 1971). "A method for estimating secondary and total
settlement." In Proc.of the Fourth Asian Regional Conf. on Soil Mechanics and
Foundation Engrg., 2, Za-Chieh Moh, ed., Bangkok, Thailand.
Yaromko, V.N. and Ses'kov, V.E. (1979). "Elastic and dissipative properties of peats and
organic muds." Translated from Osnovaniya, Fundamenty I Mekhanika Gruntov, 1,
11-13.
D.J. Bobrow 1 and S. Vaghar 2
Review of GIS Developments in Geotechnical Instrumentation on the Central

Artery/Tunnel Project, Boston, Massachusetts
REFERENCE: Bobrow, D.J., and Vaghar, S., "Recent GIS Developments in

Geotechnical Instrumentation on the Central Artery/Tunnel Project, Boston,
Massachusetts," Field Instrumentationfor Soil and Rock, ASTM STP 1358, G.N.
Durham and W.A. Marr, Eds., American Society for Testing and Materials, West
ABSTRACT: This paper discusses the Geographic Information System (GIS) application
which has been developed for rapid analyses and reporting of instrumentation and survey
data on the Central Artery/Tunnel Project, in Boston, Massachusetts. Recent
developments of the GIS application have included the addition of modules for preparing
groundwater and settlement contour plans, and integration of digital photographs into the
projectwide mapping. A module for inserting geological cross sections is underway.
KEYWORDS: geographic information system (GIS), graphical interface, wide area

network, contouring, digital photograph
The Central Artery/Tunnel (CA/T) Project is currently the largest government

funded public works project in the United States. An existing 1950s elevated highway is
being replaced by a tunnel alignment through downtown Boston. Figure 1 shows the
project location; Contract C09A4 consists of jacking prefabricated concrete tunnel
segments beneath an existing and operational railroad line.
The following sections provide an overview of the Project, the instrumentation

program, and the GIS application which has been developed to store, process, and present
the instrumentation information.
lGeotechnical Engineer, Bechtel/Parsons Brinckerhoff, Central Artery/Tunnel Project, 1

South Station, Boston, MA 02072.
2Loss Control Consultant, American International Group, 101 Federal Street, Boston, MA
02072.
BOBROW AND VAGHAR ON GIS DEVELOPMENTS 361
A full account of this application can be found in a 1997 paper prepared by the
authors (Vaghar et al, 1997).
All elevations mentioned in this paper are referred to the Central Artery/Tunnel
Project Datum (PD), which is I00 feet below the National Geodetic Vertical Datum
(Mean Sea Level).
Background
Boston is in the geographic center of the Boston Basin, an area with sedimentary,
metamorphic and igneous rocks of late Precambrian and Cambrian age, overlain by glacial
and post glacial soil deposits (Kaye 1982).
The Project site lies predominantly within an area that was once mostly covered by
seawater, tidal marshes and estuaries. Most of these areas have been filled over the years,
using clays dredged from the harbor, and sand and gravel from gravel pits and glacial
drumlins, as well as building demolition rubble. The filling began in the late 1700s and
generally ended in the 1960s (Ty 1987).
An idealized subsurface geologic profile of the Project is shown in Figure 2.
Instrumentation Program
The instrumentation program on the Project serves the following objectives

(Vaghar et al. 1997):
1. Monitoring of ground and facilities deformations resulting from the

construction, and forewarning of unforeseen conditions which may impact the
safety of the personnel, the public, and the construction work.
2. Monitoring of changes in groundwater patterns and possible resulting ground

movements, which may adversely impact the abutting facilities.
3. Providing pre-construction baseline condition data for comparison with

construction and post-construction data, to aid in the evaluation of construction
related impacts.
Figure 3 shows typical instrument installation details for selected instruments.
Overview of GIS Application
As mentioned above, a full account of the GIS application can be found in a 1997
paper prepared by the authors (Vaghar et al. 1997). The GIS application allows a user
to retrieve, display and generate output from text and graphics databases for a specific
area and then to relate this information to data collected over time for the selected
Figure 1 - Location Plan
instruments located in the specific contract. The text data are stored in Oracle, a relational
database, while the graphics data such as mapping are stored in GDS, CA/T Project
CADD system.
The Structured Query Language (SQL) used to communicate with the relational
database allows the users to create the database tables and store and manipulate
information in these tables, as well as maintain the database itself.
Once the data are entered and checked, they are available to users across a Wide
Area Network (WAN). Users unfamiliar with the location of instruments can access a
GDS graphical interface for location of the instruments.
Using GIS technology, spatial analysis capabilities allow the users to map tabular
data from the relational database, interactively query graphics for associated tabular data,
and restrict tabular data to a geographic area defined in GDS. Oracle and GDS are
integrated using a Graphical User Interface. Using the GDS module SQL CADD,
customized menus and basic programs written in GDS, the user is able to access and
perform operations on both databases to generate customized maps, without an in-depth
knowledge of either GDS or Oracle.
Use of the GDS graphical database helps with the evaluation of the data. When a
change in the data is identified, the user can locate the area geographically, and review
instrument readings. The user can zoom out to display an area as large as needed for the
review. Once the desired instrument types are displayed, an attribute query can display
specific categories of instruments within the group, such as instruments read within the
past week, vibrating wire piezometers in clay, and instruments that exceed a predefined
allowable value. The user may also perform the above mentioned queries, and color code
any instruments that meet the criteria, by using pull-down menus.
Included in all reports sent to the Resident Engineer's office is a cover letter with
an evaluation of the data. Any significant changes in readings are correlated with
construction or natural activities and factors. When instruments exceed the predefined
contract response values, a graph of the data accompanies the report. The Project uses
plotting software to plot the instrumentation data. Oracle generates a text file containing
all the data as well as default parameters used by the plotting program.
Recent Developments
The GIS application described above has been under concurrent development and
use since 1994. New features are continuously developed and added, as new needs are
identified. The following modules have recently been added.
9 A module which generates groundwater and settlement contour plans.
0~
120 --
FILL m
I00-- -- ~00
r
~- -8o p_
SAND
9< --
Q C
E
~ - --60 ~ m
z
~ - -- ~J
o
~
Q_
JACKED T L ~ N E L Z
- - ~ -n
I--
w 20-- MARINE CLAY -2o W o
w
- o_
Z r-
~o- z
I---
<
>
~ -2o - -- - 2 0 0
W 0
GLACIAL DEPOSITS
-,40- --...40
BEDROCK
70 72 74 76 78 80 8Z 84 86 88 90 92 94 96 98
STATIONS IN FEET
SCALE, VERTICAL- I0 x HORIZONTAL
Figure 2 - I d e a l i z e d C09A 4 S u b s u r f a c e P r o f i l e
OBSERVATION WELL (OWl VIIt:tATINO WIfE PIF70#/EI'ER (VWPZ|
J i,-'-'- Cil~i~II
II1~"
~ ~ "
"/---4.6 ~ Ill.
~
!i
~ ~'~
~'i
!
'i---
i F FiL'II~t
III~I_I~ {INCI_|
MLILTI ~ HEAVE ~ I1413"~)
~ ~ 61IUOUT . - - - c o 4 ~ q r a,~txn"
i-i ' ~"
,o-f ~C
4.5" TO 5.5" 9 ~,I--4.5" TO 5.5~

O~A. 80REH].E~ F : ~. ~.E
TOP OF STRATUM
~ 8k.~E ~f'ECIAL {~OUT
FI~'Y _ ~ _ _ T'ffF_8
b
rOk'RN m~a
I
~ 3 . 7 5 " MIN. O(A. ~ ,~ ', ~-.-a~ENT eaO.tr
r BOR~-IOLEF IN .
Figure 3 - Typical Instrument Details
9 A module which has enabled site photographs taken with digital cameras to be
incorporated into the GDS base mapping, to be made available to the users,
without a need to exit the GIS application.
These modules are briefly described below.
Contour Plan Module
The contouring module integrates the speed, quality assurance, and the data
crunching attributes of Oracle with the ease of use and graphic ability of GDS. Thus the
user is able to produce quick, accurate, and up-to-date contours for either groundwater or
ground settlement. As soon as new data has been entered into Oracle it is available for
contouring. Having access to data for multiple contracts allows the user to evaluate the
effect of groundwater drawdown in one contract on adjacent contracts.
The contour plans are fully customizable. The groundwater contour routine leads
the user through a series of options such as instrument type, minimum and maximum date
of readings, contour interval, spot elevations on instruments, contour colors, and geologic
strata of instruments. The output is a contour plan with labels and spot height water
levels at each instrument location. The option of selecting minimum and maximum dates
of readings allows present contours to be compared with contours prior to construction.
This also allows bracketing time so that during a pump test, for example, the immediate
response of the groundwater can be assessed.
The settlement contour routine has similar features to the groundwater contour
routine. In addition, it is possible to combine a number of instrument types, for example
deformation monitoring points on buildings and on the ground adjacent to the building can
be combined for a comprehensive plan coverage.
The advantages of using a contouring module in GDS as opposed to a

commercially available contouring software are the facile nature of plotting the tabular
data and the plans on one platform, and the ability to use in-house programmers to
enhance the module, on an as needed basis.
Digital Photograph Module
The Project tracks or monitors over 250 buildings and structures, which abut the
Project, for possible construction impact. Details of each building such as building date,
ownership, structural ranking and foundation type, along with photographs of the exterior,
are kept in a file for each building. Retrieving this information, in order to correlate the
address with the instrumentation data has been a manual and rather time-consuming task,
due in part to the nonsystematic nature of some of the building addresses.
Advantage was taken of the increasing commercial availability of digital cameras

and associated software to add a module in the GIS application to locate and retrieve the
building photographs electronically. The exterior of each building was photographed

using a digital camera. The photographs were assigned an address and placed in a
database. Through a link with GDS, these photographs can be viewed by clicking an icon
on the building footprint on the project wide location plan. The photographs can also be
printed or imported into other documents, in order to accompany instrumentation reports.
Work is underway to add the building data retrieved in the form of a pop-up table. Using
this module, buildings which the instrumentation data suggest may have been impacted,
are readily identified, for possible further action such as owner notifications, and
installation of additional instrumentation.
Future scheduled steps include the addition of interior photographs on selected

buildings, particularly those which have been internally instrumented, with a link to a
commercially available sottware which renders a "virtual reality" three dimensional
picture of the interior of the building. This function is useful when evaluating buildings
where the number of internal instrumentation exceeds several hundred, or where routine
building access is difficult.
A follow up to the above module is to link the GIS application to the network of
overhead video cameras which continuously monitor many areas of the Project and the
downtown alignment corridor. This would enable the instrumentation data to be
correlated with the causal construction activities on a real time basis.
Conclusion
Development of the GIS application has led to effective processing and evaluation
of instrumentation data. The use of in-house programmers has allowed the development
and enhancements to be directed by the user engineers, thus producing an application
which is both user friendly and focused on the needs of the Project.
Recent additions such as groundwater and settlement plans and digital photograph
modules go beyond the basic data processing, single instrument and tabular reports, and
present the next logical steps in the evaluation of the instrumentation data.
Acknowledgment
Thanks are due to the Massachusetts Highway Department for permission to

publish Project information.
References
Kaye, C.A., 1982, "Bedrock and the Quaternary Geology of the Boston Area",
Geological Society of America, Reviews in Engineering Geology, Volume V.
Ty, R.K.S., 1987, "History and Characteristics of Man-made Fill in Boston and
Cambridge", MSc. Thesis, Massachusetts Institute of Technology, pp. 132, 31
maps.
Vaghar, S., Bobrow, D.J,, and Marcotte, T.A., 1997, "Instrumentation for
Monitoring Ground Movements; Central Artery/Tunnel Project, Boston,
Massachusetts", Rapid Excavation and Tunneling Conference, Las Vegas, Nevada.
Instrumentation for Measuring Physical
Properties in the Field
Kanglin Li 1 and M. R. Reddy2
Use of Tensiometer for In Situ Measurement of Nitrate Leaching
REFERENCE: Li, K. and Reddy, M. R., "Use of Tensiometer for In Situ

Measurement of Nitrate Leaching," Field Instrumentation for Soil and Rock ASTM
STP 1358, G. N. Durham and W. A. Marr, Eds., American Society for Testing and
Abstract: In order to monitor nitrate leaching from non-point source pollution, this
study used tensiometers to measure insitu nitrate concentration and soil-moisture
potential. Instead of filling the tensiometers with pure water, the study filled the
tensiometers with nitrate ionic strength adjuster (ISA, 1 M 0NH4)2SO4). After the
installation of the tensiometers at various depths along soil profiles, a portable pressure
transducer was used to measure the soil moisture potential, and a nitrate electrode
attached to an ion analyzer was used to measure the nitrate concentration insitu. The
measurement was continuous and non-destructive. To test this method in the
laboratory, eight bottles filled with pure sand were treated with known nitrate solutions,
and a tensiometer was placed in each bottle. Measurements were taken every day for 30
days. Laboratory test showed a linear relationship between the known nitrate
concentration and the tensiometer readings (R= = 0.9990). Then, a field test was
conducted in a watermelon field with green manure mulch. Field data indicated a
potential of nitrate leaching below the soil depth of 100 cm when crop uptake of
nutrients was low.
Keywords: nitrate leaching, tensiometer, soil-moisture potential, ground water, porous

media
To identify chemical transport in a soil system, frequent measurements are

required to evaluate: 1) soil matric potential; 2) availability of plant nutrients; and 3)
quality of ground water. Both moisture and nutrients in the topsoil support the growth
of plants. However, nutrients found in deeper horizons of the soil are not desirable
because they cause pollution of the ground water. Ground water is the source of
drinking water to about 50% of the overall population and to more than 90% of the rural
1,~Assistant Professor and Professor, respectively, Department of Natural Resources and

Environmental Design, North Carolina A&T State University, Greensboro, NC
27411.
population in the USA (Office of Technology Assessment 1990). Nitrate leaching is

one of the most problematic and widespread of the vast number of potential ground
water contaminants (Keeney 1986). Ground water with a nitrate nitrogen concentration
> 10 mg L ~ is considered unsuitable for drinking in the USA (Canter 1997).
Tensiometers are widely used for irrigation management (Smajstrla and
Locascio 1996, Peterson et al. 1993, Gaussoin et al. 1990) and for measurement of soil
hydraulic potential (Hendrickx et al. 1990, Cassel and Klute 1986). Li et al. (1997)
used a series of tensiometers to determine whether water movement and solute transport
in intact saprotite materials were preferential. A tensiometer generally consists of a
plastic tube closed at the bottom by a water-permeable porous ceramic cup. The tube is
filled with pure water, then closed by a rubber septum stopper at the top. The
permeable end is inserted into the soil with the tube usually in a vertical position. Water
passes through the porous cup in either direction, depending on the water content of the
soil. Water moving out of the tube into a dry soil creates a negative pressure head in the
tensiometer. A pressure transducer is used to measure the pressure head. For a
saturated soil, tensiometer readings are near zero kPa and could be as low as -80 to -100
kPa for dry soils. Therefore, the tensiometer is a device that contains a permanent
internal volume of water capable of continuous ionic exchanges with soil solution. The
solute concentration of the tensiometer solution is a function of the soil solution.
Lysimeters are used to collect solution from porous materials in order to
measure solute concentrations in soil solution. The lysimeter is a sampler using a porous
ceramic cup with a large diameter. With the cup in contact with the soil, soil solution is
trapped in the lysimeter due to the negative pressure created inside the cup for a period
of time. Some studies have summarized the advantage of this method for analyzing soil
solution (Heinrichs et al. 1996, Moyer et al. 1996, Fraser et al. 1994). Ohte et al. (1997)
used insitu lysimeters to study inorganic nitrogen discharged from forest soils. Hempel
et al. (1995) studied mercury behavior in mercury contaminated sites by lysimeters.
Vanclooster et al. (1995) monitored solute transport in a multi-layered sandy soil by a
lysimeter. They found that all soil water was used to transport solute in the top of the
soil. Studies also found some drawbacks of this method: 1) The accumulation of
solution in the ceramic cup was irreversible (Bernhard and Schenck 1986). Ionic
exchange between the lysimeter and the soil solution was limited; 2) it was difficult to
obtain a volume of soil solution large enough for analytical purpose from a dry soil with
a hydraulic potential value lower than -60 kPa (Saragoni et al. 1990); 3) the porous cups
were clogged after a series of sampling cycles (Morisson and Lowery 1990, Grover and
Lambom 1970).
The ceramic cup of tensiometers can be used to capture solutes from soil
systems as well as measure soil matric potential. Based on the ionic equilibrium
between the soil solution and the tensiometer solution, Moutonnet et al. (1993) modified
the tensiometer and used it to measure soil hydraulic potential, and then pumped
solution from the tensiometer to measure nitrate. Finally, the solution was returned to
the tensiometer for later measurement. They found that it required 8 days to establish
an ionic equilibrium between the tensiometer and the soil solution in a silt loam soil.
Their measurements with the solution from the tensiometers correlated with the amount
LI AND REDDY ON USE OF TENSIOMETER 373
of nitrate that had been consumed by the crop. During the growing season there was a
minimal risk for contamination of water table by nitrate.
In this study, we filled a tensiometer with 1 M ammonium sulfate ((NH4)2804)
solution as an ionic strength adjuster (ISA) instead of pure water. After the
establishment of a pressure and ionic equilibrium between the tensiometer and the soil
solution, a pressure transducer was used to measure the soil hydraulic potential. Then, a
nitrate electrode was directly inserted into the tensiometer to measure the nitrate
concentration. The 1 M (NH4)2SO4 solution maintains a desirable salt concentration in
the tensiometer during the ionic exchange process (Mulvaney 1996, Greenberg et at.
1992). The objective of this study is to use tensiometers to monitor nitrate movement in
soil profiles as well as to measure soil hydraulic potential as its original function. A
field experiment was conducted in a soil with a seedless watermelon crop.
Materials and Methods
Tensiometers were assembled in the laboratory using a porous ceramic cup

(O.D. = 2.223 cm), a piece of schedule 80 PVC tube (O.D. = 2.223 cm) and a piece of
clear acrylic tube (O.D. = 1.59 cm). Various lengths of the PVC tube allowed the
tensiometers to be installed at soil depths of interest. The unit was then filled with ISA
solution (1 M (NH4)2SO4). Finally, the open end of the acrylic tube was sealed by a
rubber septum stopper.
Laboratory Calibration of the Tensiometers
Eight 1-L polypropylene bottles were filled with pure sand (diam. < 0.5 mm). A
tensiometer with a length of 30 cm was inserted into the sand 2.5 cm above the bottom
of each bottle. The bottles were then placed on a mechanical shaker to ensure a good
contact between the porous cups and the sand. Each bottle was treated with 250 ml of
solution with concentrations of 2, 10, 50, or 100 mg L % f NO3--N with a duplicate,
respectively. The bottles were capped by lids with holes for the tensiometers to go
through. Then, the holes were sealed with Teflon tape. Thus, the bottles were closed
systems and able to maintain a stable hydraulic pressure for the calibration period.
Measurement was started from the first day of the installation. A portable
pressure transducer (Marthaler et al. 1983) equipped with a hypodermic needle was used
to measure the hydraulic potential in the bottles. By removing the septum stopper, a
nitrate electrode was inserted into the solution through the open end of the acrylic tube.
A portable Orion 710 pH/ISE meter was connected to the electrode. The mitlivolt
readings of the concentrations of NO3-N in the tensiometers were read with the meter.
Each reading was plotted against its known concentration of NO3"-N to yield a
regression coefficient and a R 2. A stable regression coefficient was found when an ionic
equilibrium was established between the tensiometer solution and the sand solution.
Measurements were taken for 30 days. The meter and the measurement were capable of
being controlled by a computer.
Field Insitu Measurement of NO3-N and Hydraulic Potential
A field experiment was conducted at a site with a sustainable agricultural

cropping system for seedless watermelon grown in rotation with green manure. The site
is located at the Environmental Studies Center of North Carolina A&T State University.
The soil was identified as an Enott-Like sandy loam (fine, mixed, thermic typic
hapludalf) by the North Carolina Soil Survey Staff before the experiment was initiated.
This survey also described the properties of the soil profile prior to the experiment
(Table 1). The soil and landscape at this site represent many areas of the Piedmont
region of southeastern USA.
Table 1-- Soil Horizon, Depth, Texture, pH and

Organic Carbon Content of the Experimental Site.
HORIZON DEPTH (cm) TEXTURE pH ORGANIC CARBON (%)
AP 0 - 23 Sandy Loam 5.8 1.82

Bt 23-48 Clay 6.7 0.35
BC 48-64 S. Clay Loam 6.7 0.11
C 64-107 Sandy Loam 6.9 0.04
Cr 107-135 Sandy Loam 7.2 0.04
At the experimental site cover crops were planted in the fall of 1996 (Austrian
winter peas, hairy vetch, and rye) then mowed by a flail mower on May 20, 1997 to
leave the biomass as a uniform layer of mulch on the surface of the soil. The purpose of
the mulch was to supply nutrients to the watermelon crop, conserve soil moisture, and
control weeds and soil erosion. Tensiometers were installed in soil at the depths of 20,
40, 60, 80 and 100 cm with five replications. Seedlings of watermelon were
transplanted on June 4, 1997. After a period of time for the tensiometer's equilibration,
first measurements of hydraulic potential and nitrate concentration were taken in July
1997. A calibrated PH/ISE meter was used to measure NO3"-N concentrations in the
tensiometers at different depths. The measurements were taken each week.
Results and Discussion
Laboratory Calibration
The measured hydraulic potential remained almost a constant (-5.4 kPa) in all
sand bottles during the period of the calibration (curve 4 in Figure 1). The regression
analysis of each day resulted in a regression coefficient, an intercept and a R 2. In the
initial stage (the first day) of the calibration (curve 1 of Figure 1), nitrate ions just
started to migrate to the tensiometer through pores of the ceramic cup. The mV
readings from the pH/ISE meter did not respond to the known concentrations of the
NO3"-Nin the sand solution. Curve 2 in Figure 1 shows the mV readings against the
logarithm nitrate concentrations on the sixth day of the calibration. Although it was
non-linear, a response was detected by the pH/ISE meter. The first linear response
occurred on the 12th day with a regression coefficient of 39.7 mV (log(mg L'I))l, an
intercept of 265 mV (curve 3, Figure 1) and a R2 = 0.9986; However, the regression
parameters were not stable at this time.
Figure 1 - - Plot of the pH/ISE Meter Readings as a Function of the Known NO3--N
Concentrations Approaching the Ionic Equilibrium.
The time for tensiometers to reach an ionic equilibrium between the ISA
solution inside the tensiometers and the sand solution was determined by regression
analysis between the pH/ISE readings and the known concentrations of NO3"-N. As the
calibration progressed with time, both the regression coefficient and the intercept
approached their constants (Figure 2). Fifteen days were required for the ISA solution
in the tensiometers to reach the nitrate equilibrium with the sand solution. The
equilibrated regression coefficient was between -47 and -45 mV (log(rag L~))"~ with a
standard deviation of 0.53 mV (log(mg L'~)) "~. The intercept was between 280 and 290
mV. Its standard deviation was 4.2 inV. The regression analysis had a R2=0.9990.
Figure 2 --Slopes and Intercepts of the Regression Analysis between the pH/ISE
Readings and the Known NO3--NConcentrations.
Field Experiment
The measurements of hydraulic potentials and NOf-N concentrations were

considered as time series data. These data were analyzed by statistical method of time
series estimate. Autocorrelation coefficients for hydraulic potential and NO3-N
concentration were not significant and so general statistics and graphs are used in the
following discussion. In July 1997 at the early stage of the watermelon growth, the
surface soil (top 20 cm) maintained a high hydraulic potential. The moisture distribution
over the entire profile was relatively uniform. As the demand by plant growth for water
increased with growing season, hydraulic potential at the depths of 20 and 40 cm
decreased rapidly because watermelon growth utilized the moisture mostly from the top
40 cm of the soil (Figure 3). The soil also supplied some moisture for the watermelon
growth from as deep as 60 cm. Moisture status at depths of 80 and 100 cm fluctuated
only slightly during this time period. When the hydraulic potential of the top 20 cm soil
decreased to -60 kPa, the plots were irrigated on August 20, 1997. Also, occasional rain
events replenished the top 20 cm soil with water. In general, insignificant amount of
water from rainfall or irrigation percolated below a depth of 60 cm. Soil moisture
below 60 cm depth remained stationary until the harvest of the watermelons.
Figure 3 m The Fluctuation of Soil Hydraulic Potential at Different Depths over

Time and the Rainfall Data.
Data in Figure 4 indicate that release of NO3-N from the mulch composition was
relatively low at the initial stage of watermelon growth (before July 29). The soil
maintained a low level of nitrate concentration (< 10 mg L'~). As the decomposition
started to release NO3-N, a rapid accumulation of NO3-N occurred in the soil solution.
This accumulation caused increases of NO3-N over the entire 100-cm profile until
watermelon plants started their rapid consumption of nutrients. Most of NO3"-N for
watermelon growth was from top 20 cm soil, some from 20 to 40 cm depths. The
watermelon rarely used nutrients from 60 cm depth and below. However, a decrease of
NO3-N in soil solution below 60 cm depth happened after August 8. This implied a
tendency of NO3"-N leaching. Then a steady accumulation of NO3"-N in the soil
solution below 60 cm depth was observed. After the watermelons were ripe and
harvested (September 10), decomposition of mulch was still in progress. NO3"-N in the
soil profile resumed its pace of accumulation and leaching to deeper soil. The content
of NO3-N was 15 mg L "~at 100 cm depth. It indicated a potential risk of leaching when
crop consumption could not keep up with the release of NO3"-N from the mulch
decomposition. The results also show that each rainfall event caused sudden decreases
of NO3"-N in soil solution at depths of 20 to 60 cm due to dilution (Figure 4). But rain
did not percolate to soil depths below 60 cm during the growing season.
Figure 4 - - The Occurrences of NO:-N in Soil Solution during the Progress of Mulch
Decomposition and the Nutrient Uptake by Watermelon Growth.
The nitrate concentrations along the soil profile at four stages of the growing
season were presented in Figure 5. When the mulch decomposition and the watermelon
growth were at their initial stages (curve 1, Figure 5), the overall nitrate concentration in
soil solution was low. As the depth increased, the nitrate concentration decreased.
Then, mulch decomposition increased, but watermelon requirement for nutrients did not
increase significantly. There was a high accumulation of NO3"-N in the topsoil solution
when mulch decomposition started to release NO3"-N and watermelon growth did not
have a great demand for nutrients (curve 2, Figure 5). When watermelon growth
utilized NO3"-N mostly from the top 20-em soil, NO3"-N concentration at the 40-cm
depth was higher than that at the 20-cm depth (curve 3, Figure 5). After watermelons
were ripe (September 10), NO3"-Nstarted to build up in the soil profile because the
decomposition of the mulch was still in progress. Thus, nitrate leaching to deeper soil
horizons increased the NO3"-N concentrations to the soil depth of 100 cm, significantly
(curve 4, Figure 5).
Figure 5 - - Typical NO3"-N Profiles at Various Phases of Mulch Decomposition

and Watermelon Growth.
Conclusion
In this study, the tensiometers were filled with ISA solution instead of pure
water in order to measure hydraulic potential and nitrate concentration in soil solution
simultaneously and insitu. The tensiometer can be installed at any depth of interest.
The time to reach an ionic equilibrium between the tensiometers and the test materials
was 15 days. Results of this study showed the progressive drying of the soil due to
deficit of irrigation and the consumption of the N nutrient by the watermelon crop. The
nitrate measurement from 20 to 40 cm depths confirmed that watermelon growth
consumed nutrients from these depths. Rainfall water only percolated the soil to a depth
of about 60 cm. However, the later redistribution of nitrate with water can induce a
potential risk of nitrate leaching when the nutrient requirement of the crop is low.
The advantage of this method of nitrate measurement is its rapidity and non-
destructivity. With the electrode connected to a computer, this method is appropriate to
obtain continuous readings of nitrate concentration in deeper soil horizons and water
tables. When different ion selective electrodes are chosen, ISA solutions are available
for other ions of interest. Thus, this method can also be used to measure different ionic
contaminants in soil solution, subsurface soil or ground water.
REFERENCES
Bernhard, C., and Schenck, C., 1986, "Utilization des bougies poresues pour extraire la
solution du sol dans le ried central de 1'Ill en Alsace," Bull. Groupe Ft.
Humidimetrie Neutrinique, 20:73-85.
Canter, L. W., 1997, "Nitrate in groundwater," University of Oklahoma. Lewis
Publishers. Boca Raton, New York, London, Tokyo.
Cassel, D. K., and Klute, A., 1986, "Water potential: Tensiometry," Method of Soil
Analysis. Part 1. ASA, CSSA, and SSSA. Madison, WI., pp. 563-596.
Causoin, R. E., Murphy, J. A., and Branham, B. E., 1990, "A vertical installed, flush-
mounted tensiometer for turfgrass research," HortScience, 25(8): 928-929.
Fraser, P. M., Cameron, K. C., and Sherlock, R. R., 1994, "Lysimeter study of the fate
of nitrogen in animal urine returns to irrigated pasture," European Journal of
Soil Sci., 45(4):439.
Greenberg, A. E., Clesceri, L. S., Eaton, A. D., and Franson, M. A. H., 1992, "Standard
methods for the examination of water and wastewater," APHA, AWWA, WDFF
Publ. APHA, Washington, D. C.
Grover, B. L., and Lambom, R. E., 1970, "Preparation of porous ceramic cups to be
used for extraction of soil water having low solute concentrations," Soil Sci.
Soc. Am. Proc., 34:706-708.
Heinrichs, H., Bottcher, G., and Pohlmann, M., 1996, "Squeezed soil-pore solutes -- A
comparison to lysimeter samples and percolation experiments." Water, Air, and
Soil Pollution, 89(1/2): 189.
Hempel, M., Wilken, R. D., and Beyer, K., 1995, "Mercury contaminated sites -
behavior of mercury and its species in lysimeter experiments," Water, Air, and
Soil Pollution, 80(1/4): 1089.
Hendrickx, J. M. H., Wierenga, P. J., and Nash, M. S., 1990, "Variability of soil tension
and soil water content," Agric. Water Manage., 18:135-148.
Keeney, D., 1986, "Sources of nitrate to groundwater," CRC critical reviews in
environmental control, Vol. 16, No. 3, pp. 257-304.
Li, K., Amoozegar, A., Robarge, W. P., and Buol, S. W., 1997, "Water movement and
solute transport through saprolite," Soil Sci. Soc. Am. J., 61:1738-1745.
Mackay, A. D., Sakadevan, K., and Hedley, M. J., 1994, "An insitu mini lysimeter with
a removable ion exchange trap for measuring nutrient losses by leaching from
grazed pastures," Australian Journal of Soil Research, 32(6): 1389.
Mathaler, H.P., Vogelsanger, W., Richard, F., and Wierenga, P.J., 1983, "A pressure
transducer for field tensiometers," Soil Sci. Soc. Am. J., 47:624-627.
Morrison, R. D., and Lowery, B., 1990, "Effect of cup properties, sampler geometry,
and vacuum on the sampling rate of porous cup samplers," Soil Sci., 140:308-
316.
Moutnnet, P., Pagenel, J. F., and Fardeau, J. C., 1993, "Simultaneous field measurement
of nitrate-nitrogen and matric pressure head," Soil Sci. Soc. Am. J., 57:1458-
1462.
Moyer, J., Saporito, W., Janke, L. S., L.S., and R.honda, R., 1996, "Design,
construction, and installation of an intact soil core lysimeter," Agronomy
Journal, 88(2): 253.
Mulvaney, R. L., 1996, "Nitrogen - inorganic forms," Method of Soil Analysis. Part 3..
ASA, CSSA, and SSSA. Madison, WI., pp.1123-1184.
Office of Technology Assessment, 1990, "Beneath the bottom line: Agriculture
approaches to reduced agricultural contamination of groundwater," OTA-F-
418. U.S. Congress. U.S. Government Printing Office. Washington, D. C., pp.
3-20.
Ohte, N., Tokuchi, N., and Suzuki, M., 1997, "An insitu lysimeter experiment on soil
moisture influence on inorganic nitrogen discharge from forest soil," Journal of
Hydrology, 195: 78.
Peterson, D. L., Glenn, D. M., and Wolford, S. D., 1993, "Tensiometer-irrigation
control valve," Applied Engineering in Agriculture, 9(3): 293-297.
Saragoni, H., Poss, R., and Oliver, R., 1990, "Dynamique et lixiviation des elements
mineraux dans les terres de barre du sud du Togo," Agron. Trop. (Paris),
45:259-273.
Smajstrla, A.G., and Locascio, S. J., 1996, "Tensiometer-controlled, drip-irrigation
scheduling of tomato," Applied Engineering in Agriculture, 12(3): 315-319.
Vanclooster, M., Mallants, D., Vanderborght, J., Diets, J., Van Orshoven, J., and Feyen,
J., 1995, "Monitoring solute transport in a multi-layered sandy lysimeter using
time domain reflectometry," Soil Sci. Soc. Am. J., 59: 337-344.
R.A. Erchul1
Compaction Comparison Testing Using a Modified Impact Soil Tester And Nuclear
Density Gauge
REFERENCE: Erchul, R. A., "Compaction Comparison Testing Using a Modified

Impact Soil Tester and Nuclear Density Gauge," STP 1358, Field Instrumentation for
Soil and Rock, American Society for Testing and Materials, West Conshohocken, PA,
1999.
ABSTRACT: The purpose of this paper is to compare test results of a modified Impact
Soil Tester (IST) on compacted soil with data obtained from the same soil using a nuclear
density gauge at the U.S. Army Corp of Engineer's Buena Vista Flood Wall project in
Buena Vista, Virginia. The tests were run during construction of the earth flood wall
during the summer of 1996. This comparison testing demonstrated the credibility of the
procedure developed for the IST as a compaction testing device. The comparison data
was obtained on a variety of soils ranging from silty sands to clays. The Flood Wall
comparison compaction data for 90% Standard Proctor shows that the results of the IST
as modified are consistent with the nuclear density gauge 89 percent of the time for all
types of soil tested. However, if the soils are more cohesive then the results are consistent
with the nuclear density gauge 97 percent of the time. In addition these comparison tests
are in general agreement with comparison compaction testing using the same testing
techniques and methods on compacted backfill in utility trenches conducted earlier for the
Public Works Department, Chesterfield County, Virginia.
KEYWORDS: compaction testing, impact soil tester, nuclear density gage, comparison
testing
Introduction
The Impact Soil Tester (IST) was developed by Dr. Baden Clegg in Australia
during the 1970s (Clegg 1978). The apparatus is commonly known as the Clegg Hammer
of the Clegg Impact Soil Tester while the test is known as the Clegg Impact Test (CIT) or
simply Impact Test. Although predominantly developed as an in situ soil strength test, the
Impact Test has also been more recently used in the United Kingdom for the testing of
backfill in utility trenches to determine if this material has been properly compacted
lprofessor, Civil & Environmental Engineering Department, Virginia Military Institute,

Lexington, Virginia, USA.
ERCHUL ON COMPACTION COMPARISON TESTING 383
(Winter & Selby, 1991). The advantage of using the IST in compaction o f utility trench
backfill is its rapidity in taking a test and ease of operation.
The components of the IST are a guide tube, instrumented drop hammer and
electronic readout (See Figure 1). Recent IST models have been available with the
readout on the handles of the hammer, eliminating the meter box and cable as shown in
Figure 1. The IST measures in units of Impact Value (IV) which is based on the peak
deceleration of the 4.54kg (10 Ibm) drop hammer measured in units of 10 gravities. The
IV is measured by an internal aecelerometer located in the tester and digitally displayed on
the readout. The IST can be used on any type of soil and comes in four hammer sizes: 0.5
kg (1.1 Ibm), 2.25 kg (5 Ibm), 4.54 kg (10 Ibm) and 20 kg (44 Ibm). The choice of
hammer size is a function of strength of soil, the desired depth of soil to be tested and, to a
certain degree, the soil material types. In compaction testing the 4.54 kg (10 Ibm) is most
often used since it has an effective testing depth of approximately 150 mm (6 in) and its
Impact Value scale of 0 to 100 covers the full range of materials encountered up to a CBR
of about 70% (Clegg, 1983).
The operation procedure for using the IST is relatively straightforward, rapid and
easily conducted. The operator places at least one foot on the edge of the IST guide tube
flange to steady it during the test. The button on the readout box is depressed while the
4.54 kg (10 Ibm) hammer is raised 457 mm (18 in.) and allowed to free fall. The readout
button is held depressed while this step is repeated four consecutive times. With more
recent models, the button is pressed and released and the four drops carded out while the
circuit remains activated on a timer. In either case, the maximum value of the four blows
remains displayed on the readout. Figure 1 shows an IST test in progress. The reading
displayed after the fourth blow is commonly taken as the Impact Value and recorded.
The IST provides a strength parameter; it does not measure soil density directly.
By understanding how soil strength relates to the compaction process, Impact Value may
be used to indicate if compacted soils comply with the desired density specifications. The
standard test method and additional information for determination of the IV of a soil can
be found in ASTM D 5874-95. In addition, a modified testing procedure of measuring the
accumulated depth of penetration of the hammer at the fourth blow was also employed
during the comparison compaction testing. Testing with the IST using this modification is
explained in this paper.
A Modified Testing Procedure
Research into the use of the IST has been ongoing at the Virginia Military Institute
for several years. Both laboratory and field tests were conducted in conjunction with
Chesterfield County, VA, Department &Utilities on utility trenches, focusing on the use
of the IST in determining adequacy of trench backfill compaction. Specifically, the object
of the work had been on the use of the IST in testing the backfill of utility trenches in lieu
of other more expensive techniques, such as a nuclear density gauge.
Erchul and Meade (1990) and Erchul, et al. (1994) initially investigated the IST's
effectiveness for compaction on various soils in the laboratory and field. They observed
that an inverse non-linear relationship existed between the IV and the accumulated depth
Figure 1. Clegg Hammer
ERCHUL ON COMPACTION COMPARISONTESTING 385
of the hammer penetration into the soil. It was thought that this relationship defined the
energy curve developing from the impact of the hammer into the soil. They identified the
relationship between the IV and the accumulated depth of penetration at the fourth blow
(d) for the 4.54 kg (10 Ibm) Impact Test Hammer as being:
IV = 60 / d (1)
where the depth (d) is measured in units of 2.54 mm (0.1 in).

Erchul and Meade (1990) developed a modified testing procedure for the IST that
might permit a rapid prediction of compaction level or energy imparted into the soil in
regard to attaining maximum density. For the modified testing procedure, a numerical
scale of 0-20 in. depth units of 2.54 mm (0.1 in.) was added on the hammer shaft in order
to measure depth of hammer penetration. Two acceptance zones were developed from
previous field and laboratory testing data. These zones, Zone 90 Acceptance Criteria and
Zone 95 Acceptance Criteria (See Figures 2 and 3) show what was considered acceptable
combinations oflVdepth results for 90% and 95% Standard Proctor levels. In addition,
statistical acceptance criterion developed for a set of eight tests on one lot of soil whereby
5 of 8 test results had to fall within the acceptance zone. This was based on a 10% risk of
rejecting suitably compacted material.
Buena Vista Flood Wall Tests
Using ASTM D 5874-95 and the above mentioned methodology, six different
areas on the Buena Vista flood wall project were tested multiple times for a total of 46
trials run. Each of the six areas contained a different type of soil ranging from CH (high
plastic inorganic clays) to SM (silty sands). Initially a nuclear density gauge reading was
taken at a specific location after compaction was conducted along a section of the flood
wall. For comparison, immediately after this reading was taken the IST was used at eight
positions at cardinal and inter-cardinal headings around the nuclear gauge location,
approximately one foot away in each direction. The specifications called for 90%
Standard Proctor for acceptance. The set of eight readings of IV and depth data were
plotted on a Zone 90 Acceptance Criteria chart for each location. If more than three of
the eight tests at each location fell outside the zone of acceptance, the soil was considered
as failing to meet the specification. Conversely, if five or more results fell inside the zone,
the compacted soil was considered passable at the 90% Standard Proctor compaction
level.
The nuclear density gauge data were analyzed to determine water content and dry
density at each location. The nuclear gauge results were compared with Standard Proctor
data obtained in the laboratory on the same type of soil to determine whether the
compacted soil passed or failed specifications according to the nuclear density gauge
method.
The IST test results were then compared with the nuclear density gauge results to
determine their agreement with the compaction of the test locations.
Penetration Depth versus Inpact Value

A
.__.15
L_
0
E
EIO
N
||
C
" 5
0.
a~ 0
0 5 10 15 20 25 30
Impact Value (IV)
ERCHUL ON COMPACTIONCOMPARISON TESTING 387
Penetration Depth versus Inpact Value

Figure 3 : Zone 95
._r 15
t_
0
E
EIO
N
||
e-
= 5
~ 0
0 5 10 15 20 25 30
Impact Value (IV)
388 F I E L D I N S T R U M E N T A T I O N FOR S O I L A N D R O C K
Results
TABLE 1
RESULTS OF COMPARISON TESTS USING THE
IMPACT SOIL TESTER (IST) AND THE
NUCLEAR DENSITY GAUGE
A=ACCEPTED
R=REJECTED
TEST SITE SOIL TYPE IST GAUGE TEST

1 PROCTOR 29 CH A A 1
10 PROCTOR 39 SC A A 10
14 PROCTOR 41 SC-SM A R 14
15 PROCTOR 41 SC-SM A A 15
16 PROCTOR 41 SC-SM R A 16
19 PROCTOR 41 8C-SM A R 19
21 PROCTOR 41 8C-SM A R 21
22 PROCTOR 41 MH A A 22
30 PROCTOR 44 MH R A 30
39 PROCTOR 57 ~ ML A A 39
40 PROCTOR 57 ML A A 40
ERCHUL ON COMPACTION COMPARISONTESTING 389
All of the Zone 90 Acceptance Criteria and nuclear density gauge results are
shown in Table 1 in terms of acceptance (A) or rejection (R) according to passing 90%
Standard Proctor. Proctors 29, 33, 39, 41, 44 and 57 represent laboratory compaction
tests for the six areas tested. Of the 46 trials run, five trials (11%) had Zone 90
Acceptance Criteria results that did not agree with the nuclear density gauge results.
These data, though limited, imply that the Zone 90 Acceptance Criteria method has
approximately an 89% consistency rate compared with the nuclear density gauge for all of
the soil types tested. The following points need to be highlighted regarding the
consistency rate:
oThe nuclear density gauge was considered the standard

(i.e. there was no other independent test to check the
gauge),
oOnly a limited amount of data were taken, and
oRetesting using the IST or nuclear density gauge at each
site was not conducted due to contractor time constraints
when test results were inconsistent.
The second purpose of this paper was to analyze the inconsistent data and propose
some reasons for the differences. To do this, optimum moisture content, penetration
depth, maximum dry density and Impact Values were studied for each trial and soil type.
On examining these data only one trend was noted. Of these six areas, laboratory test
Proctors 29, 33, 39 and 44 were obtained from areas consisting mostly of cohesive soils
(CH, MH, SC). The other areas, Proctor 41 and 57, were mostly noncohesive soils (SC-
SM, ML). Of the seventeen trials run on the mostly noncohesive soils, four results
disagreed. This means that 76% of the trials run on the mostly noncohesive soils had
Zone 90 Acceptance Criteria results that were consistent with the nuclear density gauge
data. On the other hand, only one of twenty-nine trials run on the mostly cohesive soils
disagreed with the nuclear density gauge making it consistent 97% of the time.
It is also important to comment on the data comparison results of the IST as
modified and nuclear density gauge in earlier tests conducted in Chesterfield County, VA
(Erchul et al 1994). During this testing, the Zone 95 Acceptance Criteria were used in 18
different areas in the county on a variety of soils. These data were compared to nuclear
density gauge data from seven of the areas along with data obtained with another field
testing procedure: the Standard One Point Proctor Test (ASTM D 698 - Method C).
During this testing, the Zone 95 Acceptance Criteria results were consistent with the
nuclear gauge 86% of the time, taking into account that the one marginal result returned
by the Zone 95 Acceptance Criteria method would be considered as a rejection and be
cause for retesting or recompacting and retesting. These results are consistent with the
85% agreement rate of the Zone 90 Acceptance Criteria data taken at the Buena Vista
flood wall. The results using the Standard One Point Proctor Test were 71% in agreement
with the nuclear density gauge in the Chesterfield County trials. In addition it must be
commented that in these comparison tests, the nuclear density gauge was considered as
the standard against which the other methods were compared.
Summary
The Impact Soil Tester (IST) in these tests has been shown to be a safe, quick,
easy to operate, cost-effective way to test the acceptability of compacted soils. It is fairly
light and has few parts, which makes it very mobile and practical for on the job use. The
few parts of the IST also allow it to be set up and operated quickly and easily. With these
advantages, the IST requires only minimal training and skill to operate. Compared to
other soil compaction tests, the IST is cost effective because it involves little maintenance
and provides easily obtainable data at a reasonable cost. Handling the instrument requires
no special regulatory requirements or operator certificate. The IST gives direct and
immediate results with no calculations required to determine the strength parameter. By
using the Zone 90 Acceptance Criteria developed in earlier research as the basis for this
comparison, this meant that a soil sample did not have to be taken back to the laboratory
for further analysis when using the IST as modified as an acceptance tool regarding
compacted soil passing or failing 90% Standard Proctor.
The comparison test procedures used for this paper show that the consistency of
the Zone 90 Acceptance Criteria varies with the types of soils tested such as cohesive
clays and highly plastic silts versus noncohesive soils. The Zone 90 Acceptance Criteria
were more effective in testing cohesive soils when compared to a nuclear density gauge at
97% agreement but agreement decreased to 76% in noncohesive soils. Further testing
using the Zone 90 Acceptance Criteria method and nuclear density gauge in noncohesive
soils is recommended.
Using the modified IST procedure produced fairly reliable and compatible data in
terms of Impact Values and penetration depths which were in agreement with the
relationship developed in Equation 1 and other comparison tests conducted using the
method on various types of soils. As a final comment, the eight readings required in the
modified IST testing procedure allow for greater sampling which provides better results
on a statistical basis regarding the testing of a lot after compaction.
References
Clegg, B., 1978, "An Impact Soil Tester for Low Cost Roads," Proceedings, 2nd
Conference Road Engineering Association of Asia & Australia, pp. 58-65, Manila,
Philippines.
Clegg, B., 1983, "Application of an Impact Test to Field Evaluation of Marginal Base
Course Materials," Proceedings, Third International Conference Low Volume Roads,
Arizona, Transport Research Record, 898, pp. 174-181, TRB, Washington, D.C., USA.
Erchul, R. A. and Meade, R. B., 1990, "Using the Modified Clegg Impact Hammer to
Evaluate Adequacy of Compaction," prepared for the Department of Utilities,
Chesterfield County, Virginia, USA.
ERCHUL ON COMPACTIONCOMPARISONTESTING 391
Erchul, R. A., Decker, R. A., and Ackerman, P., 1994, "Comparison Results of Field
Data Using 4.5 kg and 20 kg Clegg Impact Tester (CIT) with Other Standard Compaction
Tests on Chesterfield County Soils," prepared for Department of Utilities, Chesterfield
County, Virginia, USA.
Winter, M. G. and Selby, A. R., June 1991, "Clegg Meter Performance Assessment With
Reference to the Reinstatement Environment," Highway and Transportation.
1 . 2 3
Wayne Saunders, Richard Benson, Frank Snelgrove, and Susan Soloyanis4
Selecting Surface Geophysical Methods for Geological, Hydrological, Geotechnical,

and Environmental Investigations: The Rationale for the ASTM Provisional Guide
REFERENCE: Saunders, W., Benson, R., Snelgrove, F., and Soloyanis, S., "Selecting
Surface Geophysical Methods for Geological, Hydrological, Geoteehnical, and
Environmental Investigations: The Rationale for the ASTM Provisional Guide,"
Marr, Eds., American Society for Testing and Materials, West Conshohocken, PA, 1999.
ABSTRACT: The ASTM Provisional Guide (PS 78-97) for Selecting Surface
Geophysical Methods was developed as a guide for project managers, contractors,
geologists, and geophysicists to assist in selecting the most likely geophysical method or
methods to conduct specific subsurface investigations. Numerous surface geophysical
methods and techniques exist that can be used to determine subsurface soil and rock
properties and their distribution. These same methods are also widely used to investigate
and locate manmade structures such as buried objects and landfills. This paper discusses
the general uses of surface geophysics and the use of the provisional guide. This paper is
not intended to be used as the guide.
The ASTM Provisional Guide provides direction in selecting the most appropriate
geophysical method or methods for a specific application under general site conditions.
Secondary methods are also proposed that, under certain circumstances, should be
evaluated before a final selection is made. Some typical conditions under which a
primary or secondary method might or might not provide satisfactory results are given in
the provisional guide. References for further information about selected methods and to
method-specific ASTM guides are also provided. Secondary methods usually have less
than desired performance, higher cost, or greater labor requirements as compared to the
primary methods.
KEYWORDS: surface geophysics, site characterization
RATIONALE FOR THE ASTM GUIDE

The ASTM provisional guide (PS 78-97) for selecting surface geophysical
methods was developed as a reference for project managers, contractors, geologists and
geophysicists to assist in selecting the most likely geophysical method(s) for geological,
1Senior Geophysicist, SAIC, 11251 Roger Bacon Drive, Reston, VA 20190.

2
President, Technos, Inc., 3333 Northwest 2 l~t Street, Miami, FL 33142.
3
Vice President, Geonics Limited, 1745 Meyerside Drive, Unit 8, Mississauga, Ontario, Canada L5T 1C6,
4
Principal Scientist, Mitretek Systems, 4610 Fox Road, Cascade, CO 80809.
SAUNDERS ET AL. ON SELECTING METHODS 393
hydrological, geotechnical, and environmental subsurface investigations. It was

developed by a small group of practicing geophysicists and approved by consensus
ASTM vote. The purpose of the guide is to direct program managers to the generally
acceptable surface geophysical methods for solving specific problems under simple or
ideal site conditions. The benefits and limitations of specific surface geophysical
methods are discussed in method-specific standards that have either been published or are
under development. ASTM D 5777, Guide for Using the Seismic Refraction Method for
Subsurface Investigations, is available as a published standard. Guides for ground
penetrating radar, gravity, and resistivity methods are currently being balloted and should
be available in late 1999. Guides for electromagnetic methods and seismic reflection are
under development.
This paper will discuss the general uses of surface geophysics and the use of the
provisional guide. Two examples, determining depth to bedrock and mapping a
contaminant plume, that illustrate the use of the guide are provided.
SURFACE GEOPHYSICAL METHODS
Surface geophysical methods provide a means of mapping lateral and vertical

variations in one or more physical properties of the subsurface. They can also be used to
monitor temporal changes in these properties. Each geophysical method measures or
responds to a specific physical, electrical, or chemical property of the soil, rock, pore
fluids, or subsurface structure. Therefore, for a geophysical method to detect geologic
conditions, buried structures, or contaminants there must be a sufficient variation in the
property being measured for the condition or structure of interest to be detected. A
contrast must be present in order for geophysical measurements to be meaningful. For
example, the contact between fresh water and saltwater can be detected by electrical
methods, which measure electrical properties. The contact between unconsolidated
material (soil) and unweathered rock can be detected by seismic methods, which measure
acoustical velocities. On the other hand, if a geophysical method is being using to detect
a small change in water quality or to delineate the top of weathered rock whose properties
are similar to the soil overlying it, it may not be successful. This is because the contrast
in measured properties is likely to be small or gradational and therefore may not be
detected by geophysical measurements.
Advantages of Surface GeophysicalMethods

Surface geophysical measurements generally can be made quickly, are minimally
intrusive, and enable interpolation between known points of control. Since a primary
factor affecting the accuracy of geotechnical or environmental site characterization efforts
is the number of sample points or borings, geophysical methods are commonly used to
map large areas, locate specific targets, or to provide a laterally continuous image of the
subsurface.
Types of Surface GeophysicalSurveys

Geophysical measurements are commonly made by either profiling or sounding
measurements. Profiling, either station by station or continuously, provides a means of
assessing lateral changes in subsurface conditions at a given depth. Soundings provide a
means of assessing depth and thicknesses of geologic layers or specific targets at a given
location. Most surface geophysical sounding measurements can resolve three layers.
Products of GeophysicalSurveys
Some surface geophysical methods provide data from which a preliminary
interpretation can be made in the field, for example, ground penetrating radar, frequency
domain electromagnetic (EM) profiling, DC resistivity profiling, magnetic profiling, and
metal detector profiling. A map of radar anomalies or a contour map of the EM,
resistivity, magnetics, or metal detector data can often be created in the field and used to
locate anomalous conditions. Some methods (for example, time domain electromagnetics
and DC resistivity soundings, seismic refraction, seismic reflection, and gravity
measurements) require that the data be processed before any quantitative interpretation
can be made.
Lateral and vertical boundaries determined by surface geophysical methods
usually coincide with geological boundaries, and a cross-section produced from
geophysical data may resemble a geological cross-section, although the two are not
necessarily identical.
PROCESS FOR SELECTING SURFACE GEOPHYSICAL METHODS
The guide provides a table (Figure 1) for the selection of commonly used surface
geophysical methods in four broad application areas:
Natural geologic and hydrologic conditions - This includes assessing soil and
unconsolidated layers, rock layers, depth to bedrock, depth to water table, fractures
and fault zones, voids and sinkholes, soil and rock properties, and dam and lagoon
leakage.
9 Inorganic contaminants - This includes detecting and mapping inorganic plumes from
landfills, saltwater intrusion, and soil salinity.
9 Organic contaminants - This includes detecting and mapping light nonaqueous phase
liquids, dissolved phase, and dense nonaqueous phase liquids.
Man-made buried objects - This includes locating utilities, drums and USTs, UXO,
abandoned wells, landfill and trench boundaries, forensic applications, and
archeological features.
The surface geophysical methods considered in the table are those that are
commonly used for geotechnical, hydrologic, or environmental applications and include:
seismic refraction and reflection, DC resistivity, induced polarization (IP), spontaneous
potential (SP), frequency and time domain electromagnetics, very low frequency (VLF)
electromagnetics, pipe and cable locators, metal detectors, ground-penetrating radar,
magnetics, and gravity. To use the table, identify the specific purpose of the
investigation and examine the A- and B-rated methods for that purpose. A rating of"A"
implies a primary choice of method. A rating of B implies a secondary or alternate
choice of method. The methods must be further evaluated for use under site-specific
conditions. The guide provides a brief discussion of each of the methods. References are
also provided for further reading.
The process for selecting a surface geophysical method is illustrated in Figure 2
and is conducted as follows:
GEOPHYSICALMETHODS
APPLICATIONS
,p, C~ : e q ~ : natic Pip,able Metal Penetrating
Refrac~on t R e f i ~ a i n I"T=meDomain ]VLF [Locator Detectors Magne~cs
Natural Geologic and Hydrologic
Conditions
Soil layers A B A B A B A
Rock layers e ~ B
Depth to bedrock A B B B B A B
Depth to water table A A B B B B A
Fractures and fault zones B B B A B A A B B
Voids and sinkholes B B A A A
Soil and rock properties A A A
Dam and lagoon leakage B A B B
Inorganic Contaminants C
Landfill leachate A A A B B Z
Saltwater inVu~;ion m
Soil salir.ty
Organic Contaminants
m
Light, ~ q t / e o u s phase liquids
Dissolved p~ase*
Dense, nonaqueous phase liquids*
t-
Manmade Buried Objects O
z
Utilities B A B A
Dnmls and USTs A A A A A m
UXO A A A r
m
Abandoned wells B B B A 0
Landfill and t~rch boundades B B A B A
Forensics B A B B A B
Archaeologicalfeatures B B B A A A
~ee Natural Geologic and HydrologicConditions to characterizecontaminant pathways.

m
A = pdmary choice of method
B = secondary choice of method o
Figure 1. Table of surface geophysical methods and applications (from ASTM PS 78-97)
r
LTI
Define
survey
objectives
$
~-- Select
application
from table
Review
application
information
Confirm L
application
r
Select
geophysical
methods from
table
Review
method
information
Select ~
preferred
method "
Compare ~., Evaluate

with site site
constraints " conditions
Method not L, ~[ Method

suitable ~" " suitable
Plan survey L,
i TM
Figure 2. Process for selecting surface geophysical methods using ASTM PS 78-97
Select Application
The first step is to define the objectives of the survey and relate them to an
application listed in the "Applications" column in the table. If the survey objective is not
specified directly in the "Applications" list, several applications may have to be assessed
by reviewing the section of the guide titled, "Discussion of Applications", before arriving
at the "Application" most closely resembling the objective. In some instances there may
be more than one objective that can be identified separately in the "Applications" list and
carried along to the next step.
Select Applicable Geophysical Methods

For each "Application" there are usually several methods that can be used to
conduct a geophysical survey, some of which are more generally used or preferred (A
methods) than the others (B methods). Some of the anrated methods might under some
circumstances be appropriate, although they are generally not used for the application.
In the cases where there is more than one "Application" required to cover the
survey objectives, then the first set of methods reviewed would be A methods listed for
both applications. For example, if the survey objective was to identify landfill boundaries
and determine whether these are delineated by buried metal drums, an A-rated method or
A in one category and B in the other might meet the survey requirements.
Confirm Selected Method(s)

At this stage, the selected method(s) are evaluated against the constraints that may
be imposed by the survey site, topography, accessibility, cultural interferences, surface
soils, and so on. If the method(s) selected fail at this stage, then different methods that
will work within these constraints must be evaluated. Once the method(s) selected satisfy
all the constraints then you are ready to plan the survey.
Review ASTM Guides and Referenced Literature

A review of the ASTM Guides for the selected method(s) and of the literature
referenced in the guide will provide invaluable information for planning the survey.
EXAMPLES
The following examples illustrate the process by which appropriate surface

geophysical methods to solve specific problems are chosen and implemented using the
guide. Although the example investigations were not originally conceived and conducted
using the ASTM guide, the case histories are useful to document the rationale for
choosing surface geophysical methods.
Determining Depth to Bedrock

In this example, in which the problem is to determine depth to bedrock where
there is a clay unit overlying bedrock, the A-rated method, radar, is eliminated because of
the presence of clay; the A-rated method seismic reflection is eliminated because of cost
and complexity; and the A-rated method, seismic refraction, is chosen.
Case History
The Farmington River valley in Simsbury, Connecticut, was deeply eroded by

glacial ice during the Pleistocene. Stratified drift (delta and lake-bottom deposits), as
much as 330 meters thick, accumulated in this overdeepened valley. The area between
the towns of Terrys Plain and Avon is almost entirely underlain by fine to very fine sand,
silt, and clay. Test holes and wells in the area penetrate thick deposits of very fine sand,
silt, and clay. A study was conducted in this area by the U.S. Geological Survey (MelVin
and Bingham, 1991) in order to assess the extent, lithology, and hydrogeologic
characteristics of the stratified drift. Seismic refraction methods were used to map the
depth to bedrock beneath the glacial deposits. The refraction method was chosen based
on previous success of this technique m glacial valleys (Haeni, 1986 and 1988), the
geology of the site, and amount and type of data needed. The results of one of the
refraction profiles is shown in Figure 3. The depths to the water table and bedrock
determined from seismic-refraction data agree with the data from well SI-363.
DISTANCE, IN METERS
A AI
METERS
70 --
] ~ /Coarse-grained
60 --J / Soil, stratified drift = / stratified drift,
| / 800 ohm-meters ~ / 500 ohm-meters
5O --(K / ~ ,~ / ~ Land surface
)
4O -
Water table
30
60
20
~ i 0 16 45 Fine-grained stratified drift
10 D
53
- - ~ - 63 120
Sea level 105 115
140
10 (ohm-meters) 120 q
_ 95
20 Sedimentary bedrock ~ / ' ~ ~
200 ohm-meters
30
40
Seismic-refraction datr JSed to interpret depth to water table and bedrock
50
0 500FEET
VERTICALEXAGGERATIONX5
I , I
0 90METERS
Figure 3. Cross Section A-A', Farmington River Valley (from Haeni, 1995)
Another study conducted by the U.S. Geological Survey, designed to evaluate

surface geophysical methods in glaciated valleys (Haeni, 1995), used this site to test and
compare several other geophysical methods. A DC resistivity sounding, a VLF terrain-
resistivity profile, and an inductive terrain-conductivity profile were conducted along the
same line as the previous refraction survey. Sources of cultural interference from railroad
tracks and from a construction company workshop with power and telephone lines were
present at the west end of the conductivity profile. The remainder of the area, which is
used partly for agricultural activities, was a relatively fiat flood plain free from cultural
interference.
Although the two studies in this area did not use the ASTM guide, they illustrate
the usefulness of it. In the guide, under the subheading of depth to bedrock, three
geophysical methods are listed as primary methods: seismic refraction, seismic reflection,
and ground penetrating radar. Ground penetrating radar would not have worked at this
site since the stratified drift is fine-grained (i.e., electrically conductive) and very deep
(up to 90 meters). Seismic reflection could have been used, but seismic refraction was a
more economical choice because only the depth to rock information was needed and
several miles of coverage were needed. Secondary methods in the ASTM Guide include
DC resistivity and electromagnetics. A DC resistivity sounding was conducted at this site
but was difficult to interpret due to the similar electrical properties of the silt and clay and
the sedimentary bedrock. Electromagnetic methods are also limited by the lack of
electrical contrast between the stratified drift and the bedrock.
In this example, the seismic refraction method is the best choice to find the depth
to bedrock in an area with thick, fine-grained material overlying unweathered
sedimentary bedrock.
Mapping a Contaminant Plume
In this example, in which the problem is to map leachate from an oil-field brine
pit, the leachate is comparable to the plume created by saltwater intrusion. Two EM
methods are used to define the plume.
Case History
An oil-field brine pit in Texas has been operational for forty years. The objective
of the survey is to determine the lateral extent and vertical distribution of the brine pond
leakage. There are three A-rated methods and two B-rated methods listed in the table in
Figure 1. The resolution, ease of use and limitations of the methods described in the
section of the guide titled "Discussion of the Geophysical Methods" should be evaluated
and these criteria applied to the survey requirements. The requirements could be met,
with some limitations, by the resistivity method but with such a large area to cover, the
resulting survey could be quite slow and hence costly. The vertical resolution of the
resistivity survey might also be somewhat limited.
Both lateral mapping and vertical distribution determination of the saline plume
can be accomplished in a Short time if the survey is carried out in two phases, using both
EM methods. The frequency domain EM method should be used for a rapid
reconnaissance survey to detect and map the areal extent of the plume. The time domain
EM method should be used to define the vertical distribution of the plume and to derive a
full cross-section using a series of soundings across the plume.
A frequency domain EM (ground conductivity) survey was conducted by making

several transects around the brine pond. After detecting the plume, a few additional
transects were made across the plume (Figure 4) to determine its full extent. Having
defined the plume, a transect consisting of five time domain EM soundings was
conducted across and extending beyond the plume to ensure full lateral coverage of the
plume. A couple of additional soundings, one completely out of the plume for reference
and one downgradient to determine its rate of dispersion, were also conducted. If a full
3D model of the plume had been required, several more time domain transects across the
plume would have been necessary. Figure 5 shows the time domain sounding results
(vertical resistivity profile) for a sounding in the middle of the plume compared to one
completely out of the plume. The conductive saline plume is obvious; both upper and
lower boundaries are clearly defined. Figure 6 shows the final cross-section of the plume.
In this example, two complementary EM methods are the best choice to delineate
the plume.
.Jr o
1
A ;#o.,.,'. 9 a'
MF_.ASUFEMEm'...-. -/-/~"M~ ,~,lr'-~l, d k A, ~, 9 9
9 . ~. / ~ "
/
I"
.~..,.,~,.X"~':~..~~..,~,2"~A
~.."-_'>~~/l'J
"-" . ~ , d ~ - ' - sOUNOm
B \ 1 .:
Figure 4. Location and apparent conductivity contour map of brine evaporation pit and
plume derived from measurements with Geonics EM-34 at 20 meter coil
separation and horizontal, co-planar magnetic dipoles (from Hoekstra, 1992).
CO~X~CTM;Y( ~ t o s ~ )
tl . . . . . . v "~ i . . . .
~
,~1
i-t
~ [--'l---J
L L
i tO' !
,,, ,,L ,, I _.
tO "~ 104 ~q 10"= ;0 ~
Figure 5. Apparent resistivity curves and their one-dimensional inversions for soundings
1 and 3 in Figure 4 (from Hoekstra, 1992).
A TDEM EM47 STATIONS A"

2 $ 4 8
.. ! . . . . . . ! .... 1,, , I ..... I .....
v:: lo,-i| lU~IB~I

e~,e) :~4K42)
a4c4~) 26(401
2~4o, ~'"
. . . . _ .
t~O0 9 ~e ~ F~n e Iaqw,li~ll~Ityo

~--" ~o m , . . (1251 C o m l t ~ i ~ .
Figure 6. Geoelectric cross section derived from one-dimensional inversions of

soundings along A-A' in Figure 4 (from Hoekstra, 1992).
STATUS OF GUIDE DEVELOPMENT
The Provisional Guide for Selecting Surface Geophysical Methods will be

revised and balloted as a standard guide within two years. The standard guide is then
reviewed, revised as appropriate, and reapproved every five years. Standard guides are
being developed for each of the surface geophysical methods included in the Provisional
Guide for Selecting Surface Geophysical Methods. The ASTM Guide for Planning and
Conducting Borehole Geophysical Logging (D5753-95) provides similar information for
the choice of appropriate borehole geophysical methods to solve specific problems.
CONCLUSIONS
The Provisional Guide for Selecting Surface Geophysical Methods covers the
selection of surface geophysical methods commonly applied to geologic, geotechnical,
hydrologic, and environmental investigations as well as for forensic and archaeological
applications. It is designed as a guide to individual method standard guides so that the
program manager or other professional tasked with evaluating proposals and writing
scopes of work can focus their attention on the most generally acceptable methods to
solve a specific problem. If physical properties of geologic units or pore fluids at a site
are known, the guide can also be used to rule out primary methods and select secondary
methods instead. The guide should not be used as the sole criterion for selecting
geophysical methods and does not replace professional judgement or expertise.
ACKNOWLEDGMENTS
The authors thank the other Task Group and Subcommittee D 18.01.02 members
who write, edit, and revise the draft standard guides and all other ASTM Committee D 18
members who comment on and provide input to the guides at all stages of development.
REFERENCES
Haeni, F.P., 1986, "Application of Seismic-Refraction Methods in Groundwater

Modeling Studies in New England", Geophysics, volume 51, number 2, page 236-
249.
Haeni, F.P., 1988, Application of Seismic-Refraction Techniques to Hydrologic Studies,

U.S. Geological Survey Techniques of Water-Resources Investigations Book 2,
Chapter D2, 86 pages.
Haeni, F.P., 1995, Application of Surface-Geophysical Methods to Investigations of Sand

and Gravel Aquifers in the Glaciated Northeastern United States, U.S.
Geological Survey Professional Paper 1415-A.
Hoekstra, Pieter, Lahti, Raye, Hild, Jim, Bates, C. Richard, and Phillips, David, 1992,
"Case Histories of Shallow Time Domain Electromagnetics in Environmental Site
Assessment", Ground Water Monitoring andReview, volume 12, number 4, page
110-117.
Melvin, R.L., and Bingham,J.W., 1991,Availability of Waterfrom Stratified-Drift

Aquifers in the Farmington River Valley, Simsbury, Connecticut, U.S. Geological
Survey Water-ResourcesInvestigationsReport 89-4140.
N. H. Osborne 1
The Key Role of Monitoring in Controlling the Construction of the New Jubilee
Line between Waterloo and Westminster, London
REFERENCE: Osborne, N. H., "The Key Role of Monitoring in Controlling the

Construction of the New Jubilee Line between Waterloo and Westminster, London,"
Marr, Eds., American Society for Testing and Materials, West Conshohocken, PA, 1999.
Abstract:
Crucial to the success of Contract 102 of the Jubilee Line Extension Project (JLEP)
was the detailed analysis of the impact of the construction process on the surrounding
ground, struc~res and tunnels under construction. The construction works were to pass
under or alongside some of London's most historic and prestigious buildings,
consequently stringent limits were imposed upon ground and structure movements. To
control these movements and remain within the limits set, a comprehensive monitoring
system was required, comprising a wide variety of instruments linked to a geoteehnical
database. An extensive and integrated instrumentation system was designed to produce
quality data and as no monitoring system existed that covered the unique requirements of
the contract and of the Balfour Beatty Amec Joint Venture (BBA), the option was taken
by BBA to design and develop a system in house.
The result was Geosys, a windows based monitoring system that could handle data
from the various instruments on site, fulfilling a major role in the successful
observational approach that was taken to control ground movement from both the
tunnelling and the grouting. Geosys enabled electronic data to be reviewed in "real time"
and the surveyed precise levelling results could be reviewed within half an hour of
survey, if required; feedback loops into the design resulted in a major impact upon the
construction decision making processes. The data handling and processing developed
within Geosys are described. The subsequent geotechnical benefits to the contract,
through the application of high quality processed monitoring data to the control of the
ground movements and the construction works, are discussed.
Keywords:
Jubilee, monitoring system, instrumentation, tunnelling, ground movement, data,

grouting, observation
lGeotechnical Engineer, Land Transport Authority of Singapore, 207 River Valley Road,
# 02-61 U.E. Square, 238275, Singapore.
OSBORNE ON CONTROLLINGTHE CONSTRUCTION 405
Introduction
This paper is concerned with construction of Contract 102 of the JLEP and focuses on
the monitoring of ground movements as a consequence of the construction processes
undertaken. These include over 2.5 km of twin running tunnel from Green Park to
Waterloo, connecting adits, concourses and platform tunnels; a 35 metre deep diaphragm
wall box for a new station at Westminster, a new ticket hall at Waterloo and associated
compensation and permeation grouting.
All this construction had a major impact upon the structures along the route and tight
control upon ground and building movements needed to be maintained to ensure that
damage did not occur to structures such as the British Rail Viaduct out of Charing Cross,
the RAC Club and Big Ben. The control of ground movements had the potential to be a
very complex task due to the limited clearances between structures, grouting shafts,
grouting arrays and tunnels. This control was achieved through numerous different types
of instruments, that were targeted to monitor specific construction events and scenarios
with information transferred to Geosys, a specifically designed monitoring system with a
large database storage capacity, multi-user access and a rapid processing time for data. It
had the facility to present the processed data in numerous flexible and simple formats; the
database also needed to be flexible with the ability to automatically accept and order vast
amounts of data from a wide variety of instruments in a wide variety of different formats.
Prior to the major construction works, a robust, reliable and integrated instrumentation
system was designed with specific monitoring objectives and installed. This comprised of
a wide range of instruments with in excess of 6000 precise levelling points, used as
primary control, backed up by over 2000 electrolevel beams, subsurface inclinometers,
extensometers, piezometers, strain gauges and assorted instrumentation. As construction
events impacted upon the instrumentation, readings were taken either manually, by a
surveyor, or remotely by datalogger and modem, and presented to Geosys for processing
and storage. A great number of different construction events were occurring at different
locations throughout the site. At one stage there were sixteen active tunnel faces, ten
separate locations of compensation grouting and the ongoing construction of the two
stations. The monitoring staff required to cope numbered 70, including surveyors,
monitoring engineers and technicians. Consequently the amount of monitoring data to be
processed was enormous, with a maximum of 7500 observations taken during a 24 hour
period and 102,950 electronic readings also collated during a 24 hour period. The
efficient and rapid capturing, processing and outputting of data by Geosys was of
paramount importance to allow fast response to ground movements.
Data Acquisition, Processing, Storage, Archive and Output
A sophisticated communications interface was required to enable survey results from

precise levelling, electronic data from instruments on site and records of grout injected to
be accepted into Geosys. The survey results could be sent directly from the field via
modem and mobile phone or downloaded and presented to Geosys at the office in a
Comma Separated Variable (CSV) format, these results could be accepted every three
minutes and processed immediately. Data from the electronic instruments was gained
through "real time" and hourly communications with 80 dataloggers throughout the site.
The grouting records were downloaded from the grouting sub-contractor's (Amec-
Geocisa JV) database and imported into Geosys at whichever frequency was required.
Once the data had been acquired by Geosys they were processed, this included the
validation of the data, the application of calibration factors and assignment to the correct
database in the correct chronological order. The data were stored in both raw downloaded
format and processed format, with an error log maintained of any corrupted data and any
failures in communications, for future scrutiny if the integrity of the data be questioned or
for future back analysis purposes.
Having processed the data, Geosys could produce plots in several flexible formats or
output data files for further analysis almost instantaneously. The selection of plots
available include displacement against time, displacement against distance over time,
with the worst case angle of distortion calculated and illustrated, or a combination of
ground movement and actual grout injected over a defined area produced as a contour
plot. It is these plots that are reviewed and analysed to control ground movements by
making adjustments to the ongoing construction activities.
Control of Heave Induced Through Grouting
...<y - ........
j =....---,'=2~ ......... ~..-

. . . . . .
= ,. , : ~-~'_-C_.- ...........
\ \ .> _-.
9 - Grout Locations
F i g u r e 1 - G r o u n d H e a v e (mm) Overlain b y Grout I n j e c t e d (l) at Waterloo, 04 -

09/08/1996
Two grouting techniques were utilised that created heave of the ground. Permeation
grouting which was carded out where the cover of London Clay to the crown was less
OSBORNE ON CONTROLLING THE CONSTRUCTION 407
than 6m, to reduce permeability and to increase strength of the overlying Thames
Gravels. This resulted in heaving of the ground as the grouting began to achieve its
objectives and the ground tightened. Observational compensation grouting was also
carried out in targeted areas to induce heave compensating for the post tunnelling
immediate and consolidation settlements. Both types of grouting needed to be controlled
closely by monitoring as the heave could induce stresses in structures within the
proximity leading to the risk of damage. Control was achieved by observational methods
through regular monitoring and the review of set control levels and targets. Grouting was
controlled by adjusting grout volumes, pumping pressures, grouting locations, or
changing grout mix formulations.
Survey information was presented as contour plots with the option to choose the base
date, either to show movement from contract datum or to reflect a particular sequence of
construction activities. To assist with this review the actual injected grout volumes and
locations could be overlaid (Figure 1). Note the small localised heave created in the area
targeted by the injection of 33 cubic litres of grout.
Concurrent Grouting and Tunnelling
To constrain settlement to a minimum, a concurrent grouting strategy was developed

by Amec-Geocisa JV such that compensation grout was injected ahead, behind and to the
sides of the advancing tunnel face in an attempt to compensate for as much immediate
settlement as was possible, without impacting on the safety of the tunnel below or
creating excessive heave ahead of the tunnel face. Again, close control was required of
the ground, structures and tunnel under construction with constant review of monitoring
observations, grout volumes and locations, to make any adjustments necessary to
optimise tunnel speed and for grout operations to achieve their objectives.
Figure 2 illustrates the control achieved whilst progressing under the RAC Club. The
RAC Club was particularly sensitive to heave and settlement due to the limited cover
between the tunnel crown and the building foundations, a minimum cover of 7m. As
tunnelling progressed, intense monitoring with a rapid output of relevant data for analysis
was required to enable adjustments to the grouting configuration and volumes to be
made, if necessary. Over the 120 hour period as the tunnel progressed below the building,
43 level runs were undertaken on the 50 precise levelling points within the building. All
this data was processed and plotted within an hour of the reading being taken, giving
great control over the movement of the building, maintaining its level to within 3mm of
its undisturbed level. The heave and subsequent settlement can be clearly identified
beyond the +/- lmm tolerance of the surveying which materialises after the grouting.
1
~q
0
~
-1
-2
13/02/96 14/02196 15/02196 16/02/96 17/02/96 18/02/96 19/02/96 20/02/96
Date
-~ Precise Levelling pohtt - - ~ Precise Levelling Point 9 Grouting episodes
Figure 2 - Control of Ground Movement Through Grouting
Control of Big Ben
A great deal of construction activity had an impact upon Big Ben including two tunnel
drives, two platform enlargements, the construction of a jacking chamber, shafts, a
pipejack and the construction of Westminster station box. All had the potential to create
settlement with various predictions made including K.G.Higgins et al. (1996) who
through numerical analysis predicted a maximum settlement of 45mm to the North face
of Big Ben and 24.4mm to the South giving an increase in tilt of 1:728. A separate
settlement analysis by Mott MacDonald produced settlement of 20.4mm of the North face
and 8.5mm at the South face resulting in an increase in tilt of 1:1261. Close control was
required, with to date 1380 level runs taken on each of the 35 precise levelling points
surrounding the structure. Other instrumentation used to monitor the structure were
electrolevels, electrolevel inclinometers for subsurface lateral movements, a Gedometer
to monitor the tilt at the top of the tower, an optical plum; and a total station for
verticality monitoring.
The control and correlation between different instruments achieved is depicted in the
following figures. In Figure 3, two independent methods of monitoring the changing tilt
of Big Ben at both the top and bottom of the tower are illustrated. The various excavation
events increase the tilt with grouting applied to compensate. By using this monitoring,
decisions were made as to the measures for controlling the movement of this historic
structure, changing grout volumes or locations, for example.
OSBORNE ON CONTROLLING THE CONSTRUCTION 409
7.0 35
6.0 30
~ 5.0 25~
4.0 20~.
@
~-- 3.0 15~
@
~q 2.0 lO~
[,.
@
1.0 5
II i Grouting-Episode
0.0 !,~11 , i
L I
01/01/95 31/05/95 28/10/95 26/03/96 23/08/96 20101197 19106197 16111197
Date
Figure 3 - Big Ben Movement, Gedometer Versus Level Difference
Displacement (mm)
Further effects of the -4 -2 0 2 4
grouting can be seen o

firom the sub-surface i i i ~ J
electrolevel 5 i i !
, , ,
, .4 ". i
,,
i,~
o~
,i
, ~r
I I
inclinometers. At the ~ 10 . . . . . . . . . ,., .... , r r

grouting horizon during
one particular grouting dt. 15
[Gro.t~g.o,~.l: ;: i '2 ;
episode, a lateral
movement of 3 m m ~ 20 i i i
. , r~" / ..z J,.~~ i i
towards Big Ben can be "~ 25 i t _F,, N . i 'i~. ~\ i I i

, , 9, r rP~',
observed (figure 4) as ~.
the grout displaces the ~ 30
ground in both . . . . . . :_ _ 2 - - ,.4 . . . .
horizontal and vertical 35
directions. Observation
40
of this lateral
i i i , I i
movement played an 45
--,--02/09197 -.4~-. 03/09/97 ...,... 04/09197 u 05109197]
important role in
-. 4.- 07109197 08109197
refinning the volumes
of grout required to Figure 4 - Subsurface Displacement at Big Ben
adjust the tilt of Big During Observational grouting
Ben and ensuring that the impact on adjacent structures was minimal.
Conclusion
The importance of both a robust, reliable and integrated instrumentation system that
combines with an effective and an efficient monitoring system was identified early within
the project. The instrumentation system was designed such that it targeted construction
activities with the emphasis on producing quality and meaningful data. As the monitoring
system was developed in house and on site, it was developed to meet the specific
requirements of both the instruments installed and the monitoring engineers. The system
was windows driven with unlimited access, allowing all personnel simultaneous use.
Three key features were: its capacity to accept, store and assimilate data form a number
of different sources; the rapid processing and plotting time of vast amounts of
information from these varied sources; and the capacity to present simple and flexible
outputs in the form of graphical plots or data files.
The Observational Method was applied to much of the construction works and for this
to function effectively, tight monitoring control was required to allow engineering
decisions to be made. This control was provided by a well planned and integrated
instrumentation system, providing quality data, and combining with a rapid processing
system providing relevant detailed data for analysis almost instantaneously.
References
Higgins, K.G., et al. April 1996, "Numerical Modelling of the Influence of Westminster
Station Excavation and Tunnelling on the Big Ben Clock Tower," Proc. Sym.
Geotechnical Aspects oJ'Uunderground Excavations in Soft Ground, London 1996, pp.
447-453.
Mort Macdonald, January 1996, "Settlement Assessment of Big Ben Clock Tower,"
Commissioned Report No. 24634/F& G/I O3B, Croydon, United Kingdom, 1996.
STP1358-EB/Oct. 1999
Author Index
A H
Alexander, R., 175 Hansen, D. T., 307

Anderson, B., 3 Hawkes, M., 319
Hugo, F., 121
B
Barker, R. D., 157 Jones, A. M., 51

Barley, A. D., 51
Bennett, R. M., 87 K
Benson, C. H., 267
Benson, R., 392 Khoury, M. A., 206
Bilyeu, J., 121
Bobrow, D. J., 360 L
Bosscher, P. J., 267
Brokaw, W. S., 298 Langer, J. A., 175
Byle, M. J., 175 LaPlante, C., 191
Lawrence, C. A., 236
Li, K., 371
C
M
Chen, D.-H., 121
Chowdury, R., 250 Marr, W. A., 36, 319
Mauldon, M., 87
McCullough, M. L., 175
D McGrath, T. J., 66, 101
Montanelli, F., 16
Davies, M. C. R., 51 Moo-Young, H. K., 191
Deming, P., 148
Drumm, E. C., 87 N
E Napolitano, R. A., 206

Erchul, R. A., 382 O
Osborne, N. H., 404
F
P
Fields, P. E., 285
Flentje, P., 250 Paddock, M. W., 236
Pinto, P. L., 3
G
Q
Gilbert, A. B., 135
Good, D., 148 Quiroz, J., 191
411
R V
Recalcati, P., 16
Reddy, M. R., 371 Vaghar, S., 360
Regimand, A., 135 Vasuki, N. C., 175
Rimoldi, P., 16
Rosenfarb, J. L., 206
Ross, D., 236 W
Russell, J. B., 236
S Webb, M C., 66, 101
Welch, L. R., 285
Saunders, W., 392 White, C. C., 157
Selig, E. T., 66, 101
Silvestri, V., 223
Snelgrove, F., 392 Y
Soloyanis, S., 392
Stinnette, P., 333 Yang, M. Z., 87
T
Z
Taylor, S. B., 157
Thomann, T. G., 206
Townsend, F. C., 3 Zimmie, T. F., 191
STP1358-EB/Oct. 1999
Subject Index
A Compaction testing, 382

Computer Advisor for Organic
Soil Projects (CAOSP),
AASHTO, 87 333
Accelerated pavement testing, Concrete box culverts, 87
121 Concrete, reinforced, 101
Aerial photographs, 298 arch culvert, 66
Aluminum access tubes, 223 Conductivity probe, landfill
Arizona, New Waddell Dam, 285 frost penetration
Asphalt pavement, 121 measurement, 191
ASTM Provisional Guide, Contouring, 360
Surface Geophysical Coordinate information, site, 307
Methods, 392 Cost factors, electronic monitoring
Australian landslide monitoring, for geotechnical
250 instrumentation, 36
Automated dam monitoring, 285 Culvert, earth pressure on
Automated geotechnical measurement, 87
instrumentation systems, Culvert-soil interaction, 66
uses, 36
Axial load, 51
D
B
Dam monitoring, automated, 285
Backfilling, 87, 382 Dam safety data acquisition
culvert and soil behavior system, 285
during, 66 Data management systems, 267,
pipe behavior during, 101 285, 298, 307
Backscatter, 135 Central Artery/Tunnel Project,
Balfour Beatty Amec Joint Boston, 319, 360
Venture, 404 geomatics in, 298
Bentonite, 148 Geosys, 404
Boston, Central Artery/Tunnel organic soil, 333
Project, 319, 360 Deep foundations, 3
Deflectometer, 121
Deformation, 121
C behavior, 16
culvert, 66
lateral, 206
Calibration curves, nuclear Delaware, Cherry Island solid
depth gauges, 223 waste site, 175
Calibration, nuclear density Density gauge, nuclear, 135,
gauges, 135 223, 382
Central Artery/Tunnel Project, Density, water, measurement,
Boston, 319, 360 223
Clay deposits, water content and Depth measurement, slurry-
density measurement, 223 supported excavations, 148
COM624P, 3 Dielectric constant, 267
413
Dilatometer/cone pressuremeter Geotechnical reading, cost per

test, 3 unit contract, 319
Direct transmission, 135 Global positioning satellite, 298
Drains, vertical, 236 Ground movement, 404
Dredge spoil, landfill Groundwater, 360, 371
construction with, 175 Grouting, 404
Dynamic replacement, 333
H
E
High density polyethylene, 16,
Earth pressure on concrete box 101
culverts, 87 Highway embankments, earth
Electrical geophysical pressures on culverts
techniques, 157 under, 87
Electrical resistivity, 267 Highway pavement response, 121
Electronic systems, to monitor Highway, peat and silt
geotechnical compression by, 236
Instrumentation, 36 Highway/tunnel project, 319, 360
Embankment, highway, earth Hubbardston Landfill,
pressures on culverts, 87 Massachusetts, 191
Embankment structure, New
Waddell Dam, 285
F Impact soil tester, 382
Inclinometers, 175, 250
Falling weight deflectometer, 121 slope, 3
FEM analysis, 16 Internet, site and area
Fiber optic cable, 285 information, 298
Field stress, Montone hill, 16 Ion analyzer, 371
Finite element analysis, 16, 87 Italy, Montone hill slope
Flood Wall project, Buena Vista, stabilization, 16
Virginia, 382
FloridaPier, 3
Freezing and thawing, 191, 267
Fresh Kills Landfill, New York, Jubilee Line Extension Project,
206 London, 404
Frost depth, 267
Frost penetration, landfill, 191 L
Landfill monitoring, 157
G dredge spoil instrumentation,
175
Geographic information systems, frost penetration, 191
298, 307 stability, 206
Central Artery/Tunnel Landslide
Project, Boston, 360 Australia, 250
Geogrid, 16 Montone hill, 16
Geomatics, 298 Large-span culverts, 66
Geomembrane liners, 157 Laser measuring device/culvert
Geophysical methods, selecting, wall movement, 66
392 Lateral load test, 3
Geosys monitoring system, 404 Leak detection, 157
INDEX 415
Liner leak detection systems, Pier, drilled, 148

landfill, 157 Piezometers, 36, 175, 250
Liquid settlement gauges, 175 Pipe, buried, 101
Load, axial, 51 Pipe-soil interaction, 101
Load simulator, 121 Pore pressure, 206
Load testin$, culvert behavior water, 250
momtoring during, 66 Porous media, 371
Load transfer curves, 3 Preload monitoring, 236
London, Jubilee Line Extension Pressure cell, 87, 175
Project, 404 Pressuremeter test, 3
Pressure transducers, 101
M Pull-out test, in situ, 16
Madeira Walk, slope stabilization R
at, 51
Manure mulch, field nitrate Refuse fill, 206
leaching test, 371 Reliability, electronic systems for
Marsh, 236 geotechnical
Massachusetts instrumentation, 36
Boston Central Artery/ Rheological modeling, 236
Tunnel Project, 319, 360 Rutting, 121
Hubbardston landfill, 191
Models and modeling, 307 S
rheological, 236
Winkler, 3 Seaside slope stabilization, 51
Montone hill slope stabilization, Sediment measurement, slurry-
Italy, 16 supported excavation, 148
Montreal's Island, water content Sensors, spatial location, 307
and density Settlement
determination, 223 contour plans, 360
highway, 236
N landfill, monitoring, 175, 206
ultimate, 333
New Waddell Dam, Arizona, 285 Shear movement, subsurface, 250
New York, Fresh Kills Landfill, Silt, organic, 236
206 Site characterization, 392
Nitrate leaching, 371 Slope inclinometers, 3
Nuclear density gauge, 135, Slope reinforcement, 16
223, 382 Slope stability, 175, 206, 250
Nuclear depth gauges, 223 Slope stabilization, 16, 51
Sludge barrier, paper, landfill,
O 191
Slurry-supported excavation,
Organic soil, 333 depth and sediment
measurement, 148
P Slurry trench, 148
Slurry wall, 148
Paper sludge barrier, landfill, 191 Soil-culvert interaction, 66, 87
Pavement testing, accelerated, Soil density, 101
121 Soil, electrical properties, 267
Peat, 236 Soil-moisture potential, 371
Penetration test, 3 Soil nails, 51
Soil, organic, data management U

system, 333
Soil-pile interaction, 3
Soil-pipe interaction, 101 U.S. Army Corps of Engineers,
Soil resistivity, freezing level Virginia flood wall
prediction, 191 project, 382
Soil strains, 66, 101 U.S. Bureau of Reclamation,
Soil stresses, 66, 101 Central Arizona Project,
Soil tester, impact, 382 285
Spatial location of sensors, 307 U.S. Census Bureau, 298
Stability monitoring program, U.S. Geographic Survey, 298
landfill, 206 Utility trench, 382
Stabilization
Maer slope at Madeira
Walk, 51 V
Montone hill, 16
Steel arch large span culvert, 66 Virginia Flood Wall project, 382
Standards, nuclear density
gauges, 135
Steel bars, strain gauged, 51 W
Steel, corrugated, 101
Strain gauges, 3, 51 Waste sites
Stress, vertical, 87 dredge spoil in construction,
Subsurface investigations, 175
methods selection for, 392 frost penetration, 91
Surface geophysics, 392 liner leak detection systems,
157
T stability, 206
Water content
Tensiometer, 371 clay deposit, measurement,
Thawing and freezing, 267 223
Thermistors, 175, 267 landfill paper sludge barrier,
frost penetration 191
measurement, 191 Watermelon, 371
Thermocouples, 267 Water pressure, pore, 250
Time domain reflectometry, 267 Weight system, two, for depth
Train/ground movement and sediment
monitoring, 404 measurement, 148
Tunnel project, 319, 360, 404 Wire, vibrating, 87
Copyright by ASTM Int'l (all rights reserved); Mon Jun 26 17:14:37 EDT 2017 ISBN 0-8031-2604-2

Field Instrumentation For Soil

Uploaded by

Copyright:

Available Formats

Field Instrumentation For Soil

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Field Instrumentation For Soil

Uploaded by

Copyright:

Available Formats

What is the document about?

What is the document about?

What are some of the topics discussed in the document?

What are some of the topics discussed in the document?

Copyright by ASTM Int'l (all rights reserved); Mon Jun 26 17:14:37 EDT 2017

Field Instrumentation for Soil

Gary N. Durham and W. Allen Marr, editors

ASTM Stock Number: STP 1358

TAT05 .F48 1999

Copyright 9 1999 AMERICAN SOCIETY FOR TESTING AND MATERIALS, West

Peer Review Policy

INSTRUMENTATION ASSOCIATED WITH SOIL STRUCTURE INTERACTION

An Instrumented Geogrid Reinforced Slope in Central Italy: Field Measurements

Uses of Automated Geotechnical Instrumentation Systems---w. ALLENMARR 36

Instrumentation for Monitoring Large-Span Culverts---MARK C. WEBB, ERNEST T. SELIG, AND

Measurement of Earth Pressures on Concrete Box Culverts under Highway

Instrumentation for Monitoring Buried Pipe Behavior During Backfilling--

INSTRUMENTATION SUPPORT CONSTRUCTION ACTIVITIES

Monitoring Pavement Response and Performance Using In-Situ Instrumentation--

A Comparison of Portable and Permanent Landfill Liner Leak Detection Systems---

Instrumentation of Dredge Spoil for Landfill Construction--MICHAEL J. BYLE,

Field M e a s u r e m e n t s of F r o s t Penetration into a Landfill Cover That Uses a P a p e r

Stability M o n i t o r i n g System for the F r e s h Kills Landfill in New Y o r k C i t y - -

INSTRUMENTATION FOR MONITORING SETTLEMENTAND STABILITY

Field M e a s u r e m e n t of W a t e r Contents and Densities by Nuclear M e t h o d s

M o n i t o r i n g P r e l o a d Performance---JAMES B. RUSSELL,CHRISTOPHERA. LAWRENCE,

Instrumentation for Slope S t a b i l i t y - - E x p e r i e n c e from an U r b a n Area--PHIL FLENTJE

DATA ACQUISITIONAND DATA MANAGEMENT

R e m o t e Field M e t h o d s to M e a s u r e F r o s t Depth----cRAi6 H. BENSONAND PETER J. BOSSCHER 267

Integrated Automation of the New Waddell Dam Performance Data Acquisition S y s t e m - -

Documentation of Data Collected from Field Instrumentation for Spatial Applications---

Data Acquisition and Management for Geotechnical Instrumentation on the Central

Data Management System for Organic Soil--PAMELA STINNETTE 333

Review of GIS Developments in Geotechnical Instrumentation on the C e n t r a l A r t e r y /

Use of Tensiometer for In Situ Measurement of Nitrate Leaching--

Selecting Surface Geophysical Methods for Geological, Hydrological, Geotechnical, and

9 Field Data Acquisition and Data Management

The following is an overview of the information presented in this volume.

Instrumentation Associated with Soil Structure Interaction and Construction Monitoring

Boston's Central Artery/Tunnel Project

Instrumentation to Monitor Landfills

Data Acquisition and Data Management

REFERENCE: Pinto, P. L., Anderson, B., and Townsend, F. C., "Comparison of

1Assistant Professor, Department Engenharia Civil, Universidade de Coimbra, 3004

2phO candidate, Department of Civil Engineering, University of Florida, Gainesville, FL

3professor, Department of Civil Engineering, University of Florida, Gainesville, FL

P-y Curves From Strain Gage Data

The horizontal displacements are obtained by double integration of the moment

FIG. 1--Graphical depiction of nonlinear variation offlexural rigidity with

P-y Curves From Slope Inclinometer Data

~d = ElYpredtcted(i)--Y ........ d(i)l (7)

Curves from DMT and Cone Pressuremeter

As a comparison to the curves obtained from instrumentation, curves based on in

The combined curve is linear up to a horizontal displacement of 1.1-mm (DMT B

Robertson and co-authors proposed a linear variation ofc~ to a depth of 4 diameters. At

As an example of the application of both methods, the authors selected a pile

FIG. 4--Location of piles in test group

Bending Moment (kN*m)

0 200 400 600 800 1000 1200

I" ,::i .........!~.......................