Thiem Method

Idf not destroy or throw away this publication.
If you have no further use

rrte to the Geological Survey at Washington and ask for a frank to return it
UNITED STATES DEPARTMENT OF THE INTERIOR

Harold L. Ickes, Secretary
GEOLOGICAL SURVEY
W. C. Mendenhall, Director
Water-Supply Paper 679 A
OS THIEM METHOD FOR DETERMINING

[EABILITY OF WATER-BEARING MATERIALS
ITS APPLICATION TO THE DETERMINATION
OF SPECIFIC YIELD
RESULTS OF INVESTIGATIONS IN
THE PLATTE RIVER VALLEY, NEBRASKA
BY
LELAND K. WENZEL
Prepared in cooperation with the Conservation and Survey

Division of the University of Nebraska
Contributions to the hydrology of the United States, 1935

(Pages 1-57)
Water Resources
Geological Survey,
Box 3106, Car
UNITED STATES
GOVERNMENT PRINTING OFFICE
WASHINGTON : 1936
Iby the Superintendent of Documents, Washington, D. C. Price 10 cents

CONTENTS
Abstract. ____-____----______________---_______----_-------------- 1
Introduction __ ____-________--__---__--__--_-__---_------------- 2
Investigation in the Platte Valley, Nebr______________---_____--_ 2
Acknowledgments. ____________--_--__-_-_----_---------------- 3
Hydrologic properties of water-bearing formations_______-__-____,--- 4
Outline of Thiem and pumping-test methods-_______-_---____-------- 8
Permeability. ______--_--_--________--___---_________--------- 8
Specific yield-______-____--___________-------_-_____---------- 9
Development of TMem 'a formula.__________-_-_-_----_-___--------- 10
Confirmation of Thiem's formula from other work done in the United
States..___________________________---___-_._.- IS-
Formula for determining the cone of depression-___.-______-------_---- 23
Graphic solution of Thiem's formula____________________-------_-__ 25
Pumping tests in Nebraska_________-________________--------_--_- 26
Results obtained from the pumping tests_______________--__--_-___- 32
Draw-down curves.____________-_----_________-----------____- 32
Recovery curves__________----_-___---______-_-_--------__l___ 34
Cones of depression______-___--_---_________-.-----------__---- 36
Computation of coefficients of permeability_____________________ 41
Differences between field and theoretical conditions _______________ 50
Determination of specific yield by the pumping method.___________ 53
ILLUSTKATIONS
Page
PIRATE 1. Graph for computing coefficients of permeability.____________ 24
2. A, Line SW of 1-inch observation wells; B, Measuring the depth
to the watertable_______________________________________ 28
3. A, Weir for measuring the discharge of the pumped well;
B, Pumping arrangement in second test_________________ 29
4. Typical draw-down and recovery curves for test l____-____-__ 36
5. Typical draw-down and recovery curves for test 2____________ 36
6. Contours on the water table before pumping and at several times
after pumping began__________________________________ 36
FIGURE 1. Plan and section of ideal ground-water conditions assumed
by Thiem__________-___-_-_________________________ H
2. Plan and section showing assumed ground-water conditions for
the development of the formula from horizontal water table. 16
3. Plan and section showing assumed ground-water conditions
for the development of the formula from horizontal artesian
conditions____________________________________________ 17
4. Map showing location of wells used in pumping tests._______ 27
5. Computed recovery curve for well 5, test !_________________ 37
6. Profiles of the cone of depression at several times after pumping
began and location of cylindrical sections used for comput-
ing specific yield-_______-_-__-_-----____________._____ 39
7. Relation of 2z>2 to P__________________________________ 48
in
TABLES
Page
1. Physical properties of samples of alluvium._________.--_____-___-__ 28
2. Log of well 84_-..----_-___._......___--__.__-_-._..___. 28
3. Location, diameter, depth, and altitude of wells______.......--___-_ 29
4. Record of pumping time_________________________________________ 31
5. Draw-down of the water table during test 1_.______________________ 36
6. Coefficients of permeability computed by Thiem's formula for all pos-
sible combinations of observation wells on line A_______-..--__--_ 42
7. Coefficients of permeability computed by Thiem's formula for several
combinations of observation wells on line A and for.several periods
of pumping.____._--__-______-_____-----_--___-_--__-------- 43
8. Coefficients of permeability computed by Thiem's formula, using a
50 feet_---_____.-....-._._.____.....__-____--.___ 44
9. Draw-down of water table for several distances and directions from
the pumped well._----_-____-_--------- _.__----_---___---_ 45
10. Differences in average draw-down on lines A and C____------------ 45
11. Final computation of coefficients of permeability__________-- _--_- 47
12. Computation of v* for P=975......-..._-----_------------------ 48
13. Observation equations for t/2 _______________-_-_-_____-------_--_ 49
14. Volumes of water-bearing material unwatered around the pumped
well _________-_-__-_-_--__-___--___----__-_-________------- 54
15. Computed quantities of ground water that percolated through several
concentric cylindrical sections around the pumped well-______..._ 54
16. Quantities of ground water taken from storage___-_--_-__--_-_---- 55
17. Specific yield as computed for several concentric cylindrical sections
and for several periods of pumping-____-_____--___-_---____--- 55
IV
THE THIEM METHOD FOR DETERMINING PERMEABILITY
OF WATER-BEARING MATERIALS AND ITS APPLICATION
TO THE DETERMINATION OF SPECIFIC YIELD
By LELAND K. WENZEL
ABSTRACT
The Thiem method for determining permeability of water-bearing materials *

consists of pumping a well, or, where the ground water is confined under pressure,
allowing the well to flow and observing the decline of the water table or piezo-
metric surface in nearby observation wells. The coefficient of permeability is
computed by the formula
527.7 q logio -
where P is the coefficient of permeability; q is the rate of pumping, in gallons a

minute; a and 01 are respective distances of two observation wells from the
pumped well, in feet; m, for artesian conditions, is the vertical thickness of the
water-bearing bed, in feet; m, for water-table conditions, is the average vertical
thickness, at a\ and a, of the saturated part of the water-bearing bed, in feet; and
s and si are the draw-downs at the two observation wells, in feet. This formula
is mathematically developed by assuming ideal geologic and ground-water con-
ditions, such as a uniform permeability, a uniform thickness of water-bearing bed,
s horizontal water table or piezometric surface, and a cone of depression that has
reached equilibrium in form. As these conditions are rarely approached, the
applicability of the formula and hence of the method has been regarded as
questionable.
Two rather elaborate pumping tests were made in 1931 near Grand Island,
Nebr., to ascertain the accuracy of the Thiem method and to investigate the possi-
bilities of determining specific yield by a pumping test. The behavior of the
ground water was observed over a large area around the pumped wells by measur-
ing the fluctuation of the water table in 81 observation wells during the period of
pumping and after pumping was stopped. A study of the data obtained from
these tests indicates that the Thiem method is applicable to conditions that are
found in nature. However, to obtain consistent and accurate determinations of
permeability it is necessary to employ an arbitrary procedure in computing the
coefficient. The draw-down of the water table at any distance from the discharg-
ing well should be taken as the average of the draw-down at that distance up-
gradient and down-gradient from the well. In Thiem's formula only results for
the draw-down of the water table that are obtained from the part of the cone of
depression that has reached approximate equilibrium in form can be used. The
part of the cone that has reached approximate equilibrium is determined by fre-
i Thiem, Q., Hydrologische Methoden, Leipzig, 1906.
2 CONTRIBUTIONS TO HYDROLOGY OF UNITED STATES, 1935
quent measurements of the draw-down during the period of pumping. If the

discharging well fails to penetrate through the water-bearing bed, the draw-down
of the water table close to the well should not be used, because of irregularities in
the cone of depression. Moreover, there are usually near the well some changes
in the permeability of the water-bearing material resulting from the development
of the well. In the first test described in this report the cone of depression reached
approximate equilibrium in form out to about 200 feet from the pumped well
after 48 hours of pumping and was affected by irregular conditions near the well
as far as 40 feet from the well. Hence the draw-downs that were used for
computations of permeability were selected from that part of the cone between
40 and 200 feet from the pumped well. In the second test pumping was stopped
several times, and the cone of depression did not reach approximate equilibrium
in form.
Computations were made to determine the specific yield of the water-bearing
materials from the data obtained in the pumping tests. The results show that
the specific yield can be readily determined by this method. Samples of the
material were analyzed in the laboratory for specific yield, and the results obtained
compared favorably with those determined by the pumping method.
INTRODUCTION
INVESTIGATION IN THE PLATTE VALLEY, NEBR.
An investigation of the ground-water resources of Nebraska has for

some time been in progress under the supervision of G. E. Condra,
director of the Conservation and Survey Division of the University
of Nebraska. At the request of Dr. Condra, a ground-water investi-
gation of that part of the Platte River Valley lying between Chapman
and Gothenburg, Nebr., was undertaken July 1,1930, as a cooperative
project between the Conservation and Survey Division and the United
States Geological Survey, under the general supervision of O. E.
Meinzer, geologist in charge, division of ground water in the Geological
Survey. The writer was assigned to this cooperative project and
began work July 12,1930.
The investigation has for its purpose the determination of the
source, quantity, and availability of the ground water, with a view to
accomplishing maximum recovery and utilization. Field work has
been carried on continuously since the project was begun, and compre-
hensive data have been collected concerning the occurrence and be-
havior of the ground water in that part of the valley. The area is
one in which there is rather intensive irrigation by ground water, and
it was found that determinations of permeability and specific yield
of the water-bearing materials should be made in order to obtain a
quantitative estimate of the ground-water supply. It was decided
to use the Thiem method for the determination of permeability and
the pumping method for the determination of specific yield.
Neither of these methods had been adequately verified by experi-
ments for accuracy and practicability. Hence it was necessary to
make rather elaborate tests that would determine the reliability of
these methods and at the same time yield the actual figures for perme-
THIEM METHOD FOE DETERMINING PERMEABILITY 6
ability and specific yield. Accordingly, two tests of these methods

were made in the summer of 1931 near Grand Island, Nebr.
The pumping method for determining specific yield was outlined by
Meinzer 2 independent of the Thiem method for determining permea-'
bility. However, the data necessary for the determination of specific
yield and permeability by these methods can be obtained from one
pumping test with a very small amount of additional effort. Where
the ground water is confined under pressure the Thiem method for
determining permeability may be used, but the pumping method for
determining specific yield fails because there is usually no unwatering
of the water-bearing materials. Hence the method for determining
specific yield is strictly a pumping method, but the Thiem method
applies also to areas where wells discharge water under artesian
pressure. In the area where the tests described in this report were
made the water-bearing materials consist of unconsolidated sand and
gravel. As the upper surface of the zone of saturation lies several
feet below the top of the water-bearing materials, the ground water is
not confined under pressure, and both the Thiem method for deter-
mining permeability and the pumping method for determining specific
yield could be used.
The behavior of the ground water near the pumped wells was ob-
served in detail during and after the period of pumping, and these
observations provided an opportunity to determine the effect of
differences between theoretical and observed conditions on the
computations for permeability and specific yield by these methods.
The results obtained from the tests, a review of Thiem's development
of the formula for permeability, and a theoretical review of the formula
are incorporated in this report. Another report, now in preparation,
will give the other data obtained in the Platte River investigation and
the conclusions that were reached as to the ground-water conditions in
that valley.
ACKNOWLEDGMENTS
The writer is indebted to G. E. Condra, A. L. Lugn, H. A. Waite,
H. P. Burleigh, Howard Haworth, and Keith Miller, of the Conserva-
tion and Survey Division of the University of Nebraska, and to E. E.
Brackett, chairman, and E. B. Lewis, research engineer, of the Agri-
cultural Engineering Department of the University of Nebraska, for
their assistance in making the tests, and to Fred Meyer for his coopera-
tion in making his farm and pumping plant available for carrying out
the tests. He is especially indebted to J. W. Shew, professor of
mathematics, Junior College, Grand Rapids, Mich., and to C. E. Van
Orstrand, geophysicist, United States Geological Survey, for their
review of the mathematical sections of this report; to C. E. Slichter,
2 Meinzer, O. E., Outline of methods for estimating ground-water supplies: U. S. Geol. Survey Water-
Supply Paper 638, p. 136, 1932.
dean of the Graduate School, University of Wisconsin, and G. Thiem,

consulting hydrologist, Leipzig, Germany, for their criticism of the
manuscript; and to O. E. Meinzer, geologist in charge of the division
of ground water of the Geological Survey, for his advice and sugges-
tions throughout the investigation and for his review of the manu-
script.
HYDROLOGIC PROPERTIES OF WATER-BEARING
FORMATIONS
Most ground-water investigations are concerned with the quantity
of water that is available for use by man. Perhaps the greatest
difficulty in the determination of this quantity lies in the variability
in the texture and hence in the hydrologic properties of the water-
bearing materials. The hydrologic properties vary greatly, even with
apparently slight differences in texture. Hence the ordinary geologic
descriptions are quite inadequate for hydrologic investigations, and
quantitative descriptions based on laboratory determinations have
become essential.
The two hydrologic properties of greatest significance are perme-
ability and specific yield. Mechanical analyses and determinations
porosity and moisture equivalent are useful chiefly as indirect
means of determining these two essential hydrologic properties.
About 1843 Poiseuille 3 discovered the law of flow through capillary
tubes namely, that the rate of flow is .proportional to the hydraulic
gradient. Later Darcy 4 verified this law and demonstrated its
application to water percolating through the capillary interstices of
sand and other porous media. He expressed this law by means of the
Kp'
formula v= j > in which v is. the velocity of the water through a
fi
column of permeable material, pf the difference in head at the ends
of the column, h the length of the column, and K a constant that
depends upon the character of the material, especially on the size of
the grains. Because it is usually more essential to determine the
quantity of water flowing through a certain cross section of permeable
material than to determine the velocity through the material, Darcy's
formula is sometimes expressed as
in which Q is the quantity of water discharged in a unit of time, P the

constant, which depends upon the texture of the material, / the
hydraulic gradient, and A the cross-sectional area through which the
water percolates. This formula serves as a basis for determining the
quantities of ground water that percolate from areas of recharge to
* Poiseuille, J,, Recherches expfirimentales sur le mouvement des liquides dans les tubes de trSs petits
diamStres: Acad. sci. Paris M6m. sav. fitrang., vol. 9, p. 433, 1846.
1 Darcy, H., Les fontaines publiques de la ville de Dijon, Paris, 1856.
THIEM METHOD FOR DETERMINING PERMEABILITY 5
areas of discharge, and consequently it is used for determining the

safe yields of ground-water supplies.
The constant P in equation 1 is the most difficult factor to deter-
mine. The hydraulic gradient of an area can be obtained from contour
maps of the water table or piezometric surface,5 and the cross-sectional
area of the water-bearing material can be approximately determined
from the logs of wells penetrating the material. The constant P has
been designated by different names and has been expressed in various
units. According to the present usage of the United States Geological
Survey, it is called the "coefficient of permeability", defined as the
rate of flow, in gallons a day, through a square foot of cross section,
under a hydraulic gradient of 100 percent, at a temperature of 60° F.6
In field terms the coefficient of permeability may be expressed as the
number of gallons a day at 60° F. that is conducted laterally through
each mile of the water-bearing bed under investigation (measured at
right angles to the direction of flow), for each foot of thickness of bed
and for each foot per mile of hydraulic gradient.6 Coefficients of
permeability range widely. Fine sand is in general less permeable
than coarse sand and therefore transmits less water through equal
cross-sectional areas under the same hydraulic gradient. Clay may
contain more water per unit volume than sand or gravel, but the
permeability of a clayey material is generally low, and therefore the
quantity of water transmitted through it is usually much less than
is transmitted through sand and gravel. Coefficients of permeability
ranging from 0.005 for clay to more than 20,000 for sand and gravel
have been determined in the hydrologic laboratory of the United
States Geological Survey.
The permeabilities of water-bearing materials may be determined
by laboratory tests of samples of the materials or by determinations
of ground-water velocities in the field. Hazen 7 and Slichter 8 have
studied the rate of flow of water through sand and have developed
formulas which essentially include the determination of the permea-
bility of the sand. King 9 has reviewed the results of the investigators
»The upper surface of the zone of saturation in ordinary soil or rock is called the "water table." If a well
Is sunk it remains empty until it enters a saturated permeable bed that is, until it enters the zone of
saturation. Then water flows into the well. If the rock through which the well passes is all permeable the
first water that is struck will stand in the well at about the level of the top of the zone of saturation that
is, at about the level of the water table. If the rock overlying the bed in which water is struck is impermeable
the water is generally under pressure that will raise it in the well to some point above the level at which it
was struck. In such a place there is no water table, and the imaginary surface to which the water rises under
its full head is called the "piezometric surface."
6 Stearns, N. D., Laboratory tests on physical properties of water-bearing materials: U. S. Qeol. Survey
Water-Supply Paper 596, p. 148,1928.
' Hazen, Alien, Some physical properties of sands and gravels: Massachusetts State Board of Health
24th Ann. Kept., p. 553, 1892.
s Slichter, C. S., The motions of .underground waters: U. S. Geol. Survey Water-Supply Paper 67, p. 26,
1902.
6 King, F. H., Principles and conditions of the movements of ground water: U. S. Qeol. Survey 19th Ann.
Kept., pt. 2, pp. 178-204, 1898.
18274 36 2
6 CONTRIBUTIONS TO HYDROLOGY OF DNITED STATES, 1935
on the flow of water through porous media and has described an

apparatus for measuring the flow in the laboratory.10 In the Geologi-
cal Survey the permeability of water-bearing materials is now deter-
mined in the laboratory by means of apparatus devised by Meinzer. 11
The coefficient of permeability of a water-bearing material is deter-
mined directly by measuring the rates of flow of water through a
sample of the material with known cross section and thickness under
observed differences of head. The laboratory methods are open to the
criticism that the coefficients of permeability of the samples tested
may differ widely from the average coefficient of the material as found
in nature. The material that is tested in the laboratory must
necessarily be removed from the ground, and as a result, especially
with the more unconsolidated material, the soil particles do not
remain in their original arrangement. Moreover, the coefficients of
permeability determined in the laboratory necessarily apply only to
very small samples, and unless a great number of samples are tested
an average coefficient for a large area cannot be determined. These
statements do not imply that laboratory determinations are not
significant; they are intended merely to point out some of the inherent
difficulties involved in such tests and to emphasize the importance of
carefully and thoroughly investigating a method such as the Thiem
method, which determines permeability in the field over a large area
and without disturbing the water-bearing material.
A method for determining the natural velocities of ground waters,
patterned after the method of the German hydrologist A. Thiem, was
developed by Slichter. 12 Several small wells are driven into the water-
bearing materials in such a manner that the water moves from one
well toward one or more of the other wells. A salt is introduced into
the up-gradient well and is allowed to move down-gradient with the
ground water to the other wells, where its arrival is detected electri-
cally. The rate of movement of the salt and hence the rate of move-
ment of the ground water is computed from the elapsed time between
the introduction of the salt in the central well and its detection in a
well located down-gradient. The quantity of water flowing through
a given cross-sectional area of the water-bearing material is computed
by the formula
where Q= quantity of water;

p porosity of the water-bearing material;
A= cross-sectional area;
v= average velocity of the ground water.
" King, F. H., op. cit., p. 228.
« Stearns, N. D., op. cit., p. 144.
» Slichter, C. S., op. cit., p. 48.
THIEM METHOD FOR DETERMINING PERMEABILITY /
The coefficient of permeability of the water-bearing material is

computed by equating equations 1 and 2:
=PZA and P-=.. ........ ......(3)
J.^i 4
There are difficulties in the use of this method. The method is
not satisfactorily adaptable to localities where the ground water has
low velocity, because the salt solution, whose specific gravity is higher
than that of the natural water, sinks rather rapidly and may not
reach the down-gradient wells. In using this method in such a
locality, the wells are located comparatively close to one another
usually about 4 feet apart. Under these conditions errors in deter-
mining the velocity of the ground water are often introduced by
failure to sink the wells exactly plumb, by the diffusion of the salt
solution, and by increase in the hydraulic gradient caused by the
rise of water in the up-gradient well at the time the salt is introduced.
The specific yield of a water-bearing formation is defined by
Meinzer as the ratio of (1) the volume of water which, after being
saturated, it will yield by gravity to (2) its own volume.13 It is a
measure of the quantity of water that a formation will yield when it is
drained by lowering of the water table. Thus if 100 cubic feet of
saturated water-bearing material when drained will supply 20 cubic
feet of water, the specific yield of the material is said to be 20 percent.
The practical use of the specific yield is obvious. The quantity of
water that a saturated material will furnish from storage depends upon
its specific yield. To estimate the water supply obtainable from a
material for each foot that the water table is lowered, or to estimate
the available supply represented by each foot of rise in the water table
during periods of recharge, it is necessary to determine the specific
yield.
Meinzer 14 gives seven more or less distinct methods of determining
specific yield namely, (1) saturating samples in the laboratory and
allowing them to drain; (2) saturating in the field a considerable
body of material situated above the water table and above the
capillary fringe and allowing it to drain downward naturally; (3)
collecting samples immediately above the capillary fringe after the
water table has gone down an appreciable distance, as it commonly
does in summer and autumn; (4) ascertaining the volume of sediments
drained by heavy pumping, a record being kept of the quantity of
water that is pumped; (5) ascertaining the volume of sediments
saturated by a measured amount of seepage from one or more streams;
(6) making indirect determinations in the laboratory with small
is Meinzer, O. E., Outline of ground-water hydrology: V. S. Geol. Survey Water-Supply Paper 494,
p. 28, 1923.
n Meinzer, O. E., Methods for estimating ground^ water supplies: U. S. Geol. Survey Water-Supply
Paper 638, p. 113, 1932.
samples by the application of centrifugal force; and (7) making

mechanical analyses and determinations of porosity and estimating
therefrom the specific retention and the specific yield.
Much work has been done on the determination of specific yield by
able investigators, but the methods just enumerated are still not
thoroughly developed. The method of determining specific yield
from pumping tests probably is the least developed of all.
OUTLINE OF THIEM AND PUMPING-TEST METHODS
PERMEABILITY
The Thiem method is very simple in principle. It consists of
pumping a well that penetrates water-bearing material, the perme-
ability of which is to be determined, and observing the decline of the
water table or piezometric surface around the pumped well. Ground
water obeys the law of fluids in that it always flows away from a
point of high pressure toward one of low pressure. In other words,
it flows in the direction of the hydraulic gradient. When a well is
pumped some water inevitably is taken out of storage from the well
and from the material surrounding it. This reduces the pressure,
creates a hydraulic gradient toward the well, and causes ground water
to flow into the well. If the water-bearing formation has a water
table, considerable ground water may have to be removed from
storage before a gradient will be developed that is steep enough to
make the water flow toward the well at the rate that it is pumped and
thus establish approximate equilibrium. If the formation is filled
with water under pressure only a comparatively small amount of
water has to be removed from storage in order to give the required
gradient, and hence the draw-down will be more rapid and approxi-
mate equilibrium will be more quickly established.
When, with a constant rate of pumping, equilibrium is established,
water is no longer removed from storage around the well but flows to
the well as rapidly as it is withdrawn. If before pumping begins the
water table or piezometric surface in a homogeneous formation is
horizontal, water percolates toward the pumped well equally from all
directions, and the same quantity of water percolates toward the
pumped well through each of the indefinite series of concentric
cylindrical sections around the pumped well. Because the areas of
the large cylinders through which the water percolates are greater
than the areas of the smaller cylinders, the velocity of the ground
water passing through them is proportionally less and the hydraulic
gradients are proportionally smaller.
According to equation 1 the discharge through any of the concentric
cylindrical sections of water-bearing material, Qt is equal to PiA, and
the permeability of the material, P, equals - The symbol i is used in
this report to represent the hydraulic gradient at a point on the cone of

depression around a well that is discharging water, and the symbol /is
used to represent the normal hydraulic gradient that the water table
or piezometric surface possesses when the well is idle. The two
symbols are interchangeable in equation 1, their use depending upon
whether the water table, or piezometric surface is cone-shaped or is
approximately a plane. As previously explained, after approximate
equilibrium has been reached the discharge through all concentric
cylindrical sections of water-bearing material is the same, and the
total discharge is equal to the quantity of water being pumped from
the well. The hydraulic gradient at a given distance from the pumped
well can be determined from the slope of the water table or piezometric
surface. For artesian conditions the area of the cylindrical section
through which the ground water percolates at that distance from the^
pumped well is equal to 2irxm, if x is the distance from the pumped
well and m is the thickness of the water-bearing material. For
water-table conditions the area is equal to 2irx(m-- s), where s is the
draw-down at the distance x from the pumped well. Thus the perme-
ability of the water-bearing material can be computed by substituting
these figures in the equation P = ^'
In 1906 G. Thiem,15 son of the German hydrologist A. Thiem, pub-
lished the results of his work in connection with the determination of
additional water supply for the city of Prague and its suburbs. In
this investigation he used what has since been known as the "Thiem
method" for determining permeability and sunk 10 sets of wells, each
set including 1 well that was pumped and 2 observation wells. The
observation wells-were placed in line with the pumped well but in any
convenient direction regardless of the natural hydraulic gradient. A
formula was developed for computing the permeability from the data
obtained from the pumped well and the two observation wells.
SPECIFIC YIELD
The determination of specific yield by the'pumping method is based

on the withdrawal of water from storage during the period of pumping.
Water is taken from storage until an approximate equilibrium is
reached. Thus for a time the flow through successive concentric
cylindrical sections around the pumped well will not be equal that is,
the flow through a large cylinder will be less than the flow through a
small cylinder, because a part of the ground water that percolates
through the small cylinder is derived from storage between the two
cylinders. The volume of material between any two cylinders that
is unwatered in a given time can, of course, be computed from the
" Thiem, G., Hydrologische Methoden, Leipzig, 1906.
10 CONTRIBUTIONS TO HYDBOLOGY OF UNITED STATES, 1935
draw-down of the water table as shown by successive measurements

of the depth to water in observation wells. The average hydraulic
gradient that causes the water to percolate toward the pumped well
can be determined from the same records of depth to water, provided
the altitude of the tops of the observation wells is known. The total
quantity of water that percolates through each of the two cylinders
in the given time is computed by use of the formula Q=PiA. The
difference between these quantities represents the volume of ground
water taken from storage between the two cylinders. The specific
yield of the water-bearing material is then determined by dividing
this volume by the volume of material unwatered in the same time.
DEVELOPMENT OF THIEM'S FORMULA
Thiem's formula for computing the coefficient of permeability may

be written in the convenient form
527.7g log ld-J-
p- _________^_____________________(4)
J. s \ \ /
in which P=the coefficient of permeability as defined on page 5;

g=rate of pumping, in gallons a minute;
a and ai= distances of two observation wells from the pumped
well, in feet;
m (for artesian conditions) = vertical thickness of water-
bearing bed, in feet;
m (for water-table conditions) = average vertical thickness
(at <LI and a) of the saturated part of the water-
bearing bed, in feet;
s and «i= draw-downs at the two observation wells, in feet.
Thiem assumed a region where the water table or piezometric
surface had an initial slope or hydraulic gradient before pumping
began. His final formula did not contain a factor involving this
slope, and he concluded that an initial slope of the water surface had
no effect on the coefficient of permeability as computed by his formula.
A review of the development of his formula indicates that during the
development his original system of oblique coordinates was changed
to a system of rectangular coordinates, which eliminated the factor
involving the hydraulic gradient, and the resultant formula theoreti-
cally pertains only to regions where the water table or piezometric
surface is horizontal. The following is Thiem's development, with
some added interpretation, for water-table conditions:
A water-bearing bed of uniform permeability is assumed to rest on
a relatively impervious formation, as indicated in figure 1. Water
moves through the bed under a normal hydraulic gradient that is
parallel to the slope of the underlying impervious bed. In this
ground-water stream there is a well equipped with a pump extending
to the bottom of the water-bearing material, and two observation

wells are placed in line with the pumped well. The pump is operated
at a uniform rate during a period in which the water table declines
and takes a form somewhat similar to an inverted cone around the
pumped well. The draw-down (decline of the water surface) in each
observation well, the distances of these wells from the pumped well,
the rate of discharge of the pumped well, and the thickness of the
water-bearing bed are measured. The coefficient of permeability is
computed by substituting these measurements in Thiem's formula.
PLAN
s\
Direction of ground-water movement / \
Observation well
Pumped well
SECTION
FIGTJKE 1. Plan and section of ideal ground-water conditions assumed by Thiem.
The following symbols, in addition to those previously given, are

used in the development of the formula:
I natural hydraulic gradient;
A= cross-sectional area in square feet that is, area of any
designated cylindrical section through which the water
percolates on its way to the pumped well;
H= thickness, in feet, of the saturated part of the water-bearing
bed in the undisturbed condition of the water table;
Q rate of pumping, in gallons a day;
x and y= oblique coordinates of a point, J, on the cone of depression,

with reference to the point of intersection of the imperme-
able bottom of the formation with the axis of the well as
the origin;
i= hydraulic gradient, in feet per foot, at any point, J, on the
cone of depression;
x1 distance, in feet, of point Jfrom the pumped well;
<£=the angle that a line through the pumped well and the two
observation wells makes with the uninfluenced direction
of ground-water movement;
6= the angle of inclination of the impermeable bottom;
t distance, in feet, of the projection of point J from the
pumped well measured along the uninfluenced direction
of the movement of the ground water.
In the ground plan, figure 1, assume a small sector with an angle
d$ whose apex is at the axis of the well and whose sides form an
angle 0 with the uninfluenced direction of ground-water flow. At
the point J, at a distance x from the well, the flow dQ passes through
the sector d$. By using Darcy's equation, Q=PiA (1), it is possible
to compute the flow dQ through the sector d<j> at the point «7.
The length of the arc at J=x cos Qd^>. The vertical thickness of
the saturated water-bearing material is y. So the area through
which the flow dQ passes is equal to xy cos Qd<l>.
The hydraulic gradient, i, at any point on the cone of depression
is equal to the rate of change of the coordinates. These coordinates
must be at riglit angles, so the horizontal coordinate is x1 and the
vertical coordinate is (n4-y).
x1 =x cos 6_-____ -------------- _-_(5)
and n=x sin 6__ _____________________ (6)
,
tlms *.d(n+y}_d(x
3? sin 6+y) ----------------(7)
d(zcose) m
By substituting in Darcy's equation,

sin Q-\-y)xy cos Qd(f> , .
d(x cos 9) ---- ------(8)
From the plan and section (fig. 1)

t=x cos 6 cos 0 ___________________ (9)
77
The natural hydraulic gradient is equal to , hence
,=»_- ._.___- _____._.
By equating (9) and (10),

n=xl cos 6 cos 0_ _________________ (11)
n=x sin 6 ________________________ (12)
Equating (11) and (12), we get

x sin 6=arJcos 6 cos 0 _________________ (13)
and sin 6= J cos 6 cos <f> _________________(14)
When both sides of equation 14 are squared,
sin2 6=72 cos2 6 cos2 0_ ______________ .(15)
1 sin2 6 may be substituted for cos2 6; thus equation 15 becomes
sin2 e=72 (l sin2 6) cos2 0 ______ ______.(16)
=P cos2 Q-P sin2 6 cos2 0_ ______ (17)
and 72 cos2 0=sin2 6 +P sin2 6 cos2 0___________ _ _(18)
=sin2 e(l+72 cos2 0) ____.____________(19)
Thus 72 cos2 0 = l+72 cos2 0- ----- ----- -(20)
and, by taking the square root,

sin 6 1
7cos0 T/l + P cos2 0 -- -
Equation 14 may be written
sin 6
=cos Q _____ __________________ (22)
7 cos 0
and by equating 21 and 22,
cos e=-7rr==L===_________________(23)
VI +P cos2 <f>
1 cos2 6 may be substituted for sin2 6; thus equation 15 may be
written
1 cos2 Q=P cos2 6 cos2 0________________(24)
l=cos2 e+/2 cos2 6 cos2 0___________.____(25)
l=cos2 6 (P cos2 0+l)__________________(26)
and 75cos2
P 2 0+1
... =cos2 6 ____---__________(27)
v '
By taking the square root,
\ ^TT= COS ©----- - -- ---(28)
cos2 0+1
and by multiplying both sides of equation 28 by cos 0 I,
cos 07
=cos 6 cos 07____________(29)
V72 cos2 0+1
Equation 14 may be substituted in equation 29, thus:
V72 cos2 0+1
18274 36 3
Equations 28 and 30 may be substituted in equation 8, thus:

cos
Thiem states at this point that P is very small and therefore can
be assumed to be zero, thus introducing a small error. If P Q,
equation 31 becomes
cos <j>I+y)xyd<j> __ _ __ _ (32)
dx
Pxyd<}>( cos 0 ------ ---(33)
<^----- ----- -(34)
In order to integrate equation 34, Thiem changed from an oblique

system of coordinates to a rectangular system of coordinates. Thus
7=0, n Q, and 0=0, and Thiem's final equation will pertain only
to horizontal water-table conditions. Equation 34 then becomes
By integrating with respect to $ and Q,
V (36)
and Q =27rP ------------ (37)
If the equation is now further integrated with respect to x and y,

- (38)
and +a.... .... (40)
This is Thiem's general equation. If equation 37 is integrated

between limits x=a, x=a^ and y=h, y=hi (fig. 1, section) equation
40 is developed into a more practical form. From equation 37
f 1 *-?-? f
Ja X Q Jh
r
loge# f) 9 I
L H? L *Jh
i , ll2 h2
loge di loge a= ~Q-
, ,
loge fli log. a=
and p= ^(log -log. a)
A) -- -- -(46)
and (hi h) is equal to the difference of draw-downs (s «i). Thus
V-A2 =$i+fc)(s-Si) ----- (47)
and equation 45 becomes
p=@ (loge fli loge a) -------- _ (48)
---
This is Thiem's final equation and applies to regions where the-

ground water is not confined below an impermeable bed. Thiem
developed a formula that applies to artesian conditions in the same
manner. His artesian formula differs from equation 48 only in that
(hi+h) is replaced by 2m. If m is used as defined in this paper
(p. 10), the equation for both water-table and artesian conditions
may be expressed
p ^Q (loge QI loge a) ___ _____ __ ...(49),
2irm(s Si)
Equation 49 includes factors involving natural logarithms that is,
logarithms with base e. It is developed to the more convenient form
given on page 10 in the following manner:
From equation 49
~2Trm(s Si)
logeX=2.30259 logiocc
2.30259 Qlogw ^
thus P= -= ?
2irm(s Si)
If the rate of pumping is expressed in gallons a minute, the equation?
becomes
2.30259(/l,440g Iog10-) 527.7g Iog10-
p= V g /_ <
Si) m(s Si)
which is Thiem's formula in modified form, for convenient use in
the United States.
Thiem's formula may be developed more simply by starting with

the assumption of a horizontal water table or piezometric surface
(fig. 2). For water-table conditions the demonstration is as follows:
i, at any point on the cone of depression, is equal to the slope, or

fjfij
-/> and the increment area through which the flow dQ moves is
equal to xyd<f>. Therefore
---- ................ (52)
PLAN
Pumped well
SECTION
_. . . , , Pumped well
Static water table ^ \,
FIGURE 2. Plan and section showing assumed ground-water conditions for the development of the for-
mula from horizontal water table.
With horizontal conditions, x and y are independent of $ ; hence $

may be integrated independently of x and y, and equation 52 becomes
/[*] * - -(53)
and .(54)
By integrating x and y,
Cdx 2irP C , ,.-.
I -0- I ydy--.-- - _-__(55)
27T.1
log«ic=
and « -*_*
PLAN
Observation we//\
J^
Observation we//-^
Pumped well.
SECTION
Observation
Pumping / we//s \i
Static piezometr/c surface r
Conftn/nghed'
/ / 7 / / /
FIGUKE 3. Plan and section showing assumed ground-water conditions for the development of the formula
from horizontal artesian conditions.
Equation 57 is identical with equation 40 obtained by Thiem. The

formula for computing the coefficient of permeability may be devel-
oped from equation 57 in the same manner as it was from equation 40.
Thiem's formula for artesian conditions can be developed in a
manner similar to the one that applies to water-table conditions (fig. 3).
The demonstration is as follows, again beginning with
The hydraulic gradient at any point on the cone of depression is

dlJ
equal to -£ The total area through which the flow, Q, passes is
2m-xm, where the water bed is horizontal. Therefore
and
Cttl dx 2irmP C hl
By integrating, = n <fy___ ___ _____________(60)
Ja <*' V Jh
,N ,_.,.
i h)--- ___________ (61)
ana
and
Equation 62 is developed to equation 4 in the same way that equation

4 is obtained from equation 50.
CONFIRMATION OF THIEM'S FORMULA FROM OTHER
WORK DONE IN THE UNITED STATES
The theoretical work by Slichter and by Turneaure and Russell,

done several years before Thiem's paper was published, is briefly
described below, for the purpose of showing that Thiem's formula
can be deduced from their results. Thus Thiem's formula is given
essential confirmation by these eminent hydrologists in the United
States, for the particular conditions, to which it applies namely, a
homogeneous water-bearing material, an original horizontal water
table or piezometric surface, and an original uniform thickness of
the saturated part of the water-bearing formation. Thiem's formula
has essentially been derived by the others, but because of differences
in their use of symbols and in the final form of their formulas, these
formulas have not generally been recognized as being different
manners of expressing identical conclusions. Thiem's outstanding
contribution was in the application that he made of his formula for
determining permeability.
The following discussion outlines the development of Slichter's
formula and points out its relation to Thiem's formula for artesian
conditions. 16 Slichter starts with the assumption of a homogeneous
water-bearing material overlain and underlain by impervious mate-
rial, and an artesian well that completely penetrates the material.
16 Slichter, C. S., Theoretical investigation of the motion of ground waters: U. S. Geol. Survey 19th
Ann. Kept., pt. 2, p. 359, 1899.
The following nomenclature and units are changed somewhat from

that of Slichter to correspond to the nomenclature and units used
previously in this paper:
ra, thickness of water-bearing material in feet (Slichter's a):
K, a constant defined by Slichter as "the quantity of water that
would be transmitted in unit time through a cylinder of
stone of unit length and cross section, under unit difference
in head at the ends";
v, velocity of the ground water, in feet per day;
r, radius of the well, in feet;
Q, rate of discharge of the pumped well, in gallons a day;
h, amount of lowering of water in well by pumping, in feet;
J, a point on the cone of depression;
x, distance of point J from the axis of the well, in feet;
Z, pressure at point J, in feet of water (Slichter's p) ;
B, distance, in feet, from the wall of the well at which the pressure
may be assumed to be equal to its normal value (that is,
Z=0 when x=R+r}.
The velocity at point J at distance x from the axis is given by
KdZ
The velocity varies inversely with the distance from the axis of
the well, so
v=-
x
in which c is a constant to be determined. After equating 63 and
64,
__________ __ ...__. (65)
C
From which c loge x=KZ+C1. ...... _ ._____- (66)

When x=r, Z=h, and when Z=0, x=R+r-, thus
cbg.(B+r) = C'1 -(67)
and cloge r Kh=d __ ___ __ _ ____(68)
By equating 67 and 68,
cloge (fl+r)=c log. r-l£ft______ __-__-___ (69)
Therefore c= ^-x___.________________ (70)
and by substituting equation 70 in 64,
y ................(71)
20 CONTRIBUTIONS TO HYDROLOGY OP UNITED STATES, 1935
The velocity at the wall of the well is found by placing x r. An

expression for the total amount of water flowing into the well in
unit time is obtained by multiplying the velocity at the wall of the
well by 2irrm. Therefore
At this point Slichter solved equation 72 for Q, having determined

K by means of a previously developed formula that depends upon the
laboratory analysis of the water-bearing material for the effective
size of the sand grains. However, equation 72 can be converted to
Thiem's formula for permeability by proper substitution. Solving
equation 72 for K, we get
K==
^ '___ ................ (73)
2-jrhm
K, as previously defined, is really a coefficient of permeability. There-
fore the symbol P may be substituted for K, giving
Q
P=
(R+r) corresponds to the distance ax in Thiem's formula, and r
corresponds to the distance a. Equation 74 can then be written
The term h is equal to the draw-down at the pumped well, at the dis-
tance r from the axis of the well. The draw-down at the distance
(R-\-r) was assumed by Slichter to be zero. Therefore, h represents
the difference in draw-downs between the two points on the cone of
depression r and (B-\-r) and is equivalent to Thiem's term (s «i).
Substituting in equation 75, we have
p_ $ lo&T = Qqogea1 -logea)

----- --------
Thiem's final formula for artesian conditions is identical with formula
76.
Turneaure and Russell 17 published the development of *a formula
which is similar to the simple development from horizontal water-
table conditions given on page 16. Using Darcy's law as a basis
for their development, they arrived at essentially the same equation
as 57, with the exception that the factor p (porosity of the water-
a Turneaure, F. E., and Russell, H. L., Public water supplies, 1st ed., p. 269, John Wiley & Sons, Inc.,
1901.
bearing material) is included. As given by Turneaure and Russell,

the equation is
_. __....___- _(77)
At this point c is evaluated by substituting x=r (radius of the well)
and y=h (saturated thickness of water-bearing material at the wall
.of the well). Then
c=Q\oger-TrKph2 .......... .....(78)
and by substituting in equation 77
Q logo
'+" -----"-<? >
If in equation 79 the value of x is taken to be R, or the distance from
the axis of the well at which the change in water level is inappreciable,
the corresponding value of y will be H, the original depth of water, and
equation 79 will become
and
log.
The product Kp corresponds to Thiem's coefficient of permeability,
P. Hence
In equation 82, H and h represent the thicknesses of the saturated

part of the water-bearing bed at R and r, respectively. In Thiem's
formula hi and h represent the thicknesses of the saturated part of
the water-bearing bed at di and a, respectively. The characters used
in Thiem's formula may be substituted in equation 82, and that equa-
tion then becomes
____ _ (83)
Cl 1
V °Sea^
This equation is Thiem's final formula 48, for computing the coef-
ficient of permeability from water-table conditions.
As shown above, there is little difference between the formulas of
Slichter, Turneaure and Russell, and Thiem. The principal variance
occurs in that Thiem determined the coefficient of permeability,
18274 36 4
whereas the others determined the quantity of water entering the well
and obtained the coefficient of permeability from laboratory analyses
of the water-bearing material. Thiem's formula includes the draw-
down of the water level in observation wells at two definite and meas-
urable points on the cone of depression, but the formulas of the others
contain the draw-down at the more indefinite points r and R. The
draw-down at the wall of the well at a distance r from the axis of the
well has usually been taken to be the water level in the well while
pumping was in progress. This sometimes introduces a large error,
because a part of the draw-down in the pumped well is caused by the
loss of head of the water as it enters the well. Moreover, the texture
of water-bearing material, if it is sand or gravel, is likely to be dis-
turbed for several feet around a pumped well by the development of the
well, and therefore the effective diameter r may be considerably
larger than the nominal diameter of the well.
The formulas of Slichter and of Turneaure and Russell include the
determination of the radius of the cone of depression, R. Slichter
assumed . this distance to be 600 feet, and Turneaure and Russell
determined it with a formula derived by the following reasoning:
"Assuming that all the water in the circle of influence flows into the
well, the width of the strip of the ground-water stream tributary to
the well will be 2R, and the original cross section of this portion of th&
ground-water stream is 2RH." Then from formula 1, Q=PI(2RH)
and R o PJTJ' By substituting the value of Q, from equation 82 the
formula, after reduction, becomes
_ ,..............
................(85)
2IH log."
This formula involves the draw-down in the pumped well and the
radius of the well, and therefore it is subject to the difficulties
previously enumerated in ascertaining these items. It is certain
that R is rather difficult to determine, and under some conditions it
has been known to exceed 5,000 feet.
Recently the results of laboratory experiments on the flow of
water through sand by Wyckoff, Botset, and Muskat 18 were pub-
lished. They constructed a small apparatus in which the ground-
water conditions around a pumped well were reproduced. They
observed the draw-downs of the water table and piezometric surface
at several distances from the well under various rates of flow. A
formula was prepared from the data obtained from these experiments,,
by which the flow into a well could be computed from a knowledge
is Wyckofl, R. D., Botset, H. Q., and Muskat, M., Flow of liquids through porous media under the-
action of gravity: Physics, vol. 3, no. 2, pp. 90-113, August 1932.
of the permeability of the water-bearing material and the saturated

thicknesses of the material at two points on the cone of depression.
The formula with the nomenclature altered to correspond to usage
in this report is
in which K is a coefficient of permeability, p is the density of the

fluid, g is the acceleration of gravity, and h and AI are the fluid pres-
sures at the respective distances a and ax from the pumped well.
Equation 86 may be written
Q^ -" """ '" '" '" " (87)
The density of water is essentially 1, and therefore p may be regarded

as equal to 1. Kg in equation 87 is therefore equivalent to the
coefficient of permeability, P, as contained in Thiem's formula.
The fluid pressure is probably equivalent for most conditions found
in nature to the saturated thickness of the water-bearing materials.
Therefore
.... ...............(88)
which may then be reduced to equation 48 (p. 15).

FORMULA FOR DETERMINING THE CONE OF DEPRESSION
A formula can be developed for the cone of depression from Thiem's
formula, provided the conditions are the same as those assumed in
developing Thiem's formula for artesian conditions the water-
bearing material is homogeneous and of uniform thickness, the ground
water is confined between horizontal impermeable formations, and
the piezometric surface is horizontal before pumping is started (fig. 3).
From Thiem's modified formula (equation 4, p. 15)
527.7 Io 10
Let the quantity p^ be represented by B, a constant. Then
which is the equation for the cone of depression pertaining to artesian

conditions. The draw-down, s, at any point on the cone of depression
can be computed by substituting the corresponding figure for a, the

distance from the pumped well at which the draw-down occurs.
The equation for the cone of depression pertaining to water-table
conditions is developed as follows from equation 4:
527.72 Iog10~ 1055.4g log^1

P= - = - 2 ---------(91)
1055.'_.
or h2*=hi2 p ________________(92)
Let the quantity 105p'4g be represented by F, a constant. Then
h? is equal to (fif s) 2 (fig. 2); therefore
^......__.........(94)
and s=flr- 12-^logio 1 -- ------- (95)
The draw-down, s, at any point on the cone of depression can be com-

puted by substituting the corresponding figure for a, the distance from
the pumped well at which the draw-down occurs.
The slope of the cone of depression at any point may also be com-
puted from Thiem's formula. Starting with the formula as stated in
equation 49 (p. 15),
s ~Sl ~
Let 'the quantity % p be represented by E, a constant. Then
............... .(97)
If di is a fixed point in any given pumping test, then log^ is a con-
stant in that test. Therefore the quantity E log^x is also a constant
and may be represented by L. Then
s Si=L E \ogea _____ __ ___ (98)
and s=i+Si Elogtd _ _ __.___.__. (99)
tT. 8. GEOLOGICAL SURVEY WATER-SUPPLY PAPEK 67ft PLAffl i
COEFFICIENT OF PERMEABILITY, P
oi «f II ! ! II I I I ! in
-90O
X
-SCO
i-
-TOO
-20O
-IOO
0 .1 .2 .3 .4 .5 .6 .7 .3 .9 1.0 U 1.2 1.3 IA 1.6 1.6 1.7 1.6 1.9 2.O 3.1 22 2.3 2.4 2.6 26 2.7 2.8
i ! DIFFERENCE IN DRAW-DOWNS(s-St)IN\FEET \ \
400 , 500 i 60O ! 7OO ! 800 '. 900 , 1,000 1,100 1,200 1,3OO 1,400
'DISCHARGE OF PUMPED. WELL 7, //V GALLONS PER MINUTE \
0 IO 2O 3O 40 SO 6O 7O 80 30 100 110 12O 13O 14O I5O 160 17O 180 190 ZOO 2IO 220 23O 24O 25O 26O 27O 28O
THICKNESS OF SATURATED WATER-BEARING MATERIAL, m, IN FEET
GRAPH FOR COMPUTING COEFFICIENTS OF PERMEABILITY,

THIBM METHOD FOR DETERMINING PERMEABILITY 25
Subtracting both sides of equation 99 from h, we have
By substituting the general factors (h s}=y and a=x in equation

100 (fig. 3), we get
Si _-- (101)
and by differentiating with respect to y and x,
dy=^- -__-_-__. (102)
The slope of the cone of depression is equal to dy/dx. Therefore,
and i= 2irPmx
0 ____________ _________(104)
^ J
In a similar manner it can be shown that for water-table conditions

the slope of the cone of depression at any distance, x, from the pumped
well can be computed by the formula
---- --008)
This formula differs from equation 104 for artesian conditions only in
that the thickness of the water-bearing formation, m, is replaced by
the thickness of the saturated water-bearing material, y.
The slope of the cone of depression, i, at any point may be computed
for both water table and artesian conditions by substituting for x,
the distance of the point from the pumped well.
GRAPHIC SOLUTION OF THIEM'S FORMULA
By the use of the graph presented in plate 1 the coefficient of

permeability may be determined without the usual computations,
if the factors contained in Thiem's formula are known. This graph
is particularly useful for determining the effect on the computed
coefficient of permeability of changes in the factors in Thiem's-
formula, and it provides a rapid method for calculating the per-
meability for several regions and comparing the ground-water condi-
tions of those regions. The graphic solution for permeability is
made in the following manner:
Locate the point of intersection of the horizontal line correspond-

ing to -Cb and of the vertical line corresponding to (s s:} ; move inward
or outward along the radial line through this point to the intersection
of the line with the vertical line representing the discharge of the well,
q; move horizontally from this point to the right or left to the vertical
line representing the saturated thickness of the formation; move
radially outward from this point to either the upper or the right mar-
gin of the diagram, where the coefficient of permeability, P, is read.
As an example, the graphic solution corresponding to the conditions
-=2,
df
8 s1 =1.5, 2=800, and m=100
is shown in plate 1 by a heavy dashed line. The coefficient of per-
meability so (determined is approximately 850.
PUMPING TESTS IN NEBRASKA
Two pumping tests were made near Grand Island, Nebr., during
the summer of 1931 on the farm of Fred Meyer, about 4 miles east
of Grand Island, in the NW% sec. 17, T. 11 N., R. 8 W. This location
was selected after a thorough inventory of existing irrigation wells
in the vicinity, as the one that most nearly approached the ideal
conditions desired for the pumping tests. The irrigation well used
for test 1 was in a pasture just west of a large field of corn (well
83, fig. 4). The land near the well was rather flat, although the
field of corn was slightly higher than the pasture. There was a dry
slough about 800 feet west of the well, but as no drainage had entered
it for some time preceding the pumping tests, it probably did not
affect the normal level of the ground water. Throughout the area
covered by figure 4 the water table ranged only from 2 to 10 feet
below the land surface, and hence the sinking of observation wells
was not difficult. It is probable that the water table was lowered
sjomewhat during the period of tests by drafts made on the zone of
saturation by plants, but the amount of lowering was small,, as indi-
cated by the small decline of the water table in those wells located
farthest from the pumped well. There were three irrigation wells
within a mile of the test wells, but none of them were operated during
the tests or for several days before the tests were begun.
Before the pumping tests were made a test hole was drilled near
observation well 76 to determine the thickness of the water-bearing
materials. Sand and gravel showing a great range in size and some
clay were penetrated to a depth of about 110 feet, where bedrock
was struck. The hole was continued into the bedrock to a depth of
143 feet below the ground surface. Later a well was drilled about
25 feet south of the existing irrigation well for the second pumping
test, and samples of the water-bearing materials penetrated were sent
to the hydrologic laboratory of the United States Geological Survey
for determinations of porosity, moisture equivalent, and permeability
and a mechanical analysis (table 1). A log of the materials encoun-
tered in this well (84, fig. 4) is given in table 2. This well was 12
inches in diameter and was drilled to a depth of 105 feet; the lower
48 feet and the upper 24 feet of the casing were perforated. The
existing irrigation well used for the first pumping test was 24 inches
in diameter and 40 feet deep, and all the casing was perforated.
TABLE 1. Physical properties of samples of alluvium taken from well 84, near
Grand Island, Nebr.
[Determined in the hydrologic laboratory of the U. S. Geological Survey by V. C. Fishel]
Mechanical analysis (percent by weight)
Depth (feet) Larger 0.125- 0.062- Less than

than 2.0 2.0-1.0 1.00-0.50 0.50-0.25 0.25-0.125 0.062 0.005 0.005
mm mm mm mm mm mm mm mm
6 to 10... 29.7 16.9 18.9 17.1 15.4 1.3 0.4 0.2

10 to 16.. 14.1 17.9 31.2 30.4 5.5 .3 .2 .1
16 to 20.. 16.8 15.2 25.8 29.4 10.5 1.6 .5 .1
20 to 25.. 18.6 18.8 21.3 24.8 13.8 1.9 .2
25 to 30.. 7.5 17.2 25.0 30.0 16.0 3.4 .a
30 to 39.. 36.4 20.8 21.4 15.0 4.7 .8 .5 .1
39 to 40.. 3.4 3.6 1.8 4.7 26.0 14.0 31.5 13.6
40 to 42.. 15.9 11.0 20.1 33.4 15.4 2.6 .4 .2
42 to 46.. 15.4 15.2 20.2 19.5 16.4 7.0 4.5 1.5
46 to 51.. 17.3 10.7 13.1 29.4 24.4 3.2 1.0 .4
51 to 55.. 39.6 12.8 9.5 15.7 13.5 4.7 2.5 1.0
55 to 61.. 27.4 14.9 16.3 22.4 11.8 3.7 2.1 1.0
61 to 66.. 20.6 19.6 19.7 19.1 9.7 4.3 4.4 .9
66 to 71.. 18.1 18.0 17.7 23.7 14.0 3.3 3.0 1.9
71 to 78. . 179.3 3.5 3.9 6.3 4.0 1.5 1.0 .3
78 to 86. . 14.3 11.9 18.2 25.1 18.7 6.5 3.0 1.7
86 to 92.. 36.2 10.3 14.6 17.1 11.6 4.3 3.4 2.3
92 to 99.. 15.1 10.4 22.8 31.1 13.9 3.0 2.5 1.0
99 to 105. 25.8 13.3 13.7 21.9 14.3 5.2 3.5 2.0
Moisture equivalent
Apparent Porosity Coefficient
Depth (feet) specific (percent) of perme-
gravity Percent Percent ability
by weight by volume
6 to 10... 90 27.1 1.4 2.6 480

10 to 16.. 30.9 1.5 2.7 1,685
16 to 20.. 32.3 1.1 2.0 1,460
20 to 25.. 28.5 1.4 2.6 1,095
25 to 30-. 31.0 1.4 2.6 1,095
30 to 39.. 30.6 1.0 1.9 4,350
39 to 40- 40.3 17.4 27.1 2
40 to 42.. 31.2 1.5 2.7 925
42 to 46.. 26.3 1.6 3.0 150
46 to 51.. 30.2 1.6 3.0 350
51 to 55.. 26.2 1.7 3.3 780
55 to 61-. 25.6 1.6 3.0 730
61 to 66-. 25.0 1.4 2.8 2,095
66 to 71.. 26.3 1.6 3.0 1,050
71 to 78.. 22.8 2,185
78 to 86.- 41.8 1.6 3.1 220
86 to 92.. 21.5 1.9 3.9 495
92 to 99.- 29.9 1.2 2.1 430
99 to 105- 27.6 1.5 2.9 285
J 76.0 percent larger than 5 mm.
TABLE 2. Log of well 84, drilled for second pumping test
Thick- Thick-
ness Depth ness Depth
Feet Feet Feet Feet

Top soil ___________ 16 71
38 39 7 78
Clay. - .. . .... 40 8 86
10 50 6 92
5 55 13 105
TT. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 679 PLATE 2
A. LINE SW OF 1-INCH OBSERVATION WELLS. B. MEASURING THE DEPTH TO THE WATER TABLE.
U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 679 PLATE 3
A. WEIR FOR MEASURING THE DISCHARGE OF THE PUMPED WELL.
B. PUMPING ARRANGEMENT IN SECOND TEST.

THIEM METHOD FOR DETERMINING PERMEABILITY
A transit was set up over well 83, and six radiating lines of wells
were laid out. Lines C and D were projections of lines A and B,
and line W bisected the 90° angle formed by the intersecting lines
A and B. lines N and S only approximately bisected the angles
formed by the intersections of lines A and D and lines B and C,
because the topographic features were such that actual bisections
would have been difficult. Line SW was laid out from well 84
(pl. 2, A).
More than 80 observation wells were sunk, most of them relatively
close to the pumped wells, where the decline of the water table during
pumping would be the greatest. Some of the observation wells were
1 inch in diameter and were fitted with 18-inch screen drive points.
These wells were driven into the saturated sand and gravel -to such
depths that the water table during pumping would not drop below
the bottoms of the wells. Several observation wells 3 inches in
diameter were fitted with drilling bits at their lower ends and were
jetted down with a drilling rig. Holes in the bits allowed water to
enter the wells freely. The diameters and depths of the observation
wells are recorded in table 3.
TABLE 3. Location, diameter, depth, and altitude of wells used in the pumping
Distance
Depth of meas-
Diam- of well uring Altitude of D istance
from
Distance
from
Well no. Line eter below point measuring pumped pumped
measur- above point
ing point land well 83 well 84
surface
Inches Feet Feet Feet Feet Feet

1 ___ ... A 3 21.4 0.7 1, 814. 34 24.9 42.3
2 ____________ __ __ ... A 3 10.3 .1 1, 815. 66 59.9 74.6
3. __ ............. __ .......... A 3 10.1 0 1, 815. 26 114.4 127.9
4. __ A 3 10.3 .1 1, 814. 63 164.2 177.2
5 _________ - ___ .... _ ... A 1 11.6 1.4 1, 815. 83 229.0 241.5
6 ...- _ .- . . .-. ... A 1 6.5 .5 1, 812. 35 354.1 366.4
7...... .......... ............. A 3 10.2 .2 1, 815. 39 429.3 441.2
8 _ --_ . .--...--.._.- A 1 11.4 1.2 1, 815. 52 478.9 490.4
9..--.. .. -.---......-..-. A 3 10.3 .1 1, 814. 97 604.0 616.1
A 3 10.3 1, 814. 73 754.6
11 A 3 .4 1, 814. 05 903.8 916.2
13 B 91 4 1, 814. 84 29.9 14 9
14 B 3 10.9 .3 1, 815. 17 70.0 49.6
16--- B 1 11.6 1.3 1, 816. 10 120.0 98.9
16 B Q Q 1, 815. 67 184.9 163.1
17 B 3 9.9 .2 1, 815. 46 254.7 233.0
18 B 3 in 9 .3 1, 815. 08 375.3 353.8
19 B 1A 1 .5 1, 815. 86
20 B 3 in 9 .4 499.7 477.7
21, B 3 9.9 .3 1, 816. 32 649.7 627.6
22 B 9 a .3 1, 816. 39 775.3 752.9
23 B 1 11.5 1.9 1, 816. 95 974. 3 951.8
24 . B 0
10.6 1, 817. 12 1, 149. 3 1, 127. 0
26 W 30 .3 1, 815. 39 49.7 48.8
26 W 1A ^ .3 1,814.78 170.0 164.3
27 W 3 .3 1, 815. 33 270.0 264.0
28 W .1 1. 813. 46 430.0 vIOQ A
29 W q 10.1 .3 1, 815. 68 625.0 618.0

30 W 3 10.7 .3 1,815.39 804.5 797.6
31 W 3 939.7 932.5
32 D 21.5 1.2 1, 820. 42 40.1 63.1
33 D 1 16.5 1.0 1, 819. 17 95.1 117.7
34. D 16.5 1.2 1,818.99 144.7 166.9
35 . D 1 16.5 1.2 1. 818. 93 214.3 236.7
TABLE 3. Location, diameter, depth, and altitude of wells used in the pumping
tests Continued
Distance
Depth of meas-
Diam- of well uring Altitude of Distance
from
Distance
from
Well no. Line eter below point measuring pumped pumped
measur- above point
ing point land well 83 well 84
surface
Inches Feet Feet Feet Feet Feet

36 D 1 16.5 0.9 1, 818. 31 323.8 345.8
37 D 1 16.5 .9 1, 818. 27 423.2 445.0
38 D 16.5 1.1 1, 818. 33 448.2 470.2
39 D 16.5 1.0 1, 818. 05 572.9 594.4
40 D 16.5 1.2 1, 818. 83 722.7 744.2
D 12.7 1.8 1, 818. 08 872.2 893. 6
42 D 1 1, 817. 19 1, 072. 5 1,094,2
43 D 12.7 .5 1, 816. 73 1. 197. 0 1, 218. 3
44 C 23.0 1.0 1, 820. 12 39.3 35.0
45 C 1 17.2 .5 1, 818. 37 80.5 71.9
C 11.1 .5 1, 818. 02 130.3 120.2
47 C 12.7 .5 1, 817. 37 195.6 185.0
48 C 12.6 .8 1, 817. 90 285.6 274.7
49 __ C 1 11.4 1,818.36 410.2 398.7
60 C 12.3 .5 1, 818. 30 425.2 413.9
51 C 12.4 1, 818. 80 535.4 524.2
52 C 12.7 .7 1, 819. 45 685.3 673.9
53 C 1 12.6 822.8
54-.- __ C 12.7 .5 1, 034. 7 1,022.9
55 C 12.6 .6 1, 817. 37 1, 174. 9 1, 162. 8
56. - sw 11.0 1 ST4. Q7 46 7 OE Q
57 sw 10.5 9 69.5 50.7

58 _ - sw 1 10.6 .8 1, 815. 29 93.6 75.8
59- __ _ sw 11.0 .7 1, 816. 18 118.0 100.8
60 _ _ .... ___ ...... sw 1 11.0 .8 1, 815. 88 216.9 200.7
61 ___ . ___ .. .... sw 1 10.6 .9 1, 816. 14 316.6 300.7
62 sw 10.9 400.8
63 _ . .................. sw 10.7 .8 1, 816. 47 516.5 500.9
64- __ . __ . _ .. sw 10.8 2.1 600 9
65- _____ ...... sw 1 10.9 2.4 716.5 701.0
66 _ _ .......... sw 12.5 2.6 1 Q17 4.9 816.6 801 2
67 _ . ___ .... sw 1 10.9 2.4 Qfll 9
68 ___ . _ ...... sw 1 11.1 29 1, 816. 54 1, 016. 6 I ftm *?
69- . sw 1 6.1 1.3 1, 815. 37 1, 116. 9 1, 101. 5
70-. , _ .... sw 1 11.4 O A
1 QTT QQ 1, 217. 1 1, 201. 8
71 . _ .... _ . A 1 B.19 RQ 2.6 26.0
72 ......................... A 1 12.3-l DO O
73 . ___________ . _ . s 1 Qi K QX m 105.3
74-- . _ . _______ . ..... s 12 0 .7 1 ci K nn OOK O 9fifi Q
75-- ________ ...... 8 1 12.6 .5 1, 816. 05 OKK 1
76--- ___ ... _ ...... s 1 13.0 9 3 1, 817. 74 382.7 qee A
77 N 1 6.1 1, 813. 08 Q7 1
78-- . _ ........ N 1 12.8 1.7 1, 815. 32 160.0 183.5
79 ....................... N 1 to n .7 1,815.48 OCR 0
80- ____ ............ N 12.3 1 9 1, 816. 13 342.0 365.7

81 _ ___ ....... ..... N 11.8 1 3 1, 816. 41 445.8 469 0
82- . _____ ......... SW 97 ft 0 04. Q
2 A
83.- _ . ............ 24 on K -.5 1, 812. 66 0 f)A Q
84 _ ............. ............... 12 102.0 -3.0 1 1 R14 On t)A Q
0
i First test. Altitude for second test 1,812.35 feet.
Each observation well was pumped with a pitcher pump until the
water discharged was clear, indicating that the ground water had
free access to the well and that the water level in the well showed the
level of the water table outside the well. Definite points were
established at each well from which measurements of depth to the
water level could be made, and the distance of these measuring points
above the land surface was recorded. To determine the altitude of
the measuring points, instrumental levels were run to all the obser-
vation wells and to the two pumped wells (table 3).
Both pumping tests were started early in the morning. During the
day preceding each of the tests several measurements were made of
the depth to water in the observation wells, in order to determine
the static level of the water table, and a few minutes before pumping
began additional measurements were made as a check on the measure-
ments of the day before. The measurements were made with a steel
tape graduated in hundredths of a foot. The end of the tape was
loaded with a swiveled weight, so that the tape would hang plumb, and
the lower foot or so of the tape was coated with blue carpenter's chalk,
so that the depth of immersion of the tape into the water could be
plainly seen. The period of pumping in the first test was about 48
hours, and the average rate of pumping 540 gallons a minute. During
the second pumping test the pump was stopped several times because
of trouble with the 50-horsepower gasoline engine that was used to
drive it. In order to make the two tests as comparable as possible,
pumping in the second test was continued a few hours longer than
the 48-hour period, so that the total quantity of water pumped was
about equal to the quantity pumped during the first test. Records
of pumping time are given in table 4.
TABLE 4. Record of pumping time
Well 83 (test 1) Well 84 (test 2) Continued
Started___ July 29, 1931, 6:05 a. m- Started___ Sept. 10, 1931, 9:32 a. m.
Stopped.___ July 31, 1931, 6:04 a. m. Stopped_____________ 9:36 a. m.
Started....___________ 9:38 a. m.
Well 84 (test 2)
Stopped________________ 9:39 a.m.
Started___ Sept. 9, 1931, 8:05 a. m' Started_____________ 9:40 a. m.
Stopped__________-_____ 11:18 a. m' Stopped_________-______ 9:48 a. m.
Started_______________ 11:35 a. m. Started...______________ 9:51 a. m.
Stopped________________ 12:35 p. m. Stopped________________ 11:17 a. m.
Started_________________ 12.37 p. m. Started,.____________ 11:19 a. m.
Stopped._______________ 2:00 p. m. Stopped_____-__________ 11:49 a. m.
Started_______________ 3:38 p. m. Started ------..-.---- 11:55 a. m.
Stopped_____-_-________ 5:55 p. m. Stopped...._____.-- 12:06 p. m.
Started_______________ 6:31 p. m. Started____________ 12:11 p. m.
Stopped__ Sept. 10, 1931, 4:26 a. m. Stopped_ Sept. 11, 1931, 10:28 a. m.
Started___________ 6:03 a. m. Started___________ 10:34 a. m.
Stopped____ __________ 8:57 a. m. Stopped__-_______-_---- 2:05 p. m.
Twelve men were employed during the tests to make measure-

ments of the depth to water in the observation wells, to measure the
discharge of the pumped wells, and to operate the power unit. The
men worked on alternate shifts of 6 hours during the first test and
8 hours during the second test. During each shift three men made
measurements of depth to water in the observation wells (pi. 2, B).
These measurements were continued throughout the night with the
aid of lanterns. The water from the pumped wells was discharged
into a stilling basin about 25 feet east of the wells (pi. 3, J5), from
which it flowed over a rectangular weir and through a small canal

across the field of corn (pi. 3, A). The water was then used to
irrigate the corn grown in the eastern part of the field.
After the completion of each pumping test measurements of depth
to water in the observation wells were continued for at least 24 hours,
so that the recovery of the water table could be determined. The
measurements of depth to water made during pumping and after
pumping had stopped, which number about 9,500, are on file and may
be consulted in the office of the United States Geological Survey at
Washington.
RESULTS OBTAINED FROM THE PUMPING TESTS
A considerable amount of study has been devoted to the records

obtained in the pumping tests above described in an effort to deter-
mine the best procedure to be used in future tests for determining-
permeability by the pumping method. The present tests involved
more time and expense than could ordinarily be spent on a field
determination of the permeability of a water-bearing formation, and
it was with the view of reducing the complexity of the tests that this
study was made. The conclusions given below, under the heading
"Computations of coefficients of permeability", show that satis-
factory results can be obtained by less elaborate tests if certain facts
developed in these tests are kept in mind.
The data collected in the tests are so adequate that a detailed study
could be made of the behavior of the ground water in the vicinity of
a pumped well, both during pumping and after pumping stopped.
However, the writer has been able to make only a rather cursory
examination of the whole mass of data, and hence the results here
presented are not all that could be obtained if a more intensive study
were made.
DRAW-DOWN CURVES
To obtain the draw-down of the water table at any time it is only

necessary to subtract the depth to the water level before pumping
started from the depth at that particular time. The altitude of the
water level at any time can be obtained by subtracting the draw-down
at that time from the altitude of the normal water level. A con-
tinuous curve representing the decline of the water level in a well
during the period of pumping is called a "draw-down curve."
Draw-down curves were plotted for many of the observation wells,
chiefly for the first pumping test, because the curves for the second
test show irregularities caused by interruptions in pumping. The '
draw-down curves for the first test are remarkably regular. A smooth
curve could be drawn through most of the points, and the very few
points that plotted far from the curve were obviously caused by errors
in making the measurements. The equation for the draw-down

<;urve is not known, but the general form of the curve is shown by
typical curves given in plate 4 for wells on line A. As that line was
approximately at right angles to the natural direction of ground-water
movement, the altitude of the static water level was nearly the same
in all the wells.
The slope of the draw-down curve indicates the rate at which the
water was withdrawn from storage in the sand and gravel. The form
of the draw-down curves of the wells close to the pumped well indicates
that a large volume of water was withdrawn from storage immediately
after pumping began, the amount withdrawn being in general
inversely proportional to the distance from the pumped well. The
draw-down curves of the wells comparatively far from the pumped
well indicate that there was no withdrawal of water from storage at
those distances from the pumped well for several hours after pumping
began and that the maximum rate of withdrawal of water was not
reached for some time after the first water was withdrawn. Obviously
this lag was caused by the fact that all the water necessary to supply
the pumped well was at first obtained from the sand and gravel
nearby. The draw-down curves given in plate 4 show that the water
levels in the observation wells were still declining after 48 hours of
pumping. In other words, the cone of depression around the pumped
well had not reached a condition of equilibrium.
The draw-down curves of observation wells near the pumped well
reflected unavoidable changes in the rate of pumping. The pump
was stopped several times during the second pumping test, and these
ishut-downs caused the water levels in nearby observation wells to
rise. Typical draw-down curves for the second test, given in plate 5,
indicate that water continued to percolate toward the pumped well
during the periods of interruption in pumping, and that this water
began to refill the sand and gravel that had been unwatered during
the periods of pumping. Of course, as soon as pumping was resumed
the water table was again lowered. However, this lowering was
resumed from a new level the level caused by the rise of the water
table during the period in which there was no pumping and plate 5
shows that sometimes it took several hours to lower the water level
to the point where it stood before pumping was stopped. Only the
wells comparatively close to the pumped well showed a rise of water
level during the interruptions in pumping. The wells farther away
showed a continuous decline of the water table, thus indicating that
the water which caused the rise of the water levels in wells close to the
pumped well came, at least in part, from storage in the area farther
from the pumped well. There were several interruptions of pumping
in the second test, and consequently considerable water was taken
out of storage at some distance from the pumped well and stored
temporarily in the sand and gravel close to the pumped well. This
process tended to reduce the draw-down in the wells close to the
pumped well, and therefore the coefficients of permeability computed
from draw-downs observed in the second test are believed to be
greater than the true permeability of the material. For this reason
more study has been devoted to the first test, in which pumping was
carried on at a nearly constant rate. It is probable that the
inequalities in the cone of depression in the second test would have
disappeared if pumping had been continued without interruption
for several more days.
RECOVERY CURVES
A recovery curve is a continuous curve representing the movement of

the water level in an observation well after pumping has stopped.
Depending upon the location of the well, the movement after pumping
has stopped may be either an immediate rise of the water level or a
decline that is eventually followed by a rise. Measurements of the
depth to the water level in the observation wells were continued in
both tests after pumping had stopped in order to determine the rate
and amount of recovery of the water table. Several typical recovery
curves are shown in plates 4 and 5 as continuations of the draw-down
curves. The recovery curves are usually smoother than the draw-
down curves because of the absence of irregularities caused by inter-
ruptions in pumping or variations in the rate of pumping. The
recovery curves show that in observation wells close to the pumped
well the recovery was most rapid immediately * after pumping had
stopped, whereas in observation wells comparatively far from the
pumped well the recovery was most rapid several hours after pumping
had stopped.
After pumping stopped, water continued to percolate toward the
pumped well under the hydraulic gradient set up during the period of
pumping, but instead of being discharged by the well it refilled the
interstices in the sand and gravel that had been unwatered by the
pumping. As the unwatered sand and gravel was gradually refilled
the hydraulic gradient toward the well decreased, and the flow toward
the well decreased proportionally. Thus the rate of recovery became
progressively slower. The water level in wells comparatively far from
the pumped well declined for several hours after pumping stopped.
In these areas water continued to be taken from storage to supply the
water that refilled the sediments around the pumped well. Of course,
in time there was a general equalization of water levels over the entire
region, and the water table assumed a form similar to that it had before
pumping began. However, the ultimate level of the water table was a
little lower than before pumping began, because water had been
permanently removed from the zone of saturation during the period of
pumping.
The rate of recovery of the water table is in general inversely pro-
portional to the distance from the pumped well. However, this is
true only for a short time after pumping ceases. Even though the
water table close to the pumped well initially has a greater amount to
recover, the rate of rise after a certain time is the same as the rate of
rise of the water table at greater distances. After pumping is stopped
the water table close to the well rises until the possible amount of
recovery remaining that is, the remaining draw-down is equal to
the remaining draw-down at distances farther from the pumped well.
This is specifically shown in table 5. After 48 hours of pumping the
decline of the water level in well 1, 24.9 feet from the pumped well,
was 4.03 feet, and the decline in well 2, 59.9 feet from the pumped well,
was 2.81 feet. After 2 hours of recovery the remaining draw-down in
each well was about 1.68 feet, and the rates of recovery for the next
22 hours were the same in the two wells. After 48 hours of pumping
the draw-down in well 3, 114.4 feet from the pumped well, was 2.03
feet about half of the draw-down of well 1. After 12 hours of re-
covery the remaining draw-down in all three wells was 0.77 foot, and
the rates of recovery from that time on were the same. If the measure-
ments of depth to water had been continued longer, the indicated rate
of recovery in the wells farther from the pumped well would eventually
have been nearly the same as the rate in wells 1, 2, and 3.
TABLE 5. Draw-down of the water table during test 1, at several times after pumping
stopped
Dis- Draw-down (feet) at time indicated (hours) after pumping stopped

tance
from
Well no. pumped
well 0 2 4 6 8 10 12 14 16 18 20 22 24
(feet)
24.9 4.03 1.187 1.32 1.12 0.98 0.86 0.77 0.70 0.64 0.59 0.55 0.51 0.48
2. ... ... 59.9 2.81 1.69 1.33 1.13 .99 .87 .77 .70 .64 .59 .55 .51 .47
3 _ 114.4 2.03 1.49 1.25 1.17 .94 .85 .77 .70 .64 .59 .55 .51 .48
4. 164.2 1.62 1.30 1.11 .98 .88 .80 .72 .67 .60 .56 .52 .49 .46
5. 229.0 1.14 1.01 .91 .83 .76 .70 .64 .59 .55 .51 .48 .45 .43
6 ....... 354.1 .65 .64 .60 .56 .53 .50 .48 .45 .43 .41 .39 .37 .35
7 .-.. ... 429.3 .52 .51 .49 .47 .45 .43 .41 .40 .38 .36 .35 .33 .32
8. ...... 478.9 .44 .44 .44 .43 .42 .41 .39 .38 .36 .35 .33 .32 .30
9 ....... 604.0 .26 .27 .28 .28 .28 .28 .28 .27 .27 .26 .26 .25 .24
10 754.6 .15 .16 .16 .17 .17 .17 .17 .17 .18 .18 .18 .18 .17
11 903.8 .11 .11 .11 .12 .12 .12 .12 .13 13 .13 .13 .14 .14
Table 5 also illustrates the decline of the water table after pumping
ceases in wells comparatively far from the pumped well. Recovery
started almost at once in wells 1 to 7, but in well 8 there was a lag of
a few hours, and it was 6 hours before the water level reached a point
0.01 foot above its level at the time when pumping stopped. In
well 9 there was an actual decline of 0.02 foot during the first 4 hours
of recovery, and it was not until 14 hours after pumping had stopped
that there was any recovery from this low level. In well 10 there was
a decline for 22 hours after pumping had stopped, and in well 11 the
water level was apparently still declining after 24 hours. This lag
is also shown by the recovery curves in plate 4.
The writer has devoted some time in an attempt to develop an equa-
tion for the recovery curves and their relation to the permeability of
the water-bearing material. It would seem that the rate of recovery
of the water level in an observation well is dependent on the quantity
of water pumped, the draw-down of the water level at the time pump-
ing stopped, the distance of the observation well from the pumped
well, the initial hydraulic gradient, the thickness of the water-bearing
formation, and the permeability of the formation. No equation was
found that could be used for the draw-down curves of all the obser-
vation wells. However, the following general equation is suitable
for many of the curves:
R= D
rpn l !
where R is the recovery of the water level, in feet; D is the draw-down

from static water level at the time pumping stopped, in feet; K is a
coefficient for each particular well; T is the elapsed period of recovery,
in hours; and n is an exponent. K and n contain the varying distance
factor as well as several other constants enumerated above. The
formula is based on the assumption that when T equals 0, R equals 0;
and when T equals infinity, R equals D. Of course, if pumping were
carried on over an extended period during which there were no
recharge, R would probably never equal D, because of the permanent
withdrawal of water from storage. In figure 5 the recovery curve of
well 5 is plotted as determined from the theoretical equation, and the
actual field measurements are also indicated.
CONES OF DEPRESSION
Soon after pumping begins the water table around a pumped well
assumes a form which is comparable to an inverted cone, although
it is not a true cone. Where the water-bearing material is homoge-
neous, the so-called "cone of depression" will be circular if the initial
water table is horizontal, but elliptical if the initial water table is
sloping. The form of the cone of depression at any time can be shown
by either profiles or contours on the water table. Profiles at different
angles with the direction of initial slope of the water table may differ
widely in form. The profiles in figure 6 are based on the draw-downs
in wells on lines B and D, which have nearly the same direction as
the initial slope of the water table. The development of the cone is
shown by the several profiles, and it is interesting to note the rate at
which the radius of the cone increased with the period of pumping.
The profile of the cone after 2 hours of pumping shows that most of
tJ. 8. GEOLOGICAL SURVEY
WATBB-SUPPLY PAPEB 879 PLATE 5
Probable draw-down
Pump not operating
_______________________________________- i i < i.i i i.i ' I I II I

4p.m. 6p.m. 8p.m. 10p.m. 12p.m. 2a.m. 4am. 6a.m 8a.m. fOam. 12m. 2pm. 4p.m. 6p.m. 8p.m. 10p.m. 12p.m. 2am. 4a.m. 6a.m. 8a.m. 10a.m. 12m. 2p.m. 4pm. 6p.m. 8p.m. tOpjn. 12pm. 2am. 4am. Sam. 8am.
TYPICAL DRAW-DOWN AND RECOVERY CURVES FOR TEST 2.

U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 679 PLATE 4
&IQ.4O
W0.30
Water level
Pump not operating
6a.m 8a.m. !0a.m. 12 m. 2p.m. 4p.m. 6p.m 8 p.m. 10p.m. 12p.m. 2a.m. 4a.m. 6a.m. 8a.m. 10a.m. 12m. 2p.m. -4p.m. 6p.m. 8p.m. 10p.m. 12p.m. 2a.m. 4a.m. 6a.m. 8a.m. 10a.m. 12m. 2p.m. 4p.m. 6p.m. 8pm. K)p.m. 12p.m. 2am-4a.m. 6a.m. 8a.m. 10a.m.
TYPICAL DRAW-DOWN AND RECOVERY CURVES FOR TEST 1.

C. B. GEOLOGICAL SUBVE* WATEB-SUPPLT PAFXR 878 PLAT* 6
Contour int«rv«l OJQ foot
CONTOURS ON THE WATER TABLE BEFORE PUMPING AND AT SEVERAL TIMES AFTER PUMPING BEGAN.
1, Before pumping began; 2, noood boor; 3, nidi hoar; 4, twelfth hour; 5, twenty-ftmrth hoar; 6, forty-eighth hour.
the water pumped came from the unwatered sediments within a

distance of 430 to 540 feet from the pumped well, because there was
no draw-down of the water table beyond those distances. After 6
hours of pumping the radius of the cone had increased to about 800
feet, and after 12 hours of pumping there were small but measurable
draw-downs at 1,200 feet from the pumped well. Further pumping
undoubtedly increased the radius of the cone beyond the most distant
observation wells.
It is probable that the rate at which the radius of the cone of depres-
sion develops depends somewhat upon the specific yield of the water-
bearing sand and gravel. Most of the water that is discharged by the
pumped well during the first hours of pumping is taken directly from
the sediments around the pumped well. The amount of water taken
from storage in this manner is indicated by the draw-down of the
water table, which provides a method for determining the specific
yield of the sediments (p. 9). If the water-bearing materials have
a low specific yield, the radius of the cone of depression will probably
develop more rapidly than if the materials have a high specific yield,
for if the specific yield is low, there is less water available in the sedi-
ments, and hence the effect of pumping will be transmitted outward
from the pumped well at a more rapid rate. The actual difference
in the rate at which the radius of the cone of depression will develop
in materials having different specific yields depends, of course, not
alone on the specific yield but also on the relation between the specific
yield and the permeability of the materials.
The slope of the cone of depression is steeper up-gradient from the
pumped well than down-gradient (fig. 6). This indicates that if the
permeability is the same, less water is percolating to the well from the
down-gradient side. The slope of the cone down-gradient from the
well becomes progressively less than the slope at the corresponding
distance up-gradient, until at some distance down-gradient from the
well the water table is horizontal. This point is called the ground-
water divide. All water below this divide percolates away from the
pumped well, and all water above this divide percolates toward the
well. The ground-water divide moves down-gradient as the pumping
period is increased and the cone of depression becomes larger. The
ground-water divide in test 1, as indicated in figure 6, was about 280
feet down-gradient from the pumped well after 2 hours of pumping.
The divide gradually moved to about 360 feet below the pumped well
after 6 hours of pumping, 440 feet after 12 hours, 500 feet after 24
hours, 560 feet after 36 hours, and about 600 feet after 48 hours.
Contours on the water table before pumping began and at several
times after pumping began are shown in plate 6. The limits of the
area included in these maps are somewhat less than the distance of the
farthest observation well from the pumped well, because there were
too few observation wells to furnish data for contours over a

larger area. The contours of the original water table were drawn
readily by direct interpolation between the altitudes of water levels
in the observation wells, but to draw the contours of the water table
at given times after pumping began it was necessary to plot a profile
of the cone of depression for each of the several lines and to determine
the distance of each contour from the pumped well by inspection of
the profile. A certain amount of judgment had to be used in tracing
the contours between the several lines of wells.
The map showing contours on the water table before pumping
began (map 1, pi. 6) indicates that the initial slope of the water table
was nearly in the direction of the lines B and D and that this slope was
very uniform, averaging about 6.9 feet to a mile. The contour
interval selected was 0.1 foot for all the maps. These maps were
drawn to illustrate the development of the cone of depression and
especially the movement of the ground-water divide. The cone of
depression is well denned close to the pumped well on map 2, showing
contours on the water table after 2 hours of pumping, although there
are some inequalities which are caused by the steep slope of the cone.
The contours are nearly circular close to the well but gradually become
elliptical at greater distances up to the ground-water divide. The
divide as shown on these maps is a semielliptical line between the
water which eventually enters the pumped well and that which per-
colates on down-gradient. A draw-down of the water table at some
distance from the pumped well does not necessarily indicate that the
water at that distance lies within the ground-water divide. It may
indicate only that the draw-down of the water table created a slope
of the water table that tended to move the water from its normal
path, but this water may not reach the pumped well. The contours
in map 2 show there was a draw-down of the water table for several
hundred feet down-gradient from the ground-water divide. Also
maps 2 and 3 show that there was a draw-down of the water table on
lines A and C beyond the ground-water divide. The ground-water
divide has been represented on maps 2 to 6 as the only line normal to
the contours on the water table. Its position can be drawn in from
the points of inflection of the contour lines. The steady decline of the
ground-water divide down-gradient can be observed in maps 2 to 6,
as well as by the profiles of the cone of depression in figure 6. The
contour maps show in addition the lateral development of the divide
and the increase in cross-sectional area through which water percolates
to the pumped well.
The development of the cone of depression can probably best be
seen by discussing one contour for example, the 1,809.5-foot con-
tour on map 2 and observing the position of this line on the contour
maps for subsequent times. After 2 hours of pumping this line is
somewhat elliptical, for it crosses line B at a point 85 feet up-gradient

from the pumped well and line D at a point 120 feet down-gradient
from the pumped well. After 6 hours of pumping this contour is still
closed but includes a larger area. Its intersection with line B has
advanced up-gradient to a point about 130 feet from the pumped well,
and its intersection with line D has declined down-gradient to a point
about 210 feet from the pumped well. After 12 hours of pumping this
contour is no longer closed, and during the remaining 36 hours of pump-
ing it diverges still more on the down-gradient side of the well, while
its intersection with line B steadily moves up-gradient. The develop-
ment of the cone close to the pumped well as the period of pumping
lengthens can readily be seen by observing the increase in area included
in the first closed contour, which is the same for maps 3 to 6.
After 2 hours of pumping the contours 300 to 400 feet up-gradient
from the pumped well were affected but little by the pumping. As
pumping continued the lines became closer together (indicating an
increase in the slope of the cone of .depression), and after 48 hours
they had a noticeable curvature.
The movement of the ground water was, ot course, always normal
to the contours on the water table. Continuous lines drawn normal
to the contours are called "lines of flow" of the ground water, and
they trace the path of movement of a particle of water. The lines
of flow on map 1, plate 6, are approximately parallel, but on maps 2
to 6 the lines of flow included between the ground-water divide
and the pumped well converge toward the well. The distances
through which particles of water moved hi reaching the pumped well
differed considerably. Thus a particle of water 400 feet from the
pumped well on line B traveled to the pumped well through a much
shorter path than a particle of water at the same distance from the
pumped well on line A (map 6). Moreover, the particle of water on
line B had to move with a greater velocity than the particle on line A,
because the average hydraulic gradient along its path was greater.
Consequently, the quantity of water that passed through a unit area
on line B was much greater than the quantity that passed through a
similar area on line A.
COMPUTATION OF COEFFICIENTS OF PERMEABILITY
Thiem's equation 4 was used for computing coefficients of per-

meability, and as the draw-downs at only two points on the cone of
depression are required for the computation of a coefficient, the data
obtained in the pumping tests are sufficient for a great many compu-
tations. Two general methods were used. In one method coeffi-
cients were computed by using the draw-downs that were measured
in the observation wells, and in the other method coefficients were
computed by using the interpolated draw-downs at selected distances
from the pumped well, which were obtained from profiles of the cone
of depression.
The coefficients of permeability computed by using the difference
in draw-down at any two points on the cone of depression would be
equal if the iorm of the observed cone of depression was the same as
that of the theoretical cone of depression obtained by Thiem's formula.
However, the cones of depression in both pumping tests were not
identical with the theoretical cone, and the computed permeability
ranged through wide limits. Computations of permeability were
made by using the draw-downs in all possible combinations of obser-
vation wells on line A, after 48 hours of pumping, in test 1 (table 6).
The coefficients of permeability thus computed ranged from 535 to
5,630. The equation used was
Tt
527.7X540Xlog^ Clr
in which the symbols are those given on page 10.

TABLE 6. Coefficients of permeability computed by Thiem's formula for all possible
combinations of observation wells on line A
Well no. 71 72 1 2 3 4 5 6 7 8 9 10
72...... _ .. 535
1.. ...... ...... . 615 840
2............... 668 885 919
3_ -..-.. 690 927 974 1,058
4....... ........ 729 949 1,000 1,079 1,112
5_ - 746 941 976 1,014 983 866
6 -.-_ --. 767 956 998 1,038 1,030 1,045 1,188
7...... ......... 785 984 1,027 1,096 1,098 1,262 1,830
8_. .......... ... 778 998 1,043 1,130 1,315 1,786 1,760
9............... 810 1,018 1,180 1,192 1,374 1,700 1,630 1,605
10 -- 835 1,052 1,113 1,196 1,292 1,500 1,880 1,895 1,950 2,505
11...... _ ..... 845 1,090 1,160 1,260 1,345 1,420 1,660 2,150 2,255 2,320 3,320 5,630
An inspection of table 6 shows that the computed coefficients were

the smallest when both of the observation wells selected were close
to the pumped well and the largest when both were far from the
pumped well. This indicates that the difference in draw-down of
wells close to the pumped well was great in comparison to the dif-
ference in draw-down of wells farther from the pumped well. In
other words, the cone of depression had developed very little at
distances far from the pumped well. These wide variations make the
determination of the most nearly correct coefficient almost an im-
possibility from computations of this kind. The coefficients computed
from combinations of observation wells on other lines gave similar
results.
Coefficients could be computed, of course, from draw-downs at
any time after pumping started. If the difference in draw-down
changed during the period of pumping, the computed coefficient
THIEM METHOD FOR DETERMINING PERMEABILITY
would likewise change. In order to observe the effect of the length

of the period of pumping on the computed permeability, coefficients
were computed for several combinations of wells on line A and for
several periods of time after pumping began (table 7). In general,
the computed coefficients became smaller as the period of pumping
increased, because the difference in draw-down became larger. How-
ever, the range in coefficients was smallest when the observation wells
were selected close to the pumped well. At all times the coeffi-
cients became larger as the farthest observation well was selected
farther from the pumped well, indicating that the form of the cone
of depression was not equal to that of the theoretical cone.
TABLE 7. Coefficients of permeability computed by Thiem's formula for several
combinations of observation wells on line A and for several periods of pumping
Coefficients at different times (hours) after pumping began

Well nos.
2 6 12 24 36 48
1 and 2 ____________________ 960 868 879 935 918 919

1 and 3. ___________________ 1,110 958 948 1,003 981 974
1 and 4. _ ...--.-......... -...... .. 1,180 1,008 974 1,025 1,003 1,000
1,268 1,037 987 1,008 982 976
1 and 6- ___________________ 1,413 1,132 1,040 1,040 1,010 998
1 and 7.. . .. ..... ....... 1,510 1,198 1,090 1,076 1,043 1,027
1,550 1,225 1,107 1,095 1,060 1,043
Another computation was made by using the draw-downs obtained

from profiles of the cone of depression. The draw-downs at several
distances on lines A, B, C, and D were determined after 48 hours of
pumping, and coefficients were computed by using a=50 feet and
aj equal to several distances (table 8). The computed coefficients
of permeability vary in about the same manner as those shown in
table 6, but the variation is considerably less. The greatest varia-
tion occurred on line B, but this was only from 823 to 1,180. On lines
C and D the coefficients computed for aj = 75 feet were larger than
the coefficients obtained for ai = 200 feet, but on lines A and B the
opposite was true. This fact suggests that the difference in the
variation of the coefficients may be caused by the initial slope of
the water table, because lines C and D are extensions of lines A and
B, respectively. In general, this suggestion is confirmed by aver-
aging the coefficients computed for several distances on line A with
those computed for the same distances on line C, and similarly
averaging the coefficients computed for lines B and D (table 8).
The averages for lines A and C are very nearly the same as the aver-
ages at corresponding distances on lines B and D. The differences in
the coefficients could be due to the fact that the wells penetrated
materials of varying permeability, but it does not seem probable that
the materials would be distributed areally in such a manner that all
wells on one line penetrated a material of one permeability and all

wells on another line penetrated a material of another permeability.
TABLE 8. Coefficients of permeability computed by Thiem's formula, using
a 60 feet
Coefficients
ai (feet) Average Average

Line A Line B LineC Line D of lines of lines
A and C BandD
75 957 823 1,032 1,200 995 1,012

100 - ---------- ___ . _ ............. 994 856 980 1,208 987 1,032
150 1,055 888 952 1,114 1,004 1,001
200 1,025 914 970 1,102 996 1,008
250 1,000 929 987 1,087 994 1,008
300 ....... .... _ 1,008 943 1,002 1,090 1,005 1,017
350 1,022 958 1,018 1,100 1,020 1,029
400 - . 1, 050 978 1,030 1,104 1,040 1,042
450 1,074 994 1,050 1,115 1,062 1,055
500 1,096 1,020 1,075 1,132 1,081 1,076
550 1,112 1,040 1,100 1,152 1,106 1,096
600 ...... 1,130 1,067 1,122 1,168 1,126 1,118
650 1,140 1,086 1,147 1,182 1,144 1,134
700 1,156 1,109 1,170 1,198 1,163 1,154
750 1,171 1,130 1,180 1,216 1,176 1,173
800-.. .. ...... ....... 1,192 1,147 1,210 1,237 1,201 1,192
850 1,215 1,169 1,229 1,260 1,222 1,215
900 1,232 1,180 1,224 1,280 1,228 1,230
As the variations in the coefficients of permeability are caused by

differences between observed and theoretical draw-downs it is well to
examine the observed draw-downs and to determine the manner and
amount of their deviation from theoretical draw-downs. The ob-
served draw-downs of the water table as taken from profiles of the
cone of depression after 48 hours of pumping in test 1 for several dis-
tances and directions from the pumped well are given in table 9. The
draw-downs decrease regularly with the distance from the pumped
well, but the draw-downs at equal distances from the pumped well are
not equal. However, as would be expected from the previous averag-
ing of the coefficients, the average of the draw-downs on lines A and C
at equal distances from the pumped well are very nearly equal to the
average draw-downs on lines B and D at the same distances from the
pumped well.
The computations of coefficients of permeability have indicated that
the computed coefficients become larger as the observation wells are
selected at greater distances from the pumped well. As the differences
in draw-down are substituted in the denominator of Thiem's formula,
the reason for this increase in the coefficients is that the difference in
draw-downs is relatively too small. This, in turn, indicates that the
cone of depression has not reached a condition of equilibrium. The
draw-downs at several distances on line A were averaged with the
draw-downs at corresponding distances on line C, and the differences
in draw-down (s sO were computed, s being taken as equal to the
average of the draw-downs on lines A and C at 40 feet from the
TABLE 9. Draw-down of water table, in feet, after 48 hours of pumping during test 1,
for several distances and directions from the pumped wett
Draw-down
Distance from pumped well (feet) Average Average

Line A LineB LineC Line D of lines of lines
A and C B and D
50.. ......... ......... ...... . ........... 3.06 3.14 2.98 2.93 3.02 3.08
75.. . 2.52 2.505 2.48 2.50 2.52 2.50
loo....................................... 2.17 2.11 2.08 2.20 2.13 2.16
150............ ........................... 1.74 1.57 1.52 1.68 1.63 1.63
200... ................................. 1.35 1.22 1.17 1.33 1.26 1.28
260.. . ................................. 1.03 .95 .92 1.06 .96 1.01
300. . . . ......... ... .82 .74 .73 .86 .78 .80
350.. ...... ........................... .66 .575 .57 .70 .62 .64
400.. .......... . .57 .455 .44 .56 .51 .52
450. ................................... .49 .36 .35 .445 .42 .41
500.----.. ..... .41 .29 .28 .37 .35 .33
650 .................................. .34 .23 .23 .305 .29 .27
600 .28 .195 .195 .245 .24 .23
650-. _ .. ... . .22 .16 .16 .195 .19 .18
700. .................. ............... .18 .14 .14 .155 .16 .15
750...... ................... . ....... .15 .115 .12 .13 .14 .13
800... . ........... ............. .125 .085 .10 .11 .11 .10
850 .12 .075 .075 .10 .10 .09
900 ............. ............... .105 .065 .06 .085 .08 .08
950... .......................... ...... .06 .05 .08 .07
1,000.. . ..... ................. .05 .04 .065 .06
1,050 .05 .025 .065 .06
pumped well. These differences in draw-down are given in table 10.

They increase as the distance from the pumped well increases. The
difference in draw-downs between 40 and 60 feet from the pumped
well decreased as the period of pumping increased, and for distances up
to 200 feet the difference was practically the same for the final 24 hours
of pumping. For distances greater than 200 feet the difference in
draw-downs increased throughout the period of pumping.
TABLE 10. Differences in average draw-down on lines A and C, in feet
Difference at different times (hours) after

pumping began
2 6 12 24 36 48 >48+
40 and 60 ... 0.56 0.55 0.54 0.53 0.53 0.53 0.53

40 and 80. -_.-_ .88 .95 .94 .91 .92 .91 .90
40 and 100---------....--- ----- ------------- - 1.08 1.21 1.23 1.20 1.20 1.19 1.19
40 and 120_ 1.24 1.41 1.45 1.43 1.42 1.44 1.42
40 and 140 1.37 1.57 1.64 1.61 1.61 1.63 1.62
40andl60 1.46 1.70 1.78 1.76 1.77 1.79 1.79
40andl80 1.53 1.81 1.92 1.91 1.92 1.93 1.95
40and200. 1.59 1.91 2.04 2.04 2.05 2.07 2.08
40and240.. - 1.67 2.04 2.23 2.25 2. 26 2.29 2.32
40 and 280... --------- 1.72 2.15 2.36 2.41 2.43 2.46 2.52
40and320._ - - 1.76 2.23 2.47 2.53 2.55 2.60 2.69
40and380 1.79 2.29 2.55 2.64 2.67 3.71 2.84
40aad t4ttU 1.81. 2.33 2.61 2,71 2.75 2.80 2.98
40and>;800. __ 1.83 2.39 2.72 2.85 2.92 2.99 3.27
1.84 2.42 2.78 2.93 3.02 3.11 3.51
40 and 700-.. 1.84 2.43 2.81 2.97 3.08 3.19 3.71
40 and 800. 1.84 2.43 2.82 3.00 3.12 3.23 3.88
40 and 900. _ 1.84 2.43 2.83 3.01 3.13 3.25 4.03
1.84 2.43 2.84 3.03 3.15 3.28 4.12
1 Theoretical difference in draw-down after the cone of depression reaches equilibrium.

The data given in table 10 indicate that after 48 hours of pumping

the cone of depression had reached approximate equilibrium in form
for a distance of about 200 feet from the pumped well. The cone of
depression is said to have reached "approximate" equilibrium,
because as the water table was lowered the cross-sectional area through
which the water percolated was decreased, and the form of the cone
was changed slightly to compensate for this decrease. An inspection
of table 10 shows that the cone had reached approximate equilibrium
at 160 feet from the pumped well after only 12 hours of pumping.
This means that at these distances from the pumped well the coeffi-
cients of permeability computed between 12 to 48 hours after pumping
began would be approximately equal to the coefficients computed after
a much longer period of pumping. With the equation developed for
the cone of depression, the theoretical draw-downs at various distances
from the pumped well were computed, and the differences in draw-
downs were obtained. These are given in the last column in table 10
and are the theoretical differences in draw-down that would occur
after the whole cone of depression had reached a condition of equilib-
rium. The table shows that for distances up to about 200 feet from
the pumped well the theoretical and observed draw-downs are practi-
cally equal, but that for greater distances the theoretical difference is
larger than the observed difference. Hence the coefficients of per-
meability computed from the differences in draw-down up to 200 feet
would be nearly equal, but the computed coefficients would increase
beyond that distance.
It has been shown that the observed draw-down of the water table
does not equal the theoretical draw-down at all points on the cone of
depression; hence values for the coefficient of permeability may vary
widely when computed directly from the difference in draw-down of
two observation wells. However, the observed draw-downs on one
side of the pumped well when averaged with the observed draw-downs
at corresponding distances on the opposite side of the well approached
or equaled the theoretical draw-downs at those distances. Coefficients
of permeability computed from the draw-downs averaged in this man-
ner were nearly equal lor that part of the cone of depression that had
reached approximate equilibrium in form that is, for distances from
about 40 to 200 feet from the pumped well. This indicates that the
pumping method can be used in the field and that it will yield consist-
ent results provided the observation wells are located within that
part of the cone of depression that has reached approximate equilib-
rium and on a straight line through the pumped well preferably on
a line along the natural hydraulic gradient and on both sides of the
pumped well.
The coefficients of permeability computed from various combina-
tions of the draw-downs of the water table between the limits of 40 and
200 feet from the pumped well in test 1 differ but little and average
THIEM METHOD FOB DETERMINING PERMEABILITY 47
about 997 (table 11). In table 1 are given the coefficients of permea-
bility of samples of the sand and gravel taken from well 84 during the
process of drilling, as determined in the hydrologic laboratory of the
United States Geological Survey. These coefficients reach a maximum
of 4,350, but their average, weighted as to thickness, is about 1,200.
TABLE 11. Final computation of coefficients of permeability from the average
draw-downs on lines A and C where the cone of depression had reached approxi-
mate equilibrium in form
0.1 , a\ TO S Sl
(feet) (feet) loga~ (feet) (feet) P
60 40 0.176 96.94 0.53 976

80 40 .301 97.14 .91 970
100 40 .398 97.28 1.19 980
120 40 .477 97.40 1.44 969
140 40 .544 97.50 1.63 975
160 40 .602 97.58 1.79 982
180 40 .653 97.65 1.93 987
200 40 .699 97.71 2.07 985
80 60 .125 97.40 .38 962
100 60 .222 97.54 .66 983
120 60 .301 97.67 .91 965
140 60 .368 97.76 1.10 975
160 60 .426 97.84 1.26 985
180 60 .477 97.91 1.40 992
200 60 .523 97.98 1.53 994
100 80 .097 97.73 .28 1,010
120 80 .176 97.85 .53 967
140 80 .243 97.95 .72 982
160 80 .301 98.03 .88 994
180 80 .352 98.10 1.02 1,002
200 80 .398 98.16 1.15 1,005
120 100 .079 98.00 .25 919
140 100 .146 98.09 .44 964
160 100 .204 98.17 .60 987
180 100 .255 98.24 .74 1,000
200 100 .301 98-30 .87 1,003
140 120 .067 98.22 .19 1,023
160 120 .125 98.30 .35 1,035
180 120 .176 98.37 .49 1,040
200 120 .222 98.43 .62 1,043
160 140 .058 98.39 .16 1,050
180 140 .111 98.46 .30 1,071
200 140 .155 98.52 .43 1,043
180 160 .051 98.54 .14 1,053
200 160 .097 98.60 .27 1,038
200 180 .045 98.68 .13 1,000
Average, 997.
Two methods for determining the most probable coefficient of per-

meability from the observed data are suggested by C. E. Van Ors-
trand, geophysicist of the United States Geological Survey. These
.methods as explained by Mr. Van Orstrand are as follows:
The value of P is to be determined from the equation
(40)
One method of obtaining a close approximation to the value of P

consists in assuming at least three values of P} preferably at equal
intervals and over a range so great that the true value of P falls
between the extremes. From each observed value of y we thus obtain
.a value of C as shown in table 12. The average value of C, 8,409.15,
and the corresponding value of P, 975, are then substituted in for-

mula 40, to obtain y e, the computed value of y.
TABLE 12. Computation of v2 for P=975 and Q = 777,600 gallons a day
(feet) ^log.x c y° y* »
40 913, 063. 35 8, 410. 54 96:68 96.67 +0.01

60 1, 013, 421. 91 8, 410. 37 97.21 97.20 +.01
80 1, 084, 630. 29 8, 411. 37 97.59 97.58 +.01
100 1, 139, 861. 41 8, 409. 45 97.87 97.87 .00
120 .. _ _ .................... _ _. 1,184,988.85 8, 412. 16 98.12 98.10 +.02
140 1, 223, 143. 71 8, 410. 35 98.31 98.30 +.01
160 1, 256, 194. 76 8, 407. 94 98.47 98.48 -.01
180 1, 285, 349. 88 8, 405. 62 98.61 98.63 -.02
200 1, 311, 428. 35 8, 404. 54 98.74 98.76 -.02
8, 409. 15 S V. ...... .0017
2,500
775 875 975 1,075 1,175

VALUES OF P
FIGUBE 7. Relation of S»2 to P.
The differences (r) between tile observed and computed values, the
residuals, are tabulated in the last column, and we find 2y2 =0.0017.
Proceeding in a similar manner for the values P=775 and 1,175,
we find for Sp2 the respective values 0.2835 and 0.0948. Plotting all
the values as shown in figure 7, we see at once that the value of P
THIBM METHOD FOR DETERMINING PERMEABILITY 49
that makes 2r* a minimum is slightly less than 1,000. This, in accord-
ance with the principle of least squares, is the most probable value
of P.
Another method of procedure arrives at the correct result in a
more direct manner.' Let us write down an observation equation
(40) for each observed value of y as given in table 12.
TABLE 13. Observation equations for y2
Observation equations v for y* y» »for y
C+ 91 3, 063 PI =9, 347 0 96.68 0 -0.01

C+l, 013, 422 PI =9, 450 +.01 97.21 .00 .00
C+l, 084, 630 PI =9, 524 +.02 97.59 +.01 +.01
C+l, 139, 861 Pi=9, 579 +.01 97.87 .00 .00
C+l, 184, 989 PI =9, 628 +.04 98.12 +.02 +.02
C+l, 223, 144 Pi=9, 665 +.02 98.31 +.01 +.01
C+l, 256, 195 PI =9, 696 -.01 98.47 +.00 .00
C+l, 285, 350 PI = 9, 724 -.03 98.61 -.02 -.01
C+l, 311, 428 Pi=9, 750 -.03 98.74 -.02 -.01
'4 2»2 .0014 .0009
Combining these equations by the method of least squares, we have

the normal equations .
9(7+ 10,412,082P1 = 86,363
10,412,082(7+12,186,687,602,840^=100,055,976,372
the solution of which gives P=980.083, £7=8,415.381. The residuals
corresponding to these values of P and C are given in the next to the
last column of table 13. These values are slightly erroneous, owing
to the fact that y2 instead of y was used in evaluating P and (7,
To obtain the correct values, let us compute the partial differential
coefficients in the equation
As dy is here the error in y that is, v we write the observation

equations in the form
dC Q]Qg.xdp _j.
L\ ^\ y-vrt Ifrjl U/U V,
2y 2iryP*
Substituting the appropriate values from tables 12 and 13, we have
0.005172<&7-0.004916dP= 0.00
0.005144rfCr -0.005426dP= 0.00
0.005123<#7 0.005785rfP= +0.01
0.005109d<7-0.006062dP= 0.00
0.005096c?<7 0.006286dP= +0.02
0.005086^(7 0.006476dP= +0.01
0.005078^(7- 0.006640dP= 0.00
0.005070<ft7-0.006785dP= -0.02
.005064<ft7- 0.0069 13dP= -0.02
50 CONTEIBUTIONS TO HYDROLOGY OF UNITED STATES, 1935
The solution of these equations gives dP 6.782, dC 8.162; hence

P= 980.083+6.782=986.87
(7=8,415.381+8.162=8,423.54
The substitution of these constants in equation 40 gives the values
in the last column of table 13. As shown in the table, the value of
S#2 has been reduced from 0.0014 to 0.0009 by the last solution.
Assuming that the work has been done correctly and that the num-
bers have been rounded off correctly, a further reduction hi the value
of Z,v2 is impossible.
DIFFERENCES BETWEEN FIELD AND THEORETICAL CONDITIONS
The conditions found in the field rarely approach closely the theo-
retical considerations from which Thiem's formula was developed.
The water-bearing formations usually differ from place to place in
thickness and in permeability. Wells usually do not extend to the
bottom of the formation, and the bottom is not always parallel to
the water table or piezometric surface. Most water tables and
piezometric surfaces are not horizontal, and few pumping tests are
continued until the cone of depression reaches approximate equilib-
rium in form over a large area.
The errors introduced into Thiem's formula by the differences be-
tween field and theoretical conditions can be minimized. The thick-
ness of the water-bearing material can be obtained from as many well
logs as possible, and an average value used for m. Of course, the
value used in Thiem's formula is not important if the computed
coefficient of permeability is to be used to determine the quantity of
water that percolates through some cross section of the material in
which the factor of thickness appears, because in that case m cancels.
The effect of variations in the permeability of the water-bearing ma-
terial can be lessened by selecting an average value for the computed
coefficient by one of the methods previously explained. Other
differences such as pumped wells that do not extend through the
formation, water tables that are not horizontal, and cones of depres-
sion that have not reached equilibrium in form cause part of 'the
cone of depression to differ from the theoretical cone. The effect of
these differences can be minimized by substituting in Thiem's formula
only the draw-down of the water table from that part of the cone of
depression which corresponds with the theoretical cone.
The reason why the observed cone of depression differs from the
theoretical cone may be made clear by a review of the behavior of the
water table during the first pumping test. As soon as the pump began
discharging water from the well a hydraulic gradient from all direc-
tions was established toward the well, and the water table was lowered.
The lowering of the water table unwatered a considerable volume of
sediments, the water in them gradually draining down to the new

water table and eventually entering the well. Thus most of the water
which the pumped well discharged was obtained from the unwatered
sediments. The decline in the water table close to the pumped well
was large at first, and then as a gradient adequate to discharge the
540 gallons of water a minute required for the well was formed, the
water table was lowered more slowly. Farther from the pumped
well the water table lowered less rapidly, because (1) water was being
supplied to the pumped well from sources closer to the well and (2)
the hydraulic gradient necessary to force the required amount of
water through a unit area of the sediments was smaller. Gradually
the water table around the pumped well assumed a slope that was
large enough to cause equal quantities of water to flow toward the
pumped well without withdrawing large amounts of water from
storage. Thus the slope of the cone of the depression for distances
up to 200 feet from the well reached essential equilibrium in form
after 48 hours of pumping. Beyond the distance of 200 feet, how-
ever, the slope of the water table was less than necessary to transmit
540 gallons of water a minute through cylindrical sections of the
sediments. If pumping had been continued the water table at
distances greater than 200 feet from the pumped well would have
continued to decline until the necessary hydraulic gradient had been
developed. To develop the necessary hydraulic gradient at distances
comparatively far from the pumped well, a very large quantity of
water must be taken from storage. Of course, most of the water
taken from storage must be discharged by the pumped well, and
hence the time required for the cone of depression to reach equilibrium
depends to some extent upon the specific yield of the water-bearing-
materials. The slope of the cone of depression necessary to force
540 gallons a minute through cylindrical sections is developed at
increasing distances from the pumped well only by the decline of the
water table at those distances. Thus it is evident that in order for
the cone to maintain its slope nearer the pumped well, the water
table must continue to decline there, but at a diminishing rate. This
decline in turn uriwaters more sediments, and the development of
the cone proceeds at a still slower rate.
The fact that the cone of depression reaches essential equilibrium
near the pumped well comparatively soon after pumping begins
provides the opportunity to use Thiem's formula. Fortunately the
difference in draw-down, s Si, is substituted in Thiem's equation, and
so long as this difference is constant, the coefficients will be equal.
Little error is introduced by the increase in absol ute draw-down as pump-
ing is continued. Thus, though the cone of depression has not reached
equilibrium it is still possible to use the formula. It is interesting to
observe the difference between the theoretical and observed draw-
downs shown in table 10. The differences are practically equal for
distances up to 200 feet from the pumped well, but for distances
farther from the pumped well the theoretical difference is larger.
The theoretical difference in draw-downs of the water table at 40
feet and 1,000 feet from the pumped well is 4.12 feet, and the observed
difference after 48 hours of pumping was 3.28 feet. This indicates
that with further pumping the water table at 40 feet and 1,000 feet
from the pumped well would lower in such amounts that the difference
between the declines would be increased by 0.84 foot. There is no
way of ascertaining the net decline of the water table at either of
these distances from the pumped well, as the net decline depends
upon the length of the period of pumping.
The pumped well used in test 1 did not extend through the water-
bearing sand and gravel, and it seems probable that this influenced
the cone of depression close to the pumped well. As shown in table 6,
the coefficients of permeability computed by using the draw-downs
in wells 71 and 72, 2.6 feet and 12.3 feet, respectively, from the center
of the pumped well, were comparatively small, indicating that the
draw-downs near the pumped well were relatively great. It seems
probable that these comparatively large draw-downs were an effect
of the well's failing to penetrate the entire thickness of the formation,
because the form of the cone of depression reached essential equilib-
rium from 40 feet to 200 feet from the pumped well, and the draw-
downs were of such magnitude that the computed coefficients of
permeability were practically equal, whereas the water table within
40 feet of the pumped well reached essential equilibrium but did not
correspond to the theoretical cone. A part of these large draw-downs
may also have been caused by changes in the permeability of the
formation due to the rearrangement of the sand and gravel during
the development of the well.
It is unfortunate that there were interruptions during the period .of
the second test, for these interruptions so changed the normal draw-
downs in the observation wells that the computations of permeability
from.the data obtained in this test are of doubtful value. Some
computations were made, and the coefficients averaged about 1,300
for wells 56 to 61 on line SW. Because there were interruptions
in pumping it would be expected that the draw-downs would be
smaller and consequently the coefficients larger. It is difficult to
make an intensive study of this test, because only line SW extended
through pumped well 84. The draw-down measurements made
during these tests indicate the behavior of the water table at several
distances from the pumped well when pumping is not continuous, and
the measurements of depth to the water table made after the pumping
stopped are valuable for determining the rate and amount of the
recovery of the-water table.
DETERMINATION OF SPECIFIC YIELD BY THE PUMPING METHOD
The data obtained during the first pumping test are adequate for
detailed study of the use of the pumping method for determining the
specific yield of water-bearing materials. This method was suggested
by Meinzer 19 and essentially consists of determining the ratio of (1)
the quantities of ground water that in a given time are taken from
storage between concentric cylindrical sections around the pumped
well to (2) the volumes of sediments between the cylinders that are
unwatered in that time. A preliminary report by the writer on this
method, with reference to the pumping tests described in this paper,
has been published.20 As the quantities of water taken from storage
are determined by ascertaining the difference in the quantities of
water that percolate through the cylinders in a given time, the
specific yield may be expressed by the equation
where y is the specific yield, YI is the quantity of ground water, in

cubic feet, that percolates through the smaller cylinder, Y is the quan-
tity of ground water, in cubic feet, that percolates through the larger
cylinder, and V is the volume of water-bearing material, in cubic
feet, that is unwatered between the cylinders.
The volumes of water-bearing material that were unwatered
between concentric cylindrical sections around the pumped well
during certain periods of pumping were computed by the formula
-- (107)
where V is the volume of unwatered material, in cubic feet; a is the

radius of the large cylinder, in feet; a,\ is the radius of the small
cylinder, in feet; and s and Si are the draw-downs of the water table,
in feet, at the distances a and a\, respectively, from the pumped well.
The draw-down at any given distance from the pumped well was
taken as the average draw-down at that distance on all lines. The
volumes of material that were unwatered during 6, 12, 24, 36, and 48
hours of pumping were determined by computing the volumes
unwatered between 22 cylinders whose radii ranged from 10 feet to
900 feet, but only the 'volumes of water-bearing material that were
unwatered between several selected concentric cylindrical sections
around the pumped well are given in this report (table 14).
19 Meinzer, O. E., Outline of methods for estimating ground-water supplies: U. S. Oeol. Survey Water-
Supply Paper 638, p. 136, 1932.
" Wenzel, L. K., Specific yield determined from a Thiem's pumping test: Am. Oeophys. Union
Trans., 1933, pp. 475-477.
TABLE 14. Volumes of water-bearing material, in cubic feet, that were unwatered
between concentric cylindrical sections around the pumped well for several periods
of pumping
Volume during different periods of pumping (hours)

Radii of cylinders (feet)
0-6 0-12 0-24 0-36 0-48
50 and 280. _________________________ 160, 500 227, 700 275,600 308, 300 336, 400
50 "arid 320 .- . .._ -.. .- 179,800 260, 100 319, 300 361, 400 397,000
50and»360 ..... .................... _ ............ 196,000 289,200 359, 500 411,400 455,200
50 and 400........ ................................... 209,400 315, 000 395, 800 458,200 510, 600
50 and 500......... .... __ __._- __ ... _ . _ -. 234,200 367, 300 474, 300 561, 500 632, 400
248, 900 401,000 532, 100 640,000 731,000
259, 100 421, 600 575,000 699,300 806, 700
263, 800 437,000 608,000 745, 200 865, 600
263,800 449,000 634,800 785, 300 916, 300
The quantities of water that were taken from storage between the
concentric cylindrical sections were determined by computing the
quantity of water that percolated through the smaller cylinder and
subtracting from it the quantity of ground water that percolated
through the larger cylinder. The quantities of water that percolated
through each cylinder in a given time were computed by the formula
7.48X24-------------- (108)
where Y is the quantity of ground water, in cubic feet; P is the coeffi-

cient of permeability; i is the average hydraulic gradient, in feet per
foot, at the distance a from the pumped well; h is the average thick-
ness, in feet, of the saturated water-bearing material at the distance a;
and T is the period of pumping, in hours. The quantities of water
that percolated through several selected cylindrical sections in short
periods were computed, and then a summation of these quantities
was made (table 15). The computations were first made for short
periods of time in order to decrease the error introduced by changes of
the hydraulic gradient, especially in the first few hours of pumping.
The periods of pumping given in table 15 were arbitrarily selected in
order to show the change in specific yield with the period of draining.
TABLE 15. Computed quantities of ground water, in cubic feet, that percolated
through several concentric cylindrical sections around the pumped well for several
periods of pumping
Quantity during different periods of pumping
(hours)
Jttaaius or cylinder (.leet;
0-6 0-12 0-24 0-36 0-48
SO........ ......................................... .. 23,800 51,400 104, 000 155, 400 206,800

280.... ...................................... ..... ... 8,000 22, 300 57,900 97,700 139,600
320... ............................................... 7,200 20, 300 53,000 89,800 128, 200
360............................................. _ .. 5,700 16, 700 45,200 79, 300 116,000
400.... _. , -__-..-___ __ ... 4,700 14,000 39, 600 70,400 105, 200
500... ... ................... .... .... . - _ - 2,000 8,100 26, 300 51,300 80,800
600..-............. ........... ...... ..... . 1,400 5,600 19,500 36,400 60,200
700. . . . 1,200 4,800 12,800 23,600 38,800
800..... .............................. ...-...-.-. . 700 2,700 8,800 18,100 29,400
900.... ... . ...-.-.-..-_-.__-.- _ 0 1,200 5,800 12,800 20,300
THIEM METHOD FOR DETERMINING PERMEABILITY OO
The cylinder with a radius of 50 feet was chosen as the control

cylinder, and the quantities of water that were taken from storage
were computed by determining the differences in the quantities of
water that percolated through that cylinder and the quantities that
percolated through larger cylinders (table 16). Specific yield was
then determined by dividing the quantities of ground water taken
from storage in certain periods of time (table 16) by the volumes of
water-bearing material that were unwatered in the corresponding
periods (table 14). The results are given in table 17.
TABLE 16. Quantities of ground water, in cubic feet, taken from storage between
several concentric cylindrical sections around the pumped well and for several
periods of pum.ping
Quantity during different periods of pumping

(hours)
Radii of cylinders (feet)
0-6 0-12 0-24 0-36 0-48
50 and 280.__.__ __ ................................ 15,800 29,100 46,100 57, 700 67,200

50 and 320..... ...................................... 16,600 31,100 51,000 65,600 78,600
50 ana'-SQO-L.. ............... __ . __ . _ -...-.--_. 18, 100 34,700 58,800 76,100 90,800
-50 and 400 .__.-_. --......_. __ ... _ .......... 19, 100 37,400 64,400 85,000 101,600
21,800 43,300 77, 700 104, 100 126,000
50 and 600.......... ________ __________ 22,400 45,800 84,500 119,000 146,600
50and700..... .-_..-.._.-.-.-_._._... 22,600 46,600 91,200 131. 800 168,000
SOandSOO....----- - _-.-- ..--. __ .-. 23,100 48,700 95,200 137, 300 177, 400
50 and 900.... ...... ___ .... ____ .... _ ........ 23,800 50,200 98,200 142, 600 186, 500
TABLE 17. Specific yield as computed for several concentric cylindrical sections and
for several periods of pumping
Specific yield during different periods of

pumping (hours)
Radii of cylinders, (feet)
0-6 0-12 0-24 0-36 0-48
«) and 28CL.......... ........................................ . 9.8 12.8 16.7 18.7 20.0

SOsaA-330......... .......................... .................. 9.2 11.9 16.0 18.2 19.8
50 and 360-...-..--.. __ ............... ___ ..... __ ....... 9.2 12.0 16.3 18.5 19.9
50 and 400________________________________ 9.1 11.9 16.3 18.5 19.9
SOandSOO........ ___________________________ 9.3 11.8 16.4 18.5 19.9
9.0 11.4 18.6 20.0
50 and 700.. ..... _ ........... _ . __ ... ____ ......... 8.7 11.1 15.9 18.8 20.8
.50 and 800........ __ . .. _ . ______ . ___ ------- 8.8 11.1 15.7 18.4 20.5
50 and 900-..-.- ____________________________ . 11.2 15.5 18.2 20.3
9.2 11.7 16.1 18.5 20.1
As an illustration, the specific yield will be computed for the volume

of sediments unwatered during 48 hours of pumping between concen-
tric cylindrical sections with radii of 50 and 280 feet. The average
slope of the cone of depression during the last 12 hours of pumping
was 2.59 percent at 50 feet from the pumped well and 0.369 percent
at 280 feet from the well. The quantity of water that percolated
through the 50-foot cylinder from 36 to 48 hours after pumping began
was as follows:
2X3.1416X975X0.0259X50X96.92X12
7.48X24 =51,400 cubic feet
This quantity added to the 155,400 cubic feet (determined in the same
manner) that percolated through this cylinder in the preceding 36
hours gives a total of 206,800 cubic feet ground water that perco-
lated through the 50-foot cylinder in 48 hours. The quantity of
water that percolated through the 280-foot cylinder in the last 12
hours of pumping was as follows:
V
2= 2X3.1416X975X0.00369X280X99.42X12
7 4SV24 Qnn cubic
=41,900 . . feet
,
This quantity added to the 97,700 cubic feet that percolated through
this cylinder in the preceding 36 hours gives a total of 139,600 cubic
feet of ground water that percolated through the 280-foot cylinder
in 48 hours. The quantity of ground water taken from storage
between the* 50-foot and 280-foot cylinders equals 206,800 minus
139,600, or 67,200 cubic feet; the volume of material that was un-
watered between these cylinders was 336,400 cubic feet (table 14),
Hence the specific yield is computed as follows:
67,200X100
336,400 ~^ U ' U
The computed specific yield becomes larger as the pumping in-

creases (table 17), the reason being that all the water in the material
does not drain out of it immediately. Investigations have shown
that a sample of material after being saturated will yield a very large
percentage of its water within a few hours, though it may continue
to yield small amounts for several years. When pumping first starts
a comparatively large volume of material is unwatered, partly because
only a small percentage of the water contained in the interstices of
the sediments immediately drains down to the water table. As pumping
is continued more water gradually drains out of the unwatered
sediments, and hence the specific yield computed from the first few
hours of pumping is relatively small. The specific yield computed
from the volume of water-bearing material unwatered in the last few
hours of pumping would be larger, because of the addition of water
that percolated down from the material previously unwatered. The
average values for specific yield (table 17) plotted against the periods
of pumping fall on a smooth curve. By extending this curve the con-
clusion is reached that the true specific yield lies between 22 and 23.
A sample of the water-bearing material that was unwatered during
this test was examined in the hydrologic laboratory of the United
States Geological Survey. The porosity was found to be 27.1 and
the moisture equivalent 2.6 (table 1). If the moisture equivalent is
used roughly for specific retention, as is sometimes done, the specific
yield of the sample is computed to be 24.5; if Piper's relation between
moisture equivalent and specific retention 21 is used, giving a specific

retention of about 5 for materials with a moisture equivalent of 2.6,
the specific yield is computed to be 22.1. Both of these values
compare favorably with the value determined from this pumping test.
11 Piper, A. M., Notes on the relation between the moisture equivalent and the specific yield of water-
bearing materials: Am. Geophys. Union Trails., 1933, pp. 481-487.

Thiem Method

Uploaded by

Copyright:

Available Formats

Thiem Method

Uploaded by

Document Information

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

Thiem Method

Uploaded by

Copyright:

Available Formats

Idf not destroy or throw away this publication.

If you have no further use

UNITED STATES DEPARTMENT OF THE INTERIOR

Water-Supply Paper 679 A

OS THIEM METHOD FOR DETERMINING

Prepared in cooperation with the Conservation and Survey

Contributions to the hydrology of the United States, 1935

Iby the Superintendent of Documents, Washington, D. C. Price 10 cents

The Thiem method for determining permeability of water-bearing materials *

where P is the coefficient of permeability; q is the rate of pumping, in gallons a

quent measurements of the draw-down during the period of pumping. If the

An investigation of the ground-water resources of Nebraska has for

ability and specific yield. Accordingly, two tests of these methods

dean of the Graduate School, University of Wisconsin, and G. Thiem,

in which Q is the quantity of water discharged in a unit of time, P the

areas of discharge, and consequently it is used for determining the

on the flow of water through porous media and has described an

where Q= quantity of water;

The coefficient of permeability of the water-bearing material is

samples by the application of centrifugal force; and (7) making

this report to represent the hydraulic gradient at a point on the cone of

The determination of specific yield by the'pumping method is based

draw-down of the water table as shown by successive measurements

Thiem's formula for computing the coefficient of permeability may

in which P=the coefficient of permeability as defined on page 5;

to the bottom of the water-bearing material, and two observation

FIGTJKE 1. Plan and section of ideal ground-water conditions assumed by Thiem.

The following symbols, in addition to those previously given, are

x and y= oblique coordinates of a point, J, on the cone of depression,

By substituting in Darcy's equation,

From the plan and section (fig. 1)

,=»_- ._.___- _____._.

By equating (9) and (10),

Equating (11) and (12), we get

and, by taking the square root,

V72 cos2 0+1

Equations 28 and 30 may be substituted in equation 8, thus:

<^----- ----- -(34)

In order to integrate equation 34, Thiem changed from an oblique

By integrating with respect to $ and Q,

and Q =27rP ------------ (37)

If the equation is now further integrated with respect to x and y,

and +a.... .... (40)

This is Thiem's general equation. If equation 37 is integrated

and p= ^(log -log. a)

This is Thiem's final equation and applies to regions where the-

Thiem's formula may be developed more simply by starting with

i, at any point on the cone of depression, is equal to the slope, or

With horizontal conditions, x and y are independent of $ ; hence $

Equation 57 is identical with equation 40 obtained by Thiem. The

The demonstration is as follows, again beginning with

The hydraulic gradient at any point on the cone of depression is

Equation 62 is developed to equation 4 in the same way that equation

The theoretical work by Slichter and by Turneaure and Russell,

The following nomenclature and units are changed somewhat from

From which c loge x=KZ+C1. ...... _ ._____- (66)

and by substituting equation 70 in 64,

The velocity at the wall of the well is found by placing x r. An

At this point Slichter solved equation 72 for Q, having determined

p_ $ lo&T = Qqogea1 -logea)

,=»_- ._._- ___._.