GSFLOW

GSFLOWCoupled Ground-Water and Surface-Water Flow
Model Based on the Integration of the Precipitation-Runoff

Modeling System (PRMS) and the Modular Ground-Water
Flow Model (MODFLOW-2005)
Chapter 1 of
Section D, Ground-Water/Surface-Water
Book 6, Modeling Techniques
Precipitation
Infiltration Sur
fac
e ru
Soil water nof
f
Unsaturated Gravity drainage Precipitation

zone Evapotranspiration
Recharge Inte
rflo
Water t w
ab le
Surfa Evaporation
c
runof e
Ground-w f
ater flow
Stream
Ground-water flow
Techniques and Methods 6D1
U.S. Department of the Interior

U.S. Geological Survey
GSFLOWCoupled Ground-Water and
Surface-Water Flow Model Based on the
Integration of the Precipitation-Runoff
Modeling System (PRMS) and the Modular
Ground-Water Flow Model (MODFLOW-2005)
By Steven L. Markstrom, Richard G. Niswonger, R. Steven Regan, David E. Prudic,

and Paul M. Barlow
Chapter 1 of
Section D, Ground-Water/Surface-Water
Book 6, Modeling Techniques
Techniques and Methods 6D1

DIRK KEMPTHORNE, Secretary

Mark D. Myers, Director
U.S. Geological Survey, Reston, Virginia: 2008
For product and ordering information:

World Wide Web: http://www.usgs.gov/pubprod
Telephone: 1-888-ASK-USGS
For more information on the USGS--the Federal source for science about the Earth, its natural and living resources,
natural hazards, and the environment:
World Wide Web: http://www.usgs.gov
Telephone: 1-888-ASK-USGS
Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the
U.S. Government.
Although this report is in the public domain, permission must be secured from the individual copyright owners to
reproduce any copyrighted materials contained within this report.
Suggested citation:
Markstrom, S.L., Niswonger, R.G., Regan, R.S., Prudic, D.E., and Barlow, P.M., 2008, GSFLOWCoupled ground-
water and surface-water flow model based on the integration of the Precipitation-Runoff Modeling System (PRMS)
and the Modular Ground-Water Flow Model (MODFLOW-2005): U.S. Geological Survey Techniques and Methods
6-D1, 240 p.
iii
Contents
Preface............................................................................................................................................................ix
Abstract............................................................................................................................................................1
Introduction ....................................................................................................................................................1
Integrated Hydrologic Models.............................................................................................................1
Purpose and Scope...............................................................................................................................3
Design of GSFLOW.........................................................................................................................................3
Description of PRMS.............................................................................................................................4
Description of MODFLOW-2005...........................................................................................................9
Integration of PRMS and MODFLOW-2005......................................................................................15
Cascading-Flow Procedure........................................................................................................................33
Computations of Flow...................................................................................................................................34
Distribution of Temperature, Precipitation, and Solar Radiation.................................................36
Potential Evapotranspiration.............................................................................................................43
Canopy Interception............................................................................................................................44
Snowpack.............................................................................................................................................45
Impervious Storage, Hortonian Runoff, and Infiltration................................................................47
Soil Zone................................................................................................................................................49
Unsaturated-Zone Flow......................................................................................................................62
Streams.................................................................................................................................................64
Lakes......................................................................................................................................................72
Ground Water.......................................................................................................................................78
Water Budgets..............................................................................................................................................84
Overall Water Budget ........................................................................................................................84
Computation of Budget Errors...........................................................................................................87
Assumptions and Limitations......................................................................................................................91
Discretization of Time and Space.....................................................................................................91
Canopy Zone.........................................................................................................................................92
Land-Surface Precipitation and Temperature................................................................................92
Soil Zone................................................................................................................................................92
Streams.................................................................................................................................................92
Lakes......................................................................................................................................................92
Ground Water.......................................................................................................................................93
Unsaturated Zone................................................................................................................................93
Input and Output Files .................................................................................................................................93
Input Files..............................................................................................................................................93
Output Files...........................................................................................................................................96
iv
ContentsContinued
Example Problem Sagehen Creek Watershed....................................................................................100
Description of Sagehen Creek Watershed . .................................................................................100
Calibration...........................................................................................................................................104
Model Results.....................................................................................................................................113
Sensitivity to Unsaturated-Zone and Aquifer Hydraulic Conductivity......................................120
Discussion of Results........................................................................................................................122
Acknowledgments .....................................................................................................................................124
References Cited........................................................................................................................................124
Appendix 1. Input Instructions for GSFLOW...........................................................................................131
Appendix 2. Definitions of Symbols. .......................................................................................................227
Figures
Figure1. Diagram showing distribution, flow, and interaction of water on the land and
in the subsurface 2
Figure2. Schematic diagram of a watershed and its climate inputs (precipitation, air
temperature, and solar radiation) simulated by PRMS 5
Figure3. Diagram showing components of the snowpack energy balance,
accumulation, snowmelt, and sublimation 6
Figure4. Diagrams showing delineation of hydrologic response units of a
hypothetical watershed determined from (A) topology and (B) topology,
climate, and vegetation 7
Figure5. Diagram showing computational sequence of PRMS used for simulating
streamflow response in watersheds 9
Figure6. Diagram showing flow through a hypothetical aquifer system that can be
simulated using MODFLOW-2005 10
Figure7. Diagram showing a discretized hypothetical aquifer system 11
Figure8. Diagram showing simple stream network having four segments and eight reaches
in a finite-difference model grid consisting of three rows and three columns 11
Figure9. Diagram showing simple stream and lake network having one lake, two
stream segments, and four stream reaches in a finite-difference model grid
consisting of seven rows and columns and three layers: (A) plan-view of
active finite difference grid, (B) vertical section along model row 4 13
Figure10. Diagram showing computational sequence of MODFLOW-2005 used for
simulating ground-water flow through aquifer systems 14
Figure11. Schematic diagram of the exchange of flow among the three
regions in GSFLOW 15
Figure12. Diagrams showing differences in conceptualization of soil-zone water
simulated by PRMS, where boxes are storage reservoirs representing
pore-space volumes for a given volume of soil: (A) original conceptualization
of a finite-volume soil-zone reservoir and an infinite-volume subsurface
reservoir; (B) revised conceptualization of three finite-volume reservoirs:
preferential flow, gravity, and capillary 21
Figure13. Diagrams showing conceptual changes to the soil-zone, shallow-subsurface, and
ground-water reservoirs for a PRMS-only simulation: (A) original conceptualization
of PRMS (Leavesley and others, 1983), and (B) revised conceptualization that
includes cascading flow of surface runoff and interflow from upslope hydrologic
response units to downslope hydrologic response units 23
v
FiguresContinued
Figure14. Diagram showing delineation of land-based and lake hydrologic response units
of a hypothetical watershed 24
Figure15. Diagram showing changes to MODFLOW-2005 implemented in GSFLOW
include: unsaturated zone is connected to PRMS at the soil-zone base;
unsaturated flow is simulated beneath lakes; distributed flow is routed in
streams 25
Figure16. Diagram showing (A) example distribution of gravity reservoirs in the soil zone
of hydrologic response units used to connect hydrologic response units in
PRMSwith finite-difference cells in MODFLOW. (B) includes 4hydrologic
response units, 6 finite-difference cells, and 9 gravity reservoirs 26
Figure17. Flow chart showing overall computational sequence used for an integrated
PRMS and MODFLOW-2005 simulation in GSFLOW. 31
Figure18. Diagrams showing cascading flow of surface runoff and interflow among
hydrologic response units and streams delineated from (A) topology, climate,
and vegetation; and (B) a finite-difference grid 35
Figure19. Diagram showing computational sequence used in computing flow to and from
a soil zone in a hydrologic response unit 50
Figure20. Diagram showing effect of undulations at the soil-zone base on gravity
drainage in relation to the water table in a finite-difference cell 60
Figure21. Diagram showing a wetting front moving through a uniform column of
unsaturated material affected by a decrease in flux (volumetric flow per unit
area), and results from a kinematic-wave solution of the wetting front
represented by leading and trailing waves 63
Figure22. Diagrams showing (A) Inflows to and (B) outflows from a stream reach 66
Figure23. Section showing eight paired horizontal and vertical locations used to
compute stream depth, top width, and wetted perimeter for a stream segment 70
Figure24. Section showing discretization of unsaturated zone beneath an eight-point
cross section of a stream segment within a finite-difference cell 72
Figure25. Diagram showing inflows to and outflows from a lake 73
Figure26. Diagrams showing (A) Conductance terms used to compute leakage from and
ground-water discharge to a lake; (B) representation with finite-difference cells
in MODFLOW-2005 75
Figure27. Diagrams showing (A) Flow into finite-difference cell i, j, k from cell i, j-1, k,
and (B) indices for six adjacent cells surrounding cell i, j, k (hidden) 79
Figure28. Diagram showing storage reservoirs represented in GSFLOW include the plant
canopy, snowpack, impervious surfaces, streams, and lakes at or above land
surface, and the soil, unsaturated, and saturated zones in the subsurface 85
Figure29. Selected output from a GSFLOW Water-Budget File that shows a summary listing
for December 15, 1981, from the Sagehen Creek example problem 86
Figure30. MODFLOW Name File used for the Sagehen Creek watershed
example problem 95
Figure31. Three lines of a GSFLOW Comma-Separated-Values (CSV) File that shows
variables that can be used for water-budget calculations for each time step 97
Figure32. PRMS Water-Budget File that shows annual and simulation-run summary
tables (parameters print_type set to 1 and print_freq set to 3 in the PRMS
Parameter File) for the Sagehen Creek watershed example problem 98
vi
FiguresContinued
Figure33. Selected output from the PRMS Statistic Variables (statvar) File that lists
simulated (basin_cfs) and measured (runoff) streamflow for the Sagehen
Creek watershed example problem 98
Figure34. Selected output from a PRMS Animation File that lists snowpack-water
equivalent, soil moisture, evapotranspiration, and precipitation for each HRU 99
Figure35. Map showing Sagehen Creek watershed located on the east slope of the
northern Sierra Nevada near Truckee, California101
Figure36. Map showing hydrologic response units discretized for Sagehen Creek
watershed near Truckee, California 103
Figure37. Map showing hydraulic conductivity values used in ground-water model,
Sagehen Creek watershed near Truckee, California105
Figure38. Graphs showing PRMS simulated and daily-mean streamflow at Sagehen
Creek gage for (A) calibration period (water years 198288) and (B) verification
period (water years 198995)107
Figure39. Map showing steady-state infiltration rate for the ground-water model of
Sagehen Creek watershed near Truckee, California109
Figure40. Map showing simulated steady-state water-table depth below land surface in
the Sagehen Creek watershed near Truckee, California 110
Figure41. Map showing finite-difference cells in relation to hydrologic response units
in Sagehen Creek watershed near Truckee, California111
Figure42. Graphs showing GSFLOW simulated and daily-mean streamflow at Sagehen
Creek gage for (A) calibration period (water years 198288) and (B) verification
period (water years 198995)112
Figure43. Graph showing comparison between simulated and estimated base flow for
water years 198195 113
Figure44. Graph showing number of iterations per simulation time step for GSFLOW to
converge during water years 198194 114
Figure45. Graphs showing cumulative change of (A) snowpack and soil-zone storage;
and (B) unsaturated-zone and saturated-zone (ground-water) storage during
Figure46. Graph showing major fluxes across land surface during water years 198189
that include precipitation, snowmelt, and evapotranspiration 116
Figure47. Graph showing fluxes of gravity drainage from soil zone down to unsaturated
zone and ground-water discharge from saturated zone into soil zone during
Figure48. Graph showing components of streamflow during water years 198189 117
Figure49. Maps showing distribution of ground-water discharge to the soil zone for
(A) steady-state simulation of the autumn low precipitation period, and
(B) July 8, 1983 118
Figure50. Graphs showing water-content profiles below land surface (finite-difference
cell layer 1, row 45, and column 30)(A) during April 1983, and (B) during
June 1983 119
Figure51. Graphs showing sensitivity analysis for daily volumetric flow of (A) net gravity
drainage from the soil zone, (B) recharge to the saturated zone, and (C)
ground-water discharge to land surface summed over the Sagehen Creek
watershed for water years 198283 121
Figure52. Graphs showing streamflow components for (A) 1983, a year of above-average
precipitation, and (B) 1992, a year of below-average precipitation 123
vii
Tables
Table1. Description of PRMS and GSFLOW modules and MODFLOW-2005 packages
implemented in GSFLOW, listed in computational order 17
Table2. PRMS modules and MODFLOW-2005 packages not implemented in GSFLOW 20
Table3. Topological parameters used to relate areas of gravity reservoirs in the soil
zone to hydrologic response units in PRMS and to areas of finite-difference
cells in MODFLOW-2005 27
Table4. PRMS parameters for module gsflow_prms2mf that relate hydrologic response
units in PRMS to stream segments and reaches in the Streamflow-Routing
Package of MODFLOW-2005 27
Table5. GSFLOW main program, Modular Modeling System utility functions used with
PRMS, and additional MODFLOW-2005 packages and files used in GSFLOW 29
Table6. Description of computational sequence for GSFLOW 30
Table7. Conversion of volumes, volumetric flow rates, length, and velocity for variables in
modules gsflow_prms2mf and gsflow_mf2prms in GSFLOW 33
Table8. Parameters for PRMS Cascade Module (cascade_prms) used to define
connections for routing flow from upslope to downslope hydrologic response
units and stream segments and among ground-water reservoirs 34
Table9. Sequence of steps used in the computation of flow into and out of the soil zone
used in GSFLOW 51
Table10. Variables used to transfer water from reservoirs in PRMS to finite-difference
cells and stream reaches in MODFLOW-2005 used in GSFLOW 54
Table11. Inflow, outflow, changes in storage, and errors reported in the GSFLOW
Budget Report File for GSFLOW 87
Table12. Names and descriptions of variables written to the GSFLOW
Comma-Separated-Values File at end of each daily time step for GSFLOW 88
Table13. Values and source of non-default PRMS parameters used for the
Sagehen Creek watershed example problem 102
Table14. MODFLOW-2005 packages and files used for the Sagehen Creek watershed
example problem 104
Table15. Hydraulic properties and other selected variables used in the Layer-Property
Flow, Unsaturated-Zone Flow, and Streamflow-Routing Packages for the
Table16. Annual water budget for PRMS calibration and verification periods for the
Table17. Values of PRMS parameters modified for the GSFLOW Sagehen
Creek watershed example problem 109
Table18. Annual water budget for GSFLOW simulation of the Sagehen Creek watershed
example problem, 198295120
viii
Conversion Factors and Datums

Conversion Factors
Multiply By To obtain
cubic meter (m )
3
35.31 cubic foot (ft3)
cubic meter (m3) 264.2 gallon (gal)
cubic meter per day (m3/d) 35.31 cubic foot per day (ft3/d)
cubic meter per day (m3/d) 264.2 gallon per day (gal/d)
cubic meter per second (m3/s) 70.07 acre-foot per day (acre-ft/d)
cubic meter per second (m3/s) 35.31 cubic foot per second (ft3/s)
gram (g) 0.03527 ounce, avoirdupois (oz)
gram per cubic centimeter (g/cm3) 62.4220 pound per cubic foot (lb/ft3)
kilogram (kg) 2.205 pound, avoirdupois (lb)
kilometer (km) 0.6214 mile (mi)
kilopascal (kPa) 0.2961 inch of mercury at 60F (in Hg)
kilopascal (kPa) 0.1450 pound per square inch (lb/ft2)
liter (L) 0.2642 gallon (gal)
meter (m) 3.281 foot (ft)
meter per day (m/d) 3.281 foot per day (ft/d)
meter per kilometer (m/km) 5.27983 foot per mile (ft/mi)
meter per second (m/s) 3.281 foot per second (ft/s)
meter squared per day (m2/d) 10.76 foot squared per day (ft2/d)
millimeter (mm) 0.03937 inch (in.)
square meter (m2) 10.76 square foot (ft2)
square kilometer (km2) 247.1 acre
square kilometer (km2) 0.3861 square mile (mi2)
Temperature in degrees Celsius (C) may be converted to degrees Fahrenheit (F) as follows:
F= (1.8C) +32
Temperature in degrees Fahrenheit (F) may be converted to degrees Celsius (C) as follows:
C= (F-32)/1.8
Datums
Vertical coordinate information is referenced to the North American Vertical Datum of 1988
(NAVD 88)
Horizontal coordinate information is referenced to North American Datum of 1983 (NAD 83)
Altitude, as used in this report, refers to distance above the vertical datum.
ix
Preface
This report describes the U.S. Geological Survey coupled Ground-water and Surface-
water Flow model (GSFLOW). The performance of the program has been tested in a variety of
applications. Future applications, however, might reveal errors that were not detected in the test
simulations. Users are requested to send notification of any errors found in this report or the
model program to:
Office of Ground Water

411 National Center
Reston, VA 20192
(703) 648-5001
The latest version of the model program and this report can be obtained using the Internet
at address: http://water.usgs.gov/software/.
x
This page intentionally left blank.

GSFLOWCoupled Ground-Water and Surface-Water
Flow Model Based on the Integration of the
Precipitation-Runoff Modeling System (PRMS) and the
Modular Ground-Water Flow Model (MODFLOW-2005)
By Steven L. Markstrom, Richard G. Niswonger, R. Steven Regan, David E. Prudic, and Paul M. Barlow
Abstract estimate recharge using long-term average precipitation and

evaporation, without considering flow through the soil and
The need to assess the effects of variability in climate, unsaturated zones (Cosner and Harsh, 1978; Morgan, 1988;
biota, geology, and human activities on water availability Danskin, 1998; Lindgren and Landon, 1999). Similarly,
and flow requires the development of models that couple two models used to evaluate the effects of climate variability on
or more components of the hydrologic cycle. An integrated surface-water resources typically represent ground-water flow
hydrologic model called GSFLOW (Ground-water and to streams using empirical and(or) simplified equations that
Surface-water FLOW) was developed to simulate coupled do not consider the effects of local geology and surface-water
ground-water and surface-water resources. The new model interactions (Beven and others, 1984; Jeton, 1999; Zarriello
is based on the integration of the U.S. Geological Survey and Ries, 2000; Ely, 2006).
Precipitation-Runoff Modeling System (PRMS) and the
U.S. Geological Survey Modular Ground-Water Flow Model Integrated Hydrologic Models
(MODFLOW). Additional model components were developed,
and existing components were modified, to facilitate Integrated hydrologic models are useful for analyzing
integration of the models. Methods were developed to route complex water-resources problems faced by society because
flow among the PRMS Hydrologic Response Units (HRUs) they can consider feedback processes that affect the timing
and between the HRUs and the MODFLOW finite-difference and rates of evapotranspiration, surface runoff, soil-zone flow,
cells. This report describes the organization, concepts, and ground-water interactions (fig. 1). Two general approaches
design, and mathematical formulation of all GSFLOW model have been used to design integrated models. One approach is
components. An important aspect of the integrated model referred to as fully integrated whereby equations governing
design is its ability to conserve water mass and to provide surface and subsurface flows are solved simultaneously
comprehensive water budgets for a location of interest. (VanderKwaak, 1999; Panday and Huyakorn, 2004). The
This report includes descriptions of how water budgets are second approach partitions the surface and subsurface systems
calculated for the integrated model and for individual model into separate regions and the governing equations that describe
components. GSFLOW provides a robust modeling system for flow in each region are integrated (coupled) using iterative
simulating flow through the hydrologic cycle, while allowing solution methods (Smith and Hebbert, 1983; Refsgaard and
for future enhancements to incorporate other simulation Storm, 1995; Sophocleous and Perkins, 2000). This second
techniques. approach is called the coupled regions approach.
The fully integrated approach uses the three-dimensional
form of Richards equation to simulate unsaturated and
Introduction saturated flow (Freeze, 1971; VanderKwaak, 1999; Panday
and Huyakorn, 2004; Thoms and others, 2006). Models
that use the three-dimensional form of Richards equation
Hydrologic models of ground-water and surface-water
require much finer spatial grids and smaller time steps than
systems traditionally have been developed with a focus
typically are used to simulate saturated flow, which limits their
on either the ground-water or surface-water resource. For
applicability for simulating flow through regional hydrologic
example, models used to evaluate the effects of pumping and
systems that encompass hundreds to thousands of square
surface-water interactions on ground-water resources typically
kilometers and simulation time periods of months to decades.
2 GSFLOWFlow Model Based on the Integration of the PRMS and the MODFLOW-2005
Precipitation
Infiltration Sur
fac
e ru
Soil water nof
f
Unsaturated Gravity drainage Precipitation

zone Evapotranspiration
Recharge Inte
rflo
w
Water
ta ble Surfa Evaporation
c
runof e
Ground-w f
ater flow
Stream
Ground-water flow
EXPLANATION
Zone of aeration
Zone of saturation
Soil-zone base
Figure 1. Distribution, flow, and interaction of water on the land and in the subsurface.
Thus, coupled models that do not use the three-dimensional GSFLOW can be used to evaluate the effects of such factors
form of Richards equation may be better suited for simulating as land-use change, climate variability, and ground-water
flow through regional hydrologic systems. withdrawals on surface and subsurface flow. The model
An efficient approach for simulating unsaturated flow was designed to simulate the most important processes
in integrated models takes advantage of the fact that the affecting surface-water and ground-water flow using a
dominant direction of flow within the unsaturated zone numerically efficient algorithm. The model incorporates well
usually is vertical when averaged over large areas (Mantoglou, documented methods for simulating runoff and infiltration
1992; Chen and others, 1994; Harter and Hopmans, 2004). from precipitation, as well as the interaction of surface water
Accordingly, equations can be used to simulate flow and with ground water in watersheds that range from a few square
storage in the soil, unsaturated, and saturated zones separately, kilometers to several thousand square kilometers, and for time
such that model efficiency is improved with some sacrifice of periods that range from months to several decades.
model accuracy. For this approach, separate equations are used GSFLOW version 1.0 (referred to as GSFLOW in
to simulate horizontal and vertical flow through the soil zone, the remainder of the report) is an integration of the U.S.
gravity-driven vertical flow through the unsaturated zone, and Geological Survey (USGS) Precipitation-Runoff Modeling
three-dimensional ground-water flow through the saturated System (PRMS; Leavesley and others, 1983; Leavesley and
zone. This coupled-regions approach was the approach used others, 2005) with the 2005 version of the USGS Modular
in the development of the GSFLOW (Ground-water/Surface- Ground-Water Flow Model (MODFLOW-2005; Harbaugh,
water FLOW) model described in this report and outlined by 2005). PRMS and MODFLOW have similar modular
Fenske and Prudic (1998). programming methods, which allow for their integration while
GSFLOW was developed to simulate coupled ground- retaining independence that permits substitution of additional
water/surface-water flow in one or more watersheds by PRMS modules and MODFLOW packages. Both models have
simultaneously simulating flow across the land surface a long history of support and development.
and within subsurface saturated and unsaturated materials.
nv07tac-4124_fig01
Design of GSFLOW 3
PRMS was selected because it: (1) can simulate land- were integrated for GSFLOW. The equations and order
surface hydrologic processes of evapotranspiration, runoff, of calculations used in GSFLOW are presented in section
infiltration, and interflow by balancing energy and mass Computations of Flow. Inflows, outflows, and storages are
budgets of the plant canopy, snowpack, and soil zone on discussed in section Water Budgets. Important assumptions
the basis of distributed climate information (temperature, used in the development of GSFLOW and how these
precipitation, and solar radiation); (2) can be used to analyze assumptions limit the applicability of GSFLOW are described
the effects of urbanization on the spatial distribution of in section Assumptions and Limitations. The section Input
ground-water recharge (Vaccaro, 1992; Steuer and Hunt, and Output Files describes data-input requirements and
2001); (3) can be used with other models for water-resources output options of GSFLOW files. The last section presents an
management and forecasting (Fulp and others, 1995; Wilby example simulation of the Sagehen Creek watershed in the
and others, 1999; Berris and others, 2001; Hay and others, eastern Sierra Nevada near Truckee, California. Appendix1
2002; Mastin and Vaccaro, 2002; Hay and Clark, 2003; Clark provides detailed input instructions for the files that are
and Hay, 2004); and (4) has a modular design that allows for required for a GSFLOW simulation. Users of GSFLOW are
selection of alternative hydrologic process algorithms among encouraged to review documentation for PRMS (Leavesley
existing or easily added modules. Some examples of the use and others, 1983) and MODFLOW-2005 (Harbaugh, 2005).
of PRMS in hydrologic studies are the simulation of sediment
production for semi-arid watersheds (Rankl, 1987); heat and
water transfer for seasonally frozen soils (Emerson, 1991); use
of radar data to specify rainfall input (Yates and others, 2000), Design of GSFLOW
streamflow and wetland storage (Vining, 2002); and flow-
frequency characteristics (Olson, 2002). GSFLOW simulates flow within and among three
MODFLOW was selected because it is one of the most regions. The first region is bounded on top by the plant canopy
widely used and tested ground-water flow codes, and is and on the bottom by the lower limit of the soil zone; the
capable of simulating: (1) three-dimensional saturated ground- second region consists of all streams and lakes; and the third
water flow and storage; (2) one-dimensional unsaturated flow region is the subsurface zone beneath the soil zone. PRMS is
and ground-water discharge to the land surface (Niswonger used to simulate hydrologic responses in the first region and
and others, 2006a); and (3) ground-water interactions with MODFLOW-2005 is used to simulate hydrologic processes in
streams (Prudic, 1989; Swain and Wexler, 1996; Jobson and the second and third regions. The following design principles
Harbaugh, 1999; Prudic and others, 2004; Niswonger and guided development of the model:
Prudic, 2005) and lakes (Cheng and Anderson, 1993; Council, Use existing PRMS modules and MODFLOW-2005
1998; Merritt and Konikow, 2000). Moreover, MODFLOW packages where possible;
has been coupled with other precipitation-runoff models (Ross
and others, 1997; Hunt and Steuer, 2000; Sophocleous and Use a flexible and adaptive modular design that
Perkins, 2000; Nishikawa and others, 2005; Said and others, incorporates both PRMS and MODFLOW-2005
2005) and with solute-transport models (Konikow and others, programming frameworks so that existing and new
1996; Guo and Langevin, 2002; Langevin and others, 2003), PRMS and MODFLOW-2005 simulation techniques
and has sensitivity and parameter estimation capabilities (Hill can be added to GSFLOW in the future;
and others, 2000). Use general design procedures that can be used to
The initial version of GSFLOW does not include integrate other simulation models into GSFLOW;
all capabilities of the PRMS and MODFLOW models.
Additional capabilities likely will be added to GSFLOW over Allow simulations using only PRMS or
time. Moreover, future versions of GSFLOW can include MODFLOW-2005 within the integrated model for the
alternative models to PRMS and MODFLOW for simulating purpose of initial calibration of model parameters prior
surface- and(or) ground-water flow. Additional models can be to a comprehensive calibration using the integrated
integrated to simulate other environmental and anthropogenic model;
processes, such as water quality, ecology, geochemistry, and
Solve equations governing interdependent surface-
management strategies, and pre- and post-processors will
water and ground-water flow using iterative solution
likely be developed for GSFLOW to facilitate the use of
techniques;
GSFLOW by practicing hydrologists.
Compute model-wide and detailed (for example, soil-
zone flow and storage) water balances in both time and
Purpose and Scope
space;
This report describes GSFLOW, version 1.0. The Allow flexibility in the spatial discretization of the
report first describes the design of GSFLOW and includes hydrologic response units used for PRMS and the
descriptions of PRMS and MODFLOW-2005 and how they finite-difference grid used for MODFLOW-2005; and
Allow model boundaries to be defined using standard Climate data consisting of measured or estimated
specified-head, specified-flow, and head-dependent precipitation, air temperature, and solar radiation are the
boundary conditions to account for inflows to and driving factors used to compute evaporation, transpiration,
outflows from the modeled region. sublimation, snowmelt, surface runoff, and infiltration in a
The remainder of this section provides an overview of the PRMS simulation (Leavesley and others, 1983, p. 12-18).
PRMS and MODFLOW models, as well as a description of The form of precipitation (rain, snow, or mixture of both) is
how the two models were integrated into a single model. Some determined from temperature data or can be specified as input
of the components of PRMS and MODFLOW were modified data. Precipitation can be intercepted by and evaporated from
during the integration process; these are described in the the plant canopy. Precipitation that is not intercepted by the
following sections. However, because GSFLOW can be used plant canopy, which is referred to as throughfall, is distributed
to run PRMS-only and MODFLOW-only simulations (that is, to the watershed land surface. Precipitation that reaches land
non-integrated simulations), descriptions of those components surface can accumulate as part of the snowpack, be stored
of PRMS and MODFLOW that are required for PRMS- and in impervious-zone reservoirs, infiltrate into the soil zone,
MODFLOW-only simulations also are included in this report. be evaporated, or become surface runoff. Water and energy
balances are computed for the snowpack to determine snow
accumulation, snowmelt, or sublimation (fig. 3).
Description of PRMS The subsurface is represented by a series of three
reservoirsthe soil-zone, subsurface, and ground-water
PRMS is a modular deterministic, distributed-parameter, reservoirs (fig. 2). The soil-zone and subsurface reservoirs
physical-process watershed model used to simulate and are used in PRMS to account for different fractions of water
evaluate the effects of various combinations of precipitation, in pores within the matrix of the soil. Pores can range from
climate, and land use on watershed response. Response to relatively small intergranular space between grains of clay,
normal and extreme rainfall and snowmelt can be simulated silt, sand, and gravel to relatively large pores caused by cracks
to evaluate changes in water-balance relations, streamflow from seasonal shrinking and swelling of the soil, by holes
regimes, soil-water relations, and ground-water recharge. from decaying plant matter such as roots and leaf litter, by
Each hydrologic component used to model the generation holes from animal activity such as worms and gophers, and
of streamflow is represented within PRMS by a process by cracks from landscape altering events such as earthquakes.
algorithm that is based on a physical law or an empirical The relatively larger pores commonly are called macropores
relation with measured or estimated characteristics. A (Selker and others, 1999). The two reservoirs include water
description of how a hydrologic system is conceptualized in the pores that (1) can be removed by plants, (2) can drain
and discretized in PRMS is included to provide a basic downward to ground water or flow laterally downslope
understanding of how a model can be developed to represent through relatively small intergranular space between grains
a particular hydrologic system. However, the focus is on the caused by perching of water on top of a less permeable soil
capabilities of PRMS that are integrated into GSFLOW; the horizon, and (3) can flow laterally through macropores.
reader is referred to Leavesley and others (1983), Leavesley Water-saturation levels that are above the plant wilting
and Stannard (1995), and Leavesley and others (2005) for a threshold and below field capacity are simulated within
complete description. the soil-zone reservoir. Antecedent water stored in the soil-
zone reservoir controls the partitioning of available water
(throughfall and snowmelt) at land surface to surface runoff
Representation of Watershed Hydrologic and infiltration. Water stored in the soil-zone reservoir is
Processes referred to as capillary water and can be lost to evaporation
and transpiration on the basis of plant type and cover, rooting
A watershed is defined as the area of land that drains
depth, precipitation, solar radiation, and air temperature. The
into a stream above a given location (Chow and others,
soil-zone reservoir is partitioned into two zonesthe recharge
1988, p.7). PRMS simulates the hydrologic processes of a
zone and the lower zone. The recharge zone contains water
watershed using a series of reservoirs that represent a volume
up to a specified maximum water-saturation threshold that is
of finite or infinite capacity. Water is collected and stored in
available for evaporation and transpiration. The lower zone
each reservoir for simulation of flow, evapotranspiration, and
contains water when the water-saturation level in the soil-
sublimation. Flow to the drainage network, which consists
zone reservoir exceeds the maximum threshold. Lower-zone
of stream-channel and detention-reservoir (or simple-lake)
water is available only for transpiration. Water content in the
segments, is simulated by surface runoff, interflow, and
soil-zone reservoir below the wilting threshold is assumed to
ground-water discharge (fig. 2).
be constant and is not included in water-balance computations.
This water adheres strongly to the granular maxtrix and is
unavailable to plants and relatively nonmoving (Bear, 1972,
p.3).
Design of GSFLOW 5
Solar
radiation
Precipitation
Evaporation Sublimation Air temperature
Plant canopy
interception
Rain Throughfall
Rain
Evaporation
and Snowpack Evaporation
Transpiration
Transpiration Surface runoff
Snowmelt to stream or lake
Soil-Zone Reservoir Impervious-Zone Reservoir

Recharge zone
Lower zone
Subsurface recharge
Ground-water
recharge Subsurface
Reservoir Interflow (or subsurface
flow) to stream or lake
Ground-water recharge
Ground-Water
Reservoir Ground-water discharge to stream or lake
Ground-water
sink
Figure 2. Schematic diagram of a watershed and its climate inputs (precipitation, air temperature, and solar radiation)
simulated by PRMS (modified from Leavesley and others, 1983).
Water that infiltrates into the soil-zone reservoir above A new Soil-Zone Module to simulate flow in the soil-
field capacity is distributed to the subsurface and ground- zone and subsurface reservoirs was developed to facilitate
water reservoirs (fig. 2). Water stored in the subsurface integration between PRMS and MODFLOW and allow for
reservoir(s) is available for gravity drainage to a ground-water flow through macropores and Dunnian runoff (Dunne and
reservoir and for interflow to a stream or lake. Water stored Black, 1970). The new module affects PRMS-only simulations
in the ground-water reservoirs is available for ground-water as well as integrated simulations, and is described in the
discharge to a stream or lost to the ground-water sink. SoilZone section of Computations of Flow. Flow beneath
the soil zone is simulated using MODFLOW-2005 in a
GSFLOW integrated simulation.
nv07tac-4124_fig02
Incoming
Heat of Latent and
shortwave
precipitation Longwave sensible heat
radiation
radiation
from air
Longwave
radiation
from
canopy Longwave
radiation
from
Reflected Snow snow
Sublimation shortwave accumulation
radiation
Surface layer
snowpack
Upward or downward
conduction between layers Lower layer
Snowmelt snowpack
Conduction
assumed = 0
Figure 3. Components of the snowpack energy balance, accumulation, snowmelt, and

sublimation (modified from Leavesley and others, 1983).
Watershed DiscretizationHydrologic Response results in regular rectangles; or (3) a noncontiguous approach,

which results in many unique and irregularly shaped polygons
Units determined by combinations of watershed characteristics. Two
The area of a watershed is discretized into a network examples of HRU delineation of a hypothetical watershed are
of hydrologic response units (HRUs). The discretization can shown in figure 4.
be based on hydrologic and physical characteristics such as Precipitation that falls on each HRU is routed through
drainage boundaries, land-surface altitude, slope, and aspect; the soil-zone, subsurface, and ground-water reservoirs. Water
plant type and cover; land use; distribution of precipitation, is added to and subtracted from each HRU according to input
temperature, and solar radiation; soil morphology and data and model parameters. Each HRU includes separate
geology; and flow direction. Each HRU is assumed to be reservoirs that represent water stored in the plant canopy,
homogeneous with respect to these hydrologic and physical snowpack, impervious surfaces, and soil zone. A group of
characteristics and to its hydrologic response. A water balance contiguous soil-zone reservoirs can be designated to add water
and an energy balance are computed daily for each HRU. to one subsurface reservoir. Similarly, contiguous subsurface
Each HRU is identified by a numerical index. Assignment of reservoirs can be grouped to add water to one ground-water
the index to the HRU is arbitrary, but indices must be unique, reservoir. Thus, there may be fewer subsurface and ground-
consecutive, and start with 1. water reservoirs than HRUs. Typically, a watershed is divided
The delineation of a watershed into HRUs can be into many HRUs, several subsurface reservoirs, and one or
automated with the aid of a Geographic Information System two ground-water reservoirs. Each reservoir also is assumed
(GIS) analysis. Three delineation approaches typically are to be homogeneous in hydrologic response and model
used (Viger and Leavesley, 2007): (1) a topological approach, parameterization. Water entering a reservoir is considered
which results in irregularly shaped polygons representing to be instantaneously mixed with water previously in the
hill slopes and flow planes; (2) a grid-based approach, which reservoir. The number of subsurface reservoirs must equal the
number of HRUs for GSFLOW.
nv07tac-4124_fig03
A B
11
15
3 4
14
10
2
9
1 1
5 13 18
17
2
7 6 5 16
2 8 12
21
4 3 20
2 19
1
6
1
3 3
EXPLANATION
6 Hydrologic response unit and number
Watershed boundary
1 Stream segment and number
Direction of streamflow
Streamflow gage at outlet
Figure 4. Delineation of hydrologic response units of a hypothetical watershed determined from (A) topology and (B) topology, climate, and vegetation.
nv07tac-4124_fig04
Design of GSFLOW 7
Time DiscretizationTime Steps utility software for optimization, sensitivity, forecasting,

visualization, and statistical analyses (Leavesley and others,
Simulation of watershed response to climate data is 2005, p. 160).
discretized into a series of time steps. Time steps can be MMS provides utility functions for models built within
specified equal to 1 dayreferred to as the daily modeor its framework, including: (1) file input and output; (2) time
they can be specified equal to a time increment less than 1 step looping and other model control; (3) function calls to pass
dayreferred to as the storm mode (Leavesley and others, data values between modules; (4) a graphical user interface;
1983, p. 10). The length of each time step is determined by (5) tools for model calibration and uncertainty analysis;
the smallest increment of time between consecutive measured and (6) tools for statistical and visualization analysis. For
input data values. Daily mode is used to simulate a watershed GSFLOW, only the MMS utility functions for items 1, 2, and
response to changes in daily precipitation, air temperature, and 3 are implemented.
solar radiation. Storm mode is used to simulate a watershed MMS controls the PRMS computation sequence,
response to individual storms on the basis of changes in including the time loop, using four procedures: declare,
precipitation, air temperature, and solar radiation that range initialize, run, and cleanup. These procedures are similar to
from seconds to hours. Storm-mode simulations are not MODFLOW-2005 procedures. These procedures are executed
available in GSFLOW. through a call to the main subroutine of each PRMS module,
which then calls the appropriate subroutine developed for the
Channel-Flow and Detention Reservoir Routing active MMS procedure. The declare procedure is executed first
and is used to: (1) specify the valid value range, default value,
The drainage network within the watershed is units, and definition of input parameters; (2) specify the units
characterized as a collection of stream-channel and detention and definition of output variables; and (3) allocate memory.
reservoir (simple-lake) segments for daily simulations The initialize procedure is executed second and is used to
(PRMS daily mode) and as a collection of stream-channel, initialize input parameters and output variables and open any
detention-reservoir, and junction segments for storm output files. The run procedure is executed each time step
simulations (PRMS storm mode). The PRMS allows and is used to compute a particular simulation process. The
for storm mode flow routing in the watershed drainage cleanup procedure is executed after all time steps have been
network to simulate storm events using a kinematic-wave simulated and is used to release allocated memory and close
approximation of single-direction, unsteady flow (Leavesley files.
and others, 1983, p.34-37). Outflow from as many as three PRMS modules simulate a particular hydrologic
upstream channel and detention-reservoir segments can component or computational method and are called within the
be added as inflow to a downstream segment. Two routing MMS in the proper sequence. As such, PRMS modules serve
methods are available to compute outflow from detention similar purposes as packages in MODFLOW-2005. All PRMS
reservoirs: (1) a linear storage method computed using modules include a declare and initialize procedure. Most
continuity; and (2) a modified-Puls method (U.S. Soil PRMS modules have a run procedure and a few have a cleanup
Conservation Service, 1971). Inflow to stream-channel and procedure.
detention-reservoir segments includes upstream channel
flow and surface runoff from all contiguous HRUs and flow
Computational Sequence
from subsurface and ground-water reservoirs associated with
a particular segment. Watershed outflow in daily mode Model execution begins in the PRMS main program,
includes flow from all contributing HRUs, detention-reservoir where the Control File and Parameter File are opened
segments, and subsurface and ground-water reservoirs. Flow and read. Hydrologic processes and program options are
in the drainage network is simulated using MODFLOW-2005 determined from input data followed by declaration and
in a GSFLOW integrated simulation. initialization of all required parameters and variables
(fig.5). Next, the time-step loop begins and precipitation, air
Organization temperature, and optionally solar radiation values are read
from the Data File for each measurement station and then
PRMS is implemented as a set of modules in the Modular distributed over the HRUs in the watershed. The computations
Modeling System (MMS) (Leavesley and others, 1996a and of flows and storages follows the progression of water as
1996b, http://wwwbrr.cr.usgs.gov/projects/SW_MoWS/). it moves through the plant canopy, across the land surface,
Modules are a group of subroutines written in the Fortran or into the soil zone, and through the subsurface and ground-
C programming languages that simulate a particular model water reservoirs until reaching a stream (fig. 2). Results are
process. MMS was developed to: (1) support developing, saved to output files and a new time step is begun when the
testing, and evaluation of physical-process models as computation of flow and storage has been completed for a time
compatible sets of computer code; (2) facilitate integration step (fig. 5). The computation of flow and storage continues
of user-selected codes into operational physical-process for each time step until the last time step is completed. After
models; (3) facilitate the coupling of models for application all time steps have been simulated, output files are closed and
to complex, multidisciplinary problems; and (4) provide memory is deallocated.
Design of GSFLOW 9
low-permeability regions below land

Run PRMS surface commonly are called confining units
(Lohman and others, 1972; Sun, 1986). The
Declare and Initialize: Read Parameters subsurface consists of geologic materials that
are either partially saturated (unsaturated
zone) or fully saturated (saturated zone).
Run: Time-Step Loop
The modeled region of the saturated zone
Read Input Climate Data for Stations
may have multiple aquifers separated by
(Precipitation, Air Temperature, confining units that collectively are called
and optionally Solar Radiation) an aquifer system. MODFLOW-2005 allows
for many types of processes, such as areal
Compute Climate on HRUs recharge across the water table, leakage to
aquifers from streams and lakes, subsurface
Compute Flow and Storage for HRUs inflows either at fixed rates or at rates that
vary as a function of head, discharge by
Compute Streamflow as Sum evapotranspiration from phreatophytes,
of Flow Components discharge from pumping wells, and discharge
to streams and lakes (fig. 6). Confining units
Output Results to File can be simulated explicitly as separate units
or implicitly in a quasi-three-dimensional
Check Time Step; Loop as Necessary
approach as averaged into the overall
properties of the aquifer system. Wells can be
simulated as removing water from or adding
Cleanup: Deallocate Memory
water to one or more aquifers. Although
MODFLOW-2005 originally was designed
to simulate only the saturated zone beneath
the water table, the Unsaturated-Zone Flow
Figure 5. Computational sequence of PRMS used for simulating streamflow Package (Niswonger and others, 2006a) for
response in watersheds. MODFLOW-2005 provides the capability
to simulate one-dimensional flow through
the unsaturated zone above the water table
Description of MODFLOW-2005 (fig.6).
MODFLOW-2005 (Harbaugh, 2005) is a new version

of the finite-difference, three-dimensional ground-water Aquifer-System DiscretizationFinite-Difference
flow model commonly called MODFLOW (McDonald and Cells
Harbaugh, 1988; Harbaugh and McDonald, 1996; Harbaugh
and others, 2000). MODFLOW can be used to simulate An aquifer system is discretized in MODFLOW-2005
steady-state or transient flow of constant density ground with a finite-difference grid that represents the computational
water through porous earth. A general description of how a cells used to calculate ground-water heads and flows. The
ground-water flow system is conceptualized and discretized finite-difference grid should be aligned with the major axes of
in all versions of MODFLOW is included to provide a basic hydraulic conductivity. Finite-difference cells are numbered
understanding of how a model is developed to represent a according to the indices i, j, and k, where i is the row index, jis
particular ground-water system. However, the focus herein is the column index, and k is the layer index (Harbaugh, 2005,
on the capabilities of MODFLOW-2005 that are integrated p. 2-2). Horizontal discretization is defined as a rectangular
into GSFLOW; the reader is referred to Harbaugh (2005) for a grid of rows and columns. Rows are aligned parallel to the
complete description. xaxis. Increments of the row index i correspond to decreases
in the y direction. Columns are aligned parallel to the y axis.
Increments of column index j correspond to increases in
Representation of Subsurface Processes the xdirection. Vertical discretization is defined by layers.
Layers of cells are aligned parallel to the horizontal plane.
Regions of porous and permeable rocks where water
Increments of layer index k correspond to increases in the z
flows and is stored below land surface are called aquifers
direction (fig.7). Thus, the top layer of the model corresponds
(Meinzer, 1923, p. 30). The subsurface in most regions of the
with k =1; the upper-left cell in each layer corresponds with
United States generally consists of many types of rocks or
row 1 (i = 1) and column 1 (j = 1). Layers can have uniform
geologic formations, some of which are aquifers. Saturated
thickness or can have variable thickness among cells in a layer.
124_fig05
Intermittent stream
Unsaturated flow
beneath stream Evapotranspiration
Recharge
Water Well discharge

table
Lake
Ground-w
ater flow
Stream
Saturated zone
root uptake Lake
Ground-water Stream leakage
discharge leakage
Figure 6. Flow through a hypothetical aquifer system that can be simulated using MODFLOW-2005.
Some error is introduced into the solution when the grid is Time-variant inflow and(or) outflow stress rates
not oriented with the principal axes of hydraulic conductivity, (boundary conditions) can be assigned to variable-head cells
when a cell represents a combination of aquifers and confining using MODFLOW stress packages such as the Well Package,
units, or when cells in a layer have variable thicknesses where the stress rates only can change at the beginning of a
(Harbaugh, 2005, p. 2-15). stress period. Another type of possible boundary condition
Head is computed at the center of each cell called a applied to variable-head cells is the head-dependent flow
node. The Block-Centered Flow Package (Harbaugh, 2005, boundary condition. Interaction of ground water with streams
p. 5-15), the Layer-Property Flow Package (Harbaugh, 2005, and lakes is an example of a head-dependent flow boundary
p. 5-17), or the Hydrogeologic-Unit Flow Package (Anderman condition, and are covered in detail in section Discretization
and Hill, 2000; Anderman and others, 2002) can be used to of Streams and Lakes. In addition to boundary conditions
calculate conductance terms for flow between cells and storage applied to variable-head cells representing the area of interest,
terms within cells. The Horizontal-Flow Barrier Package flow can be simulated into and(or) out of the finite-difference
(Hsieh and Freckleton, 1993) can be used with the Block- grid by using the General-Head Boundary or the Flow and
Centered Flow, Layer-Property Flow, and Hydrogeologic-Unit Head Boundary Packages.
Flow Packages to simulate the effects of ground-water flow
barriers such as a fault.
Cells used to simulate boundary conditions are grouped Discretization of Streams and Lakes
into two categoriesspecified-head (or constant-head for Because streams and lakes are a primary linkage between
certain conditions in which the specified head does not change the PRMS and MODFLOW-2005 codes, a more detailed
with time) and no-flow cells (Harbaugh, 2005, p. 2-13). A description of these MODFLOW-2005 packages is provided
head is specified in the data input for specified-head cells and here. In the GSFLOW implementation of MODFLOW-2005,
the head value does not change unless it is changed in the data ground-water interactions with streams are simulated with
input for subsequent time or stress periods. No-flow cells are the Streamflow-Routing Package (Prudic and others, 2004;
cells where flow into or out of a cell is not permitted. Ground- Niswonger and Prudic, 2005). This package determines
water head is not calculated for no-flow cells; for this reason, ground-water interactions along streams using a head-
they are referred to as no-flow cells. The remaining cells are dependent boundary condition in which the head or stage
referred to as variable-head cells because ground-water head along the stream is determined from the depth of flow in
is computed in these cells and the computed heads can vary the channel. Streamflow can be routed along a network
with time (Harbaugh, 2005, p. 2-13). No special designation of channels that may include rivers, streams, canals, or
is needed along the top or bottom of the finite-difference grid ditches, and are referred to collectively as streams. The
or where an edge of the aquifer system is coincident with the Streamflow-Routing Package originally was developed for
outside edge of the grid (fig. 7). Any cell within the finite- MODFLOW-2000 (Prudic and others, 2004) and was revised
difference grid can be designated as no-flow cells in order to by Niswonger and Prudic (2005) to simulate unsaturated
approximate no-flow boundaries of aquifer systems. No-flow flow beneath streams. The package was subsequently revised
cells can be surrounded by variable-head cells to represent for MODFLOW-2005 and then for GSFLOW to simulate
internal regions in the finite-difference grid where flow does kinematic-wave routing.
not occur.
nv07tac-4124_fig06
Design of GSFLOW 11
+y
+x
+z
COLUMNS
1 2 3 4 5 6 7 8 9 EXPLANATION
1
Aquifer boundary
S
2
W
O
R
3 Active cell
vk
4 Inactive cell
5 rj Dimension of cell along the row direction
Subscript indicates the number of the column
1
ci Dimension of cell along the column direction
2
LAYERS
Subscript indicates the number of the row

3 vk Dimension of cell along the vertical direction
ci Subscript indicates the number of the layer
4
5
rj
Figure 7. A discretized hypothetical aquifer system (modified from Harbaugh, 2005).
Streams are divided into reaches and segments COLUMNS

(Prudic and others, 2004) (fig. 8). A reach is defined as 1 2 3
a section of a stream that is associated with a particular Inflow
finite-difference cell. More than one stream reach can be Reaches
assigned to a particular finite-difference cell, but only one 1,1
1
finite-difference cell can be assigned to a single reach. 1,2
Reaches are grouped into segments that represent lengths Spring
of the stream between connections with another stream or
tributary, a lake, or a watershed boundary. User-specified 2,1
1,3
inflows to a stream that are external to inflows calculated Segment
ROWS
by the model are added to the stream at the upstream 2 junction 3,1
end of a segment. Specified outflow at the upstream or 3,2
downstream end of a segment can be used to divert water 4,1
from a stream to a pipeline or lined canal; the water that
is diverted in this way is removed from the modeled area
without ground-water interaction. The quantity of water 3 4,2
diverted is limited to the available flow entering the reach Flow
where a diversion is specified. Diverted water may be direction
added back into another segment as a specified inflow.
Four different options are available for diverting flow from Outflow
the farthest downstream reach of a stream segment; these
options are described in appendix 1 (input instructions for EXPLANATION
MODFLOW-2005). Streambed properties can vary among
1,3 Segment number and reach number
reaches within a segment but are considered uniform
within the reach (appendix 1).
Stream segments typically are numbered in Figure 8. Simple stream network having four segments
downstream order starting at the upstream segment and and eight reaches in a finite-difference model grid
ending at the farthest downstream segment (fig. 8) (Prudic consisting of three rows and three columns (modified
and others, 2004). There is flexibility in how segments can from Prudic and others, 2004).
be numbered, and an upstream tributary segment can have a A simple stream-lake network associated with a finite-
lower number than the downstream segment receiving its flow. difference grid of an aquifer having 3 layers, 7 rows, and
However, within a segment, reach numbers must increase in 7 columns is illustrated in figure 9. A total of 147 finite-
downstream order. The direction of streamflow remains the difference cells are used to define the rectangular aquifer
same throughout the simulation. Multiple segments may be and lake. Each finite-difference cell is designated first by the
joined as tributaries of a downstream segment and outflow layer number and followed by the row and column numbers,
from all segments entering a confluence are summed and respectively. Cell1,1,2 is the cell in layer 1 at row 1, and
added to the downstream segment. Only a single segment may column 2. The example includes two stream segments and one
emanate from a confluence. lake. Stream segment 1 has two reaches. Reach 1 of segment1
A simple stream network is illustrated in figure 8 with a is in cell1,1,3 and reach 2 is in cell1,2,4. Inflow to stream
rectangular finite-difference grid having 1 layer, 3 rows, and 3 segment 1 is specified at the beginning of reach 1. Outflow
columns. Each finite-difference cell is designated first by the from reach 2 in segment 1 enters the lake at cell1,3,4. Stream
layer number and followed by the row and column numbers, segment 2 also has two reaches. Reach 1 of segment 2 begins
respectively. Thus, cell1,1,2 is the cell in layer 1, row 1, and in cell1,6,5 and reach 2 is in cell1,7,6. The lake occupies
column 2. Segment 1 in this example has three reaches and nine cells in layer 1, three cells in layer 2, and zero cells in
inflow is specified for the upstream end of the first reach. layer 3. Vertical discretization of the lake is shown along
Reach 1 of segment 1 is in cell1,1,1; reach 2 is in cell1,1,2; row 4 (fig. 9B). Outflow from the lake to stream segment 2 is
and reach 3 is in cell1,2,2. Segment 2 has one reach in added as inflow to reach 1. Outflow from segment 2 exits the
cell1,2,3 and is used to represent ground-water discharge modeled area.
from a spring into a stream. Segment 3 has two reaches and
receives inflow from segment 2. Reach 1 of segment 3 is in
cell1,2,3 and reach 2 is in cell1,2,2. Segment 4 has two Time DiscretizationStress Periods and Time
reaches and receives inflow from the confluence of segments Steps
1and 3 (called a segment junction in fig. 8). Reach 1 of
Time also is discretized into increments because the
segment 4 is in cell1,2,2 and reach 2 is in cell1,3,2. When
calculation of ground-water heads can be a function of time
multiple stream reaches are assigned to a finite-difference cell
as well as space (Harbaugh, 2005, p. 2-3). Time increments
such as cell1,2,2 in figure 8, the simulated ground-water head
are allowed to vary during a simulation and are divided into
in the cell is used to calculate the flow of water between the
two levels (Harbaugh, 2005, p. 4-4). The first level is called
cell and each reach.
a stress period. Stress periods correspond to changes in the
Ground-water interactions with lakes can be simulated
specified boundary conditions such as fluctuations in the
with the Lake Package (Merritt and Konikow, 2000). The
rate of pumping from a cell. Stress periods also can be used
package determines ground-water interactions with lakes
to change user-specified inflow and outflow to streams and
using a head-dependent boundary condition in which the head
lakes, such as those resulting from diversions. A simulation
or stage in a lake is determined from the difference between
must have at least one stress period and stress-period lengths
inflows to and outflows from a lake. The package originally
can vary. Stress periods can be either steady state or transient
was developed for MODFLOW-2000 (Harbaugh and others,
(Harbaugh, 2005, p. 3-6). A steady-state stress period neglects
2000) and has been revised and converted for use with
changes in storage, whereas a transient stress period considers
MODFLOW-2005 and GSFLOW. The Lake Package also was
the effects of changes in storage in the calculation of ground-
modified to include the capability to simulate unsaturated flow
water heads. Multiple steady-state stress periods can be
beneath lakes.
interspersed with transient stress periods when GSFLOW
Lakes are represented as a volume within a group of
is run in MODFLOW only mode; however, when run in
designated finite-difference cells, referred to here as lake
integrated mode, the stress periods must all be transient after
cells. Lake cells are different than other cells as they are not
the initial stress period.
directly considered in the computation of ground-water heads
The second level of time discretization is called a time
in MODFLOW-2005. Rather, a lake stage is computed by
step. Time steps are used to divide a change in stress into finer
the Lake Package and applied to all cells that comprise the
time increments that may be needed for solution convergence
lake. This lake stage is used to calculate ground-water lake
or to accurately solve the distribution of ground-water heads.
interaction. Multiple lakes can be delineated in the finite-
In addition, time steps also can be specified for printing model
difference grid. Lakes are numbered in any order from one to
output at specific times. The solution accuracy can be tested
the total number of simulated lakes. Lakes can be grouped to
by increasing the number of time steps for a stress period
allow them to coalesce during high lake stages and separate
until ground-water heads at the end of the stress period no
during low lake stages. Surface inflows to a lake can be from
longer change with increasing time steps. A steady-state stress
precipitation, surface runoff, point sources, and streams.
period has only one time step. Transient stress periods within
Surface outflows from a lake can be from evaporation, point
GSFLOW must have time steps that are divided into uniform
withdrawals, and streams. Multiple streams can enter and
daily time increments.
leave a lake from different locations.
Design of GSFLOW 13
A
COLUMNS
1 2 3 4 5 6 7
Inflow
1,1 Reaches
1
1,2 Segment
2
junction
Lake
inflow
3 Lake 1
ROWS
5 Lake
outflow
2,1
6
2,2
7
Outflow
1 Lake 1
LAYERS
1 2 3 4 5 6 7
COLUMNS
EXPLANATION
2,2 Segment number and reach number
Flow direction
Figure 9. Simple stream and lake network having one lake, two stream
segments, and four stream reaches in a finite-difference model grid
consisting of seven rows and columns and three layers: (A) plan-view of
active finite difference grid, (B) vertical section along model row 4.
Organization Processes are divided into packages. A package is defined

as the part of a process that pertains to a single aspect of
MODFLOW-2005 is organized into processes and the models operations. Packages that are used to simulate
packages (Harbaugh, 2005, p. 1-2). Each process and package time-varying, head-dependent or specified-flow boundary
contributes to the overall capabilities of MODFLOW-2005. conditions are called stress packages, whereas packages used
A process is defined as a section of the computer code to solve the system of simultaneous finite-difference equations
that solves a major equation or set of related equations. are called solver packages. Packages consist of a group of
Thus, the section of the computer program that solves the subroutines that are accessed from the MODFLOW-2005
ground-water flow equation is called the Ground-Water main program. Adetailed description of the organization
Flow Process. GSFLOW consists of the Ground-Water Flow of MODFLOW-2005 is presented by Harbaugh (2005,
(Harbaugh, 2005) and Observation Processes (Hill and Chapter1).
others, 2000). The Observation Process is used to compare
model-calculated heads and flows to measured values of
heads and flows. Processes that work with MODFLOW but Computational Sequence
are not currently (2007) supported by GSFLOW include:
A MODFLOW-2005 simulation begins by reading input
the Sensitivity and Parameter Estimation Processes of
data that do not change with stress periods. The number of
MODFLOW-2000 (Hill and others, 2000); the Ground-Water
cells in the grid for the problem to be simulated is determined
Transport Process (Konikow and others, 1996); the Local Grid
as well as the stress-package options and solution method.
Refinement Process (Mehl and Hill, 2006); the Ground-Water
Subsequently, memory is allocated for all aspects of the
Management Process (Ahlfeld and others, 2005); and the Farm
simulation. Following memory allocation, the finite-difference
Process (Schmid and others, 2006). However, GSFLOW was
grid, stress periods and time steps, and starting ground-water
designed with the intention to support these or other processes
heads and hydraulic properties are read for each finite-
as they become needed to solve future problems.
difference cell (fig. 10).
Run MODFLOW-2005
Allocate and Read
Stress-Period Loop
Read Boundary Conditions
Time-Step Loop
Iteration Loop
Compute Flow and Storage
Check Convergence;
Loop as Necessary
Save Results to File
Check Time Step;

Loop as Necessary
Check Stress Period;

Loop as Necessary
Deallocate Memory
Figure 10. Computational sequence of MODFLOW-2005 used for simulating ground-water flow through
aquifer systems (modified from Harbaugh, 2005).
Design of GSFLOW 15
The computational sequence follows three basic loops Initial ground-water heads are required for the first stress
(Harbaugh, 2005, p. 3-1). The first loop is the stress-period period of a simulation regardless of whether the stress period
loop, during which specified stress data remain constant. is designated as steady state or transient. Initial heads are
Inside the stress-period loop is the time-step loop where needed to calculate heads at the beginning of the simulation
ground-water heads and flows to and from head-dependent and to check for no-flow cells where the initial head is beneath
boundaries are computed at the end of each time step. Iterative the bottom of the cell. Initial heads do not affect the steady-
solution methods are used to solve for the heads for each time state solution other than determining if a cell is a no-flow cell
step in the iteration loop (Harbaugh, 2005, p. 2-10). or a variable-head cell. Initial heads are critically important
Flows, ground-water heads, and a summary water budget for transient solutions, however, because they set the initial
are saved or printed at the end of a time step when requested. storage conditions available for the subsequent transient
The program continues through each time step and stress run. A steady-state stress period can be used to determine
period until the end of the simulation. However, the program the initial head distribution for a subsequent transient stress
may terminate during a time step if the difference in heads period. Initial heads that are generated from a steady-state
between iterations does not decrease below a specified error stress period help reduce unrealistic changes in heads at the
tolerance within a specified number of iterations; this type beginning of a transient stress period and provide a defensible
of program termination is referred to as convergence failure. and representative basis for model calibration.
Convergence failure may be caused by various conditions
in the model that can usually be mitigated by increasing the
number of time steps (for MODFLOW-only simulations) or by Integration of PRMS and MODFLOW-2005
changing solver input parameters or other data input.
Flow is exchanged among the three regions of GSFLOW
The choice of daily time steps for GSFLOW was
on the basis of interdependent equations that calculate flow
made considering limitations of time-step length on model
and storage of water throughout the simulated hydrologic
convergence. Experience has shown that daily time steps are
system (fig. 11). The first region includes the plant canopy,
sufficiently small for MODFLOW to converge to a solution
snowpack, impervious storage, and soil zone.
for most simulations, and GSFLOW applications should not
require reduction in time-step length less than 1 day.
Region 1Plant canopy, Surface runoff

snowpack, surface- Interflow
Region 2Streams
depression storage, and lakes
and soil zone Soil-moisture (MODFLOW-2005)
(PRMS) dependent flow
Ground-water discharge Ground-water discharge
Soil-moisture or
Head-dependent flow
head-dependent flow
Gravity drainage Leakage
Region 3Subsurface
(unsaturated and saturated
zones) beneath soil zone
(MODFLOW-2005)
Figure 11. Schematic diagram of the exchange of flow among the three regions in GSFLOW.
The dependency on soil moisture and head in the computation of flow among the regions also is shown.
The second region consists of streams and lakes. The interflow to streams and lakes or how they could affect the
subsurface or third region is beneath regions 1 and 2, which link between the soil zone in PRMS with the unsaturated and
consists of the unsaturated and saturated zones. The first saturated zones in MODFLOW-2005.
region is simulated with PRMS modules, while the second and GSFLOW allows for three simulation modesintegrated,
third regions are simulated with MODFLOW-2005 packages. PRMS-only, and MODFLOW-only. All but two PRMS
A description of how GSFLOW simulates water flow within modules can be used in the integrated mode; the two modules
all three regions is provided in section Computations of that can not be used in the integrated mode are the Ground-
Flow. Water Flow and Streamflow Modules (gwflow_casc_prms
Specified inputs of precipitation and temperature and and strmflow_prms, table 1). All MODFLOW-2005 packages
specified inputs or model-estimated potential solar radiation listed in table 1 can be used for simulation using either the
are distributed to each HRU to compute energy budgets, flow, integrated or MODFLOW-only modes. PRMS modules,
and storage within region 1. A portion of the water entering MODFLOW-2005 packages, and GSFLOW modules that
region 1 infiltrates into the soil zone, where it is evaporated have been modified or created also are denoted in table 1
and transpired back to the atmosphere, flows to streams and and modules and packages are listed in their computational
lakes (region 2), and (or) drains to the deeper unsaturated and order. The capability of having PRMS-only and MODFLOW-
saturated zones (region 3). only simulations in GSFLOW allows incremental model
The rate at which water flows from the soil zone to setup that provides flexibility in calibration. For example,
streams and lakes is dependent on: (1) the rate at which water independent calibration of the example test problem
is added to the land surface by snowmelt and rain, (2) the during the development of GSFLOW allowed for a better
rate of infiltration into the soil zone, and (3) the antecedent understanding of the sensitivity of the models to the various
soil-zone storage. Water that flows from the soil zone to the model parameters, and resulted in an easier calibration of the
unsaturated and saturated zones (region 3) is called gravity integrated model.
drainage. Gravity drainage is dependent on the vertical
hydraulic conductivity of the unsaturated zone and the volume
of water stored in the soil zone. Additionally, gravity drainage Changes to PRMS
ceases as the water table rises into the soil zone. Water also Changes to PRMS modules include: (1) a new HRU-
can flow from the saturated zone into the soil zone as ground- based distribution method for solar radiation that may improve
water discharge; the rate of discharge is dependent on the the calculations of evapotranspiration and the energy budget
hydraulic conductivity and ground-water head relative to the of the snowpack; (2) a new Cascade Module designed to
altitude of the soil-zone base. Flow between the unsaturated simulate more complex flow paths in watersheds; (3) a new
and saturated zones to streams and lakes is dependent on the Soil-Zone Module designed to link with MODFLOW-2005;
ground-water head in relation to the stream- or lake-surface and (4) a new type of HRU designed to link precipitation and
altitude, the hydraulic properties of the streambed and lakebed evaporation with the Lake Package in MODFLOW-2005.
sediments, and the hydraulic properties of the unsaturated and The new HRU-based distribution method for solar
saturated zones. radiation was developed to distribute daily solar radiation on
the basis of the average slope and aspect and the latitude of the
PRMS Modules and MODFLOW-2005 Packages centroid of each HRU. The new method computes potential
solar radiation for each Julian day of the year for each HRU.
Implemented in GSFLOW This differs from the original distribution method in which
A total of 25 PRMS modules and 16 MODFLOW-2005 each HRU was assigned to a radiation plane determined by
packages are implemented in GSFLOW, (table 1). PRMS slope, aspect, and latitude. Daily potential solar radiation for
modules used in GSFLOW were reformatted to a consistent each radiation plane also was estimated by linear interpolation
coding style and updated using Fortran 90 (American National of 13 values that represent the annual cycle of potential solar
Standards Institute, 1992) language elements. Some PRMS radiation (Leavesley and others, 1983, p. 14-15). Changes
modules were revised during this conversion to include more to the computation and distribution of solar radiation were
accurate constants and to correct existing programming implemented in modules soltab_hru_prms, ddsolrad_hru_
errors. The functionality of the revised modules is identical prms, ccsolrad_hru_prms, and potet_hamon_hru_prms
to those in the original PRMS documentation except where (table1).
noted in table 1 and described in section Changes to PRMS. The new Cascade Module (cascade_prms; table 1) is
Additionally, some MODFLOW-2005 packages were designed to route surface runoff and interflow from upslope
modified to allow linkages to PRMS flows and to correct HRUs to downslope HRUs. Surface runoff and interflow
existing programming errors. Several PRMS modules and from each HRU for the daily mode of PRMS originally was
MODFLOW-2005 packages are not implemented in the initial added directly to streamflow (fig. 2 and Leavesley and others,
version of GSFLOW (table 2). Developers that intend to add 1983, p. 25 and 31). The new method allows surface runoff
new modules and packages to GSFLOW are cautioned to and interflow to satisfy soil-zone storage of downslope HRUs
evaluate how these additions could affect surface runoff and before being added as streamflow or inflow to a lake.
Design of GSFLOW 17
Table 1. Description of PRMS and GSFLOW modules and MODFLOW-2005 packages implemented in GSFLOW, listed in computational
order.
[Module or package name: User selects only one of the modules or packages in each group indicated by a number from 1 to 7. Model mode: G is integrated
simulation, P is PRMS-only simulation, M is MODFLOW-only simulation. HRU, hydrologic response unit]
Hydrologic or New or Model

Module or package Description
computational process modified mode
PRMS modules
Basin basin_prms Defines shared watershed-wide and HRU physical parameters and G,P
variables
Cascade cascade_prms Determines computational order of the HRUs and ground-water new G,P
reservoirs for routing flow downslope
Observed data obs_prms Reads and stores observed data from all specified measurement G,P
stations
Solar table soltab_hru_prms Computes potential solar radiation for each HRU; modification of modified G,P
soltab_prms
Temperature (1) temp_1sta_prms Distributes maximum and minimum temperatures to each HRU using G,P
temperature data measured at one station and an estimated monthly
lapse rate
temp_laps_prms Distributes maximum and minimum temperatures to each HRU using G,P
temperature data measured at a base station and lapse station with
differing altitudes
xyz_dist Distributes precipitation and maximum and minimum temperatures G,P
to each HRU using a multiple linear regression of measured
data from a group of measurement stations or from atmospheric
model results. Selection requires that module also be used for
precipitation
temp_dist2_prms Distributes temperature from one or more stations to each HRU by an G,P
inverse distance weighting scheme
Precipitation (2) precip_prms Distributes precipitation from one or more stations to each HRU and G,P
determines the form of precipitation using monthly correction
factors to account for differences in altitude, spatial variation,
topography, and measurement gage efficiency
precip_laps_prms Distributes precipitation from one or more stations to each HRU G,P
and determines form of precipitation using monthly lapse rate
adjustments to account for the differences in altitude of each HRU
relative to each measurement station
xyz_dist Distributes precipitation and maximum and minimum temperatures G,P
to each HRU using a multiple linear regression of measured data
from a group of measurement stations or from atmospheric model
results. Selection requires that module also be used for temperature
precip_dist2_prms Distributes precipitation from one or more stations to each HRU by G,P
an inverse distance weighting scheme
Solar radiation (3) ddsolrad_hru_prms Distributes solar radiation to each HRU and estimates missing solar modified G,P
radiation data using a maximum temperature per degree-day
relation; modification of ddsolrad_prms
ccsolrad_hru_prms Distributes solar radiation to each HRU and estimates missing solar modified G,P
radiation data using a relation between solar radiation and cloud
cover; modification of ccsolrad_prms
Potential potet_jh_prms Determines whether current time period is one of active transpiration, G,P
evapotranspiration (4) and computes the potential evapotranspiration for each HRU using
the Jensen-Haise formulation (Jensen and others, 1969)
potet_hamon_hru_prms Determines whether current time period is one of active transpiration, modified G,P
and computes the potential evapotranspiration for each HRU using
the Hamon formulation (Hamon, 1961); modification of
potet_hamon_prms
potet_epan_prms Determines whether current time period is one of active transpiration G,P
and computes the potential evapotranspiration for each HRU using
pan-evaporation data
order.Continued
simulation, P is PRMS-only simulation, M is MODFLOW-only simulation; HRU, hydrologic response unit]

PRMS modulesContinued
Interception intcp_prms Computes volume of intercepted precipitation, evaporation from G,P
intercepted precipitation, and throughfall that reaches the soil or
snowpack
Snow snowcomp_prms Initiates development of a snowpack and simulates snow accumulation G,P
and depletion processes using an energy-budget approach
Surface runoff and srunoff_smidx_casc Computes surface runoff and infiltration for each HRU using a non- modified G,P
infiltration (5) linear variable-source-area method allowing for cascading flow;
modification of srunoff_smidx_prms
srunoff_carea_casc Computes surface runoff and infiltration for each HRU using a modified G,P
linear variable-source-area method allowing for cascading flow;
modification of srunoff_carea_prms
Soil zone soilzone_gsflow Computes inflows to and outflows from soil zone of each HRU and new G,P
includes inflows from infiltration, ground-water, and upslope HRUs,
and outflows to gravity drainage, interflow, and surface runoff to
down-slope HRUs. Inflow from ground water is not allowed when
using the PRMS-only mode; merge of smbal_prms and ssflow_prms
with enhancements
Ground water gwflow_casc_prms Sums inflow to and outflow from PRMS ground-water reservoirs; modified P
outflow can be routed to downslope ground-water reservoirs, stream
segments, and sink modification of gwflow_prms
Streamflow strmflow_prms Computes daily streamflow as the sum of surface runoff, interflow, P
detention reservoir interflow, and ground-water flow
Summary hru_sum_prms Computes daily, monthly, yearly, and total flow summaries of volumes G,P
and flows for each HRU
basin_sum_prms Sums values for daily, monthly, yearly, and total flow summaries of G,P
volumes and flows for all HRUs
MODFLOW Packages
Basic BAS Handles a number of basic administrative tasks G,M
Block centered flow (6) BCF Calculates conductance coefficients for ground-water flow equations
G,M
using a block-centered flow package
Layer property flow (6) LPF Calculates conductance coefficients for ground-water flow equations
G,M
using a layer-property flow package
Hydrostratigraphic unit HUF Calculates effective hydraulic properties for model layers using
G,M
flow (6) hydrostratigraphic units
Horizontal flow barrier HFB Simulates flow barriers by reducing horizontal conductance G,M
Well WEL Adds terms to flow equation to represent wells G,M
Multi-node well MNW Adds terms to flow equation for wells that extract or inject water in
G,M
more than one cell
General head boundary GHB Adds terms to flow equation to represent general head-dependent
G,M
boundaries
Constand head boundary CHD Adds terms to flow equation to represent constant-head boundaries G,M
Flow and head boundary FHB Adds terms to flow equation to represent flow and head boundaries G,M
Streamflow routing SFR Adds terms to flow equation to represent ground-water and stream modified G,M
interactions; modification of SFR2
Lake LAK Adds terms to flow equation to represent ground-water and lake modified G,M
interactions; modification of LAK3
Unsaturated zone flow UZF Adds terms to flow equation to represent recharge from the modified G,M
unsaturated zone, evapotranspiration, and ground-water discharge to
land surface
Design of GSFLOW 19
order.Continued
simulation, P is PRMS-only simulation, M is MODFLOW-only simulation; HRU, hydrologic response unit]

MODFLOW PackagesContinued
Gage GAG Prints time series gage output for selected stream reaches and lakes; modified G,M
modification of GAG5
Solver (7) SIP Solves simultaneous equations resulting from finite-difference G,M
approximations using the strongly implicit procedure
PCG Solves simultaneous equations resulting from finite-difference modified G,M
approximations using a preconditioned conjugate-gradient
procedure
DE4 Solves simultaneous equations resulting from finite-difference G,M
approximations using a direct solution procedure
GMG Solves simultaneous equations resulting from finite-difference G,M
approximations using a geometric multigrid solution of the
preconditioned conjugate-gradient procedure
GSFLOW modules
Computation Order gsflow_prms Controls model mode and computational sequence order of PRMS modified G,P,M
and GSFLOW modules for each time step; modification of the
PRMS call_modules.c subroutine
gsflow_modflow Controls sequence order for reading time-dependent input data and modified G,M
controls the computational sequence of calculations in the time-
step and iteration loops for MODFLOW-2005 packages and
PRMS and GSFLOW modules dependent on flows to and from
MODFLOW; modification of the MODFLOW-2005 main program
Conversion gsflow_setconv Defines variables used for converting between PRMS and new G
MODFLOW-2005 units
Integration gsflow_prms2mf Distributes gravity drainage from gravity reservoirs in soil zone new G
of PRMS to finite-difference cells in MODFLOW-2005; also
distributes surface runoff and interflow from HRUs to stream
segments and lakes in MODFLOW-2005
gsflow_mf2prms Distributes ground-water discharge from finite-difference cells in new G
MODFLOW-2005 to gravity reservoirs in soil zone of PRMS
Summary gsflow_budget Computes watershed budget for GSFLOW and adjusts final storage new G
in gravity reservoirs using flows to and from finite-difference cells
at end of each time step
gsflow_sum Computes detailed water budgets for all flow components at end of new G
each time step
Table 2. PRMS modules and MODFLOW-2005 packages not implemented in GSFLOW.
[HRU, hydrologic response unit]
Hydrologic or
computational process
PRMS modules
Solar radiation ccsolrad_prms Distributes solar radiation as determined for radiation planes to each HRU and
estimates missing solar radiation data using a relation between solar radiation and
cloud cover
ddsolrad_prms Distributes solar radiation as determined for radiation planes to each HRU and
estimates missing solar radiation data using a maximum temperature per degree-
day relation
Storm infiltration grnampt_infil_prms Compute infiltration for each HRU during storm events using a modified Green and
Ampt infiltration approach (Green and Ampt, 1911)
Storm surface runoff krout_ofpl_prms Computes surface runoff for each overland flow plane during storm events using
kinematic routing and sediment detachment and transport using a rill-interrill
concept
Surface runoff and infiltration srunoff_carea_prms Computes surface runoff and infiltration for each HRU using a non-linear variable-
source-area method
srunoff_smidx_prms Computes surface runoff and infiltration for each HRU using a linear variable-
source-area method
Soil moisture accounting smbal_prms Computes soil-moisture accounting for each HRU including addition of infiltration,
computation of actual evapotranspiration, and seepage to subsurface and ground-
water reservoirs
Subsurface flow ssflow_prms Adds inflow to subsurface reservoirs and computes outflow to ground-water
reservoirs and to streamflow
Ground water gwflow_prms Sums inflow to ground-water reservoirs and computes outflow to streamflow and to
a ground-water sink if specified
Storm channel flow krout_chan_prms Computes flow for each stream segment during storm events using kinematic
routing, reservoir routing, and sediment transport
Streamflow strmflow_st_prms Computes daily streamflow as the sum of surface, subsurface, and ground-water flow
contributions, storm runoff totals for storm periods, and daily detention reservoir
routing
MODFLOW-2005 Packages
Recharge RCH Adds terms to ground-water flow equation to represent areal recharge to ground-
water system
Evapotranspiration EVT Adds terms to ground-water flow equation to represent head-dependent
evapotranspiration from ground-water system
Segmented ETS Adds terms to ground-water flow equation to represent segmented head-dependent
evapotranspiration evapotranspiration from ground-water system
Interbed storage IBS Adds terms to ground-water flow equation to represent inelastic compaction of fine-
grained sediments
Subsidence SUB Simulates aquifer-system compaction and land subsidence
Tile drain DRT Adds terms to ground-water flow equation to represent ground-water discharge to
drains while accounting for irrigation return flows
River RIV Adds terms to ground-water flow equation to represent rivers to represent head-
dependent flow between a surface water body and a ground-water system.
Drain DRN Adds terms to ground-water flow equation to represent ground-water discharge to
drains
Reservoir RES Adds terms to ground-water flow equation to represent leakage from reservoirs
Design of GSFLOW 21
The new Soil-Zone Module (table 1) was developed to The new Soil-Zone Module replaces the finite-volume
link the soil zone in PRMS to the subsurface zone, streams, used to account for water content below field capacity in
and lake simulated by MODFLOW-2005 (fig. 12). The new the soil-zone reservoir and the infinite-volume used for the
module is a replacement of the original PRMS Soil-Moisture subsurface reservoir with three finite-volume reservoirs.
Accounting and Subsurface-Flow Modules (Leavesley and These finite-volume reservoirs are the capillary reservoir,
others, 1983, p. 19 and 31); thus, the original modules are not preferential-flow reservoir, and gravity reservoir (fig. 12B).
implemented in GSFLOW (table 2). The old Soil-Moisture The finite-volume reservoirs were developed to allow for
Accounting Module was used to compute flows into and out of saturation excess in the soil zone to become surface runoff
the soil-zone reservoir, and the Subsurface-Flow Module was (Dunne and Black, 1970; Freeze, 1972). Saturation excess
used to compute flows into and out of the subsurface reservoir is referred to here as Dunnian runoff. Dunnian runoff
(fig. 12A).
A. Original PRMS soil-zone water B. Revised PRMS soil-zone water
Wilting threshold Saturation threshold

Increasing storage
Increasing storage
Soil-zone reservoir Preferential-flow reservoir
Soil-zone water
Field-capacity threshold Preferential-flow threshold
Field-capacity threshold Preferential-flow threshold
Increasing storage
Finite part of reservoir
Infinite part of reservoir
Gravity reservoir
Increasing storage
Field-capacity threshold
Subsurface reservoir
Unsaturated- Field-capacity threshold
zone Increasing storage
water
Capillary reservoir
Wilting threshold
Figure 12. Differences in conceptualization of soil-zone water simulated by PRMS, where boxes
are storage reservoirs representing pore-space volumes for a given volume of soil: (A) original
conceptualization of a finite-volume soil-zone reservoir and an infinite-volume subsurface reservoir;
(B) revised conceptualization of three finite-volume reservoirs: preferential flow, gravity, and capillary.
differs from Hortonian runoff, which is generated when applications. Several new parameters are required for both the
the precipitation rate exceeds the infiltration rate of the gravity and preferential-flow reservoirs and must be created or
soil that may not be saturated (Horton, 1933). Hortonian added to the imported datasets. Descriptions of the parameters
runoff is computed in both PRMS Surface Runoff Modules needed for PRMS in GSFLOW are in appendix 1.
(srunoff_smidx_casc or srunoff_carea_casc; table 1). PRMS-only simulations can include fewer ground-water
Capillary, preferential-flow, and gravity reservoirs are shown reservoirs than HRUs (fig. 13). The revised Ground-Water
as separate boxes within figure 12 because they represent Flow Module (gwflow_casc_prms, table 1), computes ground-
separate storage reservoirs for computational purposes in water flow to streams and allows for cascading flow among
the model; however, in reality, they are all contained within the ground-water reservoirs. The original PRMS Streamflow
the same physical space. Another reason for separating them Module (strmflow_prms, table 1) is unchanged and associated
into separate stacked boxes is that their order in the vertical parameters must be specified in the PRMS Parameter File. This
sequence corresponds to the order in which they are filled and PRMS module, like the Ground-Water Flow Module, is limited
drained, from the bottom up. to PRMS-only simulations (table 1).
The capillary reservoir is the same as the original A designation for HRUs was added as a parameter to
soil-zone reservoir and represents water held in the soil PRMS (parameter hru_type) to distinguish between land-
by capillary forces between the wilting and field-capacity based, lake, and inactive HRUs (fig. 14, appendix 1). Land-
thresholds. Water is removed from the reservoir by evaporation based HRUs are assigned an integer value of 1 and lake HRUs,
and transpiration. which encompass the maximum surface area of a lake, are
The gravity reservoir replaces the subsurface reservoir assigned an integer value of 2. Only precipitation, potential
and represents water in the soil zone between field-capacity solar radiation, and lake evaporation are computed for lake
and saturation thresholds that is not subject to the preferential- HRUs. Other processes such as plant canopy interception,
flow threshold. This reservoir was developed to provide surface runoff, and interflows through the soil zone are
gravity drainage from the soil zone to the unsaturated zone assumed zero and snowpack is computed as a liquid equivalent.
simulated by MODFLOW-2005. The gravity reservoir also HRUs that are inactive in the simulation are assigned an integer
is capable of receiving ground-water discharge into the soil value of zero and no computations are performed.
zone whenever the ground-water head in a connected finite- Several PRMS modules are not implemented in GSFLOW
difference cell is greater than the soil-zone base (fig. 12B). (table 2). The PRMS modules ccsolrad_prms, ddsolrad_prms,
Gravity drainage is added to the ground-water reservoir in srunoff_carea_prms, srunoff_smidx_prms, and gwflow_prms
PRMS instead of to finite-difference cells when using the were replaced by modified versions for GSFLOW. The
PRMS-only mode in GSFLOW. The ground-water reservoir Soil-Zone Module (table 1) was developed for the purpose
in PRMS only can discharge water to a stream, a downslope of linking PRMS with MODFLOW-2005. The new module
ground-water reservoir, and/or a ground-water sink. Thus, replaces PRMS Soil-Moisture Accounting and Subsurface-
the gravity reservoir is not capable of receiving ground-water Flow Modules (smbal_prms and ssflow_prms, table 2).
discharge when using the PRMS-only mode. Water in the PRMS modules grnampt_infil_prms, krout_chan_prms,
gravity reservoir also is available for downslope flow within krout_ofpl_prms, and strmflow_st_prms are not available
the soil zone. The downslope flow is referred to herein as slow because PRMS storm mode is not implemented in this version
interflow. of GSFLOW. Details on the implementation of the changes
The preferential-flow reservoir is a new reservoir that is to PRMS that affect the rate and timing of surface runoff and
not part of the original concept of the soil zone. This reservoir interflow to streams and lakes, and interactions with ground
represents soil water between field capacity and saturation water are described in section, Computations of Flow.
that is available for fast interflow through relatively large
openings in the soil of each HRU (fig. 12B). The capacity of
this reservoir is depicted in figure 12B as the area between the Changes to MODFLOW-2005
preferential-flow and saturation thresholds. Both fast interflow Changes to MODFLOW-2005 needed for GSFLOW
and slow interflow are routed to downslope HRUs, streams, include refinements to existing MODFLOW-2005 packages.
and lakes at the end of a time step. The Unsaturated-Zone Flow Package (Niswonger and others,
The original conceptualization of the subsurface reservoir 2006a) was developed for GSFLOW to simulate ground-
allowed for fewer reservoirs than HRUs (fig. 13). This water recharge and discharge to land surface by accounting
conceptualization has been changed with the revision of the for vertical flow through an unsaturated zone. However, the
soil zone. At least one gravity reservoir must correspond to Unsaturated-Zone Flow Package also was designed to be
each HRU. Datasets created for PRMS using fewer subsurface used in MODFLOW-only simulations and was released as an
reservoirs than HRUs must be revised before importing the independent MODFLOW-2005 package prior to the release
dataset into GSFLOW by replacing the subsurface reservoirs of GSFLOW (Niswonger and others, 2006a). Modifications
with gravity reservoirs. Some of the original parameters used to existing packages included revisions to the Streamflow-
for the subsurface reservoirs are used for the gravity reservoirs Routing Package (Niswonger and Prudic, 2006) so that
to maintain compatibility with existing PRMS model
Design of GSFLOW 23
A.
Hydrologic response units 1 2 3 4 5 6
Infiltration Infiltration
Land surface
Soil-zone reservoir
(evapotranspiration)
segment
Stream
Subsurface reservoir
(interflow)
Ground-water reservoir Ground-water

flow to stream
Ground-water sink
B.
Hydrologic response units 1 2 3 4 5 6
Surface Surface
Infiltration runoff runoff Infiltration
Land surface
Soil zone3 reservoirs
Preferential-flow reservoir fast fast
(fast interflow) interflow interflow
Gravity reservoir
(slow interflow) slow slow
(gravity drainage) interflow interflow
segment
Stream
Capillary reservoir Gravity Gravity

(evapotranspiration) drainage drainage
Ground-water reservoir Ground-water

flow to stream
Ground-water sink
Figure 13. Conceptual changes to the soil-zone, shallow-subsurface, and ground-

water reservoirs for a PRMS-only simulation: (A) original conceptualization of PRMS
(Leavesley and others, 1983), and (B) revised conceptualization that includes cascading
flow of surface runoff and interflow from upslope hydrologic response units to downslope
hydrologic response units.
streams can receive surface runoff and interflow from HRUs surface (fig. 15). It replaces the standard Recharge and
and for kinematic-wave routing of streamflow down channels; Evapotranspiration Packages of MODFLOW-2005, and these
revisions also were made to the Lake Package (Merritt and packages are not included in GSFLOW. Three modifications
Konikow, 2000) to link a MODFLOW designated lake with were made to the original version of the Unsaturated-Zone
a lake HRU, to refill intermittent lakes from surfacewater Flow Package released with MODFLOW-2005. These
inflow, and to simulate unsaturated flow beneath lakes. changes included adding the ability to simulate unsaturated
The Unsaturated-Zone Flow Package is used in flow beneath lakes and changing the formulation used for
GSFLOW to simulate flow through the unsaturated zone, transitioning between conditions of ground-water recharge
ground-water recharge, and ground-water discharge to land and ground-water discharge within a finite-difference cell.
nv07tac-4124_fig13
1 1
1
1 1
Lake
1
2
1
1 1
EXPLANATION
1 Land-based hydrologic response unit and parameter hru_type designation number
2 Lake hydrologic response unit and parameter hru_type designation number
Watershed boundary
Stream
Figure 14. Delineation of land-based and lake hydrologic response units of a hypothetical watershed.
The latter modification was made to improve the efficiency The Streamflow-Routing Package also was revised to
of MODFLOW convergence. Both of these modifications are receive flows from PRMS as surface runoff and interflow
described in more detail in section Computations of Flow. from adjacent HRUs. Surface runoff and interflow are added
The Streamflow-Routing Package is used in GSFLOW to stream reaches by connecting HRUs to stream segments.
to route water in channels, to calculate streambed leakage, The volume of runoff and interflow are distributed to each
and to receive surface runoff and interflow from PRMS. The stream reach in a segment on the basis of the fraction of HRU
capability of routing transient streamflow down a network of associated with a stream reach. The River and Drain Packages
channels was added to the Streamflow-Routing Package using in MODFLOW-2005 are not implemented in GSFLOW
a kinematic-wave routing method. Transient routing was added because they do not have an accounting procedure for surface
to consider simulations in which changes in channel storage flows and because these head-dependent boundary types can
are important on a daily time step, such as in long rivers. be simulated with the Streamflow-Routing Package.
07tac-4124_fig14
Design of GSFLOW 25
Infiltration Evaporation
Soil zone Ho
rto
nia
nr
Unsaturated-zone flow un
off
Water table Recharge Evapotranspiration
Int
erf
low Well
Distributed discharge
Dunn Lake
ian flow
Ground-w runof
ater flow f
routed in
Saturated zone stream
root uptake
Ground-water Stream
discharge leakage
EXPLANATION
Soil-zone base
Figure 15. Changes to MODFLOW-2005 implemented in GSFLOW include: unsaturated zone is connected to PRMS at the soil-
zone base; unsaturated flow is simulated beneath lakes; distributed flow is routed in streams.
The Lake Package (Merritt and Konikow, 2000) is used Connecting Hydrologic Response Units to Finite-
in GSFLOW to simulate (1) lake storage, stream inflows
and outflows, and lakebed leakage, and (2) to receive
Difference Cells
precipitation, interflow, and surface flow from PRMS lake An important component to the coupling of the PRMS
HRUs. PRMS also calculates the evapotranspiration rate and MODFLOW models is the process used to spatially link
from the surface of lakes that is removed from available lake the HRUs used by PRMS with the finite-difference cells
storage. Other changes to the Lake Package include a new used by MODFLOW. This is done through the generation of
method for simulating dry lake conditions that occur while gravity reservoirs, which are used to transfer water between
there are inflows to and outflows from the lake. This revision HRUs and finite-difference cells (fig. 16). Because HRUs
can simulate refilling of a dry lake caused by increases in and finite-difference cells can have different spatial extents,
surface inflows. The Lake Package published by Merritt and the spatial extent of each gravity reservoir is defined by the
Konikow (2000) only allowed an empty lake to refill when intersection of the HRUs and finite-difference cells (fig. 16).
the average ground-water head in cells beneath the lake was A unique identification number is assigned to each gravity
greater than the altitude of the lake bottom. The final revision reservoir, starting with 1 and ending with the total number of
was the linking of the Unsaturated-Zone Flow Package to the gravity reservoirs. HRUs and finite-difference cells can have
Lake Package for the purpose of simulating unsaturated flow more than one gravity reservoir within their areas. In figure16,
beneath lakes. Unsaturated flow beneath lakes can occur when for example, there are 10 HRUs (including the lake HRU),
the ground-water head is beneath the lakebed. 65 active MODFLOW finite-difference cells, and 98 gravity
Three additional packages available in MODFLOW-2005 reservoirs. Any gravity reservoirs associated with lake HRUs
were not included in GSFLOW. The Reservoir Package are ignored by the model. The minimum number of gravity
(Fenske and others, 1996) was not implemented because its reservoirs in a GSFLOW simulation results when an HRU is
capabilities are available in the Lake Package. Two other specified for each finite-difference cell (the plan-view area of
packages were not implemented in GSFLOW because they each HRU and finite-difference cell corresponds); in that case,
have not been properly linked with the Unsaturated-Zone Flow the total number of gravity reservoirs would equal the number
Package. These packages are the Interbed Storage Package of HRUs.
(Leake and Prudic, 1991), and the Subsidence Package Four topologic parameters need to be specified to
(Hoffmann and others, 2003). Details on the implementation relate each gravity reservoir to their corresponding HRU
of changes to MODFLOW-2005 that affect the rate and timing and finite-difference cell (table 3). Gravity drainage from
of ground-water flow and interactions with surface water are each gravity reservoir is added to the connected finite-
described in section Computations of Flow. difference cell in GSFLOW module gsflow_prms2mf (table 1).
nv07tac-4124_fig15
MODFLOW column
A 1 2 3 4 5 6 7 8 9 10
2 Lake
3
1
4
1 2
Area of part B
MODFLOW row
2
2 3
5
3
1
4
6 4
3 1
7
8
3
9 4
B EXPLANATION
MODFLOW column
5 6 7
Inactive MODFLOW cell
Hydrologic response unit boundary

4
MODFLOW row
Watershed boundary
Model boundary
5
1 Stream reach number
Figure 16. (A) example distribution of gravity reservoirs in the soil zone of hydrologic response units used to connect
hydrologic response units in PRMS with finite-difference cells in MODFLOW. (B) includes 4hydrologic response
nv07tac-4124_fig16
units, 6 finite-difference cells, and 9 gravity reservoirs. Lake hydrologic response units do not have gravity reservoirs
because the soil zone is absent.
Design of GSFLOW 27
Similarly, ground-water discharge from a finite-difference cell Connecting Hydrologic Response Units to
is added to the connected gravity reservoir using GSFLOW
module gsflow_mf2prms (table 1). Slow interflow out of each
Streams and Lakes
gravity reservoir is aggregated by HRU. Interflow can be HRUs are connected to streams for the purpose of
routed as cascading flow from upslope HRUs to downslope routing surface runoff and interflow to stream segments
HRUs, terminating in a stream segment or lake. and reaches defined in the Streamflow-Routing Package of
The total area of all gravity reservoirs in an HRU must MODFLOW-2005. Five topologic parameters need to be
equal the total area of the HRU in order to conserve mass specified as part of the model input to relate stream segments
(that is, the sum of PRMS parameter gvr_hru_pct for each and reaches to corresponding HRUs that are contiguous to
HRU must equal 1). Similarly, the total area of all gravity them (table 4). These parameters can be determined using a
reservoirs connected to a finite-difference cell must equal the GIS analysis by identifying all unique intersections among
top area of the cell (that is, the sum of GSFLOW parameter HRUs, stream segments, and stream reaches. Multiple HRUs
gvr_cell_pct for each finite-difference cell must equal1). can contribute surface runoff and interflow to a stream
These parameters are described in appendix 1. Standard segment. Parameter hru_segment is used to route flow
analyses using GIS may not produce parameters of sufficient from HRUs to segments when the option to cascade flow is
accuracy and, consequently, the percentage of each gravity not used. Surface runoff and interflow from one HRU can
reservoir area in an HRU or connected to a finite-difference be distributed to multiple stream segments when an HRU
cell may require additional processing. The GSFLOW model cascades. Surface runoff and interflow to lakes are routed from
will issue warning messages when the absolute errors of the land-based HRUs to lake HRUs (fig. 14) on the basis of the
area-weighted sums exceed 0.000001. topological parameters for cascading flow among HRUs. Refer
to section Cascading Flow Procedure.
Table 3. Topological parameters used to relate areas of gravity reservoirs in the soil zone to hydrologic response units in PRMS and to
areas of finite-difference cells in MODFLOW-2005.
[HRU, hydrologic response unit; nhrucell, total number of intersections between HRUs and active finite-difference cells]
Parameter Array
Description
name dimension
gvr_hru_id nhrucell HRU identification number corresponding to a gravity reservoir.
gvr_cell_id nhrucell Finite-difference cell identification number corresponding to a gravity reservoir. Finite-difference cell
identification number is a unique number determined by the row, column, and layer number of the cell.
The unique cell number starts with one for a cell in row 1, column 1, and layer 1. The unique cell number
is incremented by one for each column along each row until the last column in the last row of layer 1.
The unique number is then continued using the same method through the last layer. The last unique cell
identification number is equal to the product of the number of columns, rows, and layers in the finite-
difference grid.
gvr_hru_pct nhrucell Decimal fraction of HRU occupied by gravity reservoir.
gvr_cell_pct nhrucell Decimal fraction of finite-difference cell occupied by gravity reservoir.
Table 4. PRMS parameters for module gsflow_prms2mf that relate hydrologic response units in PRMS to stream segments and
reaches in the Streamflow-Routing Package of MODFLOW-2005.
[HRU, hydrologic response unit; nhru, total number of HRUs; nreach, total number of stream reaches; and nsegment, total number of stream segments]
Parameter Array
Description
name dimension
hru_segment nhru Stream segment number corresponding to each HRU. When an HRU cascades to another HRU or
stream segment, hru_segment is set to zero in module gsflow_prms2mf.
local_reachid nreach Stream reach identification number within each stream segment listed in sequence from the first to
last stream segment.
numreach_segment nsegment Number of stream reaches in each stream segment.
reach_segment nreach Stream segment number of each stream reach.
segment_pct_area nreach Decimal fraction of HRU area contributing to a stream reach.
PRMS-calculated inflows to a stream segment or a for a steady-state stress period. Similarly any specified stream,
lake from contiguous HRUs include Hortonian and Dunnian lake, and ground-water inflows and outflows are specified in
runoff, slow interflow through the gravity reservoirs, and the respective MODFLOW package input file(s). Results from
fast interflow through the preferential-flow reservoirs. The the steady-state stress period are automatically used as initial
total contribution of surface runoff and interflow to a stream conditions for a steady-state time step of the subsequent stress
segment or lake is the area-weighted sum from all contiguous period in the integrated model. Although PRMS does not
HRUs. The total contribution of inflow to a stream segment have a steady-state capability for determining initial storages
is partitioned to individual stream reaches on the basis of the in the various reservoirs, an initial volume per unit area can
decimal fraction of the HRU associated with each stream be specified for each PRMS reservoir prior to beginning a
reach (table 4). Refer to section Computations of Flow for transient simulation. Alternatively, a spin up period can be
explanations of how runoff and interflow are calculated for included at the beginning of the simulation.
routing to HRUs, streams, or lakes.
Organization
Time DiscretizationStress Periods, Time Steps,
PRMS and MODFLOW-2005 were integrated without
and Iteration much alteration to the underlying codes because both
GSFLOW operates on a daily time step, and volumetric models have similar programming frameworks. PRMS and
flow rates are exchanged between PRMS and MODFLOW MODFLOW-2005 remain separate in the integrated model
during each daily time step. MODFLOW-2005 variable-length by continuing to handle their respective operations, such as
stress periods are used to specify changes in stress or boundary reading input files, determining simulation options, computing
conditions that are not calculated by PRMS, such as recharge solutions, and printing output results to files. Additionally,
and discharge rates specified in the Well Package input file. much of each models computational sequence remains the
However, boundary conditions applied to MODFLOW-2005 same within the integrated model, other than modifications
that are calculated by PRMS are applied on a daily basis, to make some PRMS and MODFLOW-2005 computations
such as gravity drainage through the base of the soil zone. In sequential within the iteration loop.
order to ensure synchronization between MODFLOW-2005 GSFLOW is organized using two Computation-Control
variable-length stress periods and GSFLOW daily time steps, Modulesone for PRMS (gsflow_prms; table 1) and another
GSFLOW requires that variable-length stress periods also use for MODFLOW-2005 (gsflow_modflow; table 1). The module
daily time steps. That is, although the number or time unit of gsflow_prms is used to call PRMS and GSFLOW modules in
variable-length stress periods in GSFLOW is not restricted, the the proper sequence, which varies for each simulation mode
time increment of each stress period must be an even multiple (integrated, PRMS-only, and MODFLOW-only).
of days. For example, the time increment of a stress period GSFLOW time-step incrementing is the same as that
in the input for the Discretization File (appendix 1) could used for the PRMS daily mode. Time is incremented through
be described in units of seconds, but the value must result in the total number of days in the specified simulation time
a whole number of days when divided by 86,400 seconds. period. Rather than having separate stress-period and time-
Moreover, the capability in MODFLOW-2005 to increase the step loops as in MODFLOW-2005, the conditions represented
length of time steps during a stress period is not included in by the variable-length stress periods commonly used in
GSFLOW. The multiplier that typically is used to increase MODFLOW-2005 are changed on the appropriate day of
the time step during a stress period for MODFLOW-2005 the GSFLOW simulation. Counters representing the stress-
simulations is automatically set to 1.0 in GSFLOW, and the period and time-step numbers used for MODFLOW-2005
specified time of each stress period is checked for compliance computations are incremented appropriately within the module
with the requirement of 1-day time steps. The program will gsflow_modflow according to the specified number of time
print an error message before stopping when a stress period steps within a stress period.
cannot be evenly divided into uniform time steps of 1 day. All MODFLOW-2005 procedures and the PRMS
Only the first stress period in the Discretization File Soil-Zone Module (table 1) are processed in the module
of MODFLOW-2005 can be designated as steady state for gsflow_modflow. The contents and coding style of module
integrated simulations. The steady-state designation is ignored gsflow_modflow are similar to those in the MODFLOW-2005
for all other stress periods. No computations pertaining to main program, the iteration loop is retained but the
PRMS are executed for an initial steady-state stress period. MODFLOW time-step and stress-period loops have been
Consequently, gravity drainage beneath the soil zone, as removed and are controlled through module gsflow_prms.
well as any other inflows and outflows to streams, lakes, and This module determines the current MODFLOW stress
ground water are required input for a steady-state simulation. period and calls the procedures specified in the MODFLOW
Gravity drainage beneath the soil zone is specified as an Name File in the proper sequence. The Soil-Zone Module
infiltration rate in the Unsaturated-Zone Flow Package File is called from within the MODFLOW formulate and budget
procedures. Computations of the volume of water transferred
Design of GSFLOW 29
between the Soil-Zone Module and the Unsaturated-Zone Additional source-code files related to the Modular Modeling
Flow, Lake, and Streamflow-Routing Packages are made System and the MODFLOW-2005 Ground-Water Flow and
during each cycle through the iteration loop. All other PRMS Observation Processes also are required to build the GSFLOW
computations are made before the iteration loop as described executable (table 5).
in section Computations of Flow.
Additionally, five new modules were developed for
integrating PRMS and MODFLOW-2005. Module Computational Sequence
gsflow_setconv computes unit conversions between PRMS The computational sequence of GSFLOW is illustrated
and MODFLOW-2005. Modules gsflow_prms2mf and with a flow chart shown in figure 17 and descriptions listed in
gsflow_mf2prms integrate the spatial units and transfer table 6. The sequence of these computations depends on the
dependent variables and volumetric flow rates between PRMS GSFLOW simulation mode. If a simulation does not include
and MODFLOW-2005. Module gsflow_budget is used to MODFLOW-2005, then MODFLOW-2005 packages are
compute and write an overall water budget for the simulated skipped, and the computational sequence has a single daily
system and module gsflow_sum is used to produce detailed time-step loop (fig. 5) and follows the sequence described
water budgets for all flow components. These new modules in section Description of PRMS. Similarly, if a simulation
were written in Fortran 90. does not include PRMS, then PRMS modules are skipped,
GSFLOW simulations require input for each active and the sequence of computations includes the standard
PRMS module and MODFLOW-2005 package. Input for all MODFLOW-2005 stress period, time step, and iteration
PRMS modules is contained within the PRMS Parameter loops (fig. 10) and follows the sequence described in section
File and PRMS Data File, whereas there is at least one Description of MODFLOW-2005. An integrated simulation
individual input file for each MODFLOW-2005 package. executes both PRMS modules and MODFLOW-2005
New parameters were added to the PRMS Parameter File to packages on a daily time step that includes variable-length
specify data needed for new and revised modules. These data stress periods and the iteration loop. The iteration loop
include parameters that connect HRUs to finite-difference solves the interdependent equations within GSFLOW, such
cells, lakes, and streams. A complete description of the data- as ground-water head dependent flows between PRMS and
input requirements for GSFLOW is included in appendix 1. MODFLOW-2005.
Table 5. GSFLOW main program, Modular Modeling System utility functions used with PRMS, and additional MODFLOW-2005
packages and files used in GSFLOW.
Computer Programming
Description
program file language
GSFLOW main program
gsflow_main.c C GSFLOW program main, which reads the Control File and initiates execution.
Modular Modeling System utility functions used with PRMS
mms_util.c C Utility functions that primarily handle verification, storage and retrieval of input parameters and output
variables for PRMS and GSFLOW modules.
defs.h, globals.h, C Include files that define variables and values used by the Modular Modeling System.
mms.h, nodes.h,
protos.h, structs.h
Packages and file used with MODFLOW for the ground-water flow process
openspec.inc Fortran Include file that defines values for OPEN-statements.
u2ddbl.f Fortran Package of two-dimensional array reading utility subroutines.
utl7.f Fortran Package of array reading utility subroutines that includes reading of text and optional conversions to
numeric values.
Packages used with MODFLOW for the observations process
obs2bas7.f Fortran Package of subroutines that handle observations for the Basic Package.
obs2chd7.f Fortran Package of subroutines that handle observations for the Constant-Head Package.
obs2ghb7.f Fortran Package of subroutines that handle observations for the General-Head Boundary Package.
Program control begins using Modular Modeling System gsflow_modflow for each cycle through the time-step loop
(MMS) utility functions to read the GSFLOW Control (steps 316). If specified by the user, integrated GSFLOW
File, PRMS Parameter File, and PRMS Data File. Next, simulations that begin with a steady-state stress period
the MMS declare procedure is executed in which PRMS execute MODFLOW-2005 procedures and exclude PRMS
data consistency is checked, memory requirements are computations from the initial steady-state stress period
allocated, the simulation mode is determined, and variables (step5), and data related to boundary conditions are read for
are initialized and read (steps 1 and 2; fig. 17 and table 6). some MODFLOW-2005 packages at the beginning of any
PRMS computations required for each time step are made additional stress periods not designated as an initial steady
by accessing PRMS modules within module gsflow_prms state (step 6).
(steps 3 and 7). Program control also enters the module
Table 6. Description of computational sequence for GSFLOW.

Sequence
Computation
No.
1 DeclareRun PRMS and GSFLOW declare procedures.
2 Initialize, allocate, and readRun PRMS and GSFLOW initialize procedures, MODFLOW-2005 allocate and read and
prepare subroutines, and an optional MODFLOW-2005 steady-state simulation.
3 Daily time-step loopRead measured daily values for PRMS, read data for active MODFLOW-2005 packages when new
stress period begins, and run MODFLOW-2005, advance time subroutines.
4 Check for new MODFLOW-2005 stress periodIf time step loop is at 1 and initial stress period is steady state then do step
5, otherwise if new stress period then do step 6.
5 Steady-state simulationAdd source and sink terms for all packages and Run MODFLOW-only simulation including
MODFLOW-2005 formulate and budget processes. Go to step 3 and increment daily time-step loop.
6 Read and set new MODFLOW-2005 boundary conditionsAdd source and sink terms for all packages except
Unsaturated-Zone Flow, Streamflow-Routing, and Lake Packages.
7 Compute land-surface and soil-zone hydrologic processes in PRMSDistribute precipitation, compute canopy interception
and evaporation, determine snowpack accumulation and snowmelt, impervious surface evaporation and storage, infiltration,
and Hortonian surface runoff (Horton, 1933) for each HRU.
8 Begin iteration loopSequence through coupled PRMS and MODFLOW-2005 components.
9 Compute soil-zone flow and storage in PRMSCompute evapotranspiration, interflow and Dunnian runoff (Dunne and
Black, 1970) to downslope HRUs and stream reaches, and gravity drainage to the unsaturated zone.
10 Transfer dependent variables and volumetric flow rates to MODFLOW-2005Transfer flow-dependent variables and
volumetric flow rates needed for computing unsaturated and saturated flow, streamflow routing, and lake stages and volumes
in MODFLOW-2005.
11 Begin formulate procedure in MODFLOW-2005Add source and sink terms for Unsaturated-Zone Flow, Streamflow-
Routing and Lake Packages.
12 Approximate a solution to the ground-water flow equationUse one of the MODFLOW-2005 solver packages to calculate
ground-water heads and head-dependent flows.
13 Transfer dependent variables and volumetric flow rates to PRMSTransfer flow-dependent variables and volumetric
flow rates needed to compute soil-zone flow and storage in PRMS from the Unsaturated-Zone Flow Package in
MODFLOW-2005.
14 Check for convergenceIterate (repeat steps 913) until solution has converged to a specified closure criterion for changes in
ground-water head and volumetric flow rate in MODFLOW-2005 and a specified closure criterion for changes in storage in
soil zone of PRMS.
15 Compute budgetsRun MODFLOW-2005 and GSFLOW budget procedures.
16 Write resultsRun GSFLOW, PRMS, and MODFLOW-2005 output procedures.
17 Check for end of simulationRepeat time loop (steps 314) until end of simulation period.
18 End of simulationClose files and clear computer memory.
Design of GSFLOW 31
Follow Sequence of Figure 10 Follow Sequence of Figure 5

MODFLOW-only PRMS-only
Integrated simulation
Run GSFLOW
Declare (step 1)
Initialize, Allocate, and

Read (step 2)
Daily Time-Step Loop (step 3)
Check for New MODFLOW-2005

Stress Period (step 4)
Steady-State Simulation
(step 5)
Read and Set New MODFLOW-2005

Boundary Conditions (step 6)
Compute Land Surface and Soil Zone

Time-step loop
Hydrologic Processes in PRMS (step 7)
Begin Iteration Loop (step 8)
Compute Soil-Zone Flow

and Storage in PRMS (step 9)
Iteration Loop
Transfer Variables and Flows from

PRMS to MODFLOW-2005 (step 10)
Begin Formulate Procedure in

MODFLOW-2005 (step 11)
Approximate Solution to
Ground-Water Flow Equation (step 12)
Transfer Dependent Variables and
Volumetric Flow Rates from
MODFLOW-2005 to PRMS (step 13)
Check for Convergence
(step 14)
Compute Budgets (step 15)
Write Results (step 16)
Check for End of Simulation

(step 17)
End of Simulation (step 18)
Figure 17. Overall computational sequence used for an integrated PRMS and MODFLOW-2005 simulation in GSFLOW.
(See table 6 for further explanation of sequence steps.)
nv07tac-4124_fig17
Module gsflow_modflow also includes the iteration model does not converge for a time step; however, GSFLOW
loop in which computations are made for the soil, streams, continues to the next time step using the non-converged
lakes, unsaturated, and saturated zones, and for transferring solution. Similarly, warning messages are printed to the
dependent variables and volumetric flow rates between PRMS MODFLOW-2005 Listing File if stream or lake stages do not
modules and MODFLOW-2005 packages (steps 814). converge for a time step.
Ground-water heads are calculated and model convergence
is checked at the end of iterations (step 14). Subroutines
associated with the budget procedure (step 15) are executed at Conversion of Units Between PRMS and
the end of a steady-state stress period (if specified) and prior MODFLOW-2005
to the beginning of each new time step. Results and budgets
Volumes in PRMS are expressed as acre-inch and
computed for all simulation modes can be saved at the end
volumetric flow rates are expressed in acre-inch per day.
of every time step (step 14). Volumetric flow and storage
For example, if the total soil-water equivalent depth in the
is computed each time step (steps 317) until the specified
soil zone of an HRU is 3.0 inches at the beginning time
ending time is reached (step 17). Once the simulation has
step, then 3.0 inches of water are available over the HRU for
reached the end of the simulation, all files are closed and
evapotranspiration, gravity drainage, and interflow. If the area
memory is released (step18).
of an HRU is 100 acres, then the volume of water in the soil
Climate-data distribution and land-surface hydrologic
zone is 300 acre-inch or 1,089,000 cubic feet or 30,837 cubic
processes are computed by PRMS at the beginning of each
meters. Volumes in MODFLOW-2005 are expressed in units
time step (step 7). After these processes are computed, the
of cubic length and volumetric flow rates are expressed in
coupled equations in GSFLOW (steps 814) are solved in the
units of cubic length per time. Time can be in units of seconds,
iteration loop. The computational sequence within the iteration
minutes, days, or years; length can be specified in units of
loop is important for linking PRMS to MODFLOW-2005, and
inches, feet, centimeters, or meters. Time and length units are
follows a required order for solving the coupled equations
specified by MODFLOW parameters ITMUNI and LENUNI,
representing the soil zone, streams, and lakes, and the
respectively, which are specified in the Discretization File of
unsaturated and saturated zones. Soil-zone calculations are
the Basic Package (Harbaugh, 2005; p. 8-11 and appendix1).
made prior to computing flow through streams, lakes, and
Historically, ITMUNI and LENUNI were not critical to
the unsaturated and saturated zones (step 9). Flow among the
MODFLOW calculations as the input was expected to be of
soil zone and unsaturated and saturated zones is dependent
consistent units. These parameters take on a new importance
on ground-water heads and storage in soil-zone reservoirs
in GSFLOW in that they control how units are converted
from the previous time step, for the first iteration, and the
between the integrated models.
previous iteration subsequently. Initial ground-water heads
Volumetric flow rates in PRMS are converted to units
can be obtained from the optional initial steady-state stress
specified in MODFLOW-2005 when the rates are transferred
period simulation or specified by the user. Computations for
from PRMS to MODFLOW-2005. Conversely, volumetric
the MODFLOW-2005 formulate procedure that involve head-
flow rates in the specified units in MODFLOW-2005 are
dependent boundaries (step 11) are made after the soil-zone
converted to PRMS units when the rates are transferred
computations are made, followed by computations to solve
from MODFLOW-2005 to PRMS. A set of variables is
for ground-water heads (step 12). The dependent variables
determined in the GSFLOW module gsflow_setconv (table1)
and volumetric flows computed by MODFLOW are then
to calculate the particular conversion between PRMS and
transferred to PRMS (step 13) for revising flow and storage
MODFLOW-2005 units for a simulation. The variables are
in the soil-zone reservoirs during the subsequent iteration
summarized in table 7. Due to the empirical nature of some
(step9).
algorithms in PRMS, units other than acre-inch or acre-
A solution is found after the storage in the soil zone,
inch per day are used in some PRMS modules (Leavesley
depths in streams and lakes, and ground-water heads change
and others, 1983). These different units do not affect the
and volumetric flows change less than the specified tolerances
conversion of dependent variables and volumetric flow rates
between iterations (step 14). Separate convergence tolerances
between PRMS and MODFLOW-2005.
are used for the soil zone, stream stages, lakes stages, and
ground-water heads. Non-convergence occurs if any of these
tolerances are not achieved within the specified maximum
number of iterations. A warning message is printed if the
Cascading-Flow Procedure 33
Table 7. Conversion of volumes, volumetric flow rates, length, and velocity for variables in modules
gsflow_prms2mf and gsflow_mf2prms in GSFLOW.
Multiply by
Convert from To obtain
GSFLOW variable
Volume
acre-inch in PRMS Acre_inch_to_Mfl3 cubic length in MODFLOW
cubic length in MODFLOW Mfl3_to_ft3 cubic feet in PRMS
Volumetric flow rate
cubic length per time step in MODFLOW Mfl3t_to_cfs cubic feet per second in PRMS
cubic feet per second in PRMS Sfr_conv cubic length per time step in MODFLOW
acre-inch per time step Acre_inch_to_cfs cubic feet per second in PRMS
Length
length in MODFLOW Mfl_to_inch inch in PRMS
Velocity
inch per day in PRMS Inch_to_mfl_t length per time step in MODFLOW
Cascading-Flow Procedure watershed typically used in the delineation of HRUs (fig.18A),

or, if HRUs are discretized using a finite-difference grid, flow
The cascading-flow procedure developed for routing paths among the finite-difference cells can be determined
surface runoff and interflow among HRUs and to streams from their land-surface altitudes (fig. 18B). The cascading
and lakes uses a directed, acyclic-flow network (Ford and parameters must account for all HRUs and each path must
Fulkerson, 1956). Flow paths start at the highest upslope terminate in a stream segment or lake and not include loops or
HRUs and continue through downslope HRUs until reaching circles within the path.
a stream segment or lake. Cascading flow may occur along The model checks the cascading parameters and prints
many different paths and satisfies soil-zone storage of error and warning messages prior to starting the time-step
downslope HRUs before being added as inflow to stream loop of either an integrated or PRMS-only simulation. An
segments or lakes. Surface runoff and interflow from one error message stops model execution, whereas a warning
upslope HRU may provide inflow to as many downslope message results in an adjustment of parameter values. The
HRUs, stream segments, and lakes as is required to following describes the error and warning messages as related
conceptualize drainage in the watershed. to cascading flow routing from HRUs. Similar error and
The cascading-flow procedure for routing surface warning messages are printed for cascade paths involving
runoff was added to two Surface-Runoff Modules PRMS ground-water reservoirs in a PRMS-only simulation.
(srunoff_smidx_casc and srunoff_carea_casc, table1). Error messages are printed when: (1) a lake HRU is specified
Cascading interflow was included in the new Soil-Zone to cascade to another HRU, as lake HRUs can only cascade to
Module. Additionally, the cascading-flow procedure was a stream segment; (2) the sum of the contributing area from
added to the Ground-Water Flow Module (gwflow_casc_prms, an HRU to downslope HRUs of all paths originating from
table 1) used for the PRMS-only mode of simulation to route an HRU is substantially different than the area of the HRU;
upslope PRMS ground-water reservoir discharge to downslope (3) an HRU(s) is not included in the cascading pattern; and
ground-water reservoirs until reaching a stream segment. (4) a circular path is detected. Warning messages are printed
PRMS parameters are used to define the flows between when: (1) a cascade path is found with a contributing area
one reservoir and another and the fraction of area in the from the upslope HRU that is less than PRMS parameter
upslope HRU that contributes flow to downslope HRUs cascade_ tolthe area is evenly distributed to other
(table8 and appendix 1). Input parameters can be derived paths originating from the HRU; and (2) when two or more
on the basis of flow accumulation and direction analysis of a flow paths that are specified having the same source and
Digital Elevation Model (U.S. Geological Survey, 2000) of the destinationthe paths are combined and computed as a single
path.
PRMS parameter cascade_tol (table 8) can

be used to ignore any downslope connections when the
Computations of Flow
contributing area of the upslope HRU is a negligible This section describes all of the computations that
fraction of total area of the upslope HRU (below 0.01 for are made by GSFLOW, beginning with climate inputs of
example). PRMS parameter cascade_flg can be set to temperature, precipitation, and solar radiation, and ending
a value of 1 to add all flow from an upslope HRU to one with ground water and its interactions with streams and lakes.
downslope HRU. Each downslope HRU is automatically Additional details on the computations of flow are described
determined on the basis of which downslope HRU has the in the documentation for PRMS (Leavesley and others,
greatest fraction of contributing area in the upslope HRU 1983; Leavesley and others, 1996a) and MODFLOW-2005
as specified by PRMS parameter hru_down_pct. PRMS (Harbaugh, 2005).
parameter hru_strmseg_down_id is used to define the Descriptions of parameters and variables used in
connection of HRUs to stream segments (fig. 18). Stream equations for computations of flow include units for each
segments are identified in the Streamflow-Routing Package parameter and variable. Units that are specified in a generic
of MODFLOW-2005. A separate set of PRMS parameters formatsuch as cubic lengthindicate that the units are
is used to define the direction of flow among ground-water defined by data input. Units that are specified in a particular
reservoirs, which allows for different patterns of cascading formatsuch as acre-inchindicate that the units in
flow among ground-water reservoirs than those for surface GSFLOW are predetermined. The type of units used for
runoff and interflow (table 8).
Table 8. Parameters for PRMS Cascade Module (cascade_prms) used to define connections for routing flow from upslope to
downslope hydrologic response units and stream segments and among ground-water reservoirs.
[HRU, hydrologic response unit; ncascade, total number of connections among HRUs and stream segments; ncascdgw, total number of connections among
ground-water reservoirs]
Parameter name Array dimension Description

Parameters for routing surface runoff and interflow among HRUs and to stream segments1
cascade_tol 1 Minimum contributing area of an upslope HRU considered for flow to a downslope HRU.
cascade_flg 1 Flag used to force outflow from an upslope HRU to be added as inflow to one downslope
HRU3.
hru_up_id ncascade Identification number of upslope HRU.
hru_down_id ncascade Identification number of downslope HRU.
hru_strmseg_down_id ncascade Identification number of stream segment that receives inflow from an upslope HRU.
hru_pct_up ncascade Fraction of outflow from upslope HRU to be added as inflow to a downslope HRU or stream
segment2.
Parameters for routing flow among ground-water reservoirs in PRMS-only simulations1
gw_up_id ncascdgw Identification number of upslope ground-water reservoir.
gw_down_id ncascdgw Identification number of downslope ground-water reservoir.
gw_strmseg_down_id ncascdgw Identification number of stream segment in PRMS that receives cascading flow from a
ground-water reservoir.
gw_pct_up ncascdgw Fraction of outflow from a ground-water reservoir to be added as inflow to a ground-water
reservoir or stream segment4.
1 Routing of surface runoff and interflow in the integrated simulation mode is to stream segments defined in the Streamflow-Routing Package of
MODFLOW-2005, or, when using the PRMS-only mode, to detainment reservoirs defined in the PRMS module strmflow_prms.
2 Fraction is the ratio of contributing area to total area of upslope HRU.
3 A zero allows for multiple routing of inflows and outflows among HRUs and a value of one forces routing from an upslope HRU to one downslope HRU that
has the greatest fraction of contributing area in the upslope HRU.
4 Fraction is the ratio of contributing area to total area of the ground-water reservoir that produces cascading flow.
A B
11
15 20 21 22
19 23
14
10
14 15 16 17 18
2 2
9
1 13 18 1
17
7 6 5 16
8 12 9 10 11 12 13
21
4 3 20
2 19
4 5 6 7 8
1
3 2 1
3 3
EXPLANATION
6 Hydrologic response unit and number
Watershed boundary
Direction of streamflow
Direction of runoff and interflow among
hydrologic response units
nv07tac-4124_fig18
Figure 18. Cascading flow of surface runoff and interflow among hydrologic response units and streams delineated from (A) topology, climate, and vegetation;
and (B) a finite-difference grid.
Computations of Flow 35
MODFLOW-2005 variables are specified in consistent units Z Z tsta

m
of length and time as defined by variables LENUNI for length THRU
= Tstam bmonth HRU taf HRU , (1)
and ITMUNI for time in the Discretization File. Most PRMS 1000
parameters and variables must be in units specified in the data
input instructions (appendix 1). The Definition of Symbols where
(appendix 2) lists all equation variables in alphabetical order. m
THRU iss the maximum ( or minimum ) daily
Parameters represented by equation variables are identified in temperature at each HRU for time step m,
the variable definitions in courier font. in degrees Fahrenheit or Celsius;
References to other publications are used if complete
Tstam is the measured maximum ( or minimum )
descriptions of all equations used to represent a hydrologic
process in GSFLOW are not stated. Derivations of equations daily temperature at the station for time step
only are presented when they are not available in other m, in degrees Fahrenheit or Celsius;
publications or if they better explain the hydrologic processes bmonth is the monthly maximum (or minimum)
represented by the model. daily temperature lapse rate representing
the change in maximum (or minimum) air
temperature per 1,000 feet or meters of
Distribution of Temperature, Precipitation, and altittude change for each month, January to
Solar Radiation December parameter tmax_lapse
(or tmin_lapse), in degrees Fahrenheit
Model inputs are daily precipitation, maximum and or Celsius per 1,000 length units;
minimum air temperature, and, optionally, solar radiation. Air Z HRU is the mean land-surface altitude of the HRU
temperature and solar radiation are used in the computation
parameter hru_elev, in length;
of evaporation, transpiration, sublimation, and snowmelt.
Z tsta is the altitude of the air temperature
These point data are extrapolated to each HRU using a set
of adjustment coefficients developed from regional climate measurement station parameter
data. The coefficients typically include the effects of altitude, tsta_elev, in length; and
slope, aspect, and distance to one or more measurement sites taf HRU is the maximum ( or minimum ) daily HRU
(Linsley and others, 1975). temperature adjustment factor parameter
tmax_adj (or tmin_adj), which is
estimated on the basis of slope and aspect,
Temperature in degrees Fahrenheit or Celsius.
Maximum and minimum daily air temperatures are
extrapolated to each HRU by one of four user-specified The second option (module temp_laps_prms) is used
options that depend on the number of air-temperature when data from at least two stations at different altitudes
measurement stations to be used in estimating the air are used to estimate air temperature of each HRU. Each
temperature of each HRU. The first option is used when data HRU is assigned air temperatures on the basis of computed
from only one station are used to assign air temperature to lapse rates from a pair of stations. The station that is most
each HRU. The second through fourth options are used when representative of the temperature at the HRU is designated
air temperature assigned to each HRU is based on data from as the base station. The station that is not the base station is
more than one station. Temperature units can be specified as called the lapse station and is used with the base station to
either degrees Fahrenheit or Celsius, and the altitude of the calculate the lapse rate. The two stations associated with each
measurement station to which the temperature corresponds can HRU are specified by PRMS parameters hru_tsta and
be specified as either feet or meters. Lapse rate coefficients hru_ tlaps (appendix 1). Maximum and minimum daily
that are input to the model to represent changes in temperature HRU temperatures are computed by a temperature lapse rate
with land-surface altitude must be expressed in units that are from the two associated stations according to:
consistent with the specified temperature and altitude units.
The first option (module temp_1sta_prms) is used when
data from one or more measurement stations are available.
The station associated with each HRU is specified by PRMS
parameter hru_tsta (appendix 1). The maximum and
minimum daily HRU temperatures are computed from the
measured station temperatures and a monthly lapse rate
according to:

+ (Tlapse ) ZZHRU ZZbase ( )
m
m
THRU
m
= Tbase m m
Tbase taf HRU , (2) cT 2 = T stas bt X , month X sta + btY , month Y sta + btZ , month Z sta , (3c)

lapse base
wheere
where bt X , month , btY , month , and btZ , month are the maximum
m
Tbase is the measured maximum ( or minim mum ) daily
( or minimum ) air temperature regression
temperature at the base station assigned to
coefficients for longitude, latitude, and
an HRU for time step m, in degrees Fahrenheit
altitude, respectively by month , starting
or Celsius;
with January parameter max_lapse
Tlampse is the measured maximum ( or minimum ) (or min_lapse), in degrees Fahrenheit
daily temperature at the lapse station assigned or Celsius;
to an HRU during time stepp m, in degrees
X HRU is the normalized longitude of the HRU,
Fahrenheit or Celsius;
dimensionless;
Z lapse is the altitude of the lapse station, in length; and
Y HRU is the normalized latitude of the HRU,
Z base is the altitude of the base station , in length.
dimensionless;
Z HRU is the normalized altitude of the HRU,
The third option (module xyz_dist) uses a three- dimensionless;
m
dimensional multiple-linear regression based on longitude T stas is the mean measured maximum or
(x), latitude (y), and altitude (z) to distribute temperature from ( minimum ) daily temperature of all
two or more stations (Hay and others, 2000; Hay and Clark, stations for time step m, in degrees
2003). The independent variables used in the regression are Fahrenheit or Celsius; and
normalized by subtracting the mean and dividing by the
X sta , Y sta , and Z sta are the mean normalized,
standard deviation to remove the effects of units, magnitude,
longitude, latitude, and altitude of all
and inconsistency in specification of the origin. The resulting
equation is: stations respectively, dimensionless.
m
THRU = cT 1 + cT 2 taf HRU , (3a)

cT 1 = bt X , month X HRU + btY , month Y HRU + btZ , month Z HRU , and (3b)
( ) , (5b)
The fourth option (module temp_dist2_prms) weights
(YHRU Ytempi ) + ( X HRU Xtempi )
2 2
measured daily air temperatures from two or more stations by Ldist HRU ,i = 1 /
the inverse of the square of the distance between the centroid
of an HRU and each station location (Dean and Snyder, 1977;
Bauer and Vaccaro, 1987; Vaccaro, 2007). Extrapolated values where
also are adjusted by daily maximum and minimum lapse rates Ldisti iss the inverse distance between the HRU
that are calculated using the daily values from all measurement centroid and measurrement station i,
stations. Air temperature data and adjustment parameters in length;
must be in the same units. Daily basin-average lapse rates for X HRU is the longitude of the HRU centroid
maximum and minimum temperature are first computed by: parameter hru_xlong, in length;
Xtempi is the longitude of each air temperature
(( )
ntstas 1
m
blapse = i =1
)
X i2 Ti m Ti+m1 / ( Z tsta ,i Z tsta ,i +1 ) / ntstas, (4) measurement station parameter
tsta_xlong, in length;
YHRU is the latitude of the HRU centroid-
where parameter hru_ylat, in length; and
m
blapse is the daily basin-average maximum Ytempi is the latitude of each air temperature
measurement station parameter
( or miinimum ) temperature lapse tsta_ylat, in length.
rate for time step m, in degrees
Fahrenheit or Celsius per length;
Ti m and Ti+1m are the measured maximum Precipitation
( or minimum ) temperature at Precipitation, which can be measured at one or more
measurement stations i and i +1 stations, is extrapolated to each HRU by one of four user-
for time step m, in degrees specified options that depend on the number of stations
Fahrenheit or Celsius; available. The altitude of the precipitation measurements can
Z tsta ,i and Z tsta ,i +1 are the altitudes of measurement be specified in units of either feet or meters.
stations i and i +1 parameter The form of the precipitation (rain, snow, or a mixture
tsta_elev, in length; and of both) is important to the simulation of snow accumulation,
ntstas is the number of air temperature snowmelt, infiltration, and runoff. Precipitation form on
measurement stations, each HRU can be specified or it can be estimated from
dimensionless. the HRU maximum and minimum daily air temperatures.
Precipitation is considered all snow on an HRU when the
maximum daily air temperature is less than or equal to
The daily maximum and minimum temperatures are
parameter tmax_allsnow. Precipitation is all rain when
computed on the basis of data from two or more stations
either the minimum daily air temperature is greater than or
according to:
equal to parameter tmax_allsnow, or the maximum
daily air temperature is greater than or equal to parameter
ntstas m Z HRU Z i 2
T ( Ldist HRU ,i )
m
i blapse
m i =1 1, 000
THRU = ntstas
taf HRU (5a)
( Ldist )
2
HRU , i
i =1
and
tmax_allrain.When neither condition for all snow nor The second option (module precip_laps_prms) also uses
rain is met, then precipitation on an HRU is considered a equation 7, except that the correction factor is calculated for
mixture, and the rain is assumed to occur first. The fraction of each HRU on the basis of computed lapse rates from two
the total precipitation occurring as rain is computed by: precipitation stations. The station that is most representative of
precipitation on an HRU is designated as the base station. The
station that is not the base station is called the lapse station
m TmxHRU
m
TmxsnowHRU
FrainHRU = m m pmixaf month , (6) and is used with the base station to calculate the lapse rate.
Tmx HRU TmnHRU The two stations assigned to each HRU are specified by PRMS
parameters hru_psta and hru_plaps (appendix 1). The
where monthly correction factor can be calculated as:
m
Frain HRU is the decimal fraction of total
precipiitation occurring as rain on an P lapse P base
HRU for time step m, dimensionless; ( Z HRU Z base )
Z lapse Z base
Tmx mHRU is the maximum air temperature assigned CFHRU = 1.0 + paf base , (8)
P base
to the HRU for time step m, in degrees

Fahrenheit or Celsius;
Tmn mHRU is the minimum air temperature assigned
to the HRU for time step m, in degrees where
Fahrenheit or Celsius; paf base is the mean monthly faactor used to
TmxsnowHRU is the monthly maximum air temperature adjust rain ( or snow ) lapse rate
at which preecipitation is all snow for ( usually 1.00 ) parameter
the HRU parameter tmax_allsnow w, padj_rn (or padj_sn),
in degrees Fahrenheit or Celsius; and dimensionless;
pmixaf month is the monthly rain adjustment factor for
P base is the mean monthly precipitation at
a mixed precipitation event the base station parameter
( usually 1.0 ) parameter pmn_mo, in inches per day; and
adjmix_rain, dimensionless.
P lapse is the mean monthly precipitation at
the lapse station parameter
The form of precipitation also may be explicitly specified pmn_mo,, in inches per day.
through use of the variable form_data in the PRMS Data Z HRU is the mean land-surface altitude of
File (appendix 1).
HRU parameter hru_elev,
One or more precipitation stations can be used to estimate
in length;
precipitation throughout the model domain. The first option
requires assigning a precipitation station to an HRU (module Z base and Z lapse are the land-surface altitudes of
precip_prms). The station assigned to an HRU is specified the base and lapse stations,,
by PRMS parameter hru_psta (appendix 1). A monthly respectively parameter
correction factor, which accounts for elevation, spatial psta_elev, in length.
variation, topography, gage location, deficiencies in gage catch
due to the effects of wind and other factors, is used to estimate
the daily precipitation at the HRU according to:
m
PHRU = Pstam CFHRU , (7)
where
m
PHRU is the precipitation at the HR
RU during time step
m, in inches;
Pstam is the measured precipitation at the station
during time step m, in inches; and
CFHRU is the monthly correction factor as a decimal
fraction used to adjust rain ( or snow ) at the
HRU parameter rain_adj (or snow_adj),
dimensionless.
( ) ,
The third option uses a three-dimensional multiple-linear
(YHRU Yppti ) + ( X HRU Xppti )
2 2
regression to distribute precipitation from two or more stations Ldist HRU ,i = 1 / (10b)
(Hay and others, 2000; Hay and Clark, 2003) and is computed
by module xyz_dist. Computations are similar to the multiple-
linear regression for the calculation of temperature distribution
given in equations 3a-3c: Rain HRU , month
CFHRU , rain = , for monthly fraction of (10c)
P sta , month
m
PHRU = cP1 + cP 2 CFxyz , (9a) precipitation that is rain, and
SnowHRU , month
CFHRU , snow = , for monthly fraction of (10d)
P sta , month
cP1 = bp X , month X HRU + bpY , month Y HRU + bpZ , month Z HRU , and (9b)
precipitation that is snow,
m
cP 2 = P stas (bp X , month X sta + bpY , month Y sta + bpZ , month Z sta ) , (9c) where
nptas is the number of precipitation stations,
dimensionless;
wheere Xppti is the longitude of each precipitation
CFxyz is the monthly correction factor as a measurement station parameter
decimal fraction used to adjust pstta_xlong, in length;
rain ( or snow ) values Yppti is the latitude of each precipitation
parameter adju ust_rain (or measurement station parameter
adjust_snow), dimensionless; psta_ylat, in length;
bp X , month , bpY , monnth , Rain HRU , month is the mean monthly rain on each HRU
and bpZ , month are the precipitation regression that can be obtained from National
coefficiients for longitude, latitude, Weather Service ' s spatial distribution
and altitude, respectively by of mean annual precipitation for the
month, starting with January 1971 2000 climate normal period
parameter ppt_lapse, in parameter rain_mon, in length;
inches; and SnowHRU , month is the mean monthly snow on each HRU
m
P stas is the mean measured precipitation of that can be obtained from National
all stations during time step m, in Weather Service ' s spatial distribution
inches. of mean annual precipitation for the
1971 2000 climate normal period
The fourth option (module precip_dist2_prms) weights parameter snow_mon, in length;
measured precipitation from two or more stations by the P sta , monthh is the mean monthly precipitation at
inverse of the square of the distance between the centroid of each measurement sttation
an HRU and each station location (Dean and Snyder, 1977, parameter psta_mon, in length;
Bauer and Vaccaro, 1987, Vaccaro, 2007). Estimates of daily CFHRU , rain is the monthly rain correction factor as
precipitation on an HRU are adjusted by the ratio of the a decimal fraction of preecipitation at
mean-monthly precipitation on an HRU to the mean monthly the measurement station ,
precipitation at each precipitation station. Precipitation data dimensionless; and
and the adjustment parameters must be in the same units. CFHRU , snow is the monthly snow correction factor as
Daily precipitation on an HRU is computed according to:
a decimal fraction of precipitation at
the measurement statiion ,
npstas dimensionless.
( CF P m + CFHRU , snow Pi m ) ( Ldist HRU ,i )
2
HRU , rain i

m
PHRU = i =1
npstas
(10a)
( Ldist )
2
HRU , i
i =1
Solar Radiation 1963). Sunset and sunrise times also are computed for this
surface. Finally, sunset time on the sloped surface of the
Tables consisting of daily estimates of the potential (clear HRU is taken as the earliest of the computed horizontal
sky) short-wave solar radiation for each HRU are computed on surface sunset times of the HRU and equivalent-slope surface.
the basis of hours between sunrise and sunset for each Julian Likewise, sunrise time is taken as the latest of the computed
day of the year in module soltab_hru_prms. The potential horizontal surface sunrise times. Daylight length for each
short-wave solar radiation also is computed for each Julian day HRU is computed from these sunset and sunrise times.
of the year for a horizontal plane at the surface of the centroid The angle between the local meridian and the sunset
of the modeled domain. The computations of the solar tables (or sunrise) meridian, referred to as the hour angle of sunset
were modified for GSFLOW to use double precision and more (or sunrise), for the horizontal surface of both the HRU and
accurate constants, such as decimal days per year, rotational equivalent-slope surface is calculated according to (Swift,
degrees per day, eccentricity of the Earths orbit, and the 1976):
number pi and constants based on pi such as radians and
radians per year. ss m = cos 1[ tan(lat ) tan( DM m )] , (13a)

Daily estimates of obliquity are computed from (Meeus,
1999):
sr m = ss m , and (13b)
E m = 1 [ EC *cos( jd 3) * rad ] , (11)
where
E m is the obliquity of the Sunn's ecliptic for time m
shHRU
ss m
HRU
m
srHRU 24 , (13c)
step m, in angular degrees; 2P
EC is the eccentricity of the Earth's orbit
( ~ 0.01671) , in radians; where
jd is the Julian day number (3 is subtracted as the ss m is the hour angle of sunset, measured from the
solar year begins on December 29), in days; local meridian of a horizontal surface (HRU
and or equivalent-slope surface) for time step m,
rad is the revolution speed of the Earth ( ~ 0.0172 ) , in radians;
in radians per day. sr m is the hour angle of sunrise, measured from
the local meridian of a horizontal surface
Daily estimates of solar declination are computed from (HRU or equivalentt-slope surface) for time
(Meeus, 1999): stem m, in radians;
m
ssHRU is the hour angle of sunset on the sloped
DM m 0.006918 0.399912 *cos Ert 0.070257 surface of the HRU for time step m,
in radians;
*sin Ert 0.006758 *ccos Ert m
srHRU is the hour angle of sunrise on the sloped
0.000907 *sin 2 * Ert 0.002697 surface of the HRU for time step m,
*cos 3 * Ert 0.00148 * sin 3 * Ert , (12) in radians;
m
shHRU is the daylight length on the HRU for time
step m, in hours;
where
P is the constant pi ~ 3.1415926535898 ,
DM m is the solar declination for time step m,
dimensionless; and
in angular degrees; and
lat is the latitude of the horizontal surface (basin
Ert = rad
( jd 1).
centroid } parameter basin_lat, HRU
centroid } parameter hru_lat or
Sunset and sunrise times are computed for each HRU
equivalent-slope surface), positive values
with a three-step procedure. First, the sunset and sunrise times
are computed for a horizontal plane at the centroid of the are in the northern hemisphere and negative
HRU. Next, a horizontal surface on the terrestrial spheroid is values are in the southern hemisphere,
found which is parallel to the slope and aspect of the surface in radians.
of the HRU. This is called the equivalent-slope surface (Lee,
Daily estimates of potential solar radiation for each HRU are calculated as described in Frank and Lee (1966), and Swift
(1976):
m (14a)
RspHRU = sc m (c1PSR + c 2 PSR ) ,
) shHRU
c1PSR = sin( DM m ) sin(lat HRU m
, and (14b)
) sin( ssHRU
cos( DM m ) cos(lat HRU m
) sin( srHRU
+ long HRU m
) 24
+ long HRU (14c)
c 2 PSR = ,
2
where
m
RspHRU is the potential solar radiiation on the HRU during time step m, in calories
per square centimeter per day;
is the latitude of the equivalent-slope surface of the HRU, in radians;
lat HRU
is the longitude offset between the equivalent-slope surface annd the HRU,
long HRU
in radians; and
sc m is the 60-minute period solar constant for time step m, in calories per square
centimeter per hour.
Computed daily shortwave radiation data for each m m m

RahHRU = apHRU RspHRU , and (15a)
HRU is estimated using one of two methods. The first is a
modification of the degree-day method described by Leaf and
Brink (1973) and is computed by module ddsolrad_hru_prms. m
This method was developed for the Rocky Mountain region m RahHRU
RaswHRU = 1
pptadjHRU , (15b)
of the United States. It is most applicable to this region cos(tan ( slopeHRU ))
where predominantly clear skies prevail on days without
precipitation. The computed shortwave radiation is calculated
where
according to: m
RahHRU is the measured or computed as the
horizontal plane shorttwave radiation
on the HRU during time step m, in
calories per square centimeter per day;
m
apHRU is the degree-day based ratio of actual
to potential shortwave radiation for
the HRU during time step m,
dimensionless;
m
RaswHRU is the computed shortwave radiation on
the HRU during time steep m, in
calories per square centimeter per day;
slopeHRU is the slope of the HRU parameter
hru_slope, dimensionless; and
pptadjHRU is the precipitation-day adjustment factor
to solar radiation parameters
radj_wppt ( for a winter day ) and
radj_sppt ( for a summer day ) ,
dimensionless.
The second procedure uses a relation between sky cover 2

and daily range in air temperature demonstrated by Tangborn sh m
m
PETHRU = HCHRU HRU HRU , and (18)
(1978), and a relation between solar radiation and sky cover 12
developed by Thompson (1976) and is computed by module
ccsolrad_hru_prms. This procedure is applicable to more
humid regions where extensive periods of cloud cover occur THRU
with and without precipitation. Shortwave radiation applicable 6.108e17.26939
THRU + 273.3 (19)
for humid regions is calculated as: HRU = 216.7 ,
THRU + 273.3

m
CCHRU = crnmonth (TmxHRU
m m
TmnHRU ) + crbmonth , (16a)
where
PETHmRU is the potential evapotranspiration for the
HRU duringg time step m, in inches;
RahHRU = B + (1.0 B )(1.0 CCHRU ) RspHRU , and (16b)
m m cre m
HCHRU is the Hamon monthly air-temperature

coefficient parameter
m
RahHRU hamon_coef, in inch-cubic meter
m
RaswHRU = , (17) per gram;
cos(arctan( slopeHRU ))
HRU is the saturated water-vapor density
( absolute humidity ) , in grams per
where cubic meter;
B is the constannt used in the cloud cover to e is the exponential function constant
solar radiation relation , a value can be ( ~ 2.7182818) , dimensionless; and
obtained from Thompson (1976, fig. 1)
THRU is the mean daily temperature on the HRU,
parameter crad_coef, dimensionless; in degrees Celsius.
m
CCHRU is the decimal fraction of cloud cover on the
HRU during time step m, dimensionless;
cre is the exponent used in the cloud cover to The second option (module potet_jh_prms) is the
solar radiation relation , a value of 0.61 modified Jensen-Haise formulation (Jensen and others,
is suggested by Thompson (1976 ) 1969), in which potential evapotranspiration is computed
as a function of air temperature, solar radiation, and
parameter crad_exp, dimensionless;
two coefficients that can be estimated using regional air
crnmonth is the slope in the regression equation that
temperature, altitude, vapor pressure, and plant cover
relates cloud coveer to daily minimum and
according to:
maximum air temperature by month,
starting with January parameter m
RaswHRU
ccov_slope, dimensionless; and
m
PETHRU = JH month (THRU JH HRU ) , and (20a)
crbmonth is the intercept in the regression equation 2.54 HRU
that relates cloud cover to daily minimum
and maximum air temperature by month , HRU = 597.3 (0.5653THRU ) , (20b)
starting with January parameter
ccov_intcp, dimensionless.
where
JH month is the monthly Jensen--Haise air-
temperature coefficient parameter
Potential Evapotranspiration jh_coef, in degrees Fahrenheit;
Potential evapotranspiration is computed for each 2.54 is the conversion from inch to centimeters;
HRU by one of three user-specified options. The first option JH HRU is the Jensen-Haise air-temperature
(module potet_hamon_hru_prms) is the empirical Hamon coefficient for the HRU parameter
formulation, in which PET is computed as a function of daily jh_coef_hru, in degrees Fahrenheit;
mean air temperature and possible hours of sunshine according and
to (Hamon, 1961; Murray, 1967; and Federer and Lash, 1978): HRU is the latent heat of vaporization on the
HRU for time step m, in calories per
gram.
The third option (module potet_pan_prms) is used m m

SpcaHRU = ( SpcmxHRU SpcHRU )( AHRU 'HRU )
when pan evaporation data from one or more measurement
m m
stations are available. The station associated with each HRU m m SpcaHR m SpcaHRU
Ptf HRU = PHRU U
when PHRU >
is specified by PRMS parameter hru_pansta (appendix 1). AHRU 'HRU AHRU 'HRU
Potential evapotranspiration is computed from the measured m
m m SpcaHRU
pan evaporation and a monthly coefficient according to: Ptf HRU = 0.0 when PHRU , (22)
AHRU 'HRU
m m
PETHRU = Pancoef month Panevapsta , (21)
where
m
SpcaHRU is the available storage in the plant
where
canopy of the HRU during time
Pancoef month is the moonthly pan-evaporation
stepp m, in acre-inch;
coefficient parameter
SpcmxHRU is the maximum storage in the plant
epan_coef, dimensionless; and
m canopy for snow , summer rain,
Panevapsta is the pan evaporation for the
and winter rain on each HRU
corresponding measurement station
parameter snow_intcp ( snow ) ,
for time step m, in length.
srain_intcp ( summer rain ) ,
wrain_intcp ( winter rain ) ,
Canopy Interception in inches;
m
SpcHRU is the storage in the plant canopy
Interception of precipitation by the plant canopy is (summer or winter) on the HRU
computed during a time step as a function of plant-cover during time step m, in acre-inch;
density and the storage available on the predominant plant- AHRU is the area of the HRU parameter
cover type in each HRU (Leavesley and others, 1983) using hru_area, in acres;
module intcp_prms. Throughfall precipitation, which is
'HRU is the plant canopy density as a
precipitation that is not intercepted by the plant canopy, is
decimal fraction of the HRU area
computed as:
parameter covden_sum ( summer )
covden_win ( winter ) ,
dimensionless; and
m
Ptf HRU is the precipitation throughfall on the
HRU during time step m, in inches.
The precipitation that reaches the ground during time

step m is referred to as net precipitation, and is the sum of
throughfall and precipitation on the HRU not covered by
plants. Net precipitation is calculated according to:
m m m
Pnet HRU = PHRU (1.0 'HRU ) + ( Ptf HRU 'HRU ) , (23)
where
m
Pnet HRU is the precipitaiton that reaches the
ground during time steep m, in inches.
Plant-cover density varies by season and type. The types The snowpack energy balance is computed for the day
of plant cover that can be specified are bare (no cover), grass, over two 12-hour intervals (designated day and night) and
shrubs, and trees. Intercepted rain is assumed to evaporate aggregated to a daily time step. Heat transfer between the
at a free-water surface rate. Intercepted snow is assumed to surface layer and the lower layer of the snowpack occurs by
sublimate at a rate that is expressed as a decimal fraction of conduction when the temperature of the surface layer is less
the potential evapotranspiration. than freezing (0 Celsius), as discussed by Leavesley and
others (1983, p. 39-41), and calculated as:
Snowpack m
keHRU tsnow
PRMS simulates the initiation, accumulation, and
m
Ht HRU m
= 2snowHRU Cice m
snowHRU Cice
(Tsurfm Tlowm er ) , (25)
depletion of a snowpack on each HRU (fig. 3; Leavesley
and others, 1983; Leavesley and others, 2005) using module
snowcomp_prms. A snowpack is modeled as a two-layered where
m
system that is maintained and modified on both a water Ht HRU is the daily heat transferred from the
equivalent basis and as a dynamic heat reservoir, as depicted in surface layer to the lower layer of
figure 3 (Obled and Rosse, 1977). The surface layer consists of the snowpack during time step m, in
the upper 35 cm of the snowpack, and the lower layer is the calories per square centimeter;
remaining snowpack. Only equations that describe the water m
snowHRU is the density of the snow pack, in
and energy balance of the snowpack are described herein. grams per cubic centimeter;
Refer to Leavesley and others (1983) for a comprehensive Cice is the specific heat of ice , in calories per
discussion of the PRMS snowpack calculations. gram-degree Celsius;
A snowpack develops when precipitation occurs as snow. m
keHRU is the effective thermal conductivity of
If a snowpack already exists, the precipitation received during
that time step is added to the snowpack and the new snowpack the snowpack , in calories per
depth is calculated. Snowpack depth is calculated by solving second-gram-degree Celsius;
the following ordinary differential equation with a finite- tsnow is the snow computation time step, in
difference approximation of the time derivative (Riley and 43, 200 seconds ( half-day interval ) ;
m
others, 1973): Tsurf is the temperature of the snowpack
surface layer during time step m, in
m
dDsHRU m Ps m degrees Celsius; and
+ csHRU DsHRU = HRU m
dt s Tlower is the temperature of the snowpack
cs lower layer during time step m, in
m
+ HRU DseHRU , (24) degrees Celsius.
sm
wheree
m
DsHRU is the precipitaiton that reaches the ground The effective thermal conductivity of the snowpack
during time step m, in inches. m
( keHRU ) is estimated by (Anderson, 1968, p. 22):
csHRU is the snowpack settlement-time constant
parameter settle_const, in per = 0.0077 ( HRU ) .
m m 2
keHRU (26)
day;
m
PsHRU is the net snowfall rate, as a liquid water The volumetric density fraction of the snowpack used in
equivalent , on the HRU during time equation 26 is computed as (Leavesley and other, 1983, p. 41):
step m, in inches per day;
s is the density of new-fallen snow , as m
DseHRU
a decimal fraction parameter HRU
m
= m
, (27)
DsHRU
den_init, dimensionless;
sm is the average maximum snowpack
density, as a decimal fraction of the where
liquid water equivalent parameter HRU
m
is the volumetric densitty fraction of the
den_max, dimensionless; and snowpack for the HRU during time
m
DseHRU is the snowpack liquid water equivalent step m, dimensionless.
depth for the HRU at time step m, in
inches.
The computation of the energy components available Further details of variables used in equation 28,
to melt snow when both layers of the snowpack reach assumptions made for the PRMS formulation, and calculation
isothermal conditions at 0o Celsius are described by Male and are presented by Leavesley and others (1983, p. 42-45).
Gray (1981); the energy available to melt snow is calculated The snowmelt on an HRU is computed by:
according to:
m
HmHRU
m m m m m
m
VsmHRU = m
AscHRU , (29)
Hm HRU = Hs HRU + Hl HRU + Hc HRU + He
HRU HF
m m m
(28)
+ Hg HRU + HpHRU + HqHRU ,
where
m
where VsmHRU is the volume of snowppack melted during
m
HmHRU is the energy available for snowm melt during time step m, in acre-inch;
time step m, in calories; HF is the specific latent heat of fusion to melt
m one inch of water-equuivalent ice at
HsHRU is the energy gained due to shortwave
0 Celsius, in 203.2 calories per inch; and
radiation during time step m, in calories; m
m AscHRU is the snow-covered area of the HRU
HlHRU is the energy gained due to longwave
determined by snow-cover areal
radiation during time step m, in calories;
m depletion curve (Anderson, 1973),
HcHRU is the convective or sensible heat at air-
in acres.
snow interface during time step m, in
calories;
m
HeHRU is the latent heat (sublimation and Sublimation from the snow surface and evaporation
condensation) at the air-snow interface from snow on the plant canopy are assumed to occur only
during time step m, in calories; when there is no transpiration from plants. The daily loss
m
Hg HRU is the heat gained from the ground during from the snowpack is computed as a decimal fraction of
time step m, in calories; the potential evapotranspiration defined by input parameter
m potet_sublim (see appendix 1):
HpHRU is the heat gained from precipitation during
time step m, in calories; and
m
m
subHRU m
= dfsubHRU PETHRU m
qceHRU HRU
m
, (30)
HqHRU is the heat required for internal state
change during time step m, in calories.
where
m
subHRU is the sublimation from the HRU during
time step m, in inches;
dfsubHRU is the decimal fraction of potential
evapotranspiiration that is sublimated
from snow surface parameter
pot tet_sublim, dimensionless; and
m
qceHRU is the evaporation loss from interception
storage for the HRU during time steep m,
in inches.
Impervious Storage, Hortonian Runoff, and Evaporation from impervious parts of HRUs is computed
for each time step by;
Infiltration
m m m
Snowmelt and net precipitation that reach the soil C1imper = PETHRU subHRU
surface during a time step is partitioned to the pervious
m
and impervious parts of each HRU. Surface runoff due to m m 1 m VsmHRU m
C 2imper = Dimper + Pnet HRU + + ROhupHRU
infiltration excess, hereafter referred to as Hortonian runoff, AHRU
and infiltration are computed on the pervious parts of each m
ROhimper
HRU; whereas storage, evaporation, and Hortonian runoff are
computed on the impervious parts of each HRU. AscHRU
m

m
Evapimper m
= C 2imper 1-
m m
When C1imper C 2imper (32)
AHRU
Impervious Storage and Evaporation
m m AscHRU
m
m m
If rain throughfall and snowmelt satisfy available Evapimper = C1imper 1- When C1imper < C 2imper ,
AHRU
retention storage on the impervious parts of the HRU,
Hortonian runoff is generated. Hortonian runoff from
impervious parts of each HRU is calculated from continuity where
according to: m
Evapim per is the evaporation from the impervious
U for time step m,
part of the HRU
m m 1 m
Cimper = Dimper Dimximper + Pnet HRU in inches.
m
VsmHRU m Storage on the impervious parts of the HRU is calculated
+ + ROhupHRU
AHRU according to:
(31)
m m m
ROh imper =C
imper When C imper >0 m
VsmHRU
m m 1 m
Dimper = Dimper + Pnet HRU +
m m AHRU (33)
ROhimper = 0 When Cimper 0,
m m m
where + ROhupHRU -ROhimper -Evapimper .
m
Cimper is the water available for Hortonian
runoff from the impervious part per
unit areea of the HRU during time
step m, in inches;
m
ROhimper is the Hortonian runoff from the
impervious part of the HRU per unit
area during time step m, in inches;
m 1
Dimper is the impervious storage, as calculated
by equation 33, for the last iteration
of time step m 1, as volume per unit
area for the HRU, in inches;
Dimximper is the maximum retention storage for
HRU impervious area, in inches; and
m
ROhup
pHRU is the sum of Hortonian runoff from all
upslope contrributing HRUs as a
volume per unit area of the HRU for
time step m, in inches.
Hortonian Runoff and Infiltration

Hortonian runoff for the pervious part of each HRU is computed using a contributing-area concept (Dickinson and
Whiteley, 1970; Hewlett and Nutter, 1970). Hortonian runoff from pervious parts of a HRU is related to the area where the
throughfall and snowmelt exceed the soil infiltration rate. This is represented by either a linear (module srunoff_carea_casc)
or nonlinear function (module srunoff_smidx_casc) of antecedent soil-moisture content. The linear form of computing the
contributing area for pervious runoff can be written as:
DupHRU
m

m
FpervHRU = FmnHRU + ( FmxHRU FmnHRU ) , (34a)
DupmxHRU
whereas the nonlinear form can be written as:
m m 1 m
SmidxHRU = DCPR + 0.5 Pnet HRU
( Smex m
HRU SmidxHRU )
C 3mHRU = SmcHRU *10 (34b)

m
FpervHRU = C 3mHRU when C 3mHRU FmxHRU
m
FpervHRU = FmxHRU when C 3mHRU > FmxHRU ,
where
m
FpervHRU is the surface-runoff-contributing area of the pervious parts in the HRU for
time step m, as a decimal fraction of HRU area , dimensionless;
FmnHRU is the minimum possible area contributing to surface runoff , as a decimal
fraction of HRU area parameter carea_min, dimensionless;
FmxHRU is the maximum possible area contributing to surface runoff , as a
decimal fraction of HRU area parameter carea_max, dimensionless;
m
DupHR
RU is the antecedent volume per unit area of water in capiillary reservoir that is
available for evaporation during time step m, in inches;
DupmxHRU is the maximum quantity of water in the capillary reservoir parameter
soil_rechr_ma ax, in inches;
m
SmidxHRU is the soil moisture index of the capillary reservoir for time step m, in inches;
m 1
DCPR is the volume per unit area of water in the capillary reservvoir at the last
iteration of time step m-1, in inches;
SmcHRU is a coefficient used to calculate decimal fraction of pervious surfaces
parameter smidx_coef, dimensionless; and
SmexHRU is an exponent used to calculate the decimal fraction of pervious surfaces
parameter smidx_exp, in per inch.
The runoff from the pervious part of an HRU is m

VsmHRU
calculated as: C 4mHRU = ROhup mperv + ,
AHRU
m
ROhperv m
= FpervHRU ( ROhupHRU
m m
+ Pnet HRU ) , (35) m
qsi perv = ( C 4mHRU + Pnet HRU
m
ROhpmerv ) Aperv (37)
when C 4mHRU < qsnmxHRU
where
m
ROhperv is thee runoff per unit area from the
m
qsi perv = ( qsnmxHRU + Pnet HR
m
U )
pervious part of the HRU for time when C 4mHRU qsnmxHRU ,

step m, in inches.
where
m
Hortonian runoff from an HRU is routed to one or more qsi perv is the soil infiltration over the pervious
downslope HRUs and(or) to one or more stream segments part of the HRU for time step m,
with the cascading-flow procedure (see section Cascading- in acre-inch;
Flow Procedure) and connections defined by parameters qsnmxHRU is the daily maximum snowmelt
input to the PRMS Cascade Module (cascade_prms, table 8). infiltration for the HRU parameter
The fraction of runoff from an upslope HRU that flows to a snowinfil_max, in inches; and
downslope HRU or stream segment is calculated as: Aperv is the pervious area of the HRU,
in acres.
, dwn = ( ROh perv , up + ROhimpr , up )
m m m
ROhHRU
AHRU ,up (36)
FcontribHRU ,up , Soil Zone
AHRU , dwn
The soil zone is an important link in the integration of
PRMS with MODFLOW-2005 and its function was changed
where
m considerably from the original PRMS module for use in
ROhHRU , dwn is volume per unit are ea of Hortonian GSFLOW. The new Soil-Zone Module (soilzone_gsflow,
runoff to a downslope HRU or table1) is used to calculate storage and all inflows and
stream segmeent during time step m, outflows in the soil zone for each HRU. The soil zone is
in inches; conceptualized as three reservoirs: the capillary, gravity, and
m
ROhperv ,up is the runoff per unit area from the preferential-flow reservoirs (fig. 19). These reservoirs occupy
pervious part of the upslope HRU the same physical space and represent different soil-water
for time step m, in inches; processes at different soil-water content thresholds. The
m
ROhimpr capillary reservoir is limited by the field-capacity threshold.
, up is the runoff per unit area from the
The soil-water content between field capacity and soil
impervious part of the upslope HRU
saturation defines the gravity reservoir. The preferential-
for time step m, in inches;
flow reservoir is that part of the gravity reservoir from which
FcontribHRU ,up is the decimal fraction of area in the
fast interflow occurs and is defined by the preferential-flow
upslope HRU that contributes threshold.
Hortoniaan runoff to the downslope The only outflows from the capillary reservoir are to
HRU parameter hru_pct_up, evapotranspiration and gravity reservoirs. Outflow from the
dimensionless; preferential-flow reservoir can be fast interflow and saturation
AHRU ,up is area of the downslope HRU excess surface runoff, hereafter referred to as Dunnian runoff.
parameter hru_area, in acres; and Outflow from the gravity reservoir can be slow interflow,
AHRU , dwn is area of the downslope HRU flow to the preferential-flow reservoir, replenishment of the
parameter hru_area, in acres. capillary reservoir, gravity drainage, and Dunnian runoff. Flow
in the soil zone is dependent on the simultaneous solution of
flow in the soil zone and ground-water heads. Storage and all
Infiltration occurs on the pervious areas of each HRU
inflows and outflows in the soil zone are calculated according
and includes Hortonian runoff from upslope HRUs, snowmelt,
to the sequence of steps illustrated in figure 19 and listed in
and rain throughfall. Hortonian runoff from the HRU is
table 9.
subtracted from the available water for infiltration. Infiltration
is calculated as:
Fraction of From Dunnian runoff Downslope

Infiltration GVR when filled fast interflow
1 6 12 13
Saturation threshold
Porous soil
Increasing storage
Macropore flow
Preferential-flow reservoir (PFR) Preferential-flow threshold

Downslope To unsaturated Dunnian runoff
From Ground-water To To slow or saturated when pref_flow_den is 0.0
CPR discharge CPR PFR interflow zone
12
3 4 5 6 7 8
Preferential-flow threshold
Air filled pore space

Increasing storage
Capillary water
Sand grain
Soil-zone
Gravitational water
water
Gravity reservoir (GVR) Field-capacity threshold
Fraction of Upslope Dunnian To Evapotration

Infiltration runoff and interflow GVR and transpiration
1 2 3 10 11
Field-capacity threshold
Air filled pore space

Increasing storage
Capillary water
Sand grain
9 Evaporation threshold
Capillary reservoir (CPR) Wilting threshold
EXPLANATION
8 Computational sequence listed in table 9
Figure 19. Computational sequence used in computing flow to and from a

soil zone in a hydrologic response unit. Water in the soil zone consists of three
reservoirscapillary, gravity, and preferential-flowthat represent pore-space
volumes for a given volume of soil.
Table 9. Sequence of steps used in the computation of flow into and out of the soil zone used in GSFLOW.
[Input parameters are specified for each hydrologic response unit]
Sequence
Description of flow into and out of soil zone
No.
1 Partition infiltration of precipitation, snowmelt, and Hortonian runoff (Horton, 1933) between capillary and preferential-flow
reservoirs on basis of the input parameter pref_flow_den.
2 Add upslope Dunnian runoff (Dunne and Black, 1970) and interflow to capillary reservoir.
3 Add excess water above field capacity in capillary reservoirs to gravity reservoirs on basis of input parameter soil_moist_max.
4 Add ground-water discharge from MODFLOW-2005 to gravity reservoir.
5 Replenish capillary reservoirs from gravity reservoirs if water is available and water in capillary reservoirs is below field-capacity
threshold. A fraction of the excess water can be given directly to the associated ground-water reservoir for PRMS-only
simulations when the input parameter soil2gw_max is greater than zero.
6 Add a fraction of water in gravity reservoirs to preferential-flow reservoirs on basis of input parameters pref_flow_den and
sat_threshold. Parameter sat_threshold is the volume of water per unit area between field capacity and saturation
thresholds. The value can be greater than the available porosity above field-capacity to account for surface-depression storage
on land surface.
7 Compute slow interflow from gravity reservoirs on basis of input parameters slowcoef_lin and slowcoef_sq and the volume
of water per unit area stored in the reservoir.
8 Compute gravity drainage from gravity reservoirs to unsaturated zone and/or saturated zone in MODFLOW-2005 for integrated
simulations or to ground-water reservoirs for PRMS-only simulations on basis of input parameters ssr2gw_rate,
ssr2gw_exp, and ssrmax_coef and the volume of water per unit area stored in the reservoir.
9 Partition capillary reservoirs into two zones on basis of input parameter soil_rechr_max: upper zone stores water available for
evaporation and transpiration and lower zone stores water available for transpiration only.
10 Compute transpiration from lower zone of capillary reservoir.
11 Compute evaporation and transpiration from upper zone of capillary reservoirs on basis of input parameters soil_moist_max,
soil_rechr_max, and soil_type.
12 Compute Dunnian runoff (Dunne and Black, 1970) from gravity reservoirs when parameter pref_flow_den is 0.0, that is, when
the preferential-threshold equals the saturation threshold, or from preferential-flow reservoirs when the volume per unit area in
the reservoirs exceeds the saturation threshold defined by input parameter sat_threshold.
13 Compute fast interflow from preferential-flow reservoirs on basis of input parameters fastcoef_lin and fastcoef_sq and the
volume of water per unit area stored in the reservoir.
Infiltration to Preferential-Flow and Capillary m

qsi perv
Reservoirs
m
DremCPR = (1 FdenHRU ) , (38b)
Aperv
m
A fraction of the infiltration (DhruPFR ) is apportioned to
the preferential-flow reservoir to account for fast interflow where
through large openings in the soil zone near land surface, m
m
Dhru PFR is the voluume of water per unit area
whereas the remainder of the infiltration ( DremCPR ) is added to
added to all connected prefereential-
the capillary reservoir:
flow reservoirs for time step m, in
m
inches;
FdenHRU qsi perv m

m
DhruPFR = , and (38a) DremCPR is the volume of water per unit area of
Aperv the pervious partt of the HRU that
infiltrates into the capillary reservoirr
at time step m, in inches; and
FdenHRU is the decimal fraction of the soil zone
available for preferential flow
parameter pref_flow_den,
dimensionless.
All inflows to the capillary reservoir are added to storage

Qgw
m,n
Cmf 2 prms t
in the capillary reservoir prior to calculating outflows (table9; m,n
DGVR m,n
= DexcessCPR + m 1
+ DGVR , (40)
sequence Nos. 1 and 2). Volume per unit area in a capillary A
fdc
reservoir is computed as:
m
where
DremCPR Aperv m,n
D m,n
= + ROdup m,n DGVR is the volume of water per unit area in the
CPR HRU
AHRU gravity reservoir at time step m and
(39)
+ Dslup m,n
HRU
m,n
+ DfupHRU m 1
+ DCPR , iteration n, in inches;
m,n
DexcessCPR is the volume of water per unit area of
excess water in the capillary reservoir
where
m,n for time step m and iteration n, in inches;
DCPR is the volume of water inn the capillary m,n
Qgw is the volumetric-flow rate of water that
reservoir at time step m and iteration
n, in inches; discharges from ground water to the
m1 soiil zone at time step m, iteration n, in
DCPR is the volume of water in the capillary
cubic length per time;
reservoir at time step m-1, in inches;
m,n Cmf 2 prms is the conversion from MODFLOW-2005
ROdupHRU is the volume of Dunnian runoff per unit
length per time to PRMS inches per day
area into the capillary reservoir from all
contributing HRUs at time sttep m and ( table 7 ) ;
iteration n, in inches; t is the GSFLOW time step, in one day;
m,n Afdc is the top area of the finite-difference cell,
DslupHRU is the volume of slow interflow per unit
area into the capillary reservvoir from all in length squared; and
m1
contributing HRUs at time step m and DGVR is the volume of water per unit area in the
iteration n, in inches; and gravity reservoir at tiime step m-1, in
m,n
DfupHRU is the volume of fast interflow per unit inches.
area into the capillary reservoir from alll
contributing HRUs at time step m and
iteration n, in inches.
Inflow to Gravity Reservoirs

Water in excess of the field-capacity threshold in a
capillary reservoir is added to connected gravity reservoirs
(table 9; sequence No. 3). Ground-water discharge from the
connected finite-difference cell also is added to the gravity
reservoir (table 9, sequence No. 4). The volume per unit area
in the gravity reservoir is calculated as:
Ground-water discharge from the saturated zone Total volume per unit area from the finite-difference cell
to connected gravity reservoirs is computed when to the connected gravity reservoir is computed as:
celtop 0.5 Dusz < h mfdc, n 1 as:
m , n**
DGVR m , n*
= DGVR + Dmfm , n2GVR , (42)
Q m,n
gw = CNDsz ( h m,n
fdc ( celtop 0.5 Dusz ) ) , (41)
where
m , n**
where DGVR is the revisedd volume per unit area in a
celtop is the top altitude of the finite-difference gravity reservoir after gravitty drainage
cell, in length; and before Dunnian runoff for time
Duszz is the depth of undulations at soil-zone step m and iteration n, in inches;
base, in length; Dmfm , n2GVR is the volume per unit area from the
m , n1
h fdc is the ground-water head in the finite- finite-difference cell to a connected
differencce cell for time step m, iteration gravvity reservoir for time step m and
n-1, in length; iteration n, in inches; and
m,n
h fdc is the ground-water head in the finite- m , n*
DGVR is the volume per unit area in a gravity
difference cell forr time step m and reservoir after slow interflow and
iteration n, in length; before gravity drainage foor time step m
CNDsz is the conductance across the soil-zone and iteration n, in inches.
base equal to
K v Afdc The difference between potential and net gravity drainage
0.5celthkDusz
( hmfdc,n 1 celtop + 0.5Dusz ) , is returned to the soil zone. The returned water is redistributed
in length squared per time;; back to each gravity reservoir as:
K v is the vertical hydraulic conductivity of the
finite-diifference cell, in length per
m , n**
DGVR m , n*
= DGVR m,n
+ Drej , (43)
time; and
celthk is the thickness of the finite-difference
where
cell, in length. m,n
Drej is gravity drainagee volume per unit area
that is rejected by the finite-diffeerence
Ground-water discharge across the soil-zone base is cell, in inches.
dependent on the ground-water head during iteration n of time
step m, as described in equation 41. However, the conductance
(CNDsz ) also is dependent on ground-water head, and is
linearized by using the ground-water head from iteration n-1.
Table 10. Variables used to transfer water from reservoirs in PRMS to finite-difference cells and stream reaches in MODFLOW-2005
used in GSFLOW.
[HRU, hydrologic response unit; ngwcell, number of finite-difference cells; nhru, number of HRUs; nhrucell, product of HRUs and finite-difference
cells; and nreach, number of stream reaches]
Variable name Array dimension Description

Transfer of gravity drainage from gravity reservoirs to finite-difference cells
basin_szreject 1 Watershed area-weighted average gravity drainage not added to connected
finite-difference cells.
cell_drain_rate ngwcell Gravity drainage from gravity reservoirs added to connected finite-difference cells.
gw_rejected nhru HRU area-weighted gravity drainage from gravity reservoirs not added to connected
gw_rejected_grav nhrucell Gravity drainage from gravity reservoirs not added to connected finite-difference cells.
Transfer of surface runoff and interflows from gravity and preferential-flow reservoirs to stream reaches
basin_reach_latflow 1 Watershed area-weighted average lateral inflow to stream reaches.
reach_latflow nreach Lateral inflow to each stream reach.
Unused potential evapotranspiration transferred to finite-difference cells
unused_potet nhru Unsatisfied potential evapotranspiration in each HRU available as evapotranspiration from
unsaturated and saturated zone of finite-difference cells.
Transfer of ground-water discharge from finite-difference cells to gravity reservoirs
basin_gw2sm 1 Watershed area-weighted average ground-water discharge from all connected
gw2sm nhru HRU area-weighted average ground-water discharge from connected finite-difference
cells.
gw2sm_grav nhrucell Ground-water discharge from connected finite-difference cell to gravity reservoir.
actet_gw nhru HRU area-weighted evapotranspiration from finite-difference cells.
actet_tot_gwsz nhru HRU area-weighted evapotranspiration from finite-difference cells and capillary reservoirs
in the soil zone.
GSFLOW variables that are associated with the exchange to a capillary reservoir is limited by storage in the gravity
of water between gravity reservoirs and connected finite- reservoir. Flow from gravity reservoirs to a capillary reservoir
difference cells are listed in table 10. is determined by first computing the capillary reservoir deficit,
which is the volume of water needed to replenish the capillary
reservoir to field capacity. The deficit in the capillary reservoir
Outflow from Gravity Reservoirs is calculated according to:
There are five possible processes by which outflow from
m,n m,n
gravity reservoirs can occur: flow to a capillary reservoir DdeficitCPR = Dfct HRU DCPR , (44)
(sequence No. 5, table 9), flow to a preferential-flow reservoir
(sequence No. 6, table 9), flow to downslope HRUs and(or)
where
stream segments (sequence No. 7, table 9), gravity drainage to m,n
a finite-difference cell (sequence No. 8, table 9), and Dunnian DdeficitCPR is the volume of water per unit HRU
runoff when pref_flow_den is zero (sequence No. 12, area required to replenish the
table 9). Each of these is described below. capillary reservoir to the field-
capacity threshold for time step m
Flow from Gravity Reservoirs to Capillary Reservoirs and iteration n, in inches; and
Dfct HRU is the maximum volume of water per
Gravity reservoirs replenish the capillary reservoir up unit area in the capillarry reservoir-
to field capacity whenever storage in the capillary reservoir parameter soil_moist_max,
is below the field-capacity threshold (fig. 19; and table 9,
in inches.
sequence No. 5); however, flow from a gravity reservoir
If a capillary reservoir has a deficit, water is removed Water is added to the capillary reservoir according to:
equally from each gravity reservoir associated with the
capillary reservoir to replenish the capillary reservoir to LL
AGVR , L
the field-capacity threshold. If the volume of water added
m , n*
DCPR m,n
= DCPR + DrplGVR
m,n
,L (46)
L =1 AHRU
to a capillary reservoir is limited by storage in the gravity
m,n
reservoir then the deficit in the capillary reservoir will not when DCPR < Dfct HRU ,
be replenished to field capacity. Water is removed from each
associated gravity reservoir according to:
where
m , n*
m,n DCPR is the revised volume per unit area of
DdeficitCPR AGVR
m,n
DrplGVR = water in the capillary reservoirr for time
AHRU step m, iteration n, in inches;
m,n
DdeficitCPR AGVR m , n 1 L is the counter for the gravity reservoir
when < DGVR
AHRU number, dimensionless;
m,n m , n 1
(45) LL is the total number of gravity reservoirs,
DrplGVR = DGVR
dimensionless;
m,n
DdeficitCPR AGVR m , n 1
m,n
DrplGVR,, L is the volume of water per unit area that
when DGVR ,
AHRU is removed from gravity reservoir L to
replenish the capillary reservoir during
where time step m and iteration n, in inches;
m , n 1
DGVR is thhe volume per unit area of water in the AGVR , L is the area of gravity reservoir L, in acres.
gravity reservoir during time step m
and iteration n-1, in inches; and A deficit in the capillary reservoir can occur following
DrplGmVR,n
is the volume of water per unit area that periods without precipitation and can be replenished from
is removed from the gravity reservoir gravity reservoirs that have inflow from ground-water
to replenish the capillary resservoir discharge. If the HRU area is much greater than the top area
during time step m and iteration n, of the finite-difference cell, focused ground-water discharge
in inches. can result in underestimates of soil-zone saturation and
overestimates of areas affected by ground-water discharge.
This problem is illustrated by considering flow of water from
a finite-difference cell to a gravity reservoir that represents
spring flow. This condition may result in a wetland in a
particular watershed. However, in the model, the water will
be averaged over the large volume of the capillary reservoir if
there is a deficit in the reservoir, resulting in a low saturation
in the capillary reservoir that is not indicative of a wetland.
This problem can be avoided by delineation of HRUs that
correspond to wetlands and riparian areas.
Flow from Gravity Reservoirs to Preferential-Flow The interflow (slow and fast) is developed from continuity and
Reservoirs an empirical equation written as (Leavesley and others, 1983):
Water in the gravity reservoirs is added to preferential- m,n
flow reservoirs whenever the preferential-flow threshold dDGVR m,n m,n
= qGVR , in qGVR , sif , and
(49)
(Dpft HRU ) is exceeded (fig. 19 and table 9, sequence number dt
6). The threshold is determined for each HRU as:
m,n m,n m,n 2
qGVR , sif = slwcoef lin DGVR + slwcoef sq ( DGVR ) , (50)
Dpft HRU = ( Dsat HRU Dfct HRU )(1 FdenHRU ) , (47)
where
where m,n
qGVR , in is the volumetric inflow rate per unit
Dpft HRU iss the preferential-flow threshold as
area to the gravity reservoir, in inches
volume per unit area,, in inches; and
per day;
Dsat HRU is the maximum volume of water per m,n
unit area in the soil zone-parameter qGVR , siff is the slow interflow rate per unit area
sat_threshold, in inches. from the gravvity reservoir, in inches
per day;
Any water in the gravity reservoir above the preferential- slwcoef lin is the linear flow routing coefficient for
flow threshold (Dpft HRU ) is subtracted from the gravity slow interflow parameter
reservoir and added to the preferential-flow reservoir. The sl lowcoef_lin, in per day; and
volume of water per unit area in the preferential-flow reservoir slwcoef sq is the non-linear preferential-flow
m,n
in inches (DPFR ) is calculated as: reservoir routing coefficient for slow
innterflow parameter
m, n m 1 m,n m,n
DPFR = DPFR + DGVR Dpft HRU DGVR Dpft HRU > 0 slowcoef_sq, in per inch-day.
(48)
m,n
DPFR = DPmFR1 m,n
DGVR Dpft HRU 0 ,
The right-hand side of equation 50 is substituted for
m,n
where qGVR , sif
in equation 49 and the result is integrated indefinitely:
m,n
DPFR is the volume per unitt area in the m,n
preferential-flow reservoir for time dDGVR
step m,, iteration n, in inches; and
dt = m,n
qGVR , in
m,n
slwcoef lin DGVR m,n 2 .
slwcoef sq ( DGVR )
(51)
m1
DPFR is the volume per unit area in the
preferential-flow reservoir at the last
iterration of time step m-1, in inches. The solution to equation 51 takes the form:
slwcoef lin + 2 slwcoef sq DGVR

m,n

2 tan 1
Flow from Gravity Reservoirs to Slow Interflow cx
t= + Ci , (52)
Slow interflow from the gravity reservoirs represents the cx
perching of water in the soil zone above the water table that
can occur as a result of mineralization near the bottom of the where
soil zone or when soil develops over fine-grained material.
cx = ( slwcoef lin ) + 4 slwcoef sq qGVR
2 m,n
Slow interflow can occur when the water content in the soil , in ; and
zone exceeds the field-capacity threshold (fig. 19 and table 9, Ci is a constant of integration , in days.
sequence number 7). The slow interflow equation is the same
as the fast interflow equation from preferential-flow reservoirs.
The arctangent function (tan-1) in equation 52 can be m,n

Equation 54 is rearranged in terms of DGFR and
rewritten in terms of a logarithm using the identity (Selby, substituted into equation 49. The time derivative in equation
1970; p. 166): 49 is approximated as:
1+ x m,n
dDGVR D m , n DGVR
m 1
tan 1 [ x ] = ln , (53) GVR , (55)
1 x dt t
where and the result is rearranged in terms of slow interflow from the
m,n gravity reservoir for iteration n, time step m as (Leavesley and
slwcoef lin + 2slwcoef sq DGVR others, 1983, p. 31-33):
x=
cx
slwcoef lin
1 + cs (1 e cxt )
The constant of integration in equation 52 can be cx
m,n
qGVR m,n
= qGVR ,in t + cs , (56)
m,n
determined by setting time to zero and the storage ( DGVR )
, sif
slwcoef sq
m1 1+ cs (1 e cxt )
equal to the initial storage (DGVR ) and replacing the arctan with cx
the identity in equation 53. Equation 52 then becomes:
where
slwcoef lin + 2 slwcoef sq DGVRm,n

1
+
cx
m 1
DGVR cx slwcoef lin
ln cs = .
m,n
AGVR 2 slwcoef sq
1 slwcoeflin + 2 slwcoef sq DGVR
cx
t=
m,n
cx Slow interflow (Dsif HRU ) is summed for all gravity
slwcoef lin + 2 slwcoef sq DGVRm 1
reservoirs in the HRU at the end of the iteration as:
1 +
cx k
ln m 1
m,n
Dsif HRU = qim, sif, n t , (57)
1 slwcoeflin + 2 slwcoef sq DGVR i =1
cx

+ . (54)
cx where
m,n
Dsif HRU is the volume peer unit area of slow
interflow from the gravity reservoirss
of the HRU at time step m, iteration
n, in inches;
k is the total number of gravity reservoirs
in the HRU, dimensionless; and
m,n
qi , sif is the slow interflow from gravity
reservoir i for time step m, iteration
n, in inches per day.
Slow interflow is partitioned to downslope HRUs Potential gravity drainage is averaged for iterations
and(or) stream segments during the iteration loop according n and n-1 after the second iteration, and is summed for all
to the cascading-flow procedure and connections defined gravity reservoirs contributing to a single finite-difference cell
by parameters input to the Cascade Module (cascade_prms, according to:
table 8). Slow interflow to a downslope HRU is added to the
capillary reservoir (equation 46) and(or) the stream segment at LL
wqgd , L + (1 w)qgd , L AGVR , L
m,n m , n 1
the beginning of the next iteration. Slow interflow decreases
the volume of storage in a gravity reservoir prior to computing m,n
qgdc , pot = L =1 C
prms 2 mf , (60)
gravity drainage, such that: Afdca
m , n* m.n m,n
DGVR = DGVR AHRU qGVR , sif t .
(58) where
m,n
qgdc , pot is the potential gravity drainage for time
The linear coefficient of slow interflow can be estimated step m and iteration n from all gravity
using the recession constant determined from hydrograph reservoirs connected to a finite-
separation (Linsley and others, 1975, p. 227). difference cell, in length per time;
m,n
qgd , L is the potential gravity drainage per unit
Flow from Gravity Reservoirs to Gravity Drainage area of gravity reservoir L for time
step m and iteration n, in inches per day;
Potential gravity drainage from the soil zone is computed m , n 1
as a function of storage in the gravity reservoirs after slow qgd , L is the potential gravity drainage per unit
interflow has been removed (fig. 19; and table 9, sequence area of gravity reservoir L for time
number 8). Net gravity drainage from each gravity reservoir step m and iteration n-1, in inches per day;
is dependent on the ground-water heads and the vertical w is a weighting factor used to average gravity
hydraulic conductivity of the connected finite-difference cell. drainage betweeen iterations, set to 0.5;
Potential gravity drainage per unit area is computed as: and
C prms 2 mf is the conversion from PRMS inches per
DGVR
m,n

coefex HRU
day to MODFLOW-2005 length per
m,n
qgd , pot = coeflinHRU , (59) time ( table 7 ) .
Dmx HRU
m,n
where Net gravity drainage (qgdc , pot ) to the underlying finite-
m,n difference cell is dependent on conditions in the unsaturated
qgd , pot is the potential gravity drainage per unit
and saturated zones. First, net gravity drainage is set equal to
area for time step m and iteration n, the potential gravity drainage calculated above, or in the case
in inches per day; where this value exceeds the vertical hydraulic conductivity
coeflinHRU is the linear coefficient in the equation used of the unsaturated zone (Ks), it is set to the value of Ks.
to compute gravity drainage from the Because changes in gravity drainage to the unsaturated
gravity reservoir parameter zone are represented by waves (see discussion in section
ssr2gw_rate, in inches per day; Unsaturated-Zone Flow), small changes in gravity drainage
coefex HRU is the exponent in the equation used to can result in numerous waves in the unsaturated zone causing
compute gravity drainage from the an unnecessary computational burden. This problem is avoided
gravity reservoir parameter by grouping net gravity drainage into a set of individual
ssr2gw_exp, dimensionless; and values. Individual values are determined by dividing a span
DmxHRU is the maximum amount of gravity drainage of values that range between 110-4 and 86,400 m/d into 51
from the gravity reservoir parameter log-normally distributed numbers, which represent most or all
possible values of vertical hydraulic conductivity of natural
ssrmax_coef, in inches.
porous materials (Freeze and Cherry, 1979).
The volumetric water content at the top of the unsaturated zone is computed from net gravity drainage and the Brooks and
Corey (1966) relation as:
1/
m,n
qgdc
m,n
, net
m,n
=
Ks ( S y ) + r 0 < qgdc , net K s (61a)

m , n = s m,n
K s < qgdc (61b)
, net ,

where
m , n is the water content at thee top of the unsaturated zone for time step m,
and iteration n, dimensionless;
m,n
qgdc , net is the net gravity drainage to the connected finite-difference cell for time step m,
and iteration n, in length per time;
K s is the vertical saturated hydraulic conductivity of the unsaturated zone,
in length per time;
is the Brooks-Corey exponent, dimensionless;
S y is the specific yield that is used to approximate s r , dimensionless;
r is the residual water content of the unsaturated zone, in volume of water per
volume of rock; and
s is the saturated water content of the unsaturated zone, in volume of water
per volume of rock.
Specific yield is used to approximate s r to maintain continuity across the boundary between the unsaturated and
saturated zones (Niswonger and others, 2006a).
Net gravity drainage is applied to the saturated zone instead of the unsaturated zone if the ground-water head is above the
soil-zone base minus one-half the undulation depth (fig. 20). Net gravity drainage to the saturated zone decreases as a function
of ground-water head and is calculated according to:
m , n*
qgdc , net = 0 h mfdc, n > celtop + 0.5 Dusz
m,n
qgdc , net
m , n*
qgdc , net =
celtop + 0.5 Dusz h mfdc, n celtop 0.5 Dusz h mfdc, n celtop + 0.5Dusz
Dusz (62)
m , n* m,n
qgdc , net = q gdc , net celtop 0.5Dusz > h mfdc,nn ,
where
m , n*
qgdc , net is the net gravity drainage to the saturated zone for time step m and iteration n,
in length per time.
The undulation depth defines where drainage from the soil zone becomes dependent on the ground-water head. The default
value of Dusz is 1.0 and can be changed to other values on the basis of input specifications in the Unsaturated-Zone Flow
Package.
Soil zone Undulations along

soil-zone base
Location of water table where gravity drainage
Recharge
from soil zone ceases
Height of
Soil-zone undulations
base at soil-zone
base
Location of water table where gravity drainage
Discharge from soil zone is no longer affected by water table
Unsaturated or saturated zone
Column or row width of finite-difference cell
Figure 20. Effect of undulations at the soil-zone base on gravity drainage in relation to the
water table in a finite-difference cell.
Evapotranspiration m,n
ETCPR m,n
= EAPRCPR m
PETHRU , (63e)
Evaporation from bare soil and transpiration from plants
is computed from the soil zone through the capillary reservoir where
(table 9, sequence Nos. 9-11). Evaporation and transpiration m,n
DRCPR is the soil-water content ratio in the
are subtracted from the reservoir as long as (1) there is water capillary reservoir for time step m and
in storage, and (2) the rate of potential evapotranspiration iteration n, dimensionless;
is greater than zero after evaporation from the plant canopy, m,n
EAPRCPR is the actual to potential
evaporation from impervious surfaces, and snow sublimation
evapotranspiration ratio in the capillary
has occurred. If sufficient water is available to satisfy all
remaining potential evapotranspiration after surface losses, resservoir for time step m and iteration
then that volume is removed from the capillary reservoir. n, dimensionless; and
m,n
Water is removed from the capillary reservoir during the ETCPR is the actual evapotranspiration removed
iteration loop because ground-water discharge into the gravity from the capillary reservoir for time
zone could affect the volume of water stored in the capillary step m and iteration n, in inches per
reservoir. For the purposes of computing evapotranspiration, day.
the soil water in the capillary reservoir is partitioned into
m,n
two parts. The first, termed recharge soil water, is available The calculation method of EAPRCPR is dependent on
m1
for evaporation and transpiration. The second, termed the value of DCPR and soil type (Zahner, 1967; Leavesley
lower zone soil water, is available for transpiration only, and others, 1983, p. 22-23). If the predominate soil type of
m1
and is active only during the seasonal transpiration period. the capillary reservoir is sand and DCPR is greater than 0.25,
Evapotranspiration and transpiration from both partitions of equation 63b is used; otherwise equation 63d is used. If the
m1
the capillary reservoir are computed the same way: predominate soil type is loam and DCPR is greater than 0.5,
equation 63b is used; otherwise equation 63c is used. If the
m1
m,n
m 1
DCPR predominate soil type is clay and DCPR is greater than 0.67,
DRCPR = , (63a) m1
equation 63b is used; if DCPR is greater than 0.33, but less than
Dfct HRU
0.67, equation 63c is used; otherwise equation 63d is used.
Evapotranspiration from the capillary reservoir is
subtracted from the remaining potential evapotranspiration
1.0 for each HRU. The unused potential evapotranspiration is
m,n m,n made available for the unsaturated zone and the ground-water
EAPR CPR = DRCPR and , (63b, c, d)
1 m,n
system in MODFLOW-2005 when the rooting depth of plants is
2 DRCPR beneath the soil zone.
nv07tac-4124_fig20
Dunnian Runoff Dunnian runoff from an HRU is routed to the capillary

reservoir of downslope HRUs (equation 39) and(or) to stream
Dunnian runoff is simulated from the soil zone segments with the cascading-flow procedure and connections
in an HRU when storage as a volume per unit area in defined by parameters input to the PRMS Cascade Module
the preferential-flow (or gravity reservoirs when (cascade_prms, table 8). Dunnian runoff typically occurs from
pref_flow_den is zero) exceeds the depth defined HRUs that define riparian areas next to streams or lakes or
by the saturation minus field-capacity thresholds (table9, depressions in the topography.
sequence No. 12) as:
m,n
DSZfc m , n**
= DGVR m , n 1 Fast Interflow from Preferential-Flow Reservoirs
, tot + DPFR , (64)
Fast interflow is simulated whenever water is stored in
the preferential-flow reservoir (table 9, sequence No.13).
where
m,n
It is computed every iteration of a time step using the
DSZfc is the volumme of water per unit area of the same empirical equation as slow interflow from the gravity
soil zone above field cappacity for time reservoirs (equation 56). The volumetric-flow rate per unit
step m and iteration n, in inches; m,n
area from the preferential-flow reservoir (qPFR , fif
) is computed
m , n**
DGVR , tot is the volume of water in all gravity using the fast interflow coefficients (fastcoef_lin and
reservoirs affter interflow and gravity fastcoef_sq) in equation 56 as:
drainage in an HRU, divided by the area
of the HRU for time step m and iteration fstcoef lin (67a)
1 + cs (1 e cxt )
n, in inches; and m,n m,n cx
m , n1 qPFR = qPFR , in t + cs ,
DPFR is the volume per unit area of water in the , fif
fstcoef sq
preferential-flow reservoir before fast 1+ cs (1 e cx t
)
cx
w for time step m and iteration
interflow
n-1, in inches.
where
fstcoef lin is the linear flow routing coefficient for fast
Dunnian runoff from each HRU is computed as:
interflow parameter
faast_coef_lin, in per day; and
m,n
ROd HRU m,n
= DSZfc ( Dsat HRU Dfct HRU ) fstcoef sq is the non-linear flow routing coefficient for
(65)
m,n
when DSZfc Dsat HRU Dfct HRU , fast interflow parameter
fast_ _coef_sq, in per inch-day; and
D m 1 cx fstcoef lin
where cs = PFR .
m,n
ROd HRU is the volume of Dunnian runoff per unit AHRU 2 fstcoef sq
area from an HRU for time step m and
iteration n, in inches. Flow from the preferential flow reservoir decreases the
volume of storage according to:
Dunnian runoff from an HRU comes from either the
preferential-flow reservoir when pref_flow_den is greater m m 1 m,n (67b)
DPFR = DPFR qPFR , fif t ,
than zero or from the gravity reservoirs otherwise. Dunnian
runoff from gravity reservoirs in an HRU is computed as:
where
m
LL DPFR is the volume of watter in the preferential-
m,n
ROd HRU = ROd Lm , n , (66) flow reservoir per unit area for time
L =1
step m, in inches.
where Fast interflow from an HRU is routed to the capillary

ROd Lm , n is the Dunnian runoff from gravity reservoir reservoir of downslope HRUs (equation 39) and(or) to
L for time step m, iteration n, in inches. stream segments with the cascading-flow procedure and
connections defined by parameters input to the Cascade
Module (cascade_prms, table 8).
Unsaturated-Zone Flow d
= i , (69c)
Unsaturated-zone flow is simulated in GSFLOW using dt
a kinematic-wave approximation to Richards equation that
assumes diffusive gradients (capillary pressure gradients) where
are negligible (Colbeck, 1972; Smith, 1983). This allows v() is the characteristic velocity restricted to the
Richards equation to be solved using the method of downward ( positive z ) direction , in length
characteristics, which was originally done by Smith (1983) per time.
and Charbeneau (1984). A brief description of the approach
as implemented in GSFLOW is provided here; the reader is Equation 69a provides the velocity of waves that represent
referred to Niswonger and others (2006a) for a more complete wetting and drying in the unsaturated zone. Equation
narrative. 69b provides the change in water content caused by
The approach of simulating unsaturated flow in evapotranspiration during wetting. Equation 69c provides
GSFLOW differs from previous approaches that coupled a the change in water content caused by evapotranspiration
one-dimensional finite-difference form of Richards equation during drying. Equations 69a, b, and c are separable and can
to two- or three-dimensional ground-water flow equations be integrated to find algebraic equations representing the
(Pikul and others, 1974; Refsgaard and Storm, 1995) because characteristics of wetting and drying fronts in the unsaturated
it does not have a fixed-grid structure. The method adds zone.
flexibility for simulating an unsaturated zone that can change An analytic solution for the wetting-front velocity,
in thickness through space and time. The package simulates v(), can be derived by considering the effects of diffusion
ground-water discharge directly to the soil zone or land and substituting an equivalent sharp wetting front of equal
surface as well as evapotranspiration from the unsaturated and mass (Smith 1983; Charbeneau, 1984). For simplicity,
saturated zones. evapotranspiration is left out of the derivation because it
The kinematic-wave approximation to Richards equation does not show up in the final equation for wave velocity.
can be written to include evapotranspiration losses as: The solution to equation 68 for a wetting front that considers
hydraulic diffusion can be found by integrating over a control
K () (68) volume containing a single wetting front according to (Phillip,
+ +i = 0 ,
t z 1957; Charbeneau, 1984):
z
where d 2
(70)
z2
dz + ( K () D() ) =0,
is the volumetric water content , inn volume of dt z1 z
z1
water per volume of rock;

z is the altitude in the vertical direction , in length;
K () is the unsaturated vertical hydraulic where
conductivity as a function of water coontent D() is the hydrauliic diffusion coefficient , in length
and is equal to the vertical flux , in length squared per time; and
per time; z1 and z2 are points above and below the wetting front
i is the evapotranspiration rate beneath the
at distances far enough that 0,
soil-zone base per unit depth , in length per z
time per length; and respectively, in length.
t is time.

Application of the method of characteristics to equation The term D() , can be neglected because 0 at z1
68 results in the following set of coupled ordinary differential and z2 such that: z z
equations:
z
d 2
dt z1
dz + K ( z2 ) K ( z1 ) = 0 , (71)
dz K ()
= = v() , (69a)
dt
where
d i (69b) K ( z1 ) and K ( z2 ) are thee values of K () at
= ,
dz v() z1 and z2 , respectively.
Integrating over a profile containing a sharp front with the water content and flux of the uppermost lead wave are
equivalent mass gives: maintained, resulting in rewetting. A depiction of how wetting
and drying fronts are approximated by kinematic waves is
z2 shown in figure 21.
dz =
z1
z1 ( z f z1 ) + z2 ( z2 z f ) , (72) A lead waves velocity and water content will decay
during a subsequent period of less infiltration as trailing waves
overcome the lead wave. Analytical equations that represent
where the velocity of a trailing wave can be derived. However, these
z1 is the volumetric water content at depth z1 , equations become complicated when trying to represent a
in volume of water per volume of rock; trailing wave intercepting a lead wave. A simpler approach
z2 is the volumetric water content at depth z2 , in is to divide the trailing wave into steps or increments and
calculate velocities ( v()) of each increment on the basis of a
volume of water per volume of rock; and
finite-difference approximation (Smith, 1983):
z f is the depth of the sharp front.
K () K ( )
Combining equations 71 and 72 gives: v() = , (75)

dz f
= us ( z1 , z2 ) =
( )
K z1 K z2( ), (73)
where is the change in water content between two adjacent
locations along a trailing wave, in volume of water per volume
dt z1 z2 of rock. An analytic solution of equation 69c, as shown in the
next section, is solved simultaneously with equation 65 during
evapotranspiration.
where
us is the veloccity of a sharp wetting front, in
length per time. WATER CONTENT ( 0 )
0i 02 01
Equations 69c and 73 must be solved simultaneously
because zf and z1 are both unknown when ET losses are Trailing wave
considered. represented as a
step function at
The Brooks-Corey unsaturated hydraulic conductivity time t2
function is used to represent K () and can be expressed as
(Brooks and Corey, 1966): Actual wetting front
at time equals t1

r
K () = K s
DEPTH (z)
, (74) A Leading wave with

s r B water content 01
03 at time t1
where Leading wave with

K s , r , and are defined in equation 61. Actual wetting front water content 03
at time equals t2 at time t2
Wetting and Drying Fronts in Unsaturated Zone

An increase in the infiltration rate from the soil zone
into the unsaturated zone will cause a wetting front to form, Water table
which is represented by a lead wave. A decrease in the
infiltration rate will cause a drying front to occur, which Figure 21. A wetting front moving through a uniform column of
is represented by a trailing wave. Thus, waves are used to unsaturated material affected by a decrease in flux (volumetric
represent both wetting and drying fronts. Attenuation of flow per unit area), and results from a kinematic-wave solution of
a lead wave occurs as a trailing wave of higher velocity the wetting front represented by leading and trailing waves (from
overtakes it. When a trailing wave overtakes a lead wave, the Niswonger and others, 2006a, fig. 2 and modified from Smith and
water content of the lead wave becomes equal to the water Hebbert, 1983).
content of the trailing wave. Consequently, this reduces the
velocity and water content of the lead wave. Conversely,
when a lead wave overtakes a trailing wave or another lead
wave of lower velocity, the overtaken wave is removed, and
Evapotranspiration Beneath the Soil Zone Recharge to Saturated Zone

Evapotranspiration that occurs at or below land Recharge to the saturated zone is computed at the end
surface can be simulated with the PRMS Soil Zone Module of a time step on the basis of the volumetric flow in the
(soilzone_gsflow, table1) when the bottom of the soil unsaturated zone across the water table plus any water that
zone is defined as the root depth. However, where soils may be in storage in the unsaturated zone when the water
are thin and the root depth extends beneath the soil-zone table rises. Thus, large fluctuations in the water table can
base, evapotranspiration that is not satisfied by storage in occur during a time step, but the volumetric flow and storage
the soil zone is removed from the underlying unsaturated of the unsaturated zone is kept in balance with recharge to
and saturated zones. Evapotranspiration is removed from the saturated zone even during these large fluctuations. This
the unsaturated zone by integrating equations 69b and 69c includes the disappearance and appearance of an unsaturated
and evapotranspiration is removed from the saturated zone zone. A complete description of the coupling between the
(ground water) using the same method implemented in the unsaturated and saturated zones is presented by Niswonger
MODFLOW-2005 Evapotranspiration Package (Harbaugh, and others (2006a, p. 8-10).
2005). Recharge is dependent on the location of the water table
Integrating equations 69b and 69c results in: and how its position varies with time. This is because ground-
water flow in the saturated zone does not account for storage
[ K ( z2 ) K ( z1 )] iz1 in the unsaturated zone. Rather, a specific yield is used to
z2 = , and (76a) estimate changes in ground-water storage. Thus, when the
i
water table rises, storage in the unsaturated zone in the interval
of the rise is added as recharge to the saturated zone. When
+ t = it (76b) the water table declines, the thickness of the unsaturated zone
is increased and a wetting front must advance through the
interval of the decline before there is recharge.
where
z1 and z2 are depths of two arbitrary points along a
wetting front, where z2 is deeper than Streams
z1 , in length;
and + t are the times at the beginning and end of Streams are simulated in GSFLOW by the Streamflow-
the time step, respectively; Routing Package in MODFLOW-2005. Accumulated surface
is the water content of trailing waves runoff and interflow generated in the HRUs, and base flow
above the evapotranspiration extinction generated by ground-water discharge to the stream reaches
are routed by the methods described in the following section.
depth after time t, in volume of water
Stream depths, intermittent streamflow conditions, and
per volume of rock; and
unsaturated flow conditions beneath streams also can be
+t is the water content of trailing waves
computed. The reader is referred to Prudic and others (2004)
above the evapotranspiration extinction
for a more complete narrative.
depth after time t + t , volume of
water per volume of rock.
Streamflow Routing
Evapotranspiration beneath the base of the soil zone only Streamflow is computed each iteration by stream reach
occurs when there is an evapotranspiration deficit in the soil starting from the first reach of the first stream (farthest
zone and when evapotranspiration is active in the Unsaturated- upstream) segment and sequentially continuing to the last
Zone Flow Package. The deficit is used to calculate reach of the last stream (farthest downstream) segment (fig. 8).
transpiration from the unsaturated and saturated zones Inflows to a stream reach include any user-specified inflows to
according to equations 76a and 76b. Equation 76a is used to the first reach of a stream segment, inflows from an upstream
calculate the depth of points along the leading-wave profile reach, precipitation directly on the channel, surface runoff
between z1 and z2, where z1 = m , n is the water content at the and interflows from contiguous HRUs, and ground-water
top of the unsaturated zone corresponding to the infiltration discharge across the streambed (fig. 22A). Outflows include
m,n
, net z2
rate qgdc . is unknown and must be solved for by coupling
equations 73 and 76a.
any specified diversions into pipelines at the beginning of the first reach in a stream segment or into unlined canals from the last
reach of a stream segment, evaporation directly from the channel, downward leakage across the streambed, and flow from the
end of the reach (fig. 22B). Outflows from the last reach of a stream segment are saved and added as inflow to the connecting
downstream segment or, if the segment is not connected to any downstream reaches, the outflow from its last reach exits the
modeled region. Outflows from multiple upstream segments or tributary segments can be added as inflow to a downstream
reach.
User-specified flows can be added to or subtracted from the beginning of the first reach of a segment. The volumetric flow
rates are specified by stress periods in the data input of the Streamflow-Routing Package (appendix 1). Thus, a stream that enters
the modeled region can have a specified flow at the beginning of the first reach (fig. 22A). If the first reach of a stream segment
has no streamflow entering the reach, any specified rate less than zero is ignored. Lateral inflow from contiguous HRUs is
computed to a stream reach by summing all surface runoff and interflows from each HRU as:

JJ C '
FJ , sr AJ ( ROhJm , n + ROd Jm , n + Dsif Jm , n + Dfif Jm , n ) ,

m,n prms 2 mf
Qlateral = (77)
t J =1
where
m,n
Qlateral is the volumetric flow ratee into a stream reach from all connected HRUs for time
stepp m and iteration n, in cubic length per time;
'
C prms 2 mf is the conversion factor from PRMS acre-inch per day to MODF FLOW-2005 cubic
length per time ( table 7 ) ;
FJ , sr is the decimal fraction of the total area of HRU J that contributes runoff and interflow
to a particular stream reach definedd by GSFLOW parameter segment_pct_area
in table 4, dimensionless;
AJ is the area of HRU J , in acres;
ROhJm , n is the Hortonian runoff per unit area from HRU J for time step m and iteration n,
in inches;
ROd Jm , n is the Dunnian runoff per unit area from HRU J for time step m and iteration n,
in inches;
m,n
Dfif J is the fast interflow from preferential flow reservoirs per unit area of HRU J for time
step m and iteration n, in inches;
Dsif Jm , n is the slow interflow from gravity reservoirs per unit area of HRU J for time step m and
iteration n, in inches;
J is the counter for the HRU number, dimensionless; and
JJ is the total number of HRUs that contribute surface runoff and interflow to a particullar
stream reach.
A
Tr
ibu
tar PIP
y ELIN stream
E
stream ry
uta
T rib
Flood
wave
Surface Surface
runoff runoff
Soil zone Interflow Interflow

Precipitation
table
Stream reach Water
Ground-water
discharge
Finite-difference cell
B PIPELINE
Flood
wave
Evaporation
Stream reach
Diversion
Stream leakage
Figure 22. (A) Inflows to and (B) outflows from a stream reach.
nv07tac-4124_fig22
Ground-water discharge across the streambed is option allows for the diversion of a specified fraction of flow;
computed when the underlying ground-water head in the and the fourth option allows for diversion of all flow in excess
finite-difference cell connected to the stream is greater of the specified rate. Flow from the channel at the end of a
than the stream head at the midpoint of the reach, whereas stream reach that may become inflow to a downstream reach
downward leakage is computed when the ground-water head is the difference in flow that enters and leaves the stream reach
is less than the stream head. Methods for determination of the during a time step.
stream head are discussed in section Computation of Stream Two options are available to route streamflow through
Depth. The volumetric rate of flow across the streambed is a network of channels. The first is based on steady, uniform
computed as (Prudic and others, 2004, p.4): flow for computing flow through rectangular or non-prismatic
channels (Prudic and others, 2004, p. 6). Outflow from the end
hstr
m,n
h mfdc, n of a stream reach is set equal to the sum of all inflows to the
m,n
Qsrleak = K strbed wetperstr lengthstr , (78) reach minus any upstream outflows. Precipitation, evaporation,
thick
strbed and surface runoff were originally defined by data input to the
Streamflow Routing Package (Niswonger and Prudic, 2005,
where p. 28 and described in appendix 1). However, these processes,
m,n as well as interflow through the soil zone, are computed in
Qsrleak is the volumetric-flow rate across the
GSFLOW using PRMS and, therefore, these input variables
streambed for time step m and iteration
should be set to zero for an integrated simulation. However,
n and is downward leakage when
stream inflows and outflows other than those calculated by
positive and ground-water discharrge to PRMS (for example stream diversions) can be specified in the
the stream when negative, in cubic Streamflow Routing Package input file.
length per time; Outflow at the end of the reach can be expressed as:
K strbed is the hydraulic conductivity of the
streambed, in length per time; m,n m,n m m,n
Qsrout = Qsrup + Qsrin + Qlateral
wetperstr is the wetted perimeter of the streambed,
m,n m,n
in length; + Qsrpp Qsrevp Qsmrleak
,n m,n
Qsrdvr , (79)
lengthstr is the length of the stream reach , in length;
m,n
hstr is the stream head at the midpoint of the where
stream reach for time step m and m,n
Qsrout is the volumettric flow rate from the end of
iteration n, in length; and a stream reach for time steep m and
thickstrbed is the streambed thickness in the stream iteration n, in cubic length per time;
reach , in length. m,n
Qsrup is the sum of the volumetric flow rate that
enters the strream reach from outflow of
Transient leakage across the streambed can change
upstream reaches for time stepp m and
during a time step and is dependent on both the stream and
iteration n, in cubic length per time;
ground-water heads. When the ground-water head in the finite- m
difference cell is less than the bottom of the streambed (that Qsrin is the specified volumetric flow rate at the
is, the top of streambed minus the thickness of the streambed, beginning of a sttream reach for time
commonly called percolating conditions), downward step m, in cubic length per time;
leakage across the streambed is no longer dependent on Qlmateral
,n
is the specified volumetric flow rate added
ground-water head; in this case, the ground-water head in to a stream reach as lateral inflow plus
equation 78 is replaced by the altitude of the bottom of the any runoff and interfflow coming from
streambed. adjacent HRUs, in cubic length per time;
Streamflow can be diverted from the last reach of a m,n
Qsrpp is the specified precipitation rate on the
segment and routed downstream to other segments (Prudic stream surface multiplied by the wetted
and others, 2004, p. 5 and 9). The diverted flow is subtracted plan-view area for time step m and
from the end of the last reach of a segment and any remaining
iteration n, in cubic length per time;
flow in that segment is added as inflow to the beginning of m,n
the first reach of the next downstream segment. Four different Qsrevp is the specified evaporation rate from the
options are available for diverting flow into another stream stream surfface multiplied by the wetted
segment (variable IPRIOR in the input instructions for the plan-view area for time steep m and
Streamflow-Routing Package in appendix 1). The first option iteration n, in cubic length per time; and
m,n
allows for the diversion of all available flow up to the specified Qsrdvr is the volumetric flow rate diverted from the
rate; the second option allows for the diversion of the specified end of the stream reach for time step m
rate only when flow is greater than the specified rate; the third and iteration n, in cubic length per time.
The second option is based on a kinematic-wave equation The kinematic-wave approximation neglects the dynamic
that approximates the Saint-Venant equations for computing components of flow that are represented by the derivative
flow through rectangular or non-prismatic channels. The terms in the more complete form of the momentum equation
kinematic-wave equation for routing flow in streams can be referred to as the Saint-Venant equations. Another less-
written (Lighthill and Whitham, 1955): simplified approximation of the Saint-Venant equations
combines the spatial derivative from the momentum equation
msfr
Qdwn msfr
Adwn with the continuity equation, which results in a second
+ = qstr , (80) derivative term in the continuity equation (Lighthill and
xchannel t Whitham, 1955). The second derivative term in the continuity
equation causes the flood wave to spread upstream slightly,
where and is commonly referred to as a diffusion analogy for
msfr
Qdwn is volumeetric-flow rate at the downstream the dynamic component of the momentum equation. This
approach is the basis for the surface-water model called
end of a stream reach foor the current
DAFLOW that was coupled with MODFLOW (Jobson and
streamflow-routing time step msfr , in
Harbaugh, 1999).
length squared;
msfr
The kinematic-wave equation (equation 80) was solved
Adwn is the cross-sectional area at the numerically by stream reaches using an implicit four-point,
downstream end of a stream reach for finite-difference solution technique (Fread, 1993). Because
the current streamflow-routting time the movement of a flood peak in a channel may be of interest,
step msfr , in length squared; time steps for streamflow routing can have durations less than
qstr is the sum of all lateral inflows and outflows the time steps for GSFLOW or MODFLOW-2005. An option
per unit length of stream listed on the was added that divides the GSFLOW or MODFLOW-2005
right side of equation 79, in length time step in a specified number of increments. The time
squared per time; derivative ( Astr / t ) was approximated by a forward-
xchannel is the distance along the stream channel, difference quotient centered between the end of the previous
in length; and and current time steps used for routing streamflow:
t is time.
msfr
Adwn ( Aupmsfr Aupmsfr 1 ) + ( Adwn
msfr msfr 1
Adwn )
, (82)
Momentum in the kinematic-wave approximation t 2tsfr
assumes that gravitational forces are balanced by frictional
forces such that:
where
Aupmsfr is cross-sectional area at the upstream endd
Slope0 = Slope f , (81)
of a stream reach for the current
where length squared;
Slope0 is the slope of the channel in thhe Aupmsfr 1 is cross-sectional area at the upstream end
longitudinal profile, dimensionless; of a stream reach for the preevious
and streamflow-routing time step msfr -1,
Slope f is the friction slope of the channel in the in length squared;
longitudinal profiile, dimensionless. msfr
Adwn is cross-sectional area at the downstream
end of a stream reach for the current
Several forms of the momentum equation are combined
streamflow-routing time step msfr ,
with equation 80 to route flow to downstream channels. The
in length squared;
form of the momentum equation differs in how the channel msfr 1
cross-sectional dimensions and channel slope are represented Adwn is cross-sectional area at the downstream
in Mannings equation. Another option represents momentum end of a stream reach for the prevvious
using tables of specified values from streamflow gages, or streamflow-routing time step msfr -1,
the power-law equations described by Prudic and others in length squared; and
(2004, p.6-9). Momentum caused by the water-surface slope, tsfr is the time-step for routing streamflow ,
velocity head, and acceleration is neglected. in time.
The spatial derivative in equation 80, Qstr / xchannel , is ground-water head at the beginning of the time step is beneath
approximated by a forward-difference quotient that includes the bottom of the streambed, downward leakage and recharge
time weighting of the quotient at the previous time step: to the saturated zone is summed for all time subdivisions and
added to the ground-water flow equation to compute ground-
msfr 1 water heads.
msfr
Qdwn (Qdwn Qupmsfr 1 ) + ( 1)(Qdwn
msfr
Qumsfr
p ) (83)
,
xchannel x
Computation of Stream Depth
where Stream depth varies as a function of streamflow. It is
is the time-weighting factor that ranges added to the top of the streambed to determine the stream
between 0.5 and 1, dimensionless; head, which is used to compute ground-water discharge and
Qupmsfr is the volumetric flow at the upstream end downward leakage across the streambed (equation 78). Stream
depth as a function of streamflow is computed at the midpoint
of the stream reach for the currrent
of each reach. Streamflow at the midpoint is computed as:
cubic length per time; m,n m,n m m,n m,n m,n m,n
msfr 1 Qmdpt = Qsrup + Qsrin + 0.5(Qlateral + Qsrpp Qsrevp Qsrleak ) , (84)
Qup is the volumetric flow at the upstream end
m,n
of the stream reach for the previous Qsrdvr
streamflow-routingg time step msfr -1, in
cubic length per time; and where
msfr
Qdwn is the volumetric flow at the downstream m,n
Qmdpt is the volumetric-fllow rate at the midpoint
end of the stream
m reach for the current of a stream reach for time step m and
streamflow-routing time step msfr , in iteration n, in cubic length per time.
cubic length per time;
msfr 1
Qdwn is the volumetric flow at the downstream end
of the stream reach for the previoous This equation is nonlinear because flow at the midpoint
streamflow-routing time step msfr -1, in of a reach is dependent on streambed leakage, which is
cubic length per time; dependent on stream depth. Thus, the equation is solved
x is the spatial increment for routing iteratively using a mixed bisection-Newton method until the
streamflow , equal to the reach length. calculated stream stage changes between iterations is less than
a specified tolerance. This tolerance is specified in the input
Equations 82 and 83 are combined and solved iteratively for the Streamflow-Routing Package (appendix 1). A warning
using Newtons method (Burden and Faires, 1997, p.65). message is written in the main MODFLOW Listing File for
Streamflow for a particular stream reach for all time each stream reach whenever the solution error is greater than
increments computed with the kinematic-wave equation the specified tolerance after 100 iterations.
(including those smaller than the daily time step used by the One of five options can be used to compute stream depth
rest of GSFLOW) can be saved to a file using the Stream in a reach as specified in the input for the Streamflow-Routing
Gaging Station or Gage Packages (appendix 1). Package (appendix 1). The option used to calculate stream
Because streamflow routing during each subdivision depth is defined for each segment such that the depths in all
of a GSFLOW time step is a function of the head difference reaches within a segment are calculated the same. The first
between the stream and ground water, the ground-water option (ICALC=0) allows the user to specify a stream depth at
head at the beginning of the time step is used to compute the beginning and end of each segment; these two depths are
the exchange of water across the streambed and the stream used to interpolate the depth at the midpoint of each reach in
depth for each time subdivision. If the ground-water head the segment on the basis of the distance downstream from the
at the beginning of the time step is at or above the bottom beginning of the segment. This option for stream depth is not
of the streambed, the computed stream depths for all time recommended for the integrated model because stream depth
subdivisions are averaged and used in the ground-water flow remains constant and, therefore, the volume of water in storage
equation to compute ground-water heads at the end of the in the stream network does not change as a function of inflows
GSFLOW time step. However, the depth at the end of the and outflows.
time step is used for subsequent stream calculations. If the
The second and third options (ICALC=1 and 2, 3/ 5

respectively) use the general form of Mannings equation to m,n
Qmdpt
m,n
Roughness
D = , (86)
relate streamflow as a function of depth for all reaches in a mdpt
C w ( Slope )1/ 2
m str o
stream segment (Chow and others, 1988, p. 159) as:
wheere
Cm
Astr ( Rhydraulic ) ( Slopeo ) , (85)
m,n 2/3 1/ 2
Qmdpt = wstr is the stream width, in length; and
Roughness m,n
Dmdpt is the depth of water at the midpoint of a
stream reach for time step m and
where iteration n, in length.
Cm is a constant, which is 1.0 for cubic meters
per second and 1.486 for cubic feet per The third option (ICALC=2) uses one channel cross
second; section for a stream segment that is divided into three parts
Roughness is Manning's roughness coefficient , on the basis of eight paired horizontal and vertical locations
dimensionless; and (fig. 23). Eight horizontal distances relative to the left edge
Rhydrauliic is the hydraulic radius of the stream, which of the cross section (viewed in downstream direction) and the
is equal to the stream area divided by corresponding eight vertical altitudes relative to the specified
the wetted perimeter, in length. top of streambed (lowest point or thalweg in the channel)
are used for computing stream depth, top width, and wetted
The wetted perimeter and hydraulic radius of a stream, which perimeter. Stream depth, width, and wetted perimeter also are
are needed in Mannings equation, are often complicated dependent on the slope of the streambed, and two roughness
functions of depth, and can only be solved analytically by coefficientsone for the center part of the cross section and
making assumptions to simplify the channel geometry. An another for the two outer parts that may represent overbank
analytical equation is used for ICALC=1 by assuming a wide flow. The ends of the cross section are assumed to have
rectangular channel (Shen and Julien, 1993, p. 12.13). This vertical walls. Because the stream area and hydraulic radius
assumption approximates the wetted perimeter as equal to the for this option result in implicit functions of depth, a mixed
stream width and results in an approximation of the hydraulic bisection-secant method is used to compute stream depth in
radius that is equal to stream depth. Solving equation 85 for relation to streamflow (Prudic and others, 2004, p. 7-8).
stream depth then yields (Prudic and others, 2004, p. 7):
Part 1 Part 2 Part 3

VERTICAL DISTANCE
Left bank Channel Right bank

FROM THALWEG
Stream width
Water surface
Stream
depth
Wetted perimeter
(0,0)
HORIZONTAL DISTANCE FROM LEFT EDGE OF

CROSS SECTION, VIEWED LOOKING DOWNSTREAM
Figure 23. Eight paired horizontal and vertical locations used to compute stream depth, top
width, and wetted perimeter for a stream segment (modified from Prudic and others, 2004, fig. 4).
The fourth option (ICALC=3) uses power-law equations Intermittent Streams

for a stream segment that relate stream depth and width to
streamflow using the following two equations (Leopold and Stream reaches in any segment can have intermittent
others, 1992, p. 215): flow. Downward leakage from a reach during a time step is
limited by the available flow in a stream reach. Water available
for leakage when the ground-water head is beneath the bottom
= cdpth ( Qmdpt ) = awdth ( Qmdpt )
fdpth bwdth
m,n
Dmdpt m,n m,n
, and wmdpt m,n
, (87) of the streambed is calculated:
m,n m,n m m,n m,n m,n

where Qsrleak , mx = Qsrup + Qsrin + Qlateral + Qsrpp Qsrevp , (88)
m,n
wmdpt is the stream width at reach midpoint , in
length; where
cdpth and m,n
Qsrleak , mx is the total amount of water availabble for
awdth are empirical coefficients determined from
leakage to ground water for time step m
regression methods, in time per length
and iteration n, in cubic length per time.
squared; and
fdpth and However, leakage loss across the streambed from a stream
bwdth are empirical exponents determined from reach cannot exceed the rate of gravity drainage calculated
regression methods, dimensionless. according to:
This option assumes that the wetted perimeter is equal to Dmdpt

m,n

m,n
the stream width. The regressions typically are determined Qsrleak = K strbed wetperstr lengthstr + 1 . (89)
thick strbed
at streamflow-gaging stations. Representative values of the
numerical coefficients for selected regions of the United States
are listed in Leopold and others (1992, table 7-5, p. 244). No downward leakage is computed for a reach when
The fifth option (ICALC=4) allows the user to enter there are no upstream inflows or diversions and the ground-
a look-up table of stream depth, width, and corresponding water head in the finite-difference cell is beneath the top of
streamflow for a stream segment. This option also assumes the streambed. However, a reach with no surface inflows may
that the wetted perimeter is equal to the stream width. have outflow from it when ground-water discharge along the
The values typically are determined from rating curves at reach is greater than the evaporation rate within the reach.
streamflow-gaging stations. Stream depth and width are
interpolated between two specified streamflows that bracket
the computed flow at the reach midpoint. Values of depth, Unsaturated Flow Beneath Streams
width, and streamflow are converted into their logarithms prior
to the interpolation (Prudic and others, 2004, p. 9). A simple An unsaturated zone develops beneath a stream reach
linear extrapolation is used to estimate stream depth and width whenever the ground-water head in the finite-difference cell
when the computed streamflow at the midpoint of the stream is less than the bottom of the streambed. Gravity drainage
reach is less than the first specified streamflow. A warning through the unsaturated zone is simulated using the kinematic-
message is printed when the computed streamflow exceeds the wave approximation to Richards equation similar to that
highest specified streamflow in the table and stream depth and used to simulate unsaturated-zone flow beneath the soil zone.
width are computed by extrapolation from the log values of However, downward leakage across a streambed is used in
the last two streamflows in the tabulated list. place of gravity drainage for calculating infiltration into the
unsaturated zone and no evapotranspiration can be simulated
from an unsaturated zone beneath a reach. Maximum
streambed leakage for reaches that have an underlying
unsaturated zone is limited to the product of the saturated
vertical hydraulic conductivity of the unsaturated zone and the
streambed area (stream length multiplied by wetted perimeter).
Thus, streambed leakage in reaches with an unsaturated zone
is controlled either by the saturated hydraulic conductivity
of the streambed, as shown in equation 89, or the saturated
hydraulic conductivity of the unsaturated zone.
The ability to simulate unsaturated flow beneath streams Inflow and Outflow
is limited to the two options that compute stream depth on the
basis of Mannings equationthe options that assume a wide A lake in GSFLOW is represented as both an
rectangular channel of constant wetted perimeter (ICALC=1) HRU in PRMS and a group of finite-difference cells
or an eight-point cross section that allows the wetted perimeter in MODFLOW-2005. Inflows to a lake include direct
to change as a function of streamflow and stream depth precipitation; surface runoff and interflows through the
(ICALC=2). The unsaturated zone may be discretized into soil zone from contributing areas adjacent to the lake;
compartments beneath an eight-point cross section (fig. 24). specified inflow, such as a pipeline (negative withdrawal;
Water is routed independently beneath each of the unsaturated appendix1); tributary streamflow; and ground-water
compartments (Niswonger and Prudic, 2005, p. 6) and can be discharge through the subsurface (fig. 25). Outflows
used to simulate the variability in leakage and gravity drainage from a lake include evaporation, specified withdrawal,
through the unsaturated zone across a stream channel. discharge to a stream, and leakage to ground water
(fig.25). Precipitation, evaporation, and surface runoff
were originally defined by data input to the Lake Package
Lakes (Merritt and Konikow, 2000, p. 7-8 and described in
appendix 1). However, these processes, as well as interflow
Lakes are simulated in GSFLOW by a modified
through the soil zone, are computed in GSFLOW using
implementation of the Lake Package in MODFLOW-2005.
PRMS and, therefore, these input variables for the Lake
The overall approach is described in detail by Merritt and
Package should be set to zero for an integrated simulation.
Konikow (2000); a brief description of the overall approach,
and GSFLOW-related changes to the solution routines, are
described herein.
Maximum stream width

Stage at higher streamflow
Streambed
Stage at lower streamflow
Unsaturated flow
at higher streamflow
Compartment
height
Unsaturated zone
compartment
Water table
Compartment
width
MODFLOW column or row width

(larger than maximum stream width)
Figure 24. Discretization of unsaturated zone beneath an eight-point cross section of a stream segment within a finite-
difference cell (from Niswonger and others, 2005, fig. 3).
Tributary
PIPE stream
LIN E
Outlet stream
Surface
runoff
Precipitation Evaporation Surface
runoff
Int low
erf erf
low Int
Lake
cell
Ground-water
discharge Lake leakage
L a k e b e d
Aquifer
cell
Figure 25. Inflows to and outflows from a lake. Grid represents finite-difference cells for lakes and ground water.
Surface runoff (both Hortonian and Dunnian) and fast and Any snow accumulation on a lake is expressed as a liquid
slow interflow to a lake HRU are calculated for all iterations of equivalent. Thus, the stage of all lakes is assumed to be the
a time step using the cascading-flow procedure as: liquid equivalent of the solid and liquid phases. The volume of
inflow from surface runoff and interflow from all contributing
m,n
VlakeHRU = HRUs is added to the volume of precipitation minus
evaporation on the lake HRU and converted to a volumetric
JJ
F J , lakeHRU AJ ( ROhJm , n + ROd Jm , n + Dsif Jm , n + Dfif Jm , n ) , (90)

flow rate with dimensions used by the Lake Package as:
J =1
'
m, n
C prms 2 mf
QlakeHRUtolakemf =
where t
m,n
VlakeHRU is the volume of surface runoff and interflow ( PlakeHRU
m m
EvaplakeHRU m,n
) AlakeHRU + VlakeHRU , (91)
routed to a lake HRU from contributing
HRUs for time step m and iteration n,
in acre-inch; where
m,n
FJ ,lakeHRU is the decimal fraction of the total area of QlakeHRUtolakemf is the voluumetric flow rate from a lake
HRU J that contributes runoff and HRU to a lake represented inn
interflow to a lake HRU defined by MODFLOW-2005 for time step m
GSFLOW parameter hru_pct_up and iteration n, in cubic length
in table 8, dimensionless; per time;
'
J is the counter for the HRU number, C prms 2 mf is the conversion from PRMS inch-acre
dimensionless; and to MODFLOW-2005 cubic length;
JJ is the total number of HRUs that contribute m
PlakeHRU is the precipitation on the lake HRU
surface runoff andd interflow to a particular for time step m, in inches;
lake, dimensionless. m
EvaplakeHRU is the evaporation from the lake HRU
for time step m, in inches; and
AlakeHRU is the area of the lake HRU, in acres.
nv07tac-4124_fig25
Stream inflow to a lake is computed as the sum of all

Aaqm, n
outflows from the last reach of each tributary stream segment m,n
Qlakeleak = hlake
m,n m,n
haqfd
c , (92)
(Merritt and Konikow, 2000, p. 11; Prudic and others, 2004, thicklkbd thickaq
+
p. 9). Any number of inflowing stream segments can be added K lkbd K aq
to a lake. Although most lakes have only one outlet stream,
stream outflow from a lake is the sum of any number of outlet
where
stream segments. Outflow from a lake to a stream segment m,n
can be user-specified, either as a constant flow rate or by way Qlakeleak is the volumetric flow rate acrross the
of a lookup table, or it can vary as a function of lake stage lakebed to center of aquifer finite-
and the outlet channel geometry. Additionally, non-stream- difference ceell for time step m and
related withdrawals also can be specified from a lake by the iteration n and is outward or downward
user. The withdrawals are specified every stress period in the leakage when positive and ground-water
input for the Lake Package (appendix 1). A negative value for discharge wheen negative, in cubic length
withdrawal can be used to specify the addition of water to a per time;
lake from some external source such as flow through a pipe. K lkbd is the hydraulic conductivity of the lakebed ,
The exchange of water between a lake and ground water in length per time;
is dependent on the lake stage and ground-water heads next thicklkkbd is the lakebed thickness, in length;
to and beneath the lake. Ground water discharges to a lake thickaq is the horizontal or vertical distance from
when and where the ground-water head is greater than the lakebed to the center of the adjacent
lake stage, whereas outward and downward leakage occurs finite-difference cell, in length;
when and where the ground-water head is less than the lake
Aaqm, n is the area of the lakebed covering
stage. Ground-water discharge to a lake or leakage from a lake
may occur anywhere within a lake. The distribution of these a finite-difference cell,, in length
flows will depend on the lake stage and the ground-water head squared;
calculated in aquifer cells adjoining the lake. Lakes can have K aq is the horizontal or vertical hydraulic
areas of ground-water discharge and leakage at the same time, conductivity of aquifer finite-difference
and the areas of ground-water discharge and leakage can vary cell adjaccent to lake cell, in length per
through time depending on transient variations in lake stage time;
and ground-water heads. m,n
hlake is the lake stage for time step m and
A lake occupies the entire surface area of a finite- iteration n, in length; and
difference cell and does not overlap with an aquifer within the m,n
haqfdc is the ground-water head in the finite-
cell. A lake can occupy many finite-difference cells that can
difference ceell adjacent to the lake at
extend over multiple finite-difference layers (fig. 9). A lakebed
is assumed to exist between the lake cells and the aquifer the cell node (fig. 26B) for time step m
finite-difference cells and it can have different properties than and iteration n, in length.
the underlying finite-difference cells (fig. 26A). Flow between
lakes and ground water is computed as a function of the Leakage across the lakebed can change during a time step
lakebed thickness and hydraulic conductivity and the thickness and is dependent on the lake and ground-water heads and the
and hydraulic conductivity of the finite-difference layer cross-sectional area through which ground-water discharge
beneath the lakebed (fig. 26B; Merritt and Konikow, 2000, and leakage is computed. The cross-sectional area of flow
eqs. 2 and 4). This function is: along a vertical face can vary as a function of the ground-
water head. The altitude at the top of the finite-difference cell
adjacent to the lakebed is used to calculate leakage when the
ground-water head in the finite-difference cell is beneath the
lakebed.
Lake surface
Lake bottom
Lakebed
thicklkbd
EXPLANATION
C lkbd Lakebed conductance thick aq Conductance terms
Length squared per time
C aq Aquifer conductance Klkbd
Clkbd = A aq
Length squared per time thick lkbd
K lkbd Lakebed hydraulic conductivity
Length per time K aq
K aq Aquifer hydraulic conductivity Caq = Aaq
Length per time thick aq
A aq Cross sectional area of aquifer
Length squared
thick lkbd Lakebed thicknessLength A aq Point in aquifer
thick aq Aquifer thicknessLength
AQUIFER
Aquifer cell Lake cell
Kh
Cell
node
L a k e b e d
Kv
EXPLANATION
Kh Horizontal hydraulic conductivityLength per time
Kv Vertical hydraulic conductivityLength per time
Figure 26. (A) Conductance terms used to compute leakage from and ground-water discharge to a lake; (B)
representation with finite-difference cells in MODFLOW-2005 (modified from Merritt and Konikow, 2000, fig. 1).
nv07tac-4124_fig26
Computation of Lake Stage and Outflow to Streams

Because the Lake Package as implemented in GSFLOW differs from the original version of Merritt and Konikow (2000),
an extended discussion of these differences is provided in this section. Lake stage is computed on the basis of the sum of the
inflows minus outflows and a change in lake storage for each time step. The equation can be written as (modified from Merritt
and Konikow, 2000, p. 9):
LC LC
m, n
out Qwd + Cd lc hafdc ,lc (1 ) hlake Cd lc
m, n m, n m, n m m , n m , n 1 m 1
QlakeHRUtomf + Qtotstrin Qtotstr
m 1
hlake + tmf lc lc
m 1
m, n
Alksurf
hlake = , (93)
tmf LC
1 + m 1 Cdlcm, n
Alksurf lc =1
wheere
LC is the number of finite-difference cell faces in contact with the submerged lakebed;
lc is a counter used to refer to individual finite-difference cell faces in contact with the
submerged lakebed;
m, n
QlakeHRUtomf is the sum of precipitation , evaporation , surface runoff , and interflow calculated in PRMS,
in cubic length per time;
tmf is the MODFLOW-2005 time step, in time;
m, n
Qtotstrin is the sum of all tributary stream inflows for time step m and iteration n, in cubic length per time;
m, n
Qtotstrout is the sum of lake outflows to streams for time step m and iteration n,, in cubic length per time;
Qwdm
is the specified withdrawal from a lake ( negative value adds water ) for time step m, in cubic
length per time;
m, n
Cdlc is the lakebed conductance of cell face lc for time step m and iteration n, in length squared per time;
m , n1
hafdc ,lc is the ground-water head in finite-difference cell associated with cell face lc for time step m and
iteration n 1, in length;
is the time-weighting factor that ranges between 0.5 and 1, dimensio onless; and
m1
Alksurf is the surface area of the lake for the last iteration of time step m 1, in length squared.
The method for solving transient lake stage in the Lake changed to solve for lake stage and outflow to streams within
Package (Merritt and Konikow, 2000, p. 9-10) was revised the Formulate Subroutine of the Lake Package (Merritt and
for GSFLOW because lake outflow originally was computed Konikow, 2000, p. 39-40).
using a combination of lake stage from the previous time step The revision resulted in several changes to both the Lake
and the previous iteration within the time step. This method and Streamflow-Routing Packages. In the previous version of
could result in numerical oscillations of lake stage and outflow the Lake Package (Merritt and Konikow, 2000), the relation
to streams that caused the simulation to not converge to its between lake stage and surface area was discontinuous when
closure criterion in a finite-difference cell adjacent to a lake or the lake stage fluctuated through the boundary between
beneath a stream reach downstream from the lake. MODFLOW layers. This discontinuity would, in some cases,
Newtons method of computing lake stage for a steady- result in convergence failure for the solution of lake stage.
state stress period (Merritt and Konikow, 2000, p. 12) also Thus, the relation between lake stage and surface area was
was revised to compute lake stage for both steady-state and revised such that the lake stage would linearly transition over
transient stress periods. Previously, lake outflow to a stream the boundary between MODFLOW layers. Additionally,
and its derivative were computed once in the Streamflow- the calculation for streamflow emanating from a lake as a
Routing Package only for steady-state stress periods (Prudic function of lake stage is updated during the iterative solution
and others, 2004, p. 9-10). This approach did not allow the of lake stage to improve convergence. A new subroutine was
outflow to a stream and its derivative to change as the steady- added to the Streamflow-Routing Package to compute a table
state lake stage solution was updated. The computation of of streamflow versus stream depth values for streams that
lake stage for steady-state and transient simulations was emanate from a lake.
The relation between streamflow verses stream depth Three separate simulations were used to test differences
for a stream emanating from a lake can be developed using in results between the original and revised versions of the
any of the options for computing stream depth in a segment: Lake Package with time-weighting values of 0.0, 0.5, and 1.0.
Mannings equation assuming a wide rectangular channel The solutions were consistent between the two versions of
(ICALC=1) or an eight-point cross section (ICALC=2); the Lake Package. A time-weighting factor of 0.5 resulted in
power-law equations that relate stream depth and width to a transition of the Lake Package stage between the initial and
streamflow (ICALC=3); and lists of values (a look-up table) final solution that was midway between those computed using
that relate measured streamflows to measured stream depths a time weighting factor of 0.0 and 1.0. Despite differences
and widths (ICALC=4). A list is created for each outlet stream in the transition period, the three simulations resulted in the
by subtracting the maximum lake stage from the top of the same final lake stage and lake inflow. A time-weighting factor
streambed at the beginning of the outlet stream segment, of 1.0 decreased the number of iterations needed to converge
dividing that maximum depth into 200 increments, and on a lake stage compared to a time-weighting factor of 0.5;
computing flows at the beginning of the outlet segment using however, the number of iterations needed to converge on
one of the four options for computing stream depth. The result ground-water heads increased for a time-weighting factor of
is that lake stage and outflow to streams are solved within the 1.0.
Lake Package for a consistent set of ground-water heads. The maximum number of iterations and the closure
Although the data input structure was not altered, the new criterion for the solution of lake stage can be specified
method for solving lake stage does not use the time-weighting using the Lake Package variables NSSITER and SSCNCR,
factor ( ) for steady-state stress periods, and the factor is respectively (appendix 1). These values must be specified
automatically set to a value of 1.0 for all steady-state stress for a simulation that includes a steady-state stress period and
periods such that the new lake stage is not dependent on lake are optional for simulations that have only transient stress
stage from the previous time step (equation 93). The original periods. The maximum number of iterations is set to 100
conceptualization of the time-weighting factor THETA by and the closure criterion is set to 0.0001 for all simulations
Merritt and Konikow (2000, p. 52) was to allow for ground- that have only transient stress periods. An option in the Lake
water discharge to, and leakage from, the lake to depend on a Package input was created in which values of NSSITER
combination of lake stage from the previous and current time and SSCNCR can be read by specifying a negative time-
step. weighting factor (appendix 1). A negative time-weighting
Because Newtons method for solving lake stage and factor is automatically reset to a positive value after values of
outflow to streams is used for transient stress periods, the NSSITER and SSCNCR are read.
range of the time-weighting factor is limited to values from The revised method for solving for lake stage in
0.5 to 1.0. A value of 0.5 represents the stage midway between conjunction with outflow to streams may not always converge
the previous time step and the end of the current time step. A to the closure criterion because of discontinuities among lake
value of 1.0 (fully implicit) represents the lake stage at the end stage, area, and outflow to streams. This is particularly so
of the current time step. A time-weighting factor of less than for steady-state simulations of lakes with outflow to streams.
0.5 does not perform well and a zero value is undefined when A warning message is printed that includes the last two
using Newtons method. computed lake stages of a particular lake whenever the change
Slight errors in the transient solution of lake stage, in lake stage between successive Newton iterations is not less
ground-water discharge, and lake leakage may result when than the closure criterion after reaching the maximum number
the time weighting factor is greater than 0.5. However, a of iterations.
weighting factor greater than 0.5 can be useful for dampening
oscillations (Fread, 1993). A value of 0.5 represents a semi-
implicit method that commonly is called Crank-Nicolson Intermittent Lakes
(Wang and Anderson, 1982; p. 81). Wang and Anderson The Lake Package was further revised to allow for the
(1982) present results of drawdown for a simple confined continued accounting of surface inflow into intermittent lakes
aquifer with pumping in which the Crank-Nicolson method and for the filling of an empty lake when the sum of inflows
produced the best results when compared with the Theis from precipitation, surface runoff, interflow, tributary streams,
solution for different times. A value of 0.5 is the default value and ground-water discharge exceed the sum of outflows
used in the solution of lake stage for transient stress periods. to specified withdrawal, evaporation, leakage, and outlet
Results from the revised computational method of streams. The original version of the Lake Package (Merritt
computing lake stage were compared with the first example and Konikow, 2000, p. 10-11) did not maintain a water budget
test simulation presented by Merritt and Konikow (2000, p. once a lake was empty, and the lake could only refill from
21-23). No inflow from or outflow to streams was simulated ground-water discharge. An empty lake was filled to the
in this example and lake stage rose through time as a result of average ground-water head in cells beneath the lowest lake
ground-water discharge to and precipitation on the lake. The area whenever the average head at the end of a time step was
simulation was for 5,000 days and was divided into 100 time
steps.
above the top of the lakebed. This approach for refilling the cell. Initial water contents within the unsaturated zone beneath
lake was discarded in favor of an approach that accounts for a lake can be calculated using a steady-state simulation or
all inflows and outflows to an empty lake within the GSFLOW specified in the input for the Unsaturated-Zone Flow Package.
and Lake iteration loops.
Several changes were made to the lake-stage calculation
to balance flow in and out of a lake. The calculations of Ground Water
outflow from lakes to streams were added to the Newton
Flow in the saturated zone is governed by a partial
iteration loop in the Formulate Subroutine of the Lake
differential equation for three-dimensional ground-water flow
Package. If a specified withdrawal from a lake exceeds the
of constant density (Harbaugh, 2005, p. 2-1):
sum of computed inflow less ground-water discharge, the
specified withdrawal is decreased to equal the water available
in the lake. In order to simulate rewetting of a lake, ground- h h h h
K xx + K yy + K zz + W = Ss , (94)
water discharge to an empty lake is computed assuming the x x y y z z t
lake head is equal to the top of the lakebed. If evaporation
from an empty lake exceeds the sum of ground-water
where
discharge and other inflows and outflows, evaporation is
decreased to the amount of available water. If inflows exceed K xx , K yy and K zz are the hydraulic conductivity
evaporation, leakage to ground water, and stream outflow, then tensors aligned with the x, y, and z
the lake is allowed to rewet. The computation of lake stage and coordinate axes, in length per time;
outflow to streams requires that the top of the streambed for h is the potentiometric head, in length;
all outlet streams must be equal to or greater than the top of W is volumetric flow rate per unit volume
the lakebed in the lowest part of the lake. representing sources and / or sinks of water,
W <0.0 for flow out of the ground-water
system, and W >0.0 for flow into the system,
Unsaturated Flow Beneath Lakes in per time; and
An unsaturated zone can develop beneath a finite- S s is the specific storage of the porous rock ,
difference cell represented by a lake whenever the ground- in per length.
water head in the underlying finite-difference cell is less than
the bottom of the lake cell. An unsaturated zone beneath a lake
can develop when the hydraulic conductivity of the lakebed The specific storage (Ss) and hydraulic conductivities (Kxx,
is much less than that of the underlying aquifer and/or ground Kyy, and Kzz) vary as a function of space, and the volumetric
water is pumped from a well or wells near the lake. flow rate per unit volume (W) varies as a function of space and
The Lake Package was further revised to compute time.
unsaturated flow beneath a lake using the Unsaturated-
Zone Flow Package whenever the ground-water head in
an underlying cell is beneath the bottom of the lake cell. Volume-Averaged Flow Equation
Hydraulic properties of the unsaturated zone in cells beneath
Equation 94 represents flow through a macroscopic
a lake are specified in the data input for the Unsaturated-Zone
region large enough to consider water and porous material as a
Flow Package in the same manner as any other cell that is used
single continuum; however, finite-difference cell volumes used
to simulate unsaturated flow from land surface to the water
in MODFLOW are usually much larger than this macroscopic
table (appendix 1). Unsaturated flow beneath a lake cell is
region. Thus, MODFLOW solves a volume-averaged equation
computed from the bottom of the lake cell. The bottom of a
that is derived by integrating equation 94 over an arbitrary
lake cell can be higher than the water table in the underlying
cell volume. The finite-difference method is applied following
aquifer cell, in which case leakage is routed through the
the derivation of the volume-averaged flow equation, which
unsaturated zone between the lakebed and underlying water
results in a system of algebraic equations. A convention on
table.
the basis of alphabetical indices is used for defining both the
Gravity drainage through the unsaturated zone beneath
cell center (node) and cell faces for the derivation of the finite
lakes is simulated using the same method as that used to
difference equations (fig. 27A). Both nodes and cell faces are
simulate unsaturated-zone flow beneath the soil zone and
indexed because ground-water head is calculated at a node
streams. Unsaturated flow and water-contents beneath a
and the flow is calculated across the cell face. Accordingly,
particular lake can be printed using the option variable
cell faces are referenced in the finite-difference equations by
NUZGAG in the data input for the Unsaturated-Zone Flow
adding or subtracting one-half of the node index to represent
Package (appendix 1). The maximum lake leakage through
one-half the cell distance in either the positive or negative
a lakebed is limited to the saturated vertical hydraulic
direction along the x, y, or z directions (figs. 7 and 27A).
conductivity multiplied by the top area of the finite-difference
A +y Cell i,j-1,k Cell i,j,k
+x
+z vk
qi,j-1/2,k
Cj-1 Cj ci
rj-1 rj
rj-1/2
B +y +x
i-1,j,k
+z
i,j,k-1
i,j-1,k i,j+1,k
i+1,j,k
i,j,k+1
Figure 27. (A) Flow into finite-difference cell i, j, k from cell i, j-1, k, and (B) indices for six
adjacent cells surrounding cell i, j, k (hidden) (modified from Harbaugh, 2005).
Thus, the four vertical cell faces surrounding the node i, j, k defined by dropping the indices that remain whole, such that
beginning with the side facing the x direction and moving the cell face i, j, k is referenced as face j 1/ 2 . Similarly, the
clockwise are defined as i, j, k; i, j, k; i, j+, k; and i+, distance between nodes is indexed using the same notation
j, k. The bottom and top faces below and above node i, j, k are as for cell faces, where, for example, the distance in the
i, j, k+, and i, j, k, respectively. However, faces will be x (column) direction between nodes i, j, k and i, j-1, k is
represented by rj 1/ 2 (figs. 7 and 27A).
Using the notation listed in figure 27, integrating the first Similarly, the other terms in equation 94 can be integrated
h such that the volume-integrated form of the equation can be
term of equation 94, and assuming is constant along the y
and z directions results in: x rewritten as:
face j +1/ 2 rj vk ( qi +1/ 2 qi 1/ 2 ) + ci rj ( qk +1/ 2 qk 1/ 2 )

h h (95)

ci r j vk
K xx
x
dzdydx =
x
face j 1/ 2
K xx
x
dxrj vk ,
x + ci vk ( q j +1/ 2 q j 1/ 2 ) + W = SS
h
vk ci rj , (98)
t*
where
where
rj is the length along the row direction for
qi 1/ 2 is the volumetric flow rate between cells
a finite-difference cell, in length;
i 1 and i, in cubic length per time;
ci is the length along the column direction for
qi +1/ 2 is the volumetric flow rate between cells
a finite-differeence cell, in length;
i and i + 1, in cubic length per time;
vk is the length along the layer direction for
qk 1/ 2 is the volumetric flow rate between cells
a finite-difference cell, in length;
k 1 and k , in cubic length per time; and
face j +1/ 2 is the cell face between cells j and j + 1,
qk +1/ 2 is the volumetric flow rate between cells
dimensionless; and k and k + 1, in cubic length per time.
face j 1/ 2 is the cell face between cells j 1 and j ,
dimensionless. Assuming the same sign convention as used by Harbaugh
(2005), the volumetric flow rate into the cell containing node i,
The integral can be carried further such that: j, k can be written as the sum of flows across six cell faces as:
face j +1/ 2
h h
face j +1/ 2
Q L = qi 1/ 2 + qi +1/ 2 + q j 1/ 2 + q j +1/ 2 + qk 1/ 2 + qk +1/ 2 , (99)
ci vk
face j 1/ 2
K xx
x
dx = ci vk K xx
x x face j 1/ 2
(96)
where
= ci vk ( q j +1/ 2 q j 1/ 2 ) ,
QL is th
he volumetric flow rate into a cell (negative
value is a volumetric flow out of cell), in
where cubic length per time.
q j 1/ 2 is thee volumetric flow rate between cells j 1
and j , in cubic length per time; and Combining equations 98 and 99 results in equation 2-2 of
q j +1/ 2 is the volumetric flow rate between cells j Harbaugh (2005, p. 2-3):
and j + 1, in cubic length per time.
h
Q L = SS
t*
V , (100)
The time derivative in equation 94 can be integrated over the
same volume as:
where
V is the volume of the cell, in cubic lenngth.
h h

r j ci vk
Ss
t
dzdydx = SS
t*
vk ci rj ,
(97)
where
SS is the voolume averaged specific storage, in per
length; and
h
is the change in head over a specified time
t*
interval, in leength per time.
Finite-Difference Formulation Substituting equations for each cell face (equation 101)
and equation 103 into equation 98 results in:
Darcys Law can be used to remove q from equation 98.
For example, after applying the finite- difference method for
CCi 1/ 2 ( him1, j , k him, j , k ) + CCi +1/ 2 ( him+1, j , k him, j , k )
flow across the side of the block facing the x direction the
flow can be written as: + CR j 1/ 2 ( him, j 1, k him, j , k ) + CR j +1/ 2 ( him, j +1, k him, j , k )
+ CVk 1/ 2 ( him, j , k 1 him, j , k ) + CVk +1/ 2 ( him, j , k +1 him, j , k )
q j 1/ 2 = CR j 1/ 2 (hi , j 1, k hi , j , k ) , (101)
him, j , k him, j,1k
+ W = SSi , j , k m m 1 V , (104)
where t t
* *
hi , j 1, k is head at noode i, j 1, k , in length;
hi , j , k is the head at node i, j , k , in length; and where
CR j 1/ 2 is the conductance between nodes i, j 1, k CCi 1/ 2 is the con nductance between nodes i 1, j ,
and i, j , k , in length squared per time, k and i, j , k , in length squared per time;
and is equal to : CCi +1/ 2 is the conductance between nodes i, j ,
( KR j 1/ 2 ci vk ) k and i + 1, j , k , in length squared per time;
CR j 1/ 2 = , (102) CR j 1/ 2 is the conductance between nodes i, j 1,
rj 1/ 2
k and i, j , k , in length squared per time;
where
CR j +1/ 2 is the conductance between nodes i, j , k and
KR j 1/ 2 is the hydrauliic conductivity between nodes
i, j + 1, k , in length squared per time;
i, j 1, k and i, j , k , in length per time.
CVk 1/ 2 the conductance between nodes i, j , k 1
and i, j , k , in length squared per time;
Equation 101 can be derived for all six faces surrounding node CVk +1/ 2 is the conductance between nodes i, j , k
i, j, k (fig. 27B). These equations are presented by Harbaugh and i, j , k + 1, in length squared per time;
(2005). him1, j , k is the head at node i 1, j , k , at end of
The derivative of head with respect to time is time step m, in length;
approximated using backward in time differencing according him+1, j , k is the head at node i + 1, j , k , at end of
to (Harbaugh, 2005): time step m, in length;
him, j 1, k is the head at node i, j 1, k , at end of
him, j , k him, j , k him, j,1k time step m, in length;
= , (103)
t* t*m t*m 1 him, j +1, k is the head at node i, j + 1, k , at end of
time steep m, in length;
where m
hi , j , k 1 is the head at node i, j , k 1, at end of
him, j , k
is the change in head at cell i, j , k at the time step m, in length;
t*
end of time step m over a specified time m
hi 1, j , k +1 is the head at node i, j , k + 1, at end of
interval, in length per time; time step m, in length; and
him, j1, k is the head at node i, j , k , at end of time SSi , j , k is the volume averaged specific storage for
step m 1, in length; cell i, j , k , in per length.
him, j , k is the head at node i, j , k , at end of time
step m, in length;
t*m is the time at the end of time step m, in time;
and
m 1
t* is the time at the end of time step m 1,
in time.
Equation 104 is an algebraic equation that provides the cells (interblock conductance). The horizontal conductance of
time variability of ground-water head and flow everywhere in a single cell can be calculated as:
a three-dimensional ground-water system. Equation 104 must
be solved for every variable-head cell in the model and be Tr Width fdc
Ch = , (106)
combined with initial and boundary conditions. This system of
Length fdc
equations can be written in compact form as:
C h = Chbnd + QW , where
(105)
Ch is the horizontal conductancce for a prism,
in length squared per time;
where
Tr is the transmissivity of the cell, in length
C is a matrix of conductance values for thhe row ,
squared per time;
column , and layer directions (CR, CC ,
Width fdcc is the width of the prism perpendicular to
and CV ) that are multiplied by the
the direction for which the conductance
ground-water head in each variable--head
is defined, in length; and
cell and any coefficients multiplied by a
Lengtth fdc is the length of the prism, in length.
variable-hhead cell to represent external
sources and sinks, in length squared per time;
h is a vector of ground-water heads for all For confined conditions, transmissivity is equal to the product
of the horizontal hydraulic conductivity of the cell multiplied
variable-head cells, in length;
by the saturated thickness of the cell. For unconfined
Chbnd is a vector of conductance values multiplied by
conditions, transmissivity is equal to the horizontal hydraulic
ground-water head in eachh constant-head
conductivity of the cell multiplied by the difference between
cell, in cubic length per time; and
the altitude of the water table in the cell and the altitude of the
QW is the volumetric flow rate added to cell bottom.
variable-head cells that represents external As a result of the spatial variability of hydraulic
sources and sinks to the ground-waterr system, conductivity, cell thickness, and ground-water head, there are
in cubic length per time. four methods for calculating interblock conductance between
cells that are supported by MODFLOW-2005. The first
The system of equations represented by equation 105 can be method for calculating interblock conductance for cells i, j-1,
solved using several different matrix solvers that are available k and i, j, k assumes uniform transmissivity within a cell and is
for MODFLOW-2005 (Harbaugh, 2005; p. 7-1). written as (Harbaugh, 2005; p. 53):
A complication to the solution of equation 105 for
unconfined ground-water conditions occurs because the C j 1C j
horizontal conductance values shown in equation 104 CR j 1/ 2 = , (107)
(C j 1 + C j )
are dependent on ground-water head (Harbaugh, 2005;
p. 5-5). Finite-difference equations used to solve for
head for unconfined conditions are solved iteratively where
by MODFLOW-2005 and the conductance values are C j 1 is the horizontal conducttance for cell j 1
approximated by the ground-water head from the previous in the row direction , in length squared
iteration. Certain boundary conditions used to represent per time; and
external flows to the ground-water system also add coefficients C j is the horizontal conductance for cell j in the
that are dependent on ground-water head. row direction , in length squared per time.
Conductance The other three methods for calculating horizontal

conductance between finite-difference cells are described
The conductance for a group of finite-difference cells by Goode and Appel (1992) and Harbaugh (2005). These
will vary in the x, y, and z directions for anisotropic and methods make different assumptions about the flow system
heterogeneous medium. Flow is calculated across cell faces and how transmissivity (equation 106) varies in space,
such that the conductance values used in the finite-difference including assuming that transmissivity varies linearly between
equations represent averages of the conductance values for two nodes, the aquifer is flat and homogeneous with a water table,
or the aquifer is flat with a water table in which hydraulic
conductivity varies linearly between nodes.
The vertical conductance for a cell is a function of assumes that the semi-confining unit makes no measurable
the cell thickness, horizontal area, and vertical hydraulic contribution to the horizontal conductance or the storage
conductivity. The vertical conductance of a single cell is capacity of two adjacent model layers. For this approach the
expressed as: conductance is calculated as (Harbaugh, 2005; p. 5-8):
K v Afdc (108) rj ci
Cv = , CVk +1/ 2 = , (110)
z 1 / 2thcki , j , k thckCB 1 / 2thcki , j , k +1
+ +
VK i , j , k VKCBi , j , k VK i , j , k +1
where
Cv is the vertical conductance, in length squared
per time; and where
z is the cell thickness, or, for unconfined thckCB is the thickness of thhe semi-confining unit ,
conditions, the thickness of the saturated in length; and
region of the cell, in length. VKCBi , j , k is the vertical hydraulic conductivity of the
semi-confining unit beeneath cell i, j , k ,
in length squared per time.
The vertical conductance between two nodes will be the
equivalent conductance of two one-half cells in a series for
two cells with ground-water heads that are greater than the Storage
altitude of the cell top. The vertical conductance between cell
Storage is formulated differently depending on the layer
i, j, k and i, j, k+1 for the condition of no semi-confining unit
type designation that is specified within the Block-Centered
between the two cells is calculated according to (Harbaugh,
Flow or Layer-Property Flow Packages of MODFLOW-2005.
2005; p. 5-7):
Layers in the model can be specified as confined or
convertible. The storage coefficient remains constant
rj ci throughout a simulation for layers designated as confined,
CVk +1/ 2 = , (109)
1 / 2thcki , j , k 1 / 2thcki , j , k +1 whereas the value of the storage coefficient may switch during
+ a simulation if the ground-water head passes through the top
VK i , j , k VK i , j , k +1
of a layer designated as convertible. The rate of accumulation
of water in a cell with a constant storage coefficient can be
where expressed as:
thcki , j , k and thcki , j , k +1 are the thicknesses of cells
i, j , k and i, j , k + 1, respectively, in V him, j , k him, j,1k
length these thicknesses are equal to cell = SSi , j , k (vk ci rj ) m m 1 , (111)
t t* t*
thickness for confineed conditions, and equal
to the difference between altitude of the water
table in the cell and altitude of the celll bottom where
for unconfined conditions; and V
is the rate of accumulation of water in the
VK i , j , k and VK i , j , k +1 are the vertical hydraulic t
cell, in cubic length per time.
conductivity of cells i, j , k and i, j , k + 1,
respectively, in length per time.
The Block-Centered Flow Package requires input of a
An approach for simulating the effects of semi-confining dimensionless storage coefficient instead of specific storage,
units on vertical flow through aquifers incorporates the and MODFLOW-2005 converts this coefficient to units of
thickness and vertical hydraulic conductivity of the semi- per length by dividing by the cell thickness for confined
confining unit into the calculation of the vertical conductance. conditions. The storage coefficient input for unconfined
This approach commonly is referred to as the Quasi-Three- conditions is equal to the specific yield. The Layer-
Dimensional approach (Bredehoeft and Pinder, 1970) and Property Flow Package also has an option for specifying a
dimensionless storage coefficient instead of specific storage.
External Boundary Conditions Head-dependent flow as described by equation 112a causes

additional terms to be added to the conductance matrix shown
All boundary conditions used in MODFLOW-2005 can in equation 104, whereas equation 112b does not affect the
be generalized from two equations. The first form represents conductance matrix.
known functions that are either head dependent or head
independent (Harbaugh, 2005):
ai , j , k , N = C1i , j , k , N hi , j , k + C 2i , j , k , N , (112a) Water Budgets

An important feature of the design of GSFLOW is the
where calculation of an overall water budget, as well as individual
ai , j , k , N flow from the Nth source into cell i, j , k , water budgets for all the storage reservoirs simulated by
in cubic length per time; GSFLOW. These reservoirs include snowpack, impervious
C1i , j , k , N is a constant for head-dependent boundary surfaces, surface-water bodies (streams and lakes), plant
canopy, soil zone, unsaturated zone, and saturated zone
conditions, in length squared per time;
(fig.28).
and
C 2i , j , k , N is a constant for head-independent boundary
conditions, in cubic length per time. Overall Water Budget
The overall water budget for a modeled region is
Thus, for head-dependent boundary conditions (such calculated at the end of each daily time step using:
as a stream), C1i , j , k , N is nonzero and C 2i , j , k , N may be zero or
nonzero. For head-independent boundary conditions (such as a
Stotal = t (Qin Qout ) , (113)
well), C1i , j , k , N is zero and C 2i , j , k , N is nonzero.
The Unsaturated-Zone Flow Package results in a
boundary condition that is different than that described by where
equation 112a because recharge from the unsaturated zone is Stotal is total storage change in all reservoirs
an unknown function of head. Recharge from the unsaturated from beginning to end of time step, in
zone is described by: cubic length;
Qin is total inflow to the modeled region during
ai , j , k , N = f1uzf (him, j , k , him, j,1k ) + f 2uzf (him, j,1k ) , (112b) a time step, in cubic length per time; and
Qout is total outflow from the modeled region
during a time step, in cubic length per time.
where
f1uzf is the recharge received from the unsaturated
zone thaat was stored in the unsaturated Inflows include precipitation and all surface-water and
interval through which thhe water table ground-water flows into the modeled region. Outflows include
rose during time step m, in length cubed evapotranspiration and all surface-water and ground-water
per time; and flows out of the modeled region.
uzf Ground-water heads to compute budget information are
f 2 is the recharge that flowed across the surface
calculated at the end of a GSFLOW time step. Subsequently,
defined by the water table in the cell at
flows and storages for the gravity reservoirs are calculated for
time m 1, in length cubed per time.
integrated simulations. Budget information calculated at the
end of a time step can be written to the GSFLOW Comma-
The functions f1uzf and f 2uzf in equation 112b are Separated-Values File and GSFLOW Water-Budget File.
considered unknown because they are dependent on the GSFLOW writes budget results to a user-specified output
storage and vertical flow through the unsaturated zone, file called the GSFLOW Water-Budget File that includes a
which are dependent on boundary conditions at the top of the tabular summary of internal flows and storages, as well as
unsaturated zone and ground-water head during previous time flows into and out of the modeled region. The name of this
steps. However, the values of these functions are computed file is specified using parameter gsflow_output_file
during each sequence of the iteration loop on the basis of in the GSFLOW Control File. An example output listing of
the ground-water head at the previous time step, previous water-budget information calculated by GSFLOW is shown
iteration, and the characteristics of the unsaturated zone. in figure 29. The GSFLOW overall volumetric water budget
Water Budgets 85
HYDROLOGIC RESPONSE UNITS
1 2 3 4
Plant canopy
Stream
Snow pack
Land surface Impervious surface
Soil zone
Unsaturated zone
MODFLOW layers
1
Lake
Saturated zone
2
1 2 3 4 5 6 7
MODFLOW columns or rows
EXPLANATION
PRMS surfaceComputed
PRMS surface or boundaryDefined
MODFLOW surfaceComputed
MODFLOW surface or boundaryDefined
Figure 28. Storage reservoirs represented in GSFLOW include the

plant canopy, snowpack, impervious surfaces, streams, and lakes
at or above land surface, and the soil, unsaturated, and saturated
zones in the subsurface. Adjacent reservoirs exchange water. The
soil zone is composed of three reservoirscapillary, gravity, and
preferential-flow (fig. 19).
is computed by the module gsflow_sum. These budget terms The input flows are: (1) precipitation, which includes
are computed for every model time step (day) and also as rain and snowfall; (2) streamflow that flows into the watershed
cumulative total values for the entire period of simulation. from outside the watershed; (3) ground-water boundary
The water-budget output is controlled by the parameter flow, which includes inflow from constant- and variable-
rpt_days specified in the GSFLOW Control File. Output head boundaries; and (4) flow from water injected into wells
only is printed to the Water-Budget File for integrated (fig. 29). The output flows are: (1) surface and subsurface
simulations. A value of 0 will turn off all reporting. The units evapotranspiration; (2) streamflow out of the watershed;
of this report are in the MODFLOW unit of cubic length (3) ground-water boundary flow, which includes outflow
(as specified in the Discretization File) for volumes and the from constant and variable-head boundaries; and (4) flow
MODFLOW unit of cubic length per day for volumetric rates. from pumped or artesian wells. The sum of the output flows
subtracted from the sum of the input flows is the total change
in flow within the watershed during the time step. These
flows, along with their GSFLOW budget report names are
summarized in table 11.
nv07tac-4124_fig28
SUMMARY VOLUMETRIC BUDGET FOR GSFLOW

DATE: 02 21 1984 CUMULATIVE TIME STEP: 1239
MODFLOW STRESS PERIOD 5 CURRENT TIME STEP: 692 ITERATIONS: 12
-----------------------------------------------------------------------------------
CUMULATIVE VOLUMES L**3 RATES FOR THIS TIME STEP L**3/T

------------------ ------------------------
IN IN
-- --
PRECIPITATION = 1.4834E+08 PRECIPITATION = 765880.63
STREAMFLOW = STREAMFLOW =
GW BOUNDARY FLOW = GW BOUNDARY FLOW =
WELLS = WELLS =
OUT OUT
--- ---
EVAPOTRANSPIRATION = 6.5165E+07 EVAPOTRANSPIRATION = 10703.55
STREAMFLOW = 5.7521E+07 STREAMFLOW = 24139.96
GW BOUNDARY FLOW = 815384.69 GW BOUNDARY FLOW = 657.31
WELLS = WELLS =
INFLOWS - OUTFLOWS = 2.4838E+07 INFLOWS - OUTFLOWS = 730379.81

-------- --------
STORAGE CHANGE = 2.4859E+07 STORAGE CHANGE = 730400.75

-------------- --------------
LAND SURFACE = 2.3072E+07 LAND SURFACE = 755174.81
SOIL ZONE = 2481753.00 SOIL ZONE = -1679.16
UNSATURATED ZONE = -2141891.25 UNSATURATED ZONE = -8121.79
SATURATED ZONE = 1446580.88 SATURATED ZONE = -14973.11
OVERALL BUDGET ERROR = -20668.00 OVERALL BUDGET ERROR = -20.94
PERCENT DISCREPANCY = -0.01 PERCENT DISCREPANCY = 0.00
Figure 29. Selected output from a GSFLOW Water-Budget File that shows a summary listing for December 15, 1981,
from the Sagehen Creek example problem.
The change in total storage (fig. 29) is calculated as zone; (3) unsaturated-zone storage from the Unsaturated-
the sum of the change of four reported storage components, Zone Flow Package; and (4) saturated-zone storage from
which are: (1) surface storage that is comprised of snowpack, MODFLOW-2005. Positive flow into storage results in a
intercepted precipitation, and impervious surface storage, positive change in storage. These storages, along with their
streams and lakes; (2) soil-water storage in the root GSFLOW budget report names, are summarized in table 11.
Water Budgets 87
Table 11. Inflow, outflow, changes in storage, and errors reported in the GSFLOW Budget Report File for GSFLOW.
GSFLOW budget report names Description

Inflows
Precipitation Cumulative volume of precipitation added to modeled region and volumetric flow rate for a time step.
Streamflow Cumulative volume of streamflow entering modeled region and volumetric flow rate for a time step.
Ground-water boundary flow Cumulative volume of ground water entering saturated zone along external boundaries and volumetric
flow rate for a time step.
Wells Cumulative volume of water added to saturated zone by injection wells and volumetric flow rate for a
time step.
Outflows
Evapotranspiration Cumulative volume of all evapotranspiration in modeled region and volumetric flow rate for a time step
Streamflow Cumulative volume of streamflow leaving modeled region and volumetric flow rate for a time step.
Ground-water boundary flow Cumulative volume of ground water leaving saturated zone along external boundaries and volumetric
flow rate for a time step.
Wells Cumulative volume of water removed from saturated zone by pumping wells and volumetric flow rate for
a time step.
Inflows less outflows Difference in cumulative volume of all inflows and outflows and difference in volumetric flow rate for a
time step.
Storage changes
Storage change Cumulative volume of total storage change and volumetric flow rate of change for a time step.
Surface Cumulative volume of surface storage change and volumetric flow rate of change for a time step. Surface
storage includes plant canopy, snowpack, impervious surfaces, streams, and lakes.
Soil zone Cumulative volume of soil-zone storage and volumetric flow rate of change for a time step.
Unsaturated zone Cumulative volume of unsaturated-zone storage and volumetric flow rate of change for a time step.
Saturated zone Cumulative volume of saturated-zone storage and volumetric flow rate of change for a time step.
Errors
Overall budget error Difference between inflows less outflow and total storage change.
Percent discrepancy Difference multiplied by 100 (see equation 114).
Simulation cumulative and time step errors are computed The errors, along with their GSFLOW budget report
as the difference between the total gain in flow and the total names, are summarized in table 11.
change in storage of the watershed. This difference also is
reported as a percent discrepancy, according to the following
equation: Computation of Budget Errors
GSFLOW also produces output for detailed analysis of
( in out ) t storage volumetric budget errors that is written to a user-specified
D = 100 , (114)
( (in + out )t + storage ) / 2 GSFLOW Comma-Separated-Values File. The name of this
file is specified using parameter csv_output_file in the
GSFLOW Control File. It includes storage and flows for each
where major storage zone of GSFLOW for each time step. These
D is the percent discrepancy, dimensionless; variables and their description are listed in table 12. Although
in is the total inflow to the model, in cubic this file may be useful for exporting GSFLOW output data
length per time; to other programs, it is specifically designed to be imported
out is the total outflow from the model, in cubic into a spreadsheet for detailed analysis. The units for each
length per time; and value in this file are in the MODFLOW unit of cubic length
stoorage is the total storage change, in cubic length. (as specified in the Discretization File) for volumes and the
MODFLOW unit of cubic length per day for volumetric rates.
Table 12. Names and descriptions of variables written to the GSFLOW Comma-Separated-Values File at end of each daily time step
for GSFLOW.
[Abbreviations: L3/T, cubic length per time; L3, cubic length]
Variable name Description Units

Date Month, day, and year designation MONTH/DAY/YEAR
basinppt Volumetric flow rate of precipitation on modeled region L3/T
basinpervet Volumetric flow rate of evapotranspiration from pervious areas L3/T
basinimpervevap Volumetric flow rate of evaporation from impervious areas L3/T
basinintcpevap Volumetric flow rate of evaporation of intercepted precipitation L3/T
basinsnowevap Volumetric flow rate of snowpack sublimation L3/T
basinstrmflow Volumetric flow rate of streamflow leaving modeled region L3/T
Potential volumetric flow rate of gravity drainage from the soil zone to the L3/T
basinsz2gw unsaturated zone (before conditions of the unsaturated and saturated zones are
applied)
Volumetric flow rate of ground-water discharge from the saturated zone to the soil L3/T
basingw2sz zone
Volumetric flow rate to saturated zone along external boundary (negative value is L3/T
gw_inout flow out of modeled region)
stream_leakage Volumetric flow rate of stream leakage to the unsaturated and saturated zones L3/T
uzf_recharge Volumetric flow rate of recharge from the unsaturated zone to the saturated zone L3/T
basinseepout Volumetric flow rate of ground-water discharge from the saturated zone to the soil L3/T
zone
sat_stor Volume of water in the saturated zone L3
unsat_stor Volume of water in the unsaturated zone L3
basinsoilmoist Volume of water in capillary reservoirs of the soil zone L3
basingravstor Volume of water in gravity reservoirs of the soil zone L3
basingwstor Volume of water in PRMS ground-water reservoirs (PRMS-only simulation) L3
basinintcpstor Volume of intercepted precipitation in plant-canopy reservoirs L3
basinimpervstor Volume of water in impervious reservoir L3
basinpweqv Volume of water in snowpack storage L3
basininterflow Volumetric flow rate of slow interflow to streams L3/T
basinsroff Volumetric flow rate of the sum of Hortonian (Horton, 1933) and Dunnian surface L3/T
runoff (Dunne and Black, 1970) to streams
strm_stor Volume of water in streams L3
lake_stor Volume of water in lakes L3
obs_strmflow Volumetric flow rate of streamflow measured at a gaging station L3/T
basinszreject Volumetric flow rate of gravity drainage from the soil zone not accepted due to the L3/T
conditions in the unsaturated and saturated zones
basinprefstor Volume of water stores in preferential-flow reservoirs of the soil zone L3
uzf_et Volumetric flow rate of evapotranspiration from the unsaturated and saturated zones L3/T
uzf_infil Volumetric flow rate of gravity drainage to the unsaturated and saturated zones L3/T
uzf_del_stor Change in unsaturated-zone storage L3
net_sz2gw Net volumetric flow rate of gravity drainage from the soilzone to the unsaturated L3/T
and saturated zones
sat_change_stor Change in saturated-zone storage L3
streambed_loss Volumetric flow rate of stream leakage to unsaturated and saturated zones L3/T
sfruz_change_store Change in unsaturated-zone storage under streams L3
gwflow2strms Volumetric flow rate of ground-water discharge to streams L3/T
sfruz_tot_stor Volume of water in the unsaturated zone L3
lakebed_loss Volumetric flow rate of lake leakage to the unsaturated and saturated zones L3/T
lake_change_stor Change in lake storage L3
gwflow2lakes Volumetric flow rate of ground-water discharge to lakes L3/T
basininfil Volumetric flow rate of soil infiltration including precipitation, snowmelt, and L3/T
cascading Hortonian flow
Water Budgets 89
Table 12. Names and descriptions of variables written to the GSFLOW Comma-Separated-Values File at end of each daily time step
for GSFLOW.Continued
[Abbreviations: L3/T, cubic length per time; L3, cubic length]
Variable name Description Units

basindunnian Volumetric flow rate of Dunnian runoff to streams L3/T
basinhortonian Volumetric flow rate of Hortonian runoff to streams L3/T
basinsm2gvr Volumetric flow rate of flow from capillary reservoirs to gravity reservoirs L3/T
basingvr2sm Volumetric flow rate of replenishment of capillary reservoirs from gravity reservoirs L3/T
basininfil_tot Volumetric flow rate of soil infiltration into capillary reservoirs including precipitation, L3/T
snowmelt, and cascading Hortonian and Dunnian runoff and interflow minus infiltration
to preferential-flow reservoirs
basininfil2pref Volumetric flow rate of soil infiltration into preferential-flow reservoirs including L3/T
precipitation, snowmelt, and cascading surface runoff
basindnflow Volumetric flow rate of cascading Dunnian runoff and interflow to HRUs L3/T
basinactet Volumetric flow rate of actual evapotranspiration from HRUs L3/T
basinsnowmelt Volumetric flow rate of snowmelt L3/T
basinhortonianlakes Volumetric flow rate of Hortonian runoff to lakes L3/T
basinlakeinsz Volumetric flow rate of Dunnian runoff and interflow to lakes L3/T
basinlakeevap Volumetric flow rate of evaporation from lakes L3/T
basinlakeprecip Volumetric flow rate of precipitation on lakes L3/T
kkiter Number of iterations for each time step none
A pre-programmed Excel spreadsheet included with GSFLOW (gsflowAnalysis.xls) contains water-budget calculations and
errors at the end of each daily time step for the major storage zones. Volumetric water-budgets and errors at the end of each daily
time step for each of the major storage zones are computed from the continuity equation in the form:
st = st st 1 (115a)
,
(115b)
outt int
et = st + ,
t
where
st the total storage change of the time step ending at time t , in cubic length;
st is the storage at time t , in cubic length;
st 1 is the storage at the previous time step, in cubic length;
et is the budget error at time t , in cubic length per time;
int is the total inflow at time t , in cubic length per time; and
outt is the total outflow at time t , in cubic length per time;
t is one day.
The water budget and error of the surface and near surface storage zone computed by HRUs are determined when the
following GSFLOW variables are substituted for terms in equations 115a-d to compute the change in flow and storage into and
out of the HRUs:
st = smt + gzt + ict + imt + pwt , (116a)
st 1 = smt 1 + gzt 1 + ict 1 + imt 1 + pwt 1 , (116b)

int = pt + ext , and (116c)

outt = ept + eimt + eict + espt + qif t + qsot , (116d)
where
smt and smt 1 are the volumes of water in capillary reservoirs of the soil zone ( variable basinsoilmoist)
at time t and t 1, respectively, in cubic length;
gzt and gzt 1 are the volumes of water in gravity reservoirs of the soil zone ( variable basin ngravstor)
ict and ict 1 are the volumes of intercepted precipitation in plant canopy reservoirs ( variable basinintcpstor)
imt and imt 1 are the volumes of water in impervious reservoir (vvariable basinimpervstor) at time t
and t 1, respectively, in cubic length;
pwt and pwt 1 are the volumes of water in snowpack storage ( variable basinpweqv) at time t and t 1,
respectively, in cubic length;
pt is the volumetric flow rate of precipitation on modeled region ( variable basinpp pt) at time t ,
ext is the volumetric flow rate of ground-water discharge from saturated zon ne to the soil zone
( variable basingw2sz) at time t , in cubic length per time;
ept is the volumetric flow rate of evapotranspiration from pervious areas ( variable basinpervet t)
at time t , in cubic length per time;
eimt is the volumetric flow rate of evaporation from impervious areas ( variable basinimpervevap)
eict is the volumetric flow rate of evaporation of interceppted precipitation ( variable basinintcpevap)
espt is the volumetric flow rate of snowpack sublimation ( variable basinsnowevap) at time t,,
qift is the volumetric flow rate of interflow leaving modeled region ( variable basininter rflow)
at time t , in cubic length per time; and
qsot is the volumetric flow rate of surface runoff leaving the modeleed region ( variable basinsroff)
at time t , in cubic length per time.
Assumptions and Limitations 91
A water budget error is computed for streams by could be incorporated into future versions of GSFLOW, and
summing the change in storage, inflows and outflows alternate governing equations for surface runoff, interflow, and
for all streams. The total change in storage in streams is infiltration that apply at sub-daily time steps could be used,
estimated by subtracting the volume of water in streams such as those used for PRMS storm mode (Leavesley and
(variable strm_stor) at the end of the current time others, 1983).
step from that at the end of the previous time step. The Although streamflow can be routed during sub-daily time
total inflow to streams is computed by summing interflow steps, lateral flow to streams is calculated on a daily time step,
(variable basininterflow), surface runoff (variable which results in daily-averaged hydrographs. Errors can occur
basinsroff), ground-water discharge (variable in streams that have hydrographs that change abruptly over
gwflow2strms), and any specified inflows at the end of time steps much less than 1 day. This is especially true if, in
each time step. The total outflow from streams is computed actuality, a stream floodplain is inundated for a few hours but
by summing stream leakage (variable streambed_loss) the simulated flow is unrealistically contained in the active
and outflow from streams that leave the watershed (variable channel due to daily averaging of the peak flow. Similarly,
basinstrmflow) at the end of each time step. leakage errors can occur during sub-daily hydrographs when
All or some of the gravity drainage from the soil the flows and leakages are averaged over a daily time step.
zone (variable basinsz2gw) may be rejected (variable On the basis of previous ground-water modeling studies
basinszreject) when the ground-water level is above and the test simulations presented herein, 1-day time steps
an altitude of the soil-zone base less half the undulation are sufficiently small to simulate most ground-water flow
depth. The difference between basinsz2gw and conditions. Because the equations used to simulate unsaturated
basinszreject should equal uzf_infil. The water flow beneath the soil zone are analytical, other than the daily-
budget error for the unsaturated zone is computed by summing averaged infiltration rate, their solution is unaffected by the
the storage change (variable uzf_del_stor), inflow 1-day time step used for GSFLOW. Sub-daily fluctuations in
(variable uzf_infil), and outflow (variables uzf_et and infiltration rate at land surface generally would be dampened
uzf_recharge) for the unsaturated zone beneath the soil by the soil zone. Consequently, daily changes in gravity
zone. drainage from the soil zone likely are realistic for simulating
deeper unsaturated flow.
Ideally, the size of finite-difference cells are based on
the spatial resolution required for providing model results that
Assumptions and Limitations can properly address the purpose of the model and available
computer resources. However, other criterion can constrain the
Several assumptions and limitations are associated with size of finite-difference cells. For example, a large cell width
each of the individual modules and packages within GSFLOW. relative to the width of a stream that flows over a cell can
Additionally, similar to other non-linear models, GSFLOW result in model errors and misrepresentation of ground-water
is limited by the possibility of non-convergences among interaction with a stream, especially if other sources or sinks
any or all coupled dependent variables. Non-convergence are present in cells adjacent to the stream. Alternately, the
of GSFLOW can be caused by specifying inappropriate model may not converge if a cell width is equal to or less than
input data or parameters. The number of iterations used in the width of a stream. Similar conditions can occur for other
a GSFLOW simulation can increase when there are large MODFLOW-2005 packages and the associated documentation
changes in the number of finite-difference cells that discharge report for each package provides more details regarding
ground water into the soil zone. Other causes not discussed additional constraints on cell sizes.
nor revealed through model testing also may result in an Another constraint on the size of finite-difference cells
increase in the number of iterations. applies when simulating unconfined aquifers. The model may
not converge if the thickness of a model layer representing
Discretization of Time and Space an unconfined aquifer is thin relative to changes in the
water-table altitude. If the water table fluctuates above and
Because GSFLOW has a computational time step of beneath the bottom of a finite-difference cell then the cell will
1day, all flows and storages are mean daily values. In general, alternate between being a no-flow cell and a variable-head cell
flows near land surface occur more abruptly than flows in the during the iteration loop and this condition can result in non-
subsurface. Consequently, the daily time step will result in convergence of the model.
errors due to time averaging for near-land surface flows such The delineation of the HRUs must be appropriate for
as surface runoff, infiltration, and interflow, and streamflow the module algorithms. In PRMS, each HRU is assumed
and streambed leakage. However, simulating the sub-daily to be homogeneous with respect to both input parameters
behavior of these processes is not warranted because surface and computation of flows and storages, similar to a finite-
runoff, infiltration, and interflow are handled conceptually difference cell.
in this initial version of GSFLOW. Sub-daily time steps
Canopy Zone Streams

Incomplete and dated sources of land-use and land- Numerous assumptions have been made to simplify the
cover information can cause errors in GSFLOW estimates of Streamflow-Routing Package. First, flow is one-dimensional
precipitation interception and transmission of shortwave solar in the direction of the channel. Furthermore, only the channel
radiation. GSFLOW does not have a plant growth module so slope and friction loss terms in the momentum equation are
plant cover estimates, particularly in agricultural areas, may be represented such that both kinematic and dynamic waves are
problematic. neglected for the steady-flow option, and dynamic waves
are neglected in the transient-flow option. It is assumed
that momentum can be represented by the Mannings
Land-Surface Precipitation and Temperature equation for steady flow, and that stream depth corresponds
to normal depth (Streeter and Wylie, 1985; p. 470-471).
The PRMS land-surface modules are susceptible to
It also is assumed that the stage-discharge relation can be
errors in the input time series of measured precipitation and
represented by a single-valued function. Backwater effects
temperature. These errors can have serious consequences on
are not considered, and flow occurs in one direction only.
simulation of both the volume and timing of water-balance
Steady streamflow routing (steady-uniform flow) assumes
components. These errors generally are of three forms:
piecewise steady (non-changing in discrete time periods),
(1)uncertainty in measurement; (2) errors in distribution of
uniform (non-changing in location), and constant-density
precipitation and temperature; and (3) errors in determination
streamflow, such that during all times, volumetric inflow and
of the form of precipitation. Distribution errors generally will
outflow rates are equal and no water is added to or removed
result in annual or seasonal biases, whereas precipitation-form
from storage in the surface channels. Unsteady routing using
errors generally affect individual events.
the kinematic-wave equation assumes flow disturbances only
can move in the downstream direction with no hydrograph
Soil Zone attenuation (diffusion of flood waves is neglected) and that
energy provided by the channel slope is balanced by frictional
The soil zone is represented as a series of reservoirs that losses along the channel bottom. Finally, flow is assumed to be
store and transfer water among different reservoirs within the turbulent.
soil zone and to various surface and ground-water reservoirs In addition to assumptions regarding flow in stream
(including MODFLOW finite-difference cells). Equations are channels, assumptions are made with regard to streambed
used to relate the transfer rate from a reservoir in the soil zone leakage. Because this leakage is calculated on the basis of the
to storage in the reservoir. The main assumption used in the Dupuit assumption (Anderson, 2005), head loss is neglected
soil zone is that flow into and out and changes in storage occur between the top of the streambed and the point where ground-
uniformly throughout an HRU during each daily time step. water head is computed beneath the stream. Consequently,
This means the parameters that define flow and storage in the errors in calculated leakage can occur if the streambed
soil zone remain constant in each HRU during a simulation. conductance values do not include sufficient resistance to
Thus, the delineation of HRUs is critical to a simulation in reflect the effects of vertical head loss between streams and
that it defines the flow into and out of a soil zone and affects aquifers. Leakage through streambeds that are not in direct
gravity drainage to the unsaturated zone, ground-water contact with ground water is represented assuming a unit
discharge from the saturated zone, and runoff and interflow to hydraulic gradient beneath the streambed, and leakage occurs
streams and lakes. only in the vertical direction. Finally, where streams are
Other important assumptions used to define the different small relative to the finite-difference cell, substantial errors
reservoirs within the soil zone also may limit applicability. in leakage can occur if the streambed conductance is not
An important assumption is that freezing and thawing of the sufficiently small.
soil zone does not change parameters that affect flow and
storage in the soil zone during a simulation. Another important
assumption is that the capillary reservoir is represented by a Lakes
constant value of root depth for each HRU during a simulation.
A lake is represented by a completely mixed volume
Root depth is determined on the basis of the principal plant
of water within the Lake Package. Because water is not
type within each HRU. Because plants grow and die, the size
routed within lakes, water entering a lake is assumed to
and types of plants in an HRU vary with time, and the root
be instantaneously mixed with the existing water in the
depth as well as evapotranspiration from an HRU also may
lake. Multiple lakes can combine into a single lake, and
vary in space and time. Finally, another assumption is that
when this occurs the combined lakes are represented by
perturbations such as animal burrows, decaying roots, and leaf
an instantaneously mixed volume of water. An important
litter that produce preferential flow remains constant within
assumption is that the lake occupies the volume of finite-
each HRU during a simulation.
difference cells within the MODFLOW grid. Thus, the
Input and Output Files 93
position and spatial extent of the lake area is equal to the area when the water table rises into the soil zone. If the water table
of inundated no-flow cells used to represent the lake defined is at or above the soil-zone base, evapotranspiration loss from
in the MODFLOW Discretization File. Consequently, cells ground water occurs at the net potential ET rate. If the depth
used to represent a lake should be discretized appropriately to of the water table is beneath the evapotranspiration extinction
realistically represent the area and volume of a lake. depth, evapotranspiration from the water table ceases; and
Similar to the Streamflow-Routing Package, leakage between these limits, evapotranspiration from the water table
between lakes and ground water is calculated on the basis of varies linearly with water-table altitude.
the Dupuit assumption. Consequently, head loss between the
top of the lakebed and the point where the ground-water head
is calculated beneath the lake cell is neglected. Unsaturated Zone
Flow in the unsaturated zone is simulated by the
Ground Water Unsaturated-Zone Flow Package that solves a vertical, one-
dimensional form of Richards equation and neglects capillary
Assumptions associated with GSFLOW for simulating pressure gradients. An assumption is that the unsaturated zone
ground-water flow are standard assumptions corresponding to is homogeneous in the vertical direction or else meaningful
the saturated ground-water flow equation. These assumptions averages or integrated properties can be specified for each
are that ground-water flow is laminar and inertial forces, vertical column of unsaturated material. Because capillary
velocity heads, temperature gradients, osmotic gradients, and pressure gradients are neglected, the capillary fringe is not
chemical concentration gradients are not considered. Soils and simulated and evapotranspiration in the unsaturated zone
geologic formations are assumed to be linearly and reversibly cannot cause upward flow. Sediment in the unsaturated zone
elastic and mechanically isotropic; horizontal components cannot swell or deform in any direction. Evapotranspiration
of deformation are assumed small compared to vertical is limited only by the evapotranspiration extinction depth and
deformations; and ground water is instantaneously released evapotranspiration extinction water content.
from or taken into storage.
Additional assumptions are made in the discretization of
the ground-water system into a finite-difference grid. These
assumptions are that the principal directions of anisotropy are Input and Output Files
assumed to be aligned with the coordinate directions of the
Data input for a GSFLOW simulation is specified in
grid, and each model layer corresponds to a distinct aquifer or
several files that must be prepared by the model user prior
permeable horizon. The finite-difference approximations of
to a simulation. Outputs from a GSFLOW simulation are
the ground-water flow equation in MODFLOW were derived
written to several files whose names, and in some cases whose
assuming that the grid is rectangular; however, grids can be
formats, also are specified by the user as part of the simulation
non-rectangular in the vertical direction (Harbaugh, 2005,
input data. The input and output files needed for a GSFLOW
p. 3-15) resulting in some errors in the computed ground-
simulation are described in this section.
water conditions. Because hydraulic properties are uniform
within individual cells, averages or integrated properties are
assumed to represent the aquifer material contained in a single Input Files
finite-difference cell. Flow is parallel to the coordinate axis
within a finite-difference cell, nodes are at the center of cells, Several input files are needed for a GSFLOW simulation.
and discrete changes in hydraulic conductivity occur at cell These files are the GSFLOW Control File, the PRMS Data
boundaries. Simulations involving implicit confining units File, the PRMS Parameter File, the MODFLOW Name
(Quasi-Three-Dimensional approach) assume the confining File, and the input files for each MODFLOW-2005 package
unit makes no measurable contribution to the horizontal specified for use in the Name File. Because GSFLOW is
conductance or the storage capacity of either model layer based on two different models, the input format for each file
(Bredehoeft and Pinder, 1970). The only effect of an implicit depends on the genesis of the file. The format of the GSFLOW
confining unit is to restrict vertical flow between finite- Control File is based on Modular Modeling System input
difference cells. structures, as are the PRMS Data File and PRMS Parameter
Similar to ground-water discharge to streams, ground- File. Input files for each MODFLOW-2005 package are based
water discharge to the soil zone is simulated using the on specifications defined for each package. Brief descriptions
Dupuit assumption and vertical head loss is neglected. Errors of each input file follow; detailed descriptions of the input data
caused by the Dupuit assumption increase as the vertical and required for each input file are provided in appendix 1.
horizontal dimensions of the finite-difference cell increase.
Recharge from the soil zone directly to ground water decreases
GSFLOW Control File MODFLOW-2005 Input Files

The GSFLOW Control File is used to set basic The MODFLOW-2005 packages input files consist of
administrative data values related to the PRMS and GSFLOW data values that are related to the saturated and unsaturated
modules for a GSFLOW simulation, such as specification of subsurface zones of a GSFLOW simulation (that is,
model input and output file names and simulation starting beneath the soil zone), as well as the stream and lake
and ending dates. There are four types of parameters that are components of a simulation, and solver and output control
specified in the file: (1) those related to model execution; (2) input files. Unlike the PRMS modules, in which most
those related to model input; (3) those related to model output; model-input data are specified in a single file (the PRMS
and (4) those related to initial conditions for the PRMS model. Parameter File), MODFLOW-2005 packages require model
The name of the GSFLOW Control File can be specified input to be gathered from different files. For each of the
on the command line of a GSFLOW execution. For example, if MODFLOW-2005 packages that are used in a GSFLOW
the GSFLOW executable is named gsflow.exe and the name simulation, there are one or more corresponding input files.
of the Control File is sagehen.control then the command line Table 1 summarizes the MODFLOW-2005 packages that can
to execute the code would be: be used with GSFLOW; specific input instructions for each
gsflow.exe sagehen.control input file are provided in appendix 1.
If no control-file name is given after gsflow.exe, GSFLOW The MODFLOW Name File is used to specify the input
will look for a Control File named gsflow.control in the users and output files associated with each MODFLOW-2005
current directory. package that will be used in a GSFLOW simulation. Each
line in the Name File defines one file and specifies a file type
(Ftype), unit number (Nunit), and name (Fname) as shown
PRMS Data File in figure 30. The parameter Ftype specifies the package or
The PRMS Data File is used to define available the purpose of a file. Values of Ftype correspond to a major
measured or simulated climate and streamflow data for use in MODFLOW option that is activated with the appropriate
a GSFLOW simulation. These data are specified as a time- line in the Name File. Most packages are major options and
series that defines the time extent of possible simulations. The require an input file, so, in general, each package is activated
forcing functions of the GSFLOW simulationprecipitation, by at least one line in the Name File. For example, if a well
maximum and minimum air temperature, and solar is simulated, a file having a file type of WEL is required in
radiationare specified in the file, as well as the form of the Name File, and an input file in the Well Package input
precipitation (unknown, snow, or rain), pan evaporation rates, format is required. Values of Ftype are character strings that
and measured streamflow. The name of the Data File is set can be a package acronym (for example, BAS6, UZF, or
using parameter data_file in the GSFLOW Control File. SFR) or a functional description (for example, LIST or
Time-series of daily precipitation and maximum and minimum DATA) and can be specified using uppercase or lowercase
air temperature are required for a GSFLOW simulation; letters. Nunit is an integer value used to specify the Fortran
additional data are optional. unit number used to read from or write to the named file.
Fname is a character string that identifies the file either as
a full pathname or relative pathname (a relative pathname is
PRMS Parameter File specified from the users current directory). For example, if a
well is simulated, a line in the Name File must specify Ftype
The majority of input parameters used by the PRMS equal to WEL, an integer value for Nunit that specifies
and GSFLOW modules are specified in the PRMS Parameter a valid Fortran unit number, and a character string, Fname,
File. The file also is used to specify the dimensions of several identifying the name and location of the Well package input
PRMS and GSFLOW parameters, such as the number of file.
measured precipitation, air temperature, and streamflow- A MODFLOW-2005 package is inactive if the Name File
gaging stations for which data are specified in the PRMS does not include a line having the options Ftype. An option
Data File and for spatial parameters, such as the number can be turned off either by removing the appropriate line
of hydrologic response units (HRUs) and stream segments. from the Name File or by placing a pound sign (#) at the
Input instructions for the PRMS Parameter File are provided beginning of a line. The MODFLOW-2005 Input Instructions
in appendix 1. The name of the Parameter File is set using section in appendix 1 defines the available values for Ftype
parameter param_file in the GSFLOW Control File. and Nunit. The Name File must include definitions for two
files required for each simulation involving MODFLOW-2005,
the Basic Package File (BAS6) and the Discretization File
(DIS). The Basic Package File specifies, among other values,
LIST 26 .\output\sagehen.out Listing File for MODFLOW-2005

BAS6 8 .\input\sagehen.bas Basic Package File
oc 9 .\input\sagehen.oc Output Control Option File
dis 11 .\input\sagehen.dis Discretization File
lpf 12 .\in put\sagehen.lpf Layer-Property Flow Package File
pcg 13 .\input\sagehen.pcg PCG Solver Package File
uzf 14 .\input\sagehen.uzf Unsaturated-Zone Flow Package File
sfr 15 .\input\sagehen.sfr Streamflow-Routing Package File
gage 16 .\input\sagehen.gag Gage Package File
data 58 .\output\sagehen_head.out Head output file
data 65 .\output\sagehen_uzf1.out Unsaturated zone output for row 40, column 10
data 68 .\output\sagehen_uzf4.out Unsaturated zone output summed for model
dat a 80 .\output\sagehen_sfrseg12.out Stream gaging station output, segment 12, reach 4
data 81 .\output\sagehen_sfrseg13.out Stream gaging station output, segment 13, reach 11
Ftpye Nunit Fname (includes file location)
Figure 30. MODFLOW Name File used for the Sagehen Creek watershed example problem.
which finite-difference cells are variable-head in the model DIS), or a file-unit flag can be set in the Discretization File
domain and the initial heads in each finite-difference cell. to indicate that the values for the top altitude should be read
The Discretization File specifies data values such as the from a file defined in the Name File with Ftype DATA or
dimensions of the MODFLOW finite-difference grid (number DATA(BINARY).
of layers, rows, and columns), data units, and altitude of each The full pathname of the MODFLOW Name File,
cell in the top layer of the model. An example Name File for meaning the directory location and name of the Name File,
the Sagehen Creek watershed example problem described is specified in the GSFLOW Control File using parameter
in section Example Problem is shown in figure 30. The modflow_name. Similar to the original MODFLOW-2005
MODFLOW-2005 Input Instructions in appendix 1 define all construction, the Name File pathname also can be specified
file types that can be used in a GSFLOW simulation. in a MODFLOW Batch File that consists of a single line
In addition to specifying input data by way of the actual that specifies the Name File pathname beginning in column
file, MODFLOW-2005 also allows model input to be entered 1 and having no spaces (pathnames that include spaces must
from external formatted or unformatted files. Values of Ftype be quoted, for example C:\Program Files\gsflow\data\
specified in the Name File as DATA and DATA(BINARY) sagehen.nam or C: \Documents and Setting\userid\gsflow\
(in either upper or lower case) define input or output files data\sagehen\sagehen.nam). If a file named modflow.bf is
whose file unit numbers, Nunit, can be referenced by found by GSFLOW in the users current directory, it will be
other MODFLOW-2005 packages. DATA file types are used as the MODFLOW Batch File. Otherwise, the value of
for formatted (text) files, whereas DATA(BINARY) file the control parameter modflow_name is used; the default
types are for binary (unformatted) files. The DATA and value is modflow.nam. For example, if a GSFLOW simulation
DATA(BINARY) file types can be specified for many is initiated from the users current directory that contains a
files. An example is the input of an array of variables, such as file named modflow.bf, GSFLOW will read the first line to
hydraulic conductivity, by row and column grid indices. When define the pathname of the MODFLOW Name File. For this
using external files for model input, the model input values execution scenario and for the Sagehen Creek watershed
will not be read from the package input files themselves. example problem described in this report that uses the Name
Rather, the package file specifies that the input values must File named sagehen.nam, the MODFLOW Batch File consists
be read from a different file by specifying the Fortran unit of the following line:
number listed in the Name File for that input. For example, ./sagehen.nam
the values for the top altitude can be specified directly in
the Discretization File (that is, in the file whose FTYPE is
nv07tac-4124_fig30
Output Files section Water Budgets. The file is generated when parameter
gsf_rpt is set to 1 in the GSFLOW Control File; the name
Several output files can be generated during a GSFLOW of the file is specified by parameter csv_output_file
simulation. The following output file descriptions are in the GSFLOW Control File, which defaults to the name
organized by model sourceGSFLOW, PRMS, then gsflow.csv. The first line of the file (the header) lists the names
MODFLOW. of each output variable in the sequence for which values are
included in each subsequent line; these names can be used as
spreadsheet column headers. The remaining lines of the file
GSFLOW provide the simulation time-step date as a two-digit month, a
Two GSFLOW-specific output files can be optionally two-digit day, and a four-digit year, each separated by a / or
generated during a simulation. The first, the GSFLOW in abbreviated form as MM/DD/YYYY. The date is followed
Water-Budget File, contains tabular summaries of the water by model-calculated values for each variable defined in the
budget computed by GSFLOW at user-specified frequency. header line. All values are separated by commas. One line of
The second, the GSFLOW Comma-Separated-Values (CSV) values is written for each simulation time step.
File, contains simulation results for each time step; these
can be used to compute water budgets for each GSFLOW PRMS
storage reservoir and a mass balance for the simulation.
A third non-optional GSFLOW output file, named the Three PRMS output files can be generated as part of
GSFLOW Log File or gsflow.log, is generated in the users a GSFLOW simulation: the PRMS Water-Budget File of
current directory during a simulation. This Log File contains summary tables generated at a user-specified frequency, the
information that describes the GSFLOW model, any warning PRMS Statistic Variables File of user-selected variable values
or error messages, simulation time period, execution time, and for each time step, and the PRMS Animation Variables File of
iteration statistics for each time step and simulation, and any user-selected array variables for each time step.
information that was written to the users computer display. The PRMS Water-Budget File (fig. 32) provides summary
The GSFLOW Water-Budget File (fig. 29) is a text file table(s) of the water budget for a PRMS simulation. The
that provides summary tables of flows into and out of the pathname of the Water-Budget File is specified by parameter
watershed and storage changes within the watershed by zones model_output_file in the GSFLOW Control File. Three
(land surface, soil, unsaturated, saturated). The name of the types of summary tables are available, depending on the value
file is specified by parameter gsflow_output_file in specified for parameter print_type in the PRMS Parameter
the GSFLOW Control File. The frequency at which the tables File. The first is a listing of the observed and predicted flow,
are generated is specified in the GSFLOW Control File using which is generated when print_type is set to 0. The
parameter rpt_days, which defaults to every 7 days within second is a table of water-balance computations and includes
the simulation time period beginning with the table generated the watershed-weighted averages for net precipitation,
for simulation day 7. The first line of each summary table in evapotranspiration from all sources, storage in all reservoirs,
the output file provides the calendar date and corresponding and the simulated (P-Runoff) and measured (O-Runoff) flows.
time step of the simulation. The next line provides the This report is generated when print_type is set to 1. The
MODFLOW stress period and time step within the stress third is a detailed summary of the rainfall, outflow, and state
period. The remainder of each summary table is divided into variables, and is generated when print_type is set to 2.
four columns. The last two columns list budget terms for the Any of the summaries may be requested in any combination of
current time step in units of volumetric flow rate per time (that the available time incrementsdaily, monthly, yearly, or total
is, flow rates). The first two columns list cumulative flows for the simulation. The frequency of output is specified by
(units of volume) for all time steps up to and including the parameter print_freq in the PRMS Parameter File.
current time step. All values are reported in the MODFLOW The PRMS Statistic Variables File (fig. 33) is a text
cubic length unit. Details of the GSFLOW water budget are file that provides PRMS model output that can be used with
described in section Overall Water Budget. visualization and statistics programs. The file is generated
The GSFLOW Comma-Separated-Values (CSV) File is a when parameter statsON_OFF is set to 1 in the GSFLOW
text file that provides GSFLOW simulation results in a format Control File. The name of the file is set by control parameter
that can be easily imported to spreadsheet programs. Figure31 stat_var_file. The first line of the file is the number
shows the first three lines of the CSV File generated for the of variable values that are written in the file; this value is
Sagehen Creek watershed example problem described in this specified using control parameter nstatVars. The next
report. Each line shown in figure 31 is wrapped over several group of lines (nstatVars in number) lists the names
lines to stay within the margins of the page. Values included and array index of each output variable; the output variables
in this file can be used to compute the budgets specified in
Date,basinppt,basinpervet,basinimpervevap,basinintcpevap,
basinsnowevap,basinstrmflow,basinsz2gw,basingw2sz,gw_inout,
stream_leakage,uzf_recharge,basinseepout,sat_stor,unsat_stor,
basinsoilmoist,basingravstor,basingwstor,basinintcpstor,
basinimpervstor,basinpweqv,basininterflow,basinsroff,strm_stor,
lake_stor,obs_strmflow,basinszreject,basinprefstor,uzf_et, Header
uzf_infil,uzf_del_stor,net_sz2gw,sat_change_stor,streambed_loss, line
sfruz_change_stor,gwflow2strms,sfruz_tot_stor,lakebed_loss,
lake_change_stor,gwflow2lakes,basininfil,basindunnian,
basinhortonian,basinsm2gvr,basingvr2sm,basininfil_tot,
basininfil2pref,basindnflow,basinactet,basinsnowmelt,
basinhortonianlakes,basinlakeinsz,basinlakeevap,basinlakeprecip,
kkiter
10/01/1980, 0.0000000E+00, 0.7730693E+04, 0.0000000E+00,
0.0000000E+00, 0.0000000E+00, 0.6287971E+04, 0.0000000E+00,
0.1590413E+05, -0.6488422E+03, -0.6287971E+04, 0.2106252E+05,
0.1590659E+05, 0.1638903E+09, 0.9435054E+07, 0.7787383E+05,
0.1703915E+02, 0.0000000E+00, 0.0000000E+00, 0.0000000E+00,
0.0000000E+00, 0.0000000E+00, 0.0000000E+00, 0.7615586E+03, First
0.0000000E+00, 0.5627124E+04, 0.0000000E+00, 0.0000000E+00, data
0.0000000E+00, 0.0000000E+00, -0.2102739E+05, 0.0000000E+00,
-0.1760261E+04, 0.6480000E+02, 0.0000000E+00, 0.6505124E+04,
line
0.5966707E+02, 0.0000000E+00, 0.0000000E+00, 0.0000000E+00,
0.0000000E+00, 0.0000000E+00, 0.0000000E+00, 0.0000000E+00,
0.1588709E+05, 0.0000000E+00, 0.0000000E+00, 0.0000000E+00,
0.7730693E+04, 0.0000000E+00, 0.0000000E+00, 0.0000000E+00,
0.0000000E+00, 0.0000000E+00, 7
10/02/1980, 0.0000000E+00, 0.8015101E+04, 0.0000000E+00,
0.0000000E+00, 0.0000000E+00, 0.6194819E+04, 0.0000000E+00,
0.1584995E+05, -0.6488286E+03, -0.6194818E+04, 0.2092071E+05,
0.1585241E+05, 0.1638886E+09, 0.9414168E+07, 0.8572385E+05,
0.1839345E+01, 0.0000000E+00, 0.0000000E+00, 0.0000000E+00,
0.0000000E+00, 0.0000000E+00, 0.0000000E+00, 0.7564084E+03, Second
0.0000000E+00, 0.5627124E+04, 0.0000000E+00, 0.0000000E+00, data
0.0000000E+00, 0.0000000E+00, -0.2088586E+05, 0.0000000E+00, line
-0.1743168E+04, 0.6480000E+02, 0.0000000E+00, 0.6489369E+04,
0.5966872E+02, 0.0000000E+00, 0.0000000E+00, 0.0000000E+00,
0.0000000E+00, 0.0000000E+00, 0.0000000E+00, 0.0000000E+00,
0.1586515E+05, 0.0000000E+00, 0.0000000E+00, 0.0000000E+00,
0.8015101E+04, 0.0000000E+00, 0.0000000E+00, 0.0000000E+00,
0.0000000E+00, 0.0000000E+00, 6
A data line is added for each one-day time step
Figure 31. Three lines of a GSFLOW Comma-Separated-Values (CSV) File that shows variables that can be used for
water-budget calculations for each time step.
that are listed are specified using control parameter in the GSFLOW Control File. The name of the file is set by
statVar_name. The array index for variables that are parameter ani_output_file. The first group of lines
scalar, meaning a single value, is set to 1. The remaining lines in the file, starting with pound characters (#) describes the
provide the model-calculated values of each variable for each format of the file (that is, provide Meta data that define the
simulation time step. These data lines have the following file format and contents); these lines can be used by external
order: model time-step number, year, month, day of month, programs to reformat the file. The first line beyond the Meta
hour, minute, second, and each variable value in the order data is a tab-separated list of names of the output variables
specified by the list of variable names. Each value is separated whose values are provided in a column in each data line.
by a space. These output variables are specified using parameter
The NV07tac-4124_fig31
PRMS Animation File (fig. 34) is a text file that aniOutVar_names. The next line is a tab-separated list of
provides PRMS model results as a time-series of spatial the field width and data type, defined as a single text string,
arrays that can be used with animation programs. The file of each output variable in the same sequence as the list of
is generated when parameter aniOutON_OFF is set to 1 variable names. Each value in the list is single character
appended to an integer value defining the field width.
Surface Water and Energy Budgets Simulated by PRMS Version 2.3730 2008-01-18
Start time: 1980/10/01 00:00:00
End time: 1996/09/30 00:00:00
Sum of HRUs: 6782.50 Basin_area: 6782.50
1 Year Month Day Precip ET Storage P-Runoff O-Runoff
(inches) (inches) (inches) (inches) (inches)
1981. 27.189 22.711 0.794 9.266 6.181
1982. 76.778 31.551 4.142 34.977 36.819
1983. 64.524 33.072 2.415 31.771 38.374
1984. 52.146 33.120 0.743 22.507 26.168
1985. 29.538 24.228 1.103 11.245 11.260
1986. 60.447 32.247 2.695 23.094 27.044
1987. 17.887 19.568 0.295 7.768 6.438
1988. 21.214 19.986 0.357 6.016 4.140
1989. 43.851 27.848 2.042 12.659 12.683
1990. 29.970 25.613 1.705 7.398 6.491
1991. 25.841 24.300 0.529 6.032 5.574
1992. 21.591 17.620 0.364 5.928 3.847
1993. 55.061 29.138 0.368 18.250 18.886
1994. 20.340 18.582 0.639 6.248 4.182
1995. 69.736 31.238 0.544 28.132 29.493
1996. 52.824 28.632 0.832 21.199 22.769
*************************************************************
Total for run 668.938 419.456 0.832 252.492 260.352
Figure 32. PRMS Water-Budget File that shows annual and simulation-run summary tables
(parameters print_type set to 1 and print_freq set to 3 in the PRMS Parameter File) for
the Sagehen Creek watershed example problem.
2 1
basin_cfs 1 2
runoff 1 3
1 1980 10 1 0 0 0 2.632515 2.30 4
2 1980 10 2 0 0 0 2.580404 2.30 4
3 1980 10 3 0 0 0 2.554979 2.30
4 1980 10 4 0 0 0 2.539002 2.30
5 1980 10 5 0 0 0 2.528993 2.30
6 1980 10 6 0 0 0 2.521560 2.30
7 1980 10 7 0 0 0 2.516836 2.20
8 1980 10 8 0 0 0 2.513003 2.20
9 1980 10 9 0 0 0 2.509280 2.30
10 1980 10 10 0 0 0 2.503823 2.30
11 1980 10 11 0 0 0 2.497788 2.30
12 1980 10 12 0 0 0 3.187714 4.00
13 1980 10 13 0 0 0 3.707187 2.90
14 1980 10 14 0 0 0 3.428297 3.20
15 1980 10 15 0 0 0 3.163299 3.10 4
...
366 1980 10 15 0 0 0 3.163299 3.10 4
1 Two variables are included.
2 Simulated streamflow and one column of numbers, respectively.
3 Daily mean streamflow at outlet gage and one column of numbers, repsectively.
4 Time step, year, month, day, time (HH MM SS), simulated streamflow,
and daily mean streamflow, respectively. Line 4 is repeated for each daily time
step of simulation.
Figure 33. Selected output from the PRMS Statistic Variables (statvar)
File that lists simulated (basin_cfs) and measured (runoff)
nv07tac-4124_fig32
streamflow for the Sagehen Creek watershed example problem.
# Begin DBF
# timestamp,#FIELD_ISODATETIME,19,0
# nhru,#FIELD_DECIMAL,10,2
# pkwater_equiv,#FIELD_DECIMAL,10,2
# soil_moist_pct,#FIELD_DECIMAL,10,2
# hru_actet,#FIELD_DECIMAL,10,2
# hru_ppt,#FIELD_DECIMAL,10,2
# infil_tot,#FIELD_DECIMAL,10,2
# snowmelt,#FIELD_DECIMAL,10,2
# sroff,#FIELD_DECIMAL,10,2
# tavgf,#FIELD_DECIMAL,10,2
# End DBF
#
timestamp nhru pkwater_equiv soil_moist_pct hru_actet hru_ppt ...
19d 10n 10n 10n 10n 10n ...
1980-10-01:00:00:00 1 0.000e+000 2.446e-002 1.037e-002 0.000e+000 ...
1980-10-01:00:00:00 2 0.000e+000 3.593e-002 1.458e-002 0.000e+000 ...
1980-10-01:00:00:00 3 0.000e+000 2.959e-002 1.106e-002 0.000e+000 ...
1980-10-01:00:00:00 4 0.000e+000 2.480e-002 9.699e-003 0.000e+000 ...
1980-10-01:00:00:00 5 0.000e+000 4.660e-002 1.741e-002 0.000e+000 ...
1980-10-01:00:00:00 6 0.000e+000 2.442e-002 1.054e-002 0.000e+000 ...
1980-10-01:00:00:00 7 0.000e+000 2.453e-002 9.897e-003 0.000e+000 ...
1980-10-01:00:00:00 8 0.000e+000 2.593e-002 1.096e-002 0.000e+000 ...
1980-10-01:00:00:00 9 0.000e+000 4.020e-002 1.519e-002 0.000e+000 ...
1980-10-01:00:00:00 10 0.000e+000 2.766e-002 1.151e-002 0.000e+000 ...
1980-10-01:00:00:00 11 0.000e+000 2.484e-002 9.990e-003 0.000e+000 ...
1980-10-01:00:00:00 12 0.000e+000 2.607e-002 8.646e-003 0.000e+000 ...
1980-10-01:00:00:00 13 0.000e+000 2.518e-002 7.445e-003 0.000e+000 ...
1980-10-01:00:00:00 14 0.000e+000 2.519e-002 1.053e-002 0.000e+000 ...
1980-10-01:00:00:00 15 0.000e+000 4.414e-002 1.810e-002 0.000e+000 ...
...
Variables listed at beginning of file are printed for each time step until end of simulation.
Gaps in the output listing are indicated by an ellipsis
Figure 34. Selected output from a PRMS Animation File that lists snowpack-water equivalent, soil moisture,
evapotranspiration, and precipitation for each HRU.
Name File. Output-file types are categorized according to

The single character designates the data type using the
the MODFLOW-2005 process or package that writes to the
following scheme: d=date, n=number. The remaining lines
file. Output file types are defined using the variable Ftype
contain the date and corresponding index number within the
specified in the Name File, and output-file types common
spatial feature dimension and variable values in the order
to the Ground-Water Flow process are LIST, DATA
specified by the list of variable names. A date value (or
DATA(BINARY), and GAGE. The frequency and format
timestamp) is output as a 19-character string as a four-digit
for which output is written to all output files is specified by the
year, one- or two-digit month, two-digit day, two-digit hour,
user in the Output Control Option File of the Basic Package.
two-digit minute and two-digit second. Year, month and day
The Output Control Option File has its own input file, which
are separated with a dash whereas hour, minute, and second
is specified using file type OC in the Name File (see fig. 30).
are separated with a colon. The abbreviated form is: YYYY-
The Output Control Option File can be excluded by omitting
MM-DD HH:MM:SS. The index number is an integer value
the file type OC from the Name File, which will result in
using the next 5 characters. Data values are numbers written in
output being written at the end of each stress period as 10-digit
a 10-character exponential format. All values are separated by
floating-point numbers with 3 digits following the decimal
tab characters.
point to the Listing File.
The main output file for MODFLOW-2005 is called
MODFLOW-2005 the Listing File, to which much of the MODFLOW input
data and calculated results are written during a simulation.
MODFLOW-2005 output files are generated by The primary model-calculated output for the Ground-Water
specifying the file type (Ftype), unit number (Nunit), Flow Process includes the ground-water head distribution,
directory address, and name (Fname), in the MODFLOW
nv07tac-4124_fig34
drawdown distribution, the distribution of variable-head cells, water table. Finally, recharge, evapotranspiration, ground-
and the overall volumetric budget. The Listing File is used for water discharge, and water-content profiles for selected finite-
all MODFLOW-2005 simulations and must be the first file difference cells can be written to files as formatted columns of
that is specified within the Name File. In addition to calculated values for every time step.
results, input data specified for any MODFLOW-2005
package can be written (echoed) to the Listing File to verify
proper input data specification. Budget information for
MODFLOW-2005 packages, such as the Lake, Streamflow- Example Problem Sagehen Creek
Routing, and Unsaturated-Zone Flow Packages, also can be Watershed
written to the Listing File.
Output from the Ground-Water Flow Process also can be This section describes an example GSFLOW model
separated into output files other than the Listing File. Most for simulation of the Sagehen Creek watershed. All files,
packages can write calculated results as formatted text to files including documentation, input and output data, source code,
that are specified as type DATA, or as binary (unformatted) and the GSFLOW executable model necessary to run this
data to files that are specified as type DATA(BINARY). example are available from the GSFLOW web page (Internet
Calculated results for the Lake and Streamflow-Routing address is listed in the Preface to this report on page ix).
Packages can be written to output files of type DATA The derived and calibrated input files provided with
using the Stream Gaging Station or Gage Package. Each the Sagehen Creek watershed example problem, graphs and
MODFLOW-2005 package generates its own output, and tables, and other simulation results described in this section
the details regarding output data for the Lake, Streamflow- should not be interpreted for assessing water-resource assets
Routing, and Unsaturated-Zone Flow Packages supported by in the Sagehen Creek watershed. Although the models
GSFLOW are briefly described here. A MODFLOW-2005 described herein were calibrated with available data, few data
package can write detailed information using the output option are available regarding ground-water heads in the watershed;
flags specified in the input files for each package (appendix 1). consequently, these data sets and simulation results are
The Lake and Streamflow-Routing Packages can write provided for illustrative purposes only.
calculated results to formatted files with type DATA, or to
unformatted files with type DATA(BINARY). Additionally,
calculated results for a particular lake or stream can be written Description of Sagehen Creek Watershed
as formatted text to files with type DATA, as defined in the
Name File, using the Gage Package output options. There are Sagehen Creek is a USGS Hydrologic Benchmark
four options that can be specified in the Gage Package File Network Basin located on the east slope of the northern Sierra
when using the Lake Package. These options can be used to Nevada (fig. 35). The watershed drains an area of 27 km2
write columns of values for stage, volume, runoff, inflows and ranges in altitude from 1,935 to 2,653 m. Geology of the
from and outflows to streams, and lakebed conductance for Sagehen Creek watershed consists of granodiorite bedrock
every time step. There are seven options that can be specified overlain by andesitic, tertiary volcanics, which are overlain by
in the Gage Package File when using the Streamflow-Routing till and alluvium composed of granodiorite and andesite clasts
Package. These options can be used to write columns of values (Burnett and Jennings, 1965; Rademacher and others, 2005).
for stage, outflow, runoff, streambed conductance, hydraulic Quaternary gravels also are present on the northwestern side
gradients beneath the streambed, and flow diversions every of the watershed (Burnett and Jennings, 1965).Very little is
time step. If unsaturated flow beneath streams is active, then known regarding the depths and thickness of these different
changes in unsaturated zone storage and recharge also can be geologic formations. However, the volcanics were assumed
written as well as profiles of water content. The Gage Package to make up the principal component of the watershed aquifer
input instructions in appendix 1 describe these options in more and the volcanics were assumed to range in thickness between
detail. 50 and 300 m. Alluvium is thin or nonexistent in the upper
The Unsaturated-Zone Flow Package can write recharge, parts of the watershed and thickens in the lower parts and near
evapotranspiration, and ground-water discharge to formatted stream channels (Burnett and Jennings, 1965). Alluvium was
files with type DATA, or to unformatted files with type assumed to range in thickness between 0 and 10 m.
DATA(BINARY) as specified in the MODFLOW Name Fourteen perennial springs are known in the watershed.
File using the values of output option flags. Another option The chemical and isotopic composition of these springs were
is to write time-series output for infiltration, recharge, analyzed by Rademacher and others (2005) to estimate the age
evapotranspiration, and ground-water discharge that are the of ground water seeping into Sagehen Creek. Ground-water
sum of values for all finite-difference cells that contain the discharge into the creek from springs and leakage through
Example Problem Sagehen Creek Watershed 101
SIERRA NEVADA Elevation: 2,653 meters
Model boundary
k
ee
Cr
en
h
ge
Sa
Elevation: 1,935 meters

MAP
AREA
Truckee NEVADA
CA
N
LI
FO
RN
IA
Grid: 7400 m x 6600 m Cell: 100 m x 100 m
Shaded relief base from USGS 10-meter National Elevation

Data, illumination from the northwest at 45. Universal
Transverse Mercator projection, Zone 11, North American
Datum of 1983. Perspective Information: Altitude is 6,300
meters above land surface, viewing angle is 24 degrees,
vertical exaggeration = 2x.
Figure 35. Sagehen Creek watershed located on the east slope of the northern Sierra Nevada near
Truckee, California.
the streambed was estimated to range in age between 26 and The model of the Sagehen Creek watershed relied on
33 years old (Rademacher and others, 2005). These ages and precipitation and air temperature data that were collected from
the quantity of perennial base flow in the creek indicate that four weather stations within the vicinity of the watershed.
ground-water discharge to streams is an important component Three of these weather stations are provided by the U.S.
of the water budget in the Sagehen Creek watershed. Springs Department of Agriculture and are named Independence Camp
present in the upper altitudes of the watershed also indicate (Western Region Climate Center, 2006a), Independence Lake
that the main ground-water flow direction generally follows (Western Region Climate Center, 2006b), and Independence
the slope of land surface (west to east) and leaves the Creek (Western Region Climate Center, 2006c). Data from
watershed as streamflow or ground-water flow beneath the Sagehen Creek co-operative station were provided by
stream. Ground water is assumed not to flow through the University of California, Berkeley (2005). Precipitation in
boundaries of the watershed other than beneath the stream at the Sagehen Creek watershed is highly variable in form and
the watershed outlet. intensity and generally increases with altitude. The areally
averaged annual precipitation from 1960 to 1991 was 970 mm
nv07tac-4124_fig35
(Daly and others, 1994; G.H. Taylor, Oregon Climate Service, Measured data from the Independence Lake snow-
Oregon State University, written commun., 1997). Average telemetry and the Sagehen Creek co-operative stations for
annual precipitation ranged from 865 mm at land-surface 199496 were used to develop the initial values of
altitudes less than 2,135 m to a maximum of 1,168 mm at a the temperature lapse rate parameters (tmin_lapse and
land-surface altitude of 2,576 m. Air temperature generally tmax_lapse) and the precipitation adjustment factors
decreases with land-surface altitude; however, temperature (snow_adj and rain_adj) specified in the PRMS
inversions are common. Mean annual temperature near Parameter File. The years 199496 were chosen because of
Sagehen Creek from 1980 to 2002 was 4C at an altitude of the completeness and quality of the measured data. Other
2,545 m (Western Region Climate Center, 2006b). parameters were estimated for each HRU using spatial data
Daily mean streamflow values were obtained for Sagehen sets distributed with the GIS Weasel toolbox (Viger and
Creek at the streamflow gage near the outlet of the watershed. Leavesley, 2007). Most parameters that do not vary in space
The gage is a U.S. Geological Survey hydrologic benchmark were set to PRMS default values. Estimates of other PRMS
station near Truckee, California (Mast and Clow, 2000; station parameters were from previous modeling studies of the
identification number10343500). Mean daily streamflow was Sagehen Creek watershed (Jeton, 1999) or were made by other
approximately 1 m3/s during snowmelt in May and June for methods that are described in table 13.
19532003. The maximum daily mean discharge on record
was approximately 23 m3/s on January, 1, 1997. Minimum
flows usually occur in late September when surface runoff and Table 13. Values and source of non-default PRMS parameters
interflow are low and ET is still important. Average base flow (appendix 1) used for the Sagehen Creek watershed example
was estimated using daily mean streamflow during September problem (modified from Koczot and others, 2005).
and October and was approximately 0.2 m3s-1 for 19532003.
[Abbreviations: C, parameters that cannot be estimated from available
data and are adjusted during calibration; CG, parameters that are initially
computed in GIS and are adjusted, preserving relative spatial variation
PRMS Input during calibration; L, parameters obtained from the literature as estimated or
empirical estimates]
A PRMS Data File was assembled for the PRMS
Sagehen Creek watershed model for October 1, 1980 Minimum Maximum
through April 25, 2004. This PRMS Data File included Parameter Source
value value
daily streamflow at the Sagehen Creek streamflow gage and
adjmix_rain 0.0055 2.32 L
climate data (precipitation and minimum and maximum air covden_sum .683 1.0 CG
temperature) from the Independence Lake (Natural Resources covden_win .683 1.0 CG
Conservation Service snow-telemetry station 20K05S) and dday_intcp -36.0 -10.0 L
Sagehen Creek (National Weather Service co-operative station dday_slope .31 .65 L
047641). Climate data used for estimating precipitation and fastcoef_lin .4 .4 C
temperature for the PRMS component of the model were from gwflow_coef .00365 .00365 C
the Sagehen Creek co-operative station, because this station jh_coef .0163 .0267 C
provided a longer and more complete temperature record as jh_coef_hru 13.6 15.7 CG
compared with other weather stations near the Sagehen Creek potet_sublim .75 .75 C
pref_flow_den .1 .1 C
watershed. Measured solar radiation data were not available,
rad_trncf .233 .558 CG
and were estimated by the model using the approach described
sat_threshold 3.34 5.16 C
in section Computations of Flow and by Leavesley and smidx_coef .00037 .00037 C
others (1983, p. 14-18). snarea_curve .05 1.0 L
The Sagehen Creek watershed was delineated into 128 snarea_thresh .0 106.3 CG
HRUs (fig. 36), using the USGS GIS Weasel toolbox (Viger snowinfil_max 2.75 2.75 C
and Leavesley, 2007), determined by a combination of flow soil_moist_max 2.39 3.68 CG
planes, climate and vegetation zones, and depth to the water soil_rechr_max 1.48 2.09 CG
table. This resulted in 4,691 gravity reservoirs when the srain_intcp .05 .05 L
HRUs were intersected with active cells in the top layer of the ssr2gw_exp .75 .75 C
MODFLOW grid. The PRMS-only model of Sagehen Creek ssr2gw_rate .0378 .0557 C
watershed had 128 ground-water reservoirs, one corresponding tmax_allrain 60.0 70.0 L
tmax_allsnow 38.2 38.2 L
to each HRU. The ground-water reservoirs were not used in
transp_beg 3 3 C
the integrated GSFLOW model simulations.
transp_end 11 11 C
tstorm_mo 0 1 C
wrain_intcp .05 .05 L
Model boundary
k
ee
Cr
enh
ge
Sa
Hydrologic
Response Elevation: 1,935 meters
Unit MAP
AREA
Truckee NEVADA
CA
N
LI
FO
RN
IA
Grid: 7400 m x 6600 m Cell: 100 m x 100 m
Figure 36. Hydrologic response units discretized for Sagehen Creek watershed near Truckee, California.
MODFLOW-2005 Input was decreased, especially in the upland areas, to represent

a composite of alluvial and volcanic materials. No-flow
The MODFLOW-2005 model of Sagehen Creek conditions were simulated across the bottom and sides of
watershed consisted of two model layers, each layer with the model except for six cells (three for each layer) beneath
73rows and 81 columns in which all cells had a constant and adjacent to the stream at the watershed outlet, which
width and length equal to 90 m. The top altitudes of finite- were specified as constant-head cells. The simulation period
difference cells were set equal to estimated land-surface was divided into one steady-state stress period followed by
altitudes for each cell and ranged from 1,935 to 2,653 m; the four transient stress periods, three had 182 1-day time steps,
thickness of the top model layer ranged between 20 and 70 m and the last stress period had 5,500 1-day time steps. Five
and the thickness of the bottom layer ranged between 10 and stress periods were created in this example in order to test
115 m. The upper layer was created to represent the alluvium the effects of pumping wells in the simulation; however, the
in valley lowlands; however, it was made thicker than the Well Package was not included in the version of the example
assumed thickness of the alluvium to avoid drying of cells. problem presented herein.
Consequently, the hydraulic conductivity of the upper layer
nv07tac-4124_fig36
Several packages and optional files were used in the of hydraulic conductivity of layer 1 and the unsaturated zone.
simulation of the Sagehen Creek watershed (table 14). The Unsaturated-zone properties beneath all stream reaches were
Layer-Property Flow Package (Harbaugh, 2005, p. 5-17) was constant with a vertical hydraulic conductivity of 0.3 m/d, a
used and both horizontal and vertical hydraulic conductivity saturated water content of 0.30 and a Brooks-Corey exponent
values were specified. The upper model layer represented the of 3.5 (table 15). The initial water content beneath streams
shallow alluvial material, whereas the lower layer represented was set to 0.2. This was set as a starting value only because
the volcanic material. All cells in the upper layer were the initial water content beneath streams was calculated during
specified as convertible (unconfined). The second layer was the steady-state simulation for stream reaches where the water
specified as confined because the water table was contained table in layer 1 was beneath the streambed.
within the upper layer. The horizontal hydraulic conductivity
(K) for the upper layer ranged from 0.026 m/d on the ridges to
0.39 m/d in the valleys as determined by calibration to match Calibration
mean base flow of Sagehen Creek (table 15). A lower K was
The GSFLOW model of the Sagehen Creek watershed
specified on the ridges where volcanic rocks are near land
was calibrated using a stepwise procedure: first, independent
surface (fig. 37). The horizontal K for the lower layer ranged
PRMS and MODFLOW-2005 models were calibrated
from 0.00045 to 0.027 m/d, in which the lower K values were
by running these models separately (PRMS-only and
specified beneath ridges. The hydraulic conductivity within
MODFLOW-only simulations), followed by calibration of the
each cell was assumed isotropic. Specific storage was set to 2
integrated model. These steps are described in the following
x10-6 m-1 for both layers, and the specific yield was specified
sections.
as 0.08 on the ridges and 0.15 in the valleys near streams. The
vertical hydraulic conductivity values of the unsaturated zone
were specified within the Unsaturated-Zone Flow Package File PRMS Calibration
and ranged between 0.018 and 0.27 m/d (table 15). A constant
Brooks-Corey exponent of 4 and a saturated water content Initial calibration of PRMS was done by running
ranging from 0.15 to 0.25 were assigned to all unsaturated- PRMS independently of MODFLOW-2005 (parameter
zone cells. model_mode set to PRMS in the GSFLOW Control File).
Streams were represented by 15 segments made up PRMS was calibrated using a split sample approach such that
of 201 reaches using the Streamflow-Routing Package three non-overlapping periods of the data record were used for
and included the simulation of unsaturated flow beneath model initialization, calibration, and verification. Data from
streams. All stream reaches were set to a constant width of October 1, 1980, to September 30, 1981 (water year 1981)
3 m and ranged in length from 30 to 120 m. The streambed were used to develop initial storages in the PRMS model. Data
hydraulic conductivity was set equal to 5.0 m/d (table 15). The from October 1, 1981, to September 30, 1988 (water years
streambed hydraulic conductivity was set to a large value such 198288) were used for calibrating the PRMS model, and data
that ground-water exchange with the stream was a function from October 1, 1988, to September 30, 1995 (water years
198995) were used for verifying the calibrated PRMS model.
The calibration procedure, described in detail below, consisted
of first adjusting parameters that affect the average flow
Table 14. MODFLOW-2005 packages and files used for the
of water through the watershed until the model provided a
Sagehen Creek watershed example problem.
reasonable match between the simulated and measured annual
water balance. Parameters that affect the timing and magnitude
MODFLOW-2005 packages and files
of the simulated streamflow were then adjusted until the model
Basic Package File provided a good fit between the simulated and measured daily
Output Control Option File streamflow.
Discretization File The PRMS precipitation lapse rate factors
Layer-Property Flow Package File
(snow_adj and rain_adj in the PRMS Parameter
Preconditioned Conjugate Gradient Solver Package File
File) were adjusted until the mean of the simulated annual
Unsaturated-Zone Flow Package File
Streamflow-Routing Package File precipitation approximately matched the mean of the
Gage Package File measured values for the period when the lapse rates were
computed. Then, evapotranspiration parameters were
adjusted until the simulated mean annual evapotranspiration
approximately matched a value of 27 in/yr (686 mm/yr)
that was estimated from Kahrl (1979, p. 13).
Model boundary
k
ee
Cr
enh
ge
Sa
Elevation: 1,935 meters MAP
AREA
Truckee NEVADA
CA
N
LI
FO
RN
IA
Grid: 7400 m x 6600 m Cell: 100 m x 100 m
Transverse Mercator projection, Zone 11, North American EXPLANATION
Datum of 1983. Perspective Information: Altitude is 6,300 Hydraulic conductivity valueMeters per day
Layer 1 Layer 2
0.026 0.00045
0.052 0.0009
0.065 0.0027
0.13 0.009
0.39 0.027
Figure 37. Hydraulic conductivity values used in ground-water model, Sagehen Creek watershed near
nv07tac-4124_fig37
Table 15. Hydraulic properties and other selected variables used in the Layer-Property Flow, Unsaturated-Zone Flow, and Streamflow-
Routing Packages for the Sagehen Creek watershed example problem.
[NoteInitial water content values must be specified if the first stress period is transient]
Variable Units Minimum value Maximum value

Variables assigned to the Layer-Property Flow Package
Horizontal hydraulic conductivity for upper meters per day 0.026 0.39
layer
Vertical hydraulic conductivity for upper layer meters per day .026 .39
Horizontal hydraulic conductivity for lower meters per day .00045 .027
layer
Vertical hydraulic conductivity for lower layer meters per day .00045 .027
Specific storage for both layers per meter of aquifer 2 10-6 2 10-6
Specific yield cubic meter of water per .08 .15
cubic meter of aquifer
Variables assigned to the Unsaturated-Zone Flow Package
Vertical hydraulic conductivity of the meters per day 0.018 0.27
unsaturated zone below land surface
Brooks-Corey exponent for unsaturated dimensionless 4 4
zone below land surface
Saturated water content of unsaturated cubic meter of water per cubic .15 .25
zone below land surface meter of unsaturated zone
Initial water content of unsaturated zone cubic meter of water per cubic Not required for Not required for
below land surface meter of unsaturated zone steady-state stress steady-state stress
period [see note] period [see note]
Steady-state infiltration rate meters per day .0007 .0019
Variables assigned to the Streamflow-Routing Package
Hydraulic conductivity of streambed meters per day 5.0 5.0
Streambed thickness meters 1.0 1.0
Vertical hydraulic conductivity of unsaturated meters per day .3 .3
zone beneath streams
Brooks-Corey exponent for unsaturated zone dimensionless 3.5 3.5
beneath streams
Saturated water content of unsaturated zone cubic meter of water per cubic .3 .3
beneath streams meter of unsaturated zone
Initial water content of unsaturated zone cubic meter of water per cubic .2 .2
beneath streams (see note) meter of unsaturated zone
All adjusted parameters were kept within a reasonable range and a negative value indicates that the mean measured value is
so as not to violate any physical constraints and also to provide a better predictor than the simulated hydrograph. The Nash-
a close match of simulated and measured annual streamflow. Sutcliffe coefficient for the calibrated period is 0.79 and for
The annual water budgets are summarized in table 16. The the verification period is 0.71.
measured and simulated hydrographs for the calibration and Visual inspection of the hydrographs in figure 38 indicate
verification periods are shown in figure 38. The goodness that the model generally performs well; however, there are
of fit between the simulated and measured hydrographs was several instances during winter months, where measured high
assessed with the Nash-Sutcliffe efficiency coefficient (Nash streamflows are not simulated by the model, or the model
and Sutcliffe, 1970). This coefficient is a relative measure simulates high streamflows that were not measured. This
of model predictive power. A value of 1.0 indicates a perfect indicates that when air temperature in the watershed is near
simulation; a value of 0.0 indicates the mean measured value 0oC, the model cannot accurately determine the form of the
is equivalent to the simulated hydrograph in predictive power; precipitation.
1200
STREAMFLOW, IN THOUSANDS OF
Simulated
CUBIC METERS PER DAY Daily mean
800
400
0
1982 1983 1984 1985 1986 1987 1988
B
800
STREAMFLOW, IN THOUSANDS OF
CUBIC METERS PER DAY
600
400
200
0
1989 1990 1991 1992 1993 1994 1995
WATER YEAR
Figure 38. PRMS simulated and daily-mean streamflow at Sagehen Creek gage for
(A) calibration period (water years 198288) and (B) verification period (water years
198995). Daily streamflow values are from a GSFLOW Comma-Separated Values (CSV)
File using variables basinstrmflow and obs_strmflow.
nv07tac-4124_fig38
Table 16. Annual water budget for PRMS calibration and verification periods for the Sagehen Creek watershed example problem.
[, not determined]
Annual Annual Annual storage Simulated annual Measured annual Percent difference
precipitation per evapotranspiration change per streamflow per streamflow per between simulated
Water year
drainage area per drainage area drainage area drainage area drainage area and measured
(meters) (meters) (meters) (meters) (meters) annual streamflow
Calibration period
1982 1.950 0.772 0.180 0.999 0.935 -6.6
1983 1.639 .829 -.042 .854 .975 13.2
1984 1.325 .786 -.059 .599 .665 10.4
1985 .750 .551 -.056 .257 .286 10.7
1986 1.535 .769 .090 .678 .687 1.3
1987 .454 .403 -.126 .179 .164 -8.7
1988 .539 .426 -.017 .130 .105 -21.3
Average 1.170 .648 .528 .545 3.2
Verification period
1989 1.114 0.663 0.071 0.380 0.322 -16.5
1990 .761 .590 -.019 .191 .165 -14.6
1991 .656 .562 -.042 .137 .142 3.6
1992 .548 .385 -.001 .166 .098 -51.5
1993 1.399 .728 .065 .607 .480 -23.4
1994 .517 .401 -.053 .170 .106 -46.4
1995 1.771 .782 .090 .901 .749 -18.4
Average .967 .587 .364 .294 -21.3
MODFLOW-2005 Steady-State Calibration theK(Array Reading Utility Subroutines

variable CNSTNT; Harbaugh, 2005; p. 857). The infiltration-
Initial calibration of ground-water flow was done rate factor was adjusted until the steady-state streamflow at
by running MODFLOW-2005 independently of PRMS the watershed outlet was approximately equal to the mean
(parameter model_mode set to MODFLOW in GSFLOW base flow of 0.2 m3/s for 19532003. K factors specified in
Control File). Measurements of precipitation, streamflow, and the Layer-Property Flow Package File were adjusted until the
the location of springs in the watershed were used to calibrate steady-state water table was above or at the soil-zone base near
the ground-water flow model. These data were useful in this streams and where springs are located in the watershed.
example for describing an approach for model calibration and Although no wells have been drilled on the ridges in the
to provide a realistic simulation. Calibration of the Sagehen watershed, the maximum water-table depth was assumed to
Creek ground-water flow model was done using a trial-and- be less than 75 m in the watershed. This depth is based on
error approach; however, automated procedures can be used depths measured in wells elsewhere in the northern Sierra
for complicated models when sufficient data are available. Nevada. Some ground-water levels were measured in shallow
The distribution of horizontal hydraulic conductivity (<5 m) wells near the Sagehen Creek watershed outlet close
(K) was created initially on the basis of the surface geology. to the creek. These wells indicate that the water table remains
During the steady-state calibration, spatially varying ground- close to land surface in this area. Simulated water-table depths
water recharge was assumed proportional to the spatial were 60 m beneath the soil-zone base along the ridges and
distribution of mean annual precipitation with altitude at, or slightly above, the soil-zone base in some of the upland
(fig.39). Calibration of the ground-water model was done areas where springs are located. The water-table depth in the
by adjusting the steady-state infiltration-rate (Unsaturated- lowland areas and valleys generally were at or above the soil-
Zone Flow Package input variable FINF; appendix 1) and zone base (fig. 40).
Model boundary
EXPLANATION
Steady-state infiltration
rateMeter per day
0.0024
0.0016
0.0008
N
0 5,000 METERS
Figure 39. Steady-state infiltration rate for the ground-water model of

Sagehen Creek watershed near Truckee, California.
GSFLOW Calibration Surface-Water Calibration

Initial evaluation of the GSFLOW model before
Parameters describing the overlay of PRMS HRUs and
calibration indicated two important water-balance problems.
MODFLOW finite-difference cells (fig. 41) were developed
The first was that the gravity drainage rate from the soil zone
using the USGS GIS Weasel toolbox, and the calibrated
to the unsaturated zone was insufficient to keep the transient
parameter sets from the PRMS and MODFLOW-2005 models
ground-water head levels in balance with the steady-state
of the Sagehen Creek watershed were initially used as input
heads. The second was that actual evapotranspiration rates
parameters for the GSFLOW model. GSFLOW could be
were too high during the summer months. To address these
calibrated using most of the parameters from these models
problems, two PRMS calibration parameters were manually
unchanged, while focusing on just a few of the parameters that
modified for the GSFLOW model of the Sagehen Creek
influence the flows between PRMS and MODFLOW-2005.
watershed (table 17).
The measured and simulated hydrographs for the calibration
and verification periods are presented in figure 42. The Nash-
Sutcliffe coefficient for the calibrated period is 0.81 and for Table 17. Values of PRMS parameters (appendix 1) modified for
the verification period is 0.85. Changes to parameters made the GSFLOW Sagehen Creek watershed example problem.
during calibration of the integrated GSFLOW model are
described in the following two sections. Parameter PRMS values GSFLOW values
jh_coef 0.0163 to 0.0267 0.0148 to 0.0267
ssr2gw_rate 0.0378 to 0.05571 0.189 to 0.278
nv07tac-4124_fig39
Model boundary
EXPLANATION
Water-table depthMeters
>25
12.5
N
0 5,000 METERS
Figure 40. Simulated steady-state water-table depth below land

surface in the Sagehen Creek watershed near Truckee, California.
Water-table depth below land surface calculated from a MODFLOW
formatted head output file.
Transient Ground-Water Calibration The calibrated model was able to simulate decadal
oscillations in base flow caused by years of above and
Generally, the ground-water component of GSFLOW
below normal precipitation (fig. 43), although there were
(MODFLOW-2005) was calibrated to streamflow measured
some discrepancies between simulated and estimated base
during September and October (autumn) from 1981 to 1995.
flow. Base flow (simulated and estimated) for this example
The transient ground-water model calibration was done by
is defined as the annual minimum daily mean discharge at
further adjusting the K distribution from what was estimated
the Sagehen Creek gage, which usually occurred between
using the steady-state calibration. Accordingly, the transient
September and October. For example, estimated base flow
ground-water model calibration required an iterative process
was 0.175 m3/s, whereas simulated base flow was 0.105
where the K values were adjusted for the transient calibration
m3/s during 1983 (fig. 43). These discrepancies occurred
followed by further adjustment of the recharge factor for the
steady-state simulation.
nv07tac-4124_fig40
Model boundary
k
ee
Cr
enh
ge
Sa
Hydrologic
Response Elevation: 1,935 meters
Unit and MAP
finite-difference AREA
Truckee
cell NEVADA
N
CA
LI
FO
RN
IA
Grid: 7400 m x 6600 m Cell: 100 m x 100 m
Figure 41. Finite-difference cells in relation to hydrologic response units in Sagehen Creek watershed near
during years of above-average precipitation, which caused calibration of either the surface-water and/or ground-water
the snowmelt period to extend into the autumn months. One components of GSFLOW. Decadal variations in estimated
explanation for this discrepancy is that runoff and interflow base flow are caused by consecutive years of above-average
occurred throughout the autumn months into the following precipitation and the resulting increases in unsaturated-zone
water year. Thus, the discrepancy may be due to imperfect and ground-water storage.
nv07tac-4124_fig41
10,000
STREAMFLOW, IN THOUSANDS
OF CUBIC METERS PER DAY
8,000
Simulated
Daily mean
6,000
4,000
2,000
0
1982 1983 1984 1985 1986 1987 1988
B
6,000
STREAMFLOW, IN THOUSANDS
5,000
4,000
3,000
2,000
1,000
0
1989 1990 1991 1992 1993 1994 1995
WATER YEAR
Figure 42. GSFLOW simulated and daily-mean streamflow at Sagehen Creek gage
for (A) calibration period (water years 198288) and (B) verification period (water
years 198995). Daily streamflow values are from a GSFLOW Comma-Separated
Values (CSV) File using variables basinstrmflow and obs_strmflow.
nv07tac-4124_fig42
16
STREAMFLOW, IN THOUSANDS OF CUBIC

12
Estimated base flow

METERS PER DAY
Simulated base flow
0
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
Figure 43. Comparison between simulated and estimated base flow for water years
198195. Base flow was determined from daily values of streamflow from a GSFLOW
Model Results Storages

This section presents selected output from the calibrated Storage is divided into four major compartments in the
GSFLOW simulation of the Sagehen Creek watershed. The GSFLOW model of the Sagehen Creek watershed. These
model results described herein were simulated on a laptop are snowpack storage (principal surface storage), soil-zone
personal computer, running the Microsoft Windows XP storage, unsaturated-zone storage, and saturated-zone storage
Professional operating system, with a 2.0 gigahertz AMD (fig. 45). The movement of water into and out of each
processor, and 2.0 gigabytes of random access memory. No compartment was dominated by the annual snowpack cycle
additional software was required. Run time was less than as illustrated for the calibration period from 1981 to 1989.
7minutes per year (365 days) simulated, with an average Three periods in the annual snowpack cycle generally can
of 13iterations per day. The number of iterations for each be identified: (1) the accumulation period begins October 1
simulated time step from 1981 to 1994 is shown in figure 44. and peaks toward the end of April; (2) the snowmelt period
occurs from May through July; and (3) the no-snow period
nv07tac-4124_fig43
50
NONLINEAR SOLUTION ITERATIONS

40
30
20
10
0
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995
WATER YEAR
Figure 44. Number of iterations per simulation time step for GSFLOW to converge during
water years 198194. Iteration values are from a GSFLOW Comma-Separated Values (CSV)
File using variable kkiter.
occurs from August through October. Soil-zone storage Unsaturated-zone storage generally peaks in July after all
generally is higher through periods 1 and 2 because of reduced snowmelt has occurred (fig. 45B). Saturated-zone or ground-
evapotranspiration and replenishment from snowmelt, and water storage generally peaks in August, decreasing until April
then decreases in period 3. The annual cycle is superimposed or May (fig. 45B). Variations in the unsaturated-zone storage
on a longer cycle of wet and dry years. Minimal snowpack also are affected by changes in the water-table depth and thus,
was simulated during water years 1981 and 1987 and generally trend opposite to changes in the saturated-zone
maximum snowpack was simulated during water year 1983. storage. The saturated-zone storage increased to a maximum
Storages shown in figure 45 are summed over the entire during water years 198384 and subsequently decreased
watershed; simulated storage change can occur at different during the drier years 198589, whereas the unsaturated-
rates throughout the watershed. zone storage decreased during water years 198384 because
the unsaturated-zone thickness decreased and then increased
during the drier years 198589.
nv07tac-4124_fig44
A
400 400
CUMULATIVE CHANGE IN SNOWPACK STORAGE,

CUMULATIVE CHANGE IN SOIL-ZONE STORAGE,
Cumulative change in soil-zone storage

Cumulative change in snowpack storage
IN HUNDREDS OF THOUSANDS
IN HUNDREDS OF THOUSANDS
300 300
OF CUBIC METERS
OF CUBIC METERS
200 200
100 100
0 0
B
4,000 12,000
CUMULATIVE CHANGE IN UNSATURATED-ZONE
STORAGE, IN THOUSANDS OF CUBIC METERS
STORAGE, IN THOUSANDS OF CUBIC METERS

CUMULATIVE CHANGE IN SATURATED-ZONE
Cumulative change in unsaturated-zone storage
Cumulative change in saturated-zone (ground-water) storage
2,000 6,000
0 0
-2,000 -6,000
-4,000 -12,000
1981 1982 1983 1984 1985 1986 1987 1988 1989
WATER YEAR
Figure 45. Cumulative change of (A) snowpack and soil-zone storage; and (B)
unsaturated-zone and saturated-zone (ground-water) storage during water years
198189. Water year begins October 1 of each year. Storage values are from a GSFLOW
Comma-Separated Values (CSV) File. Cumulative change in snowpack storage was
calculated from variable basinpweqv, and cumulative change in soil-zone storage
was calculated from variables basinsoilmoist, basingravstor, and
basinprefstor. Cumulative change in unsaturated-zone and saturated-zone
storage was calculated from variables uzf_del_stor and sat_change_stor,
respectively.
nv07tac-4124_fig45
Fluxes snow accumulation and snowmelt period even during the

drier years, whereas ground-water discharge exceeded gravity
The main fluxes of water across land surface for the drainage following snowmelt and before snow accumulation
Sagehen Creek watershed included precipitation, snowmelt, began in late autumn. Fluxes shown in figure 47 also are
and to a lesser extent, evapotranspiration, as illustrated areally averaged values over the entire watershed; simulated
for water years 198189 (fig. 46). Precipitation occurred gravity drainage and ground-water discharge can occur at
as rain, snow, or a mixture of both. Snowmelt occurred different rates throughout the watershed. In fact, at most time
as the snowpack ablates during the snowmelt period. steps, both occurred simultaneously at different locations in
Evapotranspiration is the total evaporation from the plant the watershed.
canopy, sublimation, and evapotranspiration from the soil Three major fluxes of water in the watershed add
zone. Evapotranspiration generally occurred during the streamflow to Sagehen Creek. These are surface runoff,
warmer periods of the year when sufficient water was available interflow, and ground-water discharge directly to streams,
in the soil zone. Fluxes shown in figure 46 are areally averaged which is commonly called base flow (fig. 48). Surface runoff
values over the entire watershed; simulated precipitation, occurred sporadically from 198289 due to rainfall on
snowmelt, and evapotranspiration can occur at different rates saturated soils. Large fluxes of surface runoff were simulated
throughout the watershed depending on lapse rates, geographic when sufficient rain fell on the snowpack or on saturated or
characteristics, and subsurface properties. frozen soils. These events are difficult to simulate. Interflow
Two major fluxes of water were simulated between was the predominant source of simulated streamflow during
the soil zone and the unsaturated and saturated zones in the most years except for those years of large surface runoff. Net
Sagehen Creek watershed. The first was gravity drainage ground-water discharge directly to streams typically was the
of water from the soil zone down to the unsaturated zone, least of the three fluxes that generate streamflow in Sagehen
whereas the second was ground-water discharge from Creek annually but was the predominant source of streamflow
the saturated zone into the soil zone, as illustrated for the following the snowmelt period. As in the figures 45-47, the
calibration period from 1981 to1989 (fig. 47). Gravity fluxes shown in figure 48 are areally averaged values over the
drainage exceeded ground-water discharge during the annual entire watershed.
5,000
Precipitation
Snowmelt
4,000
Evapotranspiration
FLUX, IN THOUSANDS OF
3,000
2,000
1,000
0
1981 1982 1983 1984 1985 1986 1987 1988 1989
WATER YEAR
Figure 46. Major fluxes across land surface during water years 198189 that include
precipitation, snowmelt, and evapotranspiration. Flux values are from a PRMS
Animation File using variables basin_ppt,
400 400
GROUND-WATER DISCHARGE TO SOIL ZONE,

IN THOUSANDS OF CUBIC METERS PER DAY
THOUSANDS OF CUBIC METERS PER DAY
Gravity drainage through soil-zone base
FLOW THROUGH SOIL-ZONE BASE, IN
Ground-water discharge to soil zone
300 300
200 200
100 100
0 0
1981 1982 1983 1984 1985 1986 1987 1988 1989
WATER YEAR
Figure 47. Fluxes of gravity drainage from soil zone down to unsaturated zone
and ground-water discharge from saturated zone into soil zone during water years
198189. Flux values are from a GSFLOW Comma-Separated Values (CSV) File using
variables basinsz2gw and basingw2sz.
1,000
THOUSANDS OF CUBIC METERS PER DAY
Runoff
STREAMFLOW CONTRIBUTIONS, IN
800
Interflow
Net ground-water discharge to streams
600
400
200
0
1981 1982 1983 1984 1985 1986 1987 1988 1989
WATER YEAR
Figure 48. Components of streamflow during water years 198189. Runoff and
interflow are from a GSFLOW Comma-Separated Values (CSV) File using variables
basinsroff, and basininterflow; streamflow is from the same file using
variable basinstrmflow; and net ground-water discharge to streams was
calculated by subtracting runoff and interflow from streamflow.
nv07tac-4124_fig47
A
0 5,000 METERS
Model boundary
EXPLANATION
Ground-water discharge to
soil zoneCubic meters per day
100
50
0
N
Figure 49. Distribution of ground-water discharge to the soil zone for (A) steady-state simulation of the
autumn low precipitation period, and (B) July 8, 1983. Cell-by-cell values of ground-water discharge are from an
Unsaturated-Zone Flow Package unformatted output file.
nv07tac-4124_fig49
Simulated ground-water discharge to the soil zone interflow. Seasonal variation in ground-water discharge to the
showed seasonal variations, as illustrated by the difference soil zone likely is realistic given the large seasonal variation in
in the steady-state distribution of ground-water discharge precipitation and streamflow in the watershed.
(simulation of mean flow during the autumn low-precipitation Simulated water-content profiles beneath the soil-zone
period) and ground-water discharge during the peak snowmelt base varied seasonally and generally were greatest during
in July 1983 (fig. 49). The total ground-water discharge to the the spring snowmelt period. Water-content profiles are
soil zone was about 0.19 m3/s during October (corresponding shown in figure 50 for a MODFLOW finite-difference cell
to the steady-state discharge in fig. 49) and 0.28 m3/s during located about midway between the peak of the watershed
June 1983. Ground-water discharge to the soil zone can and the watershed outlet that had a land-surface altitude 82
contribute interflow and surface runoff and thus, ground water m greater than the altitude of the watershed outlet (layer 1,
can contribute flow to streams through surface runoff and row 45, and column 30). Maximum saturations in the upland
A B
4/1/1983 6/1/1983
1 4/5/1983 6/5/1983
DEPTH BELOW LAND SURFACE, IN METERS
4/10/1983 6/10/1983
5
0.05 0.075 0.1 0.125 0.05 0.075 0.1 0.125
VOLUMETRIC WATER CONTENT, IN CUBIC METERS PER CUBIC METER
Figure 50. Water-content profiles below land surface (finite-difference cell layer 1, row 45, and column
30)(A) during April 1983, and (B) during June 1983. Water-content values are from an Unsaturated-Zone
Package formatted output file.
areas were approximately 67 percent of the saturated water Sensitivity to Unsaturated-Zone and Aquifer
content, which ranged between 0.15 and 0.25. The water table
fluctuated seasonally by more than 4 m, resulting in seasonal
Hydraulic Conductivity
disappearance of the unsaturated zone beneath the soil-zone Model sensitivity to unsaturated-zone and aquifer
base in places where the water-table depth was less than 4 m. hydraulic conductivity (K) was analyzed by adjusting the
value of CNSTNT (Harbaugh, 2005; p. 857), which is used to
Annual Water Budgets and Errors equally scale K values for all cells. Net gravity drainage into
the unsaturated zone increased proportionately to the increase
Volumetric budgets and errors, on an annual basis, are in unsaturated zone and aquifer K, except during periods when
presented in table 18 for water years 198295. These results soil-zone storage was depleted. The amount of water available
were obtained from the GSFLOW Water-Budget File (fig. 29), for gravity drainage into the unsaturated zone beneath the
which is described in detail in section Water Budget. The soil zone during spring snowmelt was limited by the vertical
percent discrepancy values for each year are all very good. K in the unsaturated zone for lower K values (fig. 51A). The
decrease in gravity drainage for the lowest K values was
enhanced by the increase in area where ground-water levels
were at or above land surface; saturation excess caused a
Table 18. Annual water budget for GSFLOW simulation of the Sagehen Creek watershed example problem, 198295.
[All values are reported in units of cubic meters, except percent discrepancy is dimensionless]
In Out
Water year InOut
Precipitation Evapotranspiration Streamflow Boundary flow
1982 7.2483E+07 3.7929E+07 3.0458E+07 478,282.41 3,617,919.50
1983 1.1747E+08 6.1108E+07 5.2342E+07 720,214.69 3,297,289.25
1984 1.5372E+08 8.4367E+07 6.7816E+07 962,476.62 576,307.38
1985 1.7429E+08 1.0141E+08 7.5457E+07 1,201,432.75 -3,773,800.75
1986 2.1656E+08 1.2401E+08 9.1433E+07 1,441,176.00 -329,792.00
1987 2.2903E+08 1.3783E+08 9.6804E+07 1,678,463.38 -7,284,239.50
1988 2.4382E+08 1.5183E+08 1.0093E+08 1,913,408.12 -1.0854E+07
1989 2.7405E+08 1.7128E+08 1.0964E+08 2,148,526.75 -9,014,063.00
1990 2.9528E+08 1.8905E+08 1.1472E+08 2,383,076.75 -1.0870E+07
1991 3.1330E+08 2.0617E+08 1.1894E+08 2,616,651.50 -1.4427E+07
1992 3.2835E+08 2.1855E+08 1.2299E+08 2,849,880.75 -1.6040E+07
1993 3.6674E+08 2.3905E+08 1.3559E+08 3,084,391.75 -1.0983E+07
1994 3.8064E+08 2.5197E+08 1.3991E+08 3,318,625.50 -1.4556E+07
1995 4.2954E+08 2.7376E+08 1.5948E+08 3,555,781.75 -7,263,078.00
Storage change Percent

Water year Error
Surface Soil zone Unsaturated Saturated Total discrepancy
1982 120,517.80 2,660,699.00 -1,377,943.75 2,208,153.25 3,611,416.75 6,502.75 0.01

1983 7,197.33 1,636,156.88 -2,109,868.75 3,746,286.50 3,279,756.50 17,532.75 .01
1984 -1.00 385,343.59 -2,337,493.50 2,503,117.00 550,946.69 25,360.69 .02
1985 31,096.60 627,998.62 -2,079,286.75 -2,372,487.50 -3,792,694.50 18,893.75 .01
1986 35,995.13 1,863,144.12 -873,155.62 -1,377,784.75 -351,815.53 22,023.53 .01
1987 -.50 136,199.17 -1,461,251.12 -5,969,752.00 -7,294,821.00 10,581.50 .00
1988 4,490.44 185,758.05 -706,397.12 -1.0329E+07 -1.0845E+07 -9,349.00 .00
1989 2,953.11 1,075,756.12 1,275,628.12 -1.1354E+07 -8,999,396.00 -14,667.00 -.01
1990 -1.13 1,326,907.25 1,500,342.38 -1.3677E+07 -1.0850E+07 -20,031.00 -.01
1991 -.62 316,719.00 1,608,705.62 -1.6317E+07 -1.4392E+07 -35,615.00 -.01
1992 -.53 206,177.45 2,149,828.75 -1.8342E+07 -1.5986E+07 -53,920.00 -.02
1993 2.50 196,883.77 4,265,108.50 -1.5384E+07 -1.0922E+07 -60,893.00 -.02
1994 2.47 193,511.73 2,192,072.00 -1.6876E+07 -1.4490E+07 -66,133.00 -.02
1995 39,441.46 264,386.19 3,940,972.50 -1.1445E+07 -7199875.50 -63,202.50 -.01
IN THOUSANDS OF CUBIC METERS PER DAY 280

NET GRAVITY DRAINAGE FROM SOIL ZONE,
210
140
70
0
160
B
CNSTNT = 5
RECHARGE, IN THOUSANDS OF
CNSTNT = 1
120 CNSTNT = 0.2
80
40
0
60
C
GROUND-WATER DISCHARGE, IN THOUSANDS
Figure 51. Sensitivity analysis

for daily volumetric flow of (A) net
gravity drainage from the soil zone,
(B) recharge to the saturated zone,

40 and (C) ground-water discharge
to land surface summed over
the Sagehen Creek watershed
for water years 198283. The
variable CNSTNT is multiplied by
the hydraulic conductivity of all
20
finite-difference cells (Harbaugh,
2005; p. 857). Values are from an
Unsaturated-Zone Flow Package
formatted output file.
0
WATER YEAR 1982 WATER YEAR 1983
nv07tac-4124_fig51
non-linear relation between K and net gravity drainage for the The simulated streamflow at Sagehen Creek received
Sagehen Creek watershed (fig. 51). Generally, several factors contributions from surface runoff, interflow, and net ground-
affect the net gravity drainage beneath the soil zone, including water discharge directly to streams. Net ground-water
the amount of soil-zone storage, the area where ground-water discharge is the total ground-water discharge to streams
levels are at or near land surface, and the vertical K in the minus total stream leakage to ground water upstream from the
unsaturated zone. The extent to which each of these factors Sagehen Creek gage. Surface runoff and interflow provided
affects net gravity drainage is dependent on the climate, land most of the annual streamflow from 1981 to 1995. Some of
surface topography, and subsurface properties in a watershed the interflow (and surface runoff) simulated in GSFLOW was
or basin. from ground-water discharge into the soil zone in lowland
Total daily ground-water recharge rates did not areas where the water table was greater than the soil-zone
increase proportionately to the increase in K. The medium-K base. During years of above-average precipitation, interflow
distribution resulted in the greatest ground-water recharge lasted throughout the water year and remained greater than the
rates during the snowmelt period (fig. 51B). Much of the net ground-water discharge to stream reaches into the summer;
recharge occurred where the water table was slightly beneath for 1983, interflow began to decrease about July 1 (fig. 52A).
the soil-zone base, whereas in areas with a thick unsaturated However, during years of below-average precipitation, there
zone, the total daily ground-water recharge rates were were several month-long periods during which net ground-
attenuated due to coalescence of high unsaturated-zone flow water discharge was greater than interflow and surface
rates with lower rates. Attenuation occurs to a lesser degree in runoff. Net ground-water discharge decreased during peak
thin unsaturated zones because wetting fronts reach the water streamflow because stream stage exceeded the ground-water
table before being attenuated. These results are attributed to heads in many stream reaches. However, ground-water head
the relation between the horizontal K and the thickness of the subsided slower than the streamflow due to bank storage
unsaturated zone. Generally, high horizontal K values allow effects and diffuse recharge from nearby cells. This resulted
ground water to drain out of the watershed, which thickens in net ground-water discharge back to the stream reaches to be
the unsaturated zone beneath much of the watershed. The high greatest following high streamflow (fig. 52).
recharge rates for the medium-K distribution as compared The integrated model resulted in a modest improvement
with the low-K distribution are caused by much greater gravity of the Nash-Sutcliffe coefficient compared with the PRMS-
drainage from the soil zone. The high-K distribution allowed only simulation from 0.79 to 0.81 for the calibration period
for more water to be stored in the unsaturated zone, which (water years 198288) but the integrated model resulted
recharged ground water much longer after the spring snowmelt in a marked increase from 0.71 to 0.85 for the verification
than either the low or mediumK distributions (fig. 51B). The period (water years 198995). The less rigorous approach
upland areas comprise a larger fraction of the total watershed for simulating subsurface flow and storage with PRMS-only
area compared with the valley lowland areas at Sagehen was not able to simulate the observed decadal variations
Creek; consequently, a change in K values in these areas can in base flow. The integrated GSFLOW simulation with
greatly change the timing and magnitude of recharge in the MODFLOW-2005 was required to approximate the observed
watershed. decadal variations in base flow, which improved the simulation
Ground-water discharge to the soil zone increased with of streamflow for the drier verification period. Decadal
increases in K for the range in K values used in this sensitivity variations in base flow may not be important for flood
analysis (fig. 51C). However, yearly maximum ground-water predictions, or for assessing volumes of streamflow from
discharge rates may not increase with higher values of K. snowmelt runoff. Integrated models may not be practical for
The yearly maximum ground-water discharge rates decreased such applications. However, integrated models are necessary
when K was increased by a factor of 100 for a simulation with whenever ground-water interactions with surface water are
one model layer (Niswonger and others, 2006b). of concern. Examples include (1) effects of ground-water
withdrawals on surface-water supplies and stream ecology, (2)
effects caused by stream diversions and reservoirs on ground
Discussion of Results water and associated wetland systems, and (3) effects of
conjunctive use of surface- and ground-water supplies. Thus,
Runoff and ground-water recharge at Sagehen Creek
integrated models should not be assessed solely by how well
watershed was dominated by water made available by
the model predicts streamflow at the outlet of a basin using a
snowmelt. Peak evapotranspiration rates generally occurred
single measure of goodness of fit. Rather, integrated models
slightly later than peak snowmelt and runoff. Most of the
should be evaluated on the basis of a variety of different
precipitation at Sagehen Creek watershed occurred as snow
measures to evaluate the importance of changes in surface and
and accumulated in snowpack. Water became available to
subsurface flows and storages within the modeled domain.
plants as melt water during the warmer months of the year
Such measures might include changes in snowpack depth and
when plants reach their highest transpiration rates. This is
water content, unsaturated zone moisture profiles, and ground-
dissimilar to lower altitude basins where winter precipitation
water heads.
occurs as rain and becomes runoff prior to the period of
maximum evapotranspiration rates.
1,000
A. WATER YEAR 1983
100
10
FLOW, IN THOUSANDS OF CUBIC METERS PER DAY
Streamflow
1 Runoff and interflow
Net Ground-water discharge to streams
0.1
1,000
B. WATER YEAR 1992
100
10
0.1
OCT NOV DEC JAN FEB MAR APR MAY JUNE JULY AUG SEPT
Figure 52. Streamflow components for (A) 1983, a year of above-average precipitation,
and (B) 1992, a year of below-average precipitation. Runoff and interflow are from a
GSFLOW Comma-Separated Values (CSV) File using variables basinsroff and
basininterflow; streamflow is from same file using variable basinstrmflow;
and net ground-water discharge to stream reaches was calculated by subtracting runoff
and interflow from streamflow.
nv07tac-4124_fig52
Acknowledgments References Cited

The development of GSFLOW was supported through Ahlfeld, D.P., Barlow, P.M., and Mulligan, A.E., 2005,
the USGS Ground-Water Resources Program, National GWM-a ground-water management process for the
Streamflow Information Program, the Watershed and River U.S. Geological Survey modular ground-water model
System Management Program, and National Research (MODFLOW-2000): U.S. Geological Survey Open-File
Program; the Pennsylvania Department of Environmental Report 2005-1072, 124 p.
Protection; and the Clearwater Conservancy of State College,
Pennsylvania. The authors thank George H. Leavesley (retired American National Standards Institute, 1992, American
USGS National Research Program, Lakewood, Colorado), national standard for programming languageFortran
Arlen Harbaugh (retired, USGS Office of Ground Water, Extended: X3.198-1992, 369 p.
Reston, Virginia), Wolfgang-Albert Fluegel (Friedrich-
Schiller-Universitat, Jena, Germany), Arlen Feldman (retired, Anderman, E.R., and Hill, M.C., 2000, MODFLOW-2000,
U.S. Army Corps of Engineers, Davis, California), Jon P. the U.S. Geological Survey modular ground-water
Fenske (U.S. Army Corps of Engineers, Davis, California), modelDocumentation of the Hydrogeologic-Unit Flow
and Douglas Boyle (Desert Research Institute, University of (HUF) Package: U.S. Geological Survey Open-File Report
Nevada, Reno, Nevada) for their support and encouragement 2000-342, 89 p.
throughout the development of GSFLOW. The Hydrologic Anderman, E.R., and Hill, M.C., 2003, MODFLOW-2000,
Research Center, U.S. Army Corps of Engineers provided the U.S. Geological Survey modular ground-water model
initial support in designing a method for connecting rainfall- Three additions to the Hydrogeologic-Unit Flow (HUF)
runoff models with MODFLOW, which led to the development Package: Alternative storage for the uppermost active cells,
of the Unsaturated-Zone Flow Package (Niswonger and flows in hydrogeologic units, and the hydraulic-conductivity
others, 2006a). Discussions with Greg Pohll (Desert Research depth-dependence (KDEP) capability: U.S. Geological
Institute, Reno, Nevada) led to important changes to the Survey Open-File Report 2003-347, 36 p.
conceptualization of GSFLOW. Roland Viger (USGS National
Research Program, Lakewood, Colorado) was instrumental Anderman, E.R., Kipp, K.L., Hill, M.C., Valstar, J., and
in developing programs using Geographic Information Neupauer, R.M., 2002, MODFLOW-2000, the U.S.
System (GIS) software to create input for the Sagehen Creek Geological Survey modular ground-water model
watershed example problem. Rose Medina (USGS, Carson Documentation of the model-layer variable-direction
City, Nevada) developed the three-dimensional diagrams of horizontal anisotropy (LVDA) capability of the
model input data and results used to illustrate the Sagehen Hydrogeologic-Unit Flow (HUF) Package: U.S. Geological
Creek watershed example. Survey Open-File Report 2002-409, 61 p.
The authors also thank Randall J. Hunt and Steven M.
Westenbroek (USGS, Middleton, Wisconsin) and John Fulton Anderson, E.A., 1968, Development and testing of snow pack
(USGS, Pittsburgh, Pennsylvania) and Dennis Risser (USGS, energy balance equations: Water Resources Research, v. 4,
New Cumberland, Pennsylvania) for their ideas and testing no. 1, p. 19-38.
of initial versions of GSFLOW using sample data sets. We Anderson, E.A., 1973, National Weather Service river
greatly appreciate the thorough reviews and comments that forecast systemSnow accumulation and ablation model:
greatly improved the quality of this report. These technical NOAA Tech. Memorandum NWS Hydro-17, U.S. Dept. of
reviewers were Arlen Harbaugh (retired, USGS Office of Commerce, Silver Spring, Maryland, p. 3-7.
Ground Water, Reston, Virginia), David M. Bjerklie (USGS,
East Hartford, Connecticut), Randall J. Hunt, John F. Walker, Anderson, E.I., 2005, Modeling groundwatersurface water
and Steven M. Westenbroek (USGS, Middleton, Wisconsin), interactions using the Dupuit approximation: Advances in
and Kathleen M. Flynn (USGS Office of Surface Water, Water Resources, v. 28, no. 4, p. 315-327.
Reston, Virginia).
Bauer H.H., and Vaccaro, J.J., 1987, Documentation of a deep
percolation model for estimating ground-water recharge:
U.S. Geological Survey Open-File Report 86-536, 180 p.
Bear, Jacob, 1972, Dynamics of fluids in porous media: New

York, American Elsevier Publishing Co., Inc., 764 p.
References Cited 125
Berris, S.N., Hess, G.W., and Bohman, L.R., 2001, River and Council, G.W., 1998, A lake package for MODFLOW, in
reservoir operation model, Truckee River basin, California Poeter, E., Zheng, C., and Hill, M.C., MODFLOW 98:
and Nevada, 1998: U.S. Geological Survey Water-Resources Colorado School of Mines, Golden, Colo., v. 2, p. 675-682.
Investigations Report 01-4017, 138 p.
Daly, C., Neilson, R.P., and Phillips, D.L., 1994, A statistical-
Beven, K.J., Kirkby, M.J., Schoffield, N., and Tagg, H., topographic model for mapping climatological precipitation
1984, Testing a physically-based flood forecasting model over mountainous terrain: Journal of Applied Meteorology,
(TOPMODEL) for three UK catchments: Journal of v. 33, p. 140-158.
Hydrology, v. 69, p. 119143.
Danskin, W.R., 1998, Evaluation of the hydrologic system
Bredehoeft, J.D., and Pinder, G.F., 1970, Digital analysis of and selected water-management alternatives in the Owens
areal flow in multiaquifer groundwater systems: a quasi Valley, California: U.S. Geological Survey Water-Supply
three-dimensional model: Water Resources Research, v. 6, Paper 2370-H, 175 p.
no. 3, p. 883-888.
Dean, J.D., and Snyder, W.M., 1977, Temporarily and areally
Brooks, R.H., and Corey, A.T., 1966, Properties of porous distributed rainfall: Proceedings of the American Society
media affecting fluid flow: Journal of Irrigation and of Civil Engineers, Journal of the Irrigation and Drainage
Drainage, v. 101, p. 85-92. Division, v. 103, no. IR2, p. 221-229.
Burden, R.L., and Faires, J.D., 1997, Numerical Analysis: Dickinson, W.T., and Whiteley, H.Q., 1970, Watershed
Pacific Grove, Calif., Brooks/Cole Publishing Company, areas contributing to runoff: International Association of
811 p. Hydrologic Sciences Publication 96, 1.12-11.28 p.
Burnett, J.L., and Jennings, C.W., 1965, Chico Quadrangle, Dunne, T., and Black, R.G., 1970, An experimental
scale 1:250,000: State of California, Division of Mines and investigation of runoff production in permeable soils: Water
Geology. Resources Research, v. 6, p. 478-490.
Charbeneau, R.J., 1984, Kinematic models for soil moisture Ely, D.M., 2006, Analysis of sensitivity of simulated recharge
and solute transport: Water Resources Research, v. 20, to selected parameters for seven watersheds modeled using
no.6, p. 699-706. the precipitation-runoff modeling system: U.S. Geological
Survey Scientific Investigations Report 2006-5041, 21 p.
Chen, Z., Govindaraju, R.S., and Kavvas, M.L., 1994,
Spatial averaging of unsaturated flow equations under Emerson, D.G., 1991, Documentation of a heat and water
infiltration conditions over areally heterogeneous fields-1. transfer model for seasonally frozen soils with application
Development of models: Water Resources Research, v. 30, to a precipitation-runoff model: U.S. Geological Survey
no. 2, p. 523-533. Open-File Report 91-462, 92 p.
Cheng, X., and Anderson, M.P., 1993, Numerical simulations Federer, A.C., and Lash, D., 1978, Brook: A hydrologic
of ground-water interaction with lakes allowing for simulation model for eastern forests: University of New
fluctuating lake levels: Ground Water, v. 31, no. 6, Hampshire, Water Resources Research Center, Research
p. 929-933. Report No. 19, 84 p.
Chow, V.T., Maidment, D.R., and Mays, L.W., 1988, Applied Fenske, J.P., Leake, S.A., and Prudic, D.E., 1996,
Hydrology: New York, New York, McGraw-Hill, Inc., Documentation of a computer program (RES1) to simulate
572 p. leakage from reservoirs using the modular finite-difference
ground-water flow model (MODFLOW): U.S. Geological
Clark, M.P., and Hay, L.E., 2004, Use of medium-range Survey Open-File Report 96-364, 51 p.
numerical weather prediction model output to produce
forecasts of streamflow: Journal of Hydrometeorology, v. 5, Fenske, J.P., and Prudic, D.E., 1998, Development of HMS/
no. 1, p. 15-32. MODFLOW for simulation of surface and groundwater
flow, in Poeter, E., Zheng, C., and Hill, M., MODFLOW
Colbeck, S.C., 1972, A theory of water percolation in snow: 98: Golden, Colo., Colorado School of Mines, v. 1,
Journal of Glaciology, v. 2, no. 63, p. 369-385. p. 463-470.
Cosner, O.J., and Harsh, J.F., 1978, Digital-model simulation Ford, L.R., and Fulkerson, D.R., 1956, Maximal flow through
of the glacial-outwash aquifer, Otter Creek Dry Creek a network: Canadian Journal of Mathematics, p. 399-404.
basin, Cortland County, New York: U.S. Geological Survey
Water-Resources Investigations Report 78-71, 34 p. Frank, E.C., and Lee, R., 1966, Potential solar beam
irradiation on slopes: U. S. Department of Agriculture,
Forest Service Research Paper RM-18, 116 p.
Fread, D.L., 1993, Flow routing, in Maidment, D.R., ed., Harbaugh, A.W., Banta, E.R., Hill, M.C., and McDonald,
Handbook of Hydrology: New York, McGraw-Hill, Inc., M.G., 2000, MODFLOW-2000, the U.S. Geological Survey
p. 36. modular ground-water model-modularization concepts and
the ground-water flow process: U.S. Geological Survey
Freeze, R.A., 1971, Three-dimensional, transient, saturated- Open-File Report 00-92, 121 p.
unsaturated flow in a groundwater basin: Water Resources
Research, v. 7, no. 2, p. 347-366. Harbaugh, A.W., and McDonald, M.G., 1996, Users
documentation for MODFLOW-96, and update to the U.S.
Freeze, R.A., 1972, Role of subsurface flow in generating Geological Survey modular finite-difference ground-water
overland runoff, 1Baseflow contributions to channel flow: flow model: U.S. Geological Survey Open-File Report
Water Resources Research, v. 8, no. 3, p. 609-623. 96-485, 56 p.
Freeze, R.A., and Cherry, J.A., 1979, Groundwater: Harter, T., and Hopmans, J.W., 2004, Role of vadose-
Englewood Cliffs, New Jersey, Prentice-Hall, Inc., 604 p. zone flow processes in regional scale hydrology-review,
Fulp, T.J., Vickers, W.B., Williams, B., and King, D.L., 1995, opportunities and challenges, in Feddes, R.A., De
Decision support for water resources management in the Rooij, G.H., and Van Dam, J.C., eds., Unsaturated Zone
Colorado River region, in Ahuja, L., Leppert, J., Rojas, K., ModelingProgress, Challenges and Applications:
and Seely, E., Workshop on computer applications in water Dordrecht, The Netherlands, Kluwer Academic Publisher,
management: Fort Collins, Colo., p. 24-27. Wageningen Frontis Series, p. 179-208.
Goode, D.J., and Appel, C.A., 1992, Finite-difference Hay, L.E., and Clark, M.P., 2003, Use of statistically and
interblock transmissivity for unconfined aquifers and for dynamically downscaled atmospheric model output for
aquifers having smoothly varying transmissivity: U.S. hydrologic simulations in three mountainous basins in the
Geological Survey Water-Resources Investigations Report western United States: Journal of Hydrology, v. 282,
92-4124, 79 p. p. 56-75.
Green, W.H., and Ampt, G.A., 1911, Studies on soil physics, Hay, L.E., Clark, M.P., Wilby, R.L., Gutowski, W.J.,
part I, the flow of air and water through soils: Journal of Leavesley, G.H., Pan, Z., Arritt, R.W., and Takle, E.S.,
Agricultural Science, v. 4, no. 1, p. 1-24. 2002, Use of regional climate model output for hydrologic
simulations: Journal of Hydrometeorology, v. 3, p. 571-590.
Guo, W., and Langevin, C.D., 2002, Users Guide to SEAWAT:
A computer program for simulation of three-dimensional Hay, L.E., Wilby, R.L., and Leavesley, G.H., 2000, A
variable-density ground-water flow: U.S. Geological Survey comparison of delta change and downscaled GCM scenarios
Techniques of Water-Resources Investigations, Book 6, for three mountainous basins in the United States: Journal
Chapter A7, 77 p. of the American Water Resources Association, v. 36,
p. 387-397.
Halford, K.J., and Hanson, R.T., 2002, User guide for the
drawdown-limited, multi-node well (MNW) package for Hewlett, J.D., and Nutter, W.L., 1970, The varying source
the U.S. Geological Surveys modular three-dimensional area of streamflow from upland basins, in Symposium
finite-difference ground-water flow model, versions on Interdisciplinary Aspects of Watershed Management,
MODFLOW-96 and MODFLOW-2000: U.S. Geological Montana State University, Bozeman, Mont., 1970
Survey Open-File Report 2002-293, 33 p. [Proceedings], p. 65-83.
Hamon, W.R., 1961, Estimating Potential evapotranspiration: Hill, M.C., 1990, Preconditioned Conjugate Gradient 2
Proceedings of the American Society of Civil Engineers, (PCG2), A computer program for solving ground-water
Journal of the Hydraulic Division, v. 87, no. HY3, flow equations: U.S. Geological Survey Water-Resources
p. 107-120. Investigations Report 90-4048, 43 p.
Harbaugh, A.W., 2005, MODFLOW-2005, the U.S.

Geological Survey modular ground-water model-the
Ground-Water Flow Process: U.S. Geological Survey
Techniques and Methods 6-A16, variously paginated.
Hill, M.C., Banta, E.R., and Harbaugh, A.W., 2000, Konikow, L.F., Goode, D.J., and Hornberger, G.Z., 1996, A
MODFLOW-2000, the U.S. Geological Survey Modular three-dimensional method-of-characteristics solute-transport
Ground-Water ModelUser Guide to the observations, model (MOC3D): U.S. Geological Survey Water-Resources
sensitivity, and parameter-estimation processes and three Investigations Report 96-4267, 87 p.
post-processing programs: U.S. Geological Survey Open-
File Report 00-184, 210 p. Langevin, C.D., Shoemaker, W.B., and Guo, W., 2003,
MODFLOW-2000, the U.S. Geological Survey modular
Hoffmann, John, Leake, S.A., Galloway, D.L., and Wilson, ground-water model-documentation of the SEAWAT-2000
A.M., 2003, MODFLOW-2000 Ground-Water Model version with the variable-density flow process (VDF) and
User Guide to the Subsidence and Aquifer-System the integrated MT3DMS transport process (IMT): U.S.
Compaction (SUB) Package: U.S. Geological Survey Open- Geological Survey Open-File Report 03-426, 43 p.
File Report 2003-233, 44 p.
Leaf, C.F., and Brink, G.E., 1973, Hydrologic simulation
Horton, R.E., 1933, The role of infiltration in the hydrological model of Colorado subalpine forest: U.S. Department of
cycle: American Geophysical Union Transactions, v. 23, Agriculture, Forest Service Research Paper RM-107, 23 p.
p. 479-482.
Leake, S.A., and Lilly, M.R., 1997, Documentation of a
Hsieh, P.A., and Freckleton, J.R., 1993, Documentation of a computer program (FHB1) for assignment of transient
computer program to simulate horizontal-flow barriers using specified-flow and specified-head boundaries, in
the U.S. Geological Surveys modular three-dimensional Applications of the Modular Finite-Difference Ground-
finite-difference ground-water flow model: U.S. Geological Water Flow Model (MODFLOW): U.S. Geological Survey
Survey Open-File Report 92-477, 32 p. Water-Resources Investigations Report 97-571, 50 p.
Hunt, R.J., and Steuer, J.J., 2000, Simulation of the recharge Leake, S.A., and Prudic, D.E., 1991, Documentation of a
area for Frederick Springs, Dane County, Wisconsin : U.S. computer program to simulate aquifer-system compaction
Geological Survey Water-Resources Investigations Report using the modular finite-difference ground-water flow
00-4172, 33 p. model: U.S. Geological Survey Techniques of Water-
Resources Investigations, Book 6, Chapter A2, 68 p.
Jensen, M.E., Rob, D.C.N., and Franzoy, C.E., 1969,
Scheduling irrigations using climate-crop-soil data, Leavesley, G.H., Lichty, R.W., Troutman, B.M., and Saindon,
National Conference on Water Resources Engineering of L.G., 1983, Precipitation-runoff modeling systemUsers
the American Society of Civil Engineers: New Orleans, La. manual: U.S. Geological Survey Water-Resources
[Proceedings], p. 20. Investigations Report 83-4238, 207 p.
Jeton, A.E., 1999, Precipitation-runoff simulations for Leavesley, G.H., Markstrom, S.L., Brewer, M.S., and Viger,
the upper part of the Truckee River Basin, California R.J., 1996a, The Modular Modeling System (MMS)The
and Nevada: U.S. Geological Survey Water-Resources physical process modeling component of a database-
Investigations Report 99-4282, 41 p. centered decision support system for water and power
management: Water, Air, and Soil Pollution, v. 90,
Jobson, H.E., and Harbaugh, A.W., 1999, Modifications to the p. 303-311.
diffusion analogy surface-water flow model (DAFLOW)
for coupling to the modlular finite-difference ground-water Leavesley, G.H., Markstrom, S.L., Viger, R.J., and Hay,
flow model (MODFLOW): U.S. Geological Survey Open- L.E., 2005, USGS Modular Modeling System (MMS)
File Report 99-217, 107 p. Precipitation-Runoff Modeling System (PRMS) MMS-
PRMS, in Singh, V., and Frevert, D., eds., Watershed
Kahrl, W.L. (ed.), 1979, The California Water Atlas: North Models: Boca Raton, Fla., CRC Press, p. 159-177.
Highlands, Calif., Governers Office of Planning and
Research, State of California, 118 p. Leavesley, G.H., Restrepo, P.J., Markstrom, S.L., Dixon, M.,
and Stannard, L.G., 1996b, The Modular Modeling System
Koczot, K.M., Jeton, A.E., McGurk, B.J., and Dettinger, (MMS): Users manual: U.S. Geological Survey Open-File
M.D., 2005, Precipitation-runoff processes in the Feather Report 96-151, 142 p.
River Basin, northeastern California, with prospects for
streamflow predictability, water years 1971-97: U.S. Leavesley, G.H., and Stannard, L.G., 1995, The precipitation-
Geological Survey Scientific Investigations Report runoff modeling systemPRMS, in Singh, V.P., ed.,
2004-5202, 82 p. Computer Models of Watershed Hydrology: Highlands
Ranch, Colo., Water Resources Publications, p. 281-310.
Lee, R., 1963, Evaluation of solar beam irradiation as a Mehl, S.W., and Hill, M.C., 2006, MODFLOW-2005, the
climatic parameter of mountain watersheds: Colorado State U.S. Geological Survey modular ground-water model
University Hydrology Papers, no. 2, 50 p. Documentation of shared node local grid refinement (LGR)
and the boundary flow and head (BFH) package: U.S.
Leopold, L.B., Wolman, M.G., and Miller, J.P., 1992, Fluvial Geological Survey Techniques and Methods 6-A12, 78 p.
Processes in Geomorphology (2d ed.): Mineola, New York,
Dover Publications, Inc., 522 p. Meinzer, O.E., 1923, Outline of ground-water hydrology, with
definitions: U.S. Geological Survey Water-Supply Paper
Lighthill, M.J., and Whitham, G.B., 1955, On kinematic 494, 71 p.
floodsflood movements in long rivers: Proceedings, R.
Soc. London, v. A220, p. 281-316. Merritt, M.L., and Konikow, L.F., 2000, Documentation of a
computer program to simulate lake-aquifer interaction using
Lindgren, R.J., and Landon, M.K., 1999, Effects of ground- the MODFLOW ground-water flow model and the MOC3D
water withdrawals on the Rock River and associated valley solute-transport model: U.S. Geological Survey Water-
aquifer, eastern Rock County, Minnesota: U.S. Geological Resources Investigations Report 00-4167, 146 p.
Survey Water-Resources Investigations Report 98-4157,
103 p. Morgan, D.S., 1988, Geohydrology and numerical model
analysis of ground-water flow in the Goose Lake Basin,
Linsley, R.K., Kohler, M.A., and Paulhus, J.L.H., 1975, Oregon and California: U.S. Geological Survey Water-
Hydrology for Engineers: New York, McGraw-Hill, Inc., Resources Investigations Report 87-4058, 92 p.
482 p.
Murray, F.W., 1967, On the computation of saturation vapor
Lohman, S.W., and others, 1972, Definitions of selected pressure: Journal of Applied Meteorology, v. 6, p. 203-204.
ground-water terms-revisions and conceptual refinements:
U.S. Geological Survey Water-Supply Paper, 21 p. Nash, J.E., and Sutcliffe, J.V., 1970, River flow forecasting
through conceptual models part IA discussion of
Male, D.H., and Gray, D.M., 1981, Snowcover ablation and principles: Journal of Hydrology, v. 10, no. 3, p. 282-290.
runoff, in Gray, D.M., and Male, D.H., eds., Handbook of
snow: Ontario, Pergamon Press Canada Ltd., p. 360-430. Nishikawa, T., Izbicki, J.A., Hevesi, J.A., Stamos, C.L.,
and Martin, P., 2005, Evaluation of geohydrologic
Mantoglou, A., 1992, A theoretical approach for modeling framework, recharge estimates, and ground-water flow of
unsaturated flow in spatially variable soilsEffective the Joshua Tree area, San Bernardino County, California:
flow models in finite domains and nonstationarity: Water U.S. Geological Survey Scientific Investigations Report
Resources Research, v. 28, no. 1, p. 251-267. 20045267, 127 p.
Mast, M.A., and Clow, D.W., 2000, Environmental Niswonger, R.G., and Prudic, D.E., 2005, Documentation
characteristics and water-quality of Hydrologic Benchmark of the Streamflow-Routing (SFR2) Package to include
Network stations in the Western United States: U.S. unsaturated flow beneath streamsA modification to
Geological Survey Circular 1173-D, 114 p. SFR1: U.S. Geological Survey Techniques and Methods
Mastin, M.C., and Vaccaro, J.J., 2002, Documentation of 6-A13, 62p.
Precipitation Runoff Modeling System modules for the Niswonger, R.G., Prudic, D.E., and Regan, R.S., 2006a,
Modular Modeling System modified for the Watershed Documentation of the Unsaturated-Zone Flow (UZF1)
and River Systems Management Program: U.S. Geological Package for modeling unsaturated flow between the land
Survey Open-File Report 2002-362, 5 p. surface and the water table with MODFLOW-2005: U.S.
McDonald, M.G., and Harbaugh, A.W., 1988, A modular Geological Survey Techniques and Methods 6-A19, 74 p.
three-dimensional finite-difference ground-water flow Niswonger, R.G., Prudic, D.E., Markstrom, L.S., Regan,
model: U.S. Geological Survey Techniques of Water- R.S., and Viger, R.J., 2006b, GSFLOWA basin-scale
Resources Investigations, Book 6, Chapter A-1, 586 p. model for coupled simulation of ground-water and surface-
Meeus, J., 1999, Astronomical Algorithms: Richmond, Va., water flowPart B. Concepts for modeling saturated and
Willmann-Bell, Inc., 477 p. unsaturated subsurface flow with the U.S. Geological
Survey modular ground-water model: Joint Federal
Mehl, S.W., and Hill, M.C., 2001, MODFLOW-2000, the Interagency Conference, Reno, Nev., 8 p.
U.S. Geological Survey modular ground-water model
User Guide to the Link-AMG (LMG) Package for solving Obled, C., and Rosse, B.B., 1977, Mathematical models of a
matrix equations using an algebraic multigrid solver: U.S. melting snowpack at an index plot: Journal of Hydrology,
Geological Survey Open-File Report 01-177, 33 p. v.32, p. 139-163.
Olson, S.A., 2002, Flow-frequency characteristics of Vermont Said, A., Stevens, O.K., and Sehlke, G., 2005, Estimating
streams: U.S. Geological Survey Water-Resources water budget in a regional aquifer using HSFP-MODFLOW
Investigations Report 2002-4238, 47 p. integrated model: Journal of the American Water Resources
Association, v. 41, no. 1, p. 55-66.
Panday, S., and Huyakorn, P.S., 2004, A fully coupled
physically-based spatially-distributed model for evaluating Schmid, W., Hanson, R.T., Maddock, T.I., and Leake, S.A.,
surface/subsurface flow: Advances in Water Resources, 2006, User guide for the farm process (FMP1) for the U.S.
v.27, no. 4, p. 361-382. Geological Surveys modular three-dimensional finite-
difference ground-water flow model, MODFLOW-2000:
Phillip, J.R., 1957, The theory of infiltration: 2. The profile at U.S. Geological Survey Techniques and Methods 6-A17,
infinity: Soil Science, v. 83, p. 435-448. 127 p.
Pikul, M.F., Street, R.L., and Remson, I., 1974, A numerical Selby, S.M., 1970, Standard mathematical tables (15th ed.):
model based on coupled one-dimensional Richards and Cleveland, Ohio, The Chemical Ruber Co., 664 p.
Boussinesq equations: Water Resources Research, v. 10,
no.2, p. 295-302. Selker, J..S., Keller, C.K., and McCord, J.T., 1999, Vadose
zone processes: Boca Raton, Fla., CRC Press LLC, 339 p.
Prudic, D.E., 1989, Documentation of a computer program to
simulate stream-aquifer relations using a modular, finite- Shen, H.W., and Julien, P.Y., 1993, Erosion and sediment
difference, ground-water flow model: U.S. Geological transport, in Maidment, D.R., ed., Handbook of hydrology:
Survey Open-File Report 88-729 113 p. New York, McGraw-Hill, Inc., p. 61.
Prudic, D.E., Konikow, L.F., and Banta, E.R., 2004, A new Smith, R.E., 1983, Approximate sediment water movement by
Streamflow-Routing (SFR1) Package to simulate stream- kinematic characteristics: Soil Science Society of America
aquifer interaction with MODFLOW-2000: U.S. Geological Journal, v. 47, p. 3-8.
Survey Open-File Report 2004-1042, 95 p.
Smith, R.E., and Hebbert, R.H.B., 1983, Mathematical
Rademacher, L.K., Clark, J.F., Clow, D.W., and Hudson, G.B., simulation of interdependent surface and subsurface
2005, Old groundwater influence on stream hydrochemistry hydrologic processes: Water Resources Research, v. 19,
and catchment response times in a small Sierra Nevada no.4, p. 987-1001.
catchment: Sagehen Creek, California: Water Resources
Research, v. 41, no. W02004, doi:10.1029/2003WR002805, Sophocleous, M., and Perkins, S.P., 2000, Methodology and
10 p. application of combined watershed and groundwater model
in Kansas: Journal of Hydrology, v. 236, p. 185-201.
Rankl, J.G., 1987, Analysis of sediment production from two
small semiarid basins in Wyoming: U.S. Geological Survey Steuer, J.J., and Hunt, R.J., 2001, Use of a watershed-
Water-Resources Investigations Report 85-4314, 27 p. modeling approach to assess hydrologic effects of
urbanization, North Fork Pheasant Branch basin near
Refsgaard, J.C., and Storm, B., 1995, MIKE SHE, in Singh, Middleton, Wisconsin: U.S. Geological Survey Water-
V.P., ed., Computer Models of Watershed Hydrology: Resources Investigations Report 2001-4113, 49 p.
Highlands Ranch, Colo., Water Resources Publications,
p.809-846. Streeter, V.L., and Wylie, E.B., 1985, Fluid Mechanics (8th
ed.): New York, McGraw-Hill, Inc., 586 p.
Riley, J.P., Israelsen, E.K., and Eggleston, K.O., 1973, Some
approaches to snowmelt prediction, in The role of snow and Sun, R.J., ed., 1986, Regional Aquifer-System Analysis
ice in hydrology: International Association of Hydrological Program of the U.S. Geological SurveySummary of
Sciences Publication 107, p. 956-971. projects, 1978-84: U.S. Geological Survey Circular 1002,
264 p.
Rorabaugh, M.I., 1953, Graphical and theoretical analysis
of step-drawdown test of artesian well: Proceedings of the Swain, E.D., and Wexler, E.J., 1996, A coupled surface-
American Society of Civil Engineers, v. 79, no. 362, 23 p. water and ground-water flow model (MODBRANCH) for
simulation of stream-aquifer interaction: U.S. Geological
Ross, M.A., Tara, R.D., Geurink, J.S., and Stewart, M.T., Survey Techniques of Water-Resources Investigations, Book
1997, FIPR Hydrologic Model Users Manual and Technical 6, Chapter A6, 125 p.
Documentation: Florida Institute of Phosphate Research,
FIPR-OFR-88-03-085, Barton, Fla. Swift, L.W., Jr., 1976, Algorithm for solar radiation on
mountain slopes: Water Resources Research, v. 12, no. 1,
p.108-112.
Tangborn, W.V., 1978, A model to predict short-term Vining, K.C., 2002, Simulation of streamflow and wetland
snowmelt runoff using synoptic observations of streamflow, storage, Starkweather Coulee subbasin, North Dakota, water
temperature, and precipitation in Colbeck, S.C., and years 1981-98: U.S. Geological Survey Water-Resources
Ray, M., eds., Modeling of snow cover runoff, Hanover, Investigations Report 02-4113, 28 p.
New Hampshire, 1978 [Proceedings]: U.S. Army Corps
of Engineers, Cold Region Research and Engineering Wang, H.F., and Anderson, M.P., 1982, Introduction to
Laboratory, p. 414426. groundwater modeling: San Francisco, W.H. Freeman and
Company, 237 p.
Thompson, E.S., 1976, Computation of solar radiation from
sky cover: Water Resources Research, v. 12, no. 5, p. Western Region Climate Center, 2006a, Western U.S.
859-865. Climate Historical Summaries, accessed January 15,
2006, at http://www.wcc.nrcs.usda.gov/snotel/snotel.
Thoms, R.B., Johnson, R.L., and Healy, R.W., 2006, Users pl?sitenum=539&state=ca.
guide to the Variably Saturated Flow (VSF) process for
MODFLOW: U.S. Geological Survey Techniques and Western Region Climate Center, 2006b, Western U.S.
Methods 6-A18, 58 p. Climate Historical Summaries, accessed January 15,
University of California Berkeley, 2005, Sagehen Creek pl?sitenum=541&state=ca.
Field Station, UC Berkeley, Historical Weather Data,
accessed June 15, 2005: at http://www.wrcc.dri.edu/cgi-bin/ Western Region Climate Center, 2006c, Western U.S.
rawMAIN.pl?casagh%20: Sagehen Creek, California. Climate Historical Summaries, accessed January 15,
U.S. Geological Survey, 2000, U.S. GeoData Digital Elevation pl?sitenum=540&state=ca.
Models: U.S. Geological Survey Fact Sheet 040-00,
accessed January 15, 2006 at http://pubs.er.usgs.gov/ Wilby, R.L., Hay, L.E., and Leavesley, G.H., 1999, A
usgspubs/fs/fs04000, 4 p. comparison of downscaled and raw GCM output:
implications for climate change scenarios in the San Juan
U.S. Soil Conservation Service, 1971, SCS national River basin, Colorado: Journal of Hydrology, v. 225, p.
engineering handbook, Section 4Hydrology: Washington, 67-91.
D.C., U.S. Government Printing Office, 654 p.
Wilson, J.D., and Naff, R.L., 2004, MODFLOW-2000, the
Vaccaro, J.J., 1992, Sensitivity of groundwater recharge U.S. Geological Survey modular ground-water model
estimates to climate variability and change, Columbia GMG Linear Equation Solving Package: U.S. Geological
Plateau, Washington: Journal of Geophysical Research, Survey Open-File Report 2004-1261, 47 p.
v.97, no. D3, p. 28212833.
Yates, D.N., Warner, T.T., and Leavesley, G.H., 2000,
Vaccaro, J.J., 2007, A deep percolation model for estimating Prediction of a flash flood in complex terrain: Part IIA
ground-water rechargeDocumentation of modules for the comparison of flood discharge simulations using rainfall
modular modeling system of the U.S. Geological Survey: input from radar, a dynamic model, and an automated
U.S. Geological Survey Scientific Investigations Report algorithmic system: Journal of Applied Meteorology, v. 39,
2006-5318, 30 p. no. 6, p. 815-825.
VanderKwaak, J.E., 1999, Numerical simulation of flow Zahner, R., 1967, Refinement in empirical functions for
and chemical transport in integrated surface-subsurface realistic soil-moisture regimes under forest cover, in Sopper,
hydrologic systems: Ontario, Canada, University of W.E., and Lull, H.W., eds., International Symposium of
Waterloo, Department of Earth Sciences, Ph.D. Dissertation, Forest Hydrology: New York, Pergamon Press, p. 261-274.
217 p.
Zarriello, P.J., and Ries, K.G., III, 2000, A precipitation-runoff
Viger, R.J., and Leavesley, G.H., 2007, The GIS Weasel Users model for analysis of the effects of water withdrawals
Manual: U.S. Geological Survey Techniques and Methods on streamflow, Ipswich River basin, Massachusetts: U.S.
6-B4, 201 p. Geological Survey Water- Resources Investigations Report
2000-4029, 99 p.
Appendix 1 131
Appendix 1. Input Instructions for GSFLOW.
Contents
Introduction.................................................................................................................................................133
Input Terminology.......................................................................................................................................133
Styles and Formats.....................................................................................................................................133
Modular Modeling System Based Files.........................................................................................133
MODFLOW Based Input Files..........................................................................................................134
GSFLOW Control File..................................................................................................................................134
Control Parameters Related to Model Input.................................................................................137
Control Parameters Related to Model Output...............................................................................137
Control Parameters Related to Initial Conditions for PRMS.......................................................139
PRMS Data File...........................................................................................................................................139
PRMS Parameter File.................................................................................................................................142
Parameter File Format.......................................................................................................................143
PRMS Modules..................................................................................................................................150
GSFLOW Modules..............................................................................................................................173
MODFLOW Input Files................................................................................................................................176
Name File............................................................................................................................................177
Basic Package Input Instructions...................................................................................................178
Block-Centered Flow Package........................................................................................................185
Layer-Property Flow Package.........................................................................................................188
Hydrogeologic-Unit Flow Package.................................................................................................190
Horizontal Flow Barrier Package....................................................................................................194
Well Package......................................................................................................................................195
Multi-Node Wells Package .............................................................................................................196
General-Head Boundary Package..................................................................................................199
Flow and Head Boundary Package................................................................................................200
Streamflow-Routing Package . .......................................................................................................202
Lake Package.....................................................................................................................................210
Unsaturated-Zone Flow Package ..................................................................................................213
Gage Package ...................................................................................................................................217
Strongly Implicit Procedure Package............................................................................................218
Preconditioned Conjugate-Gradient Package..............................................................................219
Direct Solver Package......................................................................................................................220
Geometric Multigrid Solver Package.............................................................................................222
Array Reading Utility Subroutines...................................................................................................224
List Utility Subroutine (ULSTRD)......................................................................................................226
This page left intentionally blank

Appendix 1 133
Introduction
This appendix describes the input data requirements for a GSFLOW simulation. Because GSFLOW is based on different
models, the format and contents of data in each file depends on whether the genesis of the file was for the PRMS or MODFLOW
model. The GSFLOW Control File, PRMS Data File(s), and PRMS Parameter File are based on the Modular Modeling System
input-data file formats (Leavesley and others, 1996b). All other files described in this section are based on MODFLOW input
file formats as documented for each MODFLOW package. The MODFLOW Name File is described separately from the
MODFLOW Packages because the file is used to identify the names of MODFLOW-based files that are read from and(or)
written to during a GSFLOW simulation.
Input Terminology
GSFLOW input is described on the basis of terminology used for PRMS and MODFLOW. The input terminology of
each program was adapted, where possible, for the purpose of consistency within GSFLOW. Thus, input terminology used
in this appendix may differ from that used in previous reports documenting PRMS, MODFLOW, and MODFLOW packages.
The terminology used to define data input for PRMS and MODFLOW differ because the models were developed by different
disciplines within the U.S. Geological Survey.
Styles and Formats

Consistent styles and formats for PRMS and MODFLOW input instructions are used for the following:
Variables and parameters that are directly related to user-supplied values are identified using 10-point, bold Courier
font.
Variables from model code not directly related to user-supplied values are identified using 10-point, Times font.
Example input variables and parameters are identified using 10-point, Courier font.
Optional input is enclosed between brackets ([ ]).

The following presents descriptions of the terms variable, parameter, dimension, and item in the context of input for PRMS
and MODFLOW, respectively.
Modular Modeling System Based Files

The Modular Modeling System input structure is used to define input data for the GSFLOW Control File, PRMS Data
File(s), and PRMS Parameter File. This input structure uses the following definitions.
Variables (or states and fluxes) may vary from one time step to another, and are either specified in the PRMS Data
File(s) (such as the daily maximum air temperature at a temperature station) or calculated (such as the soil-infiltration
rate for a HRU). Variables may have a single value or they may include multiple values (one- or two-dimensional arrays).
Parameters are user-specified data that do not change during a simulation, such as the name of the PRMS Parameter
File, the number of columns of maximum air temperature values in a PRMS Data File, and the area of a hydrologic
response unit (HRU) as specified in the GSFLOW Control File, PRMS Data File, and PRMS Parameter File,
respectively. Parameters, like variables, may have a single value or they may include multiple values (one- or two-
dimensional arrays).
Dimensions are a type of parameter that define the number of spatial features and constants, such as the number of
HRUs, number of months in a year, and the number of temperature stations for which time-series data are specified in the
PRMS Data File(s). Dimensions are specified in the PRMS Parameter File.
Items are groups of input values used to define a parameter, dimension, or variable. Items may be specified on a single
line or on multiple lines in a file.
MODFLOW Based Input Files

MODFLOW-2005 input structure is used to define input data for ground-water flow, unsaturated flow, and flow in streams
and lakes. The input structure of MODFLOW uses the following definitions.
Variables (or states and fluxes) may include a single value or multiple values. One-dimensional variables are multi-
valued variables in which the number of values is indicated by a single number in parentheses after the name. Two-
dimensional variables are multi-valued variables in which the number of values is indicated by two numbers in
parentheses after the name. The total number of values represented is the product of the two numbers.
Parameters are single values that can be used to determine data values for multiple finite-difference cells (Harbaugh,
2005, p. 8-2). Input instructions for defining parameters are not included in this appendix. Input instructions for
parameters are included in Harbaugh (2005).
Dimensions are a not used to refer to MODFLOW input data.
Items are numbered groups of input variables used to define a parameter, dimension, or variable. Items may be specified
on a single line or on multiple lines in a file.
GSFLOW Control File

The GSFLOW Control File consists of a header followed by a sequence of control-parameter items specified in the
Modular Modeling System Data File format. Each line in the Control File can include descriptive text to allow annotation of
any data, such as to describe the source of a parameter value. The descriptive text is added to lines following the required data
and preceded by at least one space and two forward slashes (//). Each line in the Control File can be up to 256 characters. The
header is a single line of text that can be used to identify the file. For example, the header used for Sagehen Creek watershed
example problem is:
GSFLOW Control File for the Sagehen Creek Watershed
Control-parameter items consist of four lines followed by parameter values, one per line, that have the following general
structure and order:
#### [ // descriptive_text]
NAME [ // descriptive_text]
N_VALUES [ // descriptive_text]
DATA_TYPE [ // descriptive_text]
VALUE(S) [ // descriptive_text]
The first line is used as a delimiter signaling the start of a control-parameter item and must specify a string of four pound
signs (####) beginning in column 1. The second line specifies the name of the control parameter (NAME). The third line
specifies the number (N_VALUES) of parameter values that are specified. The fourth line identifies the data type of the control
parameter using an integer flag (DATA_TYPE). The DATA_TYPE options are:
1 is for integer;
2 is for real (single-precision, floating decimal point);
3 is for double (double-precision, floating decimal point); and
4 is for character string.
Note, no double-precision real (option 3) control parameters are required in GSFLOW.
The next N_VALUES lines specify data values (VALUES(S)), one value per line. Thus, each control-parameter item must
consist of at least five lines. Table A1-1 is a list of the GSFLOW control parameters, and provides parameter names, definitions,
number of values, data-type flag, and whether the parameter is optional. Control-parameter names are case-sensitive and must be
specified as defined in table A1-1.
Appendix 1 135
Table A1-1. Control parameters specified in the GSFLOW Control File.

[Data Type: 1, integer; 2, single precision floating point (real); 3, double precision floating point (real); 4, character string. Abbreviation: HRU, hydrologic
response unit]
Number
Parameter name Data type Default value
Definition of values
(NAME) (Data_type) or optional
(N_value)
Parameters related to model execution
model_mode Model to run (GSFLOW, PRMS, MODFLOW) 1 4 GSFLOW
start_time1 Simulation start time specified in order as: year, month, 6 1 2000, 10, 1, 0, 0, 0
day, hour, minute, second
end_time1 Simulation end time specified in order as: year, month, 6 1 2001, 9, 30, 0, 0, 0
day, hour, minute, second
Parameters related to model input
2,3
data_file Pathname(s) for PRMS Data File(s); typically, a single Equal to the 4 prms.data
Data File is specified for a GSFLOW simulation number of
data files
param_file3 Pathname for PRMS Parameter File 1 4 prms.params
modflow_name3 Pathname for MODFLOW Name File 1 4 modflow.nam
precip_module1 Module name for precipitation-distribution method 1 4 precip_prms
(precip_prms, precip_laps_prms, xyz_dist, or
precip_dist2_prms)
temp_module1 Module name for temperature-distribution method 1 4 temp_1sta_prms
(temp_1sta_prms, temp_2sta_prms, xyz_dist, or
temp_dist2.prms)
solrad_module1 Module name for solar-radiation-distribution method 1 4 ddsolrad_hru_prms
(ccsolrad_hru_prms or ddsolrad_hru_prms)
et_module1 Module name for potential-evapotranspiration computation 1 4 potet_jh_prms
method (potet_hamon_prms, potet_jh_prms, or
potet.pan.prms)
srunoff_module1 Module name for surface-runoff/infiltration computation 1 4 srunoff_smidx_prms
method (srunoff_carea_prms or srunoff_smidx_prms)
Parameters related to model output
3
gsflow_output_file Pathname for GSFLOW Water-Budget File of summaries 1 4 gsflow.out
of each component of GSFLOW water budget
model_output_file3 Pathname for PRMS Water-Budget File of summaries of 1 4 prms.out
each component of PRMS water budget
csv_output_file3 Pathname for GSFLOW Comma-Separated-Values (CSV) 1 4 gsflow.csv
File of GSFLOW water budget and mass balance results
for each time step
gsf_rpt Switch to specify whether or not the GSFLOW Comma- 1 1 1
Separated-Values (CSV) File is generated (0=no; 1=yes)
rpt_days Frequency that summary tables are written to GSFLOW 1 1 7
Water-Budget File (0=none, >0 frequency in days, e.g.,
1=daily, 7=every 7th day)
statsON_OFF1 Switch to specify whether or not PRMS Statistic Variables 1 1 0
(statvar) File of selected time-series values is generated
(0=no; 1=yes)
stat_var_file3 Pathname for PRMS Statistic Variables (statvar) File 1 4 optional
of time-series values; required only when
statsON_OFF = 1
nstatVars1 Number of variables to include in PRMS Statistic 1 1 optional
Variables File and names specified in statVar_names;
required only when statsON_OFF = 1
Table A1-1. Control parameters specified in the GSFLOW Control File.Continued

[Data Type: 1, integer; 2, single precision floating point (real); 3, double precision floating point (real); 4, character string. Abbreviation: HRU, hydrologic
response unit]
Number
Parameter name Data type Default value
Definition of values
(NAME) (Data_type) or optional
(N_value)
Parameters related to model outputContinued
1
statVar_names List of variable names for which output is written to PRMS nstatVars 4 optional
Statistic Variables File; required only when
statsON_OFF = 1
statVar_element List of identification numbers corresponding to variables nstatVars 1 optional
specified in statVar_names file (1 to variables
dimension size); required only when statsON_OFF = 1
aniOutON_OFF1 Switch to specify whether or not PRMS Animation 1 1 0
Variables File(s) of spatially-distributed values is
generated (0=no; 1=yes)
ani_output_file1,3 Root pathname for PRMS Animation Variables File(s) 1 4 optional
to which a filename suffix based on dimension names
associated with selected variables is appended; required
only when gisOutOn_Off = 1
naniOutVars1 Number of output variables specified in the aniOutVar_ 1 1 optional
names list; required only when aniOutON_OFF = 1.
aniOutVar_names1 List of variable names for which all values of the variable naniOutVars 4 optional
(that is, the entire dimension size) for each time step are
written to PRMS Animation Variables File(s), use only
for aniOutON_OFF = 1
Parameters related to PRMS model initial conditions
1
init_vars_from_file Flag to determine if a PRMS Initial Conditions File is 1 1 0
specified as an input file (0=no; 1=yes)
var_init_file1,3 Pathname for the PRMS Initial Conditions File; only 1 4 optional
required when init_vars_from_file = 1
save_vars_to_file1 Flag to determine if a PRMS Initial Conditions File (var_ 1 1 0
save_file) will be generated at the end of simulation
(0=no; 1=yes)
var_save_file1,3 Pathname for the PRMS Initial Conditions File to be 1 4 optional
generated at end of simulation; only required when
save_vars_to_file = 1
1
Additional description of parameter provided in appendix 1 section Control Parameters Related to Model Input.
2
Multiple PRMS Data Files can be specified, although typically, only one is used with GSFLOW.
3
Pathnames can be 1 to 256 characters and must be specified as a valid pathname for the operating system.
An example control-parameter item follows. The example is for the control parameter with the NAME set to data_file
that is used to specify the pathname(s) of the PRMS Data File(s). The parameter consists of a single value (N_VALUES set to 1)
with a value specified as a character string (DATA_TYPE set to 4).
#### // PRMS Data File of time-series values

data_file
1
4
D:/gsflow/data/sagehen/input/sagehen.data
Although this example specifies that the data_file parameter is a single value, multiple PRMS Data Files can be
specified for a GSFLOW simulation by specifying N_VALUES greater than one and including that number of pathnames.
Appendix 1 137
Control-parameter items can be specified in any order in the GSFLOW Control File; table A1-1 provides a suggested order
for specifying control-parameter items grouped by whether the parameter controls functions related to (a) model execution,
(b) input, (c) output, or (d) initial conditions. Any control-parameter item not specified in the Control File that is required by a
GSFLOW simulation is assigned default values. Any control-parameter item specified in the Control File that is not required
by a GSFLOW simulation is ignored. Additional information for selected control parameters beyond the definitions provided in
table A1-1 follows.
Control Parameters Related to Model Input
start_time and end_timeare used to specify the starting and ending times of the simulation. They each are specified
using six lines of integer values in the following order: year, month, day, hour, minute, and second. The specified times
must have corresponding time-series data items of the exact same date and time in the PRMS Data File(s). GSFLOW
only allows a daily time step.
data_fileis used to specify a list of PRMS Data File(s) to use in a simulation. Typically, a single PRMS Data File is used
for GSFLOW simulations (N_VALUES set to 1). Use of multiple PRMS Data Files is described in section, PRMS Data
File in this appendix.
precip_module, temp_module, solrad_module, et_module, and srunoff_moduleare used to select a

PRMS module for simulation processes to distribute precipitation, temperature, and solar radiation and to compute
potential evapotranspiration and surface runoff, respectively. The computations for each process are described in section,
Computations of Flow in the body of this report and input requirements for each module are described below in this
appendix. If the module xyz_dist is selected for the temperature-distribution method (temp_module), this module also
must be selected as the precipitation-distribution module (precip_module), and vice versa.
Control Parameters Related to Model Output
statVar_names and aniOutVar_namesare used to specify lists of names of output variables (states and fluxes)
to include in the PRMS Statistic Variables (statvar) File and PRMS Animation Variables File(s), respectively. A
description of these files is given in section, Output Files in the body of this report. A list with definitions of the 35
most commonly used GSFLOW output variables is given in table A1-2. Many additional output variables are available
for inclusion in the PRMS Statistic Variables and Animation Variables Files for specialized analysis and debugging
purposes. The complete list of the 270 output variables with definitions can be found in the file gsflow.var_name that is
provided in the gsflow/bin directory included in the GSFLOW distribution. Users can select as many output variables
as desired to include in both files using control parameters nstatVars and naniOutVars, respectively. If control
parameter statsON_OFF is set to 0, the Statistic Variables File is not created and values of statVar_names and
nstatVars are ignored. Likewise, if control parameter aniOutON_OFF is set to 0, PRMS Animation Variables
File(s) are not created and values of aniOutVar_names and naniOutVars are ignored.
ani_output_fileis used as the root filename for PRMS Animation Variables File(s). A separate output file is
generated for each dimension size associated with the list of variables specified by aniOutVar_names when
aniOutON_OFF is set to 1. Variables with the same dimension are included in a single file. The names of each file
differ in the suffix appended to ani_output_file. The suffix for each file is set to the dimension name. For example,
if ani_output_file is specified as ./output/sagehen_ani and variables with PRMS dimension nhru and nreach
are specified in the aniOutVar_names list of variable names, the files ./output/sagehen_ani.nhru and
./output/sagehen_ani.nreach are generated.
Table A1-2. Selected GSFLOW variables for which values can be written to the PRMS Statistic Variables File and PRMS Animation
Variables File(s) for each simulation time step.
[Dimension variable: nhru, number of HRUs; nhrucell, number of intersections between HRUs and MODFLOW grid cells; one, dimension of one;
Abbreviations: MF_L, MODFLOW length unit; HRU, hydrologic response unit; cfs, cubic feet per second; T, time unit]
Dimension
Variable name Definition Units
variable
basin_cfs Streamflow out of watershed cfs one
basin_et Total evapotranspiration on watershed as sum for evaporation from snowpack, inches one
impervious areas, plant canopy, and soil zone and transpiration from soil zone
basin_gwflow_cfs Area-weighted average ground-water flow for watershed cfs one
basin_potet Area-weighted average potential evapotranspiration for watershed inches one
basin_ppt Area-weighted average precipitation for watershed inches one
basin_pweqv Area-weighted average pack-water equivalent of snowpack for watershed inches one
basin_reach_latflow Area-weighted average lateral flow into stream reaches for watershed cfs one
basin_sroff_cfs Area-weighted average Hortonian and Dunnian surface runoff into stream reaches cfs one
for watershed
basin_ssflow_cfs Area-weighted average interflow into stream reaches for watershed cfs one
basinactet Volumetric flow rate of evapotranspiration for watershed (MF_L)3T-1 one
basingravstor Total volume of soil water in gravity reservoirs of soil zone for watershed (MF_L)3 one
basingw2sz Volumetric flow rate of ground-water discharge added to soil zone for watershed (MF_L)3T-1 one
basininfilprev Volumetric flow rate of soil infiltration into preferential-flow reservoirs of soil (MF_L)3T-1 one
zone including precipitation, snowmelt, and cascading Hortonian flow for
watershed
basininfil_tot Volumetric flow rate of soil infiltration into capillary reservoirs of soil zone (MF_L)3T-1 one
including precipitation, snowmelt, and cascading Hortonian flow for watershed
basininterflow Volumetric flow rate of slow interflow to stream reaches for watershed (MF_L)3T-1 one
basinprefstor Total volume of soil water in preferential-flow reservoirs of soil zone for (MF_L)3 one
watershed
basinpweqv Total volume of water in snowpack storage for watershed (MF_L)3 one
basinsnowevap Volumetric flow rate of snowpack sublimation for watershed (MF_L)3T-1 one
basinsnowmelt Volumetric flow rate of snowmelt for watershed (MF_L)3T-1 one
basinsoilmoist Total volume of soil water in capillary reservoirs of soil zone for watershed (MF_L)3 one
basinsroff Volumetric flow rate of Hortonian and Dunnian surface runoff for watershed (MF_L)3T-1 one
basinstrmflow Volumetric flow rate of streamflow leaving the watershed (MF_L)3T-1 one
gw2sm Average ground-water discharge to soil zone in an HRU inches nhru
gwc_head Head at each MODFLOW ground-water cell MF_L ngwcell
gwflow2strms Volumetric flow rate of ground-water discharge to stream reaches (MF_L)3T-1 one
hru_ppt Adjusted precipitation on HRU inches nhru
kkiter Current iteration in GSFLOW simulation dimensionless one
obsq_cfs Streamflow at streamflow-gaging station cfs nobs
pkwater_equiv Pack-water equivalent of snowpack inches nhru
reach_cfs Streamflow leaving each stream reach cfs nreach
reach_latflow Lateral flow (surface runoff and interflow) into each stream reach cfs nreach
reach_wse Water-surface elevation in each stream reach MF_L nreach
sat_store Total storage in saturated MODFLOW cells (MF_L)3 one
sm2gw_grav Gravity drainage from each gravity reservoir to each MODFLOW cell inches nhrucell
snowcov_area Fraction of snow-covered area on HRU dimensionless nhru
snowmelt Snowmelt from the snowpack on HRU inches nhru
soil_moist Water content of capillary reservoir for HRU inches nhru
soil_moisture_pct Decimal fraction of the saturation of capillary reservoir dimensionless nhru
sroff Surface runoff to streams for HRU inches nhru
ssr_to_gw Area-weighted average gravity drainage from soil zone for HRU inches nhru
ssres_flow Interflow to streams for HRU inches nhru
ssres_stor Average gravity reservoir storage for HRU inches nhru
stream_leakage Total leakage from stream segments to associated MODFLOW cells (MF_L)3 one
swrad Computed shortwave radiation for HRU langleys nhru
Appendix 1 139
Table A1-2. Selected GSFLOW variables for which values can be written to the PRMS Statistic Variables File and PRMS Animation
Variables File(s) for each simulation time step.Continued
[Dimension variable: nhru, number of HRUs; nhrucell, number of intersections between HRUs and MODFLOW grid cells; one, dimension of one.
Abbreviations: MF_L, MODFLOW length unit; HRU, hydrologic response unit]
Dimension
Variable name Definition Units
variable
tmaxf Adjusted daily maximum temperature for HRU degrees nhru
Fahrenheit
tminf Adjusted daily minimum temperature for HRU degrees nhru
Fahrenheit
unsat_store Total storage in unsaturated MODFLOW cells as simulated by the (MF_L)3 one
uzf_infil Net gravity drainage to the unsaturated zone as simulated by the (MF_L)3 one
Control Parameters Related to Initial Conditions for PRMS
init_vars_from_file, var_init_file, save_vars_to_file, var_save_fileare used to specify whether

or not the PRMS initial conditions for all computation variables (or states) will be specified and(or) saved for use
with a GSFLOW simulation. Most initial condition values are set to 0.0 internally by PRMS modules, with some
initial conditions set to values specified in the PRMS Parameter File, such as the initial water content of the soil zone
(parameters soil_moist_init, soil_rechr_init, and ssstor_init; see section, PRMS Parameter File
in this appendix for an explanation of these and other parameters that are required for each GSFLOW and PRMS
module). Typically, the control parameters init_vars_from_file and save_vars_to_file are specified as
0 for GSFLOW simulations.
If the user has a valid PRMS Initial Conditions File that was generated during a prior GSFLOW simulation, it can be
selected for use by setting init_vars_from_file to 1 and specifying the pathname of the PRMS Initial Conditions
File using control parameter var_init_file. To generate the PRMS Initial Conditions File, the user would set
save_vars_to_file to 1 and the pathname using control parameter var_save_file.
One use of the PRMS Initial Conditions File is for forecasting hydrologic conditions. A series of predictive simulations
can be generated using the results from a single simulation driven by climate data up to a selected date. GSFLOW could
use these results to set the internal states for new simulations that are driven by different sets of predicted climate data. The
predicted climate data, for example, results from a climate model or historical climate data, can be specified in additional
PRMS Data File(s).
The PRMS Initial Conditions File created from a GSFLOW simulation only includes the variables (states) for the date and
time at the end of the simulation. Thus, the file should only be used as a continuation from that date and time.
PRMS Data File

A PRMS Data File consists of three items: a header, input-variable declaration items, and time-series data items (fig.A1-1
illustrates part of a PRMS Data File), and is specified in the Modular Modeling System Data File format. The number of Data
Files and file names of each PRMS Data File is specified using the data_file control-parameter item in the GSFLOW
Control File. If multiple PRMS Data Files are specified for use in a single GSFLOW simulation the files will be combined
according to the specified date and time of each time-series data item. If data overlap, values for the smallest time increment
between consecutive values during the overlap period are used, with the last PRMS Data File having precedence. Multiple
PRMS Data Files can be used to separate data by years or decades. Multiple PRMS Data Files must have identical input-
variable declaration items in sequence and number but may have different comment lines. The time increment between two
consecutive time-series data items sets the GSFLOW time step. GSFLOW simulates on a daily time step, therefore, time-series
data items must be specified for each day with exactly 24 hours between the date and time specified for consecutive values.
Thus, PRMS Data Files with time increments other than 24 hours cannot be used in a GSFLOW simulation.
Sagehen Data File: Independence Lake and Sagehen Creek data stations
// tmax stations are:

// INDEPENDENCE LAKE
// SAGEHEN CREEK
tmax 2
// tmin stations are:

// SAGEHEN CREEK
tmin 2
// precip stations are:

// SAGEHEN CREEK
precip 2
form_data 0
solrad 0
pan_evap 0
// runoff value is:

// 10343500 SAGEHEN C NR TRUCKEE C
runoff 1
#######################################################################
1980 10 1 0 0 0 -901.0 85.0 -901.0 30.0 0.00 0.00 2.3
1980 10 2 0 0 0 -901.0 81.0 -901.0 32.0 0.00 0.00 2.3
1980 10 3 0 0 0 -901.0 83.0 -901.0 30.0 0.00 0.00 2.3
**remaining lines have same format as the three previous lines and are not shown**
Figure A1-1. Example of a PRMS Data File.
The set of PRMS Data Files determines the range of time of a GSFLOW simulation, but not the simulation time period,
which is specified by control-parameters items start_time and end_time in the GSFLOW Control File. Both
start_time and end_time must be specified as dates and times that occurred between the earliest beginning and latest
ending time of the time-series data specified in the set of PRMS Data Files. Typically, all time-series data are contained in a
single PRMS Data File.
The header item in the PRMS Data File is a single line of text, up to 256 characters in length, which can be used to identify
the file. For example, in the Sagehen Creek watershed example problem, the following header is used:
Sagehen Data File: Independence Lake and Sagehen Creek data stations
Input-variable declaration items are used to specify the type of time-series data included in each PRMS Data File. One line
is used to identify each time-series data item. Blank lines or comment lines can be included before or after an input-variable
declaration item; comment lines begin with two backslashes (//) in columns 1 and 2 to add descriptive information about an
input variable. Each item specifies two values: (1) a character string that is the name of the input variable; and (2) an integer
value that is the number of values (or columns) specified in each time-series data item for the input variable. The number of
values must equal the size of the dimension associated with the input variable as defined by the dimensions specified in the
PRMS Parameter File. For example, the two input-variable declarations items
tmax 2
tmin 2
indicate that daily maximum (tmax) and minimum (tmin) air temperature data will be specified for two observation stations
nv07tac-4124_figA1-1
in each time-series data item (and dimension ntemp is set to 2 in the dimensions section of the PRMS Parameter File). Each
time-series data item must contain two columns of tmax values followed by two columns of tmin values. If the number of
values and the associated dimension specified in the PRMS Parameter File are different an error message is printed and model
execution stops.
Appendix 1 141
A delimiter that consists of a single line specifying at least four pound symbols (####; fig. A1) beginning in column1
signals the end of the input-variable declaration items and that the next and following lines each specify a time-series data item
for consecutive dates and times. Each time-series data item consists of columns of data separated by at least one blank space,
with the number of columns equal to six plus the sum of the number of values for each input variable. The first six columns
specify the date and time of each data item as integer values in the order: year, month, day, hour, minute, and second. For
GSFLOW integrated simulations, a time-series data item must be specified for each day with exactly 24 hours between the
date and time specified for consecutive lines for the full extent of the time-series data. The hour, minute, and second must be
specified as zero (that is, columns 4 through 6 can be 0 0 0). The remaining columns for each time-series data item specify
the data values, which must be specified in the sequence the input-variable declarations items were defined. Value(s), depending
on the specified number of values for the variable, must be specified in the corresponding column for each input variable. Extra
values on any time-series data item beyond the specified number of columns (6 + the sum of the number of values for each input
variable) are ignored.
Table A1-3 is a list of valid time-series input variables for GSFLOW, and provides definitions, units, valid range, and the
dimension name associated with each variable. The input-variable names are case-sensitive and must be specified as defined in
table A1-3.
A portion of a PRMS Data File is shown in figure A1-1. The input-variable declarations items specify that values for the
following input variables are included or are not included in each time-series data item:
two air-temperature stations with the two maximum air-temperature values specified before the two minimum air-
temperature values (input variables tmax and tmin are set to 2 and dimension ntemp is set to 2 in the PRMS
Parameter File);
two precipitation stations (input variable precip is set to 2 and dimension nrain is set to 2 in the PRMS Parameter
File);
no data are included to specify the form of precipitation (input variable form_data is set to 0 and dimension nform is
set to 0 in the PRMS Parameter File);
no data are included to specify solar radiation (input variable solrad is set to 0 and dimension nsol is set to 0 in the
PRMS Parameter File);
no data are included to specify pan evaporation (input variable pan_evap is set to 0 and dimension nevap is set to 0
in the PRMS Parameter File); and
one streamflow station (input variable runoff is set to 1 and dimension nobs is set to 1 in the PRMS Parameter File).
Thus, each time-series data item (only three are shown in fig. A1-1) specifies 6 columns for the date and time, followed by two
maximum air-temperature values, two minimum air-temperature values, two precipitation values, and one streamflow value,
for a total of 13 columns of data values on each line. For the time period shown in figure A1-1, no data were available for the
first air-temperature measurement station, thus a value of -901.0 was specified as the first maximum and first minimum air
temperatures (columns 7 and 9).
Table A1-3. Time-series data that can be specified in a PRMS Data File.
[Dimension variable: nevap, number of measurement stations that measure pan evaporation; nobs, number of streamflow gaging stations; nrain, number of
measurement stations that measure precipitation; nsol, number of measurement stations that measure solar radiation; ntemp, number of measurement stations
that measure air temperature; nform, is either 0 or 1. Abbreviation: cfs: cubic feet per second]
Dimension
Variable name Definition Units Valid range
variable
pan_evap Pan evaporation at each measurement station that measures pan evaporation inches greater than 0.0 nevap
runoff1 Streamflow at each streamflow-gaging station cfs greater than 0.0 nobs
precip1 Precipitation at each measurement station that measures precipitation inches greater than 0.0 nrain
solrad Solar radiation at each measurement station that measures solar radiation langleys greater than 0.0 nsol
tmax1 Daily maximum air temperature at each measurement station that measures air degrees Celsius or -50 to 150 ntemp
temperature Fahrenheit
tmin1 Daily minimum air temperature at each measurement station that measures air degrees Celsius or -50 to 150 ntemp
temperature Fahrenheit
form_data1 Form of precipitation (0=not known; 1=snow; 2=rain) dimensionless 0, 1, or 2 nform
rain_day Day is treated as a rain day (0=no; 1=yes) dimensionless 0 or 1 one
1
Additional description of variable provided below.
Additional information for selected input variables beyond the definitions provided in table A1-3 follows.
runoffis the measured streamflow data and can be specified in units of cubic feet per second or cubic meters per second.
Set the parameter runoff_units to a value of 1 in the parameters section of the PRMS Parameter File to indicate that
runoff values are specified in cubic meters per second. Measured streamflow is not used in simulation computations and
are thus optional. They are written to the GSFLOW Comma-Separated Values File and optionally can be written to the
PRMS Statistic Variables and Animation Variables Files for dimension nobs, if included in the output variables list for
either file.
tmax and tminare the measured daily maximum and minimum air temperatures and can be specified in units of degrees
Fahrenheit or Celsius. Set the parameter temp_units to a value of 1 in the parameters section of the PRMS Parameter
File to indicate that air-temperature values are specified in degrees Celsius. At least one column of tmax and tmin
values must be specified, which also means that dimension ntemp in the dimensions section of the PRMS Parameter
File must be set to at least 1.
precipis the measured daily precipitation and can be specified in units of inches or millimeters. Set the parameter
precip_units to a value of 1 in the parameters section of the PRMS Parameter File to indicate that precipitation
values are in millimeters. At least one column of precip values must be specified, which also means that dimension
nrain in the dimensions section of the PRMS Parameter File must be set to at least 1.
form_dataand the dimension nform in the dimensions section of the PRMS Parameter File are used to indicate whether
or not the form of the precipitation will be specified as an input column. Typically, these data are not available. If
form_ data and nform are set to 0, specified temperature data will be used to determine the form of the precipitation
(rain, if temperature is greater than or equal to parameter tmax_allrain, or snow, if temperature is less than or
equal to parameter tmax_allsnow, with both specified in the parameters section of the PRMS Parameter File). If
form_ data and nform are set to 1, the form of the precipitation must be specified in the appropriate column on
each line of the time-series data item. The form of the precipitation can be specified as unknown (temperature data
will be used to determine the form), snow, or rain, using the values 0, 1, and 2, respectively. The specified form of the
precipitation is applied to precipitation values for all HRUs. Typically, form of precipitation data are not known or is not
applicable for an entire watershed, thus form_data and nform are set to 0 for a GSFLOW simulation.
PRMS Parameter File

The PRMS Parameter File specifies dimensions and parameters required for PRMS and GSFLOW modules formatted
according to the Modular Modeling System Parameter File format. The first part of this section describes the overall format
of the PRMS Parameter File. The second part provides 22 subsections that describe the parameters that are specified for each
PRMS and GSFLOW module included in GSFLOW. These subsections describe each parameter that is required for each
PRMS and GSFLOW module. The modules that are included in GSFLOW, described in the order in which they are used
computationally, are listed in table 1 in the body of this report.
A table is provided for each module that includes the following information to define each parameter: (1) name;
(2)description; (3) dimension name; (4) measurement units (acres, degrees Fahrenheit, and so forth); (5) data type (integer, real,
or character); (6) the range of values that can be specified; and (7) the default value used by GSFLOW if the parameter is not
defined in the PRMS Parameter File. No parameters have the double-precision data type (TYPE set to 3) in GSFLOW.
For some modules, an eighth column is included that provides the equation number and variable name for parameters that
are used directly in equations documented in section, Computations of Flow in the body of this report. Additional information
for selected parameters is given in the text that accompanies each table.
Appendix 1 143
Parameter File Format

The PRMS Parameter File consists of three sections: header, dimensions, and parameters. Figure A1-2 shows a portion of
the Parameter File used in the Sagehen Creek watershed example problem that includes the header section, a partial dimensions
section (8 of 24 dimensions), and two parameter-declaration items (out of 132). The number of parameters specified in a
Parameter File equals the number of unique parameters declared for all PRMS and GSFLOW modules included in a simulation.
Each module can have different numbers of parameters. Some parameters are declared in multiple modules, but need only
to be specified once in the PRMS Parameter File. The user can select different PRMS modules to distribute air temperature,
precipitation, and potential solar radiation, and to compute potential evapotranspiration and surface runoff using control
parameters precip_module, temp_module, solrad_module, et_module, and srunoff_module, respectively.
The header section consists of two items:
TITLE
Version: 1.7
The first item is a single line that specifies a description (or title) of the PRMS Parameter File; the TITLE line can contain up to
256 characters. The second item is a single line that specifies the Modular Modeling System Parameter File format version, this
must begin in column 1 and be specified as Version: 1.7.
The dimensions section is used to define the size of dimensions that are used to allocate memory for parameters and
variables required by the GSFLOW and PRMS modules of a particular GSFLOW simulation. The dimensions section begins in
line 3 of the PRMS Parameter File with the following identifier that begins in column 1:
** Dimensions **
The identifier is followed by a series of 3-line dimension declarations items that have the following format:
####
NAME
SIZE
The first line is used as a delimiter for each of the dimension declarations, specified as a string of four pound signs (####)
that begins in column 1. The second line (NAME) is the name of the dimension, specified as a character string without spaces
using lowercase letters. The third line (SIZE) is the dimension size, specified as an integer value. Table A1-4 lists the names,
definitions, and default values of the 24 dimensions that can be specified in the dimensions section. If one or more of the
dimensions listed in table A1-4 are not specified, the default value is used. Although dimension names are listed alphabetically
within functional groups in table A1-4, dimension declarations items may be specified in any order in the dimensions section.
An example dimension declaration item for dimension nhru, which specifies the number of hydrologic response units and has a
size of 128 for this particular example, follows:
####
nhru
128
Sagehen Parameter File with 128 HRUs, 201 reaches, 15 segments, cascading flow
Version: 1.7
** Dimensions **
####
one
1
####
ncascade
317
####
nsegment
15
####
nreach
201
####
nmonths
12
####
nhru
128
####
ngwcell
5913
####
nhrucell
4691
** Parameters **
####
albset_snm 10
1
one
1
2
0.2000000029802
####
tmin_lapse 10
1
nmonths
12
2
0.0
-3.200000047684
-0.2000000029802
0.4000000059605
0.0
-2.5
-4.900000095367
-6.099999904633
-5.800000190735
-4.400000095367
-2.700000047684
-1.799999952316
Figure A1-2. Example portion of a PRMS Parameter File.
nv07tac-4124_figA1-2
Appendix 1 145
Table A1-4. Dimension-variables specified in the dimensions section of the PRMS Parameter File.
Variable name Definition Default value

Spatial dimensions
ngw Number of PRMS ground-water reservoirs (used in PRMS-only simulations) 1
ngwcell Number of MODFLOW finite-difference cells in a layer (includes active and inactive cells) 0
nhru Number of HRUs 1
nhrucell Number of unique intersections between gravity reservoirs in PRMS soil zone and MODFLOW finite- 0
difference cells
nreach Number of stream reaches on all stream segments 0
nsegment Number of stream segments 0
nsfres Number of on-channel detainment reservoirs (used in PRMS-only simulations) 0
nssr Number of PRMS subsurface reservoirs (must be specified equal to nhru) 1
Time-series input data dimensions

nevap Number of measurement stations that measure pan evaporation 0
nform Number of input columns in PRMS Data File used to specify form of precipitation (0 if no form data, 0
1 if form data)
nobs Number of streamflow-gaging stations 0
nrain Number of measurement stations that measure precipitation 1
nsol Number of measurement stations that measure solar radiation 0
ntemp Number of measurement stations that measure air temperature 1
Computation dimensions
mxnsos Maximum number of table values for computing storage in and flow from detention reservoirs using 0
Puls routing (PRMS-only simulations)
ncascade Number of cascade paths associated with HRUs 0
ncascdgw Number of cascade paths associated with PRMS ground-water reservoirs 0
ndepl Number of snow-depletion curves used for snowmelt calculations 1
ndeplval Number of snow-depletion values for each snow-depletion curve ndepl*11
Fixed dimensions
ndays Maximum number of days in a year 366
nlapse Number of lapse rates in the x, y, and z directions (used by module xyz_dist) 3
nmonths Number of months in a year 12
one A constant 1
The parameters section begins with the following identifier line that begins in column 1:
** Parameters **
The identifier is followed by a series of 7 items to declare each parameter. Each parameter declaration item has the following
structure:
####
NAME WIDTH
NO_DIMENSIONS
DIMENSION_NAMES(S) (Repeat this item for each DIMENSION_NAME)
N_VALUES
TYPE
VALUE(S) (Repeat this item N_VALUES times)
The first line is used as a delimiter for each parameter declarations item, specified as a string of four pound signs (####)
that begins in column 1. The second line specifies two values: (1) the name of the parameter (NAME), specified as a character string
without spaces; and (2) a column width (WIDTH), specified as an integer value that is used by a Modular Modeling System spreadsheet
editor that is not implemented in GSFLOW. The WIDTH value is ignored by GSFLOW, but a value, such as 10, must be specified.
The third line is used to set the number of dimensions (NO_DIMENSIONS) that defines the array used to store the parameter values.
NO_DIMENSIONS is specified as an integer value (1 for scalars and one-dimensional arrays; 2 for two-dimensional arrays). Scalar
parameters (that is, those with a single value) are stored in the Modular Modeling System as a one-dimensional array; therefore, the user
must specify a 1 for scalar variables.
The next NO_DIMENSIONS lines specify the dimension name(s) (DIMENSION_NAME(S)), one value per line, over
which the parameter is declared, as character strings without spaces (see table A1-4 for a list of valid dimension names).
DIMENSION_NAME must be specified as one if the parameter is a scalar. Most of the parameters have NO_DIMENSIONS
equal to 1, and a DIMENSION_NAME that is either one or the name of a one-dimensional array (such as nhru or nmonths).
A few parameters consist of two-dimensional arrays; that is, an array of values that consists of rows and columns. An example
parameter that consists of a two-dimensional array of values is rain_adj, which is used in the PRMS precipitation modules
precip_prms and precip_laps_prms. This parameter is defined over the dimensions nhru by nmonths; that is, the parameter
has nhru rows and nmonths columns.
The line following the line(s) specifying dimension name(s) specifies the number of values (N_VALUES) that are input
for the parameter. N_VALUES is specified as an integer value. For parameters that are scalar or a one-dimensional array,
N_VALUES is specified as the size of the associated dimension, as declared in the dimensions section of the file. The total
number of values specified for each two-dimensional parameter is equal to the product of its two dimension sizes. For the
rain_adj parameter mentioned above, N_VALUES is specified as the product of the sizes of nhru and nmonths.
The line following specification of N_VALUES specifies type (TYPE) of the parameter values as an integer value; options
are:
1 for integer
2 for real (single-precision, floating decimal point)
3 for double (double-precision, floating decimal point)
4 for character string
Note, no double-precision real (option 3) parameters are required in GSFLOW.
The line(s) following specification of TYPE, N_VALUES in number, each specifies one parameter value [VALUE(S)].
Two-dimensional array values are read column by column. For example, for the rain_adj parameter mentioned above, a value
for each HRU (nhru values) for January (nmonths set to 1) are specified first, followed by nhru values for nmonths set to
2, and so forth, until a total of the product of nhru and nmonths values are specified.
Two example parameter-declaration items follow. The first example is for the latitude of the watershed centroid
(VALUE set to 39.42900085449), parameter basin_lat, which is a scalar parameter (NO_DIMENSIONS set to 1;
DIMENSION_ NAME set to one) with a single value (N_VALUE(S) set to 1) of real type (TYPE set to 2):
####
basin_lat 15
1
one
1
2
39.42900985449
The second example is for the area of each HRU, parameter hru_area, which is a one-dimensional array declared
using dimension nhru (NO_DIMENSIONS set to 1; DIMENSION_NAME set to nhru). The parameter consists of 128 values
(N_VALUE(S) set to 128), which equals the size of nhru specified in the dimensions section. The values are of real type
(TYPE set to 2). Only the first five values are shown:
####
hru_area 15
1
nhru
128
2
75.16000366211
38.68999862671
13.11999988556
108.9700012207
26.23999977112
(remaining 123 values not shown)
Appendix 1 147
Parameter-declaration items can be listed in any order, although it may be convenient to group items alphabetically or
by each of the PRMS and GSFLOW modules. However, as some parameters are used by multiple modules, care must be
taken in grouping parameter-declaration items by module. All PRMS parameters and the modules they are used in are listed
alphabetically in table A1-5. If multiple parameter-declaration items are specified for the same parameter, the values specified
last in the Parameter File will be used. Any parameter not specified in the Parameter File that is required by a GSFLOW
simulation is assigned a default value. Any parameter specified in the Parameter File that is not required by a GSFLOW
simulation is ignored. Warning message(s) are printed in both cases.
Table A1-5. Parameters in the PRMS Parameter File listed alphabetically and their associated modules.
Parameter name Module or modules

adjmix_rain precip_dist2_prms, precip_laps_prms, precip_prms, xyz_dist
adjust_rain xyz_dist
adjust_snow xyz_dist
albset_rna snowcomp_prms
albset_rnm snowcomp_prms
albset_sna snowcomp_prms
albset_snm snowcomp_prms
basin_area basin_prms
basin_cfs_init gsflow_budget, strmflow_prms
basin_lat soltab_hru_prms
basin_solsta ccsolrad_hru_prms, ddsolrad_hru_prms
basin_tsta basin_sum_prms, ccsolrad_hru_prms, temp_1sta_prms, temp_dist2_prms, temp_laps_prms
basin_tsta_hru basin_sum_prms, xyz_dist
carea_max srunoff_carea_casc, srunoff_smidx_casc
carea_min srunoff_carea_casc
cascade_flg cascade_prms
cascade_tol cascade_prms
ccov_intcp ccsolrad_hru_prms
ccov_slope ccsolrad_hru_prms
cecn_coef snowcomp_prms
conv_flag xyz_dist
cov_type intcp_prms, snowcomp_prms, soilzone_gsflow
covden_sum hru_sum_prms, intcp_prms, snowcomp_prms, soilzone_gsflow
covden_win hru_sum_prms, intcp_prms, snowcomp_prms, soilzone_gsflow
crad_coef ccsolrad_hru_prms
crad_exp ccsolrad_hru_prms
dday_intcp ddsolrad_hru_prms
dday_slope ddsolrad_hru_prms
den_init snowcomp_prms
den_max snowcomp_prms
elev_units basin_prms, xyz_dist
emis_noppt snowcomp_prms
epan_coef intcp_prms, potet_pan_prms
fastcoef_lin soilzone_gsflow
fastcoef_sq soilzone_gsflow
freeh2o_cap snowcomp_prms
gvr_cell_id gsflow_budget, gsflow_mf2prms, gsflow_prms2mf, gsflow_setconv
gvr_cell_pct gsflow_prms2mf, gwflow_setconv
gvr_hru_id gsflow_budget, gsflow_mf2prms, gsflow_prms2mf, soilzone_gsflow
gvr_hru_pct gsflow_budget, gsflow_mf2prms, gsflow_prms2mf, soilzone_gsflow
gw_down_id cascade_prms
gw_pct_up cascade_prms
gw_strmseg_down_id cascade_prms
gw_up_id cascade_prms
gwflow_coef gwflow_casc_prms
Table A1-5. Parameters in the PRMS Parameter File listed alphabetically and their associated modules.Continued

gwsink_coef gwflow_casc_prms
gwstor_init gwflow_casc_prms
hamon_coef potet_hamon_hru_prms
hru_area basin_prms, cascade_prms, ccsolrad_hru_prms, ddsolrad_hru_prms, gsflow_budget, gsflow_mf2prms,
gsflow_prms2mf, gwflow_casc_prms, intcp_prms, potet_hamon_hru_prms, potet_jh_prms, potet_pan_prms,
precip_dist2_prms, precip_laps_prms, precip_prms, snowcomp_prms, soilzone_gsflow, srunoff_carea_casc,
srunoff_smidx_casc, strmflow_prms, temp_1sta_prms, temp_dist2_prms, temp_laps_prms, xyz_dist
hru_aspect soltab_hru_prms
hru_deplcrv snowcomp_prms
hru_down_id cascade_prms
hru_elev basin_prms, precip_laps_prms, temp_1sta_prms, temp_dist2_prms, temp_laps_prms, xyz_dist
hru_gwres cascade_prms, gwflow_casc_prms, strmflow_prms
hru_lat soltab_hru_prms
hru_pansta potet_pan_prms
hru_pct_up cascade_prms
hru_percent_imperv basin_prms, hru_sum_prms, srunoff_carea_casc, srunoff_smidx_casc
hru_plaps precip_laps_prms
hru_psta precip_laps_prms, precip_prms
hru_segment gsflow_prms2mf
hru_sfres strmflow_prms
hru_slope basin_prms, soltab_hru_prms
hru_strmseg_down_id cascade_prms
hru_tlaps temp_laps_prms
hru_tsta temp_1sta_prms, temp_laps_prms
hru_type basin_prms, cascade_prms, gsflow_budget, gsflow_mf2prms, gsflow_prms2mf, hru_sum_prms, intcp_prms,
snowcomp_prms, soilzone_gsflow, srunoff_carea_casc, srunoff_smidx_casc
hru_up_id cascade_prms
hru_x xyz_dist
hru_xlong precip_dist2_prms, temp_dist2_prms
hru_y xyz_dist
hru_ylat precip_dist2_prms, temp_dist2_prms
id_obsrunoff gsflow_sum
imperv_stor_max srunoff_carea_casc, srunoff_smidx_casc
jh_coef potet_jh_prms
jh_coef_hru potet_jh_prms
lake_hru_id gsflow_budget, gsflow_mf2prms, gsflow_prms2mf
lapsemax_max temp_dist2_prms
lapsemax_min temp_dist2_prms
lapsemin_max temp_dist2_prms
lapsemin_min temp_dist2_prms
local_reachid gsflow_prms2mf
max_lapse xyz_dist
maxmon_prec precip_dist2_prms
melt_force snowcomp_prms
melt_look snowcomp_prms
min_lapse xyz_dist
monmax temp_dist2_prms
monmin temp_dist2_prms
moyrsum hru_sum_prms
mxsziter gsflow_modflow, gsflow_prms2mf
nsos strmflow_prms
numreach_segment gsflow_prms2mf
o2 strmflow_prms
objfunc_q basin_sum_prms
padj_rn precip_laps_prms
padj_sn precip_laps_prms
Appendix 1 149

pmn_mo precip_laps_prms
pmo hru_sum_prms
potet_sublim intcp_prms, snowcomp_prms
ppt_add xyz_dist
ppt_div xyz_dist
ppt_lapse xyz_dist
ppt_rad_adj ccsolrad_hru_prms, ddsolrad_hru_prms
precip_units precip_dist2_prms, precip_laps_prms, precip_prms, xyz_dist
pref_flow_den soilzone_gsflow
print_freq basin_sum_prms
print_objfunc basin_sum_prms
print_type basin_sum_prms
psta_elev precip_laps_prms, xyz_dist
psta_freq_nuse xyz_dist
psta_mon precip_dist2_prms
psta_month_ppt xyz_dist
psta_nuse xyz_dist
psta_x xyz_dist
psta_xlong precip_dist2_prms
psta_y xyz_dist
psta_ylat precip_dist2_prms
rad_conv ccsolrad_hru_prms, ddsolrad_hru_prms
rad_trncf snowcomp_prms
radadj_intcp ddsolrad_hru_prms
radadj_slope ddsolrad_hru_prms
radj_sppt ccsolrad_hru_prms, ddsolrad_hru_prms
radj_wppt ccsolrad_hru_prms, ddsolrad_hru_prms
radmax ccsolrad_hru_prms, ddsolrad_hru_prms
rain_adj precip_prms
rain_code obs_prms, xyz_dist
rain_mon precip_dist2_prms
reach_segment gsflow_prms2mf
runoff_units basin_sum_prms, gsflow_sum
s2 strmflow_prms
sat_threshold soilzone_gsflow
segment_pct_area gsflow_prms2mf
settle_const snowcomp_prms
sfres_coef strmflow_prms
sfres_din1 strmflow_prms
sfres_init strmflow_prms
sfres_qro strmflow_prms
sfres_type strmflow_prms
slowcoef_lin soilzone_gsflow
slowcoef_sq soilzone_gsflow
smidx_coef srunoff_smidx_casc
smidx_exp srunoff_smidx_casc
snarea_curve snowcomp_prms
snarea_thresh snowcomp_prms
snow_adj precip_prms
snow_intcp intcp_prms
snow_mon precip_dist2_prms
snowinfil_max srunoff_carea_casc, srunoff_smidx_casc
soil_moist_init soilzone_gsflow
soil_moist_max soilzone_gsflow, srunoff_carea_casc, srunoff_smidx_casc
soil_rechr_init soilzone_gsflow
soil_rechr_max soilzone_gsflow, srunoff_carea_casc
soil_type soilzone_gsflow

soil2gw_max soilzone_gsflow
solrad_elev xyz_dist
srain_intcp intcp_prms
ssr2gw_exp soilzone_gsflow
ssr2gw_rate soilzone_gsflow
ssrmax_coef soilzone_gsflow
ssstor_init soilzone_gsflow
szconverge gsflow_prms2mf
temp_units potet_hamon_hru_prms, potet_jh_prms, precip_dist2_prms, precip_laps_prms, precip_prms, temp_1sta_prms,
temp_dist2_prms, temp_laps_prms, xyz_dist
tmax_add xyz_dist
tmax_adj temp_1sta_prms, temp_dist2_prms, temp_laps_prms, xyz_dist
tmax_allrain ddsolrad_hru_prms, precip_dist2_prms, precip_laps_prms, precip_prms, xyz_dist
tmax_allsnow precip_dist2_prms, precip_laps_prms, precip_prms, snowcomp_prms, xyz_dist
tmax_div xyz_dist
tmax_index ddsolrad_hru_prms
tmax_lapse temp_1sta_prms
tmin_add xyz_dist
tmin_adj temp_1sta_prms, temp_dist2_prms, temp_laps_prms, xyz_dist
tmin_div xyz_dist
tmin_lapse temp_1sta_prms
transp_beg potet_hamon_hru_prms, potet_jh_prms, potet_pan_prms
transp_end potet_hamon_hru_prms, potet_jh_prms, potet_pan_prms
transp_tmax potet_hamon_hru_prms, potet_jh_prms
tsta_elev temp_1sta_prms, temp_dist2_prms, temp_laps_prms, xyz_dist
tsta_month_max xyz_dist
tsta_month_min xyz_dist
tsta_nuse xyz_dist
tsta_x xyz_dist
tsta_xlong temp_dist2_prms
tsta_y xyz_dist
tsta_ylat temp_dist2_prms
tstorm_mo snowcomp_prms
upst_res1 strmflow_prms
wrain_intcp intcp_prms
x_add xyz_dist
x_div xyz_dist
y_add xyz_dist
y_div xyz_dist
z_add xyz_dist
z_div xyz_dist
PRMS Modules
Input instructions for those modules that relate to a single process are grouped together, such as modules temp_1sta_prms,
temp_laps_prms, and temp_dist2_prms, which relate to the distribution of temperature across the modeled area. The module
xyz_dist is listed separately because it is used to distribute both temperature and precipitation across the modeled area. Several
parameters are used by more than one module. Because the modules are defined independently of one another, the parameter
description is repeated for each module. For example, the parameter hru_area is used by several modules, including modules
basin_prms and temp_1sta_prms.
Appendix 1 151
Basin Module
The PRMS Basin Module (basin_prms) computes shared watershed-wide variables. Shared variables include the area of
each HRU that is pervious and impervious determined on the basis of the fraction in each HRU that is impervious (parameter
hru_precent_imperv), the total area of the watershed determined as the sum of the area in each HRU (parameter
hru_ area), and the total area occupied by lake HRUs and land HRUs determined as the sum of the area in each lake and
land HRU (parameter hru_type), respectively. Input parameters for the Basin Module are defined in table A1-6. Checks for
consistency of watershed-wide variables are done with the input parameters.
Table A1-6. Input parameters required for the PRMS Basin Module: basin_prms.
[HRU, hydrologic response unit; nhru, number of HRUs]
Dimension Default
Parameter name Description Units Type Range
variable value
basin_area Total area of watershed one acres real 0.1 to 1.0e9 1.0
elev_units Units of altitude (0=feet; 1=meters) one dimensionless integer 0 or 1 0
hru_area Area of HRU nhru acres real 0.1 to 1.0e9 1.0
hru_elev Mean land-surface altitude of HRU nhru elev_units real -300.0 to 30,000.0 0.0
hru_percent_imperv Decimal fraction of HRU area that is nhru dimensionless real 0.0 to 1.0 0.0
impervious
hru_slope Slope of HRU, specified as change in nhru dimensionless real 0.0 to 10.0 0.0
vertical length divided by change in
horizontal length
hru_type Type of HRU (0=inactive; 1=land; 2=lake) nhru dimensionless integer 0 to 2 1
Cascade Module
The PRMS Cascade Module (cascade_prms) determines the computational order of the HRUs and ground-water reservoirs
for routing flow downslope in a cascading pattern. Input parameters for Cascade Module are defined in table A1-7; parameters
are arranged by type of routing (HRU or ground-water reservoir) with those that are common to both routing patterns listed first.
Each link in an HRU cascade path is specified by four parameters (hru_up_id, hru_down_id, hru_strmseg_ down_ id,
and hru_pct_up). The number of HRU cascade links must equal dimension ncascade. Likewise, each ground-water
reservoir cascade link is specified by four parameters (gw_up_id, gw_down_id, gw_strmseg_down_id, and
gw_ pct_ up). The number of ground-water reservoir cascade links must equal dimension ncascdgw.
Additional information for selected parameters defined in table A1-7 follows:
cascade_flgrouting switch. A value of 0 allows for multiple routing paths of inflows and outflows among HRUs (many-
to-many routing) and to stream segments. A value of 1 forces routing to be from an upslope HRU to one downslope
HRU (one-to-one routing) and(or) stream segment, even if multiple cascade links are specified from an upslope HRU.
The cascade link with the greatest fraction of contributing area (parameter hru_pct_up) in the upslope HRU is
selected as the single outflow from the HRU.
cascade_tolarea tolerance for cascade links. This parameter can be used to ignore cascade links involving small
contributing areas in the upslope HRU or ground-water reservoir that may be generated in a GIS analysis to produce
the cascade parameters. Contributing areas less than cascade_tol are evenly distributed to all other cascade paths
originating in the upslope HRU. Contributing areas for an HRU cascade link equals the area of the HRU (parameter
hru_area) times the fraction of outflow for the path (parameter hru_pct_up). Contributing areas for a ground-water
reservoir cascade link equals the area of the ground-water reservoir (variable gwres_area computed in the Cascade
Module) times the fraction of outflow for the link (parameter gw_pct_up).
hru_pct_up and gw_pct_upfraction of total outflow from an upslope HRU and ground-water reservoir, respectively, to
be added as inflow to a downslope HRU (or ground-water reservoir) or stream segment. This fraction represents the ratio
of the contributing area to the total area of the upslope HRU (or ground-water reservoir) to be routed to a downslope
HRU (or ground-water reservoir).
hru_strmseg_down_id and gw_strmseg_down_ididentification number of a stream segment that receives outflow
from an HRU and ground-water reservoir, respectively. When either parameter is greater than zero, the specified
stream segment receives the outflow for the cascade path and the corresponding parameter (hru_down_id and
gw_ down_id) is ignored.
Table A1-7. Input parameters required for the PRMS Cascade Module: cascade_prms.
[Equation number refers to equations listed in the main body of report equation variable of parameter name is defined in first listed equation. HRU: hydrologic
response unit; nhru, number of HRUs; ncascade, number of cascade paths associated with HRUs; ncascdgw, number of cascade paths associated with
PRMS ground-water reservoirs; ngw, number of PRMS ground-water reservoirs ; one, a dimension of one]
Dimension Default Equation

variable value number
hru_area nhru
Area of HRU acres real 0.1 to 1.0e9 1.0 22, 37
casacade_flg1 Type of cascade routing (0=allow one dimensionless integer 0 or 1 0
many-to-many; 1=only allow
one-to-one)
cascade_tol1 Minimum area of upslope HRU for one acres real 0.0 to 99.0 5.0
computing cascading flow
Parameters for routing surface runoff and interflow among HRUs and to stream segments
hru_up_id Identifier of HRU that contributes ncascade dimensionless integer 1 to nhru 1
flow for each cascade link
hru_strmseg_down_id1 Identifier of stream segment that ncascade dimensionless integer 0 to 0
receives flow for each cascade nsegment + 1
link
hru_down_id Identifier of HRU that receives ncascade dimensionless integer 1 to nhru 1
flow for each cascade link; if
hru_strmseg_down_id is not 0
for a cascade link, hru_down_id
is ignored
hru_type Type of each HRU (0=inactive; nhru dimensionless integer 0 to 2 1
1=land; 2=lake)
hru_pct_up1 Decimal fraction of area in the ncascade dimensionless real 0.0 to 1.0 1.0 36
upslope HRU that contributes
Hortonian runoff to the downslope
HRU
Parameters for routing flow among ground-water reservoirs and to streams, used for PRMS-only simulations
1
gw_pct_up Decimal fraction of source ground- ncascdgw dimensionless real 0.0 to 1.0 1.0
water reservoir associated with
each ground-water cascade link
hru_gwres Identifier of ground-water reservoir nhru dimensionless integer 1 to ngw 1
associated with an HRU
gw_strmseg_down_id1 Identifier of stream segment that ncascdgw dimensionless integer 0 to 0
receives flow for each ground- nsegment + 1
water cascade link
gw_up_id Identifier of HRU that contributes ncascdgw dimensionless integer 1 to ngw 1
flow to each ground-water
cascade link
gw_down_id Identifier of ground-water reservoir ncascdgw dimensionless integer 1 to ngw 1
that receives flow from each
ground-water cascade link;
gw_down_id is ignored if gw_
strmseg_down_id is not 0 for a
cascade link
1
Additional description of parameter provided on p. 151.
Appendix 1 153
Observed Data Module

The PRMS Observed Data Module (obs_prms) makes available the measured data for each time step specified in the
PRMS Data File(s), and verifies that the data are within a valid range. If measured values are outside the valid range for an
input variable, a warning message is printed. Data that are made available are precipitation and maximum and minimum air
temperature. Optional measured data if specified in the PRMS Data File include solar radiation and streamflow. A single
parameter is used by the Observed Data Module and it is only needed when the xyz_dist module is active. The input parameter
for the module is defined in table A1-8.
Table A1-8. Input parameter required for the PRMS Observed Data Module: obs_prms.
[nmonths, number of months in a year]
Dimension Default
variable value
rain_code Use of measured precipitation values (1=if psta_nuse nmonths dimensionless integer 1 to 5 2
stations have precipitation; 2=if any precipitation
station has precipitation; 3=if xyz regression indicates
precipitation; 4=if rain_day variable is set to 1 in a
PRMS Data File; 5=if psta_freq_use stations have
precipitation)
Potential Solar-Radiation Module

The PRMS Potential Solar-Radiation Module (soltab_hru_prms) calculates tables of 366 (maximum number of days
in a year) values of potential solar radiation and hours of sunlight for each HRU on the basis of representative slope, aspect,
and latitude of each HRU. The module also computes a table of the potential solar radiation at the watershed centroid with a
horizontal slope. Input parameters for the module are defined in table A1-9.
Table A1-9. Input parameters required for the PRMS potential solar-radiation module: soltab_hru_prms.
[Equation number refers to equations listed in the main body of reportparameter name is defined following the equation. HRU, hydrologic response unit;
nhru, number of HRUs]
Parameter Dimension Default Equation

Description Units Type Range
name variable value number
basin_lat Latitude of watershed centroid one degrees latitude real -90.0 to 90.0 40.0 13a
hru_aspect Aspect of HRU nhru degrees azimuth real 0.0 to 360.0 0
hru_lat Latitude of HRU centroid nhru degrees latitude real -90.0 to 90.0 40.0 13a
hru_slope HRU slope, specified as change in vertical nhru dimensionless real 0.0 to 10.0 0.0 15b
length divided by change in horizontal
length
Temperature Distribution Modules

Four PRMS Temperature Distribution Modules (temp_1sta_prms, temp_laps_prms, temp_dist2_prms, and xyz_dist) are
included in GSFLOW. The modules are used to distribute maximum and minimum temperatures to each HRU. The user selects
one of the four modules for a particular simulation by setting the control parameter temp_module in the GSFLOW Control
File to the one of the temperature module names. If module xyz_dist is selected for distributing temperature then it also must
be selected for distributing precipitation. Modules temp_1sta_prms, temp_laps_prms, and temp_dist2_prms are explained in
this section and module xyz_dist is described in the subsequent section Temperature and Precipitation Distribution Module
because when specified it also must distribute precipitation to each HRU.
Module temp_1sta_prms distributes temperatures to HRUs using temperature data measured at one station and a monthly
parameter based on a lapse rate with elevation. Module temp_laps_prms distributes temperatures to HRUs using a lapse rate
computed from daily measured data at two temperature stations. Module temp_dist2_prms distributes temperatures to HRUs
using a lapse rate computed from from two or more stations weighted by the inverse of the square of the distance between
the centroid of an HRU and each station location. Input parameters for modules temp_1sta_prms, temp_laps_prms, and
temp_dist2_prms are defined in table A1-10. Some parameters are common to the three modules and are listed first, followed by
parameters specific for module temp_1sta_prms, then for temp_laps_prms, and last for temp_dist2_prms.
Table A1-10. Input parameters required for the PRMS Temperature Distribution Modules: temp_1sta_prms, temp_laps_prms, and
temp_dist2_prms.
nhru, number of HRUs; ntemp, number of measurement stations that measure air temperature; nmonths, number of months in a year; elev_units, PRMS
Basin Module parameter to define units of feet (0) or meters (1)]

Parameters common to three modules
basin_tsta Identifier of the measurement station used one dimensionless integer 1 to ntemp 1
to compute basin air temperature
hru_elev Mean land-surface altitude of HRU nhru elev_units real -300.0 to 0.0 1, 2, 5a
30,000.0
temp_units Units of measured air temperature one dimensionless integer 0 or 1 0.0
(0=degrees Fahrenheit; 1=degrees
Celsius)
tmax_adj1 Maximum daily HRU temperature nhru temp_units real -10.0 to 10.0 0.0 1, 2, 5a
adjustment factor, which is estimated on
the basis of slope and aspect
tmin_adj1 Minimum daily HRU temperature nhru temp_units real -10.0 to 10.0 0.0 1, 2, 5a
adjustment factor, which is estimated on
the basis of slope and aspect
tsta_elev Altitude of the air temperature measurement ntemp elev_units real -300.0 to 0.0 1, 4, 5a
station 30,000.0
Additional parameters for module temp_1sta_prms
hru_tsta Identifier of the measurement station used nhru dimensionless integer 0 to ntemp 1
to compute HRU daily maximum and
minimum air temperatures
tmax_lapse1 Monthly maximum air temperature lapse nmonths temp_units real -10.0 to 10.0 3.0 1
rate, representing the change in maximum
air temperature per 1,000 feet or meters
of altitude change for each month,
January to December
tmin_lapse1 Monthly minimum air temperature lapse nmonths temp_units real -10.0 to 10.0 3.0 1
rate, representing change in minimum air
temperature per 1,000 feet or meters of
altitude change for each month, January
to December
Appendix 1 155
Table A1-10. Input parameters required for the PRMS Temperature Distribution Modules: temp_1sta_prms, temp_laps_prms, and
temp_dist2_prmsContinued.
nhru, number of HRUs; ntemp, number of measurement stations that measure air temperature; nmonths, number of months in a year; elev_units,
PRMS Basin Module parameter to define units of feet (0) or meters (1)]

Additional parameters for module temp_laps_prms
hru_tlaps Identifier of lapse measurement station used nhru dimensionless integer 1 to ntemp 1
for air-temperature lapse rate calculations
hru_tsta Identifier of base measurement station used nhru dimensionless integer 1 to ntemp 1
for air temperature lapse rate calculations
Additional parameters for module temp_ dist2_prms
hru_xlong Longitude of HRU centroid nhru feet real -1e+09 to 0.0 5b
1e+09
hru_ylat Latitude of HRU centroid nhru feet real -1e+09 to 0.0 5b
1e+09
lapsemax_max Monthly maximum lapse rate from nmonths temp_units real -2.0 to 4.0 3.0
historical data used to constrain highest
daily maximum lapse rate
lapsemax_min Monthly minimum lapse rate from historical nmonths temp_units real -7.0 to -3.0 -6.5
data used to constrain lowest daily
maximum lapse rate
lapsemin_max Monthly maximum lapse rate from nmonths temp_units real -2.0 to 4.0 3.0
historical data to constrain highest daily
minimum lapse rate
lapsemin_min Monthly minimum lapse rate from historical nmonths temp_units real -7.0 to -3.0 -4.0
data used to constrain lowest daily
minimum lapse rate
monmax Monthly maximum air temperature from nmonths temp_units real 45.0 to 115.0 100.0
historical data used to constrain lowest
daily maximum air temperatures
monmin Monthly minimum air temperature from nmonths temp_units real -35.0 to 45.0 -20.0
historical data used to constrain lowest
daily minimum air temperatures
tsta_xlong Longitude of measurement station that ntemp feet real -1e+09 to 0.0 5b
measures air temperature 1e+09
tsta_ylat Latitude of measurement station that ntemp feet real -1e+09 to 0.0 5b
measures air temperature 1e+09
1
Additional description of parameter provided below.
Additional information for four parameters shown in table A1-10 follows:

tmax_adj and tmin_adjmaximum and minimum air-temperature adjustments, respectively, for each HRU based on slope
and aspect of the HRU, in degrees Fahrenheit or degrees Celsius.
tmax_lapse and tmin_lapsemonthly maximum and minimum air temperature lapse rates, respectively, in degrees
Fahrenheit or degrees Celsius. Twelve values each are specified that define the change in maximum and minimum air
temperature per 1,000 feet or meters elevation change during each month from January to December.
Temperature and Precipitation Distribution Module

The Temperature and Precipitation Distribution Module (xyz_dist) distributes daily precipitation and minimum and
maximum air temperatures to each HRU using a multiple linear regression of measured data from a group of measurement
stations or from the results of an atmospheric model. The user selects the xyz_dist module for a particular simulation by setting
the control parameters temp_module and precip_module in the GSFLOW Control File to xyz_dist. Important
geographic factors affecting the spatial distribution of precipitation and air temperature distributions within a watershed are
longitude (X), latitude (Y), and elevation (Z). The method used by module xyz_dist is described in Hay and others (2000). Input
parameters for the module xyz_dist are defined in table A1-11.
Table A1-11. Input parameters required for the PRMS Temperature and Precipitation Module: xyz_dist.
[Equation number refers to equations listed in the main body of report equation variable of parameter name is defined in first listed equation. HRU, hydrologic
response unit; nhru, number of HRUs; ntemp, number of measurement stations that measure air temperature; nrain, number of measurement stations that
measure precipitation; nlapse, number of lapse rates in the X, Y, and Z directions; nmonths, number of months in a year]

adjmix_rain1 Monthly adjustment factor for a mixed nmonths dimensionless real 0.0 to 3.0 1.0 6
precipitation event as a decimal
fraction
adjust_rain1 Monthly factor as a decimal fraction nmonths dimensionless real 0.0 to 1.0 0.01 9a
used to adjust rain values
adjust_snow1 Monthly factor as a decimal fraction nmonths dimensionless real 0.0 to 1.0 0.01 9a
used to adjust snow values
basin_tsta_hru Identifier of HRU used to compute one dimensionless integer 0 to nhru 1
watershed air temperatures
conv_flag Conversion of altitude (0=no one dimensionless integer 0 to 2 0
conversion; 1=feet to meters;
2=meters to feet)
elev_units Units of altitude (0=feet; 1=meters) one dimensionless integer 0 or 1 0
precip_units Specify units of precipitation (0=inches; one dimensionless integer 0 or 1 0
1=millimeters)
rain_code Indicates use of XYZ distribution nmonths dimensionless integer 1 to 5 2
technique for each time step (1=if
psta_nuse stations have precipitation;
2=if any precipitation station has
precipitation; 3=always; 4=if
rain_day variable is set to 1 in a
PRMS Data File; 5=if psta_freq_use
stations have precipitation)
solrad_elev Altitiude of each measurement station one meters real 1000.0 to 0.0
that measures solar radiation and 10,000.0
used in calculating degree-day curves
temp_units Units of air temperature (0=degrees one dimensionless integer 0 or 1 0
Fahrenheit; 1=degrees Celsius)
tmax_adj Adjustment to maximum air nhru temp_units real -10.0 to 10.0 0.0 1,3a
temperature for HRU, estimated on
basis of slope and aspect
tmax_allrain Monthly minimum air temperature at an nmonths temp_units real 0.0 to 90.0 40.0
HRU that results in all precipitation
during a day being rain
tmax_allsnow Monthly maximum air temperature at one temp_units real -10.0 to 40.0 32.0 6
which precipitation is all snow for
the HRU
tmin_adj Minimum daily temperature adjustment nhru temp_units real -100.0 to 100.0 0.0 1,3a
factor
Appendix 1 157
Table A1-11. Input parameters required for the PRMS Temperature and Precipitation Module: xyz_distContinued

Location and altitude parameters (x, y, and z)
1
hru_elev Mean land-surface altitude of HRU nhru elev_units real -300.0 to 0.0
30,000.0
hru_x1 Longitude (X) for HRU in albers nhru meters real -1.e-7 to 1.e7 0.0
projection
hru_y1 Latitude (Y) for HRU in albers nhru meters real -1.e-7 to 1.e7 0.0
projection
psta_elev1 Altitude of each measurement station nrain elev_units real -300.0 to 0.0
that measures precipitation 30,000.0
psta_y1 Latitude (Y) for each measurement nrain meters real -1.e-7 to 1.e7 0.0
station that measures precipitation in
albers projection
tsta_elev1 Altitude of each measurement station ntemp elev_units real -300.0 to 0.0
that measures air temperature 30,000.0
tsta_x1 Longitude (X) for each measurement ntemp meters real -1.e-7 to 1.e7 0.0
station that measures air temperature
in albers projection
tsta_y1 Latitude (Y) for each measurement ntemp meters real -1.e-7 to 1.e7 0.0
station that measures air temperature
in albers projection
Multiple linear regression parameters
max_lapse1 Maximum air temperature regression nlapse by temp_units real -100.0 to 100.0 0.0 3b
coefficient for longitude, latitude, nmonths
and altitude, respectively by month,
starting with January
min_lapse1 Minimum air temperature regression nlapse by temp_units real -100.0 to 100.0 0.0 3b
coefficient for longitude, latitude, nmonths
and altitude, respectively by month
starting with January
ppt_lapse1 Precipitation regression coefficient nlapse by inches real -10.0 to 10.0 0.0 9b, 9c
for longitude, latitude, and altitude, nmonths
respectively by month, starting with
January
Designated station parameters
1
psta_freq_nuse Defines measurement stations used to nrain dimensionless integer 0 or 1 1
determine if precipitation is occurring
in watershed (0=no; 1=yes)
psta_nuse Defines which measurement stations nrain dimensionless integer 0 or 1 1
will be used in the distribution
regression of precipitation (0=no;
1=yes)
tsta_nuse Defines which measurement stations ntemp dimensionless integer 0 or 1 1
will be used in distribution regression
of air temperatures (0=no; 1=yes)
Table A1-11. Input parameters required for the PRMS Temperature and Precipitation Module: xyz_distContinued

Mean monthly climate parameters
1
psta_month_ppt Monthly average precipitation at each nrain by precip_units real 0.0 to 200.0 0.0
measurement station nmonths
tsta_month_max1 Monthly average maximum air ntemp by temp_units real -100.0 to 200.0 0.0
temperature at measurement station nmonths
tsta_month_min1 Monthly average minimum air ntemp by temp_units real -100.0 to 200.0 0.0
temperature at each measurement nmonths
station
Transformation parameters for dependent variables
1
ppt_add Calculated mean of precipitation for one precip_units real -10.0 to 10.0 0.0
watershed
ppt_div1 Calculated standard deviation of one precip_units real -10.0 to 10.0 0.0
precipitation for watershed
tmax_add1 Calculated mean of maximum air one temp_units real -100.0 to 100.0 0.0
temperature for watershed
tmax_div1 Calculated standard deviation of one temp_units real -100.0 to 100.0 0.0
maximum air temperature for
watershed
tmin_add1 Calculated mean of minimum air one temp_units real -100.0 to 100.0 0.0
temperature for watershed
tmin_div1 Calculated standard deviation of one temp_units real -100.0 to 100.0 0.0
minimum air temperature for
watershed
Transformation parameters for independent variables
x_add1 Calculated mean of measurement one meters real -1.e-7 to 1.e7 0.0
station longitude (X) coordinates for
watershed
x_div1 Calculated standard deviation of one meters real -1.e-7 to 1.e7 0.0
measurement station longitude (X)
coordinates for watershed
y_add1 Calculated mean of measurement one meters real -1.e-7 to 1.e7 0.0
station latitude (Y) coordinates for
watershed
y_div1 Calculated standard deviation of one meters real -1.e-7 to 1.e7 0.0
measurement station latitude (Y)
z_add1 Calculated mean of measurement one meters real -1.e-7 to 1.e7 0.0
station altitude (Z) coordinates for
watershed
z_div1 Calculated standard deviation of one meters real -1.e-7 to 1.e7 0.0
measurement station altitude (Z)
Appendix 1 159

adjmix_rainadjustment factor to the fraction of measured daily precipitation in the PRMS Data File that is estimated to be
rain. The factor is used to adjust rain in relation to precipitation before computing daily volumes of rain and snow in each
HRU.
adjust_rain and adjust_snowadjustment factors to precipitation values (rain and snow, respectively) specified in the
PRMS Data File. The adjustments are done before computing daily volumes of rain and snow in each HRU.
hru_elev, hru_x, hru_y, psta_elev, psta_x, psta_y, tsta_elev, tsta_x, tsta_ylocation

and elevation (longitude (x), latitude (y), and altitude (elev)) for the centroid of each HRU, and location and elevation of
precipitation (psta) and air-temperature (tsta) measurement stations. These values are computed for the watershed and
should not be adjusted in any model-calibration process.
max_lapse, min_lapse, and ppt_lapsemaximum (max) and minimum (min) air-temperature and precipitation
(ppt) lapse rates, respectively. These two-dimensional parameters (dimensions: nlapse by nmonths, or 3 by 12) are
used to define a multiple-linear regression (MLR) coefficient for each independent variable (longitude (x), latitude (y), or
altitude (z)) by month.
psta_month_pptMean monthly precipitation at a precipitation station. The values are used in place of the station mean
when data are missing at a precipitation station.
tsta_month_max and tsta_month_min Mean monthly maximum and minimum air temperature at a measurement
station, respectively. The values are used in place of the station mean when data are missing at a temperature station.
ppt_add, tmax_add, and tmin_addmean values of the precipitation (precip) and maximum (tmax) and minimum
(tmin) air-temperature for the watershed. These values are computed for the watershed and should not be adjusted in
any model calibration process.
ppt_div, tmax_div, and tmin_divstandard deviation values of the precipitation (precip) and maximum (tmax)
and minimum (tmin) air-temperature for the watershed. These values are computed for the watershed and should not be
adjusted in any model-calibration process.
x_add, y_add, and z_addmean values of the longitude (x), latitude (y), and altitude (z) for the climate-stations used in
the regression analysis. These values are computed for the watershed and should not be adjusted in any model-calibration
process.
x_div, y_div, and z_divstandard deviation values of the longitude (x), latitude (y), and altitude (z) for the climate-
stations used in the regression analysis. These values are computed for the watershed and should not be adjusted in any
model-calibration process.
Precipitation Distribution Modules

Four PRMS Precipitation Distribution Modules (precip_prms, precip_laps_prms, precip_dist2_prms, and xyz_dist) are
included in GSFLOW. The modules are used to distribute precipitation to each HRU and to determine the form of precipitation
(rain, snow, or a mixture of both). The user selects only one of the modules for a simulation by setting the control parameter
precip_module in the GSFLOW Control File to one of those four precipitation module names. If module xyz_dist is selected
for distributing precipitation then it also must be selected for distributing temperature. Modules precip_prms, precip_laps_prms,
and precip_dist2_prms are explained in this section and module xyz_dist is described in the preceding section Temperature
and Precipitation Distribution Module because it also is used to distribute temperature to each HRU. Module precip_prms
distributes precipitation to HRUs using precipitation data measured at one station and a monthly adjustment factor for the form
of precipitation. Module precip_laps_prms distributes precipitation to HRUs using a lapse rate computed from daily measured
data at two precipitation stations. Module precip_dist2_prms distributes precipitation to HRUs using a lapse rate computed from
from two or more stations weighted by the inverse of the square of the distance between the centroid of an HRU and each station
location. Input parameters for modules precip_prms, precip_laps_prms, and precip_dist2_prms are defined in table A1-12.
Table A1-12. Input parameters required for the PRMS Precipitation Distribution Modules: precip_prms, precip_laps_prms, and
precip_dist2_prms.
measure precipitation; nlapse, number of lapse rates in the X, Y, and Z directions; nmonths, number of months in a year; elev_units, PRMS Basin
Module parameter to define units of feet (0) or meters (1)]

Parameters common to all three modules
adjmix_rain Monthly rain adjustment factor for a nmonths dimensionless real 0.0 to 3.0 1.0 6
mixed precipitation event (usually
1.0)
precip_units Units of precipitation (0=inches; one dimensionless integer 0 or 1 0
1=millimeters)
temp_units Units of air temperature (0=degrees one dimensionless integer 0 or 1 0
Fahrenheit; 1=degrees Celsius)
tmax_allsnow Monthly maximum air temperature at one temp_units real -10.0 to 40.0 32.0 6
the HRU
Additional parameters for module precip_prms
hru_psta Identifier of measurement station used nhru dimensionless integer 1 to nrain 1
as base in calculating precipitation
lapse rate
rain_adj1 Monthly factor as a decimal fraction nhru by dimensionless real 0.2 to 5.0 1.0 7
used to adjust rain at the HRU nmonths
snow_adj1 Monthly factor as a decimal fraction nhru by dimensionless real 0.2 to 5.0 1.0 7
used to adjust snow at the HRU nmonths
Additional parameters for module precip_laps_prms

hru_elev Mean land-surface altitude of HRU nhru elev_units real -300.0 to 0.0 1, 8
30,000.0
hru_plaps Identifier of lapse measurement station nhru dimensionless integer 1 to nrain 1
used in calculating precipitation
lapse rate
hru_psta Identifier of measurement station used nhru dimensionless integer 1 to nrain 1
as base in calculating precipitation
lapse rate
padj_rn1 Mean monthly factor to adjust rain nrain by inches per day real -2.0 to 10.0 1.0 8
lapse rate computed between nmonths
station_psta and
station_ plaps when precipitation
is rain (positive factors are used as
multipliers and negative factors are
made positive and substituted for the
computed lapse rate)
padj_sn1 Mean monthly factor to adjust snow nrain by inches per day real -2.0 to 10.0 1.0 8
lapse rate computed between nmonths
station_psta and
station_ plaps (positive factors
are used as multipliers and negative
factors are made positive and
substituted for the computed lapse
rate)
Appendix 1 161
Table A1-12. Input parameters required for the PRMS Precipitation Distribution Modules: precip_prms, precip_laps_prms, and
precip_dist2_prms.Continued
measure precipitation; nlapse, number of lapse rates in the X, Y, and Z directions; nmonths, number of months in a year; elev_units, PRMS Basin
Module parameter to define units of feet (0) or meters (1)]

Additional parameters for module precip_laps_prmsContinued
pmn_mo Mean monthly precipitation at base nrain by inches real 0.0 to 100.0 1.0 8
and lapse stations nmonths
psta_elev Land surface altitude of base and nrain elev_units real -300.0 to 0.0 8
lapse stations 30,000.0
Additional parameters for module precip_dist2_prms
hru_xlong Longitude of HRU centroid nhru feet real -1e+09 to 0.0 10b
1e+09
hru_ylat Latitude of HRU centroid nhru feet real -1e+09 to 0.0 10b
1e+09
maxmon_prec1 Maximum monthly precipitation at nmonths inches real 0.0 to 15.0 5.0
all measurement stations
psta_mon Mean monthly precipitation at each nrain by inches real 0.0 to 50.0 1.0 10c, d
measurement station nmonths
psta_xlong Longitude of each measurement nrain feet real -1e+09 to 0.0 10b
station that measures precipitation 1e+09
psta_ylat Latitude of each measurement station nrain feet real -1e+09 to 0.0 10b
that measures precipitation 1e+09
rain_mon Mean monthly rain on each HRU nhru by inches real 0.0 to 50.0 1.0 10c
that can be obtained from National nmonths
Weather Service's spatial distribution
of mean annual precipitation for the
1971-2000 climate normal period
snow_mon Mean monthly snow on each HRU nhru by inches real 0.0 to 50.0 1.0 10d
that can be obtained from National nmonths
Weather Service's spatial distribution
of mean annual precipitation for the
19712000 climate normal period
1

rain_adj and snow_adjmonthly factors that adjust measured rainfall and snowfall at a precipitation station, respectively,
to an HRU. The factors account for differences in elevation, spatial variation, topography, and gage catch efficiency.
padj_rn and padj_snmonthly rain and snow adjustment factors, respectively, for each precipitation station. These
monthly factors are used to adjust the precipitation lapse rate (hru_plaps) computed between stations. Positive factors
are multiplied by the lapse rate. Negative factors are made positive and substituted for the computed lapse rate.
maxmon_precmaximum monthly precipitation. Precipitation is assumed to be in error if measured precipitation at any

measurement station is greater than the maxmon_prec value for the month.
Solar-Radiation Distribution Modules

Two PRMS Solar-Radiation Distribution Modules (ccsolrad_hru_prms and ddsolrad_hru_prms) are included in GSFLOW.
The modules are used to distribute solar radiation to each HRU. The user selects one module for a particular simulation by
setting the control parameter solrad_module in the GSFLOW Control File to the one of the module names. Both modules
include algorithms for estimating solar radiation when solar-radiation data are not available. The estimation algorithm in
module ccsolrad_hru_prms uses a relation between solar radiation and cloud cover, whereas the estimation algorithm in module
ddsolrad_hru_prms uses a relation between maximum air temperature and degree-day. Input parameters for the two modules are
defined in table A1-13.
Table A1-13. Input parameters required for PRMS Solar-Radiation Modules: ccsolrad_hru_prms and ddsolrad_hru_prms.
response unit; nhru, number of HRUs; ntemp, number of measurement stations that measure air temperature; nsol, number of measurement stations that
measure solar radiation; nmonths, number of months in a year; temp_units, PRMS Temperature Distribution Modules parameter to define units of degrees
Fahrenheit (0) or Celsius (1)]

Parameters common to both modules
basin_solsta Identifier of measurement station used in one dimensionless integer 1 to nsol 1
computing solar radiation
ppt_rad_adj Precipitation threshold used to determine nmonths inches real 0.0 to 0.5 0.02
if solar radiation is adjusted for cloud
cover.
rad_conv1 Factor to convert measured solar radiation one Converts real 0.1 to 100.0 1.0
to langleys measured solar
radiation to
langleys
radj_sppt Precipitation-day adjustment factor to one dimensionless real 0.0 to 1.0 0.44 15b
solar radiation for a summer day with
precipitation greater than ppt_rad_adj
as a decimal fraction
radj_wppt Precipitation-day adjustment factor to one dimensionless real 0.0 to 1.0 0.5 15b
solar radiation for a winter day with
precipitation greater than ppt_rad_adj
as a decimal fraction
radmax Maximum fraction of potential solar one dimensionless real 0.0 to 1.0 0.8
radiation that reaches land surface as a
decimal fraction
Additional parameters for module ccsolrad_hru_prms
basin_tsta Identifier of measurement station used in one dimensionless integer 1 to ntemp 1
computing air temperature
ccov_intcp1 Intercept in the regression equation that nmonths dimensionless real 0.0 to 5.0 1.83 16a
relates cloud cover to daily minimum
and maximum air temperature by
month, starting with January
ccov_slope1 Slope in the regression equation that nmonths dimensionless real -0.5 to -0.01 -0.13 16a
relates cloud cover to daily minimum
and maximum air temperature by
month, starting with January
crad_coef1 Constant used in the cloud-cover to one dimensionless real 0.1 to 0.7 0.4 16b
solar-radiation relation, a value can be
obtained from Thompson (1976, fig. 1)
Appendix 1 163
Table A1-13. Input parameters required for PRMS Solar-Radiation Modules: ccsolrad_hru_prms and ddsolrad_hru_prmsContinued
response unit; nhru, number of HRUs; ntemp, number of measurement stations that measure air temperature; nsol, number of measurement stations that
measure solar radiation; nmonths, number of months in a year; temp_units, PRMS Temperature Distribution Modules parameter to define units of degrees
Fahrenheit (0) or Celsius (1)]

Additional parameters for module ccsolrad_hru_prmsContinued
1
crad_exp Exponent used in the cloud-cover to one dimensionless real 0.2 to 0.8 0.61 16b
solar-radiation relation, a value of
0.61 is suggested by Thompson
(1976)
Additional parameters for module ddsolrad_hru_prms
1
dday_slope Slope of monthly degree-day to nmonths degree-day per real 0.2 to 0.7 0.4
temperature relation temp_units
dday_intcp1 Intercept of monthly degree-day to nmonths degree-day real -60.0 to 4.0 -10.0
temperature relation
radadj_intcp1 Intercept of solar radiation adjustment one degree-day real 0.0 to 1.0 0.0
to temperature
radadj_slope1 Slope of solar radiation adjustment to one degree-day per real 0.0 to 1.0 0.0
temperature temp_units
tmax_index Maximum monthly air temperature nmonths temp_units real -10.0 to 110.0 50.0
used to adjust solar radiation for
precipitation
1

ccov_intcp and ccov_slopeintercept and slope, respectively, in the temperature/cloud-cover relation:
cloud cover=ccov_intcp+ccov_slope(tmaxf-tminf)
See table A1-2 for definition of tmaxf and tminf.
crad_coef and crad_expcoefficient and exponent, respectively, in the cloud cover/solar radiation relation (Thompson,
1976):
solar radiation=crad_coef+(1-crad_coef)(1-cloud cover)crad_exp
dday_intcp and dday_slopeintercept and slope, respectively, in the temperature/degree day relation:
degree-day coefficient=(dday_slope)(tmaxf)+dday_intcp
See table A1-2 for definition of tmaxf.
radadj_intcp and radadj_slopeintercept and slope, respectively, in the temperature range adjustment to solar
radiation:
ppt_adj=(radadj_slope(tmax-tmin))+radadj_intcp
See table A1-2 for definition of tmaxf and tminf.
rad_convconversion factor to langleys for observed radiation. For example, if the units for solar radiation are watt-hours
per square meter, the multiplication factor rad_conv should be set to 0.0860.
Potential Evapotranspiration Modules

Three PRMS Potential Evapotranspiration Modules (potet_hamon_hru_prms, potet_jh_prms, and potet_pan_prms) are
included in GSFLOW. The modules are used to calculate the amount of potential evapotranspiration and to determine if a
time step is one of active transpiration in an HRU. The user selects one of the modules for a simulation by setting the control
parameter et_module in the GSFLOW Control File to one of the module names. Module potet_hamon_hru_prms uses the
Hamon formulation (Hamon, 1961; Murray, 1967; and Federer and Lash, 1978) to calculate potential evapotranspiration.
Module potet_jh_prms uses the Jensen-Haise formulation (Jensen and others, 1969) to calculate potential evapotranspiration.
Module potet_pan_prms uses measured pan evaportation and a monthly coefficient to calculate potential evapotranspiration.
Input parameters for the three modules are defined in table A1-14.
Table A1-14. Input parameters required for the PRMS Potential-Evapotranspiration Modules: potet_hamon_hru_prms, potet_jh_prms,
and potet_pan_prms.
response unit; nhru, number of HRUs; nevap, number of measurement stations that measure air temperature; nmonths, number of months in a year;
temp_units, PRMS Temperature Distribution Modules parameter to define units of degrees Fahrenheit (0) or Celsius (1)].

Parameters common to all three modules
transp_beg Begin month for transpiration nhru month integer 1 to 12 4
computations at HRU
transp_end Last month for transpiration nhru month integer 1 to 12 10
computations at HRU
Additional parameters for module potet_hamon_hru_prms
hamon_coef Monthly air temperature coefficient nmonths inch-cubic meter real 0.004 to 0.008 0.0055 18
used in the Hamon potential per gram
evapotranspiration equation
temp_units Units of measured air temperature one dimensionless integer 0 or 1 0
Celsius)
transp_tmax Maximum temperature used to nhru degree-day real 0.0 to 1000.0 500.0
determine when transpiration begins
in an HRU
Additional parameters for module potet_jh_prms
jh_coef Monthly air temperature coefficient nmonths temp_units real 0.005 to 0.06 0.014 20a
used in Jensen-Haise potential
evapotranspiration equation
jh_coef_hru Air temperature coefficient used nhru temp_units real 5.0 to 20.0 13.0 20a
in Jensen-Haise potential
evapotranspiration equation for each
HRU
temp_units Units of measured air temperature one dimensionless integer 0 or 1 0
Celsius)
transp_tmax Maximum temperature used to nhru degree-day real 0.0 to 1000.0 500.0
determine when transpiration begins
in each HRU
Additional parameters for module potet_pan_prms
hru_pansta Identifier of measurement station that nhru dimensionless integer 1 to nevap 1
measures pan evaporation
epan_coef Monthly pan evaporation coefficient nmonths dimensionless real 0.2 to 3.0 1.0 21
used to convert value to potential
evapotranspiration
Appendix 1 165
PRMS Canopy Interception Module: intcp_prms

The PRMS Canopy Interception Module (intcp_prms) calculates the amount of rain and snow that is intercepted by
vegetation, the amount of evaporation of intercepted rain and snow, and the amount of net rain and snow throughfall that reaches
the soil or snowpack. Input parameters for the module are defined in table A1-15.
Table A1-15. Input parameters required for PRMS Canopy Interception Module: intcp_prms.
response unit; nhru, number of HRUs; nmonths, number of months in a year]

cov_type Plant type on HRU (0=bare soil; nhru dimensionless integer 0 to 3 3
1=grasses; 2=shrubs; 3=trees)
covden_sum Summer plant canopy density as a nhru dimensionless real 0.0 to 1.0 0.5 22, 23
covden_win Winter plant canopy density as a nhru dimensionless real 0.0 to 1.0 0.5 22, 23
epan_coef Monthly evaporation pan coefficient nmonths dimensionless real 0.2 to 3.0 1.0
hru_area Area of HRU nhru acres real 0.1 to 1.0e9 1.0 22
hru_type Type of HRU (0=inactive; 1=land; nhru dimensionless integer 0 to 2 1
2=lake)
potet_sublim Fraction of potential evapotranspiration one dimensionless real 0.1 to 0.75 0.5
sublimated from snow surface as a
decimal fraction
snow_intcp Maximum snow storage in the plant nhru inches real 0.0 to 5.0 0.1 22
canopy for plant type on HRU
srain_intcp Maximum summer rain storage in the nhru inches real 0.0 to 5.0 1 22
plant canopy for plant type on HRU
wrain_intcp Maximum winter rain storage in the nhru inches real 0.0 to 5.0 0.1 22
plant canopy for plant type on HRU
Snow Computation Module

The PRMS Snow-Computation Module (snowcomp_prms) initiates development of a snowpack and simulates snow
accumulation and depletion processes using an energy-budget approach. Input parameters for the module are defined in
tableA1-16.
Table A1-16. Input parameters required for PRMS Snow-Computation Module: snowcomp_prms.
response unit; nhru, number of HRUs; nmonths, number of months in a year; ndepl, number of snow depletion curves; temp_units, PRMS Temperature
Distribution Modules parameter to define units of degrees Fahrenheit (0) or Celsius (1)]

albset_rna Decimal fraction of rain in a mixed rain one dimensionless real 0.0 to 1.0 0.8
and snow event above which snow
albedo is not reset (applied when
snowpack is accumulating)
albset_rnm Decimal fraction of rain in a mixed rain one dimensionless real 0.0 to 1.0 0.6
and snow event above which snow
albedo is not reset (applied when
snowpack is melting)
albset_sna Minimum snow fall, in water one dimensionless real 0.001 to 1.0 0.05
equivalent, needed to reset
snow albedo when snowpack is
accumulating as a decimal fraction
albset_snm Minimum snow fall, in water one dimensionless real 0.001 to 1.0 0.2
equivalent, needed to reset snow
albedo when snowpack is melting as
a decimal fraction
cecn_coef Monthly convection-condensation nmonths calories per degree real 0.0 to 20.0 5.0
energy coefficient Celsius above 0
cov_type Plant cover type for HRU (0=bare soil; nhru dimensionless integer 0 to 3 3
covden_sum Summer plant cover density for plant nhru dimensionless real 0.0 to 1.0 0.5
type on HRU as a decimal fraction
covden_win Winter plant cover density for plant nhru dimensionless real 0.0 to 1.0 0.5
type on HRU as a decimal fraction
den_init Density of new-fallen snow as a one dimensionless real 0.01 to 0.5 0.10 24
decimal fraction
den_max Average maximum snowpack density as one dimensionless real 0.1 to 0.8 0.6 24
a decimal fraction of the liquid water
equivalent
emis_noppt Emissivity of air on days without one dimensionless real 0.757 to 1.0 0.757
precipitation
freeh2o_cap Free-water holding capacity of one dimensionless real 0.01 to 0.2 0.05
snowpack expressed as decimal
fraction of total snowpack water
equivalent
hru_deplcrv Identifier of snowpack areal-depletion nhru dimensionless integer 1 to ndepl 1
curve for HRU
2=lake)
melt_force1 Julian date to force snowmelt one Julian day integer 1 to 366 90
melt_look1 Julian date to start looking for when one Julian day integer 1 to 366 90
snowmelt begins
Appendix 1 167
Table A1-16. Input parameters required for PRMS Snow-Computation Module: snowcomp_prmsContinued.
response unit; nhru, number of HRUs; nmonths, number of months in a year; ndepl, number of snow depletion curves; temp_units, PRMS Temperature
Distribution Modules parameter to define units of degrees Fahrenheit (0) or Celsius (1)]

potet_sublim Decimal fraction of potential one dimensionless real 0.1 to 0.75 0.5 30
evapotranspiration that is sublimated
from snow surface
rad_trncf Transmission coefficient for short-wave nhru dimensionless real 0.0 to 1.0 0.5
radiation through winter plant canopy
on an HRU as a decimal fraction
settle_const Snowpack settlement-time constant one per day real 0.01 to 0.5 0.10 24
snarea_curve Snow area-depletion curve values, 11 11 by dimensionless real 0.0 to 1.0 1.0
for each curve as a decimal fraction ndepl
snarea_thresh Maximum water equivalent threshold, nhru inches real 0.0 to 200.0 50.0
water equivalent in an HRU less than
threshold results in use of snow-
covered-area curve
tmax_allsnow Monthly maximum air temperature at one temp_units real -10.0 to 40.0 32.0
the HRU
tstorm_mo Monthly storm prevalence (0=frontal nmonths dimensionless integer 0 or 1 0
storms prevalent; 1=convective
storms prevalent)
1

melt_forceJulian day to force the use of the isothermal form of the snowpack energy equation. This date should always be
later than melt_look, varies by region, and is dependent on the persistence of the snowpack.
melt_lookJulian day when the snowpack simulation changes from the accumulation phase to the melt phase. This
primarily affects albedo computations. April 1 (Julian day = 90) works well for most western mountain watersheds. For
eastern and some western watersheds, where rain-on-snow or intermittent snowpacks occur during the snow season, the
melt phase may be more appropriate for the whole season. In this case, set melt_look to the start of the snow season.
At the beginning of the water year, the snowpack simulation is reset to the accumulation phase.
Surface Runoff and Infiltration Modules

Two different PRMS Surface Runoff and Infiltration Modules (srunoff_smidx_casc and srunoff_carea_casc) are included
in GSFLOW. The modules are used to compute surface runoff and infiltration for each HRU. The user selects one of the two
modules for a particular simulation by setting the control parameter srunoff_module in the GSFLOW Control File to either
module name. Module srunoff_smidx_casc uses antecedent soil moisture and a non-linear variable-source-area method, whereas
module srunoff_carea_casc uses antecedent soil moisture and a linear variable-source-area method. Input parameters for the two
modules are defined in table A1-17.
Table A1-17. Input parameters required for the two PRMS Surface Runoff and Infiltration Modules: srunoff_carea_casc and
srunoff_smidx_casc.
response unit; nhru, number of HRUs]

carea_max Maximum possible area contributing nhru dimensionless real 0.0 to 1.0 0.6 34
to surface runoff, expressed as a
decimal fraction of HRU area
hru_area Area of HRU nhru acres real 0.1 to 1.0e9 1.0 22, 31,32,
33, 36, 37
impervious
2=lake)
imperv_stor_max Maximum retention storage for HRU nhru inches real 0.0 to 10.0 0.0 31
impervious area
snowinfil_max Daily maximum snowmelt infiltration nhru inches real 0.0 to 20.0 2.0 37
for the HRU
soil_moist_max Maximum available capillary water- nhru inches real 0.0 to 20.0 6.0
holding capacity of soil zone in an
HRU
Additional parameters for module srunoff_carea_casc
carea_min Minimum possible area contributing to nhru dimensionless real 0.0 to 1.0 0.2 34
surface runoff, as a decimal fraction
of HRU area
soil_rechr_max Maximum quantity of water in the nhru inches real 0.0 to 10.0 2.0 34a
capillary reservoir (value must be less
than or equal to soil_moist_max)
Additional parameters for module srunoff_smidx_casc
1
smidx_coef Coefficient in non-linear contributing nhru dimensionless real 0.0001 to 1.0 0.01 34b
area algorithm
smidx_exp1 Exponent in non-linear contributing nhru per inch real 0.2 to 0.8 0.3 34b
area algorithm
1

smidx_coef and smidx_expcoefficient and exponent, respectively, in the non-linear contributing area algorithm.
Appendix 1 169
Soil Zone Module

The PRMS Soil Zone Module (soilzone_gsflow) calculates inflows to and outflows from the soil zone of each HRU and
includes inflows from infiltration, ground water, and upslope HRUs, and outflows to gravity drainage, interflow, and surface
runoff to downslope HRUs. Inflow from ground water is not allowed when using the PRMS-only model. The Soil Zone Module
is a combination and extension of the PRMS Soil-Moisture Balance (smbal_prms) and Subsurface Flow (ssflow_prms) Modules.
For more information about this module see section, Changes to PRMS in body of this report. Input parameters for the module
are defined in table A1-18.
Table A1-18. Input parameters required for PRMS Soil-Zone Module: soilzone_gsflow.
[Equation number refers to equations listed in the main body of reportequation variable of parameter name is defined in first listed equation. HRU, hydrologic
response unit; nhru, number of HRUs; nhrucell, number of unique intersections between gravity reservoirs in PRMS soil zone and MODFLOW finite-
difference cells; nssr, number of PRMS subsurface reservoirs]

cov_type Plant cover type (0=bare soil; nhru dimensionless integer 0 to 3 3
covden_sum Summer plant cover density for plant nhru dimensionless real 0.0 to 1.0 0.5
type as a decimal fraction
covden_win Winter plant cover density for plant nhru dimensionless real 0.0 to 1.0 0.5
type as a decimal fraction
fastcoef_lin Linear flow-routing coefficient for fast nhru per day real 0.0 to 1.0 0.1 67a
interflow
fastcoef_sq Non-linear flow-routing coefficient for nhru per inch-day real 0.0 to 1.0 0.8 67a
fast interflow
gvr_hru_id Identifier of HRU corresponding to nhrucell dimensionless integer 1 to nhru 1
each gravity reservoir
gvr_hru_pct Decimal fraction of HRU area nhrucell dimensionless real 0.0 to 1.0 0.0
associated with gravity reservoir
hru_area Area of HRU nhru acres real 0.1 to 1.0e9 1.0 22, 39, 45,
46, 58
2=lake)
pref_flow_den Decimal fraction of the soil zone nhru dimensionless real 0.0 to 1.0 0.2 38, 47
available for preferential flow
sat_threshold Maximum volume of water per unit nhru inches real 1.0 to 999.0 999.0 47, 65
area in the soil zone (set to 999.0
for infinite volume)
slowcoef_lin Linear flow-routing coefficient for nhru per day real 0.0 to 1.0 0.015 50, 56
slow interflow
slowcoef_sq Non-linear flow-routing coefficient for nhru per inch-day real 0.0 to 1.0 0.1 50, 56
slow interflow
soil_moist_init Initial value of available water in the nhru inches real 0.0 to 20.0 3.0
capillary reservoir
soil_moist_max Maximum volume of water per unit nhru inches real 0.0 to 20.0 6.0 44, 46,
area in the capillary reservoir 63a
soil_rechr_init Initial value in capillary reservoir nhru inches real 0.0 to 10.0 1.0
where evaporation and transpiration
can occur simultaneously (value
must be less than or equal to
soil_ moist_max)
soil_rechr_max Maximum value in capillary reservoir nhru inches real 0.0 to 10.0 2.0
where evaporation and transpiration
can occur simultaneously (value
must be less than or equal to
soil_ moist_max)
soil_type Soil type in HRU (1=sand; 2=loam; nhru dimensionless integer 1 to 3 2
3=clay)
Table A1-18. Input parameters required for PRMS Soil-Zone Module: soilzone_gsflowContinued.
[Equation number refers to equations listed in the main body of reportequation variable of parameter name is defined in first listed equation. HRU, hydrologic
response unit; nhru, number of HRUs; nhrucell, number of unique intersections between gravity reservoirs in PRMS soil zone and MODFLOW finite-
difference cells; nssr, number of PRMS subsurface reservoirs]

soil2gw_max Maximum value of soil-water excess nhru inches real 0.0 to 5.0 0.0
routed directly to PRMS ground-
water reservoir
ssr2gw_exp Exponent in the equation used to nssr or dimensionless real 0.0 to 3.0 1.0 59
compute gravity drainage to nhrucell
PRMS ground-water reservoir or
MODFLOW finite-difference cell
ssr2gw_rate Linear coefficient in the equation nssr or inches per day real 0.0 to 1.0 0.1 59
used to compute gravity drainage nhrucell
to PRMS ground-water reservoir or
ssrmax_coef Maximum amount of gravity drainage nssr or inches real 1.0 to 20.0 1.0 59
to PRMS ground-water reservoir or nhrucell
ssstor_init Initial storage in PRMS subsurface nssr or inches real 0.0 to 20.0 0.0
reservoir or gravity reservoir nhrucell
Ground-Water Flow Module

The PRMS Ground-Water Flow Module (gwflow_casc_prms) sums inflow to and computes storage and outflow from each
PRMS ground-water reservoir. Outflows can be routed to downslope ground-water reservoirs or to a stream. This module is used
for PRMS-only simulations. Input parameters for module gwflow_casc_prms are defined in table A1-19.
Table A1-19. Input parameters required for PRMS Ground-Water Flow Module: gwflow_casc_prms, included with PRMS-only
simulations.
[HRU, hydrologic response unit; nhru, number of HRUs; nssr, number of PRMS subsurface reservoirs; ngw, number of PRMS ground-water reservoirs]
Parameter Dimension Default

name variable value
hru_gwres Identifier of PRMS ground-water reservoir nhru dimensionless integer 1 to ngw 1
associated with HRU
gwstor_init Initial storage in ground-water reservoir ngw inches real 0.0 to 20.0 0.1
gwflow_coef1 Linear coefficient to route water in ground-water ngw per day real 0.0 to 1.0 .015
reservoir to streams
gwsink_coef1 Linear coefficient to route water in ground-water ngw per day real 0.0 to 1.0 0.0
reservoir to ground-water sink
1

gwflow_coef ground-water routing coefficient. The coefficient is multiplied by the storage in the ground-water reservoir to
calculate ground-water flow computation to downslope HRUs and(or) to streamflow.
gwsink_coefground-water sink coefficient. The coefficient is multiplied by the storage in the ground-water reservoir to
calculate seepage from each ground-water reservoir to the ground-water sink. Water in the ground-water sink is no longer
available for flow within the watershed.
Appendix 1 171
Streamflow Module
The PRMS Streamflow Module (strmflow_prms) calculates daily streamflow as the sum of surface runoff, interflow, flow
from detainment reservoirs, and ground-water flow. The module is used for PRMS-only simulations. Input parameters for the
module are defined in table A1-20.
Table A1-20. Input parameters required for PRMS Streamflow Module: strmflow_prms, included with PRMS-only simulations.
[HRU, hydrologic response unit; nhru, number of HRUs; nssr, number of PRMS subsurface reservoirs; ngw, number of PRMS ground-water reservoirs;
nsfres, number of on-stream detainment reservoirs; cfs, cubic feet per second]

name variable value
basin_cfs_init Initial streamflow at watershed outlet, required one cfs real 0.0 to 1e+09 0.0
if first time step is a storm period
hru_gwres Identifier of PRMS ground-water reservoir nhru dimensionless integer 1 to ngw 1
associated with HRU
Parameters only required when detainment reservoirs are included in a PRMS-only simulation
hru_sfres Identifier of detainment reservoir associated nhru dimensionless integer 0 to nsfres 0
with HRU
mxnsos Maximum number of values in table for Puls nsfres dimensionless integer 0 to 10 0
routing
o2 Outflow from detainment reservoir in table for mxnsos by cfs real 0.0 to 100,000.0 0.0
Puls routing nsfres
s2 Storage in detainment reservoir in table for Puls mxnsos by cfs-days real 0.0 to 100,000.0 0.0
routing nsfres
sfres_coef Linear coefficient to route detainment reservoir nsfres per day real 0.0 to 1.0 0.1
storage to streams
sfres_din1 Inflow to detainment reservoir from previous nsfres cfs real 0.0 to 1.0 0.1
time step
sfres_init Initial storage in detainment reservoir nsfres cfs-days real 0.0 to 0.0
2,000,000.0
sfres_qro Initial daily mean outflow from detainment nsfres cfs real 0.0 to 1.0 0.1
reservoir
sfres_type Type of detainment reservoir (8=Puls; nsfres dimensionless integer 8 or 9 8
9=linear)
upst_res1 Identifier of first upstream detainment reservoir nsfres dimensionless integer 0 to nsfres 0
upst_res2 Identifier of second upstream detainment nsfres dimensionless integer 0 to nsfres 0
reservoir
upst_res3 Identifier of third upstream detainment reservoir nsfres dimensionless integer 0 to nsfres 0
Hydrologic-Response-Unit Summary Module

The PRMS Hydrologic-Response-Unit Summary Module (hru_sum_prms) calculates daily, monthly, yearly, and total
summaries of volumes and flows for each HRU. Input parameters are defined in table A1-21.
Table A1-21. Input parameters required for PRMS Hydrologic-Response-Unit Summary Module: hru_sum_prms.
[HRU, hydrologic response unit; nhru, number of HRUs]

name variable value
covden_sum Summer plant cover density for plant type nhru dimensionless real 0.0 to 1.0 0.5
on HRU as a decimal fraction
covden_win Winter plant cover density for plant type on nhru dimensionless real 0.0 to 1.0 0.5
HRU as a decimal fraction
impervious
moyrsum Monthly and yearly summaries for each one dimensionless integer 0 or 1 0
HRU (0 for no summaries; 1 for
summaries)
pmo Month for which monthly summary one dimensionless integer 0 to 12 0
is written to PRMS Water-Budget
(0=no monthly summary; 1=January,
2=February, and so forth)
Basin Summary Module

The PRMS Basin Summary Module (basin_sum_prms) sums daily, monthly, yearly, and total summary values of volumes
and flows as area-weighted watershed averages. Input parameters for the module are defined in table A1-22.
Table A1-22. Input parameters required for PRMS Basin Summary Module: basin_sum_prms.
[HRU, hydrologic response unit; nhru, number of HRUs; ntemp, number of measurement stations that measure air temperature; nobs, number of streamflow-
gaging stations]

name variable value
basin_tsta Identifier of measurement station used in one dimensionless integer 0 to ntemp 1
computing air temperature
basin_tsta_hru Identifier of HRU used in computing watershed one dimensionless integer 0 to nhru 1
temperatures
objfunc_q Streamflow-gaging station used in objective one dimensionless integer 0 to nobs 1
function calculations
print_freq Frequency of output written in PRMS Water- one dimensionless integer 0 to 15 1
Budget File (0=none; 1=simulation totals;
2=yearly; 4=monthly; 8=daily; or additive
combinationsfor example, use 3 for output
of yearly and simulation totals)
print_objfunc Objective function output in PRMS Water- one dimensionless integer 0 to 1 0
Budget File (0=no; 1=yes)
print_type Type of output written in PRMS Water-Budget one dimensionless integer 0 to 2 1
File (0=measured and predicted flow only;
1=water balance table; 2=detailed output)
runoff_units Units of measured streamflows written in one dimensionless integer 0 or 1 0
PRMS Water-Budget File (0=cubic feet per
second; 1=cubic meters per second)
Appendix 1 173
GSFLOW Modules
Input instructions for GSFLOW modules are described next. Several parameters are used by more than one module.
Because the modules are defined independently of one another, descriptions of parameters used by more than one module are
repeated in each module. For example, the parameter gvr_cell_id is used by GSFLOW modules gsflow_setconv,
gsflow_prms2mf, gsflow_mf2prms, and gsflow_budget and a description of the parameter is repeated in each module.
GSFLOW Computation-Control Modules

Two GSFLOW modules are used to control the sequence of computations within GSFLOW. The module gsflow_prms is
used to call PRMS and GSFLOW modules for each daily time step in the proper sequence, which varies for each simulation
mode (integrated, PRMS-only, and MODFLOW-only). This module replaces the call_modules.c subroutine in the Modular
Modeling System. Module gsflow_modflow is used to call the MODFLOW-2005 packages and run procedures of the
PRMS Soil-Zone Module (soilzone_gsflow) as well as other GSFLOW modules in the proper sequence. It replaces the
MODFLOW-2005 main program. Module gsflow_prms requires one parameter, model_mode, which is specified in the
GSFLOW Control File. Module gsflow_modflow requires two parameters: (1) mxsziter, which is specified in the PRMS
Parameter File; and (2) modflow_name, which is specified in the GSFLOW Control File. Table A1-23 defines the input
parameters for the two GSFLOW Computation-Control Modules (gsflow_prms and gsflow_modflow).
Table A1-23. Input parameters required for GSFLOW Computation-Control Modules: gsflow_prms and gsflow_modflow.

name variable value
Parameter for gsflow_prms
1
model_mode Model to run one dimensionless integer GSFLOW, PRMS, GSFLOW
or MODFLOW
Parameters for gsflow_modflow
1
model_mode Model to run one dimensionless integer
GSFLOW, PRMS, GSFLOW
or MODFLOW
modflow_name1 Pathname of MODFLOW Name File one dimensionless character 1 to 256 modflow.nam
mxsziter2 Maximum number of iterations soil-zone one dimensionless integer 2 to 200 15
flow to MODFLOW finite-difference
cells are computed each time step
1
Parameter specified in GSFLOW Control File.
2
Parameter is not required in MODFLOW-only simulations.
GSFLOW Conversion Factors Module

A set of variables are determined in GSFLOW Conversion Factors Module (gsflow_setconv) that are used in other
GSFLOW modules to convert units between PRMS and MODFLOW-2005 during a simulation. The variables are summarized in
table 7 in the main body of this report. Input parameters for the module are defined in table A1-24.
Table A1-24. Input parameters required for GSFLOW Conversion Factors Module: gsflow_setconv.
[HRU, hydrologic response unit; nhru, number of HRUs; nhrucell, number of unique intersections between gravity reservoirs in PRMS soil zone and
MODFLOW finite-difference cells]

name variable value
gvr_cell_id Finite-difference cell associated with a gravity nhrucell dimensionless integer 0 to ngwcell 0
reservoir
gvr_cell_pct Decimal fraction of HRU area associated with nhrucell dimensionless real 0.0 to 1.0 0.0
a finite-difference cell
GSFLOW Integration Modules

Two GSFLOW modules (gsflow_prms2mf and gsflow_mf2prms) are used to integrate the spatial units and transfer
dependent variables (model states and fluxes) and volumetric flow rates between PRMS and MODFLOW-2005. Module
gsflow_prms2mf distributes gravity drainage from gravity reservoirs in the soil zone of PRMS to finite-difference cells
in MODFLOW-2005. The module also distributes surface runoff and interflow from HRUs to stream segments and
lakes in MODFLOW-2005. Module gsflow_mf2prms distributes ground-water discharge from finite-difference cells in
MODFLOW-2005 to gravity reservoirs in the soil zone of PRMS. Input parameters for the two modules are defined in
tableA1-25.
Table A1-25. Input parameters required GSFLOW Integration Modules: gsflow_prms2mf and gsflow_mf2prms.
MODFLOW finite-difference cells; ngwcell, number of MODFLOW finite-difference cells in a layer (includes active and inactive cells; nreach, number of
MODFLOW stream reaches; nsegment, number of MODFLOW stream segments]

name variable value
gvr_cell_id Finite-difference cell associated with a nhrucell dimensionless integer 0 to ngwcell 0
gravity reservoir
gvr_hru_id HRU associated with a gravity reservoir nhrucell dimensionless integer 1 to nhru 1
gvr_hru_pct Decimal fraction of HRU area associated nhrucell dimensionless real 0.0 to 1.0 0.0
with a gravity reservoir
lake_hru_id MODFLOW lake number associated with nhru dimensionless integer 0 to nhru 0
an HRU
Additional parameters for gsflow_prms2mf
gvr_cell_pct Decimal fraction of HRU area associated nhrucell dimensionless real 0.0 to 1.0 0.0
with a finite-difference cell
hru_segment HRU associated with a stream segment nhru dimensionless integer 0 to nsegment
local_reachid Stream reach within a stream segment nreach dimensionless integer 0 to nreach 0
mxsziter Maximum iterations for computing soil-zone one dimensionless integer 2 to 200 15
flow to finite-difference cells during a
time step
numreach_segment Number of stream reaches in a stream nsegment dimensionless integer 0 to nreach 1
segment
reach_segment Stream segment associated with a stream nreach dimensionless integer 0 to nsegment 0
reach
segment_pct_area Decimal fraction of HRU area that nreach dimensionless real 0.0 to 1.0 0.0
contributes flow to a stream reach
szconverge1 Convergence criterion for checking soil-zone one inches real 1.e-15 to 1.e-1 1.e-8
flows
1
Additional information for one of the parameters defined in table A1-25 follows:
szconvergeconvergence criterion for checking soil-zone states that is specified as the smallest change in gravity drainage
from the PRMS gravity-flow reservoir between iterations.
Appendix 1 175
GSFLOW Budget Module

The GSFLOW Budget Module (gsflow_budget) calculates a watershed budget for GSFLOW and adjusts the final storage in
gravity reservoirs using flows to and from finite-difference cells at the end of each time step. Input parameters for gsflow_budget
are defined in table A1-26.
Table A1-26. Input parameters required for GSFLOW Budget Module: gsflow_budget.
MODFLOW finite-difference cells; ngwcell, number of MODFLOW finite-difference cells in a layer (includes active and inactive cells); cfs, cubic foot per
second]

name variable value
basin_cfs_init Initial streamflow at outlet one cfs real 0.0 to 1e+09 0.0
gvr_cell_id Identifier of finite-difference cell associated nhrucell dimensionless integer 0 to ngwcell 0
gvr_hru_id Identifier of HRU associated with a gravity nhrucell dimensionless integer 1 to nhru 1
reservoir
gvr_hru_pct Decimal fraction of HRU area associated nhrucell dimensionless real 0.0 to 1.0 .0
lake_hru_id MODFLOW lake number associated with nhru dimensionless integer 0 to nhru 0
an HRU
GSFLOW Summary Module

The GSFLOW Summary Module (gsflow_sum) calculates summary tables of the water balance at the end of each
time step. Input parameters for gsflow_sum are defined in table A1-27. Parameters id_obsrunoff, and runoff_units
are specified in the PRMS Parameter File. Parameters csv_output_file, gsf.rpt, gsflow_output_file,
model_output_file, and rpt_days are specified in the GSFLOW Control File.
Table A1-27. Input parameters required for GSFLOW Summary Module: gsflow_sum.
[HRU, hydrologic response unit; nhru, number of HRUs; ntemp, number of measurement stations that measure air temperature; nobs, number of streamflow
gaging stations]

name variable value
csv_output_file1 Pathname for GSFLOW Comma- one dimensionless character 1 to 256 gsflow.csv
Separated-Values (CSV) File
gsf_rpt1 Switch to specify whether or not the one dimensionless integer Dor1 1
GSFLOW Comma-Separated-Values
(CSV) File is generated (0=no, 1=yes)
gsflow_output_file1 Pathname for GSFLOW Water Budget one dimensionless character 1 to 256 gsflow.out
File
id_obsrunoff Identifier for streamflow-gaging station one dimensionless integer 0 to nobs 0
at outlet
model_output_file1 Pathname for PRMS Water Budget File one dimensionless character 1 to 256 prms.out
rpt_days1 Frequency that summary tables are one days integer 1 to 365 7
written to GSFLOW Water-Budget
File (0=none, >0 frequency in days,
e.g., 1=daily, 7=every 7th day)
runoff_units Units of measured streamflow (0=cubic one dimensionless integer 0 or 1 0
feet per second; 1=cubic meters per
second)
1
Parameter specified in GSFLOW Control File.
MODFLOW Input Files

Input files for most of the MODFLOW packages were not changed for use with GSFLOW and the following input
instructions were extracted from existing MODFLOW documentation. However, GSFLOW does restrict the values of some
MODFLOW input variables. Information regarding these input data restrictions is differentiated from the original input
instructions in the following sections by placing braces ({ }) around the added text. Additionally, input instructions related
to parameters were removed for simplicity, and no description of the removed parameter items is included in this appendix.
MODFLOW packages within GSFLOW that previously read data using parameters can still be used in GSFLOW. Input
instructions for using parameters in MODFLOW are provided by Harbaugh (2005). Some parameter-related variables are
described in this appendix and must be assigned a zero value when parameters are not used in a simulation.
Input instructions for the Ground-Water Flow Process of MODFLOW-2005 are given separately for each package used
by GSFLOW. The Basic Package reads several files. Each file is described in a separate subsection. The final sections in this
chapter describe the Array Reading Utility Subroutines and the List Utility Subroutine.
Data for cells in the model grid can be categorized in two ways according to the density or sparseness of the data. Layer
data refers to any type of data for which a value is required for every cell in one or more layers of the grid. Examples of layer
data include hydraulic conductivity and specific storage. List data refers to any type of data for which data values are required
for selected cells in the grid. Examples include the well withdrawal rate as simulated with the Well (WEL) Package and head-
dependent flow as simulated using a specified head and hydraulic conductance with the General-Head Boundary (GHB)
Package.
Layer data are described by the number of columns and rows in parentheses after the variable. Although the program
contains many three-dimensional variables that are defined through user input, these are treated as multiple-layer variables
for input purposes. For example, IBOUND(NCOL, NROW) indicates that IBOUND is a layer variable. An IBOUND variable is
read for each layer of the grid. Most multiple-layer variables are read by utility subroutines. which are named along with the
variable. The utility subroutine U1DREL (Utility-1D-Real) is used for reading a one-dimensional array of variables that are
real numbers, U2DREL is used for reading an array of two-dimensional variables that are real numbers, and U2DINT is used for
reading an array of two-dimensional variables that are integers.
List data of variables include the layer, row, and column indices that are used to identify a particular cell and multiple lines
of data can be read. List data are usually read by the list utility subroutine, ULSTRD (Utility-List-Read). The use of ULSTRD is
mentioned as part of the note that describes the repeated item.
The input data for each item must start on a new line. All variables for an item are assumed to be contained in a single line
unless the item consists of a multi-valued variable. That is, a multi-valued variable may occupy multiple lines. When reading a
layer variable, the data for each row must start on a new line.
Each input variable has a data type, which can be Real, Integer, or Character. Integers are whole numbers and must not
include a decimal point or exponent. Real numbers can include a decimal point and an exponent. If no decimal point is included
in the entered value, then the decimal point is assumed to be at the right side of the value. Any printable character is allowed
for character variables. All variables starting with the letters I-N are integers. Variables starting with the letters A-H and O-Z
are either real numbers or character data, and most of these variables are real numbers; that is, there are few character variables
in MODFLOW. A variable starting with A-H or O-Z is assumed to be a real number unless otherwise stated in the variable
definitions. This data-type convention is used for all input variables regardless of whether a variable is an actual program
variable.
Both free and fixed format is used as specified throughout the input instructions. With fixed format, a data value has a
specified column location, called the field, within a line. If a numeric value does not require the entire field width, the value
should be right justified within the field, and blanks should fill the remaining width of the field. The justification of character
data within a field does not matter unless specifically noted.
Free format is similar to Fortan list directed input. With free format, values are not required to occupy a fixed number of
columns in a line. Each value can occupy one or more columns as required to represent the value; however, the values must
be included in the prescribed order. One or more spaces, or a single comma optionally combined with spaces, must separate
adjacent values. Also, a numeric value of zero must be explicitly represented with 0 and not by one or more spaces when free
format is used. Free-format input implemented in MODFLOW-2005 does not allow for null values in which input values are left
unchanged from their previous values. A forward slash ( / ) cannot be used to terminate an input item without including values
for all the variables; required data values for all input variables must be explicitly specified on an input item. MODFLOW does
not require apostrophes to delineate apostrophes unless a blank or a comma is part of a character variable.
Appendix 1 177
Name File
Although the Name File is actually used as part of the Basic Package, it is described separately because it is fundamental
for identifying which input files will be used in the MODFLOW-components of GSFLOW. The Name File contains the names
of the input and output files used for MODFLOW in a GSFLOW simulation and controls whether or not parts of the model
program are active. (Temporary OPEN/CLOSE files, described in section Array Reading Utility Subroutines, are not included
in the Name File.) The Name File is read on FORTRAN unit 99. The Name File is constructed as follows.
FOR EACH SIMULATION
1. Ftype Nunit Fname [Fstatus]

The Name File contains one of the above lines (item 1) for each file. All variables are free format. The length of each
line must be 299 characters or less. The lines can be in any order except for the line where Ftype (file type) is LIST as
described in the subsequent Explanation of Variables.
Comment lines are indicated by the # character in column one and can be located anywhere in the file. Any text
characters can follow the # character. Comment lines have no effect on the simulation; their purpose is to allow users
to provide documentation about a particular simulation. All comment lines after the first item-1 line are written in the
listing file.
The Name File must include two files required for each simulation, the BAS Package File and the Discretization File.
Explanation of Variables in the Name File:

Ftypeis the file type, which must be one of the following character values. Ftype may be entered in any combination of
uppercase and lowercase. {Note that the file types listed below only include those that can be used in the MODFLOW-only or
GSFLOW mode of a GSFLOW simulation; see table 1 in body of report.}
LIST for the Listing Filethis type must be present and must be the first file in the Name File.
DIS for the Discretization File
BAS6 for the Basic Package
OC for the Output Control Option
CHD for the Time-Variant Specified-Head Option
BCF6 for the Block-Centered Flow Package
LPF for the Layer Property Flow package
HFB6 for the Horizontal Flow Barrier Package
WEL for the Well Package
MNW1 for the Multi-Node Well Package
GHB for the General-Head Boundary Package
FHB for the and Flow and Head Boundary Package
SFR for the Streamflow Routing (SFR2) Package
LAK for Lake Package
GAGE for the Gage Package
UZF for the Unsaturated-Zone Flow Package
SIP for the Strongly Implicit Procedure Package
PCG for the Preconditioned Conjugate-Gradient Package
DE4 for the Direct Solution Package
DATA(BINARY) for binary (unformatted) files, such as those used to save cell-by-cell budget data and binary
(unformatted) head and drawdown data.
DATA for formatted (text) files, such as those used to save formatted head and drawdown and for input of data from files
that are separate from the primary package input files.
Nunitis the Fortran unit to be used when reading from or writing to the file. Any legal unit number on the computer being
used can be specified except unit 99. Unit 99 is used for the Name File and for reading multi-valued variables using the
OPEN/CLOSE option of the utility subroutines (see section, Array Reading Utility Subroutines). The unit number for
each file must be unique.
Fnameis the name of the file, which is a character value. Pathnames may be specified as part of Fname.
Fstatusis the optional file status, which applies only to file types Data and Data(Binary). Two values are allowed: OLD
and REPLACE. OLD indicates that the file should already exist. REPLACE indicates that if the file already exists, then it
should be deleted before opening a new file. The default is to open the existing file if the file exists or create a new file if
the file does not exist.
Basic Package Input Instructions

The Basic Package reads several files: Name File (described above), Discretization File, Basic Package file, Multiplier
Array File, Zone Array File, Output Control Option File, and the Constant-Head Boundary (that is, the Time-Variant Specified-
Head) Option File. The Name File, Discretization File, and Basic Package File are required for all simulations. The other
files are optional. Input instructions for the Multiplier Array and Zone Array Files are part of MODFLOW parameters and are
described in Harbaugh (2005, p. 8-15 and 8-16) but not herein.
Discretization File
Discretization information is read from the file that is specified by DIS as the file type.
FOR EACH SIMULATION
0. [#Text]
Item 0 is optional# must be in column 1. Item 0 can be repeated multiple times.
1. NLAY NROW NCOL NPER ITMUNI LENUNI
2. LAYCBD(NLAY)
3. DELR(NCOL) - U1DREL
Figure A1-3 illustrates the
C O LU M N S
orientation of DELR and DELC. 1 2 3 N COL
4. DELC(NROW) - U1DREL
5. Top(NCOL,NROW) - U2DREL 1 D E LC
1
6. BOTM(NCOL,NROW) - U2DREL
2 D E LC 2
Item 6 is repeated for each model
layer and Quasi-3D confining bed
ROW S 3 D E LC
in the grid. Thus, the number of 3
BOTM variables must be NLAY plus

the number of Quasi-3D confining
beds. The BOTM variables are read N ROW D E LC
in sequence going down from the NROW
top of the system. For example, in

D E LR 1 D E LR2 D E LR3 D E LRNC O L
a 3-layer model with a Quasi-3D
confining bed below layer 2, there
would be four BOTM arrays. The
arrays would be the bottom of Figure A1-3. Plan view of grid showing DELR and DELC values.
layer 1, the bottom of layer 2, the
bottom of the Quasi-3D confining
bed below layer 2, and the bottom
of layer 3.
Appendix 1 179
FOR EACH STRESS PERIOD
7. PERLEN NSTP TSMULT Ss/Tr
Explanation of Variables Read from the Discretization File:

Textis a character variable (199 characters) that starts in column 2. Any characters can be included in Text. The # character
must be in column 1. Lines beginning with # are restricted to the first lines of the file. Text is written to the Listing File.
NLAYis the number of layers in the model grid.
NROWis the number of rows in the model grid.
NCOLis the number of columns in the model grid.
NPERis the number of stress periods in the simulation.
ITMUNIindicates the time unit of model data, which must be consistent for all data values that involve time. For example, if
the chosen time unit is years (ITMUNI=5), then stress-period length, time-step length, transmissivity, and so on, must
all be expressed using years for their time units. {Unlike a normal MODFLOW run, ITMUNI cannot be undefined when
using GSFLOW}. When the time unit is defined, MODFLOW uses it to print a table of elapsed simulation time:
1 seconds 4 - days
2 minutes 5 - years
3 hours
LENUNIindicates the length unit of model data, which must be consistent for all data values that involve length. For example,
if feet is the chosen length unit, grid spacing, head, hydraulic conductivity, water volumes, and so forth, must all be
expressed using feet for their length units. {Unlike a normal MODFLOW run, ITMUNI cannot be undefined when using
GSFLOW}. Be sure to use consistent units for all input data even when LENUNI indicates an undefined length unit:
1 feet
2 meters
3 centimeters
LAYCBDis a flag, with one value for each model layer, which indicates whether or not a layer has an implicit or Quasi-three-
dimensional confining bed below it. 0 indicates no confining bed, and not zero indicates a confining bed. LAYCBD for
the bottom layer must be 0.
DELRis the cell width along rows. Read one value for each of the NCOL columns. This is a multi-value one-dimensional
variable with one value for each column.
DELCis the cell width along columns. Read one value for each of the NROW rows. This is a multi-value one-dimensional
variable with one value for each row.
Top{in a normal MODFLOW simulation, Top is the top elevation of layer 1. However, for a GSFLOW simulation, Top
represents the elevation of the soil-zone base}
BOTMis the bottom elevation of a model layer or a Quasi-3d confining bed.
PERLENis the length of a stress period.
NSTPis the number of time steps in a stress period.
TSMULT{in a normal MODFLOW simulation, TSMULT is the multiplier for the length of successive time steps. However, for
a GSFLOW simulation, TSMULT must be set to a value of 1.0. NSTP and PERLEN must be specified such that the
time step length is equal to 1day and the period length is equal to multiples of 1 day.}
Ss/Tris a character variable that indicates whether the stress period is transient or steady state. The only allowed options are
Ss and Tr, but these are case insensitive. {In a GSFLOW simulation, Ss/Tr can be set to Ss for the first stress period
only.}
Basic Package File

The Basic (BAS) Package file is specified with BAS6 as the file type.
FOR EACH SIMULATION
0. [#Text]
1. Options (199 text characters)
If there are no options to specify, then a blank line must be included for item 1.
2. IBOUND(NCOL,NROW) or (NCOL,NLAY) U2DINT
3. HNOFLO (10-space field unless item 1 contains FREE.)
4. STRT(NCOL,NROW) or (NCOL,NLAY) U2DREL
A layer variable is read for each layer in the grid.
Explanation of Variables Read from the BAS Package File:

Textis a character variable that starts in column 2. The first two comment lines will become variable HEADNG, which is
used as a printout title throughout the program. (If there are no comment lines, then HEADNG will be blank.) HEADNG is
limited to 80 columns, but subsequent Text lines can be up to 199 columns. Any characters can be included in Text.
The # character must be in column 1. Lines beginning with # are restricted to the first lines of the file. Text is written to
the Listing File.
Optionsis a character variable that is scanned for words (separated by one or more spaces) that specify program options.
Two options are currently (2007) recognized. Unrecognized words are ignored, and a word may be specified in either
uppercase or lowercase. A blank line is acceptable and indicates no options. The option XSECTION(courier) is not
supported in GSFLOW.
CHTOCH indicates that flow between adjacent constant-head cells should be calculated.
FREE indicates that free format is used for input variables throughout the Basic Package and other packages as indicated
in their input instructions. Be sure that all variables read using free format have a non-blank value and that a comma or
at least one blank separates all adjacent values.
IBOUNDis the boundary variable. One value is read for every finite-difference cell. These values are read one layer at a time.
Note that although IBOUND is read as one or more two-dimensional variables, IBOUND is stored internally as a three-
dimensional variable.
If IBOUND(J,I,K) < 0, cell J,I,K has a constant head.
If IBOUND(J,I,K) = 0, cell J,I,K is no flow.
If IBOUND(J,I,K) > 0, cell J,I,K is variable head.
HNOFLOis the value of head to be assigned to all no-flow cells (IBOUND = 0). Because head at no-flow cells is unused in
model calculations, this does not affect model results but serves to identify no-flow cells when head is printed. This value
is used also as drawdown at no-flow cells if the drawdown option is used. Even if the user does not anticipate having
no-flow cells, a value for HNOFLO must be entered.
STRTis initial (starting) headthat is, head at the beginning of the simulation. STRT must be specified for all simulations,
including steady-state simulations. One value is read for every finite-difference cell. For simulations in which the first
stress period is steady state, the values used for STRT generally do not affect the simulation (exceptions may occur if
cells go dry and (or) rewet). The execution time, however, will be less if STRT includes hydraulic heads that are close to
the steady-state solution.
Appendix 1 181
Output Control Option File

Input to the Output Control Option File of the Ground-Water Flow Process is read from the file that is specified as type
OC in the Name File. If no OC file is specified, default output control is used. Under the default, head and overall budget are
written to the Listing File at the end of every MODFLOW stress period. The default printout format for head and drawdown is
10G11.4.
Output Control data may be specified as words or numeric codes. One of these methods must be used throughout any
simulation.
Output control using words

Recognized words are shown in BOLD ITALICS; these words must be entered exactly as shown except that they may be
entered in either uppercase or lowercase. Optional parts of lines are shown in brackets. One or more spaces must separate each
word or variable, and the total line length must not exceed 199 characters.
FOR EACH SIMULATION
0. [#Text]
1. Any combination of the following lines:
HEAD PRINT FORMAT IHEDFM
Specifies the format for writing head to the Listing File.
HEAD SAVE FORMAT CHEDFM [LABEL]
Specifies the format for writing head to a file other than the Listing File. Omit this line to obtain a binary (unformatted)
file. Binary files usually are smaller than text files, but they are not generally transportable among different computer
operating systems or different Fortran compilers.
HEAD SAVE UNIT IHEDUN
Specifies the file unit for writing head to a file other than the Listing File.
DRAWDOWN PRINT FORMAT IDDNFM
Specifies the format for writing drawdown to the Listing File.
DRAWDOWN SAVE FORMAT CDDNFM [LABEL]
Specifies the format for writing drawdown to a file other than the Listing File. Omit this line to obtain an unformatted
(binary) file. Binary files usually are smaller than text files, but they are not generally transportable among different
computer operating systems or different Fortran compilers.
DRAWDOWN SAVE UNIT IDDNUN
Specifies the file unit for writing drawdown to a file other than the Listing File.
IBOUND SAVE FORMAT CBOUFM [LABEL]
Specifies the format for writing IBOUND to a file.
IBOUND SAVE UNIT IBOUUN
Specifies the file unit for writing IBOUND to a file.
COMPACT BUDGET [AUX or AUXILIARY]
COMPACT BUDGET indicates that the cell-by-cell budget file(s) will be written in a more compact form than is used
in the 1988 version of MODFLOW (referred to as MODFLOW-88) (McDonald and Harbaugh, 1988); however,
programs that read these data in the form written by MODFLOW-88 will be unable to read the new compact file. If
this option is not used, MODFLOW-2005 will write the files using the MODFLOW-88 form. The optional word AUX
(or AUXILIARY) indicates that auxiliary data that are defined in packages (see input data for the RIV, WEL, DRN,
and GHB Packages) should be saved in the budget file along with budget data.
FOR EACH TIME STEP FOR WHICH OUTPUT IS DESIRED
2. PERIOD IPEROC STEP ITSOC

3. Any combination of the following lines:
PRINT HEAD [list layers if all layers not desired]
Head is written to the Listing File.
PRINT DRAWDOWN [list layers if all layers not desired]

Drawdown is written to the Listing File.
PRINT BUDGET
Overall volumetric budget is written to the Listing File.
SAVE HEAD [list layers if all layers not desired]
Head is written to a file other than the Listing File.
SAVE DRAWDOWN [list layers if all layers not desired]
Drawdown is written to a file other than the Listing File.
SAVE IBOUND [list layers if all layers not desired]
IBOUND is written to a file other than the Listing File. This option is provided to allow changes in IBOUND to be
recorded in simulations where IBOUND changes during a simulation.
SAVE BUDGET
Cell-by-cell budget data are written to the files that are designated in the packages that compute budget terms.
Item 2 and one or more item-3 lines are specified each time for which output is desired. These lines must be in the order of
increasing simulation time.
Explanation of Variables Read by Output Control Using Words:

IHEDFMis a code for the format in which heads will be printed. (Positive values indicate wrap format; negative values
indicate strip format.)
0 10G11.4 11 20F5.4
1 11G10.3 12 10G11.4
2 9G13.6 13 10F6.0
3 15F7.1 14 10F6.1
4 15F7.2 15 10F6.2
5 15F7.3 16 10F6.3
6 15F7.4 17 10F6.4
7 20F5.0 18 10F6.5
8 20F5.1 19 5G12.5
9 20F5.2 20 6G11.4
10
20F5.3 21 7G9.2
CHEDFMis a character value that specifies the format for saving heads, and can only be specified if the word method of output
control is used. The format must contain 20 characters or less and must be a valid Fortran format that is enclosed in
parentheses. The format must be enclosed in apostrophes if it contains one or more blanks or commas. The optional word
LABEL after the format is used to indicate that each layer of output should be preceded with a line that defines the output
(simulation time, the layer being output, and so forth). If there is no line specifying CHEDFM, then heads are written to a
binary (unformatted) file. Binary files are usually more compact than text files, but they are not generally transportable
among different computer operating systems or different Fortran compilers.
IHEDUNis the unit number on which heads will be saved.
IDDNFMis a code for the format in which drawdowns will be printed. The codes are the same as for IHEDFM.
CDDNFMis a character value that specifies the format for saving drawdown, and can only be specified if the word method of
output control is used. The format must contain 20 characters or less and must be a valid Fortran format that is enclosed
in parentheses. The format must be enclosed in apostrophes if it contains one or more blanks or commas. The optional
word LABEL after the format is used to indicate that each layer of output should be preceded with a line that defines the
output (simulation time, the layer being output, and so forth). If there is no line specifying CDDNFM, then drawdown is
written to a binary (unformatted) file. Binary files are usually more compact than text files, but they are not generally
transportable among different computer operating systems or different Fortran compilers.
Appendix 1 183
IDDNUNis the unit number on which drawdowns will be saved.
CBOUFMis a character value that specifies the format for saving IBOUND, and can only be specified if the word method
of output control is used. The format must contain 20 characters or less and must be a valid Fortran format that is
enclosed in parentheses. The format must be enclosed in apostrophes if it contains one or more blanks or commas. The
optional word LABEL is used to indicate that each layer of output should be preceded with a line that defines the output
(simulation time, the layer being output, and so forth). If there is no line specifying CBOUFM, then IBOUND is written
using format (20I4).
IBOUND is never written as a binary (unformatted) file. HEAD PRINT FORMAT 15

HEAD SAVE FORMAT (20F10.3) LABEL
IBOUUNis the unit number on which IBOUND will be saved. HEAD SAVE UNIT 30
COMPACT BUDGET
IPEROCis the stress period number at which output is desired. DRAWDOWN PRINT FORMAT 14
PERIOD 1 STEP 1
ITSOCis the time step number (within a stress period) at which output
PRINT HEAD 2 6
is desired.
PRINT DRAWDOWN
Figure A1-4 shows an example of output control using words. Note PRINT BUDGET
that the first line cannot be blank, but after the first line, blank lines are SAVE BUDGET
ignored when the word method is used to specify Output Control data. SAVE HEAD
Indented lines are allowed because of the use of free format input. PERIOD 1 STEP 7
SAVE HEAD 1 3 5
Output Control Using Numeric Codes PRINT DRAWDOWN
SAVE BUDGET
All variables are free format if the word FREE is specified in PERIOD 2 STEP 5
item 1 of the Basic Package input file; otherwise, the variables all have PRINT HEAD
10-character fields. PRINT BUDGET
SAVE BUDGET
FOR EACH SIMULATION SAVE HEAD
0. [#Text] Figure A1-4. Example Output Control input using
Item 0 is optional# must be in column 1. Item 0 can be words.
repeated multiple times.
1. IHEDFM IDDNFM IHEDUN IDDNUN
FOR EACH TIME STEP
2. INCODE IHDDFL IBUDFL ICBCFL

3. Hdpr Ddpr Hdsv Ddsv
(Item 3 is read 0, 1, or NLAY times, depending on the value of
INCODE.)
Explanation of Variables Read by Output Control Using Numeric

Codes:
Textis a character variable (199 characters) that starts in column 2.
Any characters can be included in Text. The # character must be
in column 1. Lines beginning with # are restricted to the first lines
of the file. Text is written to the Listing File.
IHEDFMis a code for the format in which heads will be printed. See
the description above in the explanation of variables read by
output control using words.
IDDNFMis a code for the format in which drawdowns will be printed.

The codes are the same as for IHEDFM.
IHEDUNis the unit number on which heads will be saved.
IDDNUNis the unit number on which drawdowns will be saved.
INCODEis the code for reading item 3.

If INCODE < 0, item 3 flags are used from the last time step. Item 3 is not read.
If INCODE = 0, all layers are treated the same way. Item 3 will consist of one line.
If INCODE > 0, item 3 will consist of one line for each layer.
IHDDFLis a head and drawdown output flag. This flag allows item 3 flags to be specified in an early time step and then used
or not used in subsequent time steps. Thus, IHDDFL can be used to avoid resetting item 3 flags every time step.
If IHDDFL = 0, no heads or drawdowns will be printed or saved regardless of which item 3 flags are specified.
If IHDDFL 0, heads and drawdowns will be printed or saved according to the item 3 flags.
IBUDFLis a budget print flag.
If IBUDFL = 0, overall volumetric budget will not be printed.
If IBUDFL 0, overall volumetric budget will be printed.
ICBCFLis a flag for writing cell-by-cell flow data.
If ICBCFL = 0, cell-by-cell flow terms are not written to any file.
If ICBCFL 0, cell-by-cell flow terms are written to the Listing File or a Budget File depending on flags set in the
component of flow packages, that is, IWELCB, IRCHCB, and so forth.
Hdpris the output flag for head printout.
If Hdpr = 0, head is not printed for the corresponding layer.
If Hdpr 0, head is printed for the corresponding layer.
Ddpris the output flag for drawdown printout.
If Ddpr = 0, drawdown is not printed for the corresponding layer.
If Ddpr 0, drawdown is printed for the corresponding layer.
Hdsvis the output flag for head save.
If Hdsv = 0, head is not saved for the corresponding layer.
If Hdsv 0, head is saved for the corresponding layer.
Ddsvis the output flag for drawdown save.
If Ddsv = 0, drawdown is not saved for the corresponding layer.
If Ddsv 0, drawdown is saved for the corresponding layer.
Constant-Head (Time-Variant Specified-Head) Option File

Input to the Constant-Head (CHD; Time-Variant Specified-Head) Option File is read from the file that has file type CHD
in the Name File. Optional variables are shown in brackets. All variables are free format if the option FREE is specified in
the Basic Package input file; otherwise, the non-optional variables have 10-character fields and the optional variables are free
format.
Once a cell is made constant head, the cell stays constant head throughout the remainder of the simulation. For example,
if a cell is listed in the CHD file as constant head in stress period 1 and not listed in stress period 2, then the cell continues to
be constant head in stress period 2 and throughout the remainder of the stress periods. The head is adjusted only in the stress
periods in which a cell is listed. For the stress periods in which a constant-head cell is not listed, the head stays at the value that
it had at the end of the previous stress period.
Appendix 1 185
FOR EACH SIMULATION
0. [#Text]
1. MXACTC [Option]
2. ITMP NP
3. Layer Row Column Shead Ehead [xyz]
ITMP repetitions of item 3 are read by subroutine ULSTRD if ITMP > 0. SFAC of the ULSTRD utility subroutine
applies to Shead and Ehead. Item 3 is not read if ITMP is negative or 0.
Explanation of Variables Read by the CHD Option:
MXACTCis the maximum number of constant-head boundary cells in use during any stress period.
Optionis an optional list of character values.

AUXILIARY abc or AUX abcdefines an auxiliary variable, named abc, which will be read for each constant-head
boundary as part of item 3. Up to five variables can be specified, each of which must be preceded by AUXILIARY or
AUX. These variables will not be used by the Ground-Water Flow Process, but they will be available for use by other
processes. The auxiliary variable values will be read after the Ehead variable.
ITMPis a flag and a counter.

If ITMP < 0, data from the preceding stress period will be reused.
If ITMP 0, ITMP is the number of constant-head boundaries read for the current stress period.
NP{this value should be set to zero when parameters are not used.}
Layeris the layer number of the constant-head boundary.
Rowis the row number of the constant-head boundary.
Columnis the column number of the constant-head boundary.
Sheadis the head at the boundary at the start of the stress period.
Eheadis the head at the boundary at the end of the stress period.
[xyz]represents the values of the auxiliary variables for a constant-head boundary that have been defined in item 1. The
values of auxiliary variables must be present in each repetition of items 2 and 3 if they are defined in item 1. The values
must be specified in the order used to define the variables in item 1.
Block-Centered Flow Package

Input for the Block-Centered Flow (BCF) Package is read from the file that is type BCF6 in the Name File. The BCF
Package is an alternative to the LPF and HUF Packages. Only one of these packages should be used during a simulation.
FOR EACH SIMULATION
1. IBCFCB HDRY IWDFLG WETFCT IWETIT IHDWET

These six variables are free format if the option FREE is specified in the Basic Package input file; otherwise, the
variables all have 10-character fields.
2. Ltype(NLAY)
Read one value for each layer. These values are free format if the word FREE is specified in item 1 of the Basic Package
input file; otherwise, the values are read using fixed format fields that are each 2 characters wide with 40 values per
line. Use only as many lines as required for the number of model layers.
3. TRPY(NLAY) U1DREL
A subset of the following two-dimensional variables is used to describe each layer. The variables needed for each layer
depend on the layer-type code (LAYCON, which is defined as part of Ltype), whether the simulation has any transient
stress periods (at least one stress period defined in the Discretization File specifies Ss/Tr as TR), and if the wetting
capability is active (IWDFLG not 0). Unneeded variables must be omitted. In no situation will all variables be required.
The required variables (items 4-9) for layer 1 are read first; then the variables for layer 2 and so forth.
4. [Sf1(NCOL,NROW)] U2DREL If there is at least one transient stress period.
If LAYCON is 0 or 2 (see Ltype ), then read item 5.
5. [Tran(NCOL,NROW)] U2DREL.
Otherwise, if LAYCON is 1 or 3 (see Ltype), read item 6.
6. [HY(NCOL,NROW)] U2DREL
7. [Vcont(NCOL,NROW)] U2DREL. If not the bottom layer.
8. [Sf2(NCOL,NROW)] U2DREL. If there is at least one transient stress period and LAYCON (see Ltype) is 2 or 3.
9. [WETDRY(NCOL,NROW)] U2DREL. If IWDFLG is not 1 and LAYCON is 1 or 3 (see Ltype).
Explanation of Variables Read by the BCF Package:

IBCFCBis a flag and a unit number.
If IBCFCB > 0, cell-by-cell flow terms will be written to this unit number when SAVE BUDGET or a non-zero value for
ICBCFL is specified in Output Control. The terms that are saved are storage, constant-head flow, and flow between
adjacent cells.
If IBCFCB = 0, cell-by-cell flow terms will not be written.
If IBCFCB < 0, cell-by-cell flow for constant-head cells will be written in the listing file when SAVE BUDGET or a non-
zero value for ICBCFL is specified in Output Control. Cell-by-cell flow to storage and between adjacent cells will not
be written to any file.
HDRYis the head that is assigned to cells that are converted to dry during a simulation. Although this value plays no role in
the model calculations, HDRY values are useful as indicators when looking at the resulting heads that are output from the
model. HDRY is thus similar to HNOFLO in the Basic Package, which is the value assigned to cells that are no-flow cells
at the start of a model simulation.
IWDFLGis a flag that determines if the wetting capability is active.

If IWDFLG = 0, the wetting capability is inactive.
If IWDFLG is not 0, the wetting capability is active.
WETFCTis a factor that is included in the calculation of the head that is initially established at a cell when that cell is
converted from dry to wet. (See IHDWET.)
IWETITis the iteration interval for attempting to wet cells. Wetting is attempted every IWETIT iteration. This applies to
outer iterations and not inner iterations. If IWETIT is 0, the value is changed to 1.
IHDWETis a flag that determines which equation is used to define the initial head at cells that become wet:
If IHDWET = 0, equation 5-32A in Harbaugh (2005) is used: h = BOT + WETFCT (hn - BOT)
If IHDWET is not 0, equation 5-32B in Harbaugh (2005) is used: h = BOT + WETFCT (THRESH)
Ltypecontains a combined code for each layer that specifies both the layer type (LAYCON) and the method of computing
interblock conductance. Use as many lines as needed to enter a value for each layer. Values are two-digit numbers.
Appendix 1 187
The left digit defines the method of calculating interblock transmissivity. The methods are described by Goode and Appel
(1992).
0 or blankharmonic mean (the method used in MODFLOW-88).

1arithmetic mean
2logarithmic mean
3arithmetic mean of saturated thickness and logarithmic-mean hydraulic conductivity.
The right digit defines the layer type (LAYCON), which is the same as in MODFLOW-88:
0confinedTransmissivity and storage coefficient of the layer are constant for the entire simulation.
1unconfinedTransmissivity of the layer varies and is calculated from the saturated thickness and hydraulic
conductivity. The storage coefficient is constant. This type code is valid only for layer1.
2confined/unconfinedTransmissivity of the layer is constant. The storage coefficient may alternate between confined
and unconfined values. Vertical flow from above is limited if the layer desaturates.
3confined/unconfinedTransmissivity of the layer varies and is calculated from the saturated thickness and hydraulic
conductivity. The storage coefficient may alternate between confined and unconfined values. Vertical flow from above
is limited if the aquifer desaturates.
TRPYis a one-dimensional variable containing a horizontal anisotropic factor for each layer and is the ratio of transmissivity
or hydraulic conductivity (whichever is being used) along a column to transmissivity or hydraulic conductivity along a
row. Set to 1.0 for isotropic conditions. This is a single variable with one value per layer. Do not read a variable for each
layerthat is, include only one array control line for the entire variable.
Sf1is the primary storage coefficient. Read only if one or more transient stress periods are specified in the Discretization
File. For LAYCON equal to 1, Sf1 will always be specific yield, whereas for LAYCON equal to 2 or 3, Sf1 will always
be confined storage coefficient. For LAYCON equal to 0, Sf1 would normally be confined storage coefficient; however,
a LAYCON value of 0 also can be used to simulate water-table conditions where drawdowns everywhere are expected to
remain a small fraction of the saturated thickness, and where there is no layer above, or flow from above is negligible. In
this case, specific yield values would be entered for Sf1.
Tranis the transmissivity along rows. Tran is multiplied by TRPY to obtain transmissivity along columns. Read only for
layers where LAYCON is 0 or 2.
HYis the hydraulic conductivity along rows. HY is multiplied by TRPY to obtain hydraulic conductivity along columns. Read
only for layers where LAYCON is 1 or 3.
Vcontis the vertical hydraulic conductivity divided by the thickness from a layer to the layer below (also called leakance).
The value for a cell is the hydraulic conductivity divided by thickness for the material between the node in that cell and
the node in the cell below. Because there is not a layer beneath the bottom layer, Vcont cannot be specified for the
bottom layer.
Sf2is the secondary storage coefficient. Read only for layers where LAYCON is 2 or 3 and only if there are one or more
transient stress periods specified in the Discretization File. The secondary storage coefficient is always specific yield.
WETDRYis a combination of the wetting threshold (THRESH) and a flag to indicate which neighboring cells can cause a cell
to become wet. If WETDRY < 0, only the cell below a dry cell can cause the cell to become wet. If WETDRY > 0, the cell
below a dry cell and the four horizontally adjacent cells can cause a cell to become wet. If WETDRY is 0, the cell cannot
be wetted. The absolute value of WETDRY is the wetting threshold. When the sum of BOT and the absolute value of
WETDRY at a dry cell is equaled or exceeded by the head at an adjacent cell, the cell is wetted. Read only if LAYCON is 1
or 3 and IWDFLG is not 0.
Layer-Property Flow Package

Input for the Layer-Property Flow (LPF) Package is read from the file that is type LPF in the Name File. Free format is
used for reading all values. The LPF Package is an alternative to the BCF and HUF Packages. Only one of these packages can be
used during a simulation.
FOR EACH SIMULATION
0. [#Text]
1. ILPFCB HDRY NPLPF [Options]
2. LAYTYP(NLAY)
3. LAYAVG(NLAY)
4. CHANI(NLAY)
5. LAYVKA(NLAY)
6. LAYWET(NLAY)
7. [WETFCT IWETIT IHDWET]
(Include item 7 only if LAYWET indicates at least one wettable layer.)
8. and 9. These items are not included in these instructions because they relate to parameters.
A subset of the following two-dimensional variables is used to describe each layer. All the variables that apply to layer 1 are
read first, followed by layer 2, followed by layer 3, and so forth. A variable not required due to simulation options (for example,
Ss and Sy for a completely steady-state simulation) must be omitted from the input file.
These variables are read by the array-reading utility subroutine, U2DREL.
10. HK(NCOL,NROW)
11. [HANI(NCOL,NROW)]
Include item 11 only if CHANI is less than or equal to 0.
12. VKA(NCOL,NROW)
13. [Ss(NCOL,NROW)]
Include item 13 only if at least one stress period is transient.
14. [Sy(NCOL,NROW)]
Include item 14 only if at least one stress period is transient and LAYTYP is not 0.
15. [VKCB(NCOL,NROW)]
Include item 15 only if LAYCBD (in the Discretization File) is not 0.
16. [WETDRY(NCOL,NROW)]
Include item 16 only if LAYWET is not 0 and LAYTYP is not 0.
Explanation of Variables Read by the LPF Package:

ILPFCBis a flag and a unit number.

If ILPFCB > 0, cell-by-cell flow terms will be written to this unit number when SAVE BUDGET or a non-zero value for
adjacent cells.
If ILPFCB = 0, cell-by-cell flow terms will not be written.
If ILPFCB < 0, cell-by-cell flow for constant-head cells will be written in the listing file when SAVE BUDGET or a non-
Appendix 1 189
NPLPF{this value should be set to zero when parameters are not used.}
Optionsare optional key words that activate options:
STORAGECOEFFICIENT indicates that variable Ss is read as storage coefficient rather than specific storage.
CONSTANTCV indicates that vertical conductance for an unconfined cell is computed from the cell thickness rather than
the saturated thickness.
THICKSTRT indicates that layers having a negative LAYTYP are confined, and their cell thickness for conductance
calculations will be computed as STRT-BOT rather than TOP-BOT, where STRT is the ground-water head at the
beginning of the simulation.
NOCVCORRECTION indicates that vertical conductance is not corrected when the vertical flow correction is applied.
LAYTYPcontains a flag for each layer that specifies the layer type.
0 confined
>0 convertible
<0 convertible unless the THICKSTRT option is in effect. When THICKSTRT is in effect, a negative value of LAYTYP
indicates that the layer is confined, and its saturated thickness will be computed as STRT-BOT, where STRT is the
ground-water head at the beginning of the simulation.
LAYAVGcontains a flag for each layer that defines the method of calculating interblock transmissivity.
0harmonic mean
1logarithmic mean
2arithmetic mean of saturated thickness and logarithmic-mean hydraulic conductivity.
CHANIcontains a value for each layer that is a flag or the horizontal anisotropy. If CHANI 0, then variable HANI defines
horizontal anisotropy. If CHANI is >0, then CHANI is the horizontal anisotropy for the entire layer, and HANI is not
read.
LAYVKAcontains a flag for each layer that indicates whether variable VKA is vertical hydraulic conductivity or the ratio of
horizontal to vertical hydraulic conductivity.
0indicates VKA is vertical hydraulic conductivity
not 0indicates VKA is the ratio of horizontal to vertical hydraulic conductivity, where the horizontal hydraulic
conductivity is specified as HK in item 10.
LAYWETcontains a flag for each layer that indicates whether wetting is active.
0indicates wetting is inactive
not 0indicates wetting is active
WETFCTis a factor that is included in the calculation of the head that is initially established at a cell when the cell is converted
from dry to wet. (See IHDWET.)
IWETITis the iteration interval for attempting to wet cells. Wetting is attempted every IWETIT iteration. If using the PCG
solver (Hill, 1990), this applies to outer iterations, not inner iterations. If IWETIT 0, the value is changed to 1.
If IHDWET is not 0, equation 5-32B in Harbaugh (2005) is used: h = BOT + WETFCT(THRESH)
HKis the hydraulic conductivity along rows. HK is multiplied by horizontal anisotropy (see CHANI and HANI) to obtain
hydraulic conductivity along columns.
HANIis the ratio of hydraulic conductivity along columns to hydraulic conductivity along rows, where HK of item 10 specifies
the hydraulic conductivity along rows. Thus, the hydraulic conductivity along columns is the product of the values in HK
and HANI. Read only if CHANI 0.
VKAis either vertical hydraulic conductivity or the ratio of horizontal to vertical hydraulic conductivity depending on the value
of LAYVKA. If LAYVKA is 0, VKA is vertical hydraulic conductivity. If LAYVKA is not 0, VKA is the ratio of horizontal to
vertical hydraulic conductivity. In this case, HK is divided by VKA to obtain vertical hydraulic conductivity, and values of
VKA typically are 1.0.
Ssis specific storage unless the STORAGECOEFFICIENT option is used. When STORAGECOEFFICIENT is used, Ss is
confined storage coefficient. Read only for a transient simulation (at least one transient stress period).
Syis specific yield. Read only for a transient simulation (at least one transient stress period) and if the layer is convertible
(LAYTYP is not 0).
VKCBis the vertical hydraulic conductivity of a quasi-three-dimensional confining bed below a layer. Read only if there is a
confining bed. Because the bottom layer cannot have a confining bed, VKCB cannot be specified for the bottom layer.
WETDRYis a combination of the wetting threshold and a flag to indicate which neighboring cells can cause a cell to become
wet. If WETDRY < 0, only the cell below a dry cell can cause the cell to become wet. If WETDRY > 0, the cell below a dry
cell and the four horizontally adjacent cells can cause a cell to become wet. If WETDRY is 0, the cell cannot be wetted.
The absolute value of WETDRY is the wetting threshold. When the sum of BOT and the absolute value of WETDRY at
a dry cell is equaled or exceeded by the head at an adjacent cell, the cell is wetted. Read only if LAYTYP is not 0 and
LAYWET is not 0.
Hydrogeologic-Unit Flow Package

Input for the Hydrologic-Unit Flow (HUF) Package is read from the file that is type HUF2 in the Name File. Free format
is used for reading all values. The HUF Package is an alternative to the BCF and LPF Packages. The three packages cannot be
used simultaneously. The HUF Package is documented in Anderman and Hill (2000). Additional capbilities of the HUF Package
are described in Anderman and others (2002) and Anderman and Hill (2003). Users of GSFLOW are encouraged to review these
documents when using the HUF Package. Input instructions are from Anderman and Hill (2003, appendix A).
FOR EACH SIMULATION
0. [#Text]
1. IHUFCB HDRY NHUF NPHUF IOHUF IOHUFHEADS IOHUFFLOWS
2. LTHUF(NLAY)
3. LAYWT(NLAY)
4. [WETFCT IWETIT IHDWET]
Include item 4 only if LAYWT indicates at least one wettable layer.
5. WETDRY(NCOL,NROW)
Repeat Item 5 for each layer for which LAYWET is not 0. Arrays are read by the array-reading utility module, U2DREL.
6. HGUNAM
7. TOP(NCOL,NROW)
8. THCK(NCOL,NROW)
Items 6-8 are repeated for each hydrogeologic unit (NHUF times). Items 7 and 8 are read by the array-reading utility
subroutine, U2DREL.
9. HGUNAM HGUHANI HGUVANI
Repeat Item 9 for each hydrogeologic unit. If HGUNAM is set to ALL, HGUHANI and HGUVANI are set for all
hydrogeologic units and only one Item 9 is necessary. Otherwise, HGUNAM must correspond to one of the names
defined in Item 6, and there must be NHUF repetitions of Item 9. The repetitions can be in any order.
10. PARNAM PARTYP Parval NCLU
11. HGUNAM Mltarr Zonarr IZ
Each Item 11 record is called a parameter cluster. Repeat Item 11 NCLU times.
Repeat Items 10-11 for each parameter to be defined (NPHUF times).
Appendix 1 191
12. [PRINT HGUNAM PRINTCODE PRINTFLAGS]

Item 12 is optional and is included only for hydrogeologic units for which printing is desired. Item 12 must start with the
word PRINT. If HGUNAM is set to ALL, PRINTCODE and PRINTFLAGS are set for all hydrogeologic units, and only
one Item 12 is necessary. Otherwise, HGUNAM must correspond to one of the names defined in Item 6.
Explanation of Variables Read by the HUF Package:

IHUFCBis a flag and a unit number.

If IHUFCB > 0, cell-by-cell flow terms will be written to this unit number when SAVE BUDGET or a non-zero value for
adjacent cells.
If IHUFCB = 0, cell-by-cell flow terms will not be written.
If IHUFCB < 0, cell-by-cell flow for constant-head cells will be written in the listing file when SAVE BUDGET or a non-
NHUF is the number of hydrogeologic units defined using the HUF package.
NPHUF{this value should be set to zero when parameters are not used.
IOHUFis a flag and a unit number.
0 interpolated heads will not be written.
>0 calculated heads will be interpolated and written on unit IOHUF for each hydrogeologic unit using the format
defined in the output-control file.
IOHUFHEADSis a flag and unit number.

0 interpolated heads will not be written.
>0 calculated heads will be interpolated and written on unit IOHUFHEADS for each hydrogeologic unit using the
format defined in the output-control file.
IOHUFFLOWSis a flag and unit number.
0 interpolated flows will not be written.
>0 calculated cell-by-cell flows will be interpolated and written on unit IOHUFFLOWS for each hydrogeologic unit
using the format defined in the output-control file.
LTHUFis a flag specifying the layer type. Read one value for each layer; each element holds the code for the respective layer.
There is a limit of 200 layers. Use as many records as needed to enter a value for each layer.
0 indicates a confined layer
not 0 indicates a convertible layer
LAYWTis a flag that indicates if wetting is active. Read one value per layer.
0 indicates wetting is inactive
1 indicates wetting is active
WETFCTis a factor that is included in the calculation of the head that is initially established at a cell when the cell is converted
from dry to wet. (See IHDWET.)
IWETITis the iteration interval for attempting to wet cells. Wetting is attempted every IWETIT iteration. If using the PCG
solver (Hill, 1990), this applies to outer iterations, not inner iterations. If IWETIT is 0, the value is changed to 1.
If IHDWET is not 0, equation 5-32B in Harbaugh (2005) is used: h = BOT + WETFCT(THRESH).
WETDRYis a combination of the wetting threshold and a flag to indicate which neighboring cells can cause a cell to become
wet. If WETDRY < 0, only the cell below a dry cell can cause the cell to become wet. If WETDRY > 0, the cell below a dry
cell and the four horizontally adjacent cells can cause a cell to become wet. If WETDRY is 0, the cell cannot be wetted.
The absolute value of WETDRY is the wetting threshold. When the sum of BOT and the absolute value of WETDRY at
a dry cell is equaled or exceeded by the head at an adjacent cell, the cell is wetted. Read only if LTHUF is not 0 and
LAYWT is 1.
HGUNAM is the name of the hydrogeologic unit. This name can consist of up to 10 characters and is not case sensitive.
TOPis the elevation of the top of the hydrogeologic unit.
THCK is the thickness of the hydrogeologic unit.
HGUHANI is a flag and a horizontal anisotropy value for a hydrogeologic unit. Horizontal anisotropy is the ratio of hydraulic
conductivity along columns to hydraulic conductivity along rows. Read one value for each hydrogeologic unit unless
HGUNAM is set to ALL.
0 indicates that horizontal anisotropy will be defined using a HANI parameter.
> 0 HGUHANI is the horizontal anisotropy of the entire hydrogeologic unit.
HGUVANI is a flag that indicates whether array VK is vertical hydraulic conductivity or the ratio of horizontal to vertical
hydraulic conductivity. Read only one value for each hydrogeologic unit unless HGUNAM is set to ALL.
0 indicates VK is hydraulic conductivity (VK parameter must be used).
> 0 indicates VK is the ratio of horizontal to vertical hydraulic conductivity and HGUVANI is the vertical anisotropy of
the entire hydrogeologic unit. Value is ignored if a VANI parameter is defined for the corresponding hydrogeologic
unit.
PARNAM is the name of a parameter to be defined. This name can consist of up to 10 characters and is not case sensitive.
PARTYP is the type of parameter to be defined. For the HUF Package, the allowed parameter types are:
HKdefines variable HK, horizontal hydraulic conductivity.
HANIdefines variable HANI, horizontal anisotropy.
VKdefines variable VK, vertical hydraulic conductivity, for units for which HGUVANI is set to zero.
VANIdefines variable VANI, vertical anisotropy, for units for which HGUVANI is set to greater than zero.
SS defines variable Ss, the specific storage.
SY defines variable Sy, the specific yield.
SYTP when all model layers are confined, defines the storage coefficient for the top active cell at each row, column
location. The value specified is not multiplied by model layer thickness. When SYTP is specified as the parameter
type, HGUNAM of item 11 needs to be set to SYTP.
Parval is the initial value of the parameter; however, this value can be replaced by a value specified in the Sensitivity
Process input file.
NCLU is the number of clusters required to define the parameter. Each Item-11 record is a cluster (variables HGUNAM,
Mltarr, Zonarr, and IZ).
HGUNAM is the hydrogeologic unit to which the parameter applies. When PARTYP = SYTP, HGUNAM must be set to SYTP.
Mltarr is the name of the multiplier array to be used to define array values that are associated with a parameter. The name
NONE means that there is no multiplier array, and the array values will be set equal to Parval.
Appendix 1 193
Zonarr is the name of the zone array to be used to define array elements that are associated with a parameter. The name
ALL means that there is no zone array and that all elements in the hydrogeologic unit are part of the parameter.
IZ is up to 10 zone numbers (separated by spaces) that define the array elements that are associated with a parameter. The
first zero or non-numeric value terminates the list. These values are not used if Zonarr is specified as ALL.
PRINTCODE determines the format for printing the values of the hydraulic-property arrays for the hydrogeologic unit as
defined by parameters. The print codes are the same as those used in an array control record (Harbaugh and others, 2000,
p. 87).
PRINTFLAGS determines the hydraulic-property arrays to be printed and must be set to ALL or any of the following: HK,
HANI, VK, SS, or SY. Arrays will be printed only for those properties that are listed. When VK is specified, the
property printed depends on the setting of HGUVANI.
Model-Layer Variable-Direction Horizontal Anisotropy (LVDA) Capability

Input for the Model-Layer Variable-Direction Horizontal Anisotropy (LVDA) Capability within the HUF Package is read
from the file that has type LVDA in the Name File and must be used with the HUF Package. Free format is used for reading all
values. This file defines the horizontal anisotropy direction for each cell of the finite-difference grid. This additional capability
of the HUF Package is described in Anderman and others (2002). Users of GSFLOW are encouraged to review this document
when using the HUF Package with the LVDA Capability. Although the HUF input file is not changed when using the LVDA
capability, the definition of two parameter types (HK and HANI) does change. The new definitions of the two parameters are
described in Explanation of Variables Read by the LVDA Capability.
FOR EACH SIMULATION
0. [#Text]
1. NPLVDA
2. PARNAM PARTYP Parval NCLU
3. Layer Mltarr Zonarr IZWT
Each Item 3 record is called a parameter cluster. Repeat Item 3 NCLU times. Repeat Items 2 and 3 for each parameter to
be defined (that is, NPLVDA times).
Explanation of Variables Read by the LVDA Capability
NPLVDA is the number of LVDA parameters.
PARNAM is the name of a parameter to be defined. This name can consist of up to 10 characters and is not case sensitive.
PARTYP is the type of parameter to be defined. For the LVDA Capability, the only allowed parameter type is:
LVDA defines the angle between the grid axis and the principal direction of horizontal hydraulic conductivity. Angle
is positive in a clockwise direction with the positive x direction being zero. The angle can vary between 90 and 90
degrees. Hydraulic conductivity along the principal axis (Kmax of figure 2 in Anderman and others, 2002) is defined
in the HUF Package input file by an HK parameter; the hydraulic conductivity along the minor axis (Kmin of figure 2
in Anderman and others, 2002) axis is defined by a HANI parameter in conjunction with the HK parameter. Using the
LVDA Capability, HANI values need to be 1.
Parval is the initial value of the parameter; however, this value can be replaced by a value specified in the Sensitivity
Process input file.
NCLU is the number of clusters required to define the parameter. Each Item-3 record is a cluster (variables Layer, Mltarr,
Zonarr, and IZ).
Layer is the layer to which the direction applies.
Mltarr is the name of the multiplier array to be used to define array values that are associated with a parameter. The name
NONE means that there is no multiplier array, and the array values will be set equal to Parval.
Zonarr is the name of the zone array to be used to define array elements that are associated with a parameter. The name
ALL means that there is no zone array and that all elements in the hydrogeologic unit are part of the parameter.
IZWT is up to 10 zone numbers (separated by spaces) that define the array elements that are associated with a parameter. The
first zero or non-numeric value terminates the list. These values are not used if Zonarr is specified as ALL.
Horizontal Flow Barrier Package

Input for the Horizontal Flow Barrier (HFB) Package is read from the file that has file type HFB6 in the Name File. All
variables are read in free format.
FOR EACH SIMULATION
0. [#Text]
1. NPHFB MXFB NHFBNP
2. and 3. These items are not included in these instructions because they relate to parameters.
4. Layer IROW1 ICOL1 IROW2 ICOL2 Hydchr

NHFBNP repetitions of item 2 are read.
5. NACTHFB
NPHFB {this value should be set to zero when parameters are not used.}
MXFB {this value should be set to zero when parameters are not used.}
NHFBNPis the number of HFB barriers.
Layeris the number of the model layer in which the horizontal flow barrier is located.
IROW1is the row number of the cell on one side of the horizontal flow barrier.
ICOL1is the column number of the cell on one side of the horizontal flow barrier.
IROW2is the row number of the cell on the other side of the horizontal flow barrier.
ICOL2is the column number of the cell on the other side of the horizontal flow barrier.
Hydchris the hydraulic characteristic of the horizontal flow barrier. The hydraulic characteristic is the barrier hydraulic
conductivity divided by the width of the horizontal flow barrier.
NACTHFB{this value must be zero because parameters are not supported in GSFLOW.}
Appendix 1 195
Well Package
Input to the Well (WEL) Package is read from the file that has type WEL in the Name File. Optional variables are shown in
brackets. All variables are free format if the option FREE is specified in the Basic Package input file; otherwise, the non-optional
variables have 10-character fields and the optional variables are free format.
FOR EACH SIMULATION
0. [#Text]
1. MXACTW IWELCB [Option]
2. ITMP NP
3. Layer Row Column Q [xyz]
ITMP repetitions of item 3 are read by subroutine ULSTRD if ITMP > 0. (SFAC of the ULSTRD utility subroutine
applies to Q.) Item 3 is not read if ITMP is negative or zero.
Explanation of Variables Read by the WEL Package:

MXACTWis the maximum number of wells in use during any stress period.
IWELCBis a flag and a unit number.

If IWELCB > 0, cell-by-cell flow terms will be written to this unit number when SAVE BUDGET or a non-zero value for
ICBCFL is specified in Output Control.
If IWELCB = 0, cell-by-cell flow terms will not be written.
If IWELCB <0, well recharge for each well will be written to the listing file when SAVE BUDGET or a non-zero value for
AUXILIARY abc or AUX abcdefines an auxiliary variable, named abc, which will be read for each well as part of
item 3. Up to five variables can be specified, each of which must be preceded by AUXILIARY or AUX. These variables
will not be used by the Ground-Water Flow Process, but they will be available for use by other processes. The auxiliary
variable values will be read after the Q variable.
CBCALLOCATE or CBCindicates that memory should be allocated to store cell-by-cell flow for each well in order to
make these flows available for use in other packages.
If ITMP < 0, well data from the last stress period will be reused.
If ITMP 0, ITMP will be the number of wells read for the current stress period.
NP{ this value must be zero because parameters are not supported in GSFLOW.}
Layeris the layer number of the finite-difference cell that contains the well.
Rowis the row number of the finite-difference cell that contains the well.
Columnis the column number of the finite-difference cell that contains the well.
Qis the volumetric recharge rate. A positive value indicates recharge and a negative value indicates discharge (pumping).
[xyz]represents the values of the auxiliary variables for a well that have been defined in item 1. The values of auxiliary
variables must be present in each repetition of item 3 if they are defined in item 1. The values must be specified in the
order used to define the variables in item 1.
Multi-Node Wells Package

Input for the Multi-Node, Drawdown-Limited Well (MNW) Package is initiated by specifying MNW1 in the Name File.
Data are read from MNW Package input files as 256-character-wide, alphanumeric lines to facilitate the addition of comments
within the model input files and the use of keys to identify input variables. All integer, real, and character variables are read
from the alphanumeric lines. The lines are initially read by the subroutine NCREAD. Lines that begin with a # sign in the first
column are treated as comments, are not passed to any other routines, and are discarded. Once NCREAD has acquired a valid
data line, the line is checked for a !! sign that designates the beginning of any in-line comments. If a !! sign is detected, the
!! sign and all text to the right of the !! sign are removed from the line before passing it to any other routines.
Alphanumeric strings are used in the MNW Package to identify variables (keys) and make logical decisions (flags).
Specification of these keys and flags is case insensitive because all letters are capitalized before performing any logical tests.
Keys precede the variable to be read, which is acquired by identifying the key and reading the first value that follows the key.
Logical decisions are based on the presence (true) or absence (false) of a flag. In this report, bold, uppercase letters are used
to denote the part of the key that is tested. Key:data pairs that are not delimited by parentheses are mandatory and must be
included, and Key:data pairs that are delimited within parentheses are optional because default values are used if they are not
specified by the user.
The MNW Package is documented in Halford and Hanson (2002) and users of GSFLOW are encouraged to review that
document when using the MNW Package.
FOR EACH SIMULATION
1. MXMNW IWL2CB IWELPT REFerence SP: kspref

Item 1 is required.
2. LOSSTYPE (PLossMNW)
Item 2 is required.
3a. KEY:DATA FILE:filename WEL1:iunw1
3b. KEY:DATA FILE:filename BYNODE:iunby ALLTIME
3c. KEY:DATA FILE:filename QSUM:iunqs ALLTIME
Items 3a, 3b, and 3c are optional.
4. ITMP ADD
5. Layer Row Column Qdes (MN or MULTI) QWval Rw Skin Hlim Href (DD)
Iqwgrp Cp: C (QCUT or Q-%CUT: Qfrcmn, Qfrcmx) DEFAULT SITE: MNWsite
The first four values in item 5 for the variables Layer, Row, Column, and Qdes are read initially as a free format.
If this fails, the four values are read as fixed format entries from the first 40 columns using the format I10, I10, I10,
and F10.0. In all instances, these values must be specified. The following eight values for the remaining variables are
optional; space-delimited or comma-delimited entries but must be entered in the sequence specified for item 5. The
alphanumeric flags MN and DD can appear anywhere between columns 41 and 256, inclusive. Input item 5 normally
consists of one line for each well cell defined or modified. If ITMP is 0 or less, item 5 is not read and should not be
specified.
Explanation of Variables Read by the MNW Package:

MXMNWis an integer variable equal to the maximum number of well cells to be defined.
IWL2CBis a flag and an integer unit number.

If IWL2CB > 0, it is the unit number on which cell-by-cell flow terms will be recorded whenever ICBCFL is set.
If IWL2CB = 0, cell-by-cell flow terms will not be printed or recorded.
If IWL2CB < 0, well recharge, water levels in the well and cell, drawdown in the well, and the flow-rate-weighted water-
quality value of the Iqwgrp will be printed whenever ICBCFL is set.
Appendix 1 197
IWELPTis a flag. If IWELPT is not equal to 0, no well information will be printed.
REFerence SP: ksprefis the set of water levels in the HNEW matrix at the beginning of the stress period kspref that
will be used as default reference values for calculating drawdown. kspref defaults to 1 if it is not specified by the user.
LOSSTYPEis a flag to determine the user-specified model for well loss.

If LOSSTYPE is set to SKIN, head loss is defined with a coefficient representing skin effects (Halford and Hanson, 2002,
p. 8). Model is linear.
If LOSSTYPE is set to NONLINEAR, head loss is defined with a coefficient representing skin effects and a non-linear
well loss coefficient (Halford and Hanson, 2002, p. 8).
PLossMNWis variable P in equation 2 of Halford and Hanson (2002), and is the power of the nonlinear discharge component
of well loss, which usually varies between 1.5 and 3.5 (Rorabaugh, 1953).
FILE:filenameis the name of an auxiliary output file.
WEL1:iunw1is a unit number. The unit number will be associated with the auxiliary output file defined by the WEL1
filename. Output is a WEL1 input file with the flow rates specified at the end of each stress period.
BYNODE:iunbyis a unit number. The unit number will be associated with the auxiliary output file defined by the BYNODE
filename. Output is flow rate at each well node.
QSUM:iunqsis a unit number. The unit number will be associated with the auxiliary output file defined by the QSUM
filename. Output is total flow rate from each multi-node well.
ALLTIMEis a flag that indicates flow rates should be written to BYNODE or QSUM files at every time step regardless of the
settings in the output control (OC) file.
ITMPis an integer flag and a counter.

If ITMP < 0, wells from the previous stress period will be reused and input from item 4 will not be read.
If ITMP = 0, no wells will be simulated and input from item 4 will not be read.
If ITMP > 0, is the number of items of drawdown-limited well data that will be read for the current stress period. If the
key ADD is not detected on item 4, the maximum number of drawdown-limited wells for the current stress period will
be ITMP. If the key ADD is detected on item 4, ITMP wells will be added to the existing list of drawdown-limited
wells.
ADDis a flag that indicates whether or not the well cells read for the current stress period will augment or replace the well cells
that were previously defined.
Layeris the layer number of the finite-difference cell that contains the well.
Rowis the row number of the finite-difference cell that contains the well.
Columnis the column number of the finite-difference cell that contains the well.
Qdesis the desired volumetric pumping or recharge rate (variable type is real). A positive value indicates recharge and a
negative value indicates discharge. The actual volumetric recharge rate will range from 0 to Qdes and is not allowed to
switch directions between discharge and recharge conditions during any stress period.
(MN)is a flag that indicates this entry is part of a multi-node well. The flag MN is not included on the first entry of a multi-node
well and is exclusive of the flag MULTI.
(MULTI)is a flag that indicates this entry is the end of a multi-node well and all intervening nodes between this entry and the
previous MULTI flag are part of a multi-node well. Intervening nodes will be assigned the same cell-to-well conductance
that was specified in this entry. The flag MULTI is not included on the first entry of a multi-node well and is exclusive of
the flag MN.
Qwvalis a real variable equal to the water-quality value that is to be flow-rate averaged amongst wells in the same Iqwgrp.
Negative water-quality values and positive flow terms are not averaged. Water-quality values can be re-specified for each
stress period.
Rwis a real-valued flag and a variable used to define the cell-to-well conductance.
If Rw > 0, the variable represents the radius of the well and the cell-to-well conductance is calculated with equation 5 in
Halford and Hanson (2002).
If Rw = 0, the head in the cell is assumed to be equivalent to the head in the well and the cell-to-well conductance is set
to 1,000 times the transmissivity of the cell. The cell is not allowed to be part of a multi-node well.
If Rw < 0, the absolute value of the variable is the cell-to-well conductance.
Skinis a real variable that defines the friction losses to the well owing to the screen and to formation damage. The variable is
either a skin or the coefficient B depending on the LOSSTYPE, and is used in equation 5 in Halford and Hanson (2002)
when Rw > 0.
Hlimis a real variable equal to the limiting water level, which is a minimum for discharging wells and a maximum for
recharging wells. If the flag DD is set, the value of Hlim read is a drawdown from the reference elevation. For Qdes < 0,
Hlim = Href Hlim and for Qdes > 0, Hlim = Href + Hlim.
Hrefis a real variable equal to the reference elevation. If the value of Href read is greater than the maximum water level
from the HNEW matrix at the beginning of the stress period kspref, Href is set to the simulated water level at the
location of the drawdown-limited well.
DDis a flag that indicates the value of Hlim read is a drawdown or build-up from the reference elevation.
Iqwgrpis an integer variable equal to the water-quality identifier. Flow-rate averaged water-quality values are reported for
each group of wells with the same Iqwgrp and Qwval entries that are not negative.
Cp: C is a real variable that is a coefficient for nonlinear head losses (see equation 2 in Halford and Hanson, 2002). The
variable is used only when the LOSSTYPE is NONLINEAR. Default value is 0 if not specified.
QCUTis a flag that indicates pumping limits will be specified as a rate (units of length cubed per time).
Q-%CUTis a flag that indicates pumping limits will be specified as a percentage of the specified rate.
Qfrcmnis a real variable equal to the minimum pumping rate that a well must exceed to remain active.
Qfrcmxis a real variable equal to the minimum potential pumping rate that must be exceeded to reactivate a well.
DEFAULTis a flag that sets this entry of Qfrcmn and Qfrcmx as the new default values.
SITE: MNWsiteis an optional label for identifying wells. An individual file of time, discharge, water level in well,
concentration, net-inflow, net-outflow, and node-by-node flows will be written for each well with a unique MNW site
label. Individual well files are tab delimited. Only one label should be applied to a multi-node well.
Appendix 1 199
General-Head Boundary Package

Input to the General-Head Boundary (GHB) Package is read from the file that has file type GHB in the Name File. Optional
variables are shown in brackets. All variables are free format if the option FREE is specified in the Basic Package input file;
otherwise, the non-optional variables have 10-character fields and the optional variables are free format.
FOR EACH SIMULATION
0. [#Text]
1. MXACTB IGHBCB [Option]
2. ITMP NP
3. Layer Row Column Bhead Cond [xyz]
ITMP repetitions of item 3 are read by subroutine ULSTRD if ITMP > 0. (SFAC of the ULSTRD utility subroutine
applies to Cond.) Item 3 is not read if ITMP is negative or 0.
Explanation of Variables Read by the GHB Package:

MXACTBis the maximum number of general-head boundary cells in use during any stress period.
IGHBCBis a flag and a unit number.

If IGHBCB > 0, cell-by-cell flow terms will be written to the unit number when SAVE BUDGET or a non-zero value for
If IGHBCB = 0, cell-by-cell flow terms will not be written.
If IGHBCB < 0, boundary leakage for each GHB cell will be written to the listing file when SAVE BUDGET or a non-zero
value for ICBCFL is specified in Output Control.
AUXILIARY abc or AUX abc defines an auxiliary variable, named abc, which will be read for each general-head
boundary as part of item 3. Up to five variables can be specified, each of which must be preceded by AUXILIARY or
AUX. These variables will not be used by the Ground-Water Flow Process, but they will be available for use by other
processes. The auxiliary variable values will be read after the Cond variable.
CBCALLOCATE or CBC indicates that memory should be allocated to store cell-by-cell flow for each general-head
boundary in order to make these flows available for use in other packages.
If ITMP < 0, general-head boundary data from the preceding stress period will be reused.
If ITMP 0, ITMP is the number of general-head boundaries read for the current stress period.
NP{this value must be zero because parameters are not supported in GSFLOW.}
Layeris the layer number of the cell affected by the head-dependent boundary.
Rowis the row number of the cell affected by the head-dependent boundary.
Columnis the column number of the cell affected by the head-dependent boundary.
Bheadis the boundary head.
Condis the hydraulic conductance of the interface between the aquifer cell and the boundary.
[xyz]represents the values of the auxiliary variables for a boundary that have been defined in item 1. The values of auxiliary
variables must be present in each repetition of item 3 if they are defined in item 1. The values must be specified in the
order used to define the variables in item 1.
Flow and Head Boundary Package

Input for the Flow and Head Boundary (FHB) Package is initiated by specifying FHB in the Name File. All input is free
format, which requires each of the numbered data groups to start on a new input item. More than one item can be used for
any data group and numbers within data groups must be separated by at least one space or a comma. Integer data types cannot
include a decimal point. Blank spaces are not treated as zeros.
The FHB Package is documented in Leake and Lilly (1997) and users of GSFLOW are encouraged to review that document
when using the FHB Package
FOR EACH SIMULATION
1. NBDTIM NFLW NHED IFHBSS IFHBCB NFHBX1 NFHBX2

Omit item 2 if NFHBX1 = 0. Input item 2 consists of one line for each of NFHBX1 auxiliary variables.
2. VarName Weight
Omit item 3 if NFHBX2 = 0. Input item 3 consists of one line for each of NFHBX2 auxiliary variables.
3. VarName Weight
Items 4a and 4b are required for all simulations. Include NBDTIM times in item 4b.
4a. IFHBUN CNSTM IFHBPT
4b. BDTIM(NBDTIM)
Omit items 5a and 5b if NFLW = 0. Input item 5b consists of one set of numbers for each of NFLW cells. Each set of
numbers includes layer, row, and column indices, an integer auxiliary variable, and NBDTIM values of specified flow.
5b. Layer Row Column IAUX FLWRAT(NBDTIM)
Omit items 6a and 6b if NFHBX1 = 0 or if NFLW = 0. Include one set of items 6a and 6b for each of NFHBX1 auxiliary
variables. Input item 6b consists of one set of numbers for each of NFLW cells. Each set includes NBDTIM values of
the variable.
6b. AuxVar(NBDTIM)
Omit items 7a and 7b if NHED = 0. Input item 7b consists of one set of numbers for each of NFLW cells. Each set of
numbers includes layer, row, and column indices, an integer auxiliary variable, and NBDTIM values of specified head.
7b. Layer Row Column IAUX SBHED(NBDTIM)
Omit items 8a and 8b if NFHBX2 = 0 or if NHED = 0. Include one set of items 8a and 8b for each of NFHBX2 auxiliary
variables. Input item 8b consists of one set of numbers for each of NHED cells. Each set includes NBDTIM values of
the variable.
8b. AuxVar(NBDTIM)
Explanation of Variables Read by the FHB Package:

NBDTIMis an integer variable equal to the number of times at which flow and head will be specified for all selected cells.
If NBDTIM=1, specified flow and head values will remain constant for the entire simulation.
If NBDTIM>1, specified flow and head values will be computed for each time step using linear interpolation.
NFLWis an integer variable equal to the number of cells at which flows will be specified.
NHEDis an integer variable equal to the number of cells at which head will be specified.
IFHBSSis an integer variable equal to the FHB steady-state option flag. If the simulation is transient, the flag is read but not
used. For steady-state simulations, the flag controls how specified-flow, specified-head, and auxiliary-variable values will
be computed for each steady-state solution.
Appendix 1 201
If IFHBSS = 0, values of flow, head, and auxiliary variables will be taken at the starting time of the simulation. This
results in use of the first value in arrays FLWRAT, SBHED, and AuxVar for each respective boundary cell.
If IFHBSS 0, values of flow, head and auxiliary variables will be interpolated in the same way that values are
computed for transient simulations.
IFHBCBis a flag and an integer unit number.
If IFHBCB > 0, it is the unit number on which cell-by-cell flow terms will be recorded whenever ICBCFL is set (see
Harbaugh, 2005, p. 8-15).
If IFHBCB 0, cell-by-cell flow terms will not be recorded.
NFHBX1is an integer variable equal to the number of auxiliary variables whose values will be computed for each time step for
each specified-flow cell.
NFHBX2is an integer variable equal to the number of auxiliary variables whose values will be computed for each time step for
each specified-head cell.
VarNameis the name of an auxiliary variable (a character variable). Name can include up to 16 characters with no embedded
blank characters.
Weightis a real variable equal to the time-weighting factor for an auxiliary variable specifying the fraction of each time step
at which the value of the variable will be interpolated. Value must be in the range from 0.0 to 1.0.
IFHBUNis the unit number on which data lists will be read. The same or different unit numbers can be used to read lists in
items 4b, 5b, 6b, 7b, and 8b.
CNSTMis a constant multiplier (a real variable) for data list BDTIM (item 4b), FLWRAT (part of item 5b), SBHED (part of item
7b), and auxiliary variables in items 6b and 8b.
IFHBPTis an integer flag for printing values of data lists in items 4b, 5b, 6b, 7b, and 8b.
If IFHBPT >0, data list read at the beginning of the simulation will be printed
If IFHBCB 0, data list read at the beginning of the simulation will not be printed.
BDTIMis simulation time (a real variable) at which values of specified flow and (or) values of specified head will be read.
NBDTIM values are required.
Layeris the layer index of specified-flow cell (item 5b) or specified-head cell (item 7b).
Rowis the row index of specified-flow cell (item 5b) or specified-head cell (item 7b).
Columnis the column index of specified-flow cell (item 5b) or specified-head cell (item 7b).
IAUXis an integer auxiliary variable associated with each specified-flow and specified-head boundary cell.
FLWRATis a volumetric rate of flow (a real variable) at specified-flow cells. A list of NBDTIM values must be specified for
each of NFLW specified-flow cells.
AuxVaris the value of real auxiliary variable at specified-flow and specified-head cells. A list of NBDTIM values must be
specified for each of NFLW specified-flow cells and for each of NHED specified-head cells.
SBHEDis an array of real values containing NBDTIM values of the head for each specified-head cell.
Streamflow-Routing Package
Input to the Streamflow-Routing (SFR7) Package is read from the file that has file type SFR in the Name File. The user can
optionally specify that stream gages and monitoring stations along a stream channel are to be written using the Gage Package by
including GAGE in the Name File (see Gage Package section). The modifications in SFR2 do not require any changes to the
data input for SFR1.
The modification of SFR2 to simulate unsaturated flow relies on the specific yield values as specified in the Layer Property
Flow (LPF) Package, the Hydrogeologic-Unit Flow (HUF) Package, or the Block-Centered Flow (BCF) Package. When the
option to use vertical hydraulic conductivity in the LPF Package is specified, the layer(s) that contain cells where unsaturated
flow will be simulated must be specified as convertible. That is, the variable LAYTYP specified in LPF must not be equal to
zero, otherwise the model will print an error and stop execution.
Additional variables that must be specified to define hydraulic properties of the unsaturated zone are all included within the
SFR2 input file. All values are entered in free format. Data input for SFR1 works without modification if unsaturated flow is not
simulated.
The Streamflow Routing Package is documented in Prudic and others (2004) and Niswonger and Prudic (2005) and users
of GSFLOW are encouraged to review those documents when using the SFR Package. However, unlike previous versions of the
Streamflow Routing Package (Prudic and others, 2004; Niswonger and Prudic, 2005) input data for SFR2 cannot be specified
using parameters in GSFLOW.
FOR EACH SIMULATION
0. [#Text]
1. NSTRM NSS NSFRPAR NPARSEG CONST DLEAK ISTCB1 ISTCB2 [ISFROPT]
[NSTRAIL] [ISUZN] [NSFRSETS] [IRTFLG]
The first two variables (NSTRM and NSS) are used for dimensioning arrays, and must be equal to the actual number of
stream reaches defined in Item 2 and the number of segments that define the complete stream network, respectively. If
NSTRM is negative, then unsaturated flow is simulated beneath the stream segments; the absolute value of NSTRM is
used to define the number of stream reaches and for dimensioning of arrays.
If BCF is used and unsaturated flow is active, then ISFROPT must equal 3 or 5.
SFR2 differs from the Stream (STR1) Package (Prudic, 1989) because the new package solves for stream depth at the
midpoint of each reach instead of at the beginning of the reach. To solve for depth at the midpoint of each reach, like
SFR1, SFR2 uses Newtons iterative method and consequently, a tolerance (DLEAK) is used for stopping the iterative
process.
ONE LINE FOR EACH STREAM REACH
2. KRCH IRCH JRCH ISEG IREACH RCHLEN [STRTOP] [SLOPE] [STRTHICK]

[STRHC1] [THTS] [THTI] [EPS] [UHC]
Reach information is read in sequential order from upstream to downstream, first by segments and then by reaches.
Segments should be numbered in downstream order, although this is not necessary. However, if the segments are not
numbered in downstream order, the inflows and outflows from each segment will still be computed, but the computed
inflows into a segment from upstream tributary streams having a higher segment number will be from the previous
iteration. Reaches must be listed and read in sequentially because the order determines the connections of inflows and
outflows within a stream segment.
The stream network is assumed to remain fixed geometrically over the duration of a simulation. However, the active part
of the stream network can be made to vary over time by making selected stream segments inactive for selected stress
periods. This would be implemented by setting the streambed hydraulic conductivity, segment inflow, overland runoff,
and direct precipitation to zero for the inactive segments in item 4 for the specific stress periods when they are known
to be inactive or dry.
If the finite-difference cell corresponding to a stream reach is inactive, the program will search for the uppermost
active cell in the vertical column to apply the leakage. If there are no active cells or if the cell is a constant head, no
interaction is allowed and flow in the reach is passed to the next reach.
When STRTOP, SLOPE, STRTHICK, and STRHC1 are specified for each reach, then HCOND1, THICKM1, ELEVUP,
HCOND2, THICKM2, and ELEVDN are not read using items 4b or 4c.
Appendix 1 203
The residual water content for each cell is not specified by the user because it is calculated based on the specified
saturated water content minus the specific yield of the active finite-difference cell corresponding to the stream reach.
The calculation is performed internally to assure continuity between unsaturated and saturated zone storage.
Although unsaturated flow variables will not be used for reaches that are designated as ICALC = 0, 3, and 4 within the
segment information, values must be included for all reaches when ISFROPT = 2 or 3. Dummy values may be used
for reaches that are designated as ICALC = 0, 3, and 4.
3. ITMP IRDFLG IPTFLG
If ITEMP > 0:
4a. NSEG ICALC OUTSEG IUPSEG [IPRIOR] [NSTRPTS] FLOW RUNOFF ETSW
PPTSW [ROUGHCH] [ROUBHBK] [CDPTH] [FDPTH] [AWDTH] [BWDTH]
4b. [HCOND1] [THICKM1] [ELEVUP] [WIDTH1] [DEPTH1] [THTS1] [THTI1]
[EPS1] [UHC1]
These variables are read for each stress period when NSTRM is positive. THICKM1 and ELEVUP are read only for the
first stress period when ICALC is 1 or 2 and ISFROPT is 4 or 5. WIDTH1 is read for all stress periods when ICALC
is 0 or 1 and ISFROPT is 1, but is only read for the first stress period when ISFROPT is >1. WIDTH1 and DEPTH1
are read for all stress periods when ICALC is 0 and are not dependent on the value of ISFROPT.
4c. [HCOND2] [THICKM2] [ELEVDN] [WIDTH2] [DEPTH2] [THTS2] [THTI2] [EPS2]
[UHC2]
The same options apply to these variables as explained for the upstream variables in item 4b.
If ICALC = 2:
4d. XCPT1 XCPT2 XCPT8

ZCPT1 ZCPT2 ZCPT8
If ICALC = 4:
4e. FLOWTAB(1) FLOWTAB(2) FLOWTAB(NSTRPTS)

DPTHTAB(1) DPTHTAB(2) DPTHTAB(NSTRPTS)
WDTHTAB(1) WDTHTAB(2) WDTHTAB(NSTRPTS)
Item 4 must be repeated ITMP times. The data need not be defined in sequential order by stream segment number. If
ITMP 0, then stream segment data for the previous stress period will be reused.
Item 4a will contain 8 to 13 variables; depending on the values of ICALC and IUPSEG. ICALC determines how stream
depth is to be calculated; when ICALC is 1 or 2, depth is calculated using Mannings equation, which, in turn, requires
a channel roughness coefficient (ICALC = 1) or a channel and bank roughness coefficient (ICALC = 2).
Items 4b and 4c may include no input when all are defined by stream reaches in item 2 or they may include as many as 9
variables depending on the value of ICALC in item 4a and(or) ISFROPT in item 1.
A stream segment that receives inflow from upstream segments is allowed to have as many as 10 upstream segments
feeding it, as defined by the respective values of OUTSEG in item 4a.
Stream properties and stresses are assumed constant and uniform within a single stream segment. Additionally, hydraulic
conductivity, streambed thickness, elevation of top of streambed, stream width, and stream depth may vary smoothly
and linearly within a single stream segment. For these variables, data values at the upstream end of the segment are
described in item 4b and data values at the downstream end of the segment are described in item 4c. Values of these
variables for individual reaches of a segment are estimated using linear interpolation. To make any variable the same
throughout the segment, simply specify equal values in items 4b and 4c. The two elevations in items 4b and 4c are
used in conjunction with the total length of the stream segment (calculated from RCHLEN given for each reach in
item 2) to compute the slope of the stream and the elevations for any intermediate reaches. The streambed thickness
is subtracted from the top of streambed elevations to calculate the elevations of the bottom of the streambed (used in
calculations of leakage).
If item 4d is included (for ICALC = 2), it is assumed that the cross-sectional geometry defined by these data is the same
over the entire length of the segment. Similarly, if item 4e is included (for ICALC = 4), it is assumed the tabulated
relation between streamflow and stream depth and width is the same over the entire length of the segment.
If the Lake (LAK3) Package (Merritt and Konikow, 2000) also is implemented, then flow out of the lake into a stream
segment is dependent on the option used to compute stream depth (ICALC = 1, 2, 3, or 4). Constant discharge from
a lake can be simulated no matter what value of ICALC is assigned to the stream segment emanating from the lake by
assigning a positive value to FLOW in item 4a.
If a diversionary flow is large enough to warrant representation in the model, but is discharged into a pipeline, lined
canal, or other structure or system that does not interact with the aquifer and the flow might exceed the available
streamflow, then there is an alternative means to represent it. Instead of specifying a negative value of FLOW, we
suggest representing the withdrawal by a single-reach diversionary stream segment, which would be located in the
same finite-difference cell as the reach from the upstream segment (IUPSEG) from which the diversion is made;
specifying the segments streambed hydraulic conductivity equal to 0 will preclude interaction with the aquifer and
setting OUTSEG = 0 will remove the flow from the system. The diversion will then be subject to the constraints
associated with the value of IPRIOR.
Explanation of Variables Read by the SFR Package

Textis a character variable (up to 199 characters) that starts in column 2. Any characters can be included in Text. The #
character must be in column 1. Lines beginning with # are restricted to the first lines of the file. Text is written to the
Listing File.
ITEM 1 VARIABLES:
NSTRMan integer value equal to the number of stream reaches (finite-difference cells) that are active during the simulation.
The value of NSTRM also represents the number of lines of data to be included in item 2. When NSTRM is specified as a
negative integer, it also is used as a flag to read the additional variables ISFROPT, NSTRAIL, ISUZN, NSFRSETS, and
IRTFLG, or changing data input simulating unsaturated flow beneath streams, and for simulating transient routing of
streamflow. A negative value is automatically reset to a positive integer.
NSSan integer value equal to the number of stream segments (consisting of one or more reaches) that are used to define the
complete stream network.
NSFRPAR{this variable must be zero because parameters are not supported in GSFLOW.}
NPARSEG {this variable must be zero because parameters are not supported in GSFLOW.}
CONSTa real value (or conversion factor) used in calculating stream depth for a stream reach. If stream depth is not calculated
using Mannings equation for any stream segment (that is, ICALC does not equal 1 or 2), then a value of zero can
be entered. If Mannings equation is used, a constant of 1.486 is used for flow units of cubic feet per second, and a
constant of 1.0 is used for units of cubic meters per second. The constant must be multiplied by 86,400 when using time
units of days in the simulation. An explanation of time units used in MODFLOW is given by Harbaugh and others (2000,
p. 10.)
DLEAKa real value equal to the tolerance level of stream depth used in computing leakage between each stream reach and
active finite-difference cell. Value is in units of length. Usually a value of 0.0001 is sufficient when units of feet or
meters are used in model.
ISTCB1an integer value used as a flag for writing stream-aquifer leakage values. If ISTCB1 > 0, it is the unit number to
which unformatted leakage between each stream reach and corresponding finite-difference cell will be saved to a file
whenever the cell-by-cell budget has been specified in Output Control (see Harbaugh and others, 2000, p. 52-55). If
ISTCB1 = 0, leakage values will not be printed or saved. If ISTCB1 < 0, all information on inflows and outflows
from each reach; on stream depth, width, and streambed conductance; and on head difference and gradient across the
streambed will be printed in the main listing file whenever a cell-by-cell budget has been specified in Output Control.
Appendix 1 205
ISTCB2an integer value used as a flag for writing to a separate formatted file all information on inflows and outflows
from each reach; on stream depth, width, and streambed conductance; and on head difference and gradient across the
streambed. If ISTCB2 > 0, then ISTCB2 also represents the unit number to which all information for each stream reach
will be saved to a separate file when a cell-by-cell budget has been specified in Output Control. If ISTCB2 < 0, it is the
unit number to which unformatted streamflow out of each reach will be saved to a file whenever the cell-by-cell budget
has been specified in Output Control.
If NSTRM <0:
ISFROPTis an integer value that defines the input structure:
ISFROPT = 1, no vertical unsaturated flow beneath streams. Streambed elevation, stream slope, streambed thickness,
and streambed hydraulic conductivity are read for each reach only once at the beginning of the simulation.
ISFROPT = 2, streambed and unsaturated-zone properties are read for each reach only once at the beginning of
the simulation except saturated vertical hydraulic conductivity for the unsaturated zone is the same as the vertical
hydraulic conductivity of the corresponding layer in LPF and is not read in separately.
ISFROPT = 3, same as 2 except saturated vertical hydraulic conductivity for the unsaturated zone is read for each reach.
ISFROPT = 4, streambed and unsaturated-zone properties are read for the beginning and end of each stream segment.
Streambed properties can vary each stress period. Saturated vertical hydraulic conductivity for the unsaturated zone is
the same as the vertical hydraulic conductivity of the corresponding layer in LPF and is not read in separately.
ISFROPT = 5, same as 4 except saturated vertical hydraulic conductivity for the unsaturated zone is read for each
segment at the beginning of the first stress period only.
When ISFROPT >1, read the following variables:
NSTRAILan integer value that is the number of trailing-wave increments used to represent a trailing wave. Trailing waves
are used to represent a decrease in the surface infiltration rate. The value can be increased to improve mass balance in
the unsaturated zone. Values between 20 and 10 work well and result in unsaturated-zone mass balance errors beneath
streams ranging between 0.001 and 0.01 percent. See Smith (1983) for further details.
ISUZNan integer value that is the maximum number of vertical cells used to define the unsaturated zone beneath a stream
reach. If ICALC is 1 for all segments, then ISUZN should be set to 1. ISUZN may affect model run time.
NSFRSETSan integer value that is the maximum number of different sets of trailing waves used to allocate arrays. Arrays
are allocated by multiplying NSTRAIL by NSFRSETS. A value of 30 is sufficient for problems where the stream depth
varies often. NSFRSETS does not affect model run time.
If NSTRM <0:
IRTFLGan integer value that flags whether transient streamflow routing is active. If IRTFLG > 0 then streamflow will be
routed using the kinematic-wave equation.
ITEM 2 VARIABLES:
KRCHan integer value equal to the layer number of the cell containing the stream reach.
IRCHan integer value equal to the row number of the cell containing the stream reach.
JRCHan integer value equal to the column number of the cell containing the stream reach.
ISEGan integer value equal to the number of the stream segment in which this reach is located. Stream segments contain one
or more reaches and are assumed to have uniform or linearly varying characteristics.
IREACHan integer value equal to the sequential number in a stream segment of this reach (where a reach corresponds to a
single cell in the model). Numbering of reaches in a segment begins with 1 for the farthest upstream reach and continues
in downstream order to the last reach of the segment.
RCHLENa real number equal to the length of channel of the stream reach within this finite-difference cell. The length of a
stream reach can exceed the finite-difference cell dimensions because of the meandering nature of many streams. The
length is used to calculate the streambed conductance for this reach. Also, the sum of the lengths of all stream reaches
within a segment is used to calculate the average slope of the channel for the segment and subsequently other values,
such as the elevation of the streambed and stream stage.
STRTOPa real number equal to the top elevation of the streambed. This variable is read when ISFROPT is 1, 2, or 3.
SLOPEa real number equal to the stream slope across the reach. This variable is read when ISFROPT is 1, 2, or 3.
STRTHICKa real number equal to the thickness of the streambed. This variable is read when ISFROPT is 1, 2, or 3.
STRCH1a real number equal to the hydraulic conductivity of the streambed. This variable is read when ISFROPT is 1, 2, or
3.
THTSa real number equal to the saturated volumetric water content in the unsaturated zone. This variable is read when
ISFROPT is 2 or 3.
THTIa real number equal to the initial volumetric water content. THTI must be less than or equal to THTS and greater than or
equal to THTS minus the specific yield defined in either LPF or BCF. This variable is read when ISFROPT is 2 or 3.
EPSa real number equal to the Brooks-Corey exponent used in the relation between water content and hydraulic conductivity
within the unsaturated zone (Brooks and Corey, 1966). This variable is read when ISFROPT is 2 or 3.
UHCa real number equal to the vertical saturated hydraulic conductivity of the unsaturated zone. This variable is necessary
when using BCF, whereas it is optional when using LPF. This variable is read when ISFROPT is 3.
ITEM 3 VARIABLES:
ITMPan integer value for reusing or reading stream segment data that can change each stress period. ITMP must be equal to
the number of stream segments (NSEG) for the first stress period of a simulation. For subsequent stress periods, If ITMP
< NSEG then ITMP segments will be defined for the stress period and the remaining segments will be defined based on
data from the previous stress period.
IRDFLGan integer value for printing input data specified for this stress period. If IRDFLG = 0, input data for this stress
period will be printed. If IRDFLG > 0, then input data for this stress period will not be printed.
IPTFLGan integer value for printing streamflow-routing results during this stress period. If IPTFLG = 0, or whenever the
variable ICBCFL is specified, the results for specified time steps during this stress period will be printed. If IPTFLG >
0, then the results during this stress period will not be printed.
ITEM 4a VARIABLES:
NSEGan integer value of the stream segment for which information is given to identify inflow, outflow, and computation of
stream depth.
ICALCan integer value used to indicate method used to calculate stream depth in this segment:
ICALC 0, stream depth in each reach is specified at the beginning of a stress period and remains unchanged unless
flow at the midpoint of a reach is zero, then depth is set to zero in that reach. No unsaturated flow is allowed.
ICALC > 0, stream depth is calculated and updated each iteration of the MODFLOW solver within a time step.
ICALC = 1, stream depth is calculated using Mannings equation and assuming a wide rectangular channel. Unsaturated
flow is simulated when ISFROPT > 1.
ICALC = 2, stream depth is calculated using Mannings equation and assuming an eight-point channel cross section for
each segment (which allows for the computation of a wetted perimeter and for changing hydraulic conductance of the
streambed in relation to changes in flow). Unsaturated flow is simulated when ISFROPT > 1.
Appendix 1 207
ICALC = 3, stream depth and width are calculated using a power function relating each to streamflow (Q) using
equations 8 and 9 in Prudic and others (2004), where DEPTH(y) = CDPTH QFDPTH and WIDTH(w) = AWDTH
QBWDTH. No unsaturated flow is allowed.
ICALC = 4, stream depth and width are calculated using a table relating streamflow to depth and width. No unsaturated
flow is allowed.
OUTSEGan integer value of the downstream stream segment that receives tributary inflow from the last downstream reach
of this segment. If this segment (identified by NSEG) does not feed (or discharge into) another downstream (tributary)
segment, then enter a value of 0 for this variable. If the segment ends within the modeled grid and OUTSEG = 0, outflow
from the segment is not routed anywhere and is no longer part of the stream network. One may wish to use this if all
flow in the stream gets diverted into a lined canal or into a pipe. If the flow out of this segment discharges into a lake, set
OUTSEG equal to the negative value of the lake identification number (where the minus sign is used as a flag to tell the
model that flow enters a lake rather than a tributary stream segment).
IUPSEGan integer value of the upstream segment from which water is diverted (or withdrawn) to supply inflow to this
stream segment if this segment originates as a diversion from an upstream segment. If the source of a stream segment is
discharge from a lake, set IUPSEG equal to the negative value of the lake identification number (where the minus sign
is used as a flag to tell the model that streamflow into this segment is derived from lake outflow rather than a stream
segment). If this stream segment (identified by NSEG) does not receive inflow as a diversion from an upstream segment,
then set IUPSEG = 0.
IPRIORan integer value that only is specified if IUPSEG > 0 (do not specify a value in this field if IUPSEG = 0 or IUPSEG
< 0). IPRIOR defines the prioritization system for diversion, such as when insufficient water is available to meet all
diversion stipulations, and is used in conjunction with the value of FLOW (specified below).
IPRIOR = 0, then if the specified diversion flow (FLOW) is greater than the flow available in the stream segment from
which the diversion is made, the diversion is reduced to the amount available, which will leave no flow available for
tributary flow into a downstream tributary of segment IUPSEG.
IPRIOR = -1, then if the specified diversion flow (FLOW) is greater than the flow available in the stream segment from
which the diversion is made, no water is diverted from the stream. This approach assumes that once flow in the stream
is sufficiently low, diversions from the stream cease, and is the priority algorithm that originally was programmed
into the STR1 Package (Prudic, 1989).
IPRIOR = -2, then the amount of the diversion is computed as a fraction of the available flow in segment IUPSEG; in
this case, 0.0 FLOW 1.0.
IPRIOR = -3, then a diversion is made only if the streamflow leaving segment IUPSEG exceeds the value of FLOW.
If this occurs, then the quantity of water diverted is the excess flow and the quantity that flows from the last reach of
segment IUPSEG into its downstream tributary (OUTSEG) is equal to FLOW. This represents a flood-control type of
diversion, as described by Danskin and Hanson (2002).
NSTRPTSan integer value specified only when ICALC = 4. It is used to dimension a table relating streamflow with stream
depth and width as specified in item 4e. NSTRPTS must be at least 2 but not more than 50. If the table exceeds 3 50
(for streamflow, stream depth, and width) values, then MAXPTS in the allocation subroutine GWF1SFR1ALP will need
to be increased from 3 50 to 3 (the desired maximum value).
FLOWa real number that is the streamflow (in units of volume per time) entering or leaving the upstream end of a stream
segment (that is, into the first reach).
If the stream is a headwater stream, FLOW defines the total inflow to the first reach of the segment. The value can be any
number 0.
If the stream is a tributary stream, FLOW defines additional specified inflow to or withdrawal from the first reach of the
segment (that is, in addition to the discharge from the upstream segment of which this is a tributary). This additional
flow does not interact with the ground-water system. For example, a positive number might be used to represent direct
outflow into a stream from a sewage treatment plant, whereas a negative number might be used to represent pumpage
directly from a stream into an intake pipe for a municipal water treatment plant.
If the stream is a diversionary stream, and the diversion is from another stream segment, FLOW defines the streamflow
diverted from the last reach of stream segment IUPSEG into the first reach of this segment. The diversion is computed
or adjusted according to the value of IPRIOR.
If the stream is a diversionary stream, and the diversion is from a lake, FLOW defines a fixed rate of discharge diverted
from the lake into the first reach of this stream segment (unless the lake goes dry) and flow from the lake is not
dependent on the value of ICALC. However, if FLOW = 0, then the lake outflow into the first reach of this segment will
be calculated on the basis of lake stage relative to the top of the streambed for the first reach using one of the methods
defined by ICALC.
RUNOFFa real number that is the volumetric rate of the diffuse overland runoff that enters the stream segment (in units of
volume per time). The specified rate is apportioned to each reach of the segment in direct relation to the fraction of the
total length of the stream channel in the segment that is present in each reach.
ETSWa real number that is the volumetric rate per unit area of water removed by evapotranspiration directly from the stream
channel (in units of length per time). ETSW is defined as a positive value.
PPTSWa real number that is the volumetric rate per unit area of water added by precipitation directly on the stream channel (in
units of length per time).
ROUGHCHa real number that is Mannings roughness coefficient for the channel in all reaches in this segment. This variable is
only specified if ICALC = 1 or 2.
ROUGHBKa real number that is Mannings roughness coefficient for the overbank areas in all reaches in this segment. This
variable is only specified if ICALC = 2.
CDPTHa real number that is the coefficient used in the equation: (DEPTH = CDPTH QFDPTH) that relates stream depth in all
reaches in this segment to streamflow. This variable is only specified if ICALC = 3.
FDPTHa real number that is the coefficient used in the equation: (DEPTH = CDPTH QFDPTH) that relates stream depth in all
AWDTHa real number that is the coefficient used in the equation: (WIDTH = AWDTH QBWDTH) that relates stream width in all
BWDTHa real number that is the coefficient used in the equation: (WIDTH = AWDTH QBWDTH) that relates stream width in all
ITEM 4b VARIABLES:
HCOND1a real number that is a factor used to calculate hydraulic conductivity of the streambed at the upstream end of this
segment from the parameter value (in units of length per time).
THICKM1a real number that is the thickness of streambed material at the upstream end of this segment (in units of length).
ELEVUPa real number that is the elevation of the top of the streambed at the upstream end of this segment (in units of length).
WIDTH1a real number that is the average width of the stream channel at the upstream end of this segment (in units of length).
This variable is only specified if ICALC < 1.
DEPTH1a real number that is the average depth of water in the channel at the upstream end of this segment (units of length).
This variable is only specified if ICALC = 0, in which case the stream stage in a reach is assumed to equal the elevation
of the top of the streambed plus the depth of water.
THTS1Saturated volumetric water content in the unsaturated zone beneath the upstream end of this segment. This variable is
read for the first stress period when ICALC is 1 or 2 and ISFROPT is 4 or 5.
THTI1Initial volumetric water content beneath the upstream end of this segment. THTI1 must be less than or equal to THTS
and greater than or equal to the THTS minus the specific yield defined in either LPF or BCF. This variable is read for the
first stress period when ICALC is 1 or 2 and ISFROPT is 4 or 5.
Appendix 1 209
EPS1Brooks-Corey exponent used in the relation between water content and hydraulic conductivity within the unsaturated
zone beneath the upstream end of this segment. This variable is read for the first stress period when ICALC is 1 or 2 and
ISFROPT is 4 or 5.
UHC1Vertical saturated hydraulic conductivity of the unsaturated zone beneath the upstream end of this segment. This
variable is necessary when using BCF, whereas it is optional when using LPF. This variable is read only for the first
stress period when ICALC is 1 or 2 and ISFROPT is 5.
ITEM 4c VARIABLES:
HCOND2a real number that is a factor used to calculate hydraulic conductivity of the streambed at the downstream end of this
segment from the parameter value (in units of length per time).
THICKM2a real number that is the thickness of streambed material at the downstream end of this segment (in units of length).
ELEVDNa real number that is the elevation of the top of the streambed at the downstream end of this segment (in units of
length).
WIDTH2a real number that is the average width of the stream channel at the downstream end of this segment (in units of
length). This variable is only specified if ICALC < 1.
DEPTH2a real number that is the average depth of water in the channel at the downstream end of this segment (units of
length). This variable is only specified if ICALC = 0, in which case the stream stage in a reach is assumed to equal the
elevation of the top of the streambed plus the depth of water.
THTS2Saturated volumetric water content in the unsaturated zone beneath the downstream end of this segment. This variable
is read for the first stress period when ICALC is 1 or 2 and ISFROPT is 4 or 5.
THTI2Initial volumetric water content beneath the downstream end of this segment. THTI1 must be less than or equal to
THTS and greater than or equal to the THTS minus the specific yield defined in either LPF or BCF. This variable is read
for the first stress period when ICALC is 1 or 2 and ISFROPT is 4 or 5.
EPS2Brooks-Corey exponent used in the relation between water content and hydraulic conductivity within the unsaturated
zone beneath the downstream end of this segment. This variable is read for the first stress period when ICALC is 1 or 2
and ISFROPT is 4 or 5.
UHC2Vertical saturated hydraulic conductivity of the unsaturated zone beneath the downstream end of this segment. This
variable is necessary when using BCF, whereas it is optional when using LPF. This variable is read only for the first
stress period when ICALC is 1 or 2 and ISFROPT is 5.
ITEM 4d VARIABLES:
XCPTia real number that is the distance relative to the left bank of the stream channel (when looking downstream) for the
eight points (XCPT1 through XCPT8) used to describe the geometry of this segment of the stream channel. By definition,
location XCPT1 represents the left edge of the channel cross section, and its value should be set equal to 0.0; values
XCPT2 through XCPT8 should be equal to or greater than the previous distance.
ZCPTia real number that is the height relative to the top of the lowest elevation of the streambed (thalweg). One value
(ZCPT1 through ZCPT8) is needed for each of the eight horizontal distances defined by XCPTi. The location of the
thalweg (set equal to 0.0) can be any location from XCPT2 through XCPT7.
ITEM 4e VARIABLES:
FLOWTABa real number that is the streamflow (units of volume per time) related to a given depth and width. One value is
needed for each streamflow that has a corresponding value of depth and width up to the total number of values used to
define the tableFLOWTAB(1) through FLOWTAB(NSTRPTS). NSTRPTS is defined in item 4a.
DPTHTABa real number that is the average depth (units of length) corresponding to a given flow. The number and order of
values, DPTHTAB(1) through DPTHTAB(NSTRPTS) must coincide with FLOWTAB.
WDTHTABa real number that is the stream width (units of length) corresponding to a given flow. The number and order of
values, WDTHTAB(1) through WDTHTAB(NSTRPTS), must coincide with FLOWTAB.
Lake Package
Input to the Lake Package is read from the file that has file type LAK in the Name File. The user can optionally specify that
gages on a lake are to be written using the Gage Package by including GAGE in the Name File (see Gage Package section).
Input for the Lake Package consists of nine separate items, each consisting of one or more lines. These data are used to specify
information about the physical geometry of the lakes, hydraulic properties of the lakebeds, and the degree of hydraulic stress
originating from atmospheric and anthropogenic sources, as well as specifying certain output control variables. Spatial and
temporal units of input data specifications should be consistent with other data input for the MODFLOW run.
All input variables are read using free formats, unless specifically indicated otherwise. In free format, variables are
separated by one or more spaces, or by a comma and, optionally, one or more spaces. It is important to note that, in free format,
blank spaces are not read as zeroes and a blank field cannot be used to set a parameter value to zero. The Lake Package is
documented in Merritt and Konikow (2000) and users of GSFLOW are encouraged to review that document when using the
Lake Package.
FOR EACH SIMULATION
1. NLAKES ILKCB
Sublakes of multiple-lake systems are considered separate lakes for input purposes. The variable NLAKES is used, with
certain internal assumptions and approximations, to dimension arrays for the simulation.
If data are being read using the fixed format mode, then each field should be entered using I10 format.
2. THETA [NSSITR] [SSCNCR]
In the original USGS Lake Package (Merritt and Konikow, 2000), NSSITR and SSCNCR were not used for transient
solutions and were omitted for simulations that did not include a steady-state stress period. Although NSSITR and
SSCNCR are not required input for transient stress periods, changes were made to the Lake Package such that NSSITR
and SSCNCR are used for both transient and steady-state stress periods. If values for NSSITR and SSCNCR are not
input, then default values of 100 and 0.0001 are assumed, respectively. For mixed steady-state/transient runs,
NSSITR and SSCNCR must be included, even if the steady-state stress period is not the first one. If more than one
steady-state stress period is included in the total simulation period, then the initial values of NSSITR and SSCNCR
will apply to all subsequent steady-state stress periods. If there are steady-state stress periods following the first stress
period of a simulations then values of NSSITR and SSCNCR are included in item 9.
If data are being read using the fixed format mode, then the data should be entered using format (F10.4, I10, F10.4).
Appendix 1 211
FOR THE FIRST STRESS PERIOD ONLY
3. STAGES [SSMN] [SSMX]

This data set should consist of one line for each lake, where line 1 includes data for lake 1, and line n includes data for
lake n. There must be exactly NLAKES lines of data.
In the original USGS Lake Package (Merritt and Konikow, 2000), SSMN and SSMX were not needed for a transient run
and were omitted when a simulation did not include a steady-state stress period. However, the Lake Package was
modified such that SSMN and SSMX are used for all stress periods. If a simulation does not include a steady-state stress
period, then SSMN and SSMX are omitted and they are set equal to the lake bottom and top altitudes.
If data are being read using the fixed format mode, then each field should be entered using F10.4 format.
FOR EACH REMAINING STRESS PERIOD
4. ITMP ITMP1 LWRT

ICBCFL 0 also suppresses printout from the lake package.
If data are being read using the fixed format mode, then each field should be entered using I10 format.
If ITMP > 0:
5. LKARR(NCOL, NROW)
An NCOL by NROW array is read for each layer in the grid by module U2DINT.
6. BDLKNC(NCOL, NROW)
An NCOL by NROW array is read for each layer in the grid by module U2DREL.
7. NSLMS
If data are being read using the fixed format mode, then NSLMS should be entered using format I5.
If ITMP > 0 and NSLMS > 0:
8a. IC ISUB(1) ISUB(2) ISUB(IC)

8b. SILLVT(1) SILLVT(IC)
A pair of items (items 8a and 8b) is read for each multiple-lake system, i.e., NSLMS pairs of items. However, IC = 0 will
terminate the input.
If data are being read using the fixed format mode, then each field of item 8a should be entered using I5 format and each
field of item 8b should be entered using F10.4 format.
If ITMP1 0:
9. PRCPLK EVAPLK RNF WTHDRW [SSMN] [SSMX]

The variables SSMN and SSMX are optional. The capability for using mixed steady-state/transient runs with the Lake
Package does not depend on the sequence of steady-state and transient stress periods. If the second or a subsequent
stress period is steady-state, then SSMN and SSMX for those stress periods are defined in item 9a.
Explanation of Variables Read by the Lake Package:

NLAKESis the number of separate lakes.
ILKCBis a flag indicating whether or not to write cell-by-cell flows. If ILKCB >0, flows will be written; otherwise, flows
will not be written. If ILKCB <0 and ICBCFL is not equal to 0, the cell-by-cell flows will be printed in the standard
output file.
THETAis a real variable equal to the semi-implicit (0.5 THETA < 1.0) or implicit (THETA = 1.0) solution approach for
lake stage. If a negative value of THETA (abs(-0.5) THETA abs(-1.0)) is specified, then NSSITR and SSCNCR
are required input regardless of whether or not the simulation includes a steady-state stress period. If THETA is negative,
then an absolute value of THETA will be used for calculating lake stage. If the absolute value of THETA is less than 0.5
or greater than 1.0, then THETA is set equal to 0.5.
NSSITRis the maximum number of iterations for Newtons solution method for lake stages during each MODFLOW iteration
for steady-state aquifer head solution.
SSCNCRis the convergence criterion for equilibrium lake stage solution by Newtons method (L).
STAGESis the initial stage of each lake at the beginning of the run (L).
SSMNis the minimum allowed lake stage for each lake (L).
SSMXis the maximum allowed lake stage for each lake (L).
ITMPoptions are:
ITMP > 0, read lake definition data (items 5 and 6, and, optionally, items 7 through 9).
ITMP = 0, no lake calculations this stress period.
ITMP < 0, use lake definition data from last stress period.
ITMP1options are:
ITMP1 0, read new recharge, evaporation, runoff, and withdrawal data for each lake.
ITMP1 < 0, use recharge, evaporation, runoff, and withdrawal data from last stress period.
LWRTis a flag; if LWRT > 0, suppresses printout from the Lake Package.
LKARRa value is read for every grid cell. If LKARR(I,J,K) = 0, the grid cell is not a lake volume cell. If LKARR(I,J,K) >
0, its value is the identification number of the lake occupying the grid cell. LKARR(I,J,K) must not exceed the value
NLAKES. If it does, or if LKARR(I,J,K) < 0, LKARR(I,J,K) is set to zero.
BDLKNCa value is read for every grid cell. The value is the lakebed leakance (units of per time or 1/T) that will be assigned to
lake/aquifer interfaces that occur in the corresponding grid cell.
NSLMSis the number of sublake systems if coalescing/dividing lakes are to be simulated (only in transient runs). Enter 0 if no
sublake systems are to be simulated.
ICis the number of sublakes, including the center lake, in the sublake system being described in this item.
ISUBis the identification number of the sublakes in the sublake system being described in this item. The center lake number
is listed first.
SILLVTis the sill elevation (L) that determines whether the center lake is connected with a given sublake. One value is
entered in this item for each sublake in the order the sublakes are listed in the previous item.
PRCPLKis the rate of precipitation at the surface of a lake (L/T).
EVAPLKis the rate of evaporation from the surface of a lake (L/T).
RNFis the volumetric rate of overland runoff from an adjacent watershed entering the lake (L3/T). If RNF > 0, it is specified
directly as a volume. If RNF < 0, it is used as a multiple applied to the product of the lake evaporation rate and the
volume of the lake at its full stage (occupying all layer 1 lake cells). This is to account for the percentage of rainfall
becoming runoff and the size of the watershed contributing the runoff.
WTHDRWis the rate of water removal from a lake by means other than rainfall, evaporation, surface outflow, or ground-water
seepage (L3/T). A negative value indicates augmentation. Normally, this would be used to specify the rate of artificial
withdrawal from a lake for human water use, or, if negative, artificial augmentation of a lake volume for aesthetic or
recreational purposes.
Appendix 1 213

Input to the Unsaturated-Zone Flow (UZF) Package is read from the file that has file type UZF in the Name File. The
user can optionally specify unsaturated-zone water budgets and water content profiles for selected finite-difference cells by
specifying the file type DATA in the MODFLOW Name File. The row and column cell indices for each selected finite-difference
cell are included in the UZF output file. Three types of information may be printed to the specified file for each finite-difference
cell depending on the value of IUZOPT specified in the UZF input file. The three output options are: option 1 prints volumes
of water entering, leaving, and stored in the unsaturated zone; option 2 prints volumes and rates for water entering, leaving,
and stored within the unsaturated zone; and option 3 prints the water content profile between land surface and the water table.
Additionally, a time series of infiltration, unsaturated zone evapotranspiration, recharge, and ground-water discharge summed
over the model domain may be printed to a specified file. The UZF Package is documented in Niswonger and others (2006a) and
users of GSFLOW are encouraged to review that document when using the UZF Package.
The UZF Package input file consists of items numbered from 0 through 16, each consisting of one or more items.
These data are used to specify information about the hydraulic properties of the unsaturated zone, the infiltration rate,
evapotranspiration, and certain output control options. All input variables are read using free format unless indicated otherwise.
In free format, variables are separated by one or more spaces or by a comma and optionally one or more spaces. A blank field
cannot be used to set a variable value to zero. Units of length (L) and time (T) are used to define the dimensions of each variable.
The UZF Package relies on the specific yield values as specified in the Layer-Property Flow (LPF) Package, the Block-
Centered Flow (BCF) Package, or the Hydrogeologic Unit Flow (HUF) Package. When the option to use vertical hydraulic
conductivity in the LPF Package is specified, the layer(s) that contain cells where unsaturated flow will be simulated must be
specified as convertible. That is, the variable LAYTYP specified in the LPF Package must not be equal to zero, otherwise the
model will print an error and stop execution. The variable VKA in LPF may be specified as the ratio of horizontal to vertical
hydraulic conductivity (LAYVKA is not zero) or as the vertical hydraulic conductivity (LAYVKA=0). When using the BCF
Package, the right digit of LTYPE (LAYCON) must be greater or equal to one or the model will print an error and stop execution.
Neither parameter estimation nor solute transport can be simulated with the UZF Package.
FOR EACH SIMULATION
0. [#Text]
1. NUZTOP IUZFOPT IRUNFLG IETFLG IUZFCB1 IUZFCB2 NTRAIL2 NSETS2 NUZGAG
SURFDEP
The variables NTRAIL2 and NSETS2 are used for dimensioning arrays.
2. [IUZFBND(NCOL,NROW)] U2DINT
If IRUNFLG > 0, read item 3:

3. [IRUNBND(NCOL,NROW)] U2DINT
If IUZFOPT = 1, read item 4:
4. [VKS(NCOL,NROW)] U2DREL
Vertical hydraulic conductivity is optional when either the LPF or BCF Package is used to define vertical hydraulic
conductivity of cells (IUZFOPT=2).
5. [EPS(NCOL,NROW)] U2DREL
6. [THTS(NCOL,NROW)] U2DREL
7. [THTI(NCOL,NROW)] U2DREL
If NUZGAG > 0, item 8 is repeated NUZGAG times:
8. [IUZROW] [IUZCOL] [IFTUNIT] [IUZOPT]

Item 8 must include exactly NUZGAG lines of data. If NUZGAG > 1, it is permissible to repeat the same cell if a summary
of volumes and rates and water contents are desired for a given cell. The lines within item 8 can be listed in any order.
A unique unit number for IFTUNIT must be specified for each line in item 8 and each number must match that used in
the MODFLOW Name File as Ftype DATA file types and file names (see Harbaugh, 2005).
9. NUZF1
If NUZF1 > 0:
10. [FINF(NCOL,NROW)] U2DREL

If FINF is specified as being greater than the vertical hydraulic conductivity, then FINF is set equal to the vertical
unsaturated hydraulic conductivity. Excess water is routed to streams or lakes when IRUNFLG is not zero, and if SFR2
or LAK3 is active.
If IETFLG > 0, read items 11 through 16:
11. NUZF2
If NUZF2 > 0:
12. [PET(NCOL,NROW)] U2DREL

13. NUZF3
If NUZF3 > 0:
14. [EXTDP(NCOL,NROW)] U2DREL

The quantity of ET removed from a cell is limited by the volume of water stored in the unsaturated zone above the
extinction depth. If ground water is within the ET extinction depth, then the rate removed is based on a linear decrease
in the maximum rate at land surface and zero at the ET extinction depth. The linear decrease is the same method used
in the Evapotranspiration Package (McDonald and Harbaugh, 1988, chap. 10).
15. NUZF4
If NUZF4 > 0:
16. [EXTWC(NCOL, NROW)] U2DREL

EXTWC must have a value between (THTS-Sy) and THTS, where Sy is the specific yield specified in either the LPF or
BCF Package.
Explanation of Variables Read by the UZF Package

Textis a character variable (up to 199 characters) that starts in column 2. Any characters can be included in Text. The #
character must be in column 1. Lines beginning with # are restricted to the first lines of the file. Text is written to the
Listing File.
ITEM 1 VARIABLES:
NUZTOPis an integer value used to define which cell in a vertical column for which recharge and discharge is simulated.
If NUZTOP = 1, recharge to and discharge from only the top model layer. This option assumes land surface is defined as
top of layer 1.
If NUZTOP = 2, recharge to and discharge from the specified layer in variable IUZFBND. This option assumes land
surface is defined as top of layer specified in IUZFBND.
Appendix 1 215
If NUZTOP = 3, recharge to and discharge from the highest active cell in each vertical column. Land surface is
determined as top of layer specified in IUZFBND. A constant head node intercepts any recharge and prevents deeper
percolation.
IUZFOPTan integer value equal to 1 or 2. A value of 1 indicates that the vertical hydraulic conductivity will be specified
within the UZF Package input file using array VKS. A value of 2 indicates that the vertical hydraulic conductivity will be
specified within the LPF Package input file. IUZFOPT must be 1 when using the BCF Package.
IRUNFLGan integer value that specifies whether ground water that discharges to land surface will be routed to stream
segments or lakes as specified in the IRUNBND array (IRUNFLG not equal to zero) or if ground-water discharge is
removed from the model simulation and accounted for in the ground-water budget as a loss of water (IRUNFLG=0). The
Streamflow-Routing (SFR2) and(or) the Lake (LAK3) Packages must be active if IRUNFLG is not zero. This option is
available only for steady-state stress periods in GSFLOW.
IETFLGan integer value that specifies whether or not evapotranspiration (ET) will be simulated. ET will not be simulated if
IETFLG is zero, otherwise it will be simulated. This option is for simulating ET in addition to ET simulated in PRMS.
If not using this option then ET is simulated only by PRMS above the soil-zone base. If using this option then any ET
demand not met by storage in the soil zone will be applied as ET demand below the base of the soil zone.
IUZFCB1an integer value used as a flag for writing rates of ground-water recharge, ET, and ground-water discharge to land
surface to a separate unformatted file using subroutine UBUDSV. If IUZFCB1 is > 0, it is the unit number to which the
cell-by-cell rates will be written when SAVE BUDGET or a non-zero value for ICBCFL is specified in Output Control. If
IUZFCB1 is 0, cell-by-cell rates will not be written to a file.
IUZFCB2an integer value used as a flag for writing rates of ground-water recharge, ET, and ground-water discharge to land
surface to a separate unformatted file using module UBDSV3. If IUZFCB2 is > 0, it is the unit number to which cell-
by-cell rates will be written when SAVE BUDGET or a non-zero value for ICBCFL is specified in Output Control. If
IUZFCB2 is 0, cell-by-cell rates will not be written to file.
NTRAIL2an integer value that is the number of trailing-wave increments used to represent a trailing wave. Trailing waves
are used to represent a decrease in the surface infiltration rate. The value can be increased to improve mass balance in
the unsaturated zone. Values between 20 and 10 work well and result in unsaturated-zone mass balance errors ranging
between 0.001 and 0.01 percent, although for large problems fewer trailing waves between 10 and 5 can be used and
result in mass balance errors ranging between 0.01 and 0.1 percent. See Smith (1983) for further details.
NSETS2an integer value equal to the number of wave sets used to simulate multiple infiltration periods. The number of
wave sets should be set to 20 for most problems involving time varying infiltration. The total number of waves allowed
within an unsaturated zone cell is equal to the product of NTRAIL2 and NSETS2. A warning message will be printed
to the MODFLOW LIST file if the number of waves in a cell exceeds this value. NSETS2 affects memory usage but not
computational requirements.
NUZGAGan integer value equal to the number of locations (one cell per vertical column on layers) that will be specified for
printing detailed information on the unsaturated zone water budget and water content. This output option also may
be used to print time series of infiltration, evapotranspiration, unsaturated-zone storage, recharge, and ground-water
discharge summed over all finite-difference cells. See definition of the UZF variable IFTUNIT for more explanation.
SURFDEPa real value equal to the average undulation depth of the soil-zone base altitude. This value is used for calculating
recharge and ground-water discharge to the soil zone when the ground-water altitude is less than a distance of SURFDEP
above or below the soil-zone base.
ITEMS 2-7 VARIABLES:
IUZFBNDis an array of integer values used to define the areal extent of the active model in which recharge and discharge will
be simulated.
IRUNBNDis an array of integer values used to define the stream segments within the Streamflow-Routing (SFR2) Package or
lake numbers in the Lake (LAK3) Package that overland runoff from excess infiltration and ground-water discharge to
land surface will be added. A positive integer value identifies the stream segment and a negative integer value identifies
the lake number. {This option is available only for steady-state stress periods in GSFLOW.}
VKSis an array of positive real values used to define the saturated vertical hydraulic conductivity of the unsaturated zone
(L/T).
EPSis an array of positive real values for each finite-difference cell used to define the Brooks-Corey epsilon of the unsaturated
zone. Epsilon is used in the relation of water content to hydraulic conductivity (Brooks and Corey, 1966).
THTSis an array of positive real values used to define the saturated water content of the unsaturated zone in units of volume of
water to total volume (L3/L3).
THTIis an array of positive real values used to define the initial water content for each vertical column of cells in units of
volume of water at start of simulation to total volume (L3/L3). THTI is not specified for steady-state simulations.
ITEM 8 VARIABLES:
IUZROWis an integer value equal to the row number of the cell that unsaturated-zone information will be printed for each time
step. The variable is not used when IFTUNIT is negative.
IUZCOLis an integer value equal to the column number of the cell that unsaturated-zone information will be printed for each
time step. The variable is not used when IFTUNIT is negative.
IFTUNITis an integer value equal to the unit number of the output file. A positive value is for output of individual cells,
whereas a negative value is for output that is summed over all finite-difference cells. The summed output includes
applied infiltration, runoff, actual infiltration, ground-water discharge to land surface, ET from the unsaturated zone, ET
from ground water, recharge, and change in unsaturated-zone storage.
IUZOPTis an integer value that is a flag for the type of expanded listing desired in the output file. The variable is not used
when IFTUNIT is negative.
If IUZOPT = 1, prints time, ground-water head, and thickness of unsaturated zone, and cumulative volumes of
infiltration, recharge, storage, change in storage and ground-water discharge to land surface.
If IUZOPT = 2, same as option 1 except rates of infiltration, recharge, change in storage, and ground-water discharge
also are printed.
If IUZOPT = 3, prints time, ground-water head, thickness of unsaturated zone, followed by a series of depths and water
contents in the unsaturated zone.
ITEMS 9-16 VARIABLES:
NUZF1is an integer value for reusing or reading infiltration rates that can change each stress period. If NUZF1 > 0, then
infiltration rates at land surface are specified. If NUZF1 < 0, then infiltration rates from the previous stress period are
used. This variable only is used during steady-state stress periods for integrated simulations. A dummy value must be
specified for stress periods that are not steady state.
FINFis an array of positive real values used to define the infiltration rates (L/T) at land surface for each vertical column of
cells. This variable only is used during steady-state stress periods.
Appendix 1 217
NUZF2is an integer value for reusing or reading ET demand rates that can change each stress period. If NUZF2 > 0, then ET
demand rates are specified. If NUZF2 < 0, then ET demand rates from the previous stress period are used. {This variable
is not used in GSFLOW simulations because the ET demand below the soil-zone base is calculated by PRMS. However,
if ET is to be simulated below the soil-zone base (IETFLG > 0), NUZF2 should be set to a value less than zero.}
PETis an array of positive real values used to define the ET demand rates (L/T) within the ET extinction depth interval for
each vertical column of cells. {PET is not used for GSFLOW simulations.}
NUZF3is an integer value for reusing or reading ET extinction depths that can change each stress period. If NUZF3 > 0, then
ET extinction depths are specified. If NUZF3 < 0, then depths from the previous stress period are used. {If ET is to be
simulated below the soil-zone base (IETFLG > 0) NUZF3 must be specified.}
EXTDPis an array of positive real values used to define the ET extinction depths {relative to the altitude of the soil-zone base
for GSFLOW simulations. If ET is to be simulated below the soil-zone base (IETFLG > 0) EXTDP must be specified for
the first stress period at a minimum.}
NUZF4is an integer value for reusing or reading the extinction water content that can change each stress period. If NUZF4 >
0, then extinction water contents are specified. If NUZF4 < 0, then the extinction water contents from the previous stress
period are used. {If ET is to be simulated below the soil-zone base (IETFLG > 0) NUZF4 must be specified.}
EXTWCis an array of positive real values used to define the extinction water content below which ET cannot be removed
from the unsaturated zone {below the soil-zone base.} {If ET is to be simulated below the soil-zone base (IETFLG > 0)
EXTWC must be specified for the first stress period at a minimum.}
Gage Package
The input file for specifying stream and lake gaging-station locations is read if file type (Ftype) GAGE is included in
the MODFLOW Name File. A particular stream reach or lake can be designated for a gaging or monitoring station. At each
designated stream reach or lake, the time, stream or lake stage, and streamflow out of the reach or lake volume will be written to
a separate output file to facilitate model output evaluation and graphical post processing of the calculated data. Several options
are available to also print additional information about a stream reach or lake. The output file will contain two header lines that
provide relevant information and the text will be contained within quotes.
FOR EACH SIMULATION
1. NUMGAGE
FOR EACH GAGING STATION
Items 2a and 2b are used to specify information for stream or lake gaging stations, respectively. The total number of lines
specified for items 2a and 2b must equal NUMGAGE. Items 2a and 2b can be entered in any arbitrary order, and it is
permissible to interleaf items for stream-gaging stations with those for lake gaging stations if NUMGAGE > 1.
FOR EACH STREAM GAGING STATION
2a. GAGESEG GAGERCH UNIT OUTTYPE
FOR EACH LAKE GAGING STATION
2b. LAKE UNIT [OUTTYPE]
Explanation of Variables Read by the GAGE Package:

NUMGAGEis the total number of stream and lake gaging stations.
GAGESEGis an integer value that is the stream segment number where the gage is located.
GAGERCHis an integer value that is the stream reach number where the gage is located.
LAKEis a negative integer value that is the lake number where the gage is located.
UNITis an integer value that is the unit number of the output file for this gage. A unique unit number must be specified for
each gage and the unit numbers must match data file types and file names in the MODFLOW Name File (see Harbaugh
and others, 2000, p. 42-44).
OUTTYPEfor a stream gage, OUTTYPE is an integer value that is a flag for the type of expanded listing desired in output file.
For a stream gage the options are:
OUTTYPE = 0, use standard default listing of time, stream stage, and outflow.
OUTTYPE = 1, default values plus depth, width, flow at midpoint, precipitation, evapotranspiration, and runoff.
Computed runoff from the Unsaturated-Zone Flow (UZF) Package is added when the UZF Package is active.
OUTTYPE = 2, default values plus streambed conductance for the reach, head difference across streambed, and hydraulic
gradient across streambed.
OUTTYPE = 3, {This option is not used in GSFLOW.}
OUTTYPE = 4, all of the above.
OUTTYPE = 5, use for diversions to provide a listing of time, stage, flow diverted, maximum assigned diversion rate, and
flow at end of upstream segment prior to diversion.
OUTTYPE = 6, use for unsaturated flow routing to provide a listing of time, stream stage, ground-water head beneath
stream, streambed seepage, change in unsaturated zone storage, and recharge to ground water.
OUTTYPE = 7, use for unsaturated flow routing to provide a listing of time and the unsaturated-water content profile
beneath the stream. Two profiles are printed. The first is the volume averaged water content of all unsaturated zone
cells (multiple unsaturated zone cells are allowed) beneath a stream reach. The second profile is the volume averaged
water content beneath the low-flow channel only. The two profiles are identical when only one unsaturated zone cell is
assigned beneath a stream reach.
For a lake gage, OUTTYPE is an integer value that is a flag for type of expanded listing desired in output file. It is read only
when UNIT is a negative value. Options are:
OUTTYPE = 0, standard default listing of time, lake stage, and lake volume.
OUTTYPE = 1, default values plus all inflows to and outflows from lake (as volumes during time increment), and total
lake conductance. Computed runoff from the UZF Package is added whenever the UZF Package is active.
OUTTYPE = 2, default values plus changes in lake stage and lake volume.
OUTTYPE = 3, all of the above.
Note: Total lake conductance (OUTTYPE = 1 and 3 for a lake gaging station) is the sum of the conductances of each seepage
interface for each lake. Changes in lake stage and volume (OUTTYPE = 2 or 3) are listed as incremental changes from previous
time increment and as cumulative change since the start of the simulation.
Strongly Implicit Procedure Package

Input to the Strongly Implicit Procedure (SIP) Package is read from the file that is type SIP in the Name File. All numeric
variables are free format if the option FREE is specified in the Basic Package input file; otherwise, all the variables have
10-character fields. The SIP Package is documented in Harbaugh (2005) and users of GSFLOW are encouraged to review that
document when using the SIP Package.
FOR EACH SIMULATION
0. [#Text]
1. MXITER NPARM
2. ACCL HCLOSE IPCALC WSEED IPRSIP
Appendix 1 219
Explanation of Variables Read by the SIP Package:

MXITERis the maximum number of times through the iteration loop in one time step in an attempt to solve the system of
finite-difference equations.
NPARMis the number of iteration variables to be used. Five variables generally are sufficient.
ACCLis the acceleration variable, which must be greater than zero and is generally equal to one. If a zero is entered, it is
changed to one.
HCLOSEis the head change criterion for convergence. When the maximum absolute value of head change from all nodes
during an iteration is less than or equal to HCLOSE, iteration stops.
IPCALCis a flag indicating where the seed for calculating iteration variables will come from.
0the seed entered by the user will be used.
1the seed will be calculated at the start of the simulation from problem variables.
WSEEDis the seed for calculating iteration variables. WSEED is always read but is used only if IPCALC is equal to zero.
IPRSIPis the printout interval for SIP. IPRSIP, if equal to zero, is changed to 999. The maximum head change (positive
or negative) is printed each iteration of a time step whenever the time step is an even multiple of IPRSIP. This printout
also occurs at the end of each stress period regardless of the value of IPRSIP.
Preconditioned Conjugate-Gradient Package

Input to the Preconditioned Conjugate-Gradient (PCG) Package is read from the file that is type PCG in the Name File. All
numeric variables are free format if the option FREE is specified in the Basic Package input file; otherwise, all variables have
10-character fields. The PCG Package is documented in Harbaugh (2005) and users of GSFLOW are encouraged to review that
document when using the PCG Package.
FOR EACH SIMULATION
0. [#Text]
1. MXITER ITER1 NPCOND
2. HCLOSE RCLOSE RELAX NBPOL IPRPCG MUTPCG DAMPPCG [DAMPPCG2]
Explanation of Variables Read by the PCG Package:

MXITERis the maximum number of outer iterations (Hill, 1990). For a linear problem MXITER should be 1, unless more than
50 inner iterations are required, when MXITER could be as large as 1000.
ITER1is the number of inner iterations (Hill, 1990). For nonlinear problems, ITER1 usually ranges from 10 to 30; a value
of 30 will be sufficient for most linear problems.
NPCONDis the flag used to select the matrix conditioning method:

1is for Modified Incomplete Cholesky (for use on scalar computers)
2is for Polynomial (for use on vector computers or to conserve computer memory)
HCLOSEis the head change criterion for convergence, in units of length. When the maximum absolute value of head change
from all nodes during an iteration is less than or equal to HCLOSE, and the criterion for RCLOSE also is satisfied,
iteration stops.
RCLOSEis the residual criterion for convergence, in units of cubic length per time. The units for length and time are the
same as established for all model data. (See LENUNI and ITMUNI input variables in the Discretization File.) When the
maximum absolute value of the residual at all nodes during an iteration is less than or equal to RCLOSE, and the criterion
for HCLOSE also is satisfied, iteration stops.
For nonlinear problems, convergence is achieved when the convergence criteria are satisfied for the first inner iteration.
RELAXis the relaxation parameter used with NPCOND = 1. Usually, RELAX = 1.0, but for some problems a value of 0.99,
0.98, or 0.97 will reduce the number of iterations required for convergence. RELAX is not used if NPCOND is not 1.
NBPOLis used when NPCOND = 2 to indicate whether the estimate of the upper bound on the maximum eigen value is 2.0,
or whether the estimate will be calculated. NBPOL = 2 is used to specify the value is 2.0; for any other value of NBPOL,
the estimate is calculated. Convergence is generally insensitive to this NBPOL. NBPOL is not used if NBPOL is not 2.
IPRPCGis the printout interval for PCG. IPRPCG, if equal to zero, is changed to 999. The maximum head change (positive
or negative) and residual change are printed for each iteration of a time step whenever the time step is an even multiple of
IPRPCG. This printout also occurs at the end of each stress period regardless of the value of IPRPCG.
MUTPCGis a flag that controls printing of convergence information from the solver:
0is for printing tables of maximum head change and residual each iteration
1is for printing only the total number of iterations
2is for no printing
3is for printing only if convergence fails
DAMPPCGis the damping factor. A value of one indicates no damping. A value less than 1 and greater than 0 causes damping.
{Because steady-state stress periods often require lower values of DAMPPCG for convergence, the option of using a
separate damping factor for transient stress periods was added for GSFLOW. If DAMPPCG is specified as a negative value
then an additional damping factor will be read for transient stress periods and DAMPPCG will be converted to a positive
value and will only be used for steady-state stress periods.}
{DAMPPCG2is the damping factor used for transient stress periods. DAMPPCG2 only is read if DAMPPCG<0. A value of one
indicates no damping. A value less than 1 and greater than 0 causes damping.}
Direct Solver Package

Input to the Direct Solver (DE4) Package is read from the file that is type DE4 in the Name File. All numeric variables
are free format. The DE4 Package is documented in Harbaugh (2005) and users of GSFLOW are encouraged to review that
document when using the DE4 Package.
FOR EACH SIMULATION
0. [#Text]
1. ITMX MXUP MXLOW MXBW
2. IFREQ MUTD4 ACCL HCLOSE IPRD4
Explanation of Variables Read by the DE4 Package:

Appendix 1 221
ITMXis the maximum number of iterations for each time step. Specify ITMX = 1 if iteration is not desired. Ideally,
iteration would not be required for direct solution; however, iteration is necessary if the flow equation is nonlinear (see
explanation for IFREQ = 3) or if computer precision limitations result in inaccurate calculations as indicated by a large
water-budget error. For a nonlinear flow equation, each iteration is equally time consuming because [A] (eq. 2-27 in
Harbaugh, 2005) changes each iteration and Gaussian elimination is required after each change. This is called external
iteration. For a linear equation, iteration is substantially faster because [A] is changed at most once per time step; thus,
Gaussian elimination is required at most once per time step. This is called internal iteration.
MXUPis the maximum number of equations in the upper part of the equations to be solved. This value affects the amount of
memory used by the DE4 Package. If specified as 0, the program will calculate MXUP as half the number of cells in the
model, which is an upper limit. The actual number of equations in the upper part will be less than half the number of
cells whenever no-flow and constant-head cells are included because flow equations are not formulated for these cells.
The DE4 Package prints the actual number of equations in the upper part when it runs. The printed value can be used for
MXUP in future runs to minimize memory usage.
MXLOWis the maximum number of equations in the lower part of equations to be solved. This value affects the amount of
memory used by the DE4 Package. If specified as 0, the program will calculate MXLOW as half the number of cells in
the model, which is an upper limit. The actual number of equations in the lower part will be less than half the number of
cells whenever no-flow and constant-head cells are included because flow equations are not formulated for these cells.
The DE4 Package prints the actual number of equations in the lower part when it runs. The printed value can be used for
MXLOW in future runs to minimize memory usage.
MXBWis the maximum bandwidth plus 1 of the [AL] matrix. This value affects the amount of memory used by the DE4
Package. If specified as 0, the program will calculate MXBW as the product of the two smallest grid dimensions plus 1,
which is an upper limit. The DE4 Package prints the actual bandwidth plus 1 when it runs. The printed value can be used
for MXBW in future runs to minimize memory usage.
IFREQis a flag indicating the frequency at which coefficients in [A] change. This affects the efficiency of solution; much
work can be avoided if the user knows that [A] remains constant all or part of the time.
IFREQ = 1 indicates that the flow equations are linear and that coefficients of simulated head for all stress terms are
constant for all stress periods. To meet the linearity requirement, all model layers must be confined (which is specified
in the Block-Centered Flow Package by setting LAYCON equal to 0 for all layers or in the Layer-Property Flow
Package by setting LAYTYP equal to 0 for all layers), and formulations must not change based upon head (such as
seepage from a river changing from head-dependent flow to a constant flow when head drops below the bottom of the
riverbed). Examples of coefficients of simulated head for stress terms are riverbed conductance, drain conductance,
maximum evapotranspiration rate, evapotranspiration extinction depth, and general-head boundary conductance.
IFREQ = 2 indicates that the flow equations are linear as described for IFREQ = 1, but coefficients of simulated head for
some stress terms may change at the start of each stress period. Examples of coefficients of simulated head for stress
terms are riverbed conductance, drain conductance, maximum evapotranspiration rate, evapotranspiration extinction
depth, and general-head boundary conductance. For a simulation consisting of only one stress period, IFREQ = 2 has
the same meaning as IFREQ = 1.
IFREQ = 3 indicates that a nonlinear flow equation is being solved, which means that some terms in [A] depend on
simulated head. Examples of head-dependent terms in [A] are transmissivity for water-table layers, which is based on
saturated thickness; flow terms for rivers, drains, and evapotranspiration if they convert between head-dependent flow
and constant flow; and the change in storage coefficient when a cell converts between confined and unconfined. When
a nonlinear flow equation is being solved, external iteration (ITMX > 1) is normally required to accurately approximate
the nonlinearities. Note that when nonlinearities caused by water-table calculations are part of a simulation, obvious
signs may not be present in the output from a simulation that does not use external iteration to indicate that iteration is
needed. In particular, the budget error may be acceptably small without iteration even though large error in head exists
because of nonlinearity. Additional information about this issue is contained in Chapter 7 of Harbaugh (2005).
MUTD4is a flag that indicates the quantity of information that is printed when convergence information is printed for a time
step.
MUTD4 = 0 indicates that the number of iterations in the time step and the maximum head change each iteration are
printed.
MUTD4 = 1 indicates that only the number of iterations in the time step is printed.
MUTD4 = 2 indicates no information is printed.
ACCLis a multiplier for the computed head change each iteration. Normally this value is 1. A value greater than 1 may be
useful for improving the rate of convergence when using external iteration to solve nonlinear problems (IFREQ = 3).
ACCL should always be 1 for linear problems. When ITMX=1, ACCL is changed to 1 regardless of the input value;
however, a value must always be specified.
HCLOSEis the head change closure criterion. If iterating (ITMX > 1), iteration stops when the absolute value of head change
at every node is less than or equal to HCLOSE. HCLOSE is not used if not iterating, but a value must always be specified.
IPRD4is the time step interval for printing out convergence information when iterating (ITMX > 1). For example, if IPRD4
is 2, convergence information is printed every other time step. A value must always be specified even if not iterating.
Geometric Multigrid Solver Package

Input to the Geometric Multigrid Solver (GMG) Package is read from the file that is type GMG in the Name File. All
variables in the input list are free format. The GMG Package is documented in Wilson and Naff (2004) and users of GSFLOW
are encouraged to review that document when using the GMG Package. The GMG Package is based on the PCG Package
algorithm.
FOR EACH SIMULATION
0. [#Text]
Item 0 is optional# must be in column 1. Item 0 can be repeated multiple times and between items 1 through 4.
1. RCLOSE IITER HCLOSE MXITER
2. DAMP IADAMP IOUTGMG
3. ISM ISC
4. RELAX
Explanation of Variables Read by the GMG Package:

RCLOSE is the residual convergence criterion for the inner iteration. The PCG algorithm computes the l2-norm of the residual
and compares it against RCLOSE. Typically, RCLOSE is set to the same value as HCLOSE. If RCLOSE is set too high,
then additional outer iterations may be required due to the linear equation not being solved with sufficient accuracy.
Conversely, too restrictive setting for RCLOSE for nonlinear problems may force an unnecessarily accurate linear
solution. This may be alleviated with the IITER parameter or with damping.
The convergence criteria for the GMG Package may be quite different from the other solver packages. Please refer to
the Convergence Criteria section of Wilson and Naff (2004, p. 9).
IITER is the maximum number of PCG iterations for each linear solution. A value of 100 is typically sufficient. It is
frequently useful to specify a smaller number for nonlinear problems so as to prevent an excessive number of inner
iterations.
HCLOSE is the head-change convergence criterion for nonlinear problems. After each linear solve (inner iteration), the max-
norm of the head change is compared to HCLOSE. HCLOSE can be set to a large number for linear problems; HCLOSE
is ignored if MXITER=1.
MXITER is the maximum number of outer iterations. For linear problems, MXITER can be set to 1. For nonlinear problems,
MXITER needs to be larger, but rarely more than 100.
Appendix 1 223
DAMP is the value of the damping parameter. For linear problems, a value of 1.0 should be used. For nonlinear problems,
a value less than 1.0 but greater than 0.0 may be necessary to achieve convergence. A typical value for nonlinear
problems is 0.5. Damping also helps control the convergence criterion of the linear solve to alleviate excessive PCG
iterations [see equation 20 in Wilson and Naff (2004, p. 9)].
IADAMP is a flag that controls adaptive damping. The possible values of IADAMP and their meanings are as follows:
If IADAMP = 0, then the value assigned to DAMP is used as a constant damping parameter. time of the simulation. This
results in use of the first value in arrays FLWRAT, SBHED, and AuxVar for each respective boundary cell.
If IADAMP 0, then the value of DAMP is used for the first nonlinear iteration.
[see Nonlinear Solution section in Wilson and Naff (2004, p. 8)]. The damping parameter is adaptively varied on the
basis of the head change, using Cooleys method as described in Mehl and Hill (2001), for subsequent iterations.
IOUTGMG is a flag that controls the output of the GMG solver. The possible values of IOUTGMG and their meanings are as
follows:
0is for printing only the solver inputs.
1is for printing of each linear solve, the number of PCG iterations, the value of the damping parameter, the l2-norm of
the residual, and the maxnorm of the headchange and its location (column, row, layer). At the end of a time/stress period,
the total number of GMG calls, PCG iterations, and a running total of PCG iterations for all time/stress periods are
printed.
2is for printing the convergence history of the PCG iteration that includes the 12-norm of the residual and the
convergence factor for each iteration.
3is the same as 1, except output is sent to the terminal instead of the MODFLOW list file.
4is the same as 2, except output is sent to the terminal instead of the MODFLOW list file.
ISM is a flag that controls the type of smoother used in the multigrid preconditioner. The possible values for ISM and their
meanings are as follows:
If ISM = 0, then ILU(0) smoothing is implemented in the multigrid preconditioner. This smoothing requires an
additional vector on each multigrid level to store the pivots in the ILU factorization.
If ISM = 1, then Symmetric GaussSeidel (SGS) smoothing is implemented in the multigrid preconditioner. No additional
storage is required for this smoother; users may want to use this option if available memory is exceeded or nearly
exceeded when using ISM =0. Using SGS smoothing is not as robust as ILU smoothing; additional iterations are likely
to be required in reducing the residuals. In extreme cases, the solver may fail to converge as the residuals cannot be
reduced sufficiently.
ISC is a flag that controls semicoarsening in the multigrid preconditioner. The possible values of ISC and their meanings are
given as follows:
If ISC = 0, then the rows, columns, and layers are all coarsened.
If ISC = 1, then the rows and columns are coarsened, but the layers are not.
If ISC = 2, then the columns and layers are coarsened, but the rows are not.
If ISC = 3, then the rows and layers are coarsened, but the columns are not.
If ISC = 4, then there is no coarsening.
Typically, the value of ISC should be 0 or 1. In the case that there are large vertical variations in the hydraulic
conductivities, then a value of 1 should be used [see Remark 9 in Wilson and Naff (2004, p. 18). If no coarsening is
implemented (ISC = 4), then the GMG solver is comparable to the PCG2 ILU(0) solver described in Hill (1990) and
uses the least amount of memory.
RELAX is a relaxation parameter for the ILU preconditioned conjugate gradient method. The RELAX parameter can be used to
improve the spectral condition number of the ILU preconditioned system. The value of RELAX should be approximately
one. However, the relaxation parameter can cause the factorization to break down. If this happens, then the GMG solver
will report an assembly error and a value smaller than one for RELAX should be tried. This item is read only if ISC = 4.
Array Reading Utility Subroutines

The array reading utility subroutines provide a common way for all packages to read variables that have multiple values.
The term array is simply a programming term for a variable that contains multiple values. There are three subroutines:
U2DREL, U2DINT, and U1DREL. U2DREL reads real two-dimensional variables, U2DINT reads integer two-dimensional
variables, and U1DREL reads real one-dimensional variables. All of these subroutines work similarly. They read one array-
control line and, optionally, a data array in a format specified on the array-control line. Several alternate structures for the control
line are provided. The original fixed-format control lines work as documented in McDonald and Harbaugh (1988), and four free-
format versions have been added. The free-format versions are described first because they are easier to use.
FREE-FORMAT CONTROL LINES FOR ARRAY READERS:
Values in bold italics are keywords that can be specified as uppercase or lowercase. Each control line is limited to a length
of 199 characters.
1. CONSTANT CNSTNT
All values in the array are set equal to CNSTNT.
2. INTERNAL CNSTNT FMTIN IPRN
The individual array elements will be read from the same file that contains the control line.
3. EXTERNAL Nunit CNSTNT FMTIN IPRN
The individual array elements will be read from the file unit number specified by Nunit. The name of the file associated
with this file unit must be contained in the Name File.
4. OPEN/CLOSE FNAME CNSTNT FMTIN IPRN
The array will be read from the file whose name is specified by FNAME. This file will be opened on unit 99 just prior
to reading the array and closed immediately after the array is read. This file should not be included in the Name File. A
file that is read using this control line can contain only a single array.
FIXED-FORMAT CONTROL LINE FOR ARRAY READERS:
A fixed-format control line contains the following variables:

LOCAT CNSTNT FMTIN IPRN
These variables are explained below. LOCAT, CNSTNT, and IPRN are 10-character numeric fields. For U2DREL and U1DREL,
CNSTNT is a real number. For U2DINT, CNSTNT is an integer and must not include a decimal. FMTIN is a 20-character text
field. All four variables are always read when the control line is fixed format; however, some of the variables are unused in some
situations. For example when LOCAT = 0, FMTIN and IPRN are not used.
Explanation of Variables in the Control Lines:

LOCATindicates the location of the array values for a fixed-format array control line. If LOCAT = 0, all elements are set equal
to CNSTNT. If LOCAT > 0, it is the unit number specified in the MODFLOW NAME file for reading formatted lines
using FMTIN as the format. If LOCAT < 0, it is the unit number for binary (unformatted) lines, and FMTIN is ignored.
Also, when LOCAT is not 0, the array values are multiplied by CNSTNT after they are read.
Appendix 1 225
CNSTNTis a real-number constant for U2DREL and U1DREL, and an integer constant for U2DINT. If the array is being
defined as a constant, CNSTNT is the constant value. If individual elements of the array are being read, the values are
multiplied by CNSTNT after they are read. CNSTNT, when used as a multiplier and specified as 0, is changed to 1.
FMTINis the format for reading array elements. The format must contain 20 characters or less. The format must either be
a standard Fortran format that is enclosed in parentheses, (FREE) which indicates free format, or (BINARY) which
indicates binary (unformatted) data. When using a free-format control line, the format must be enclosed in apostrophes
if it contains one or more blanks or commas. A binary file that can be read by MODFLOW may be created in only two
ways. The first way is to use MODFLOW to create the file by saving heads in a binary file. This is commonly done when
the user desires to use computed heads from one simulation as initial heads for a subsequent simulation. The other way
to create a binary file is to write a special program that generates a binary file, and compile this program using a Fortran
compiler that is compatible with the compiler used to compile MODFLOW. (FREE) and (BINARY) only can be specified
in free-format control lines. Also, (BINARY) can be specified only when using U2DREL or U2DINT, and only when
the control line is EXTERNAL or OPEN/CLOSE. When the (FREE) option is used, be sure that all array elements have a
non-blank value and that a comma or at least one blank separates adjacent values.
IPRNis a flag that indicates whether the array being read should be written to the Listing File after the array has been read
and a code for indicating the format that should be used when the array is written. The format codes are different for
each of the three array-reading subroutines as shown below. IPRN is set to zero when the specified value exceeds those
defined. If IPRN is less than zero, the array will not be printed.
IPRN U2DREL U2DINT U1DREL
0 10G11.4 10I11 10G12.5

1 11G10.3 60I1 5G12.5
2 9G13.6 40I2
3 15F7.1 30I3
4 15F7.2 25I4
5 15F7.3 20I5
6 15F7.4 10I11
7 20F5.0 25I2
8 20F5.1 15I4
9 20F5.2 10I6
10 20F5.3
11 20F5.4
12 10G11.4
13 10F6.0
14 10F6.1
15 10F6.2
16 10F6.3
17 10F6.4
18 10F6.5
19 5G12.5
20 6G11.4
21 7G9.2
Nunitis the unit for reading the array when the EXTERNAL free-format control line is used.
Examples of Free-Format Control Lines
The following examples use free-format control lines for reading an array. The example array is a real array consisting of 4
rows with 7 columns per row:
CONSTANT 5.7 This sets an entire array to the value 5.7.

INTERNAL 1.0 (7F4.0) 3 This reads the array values from the
1.2 3.7 9.3 4.2 2.2 9.9 1.0 file that contains the control line.
3.3 4.9 7.3 7.5 8.2 8.7 6.6 Thus, the values immediately follow the
4.5 5.7 2.2 1.1 1.7 6.7 6.9 control line.
7.4 3.5 7.8 8.5 7.4 6.8 8.8
EXTERNAL 52 1.0 (7F4.0) 3 This reads the array from the formatted file opened on unit 52.
EXTERNAL 47 1.0 (BINARY) 3 This reads the array from the binary file opened on unit 47.
OPEN/CLOSE test.dat 1.0 (7F4.0) 3 This reads the array from the file named test.dat.
List Utility Subroutine (ULSTRD)

Subroutine ULSTRD reads lists that are any number of repetitions of an input item that contains multiple variables.
Examples of packages that make use of this subroutine are the General-Head Boundary, Well, and Streamflow-Routing
Packages.
1. [EXTERNAL IN ] or [OPEN/CLOSE FNAME]
If Item 1 is not included, then the list is read from the package file. Item 1 must begin with the keyword EXTERNAL or
the keyword OPEN/CLOSE (not both).
2. [SFAC Scale]
3. List
Explanation of Variables Read by the List Utility Subroutine:

INis the unit number for a file from which the list will be read. The name of the file associated with this file unit must be
contained in the Name File, and its file type must be DATA in the Name File.
FNAMEis the name of a file from which the list will be read. This file will be opened on unit 99 just before reading the list and
closed immediately after the list is read. This file should not be included in the Name File.
Scaleis a scale factor that is multiplied times the value of one or more variables within every line of the list. The input
instructions that define a list, which will be read by ULSTRD, should specify the variables to which SFAC applies. If
item 2 is not included, then Scale is 1.0. If item 2 is included, it must begin with the keyword SFAC. The values of
the list variables that are printed to the listing file include the effect of Scale.
Listis a specified number of lines of data in which each line contains a specified number of variables. The first three
variables are always layer, row, and column. The other fields vary according to which package is calling this subroutine.
Appendix 2 227
Appendix 2. Definitions of Symbols.

Symbol Equation Definition
Aaqm, n 92 Area of the lakebed covering a finite-difference cell, in length squared
msf 1r
Adwn 82 Cross-sectional area at the downstream end of a stream reach for the previous, streamflow-routing
time step msfr-1, in length squared
Afdc 40, 42, 108 Top area of the finite-difference cell, in length squared
Afdca 42, 60 Top area of finite-difference cell, in acres
AGVR 42, 45, 56 Area of gravity reservoir, in acres
AGVR,L 46 Area of gravity reservoir L, in acres
AHRU 22, 31, 32, 33, 37, Area of the HRU, in acres
39, 45, 46, 58, 67a
AHRU,dwn 36 Area of the downslope HRU, in acres
AHRU,up 36 Area of the upslope HRU, in acres
AJ 77, 90 Area of HRU J, in acres
ai,j,k,N 112a, 112b Flow from the Nth source into cell i, j, k, in cubic length per time
m1
A lksurf
93 Surface area of the lake for the last iteration of time step m-1, in length squared
AlakeHRU 91 Area of the lake HRU, in acres

Aperv 37, 38a, 38b, 39 Pervious area of the HRU, in acres
A msfr
up
82 Cross-sectional area at the upstream end of a stream reach for the current streamflow-routing time
step msfr, in length squared
Aupmsfr 1 82 Cross-sectional area at the upstream end of a stream reach for the previous streamflow-routing
time step msfr-1, in length squared
m
apHRU 15a Degree-day based ratio of actual to potential shortwave radiation for the HRU during time step m,
dimensionless
m
AscHRU 29, 32 Snow-covered area of the HRU determined by snow-cover areal depletion curve (Anderson,
1973), in acres
awdth 87 Empirical coefficients determined from regression methods, undefined dimension
B 16b Constant used in the cloud-cover to solar-radiation relation, a value can be obtained from
Thompson (1976, fig. 1), dimensionless
bmonth 1 Monthly maximum (or minimum) daily temperature lapse rate representing the change in
maximum (or minimum) air temperature per 1,000 feet or meters of altitude change for each
month, January to December, in degrees Fahrenheit or Celsius per 1,000 length units
bX,month 3b, 3c Maximum (or minimum) air-temperature regression coefficients for longitude, latitude, and
bY,month altitude, respectively by month, starting with January, in degrees Fahrenheit or Celsius
and bZ,month
bwdth 87 Empirical exponent determined from regression methods, dimensionless
C 105 Matrix of conductance values for the row, column, and layer directions (CR, CC, and CV)
that are multiplied by the ground-water head in each variable-head cell and any coefficients
multiplied by a variable-head cell to represent external sources and sinks, in length squared per
time
Cj 107 Horizontal conductance for cell j in the row direction, in length squared per time
Cj-1 107 Horizontal conductance for cell j-1 in the row direction, in length squared per time
Ch 106 Horizontal conductance for a prism, in length squared per time
Cice 25 Specific heat of ice, in calories per gram-degree Celsius
228 GSFLOWFlow Model Based on he Integration of the PRMS and the MODFLOW-2005

m
C imper
31 Water available for Hortonian runoff from the impervious part per unit area of the HRU during
time step m, in inches
Cm 85, 86 Constant, which is 1.0 for cubic meters per second and 1.486 for cubic feet per second
Cmf2prms 40, 42 Conversion from MODFLOW-2005 length per time to PRMS inches per day
Cprms2mf 60 Conversion from PRMS inches per day to MODFLOW-2005 length per time
Cv 108 Vertical conductance, in length squared per time
C1i,j,k,N 112a Constant for head-dependent boundary conditions, in length squared per time
C2i,j,k,N 112a Constant for head-independent boundary conditions, in cubic length per time
CCi-1/2 104 Conductance between nodes i-1, j, k and i, j, k, in length squared per time
CCi+1/2 104 Conductance between nodes i, j, k and i+1, j, k, in length squared per time
m
CC HRU
16a, 16b Decimal fraction of cloud cover on the HRU during time step m, dimensionless
Cd lcm , n 93 Lakebed conductance of cell face lc for time step m and iteration n, in length squared per time
Ci 52 Constant of integration, days

cdpth 87 Empirical coefficients determined from regression methods, undefined dimension
Celthk 41 Thickness of the finite-difference cell, in length
Celtop 41, 62 Top altitude of the finite-difference cell, in length
CFHRU 7, 8 Monthly correction factor as a decimal fraction used to adjust rain (or snow) at the HRU
CFHRU, rain 10a, 10d Monthly snow correction factor as a decimal fraction of precipitation at the measurement station
CFxyz 9a Monthly correction factor as a decimal fraction used to adjust rain (or snow)
Chbnd 105 Vector of conductance values multiplied by ground-water head in each constant-head cell, in cubic
length per time
K v Afdc
CNDsz 41 Conductance across the soil-zone base equal to
0.5celthkDusz
(
m , n 1
h fdc celtop + 0.5 Dusz ) , in length
squared per time
coefexHRU 59 Exponent in the equation used to compute gravity drainage from the gravity reservoir,
dimensionless
coeflinHRU 59 Linear coefficient in the equation used to compute gravity drainage from the gravity reservoir, in
inches per day
CRj-1/2 101, 102, 104, 107 Conductance between nodes i, j-1, k and i, j, k, in length squared per time
CRj+1/2 104 Conductance between nodes i, j, k and i, j+1, k, in length squared per time
crbmonth 16a Slope of the regression equation that relates cloud cover to daily minimum and maximum air
temperature by month starting in January, dimensionless
csHRU 24 Snowpack settlement-time constant, per day
CVk-1/2 104, 109, 110 Vertical conductance between cells i, j, k-1 and i, j, k, in length squared per time
CVk+1/2 104, 109, 110 Vertical conductance between cells i, j, k and i, j, k+1, in length squared per time
D 114 Percent discrepancy, dimensionless
D m1
CPR
34b, 39, 63a Volume per unit area of water in the capillary reservoir at the last iteration of time step m-1, in
inches
m, n
DCPR 39, 44, 46 Volume of water in the capillary reservoir at time step m and iteration n, in inches
m , n*
DCPR 46 Revised volume per unit area of water in the capillary reservoir for time step m, iteration n, in
inches
m1
DGVR 40, 54, 55, 56 Volume of water per unit area in the gravity reservoir at time step m-1, in inches
m, n
DGVR 40, 48, 49, 50, 51, 52, 53, Volume of water per unit area in the gravity reservoir at time step m and iteration n, in inches
54, 55, 58, 59
Appendix 2 229
D m , n 1
GVR
45 Volume per unit area of water in the gravity reservoir for time step m and iteration n-1, in acre-
inch
m , n*
DGVR 43, 58 Volume per unit area in a gravity reservoir after slow interflow and before gravity drainage for
time step m and iteration n, in inches
m , n **
DGVR 43 Revised volume per unit area in a gravity reservoir after gravity drainage and before Dunnian
runoff for time step m and iteration n, in inches
m , n **
DGVR , tot
64 Volume of water in all gravity reservoirs after interflow and gravity drainage in an HRU, divided
by the area of the HRU for time step m and iteration n, in inches
m
Dimper 33 Impervious storage as volume per unit area for the HRU, during time step m, in inches
m1
Dimper 31, 32, 33 Impervious storage for the last iteration of time step m-1, as volume per unit area for the HRU, in
inches
m, n
Dmdpt 86, 87, 89 Depth of water at the midpoint of a stream reach for time step m and iteration n, in length
Dmfm , n2GVR 42, 43 Volume per unit area from the finite-difference cell to a connected gravity reservoir for time step
m and iteration n, in inches
m
DPFR 67 Volume of water in the preferential-flow reservoir per unit area for time step m, in inches
m, n
D PFR 48 Volume per unit area in the preferential-flow reservoir for time step m, iteration n, in inches
m , n 1
DPFR 64 Volume per unit area of water in the preferential-flow reservoir before fast interflow for time step
m and iteration n-1, in inches
m1
DPFR 48, 67 Volume per unit area in the preferential-flow reservoir at the last iteration of time step m-1, in
inches
m,n
Drej 43 Gravity drainage volume per unit area that is rejected by the finite-difference cell, in inches
m, n
DSZfc 64, 65 Volume of water per unit area of the soil zone above field capacity for time step m and iteration n,
in inches
Dusz 41, 62 Depth of undulations at soil-zone base; in length
m, n
Ddeficit CPR 44, 45 Volume of water per unit HRU area required to replenish the capillary reservoir to the field-
capacity threshold for time step m and iteration n, in inches
m, n
DexcessCPR 40 Volume of water per unit area of excess water in the capillary reservoir for time step m and
iteration n, in inches
Dfif Jm , n 77, 90 Fast interflow from preferential flow reservoirs per unit area of HRU J for time step m and
DfctHRU 44, 46, 47, 63a, 65 Maximum volume of water per unit area in the capillary reservoir, in inches
dfsubHRU 30 Decimal fraction of potential evapotranspiration that is sublimated from the snow surface,
dimensionless
m, n
DfupHRU 39 Volume of fast interflow per unit area into the capillary reservoir from all contributing HRUs at
m
Dhru PFR 38a Volume of water per unit area added to all connected preferential-flow reservoirs for time step m,
in inches
m1
Dhru PFR Volume of water per unit area added to all connected preferential-flow reservoirs for time step
m-1, in inches
Dimximper 31 Maximum retention storage for HRU impervious area, in inches
DM m
12, 13a, 14b, 14c Solar declination for time step m, angular degrees
DmxHRU 59 Maximum amount of gravity drainage from the gravity reservoir, in inches
DpftHRU 47, 48 Preferential-flow threshold as volume per unit area, in inches
m, n
DRCPR 63 Soil-water content ratio in the capillary reservoir for time step m and iteration n, dimensionless

m
Drem CPR 38b, 39 Volume of water per unit area of the pervious part of the HRU that infiltrates into the capillary
reservoir at time step m, in inches
m, n
DrplGVR 45 Volume of water per unit area that is removed from the gravity reservoir to replenish the capillary
reservoir during time step m and iteration n, in inches
m, n
Drpl GVR ,L 46 Volume of water per unit area that is removed from gravity reservoir L to replenish the capillary
reservoir during time step m and iteration n, in inches
m
DsHRU 24, 25 Total snowpack depth for the HRU during time step m, in inches
DsatHRU 47, 65 Maximum volume of water per unit area in the soil zone, in inches
m, n
Dsif HRU
57 Volume per unit area of slow interflow from the gravity reservoirs of the HRU for time step m,
m
DseHRU 24, 25 Snowpack liquid water equivalent depth for the HRU at time step m, in inches
m, n
DslupHRU 39 Volume of slow interflow per unit area into the capillary reservoir from all contributing HRUs at
m
DupHRU 34a Antecedent volume per unit area of water in the capillary reservoir that is available for evaporation
during time step m, in inches
DupmxHRU 34a Maximum volume per unit area of water in upper capillary reservoir, in inches
D() 70 Hydraulic diffusion coefficient, in length squared per time
e 19, 56 Exponential function constant (~2.7182818), dimensionless
E m
11 Obliquity of the Suns ecliptic for time step m, angular degrees
et 115b Budget error at time t, in cubic length per time
m, n
EAPR CPR
63 Actual to potential evapotranspiration ratio in the capillary reservoir for time step m and iteration
n, dimensionless
EC 11 Eccentricity of the Earths orbit (~0.01671), radians
eict 116d Volumetric flow rate of evaporation of intercepted precipitation at time t, in cubic length per time
eimt 116d Volumetric flow rate of evaporation from impervious areas at time t, in cubic length per time
ept 116d Volumetric flow rate of evapotranspiration from pervious areas at time t, in cubic length per time
espt 116d Volumetric flow rate of snowpack sublimation at time t, in cubic length per time
m
ET GVR
63e Actual evapotranspiration removed from the capillary reservoir for time step m and iteration n, in
inches per day
m
Evapimper 32, 33 Evaporation from the impervious part of the HRU for time step m, in inches
ext 116c Volumetric flow rate of ground-water discharge at time t, in cubic length per time
FJ ,sr 77 Decimal fraction of the total area of HRU J that contributes runoff and interflow to a particular
stream reach, dimensionless
FJ ,lakeHRU 90 Decimal fraction of the total area of HRU J that contributes runoff and interflow to a lake HRU,
dimensionless
f1uzf 112b Recharge received from the unsaturated zone that was stored in the unsaturated interval through
which the water table rose during time step m, in length cubed per time
f 2uzf 112b Recharge that flowed across the surface defined by the water table in the cell at time m-1, in cubic
length per time
facej+1/2 95, 96 Cell face between cells j and j+1, dimensionless
facej-1/2 95, 96 Cell face between cells j-1 and j, dimensionless
FcontribHRU,up 36 Decimal fraction of area in the upslope HRU that contributes Hortonian runoff to the downslope
HRU, dimensionless
FdenHRU 38a, 38b, 47 Decimal fraction of the soil zone available for preferential flow, dimensionless
fdpth 87 Empirical exponent determined from regression methods, dimensionless
Appendix 2 231

Minimum possible contributing area for pervious runoff, as a decimal fraction of HRU area,
FmnHRU 34a
dimensionless
Maximum possible contributing area for pervious runoff, as a decimal fraction of HRU area,
FmxHRU 34a, 34b
dimensionless
m
FpervHRU Decimal fraction of the surface-runoff-contributing area of the pervious parts in the HRU for
34a, 34b, 35
time step m, dimensionless
Frain mHRU Decimal fraction of total precipitation occurring as rain on an HRU for time step m,
6
dimensionless
fstcoeflin 67a Linear flow-routing coefficient for fast interflow, per day
fstcoefsq 67a Nonlinear flow-routing coefficient for fast interflow, in day per inch
gzt 116a Volume of water in gravity reservoirs of the soil zone at time t, in cubic length
gzt-1 116b Volume of water in gravity reservoirs of the soil zone at time t-1, in cubic length
h 94, 95, 96, 97 Potentiometric head, in length
h 105 Vector of ground-water heads for all variable-head cells, in length
hi,j-1,k 101 Head at node i, j1, k, length
hi,j,k 101, 112a Head at node i, j, k, in length
h m
i, j,k
103, 104, 111, 112b Head at node i, j, k, at end of time step m
him1, j , k 104 Head at node i-1, j, k, at end of time step m
him+1, j , k 104 Head at node i+1, j, k, at end of time step m
him, j 1, k 104 Head at node i, j-1, k, at end of time step m
him, j +1, k 104 Head at node i, j+1, k, at end of time step m
him, j , k 1 104 Head at node i, j, k-1, at end of time step m
him, j , k +1 104 Head at node i, j, k+1, at end of time step m
m, n
haqfdc 92 Ground-water head in the finite-difference cell adjacent to the lake at the cell node for time step m
and iteration n, in length
h mfdc, n 41, 62, 78 Ground-water head in the finite-difference cell for time step m and iteration n, in length
h mfdc, n 1 41 Ground-water head in the finite-difference cell for time step m, iteration n-1; in length
m, n
hlake 92, 93 Lake stage for time step m and iteration n, in length
m1
hlake 93 Lake stage for time step m-1, in length
m , n 1
hafdc , lc
93 Ground-water head in finite-difference cell associated with cell face lc for time step m and
iteration n-1, in length
m, n
hstr 78 Stream head at the midpoint of the stream reach for time step m and iteration n, in length
HCHRU 18 Hamon monthly air-temperature coefficient, in inch-cubic meter per gram

m
Hc HRU 28 Convective or sensible heat at air-snow interface during time step m, in calories
m
HeHRU 28 Latent heat (sublimation and condensation) at the air-snow interface during time step m, in
calories
HF 29 Specific latent heat of fusion to melt 1-inch of water-equivalent ice at 0Celsius, 203.2 calories per
inch
m
Hg HRU 28 Heat gained from the ground during time step m, in calories
m
HmHRU 28, 29 Energy available for snowmelt during time step m, in calories

m
Hp HRU 28 Heat gained from precipitation during time step m, in calories
m
HqHRU 28 Heat required for internal state change during time step m, in calories
m
HsHRU 28 Energy gained due to shortwave radiation during time step m, in calories
m
Ht HRU 25 Daily heat transferred from the surface layer to the lower layer of the snowpack during time step
m, in calories per square centimeter
i 68, 69b, 69c, 76a, 76b Evapotranspiration rate beneath the soil-zone base per unit depth, in length per time per length
ict 116a Volume of intercepted precipitation in plant canopy reservoirs at time t, in cubic length
ict-1 116b Volumes of intercepted precipitation in plant canopy reservoirs at time t-1, in cubic length
imt 116a Volume of water in impervious reservoir at time t, in cubic length
imt-1 116b Volume of water in impervious reservoir at time t-1, in cubic length
in 114 Total inflow to the model, cubic length per time
int 115b, 116c Total inflow at time t, in cubic length per time
J 77, 90 Counter for the HRU number, dimensionless
jd 11 Julian day number (3 is subtracted as the solar year begins on December 29), days
JHHRU 20a Jensen-Haise air-temperature coefficient for the HRU, in degrees
JHmonth 20a Monthly Jensen-Haise air-temperature coefficient, per degrees Fahrenheit
JJ 77, 90 Total number of HRUs that contribute surface runoff and interflow to a particular stream reach
(77) or lake (90)
k 57 Total number of gravity reservoirs in the HRU, dimensionless
Kaq 92 Horizontal or vertical hydraulic conductivity of aquifer finite-difference cell adjacent to lake cell,
in length per time
Klkbd 92 Hydraulic conductivity of the lakebed, in length per time
Ks 61a, 61b, 74 Vertical saturated hydraulic conductivity of the unsaturated zone, in length per time
Kstrbed 78, 89 Hydraulic conductivity of the streambed, in length per time
Kv 41, 108 Vertical hydraulic conductivity of the finite-difference cell, in length per time
m
ke HRU
25, 26 Effective thermal conductivity of the snowpack, in calories per second-gram-degree Celsius
Kxx 94, 95, 96 Hydraulic conductivity tensor aligned with the x coordinate axes, in length per time
Kyy 94 Hydraulic conductivity tensor aligned with the y coordinate axes, in length per time
Kzz 94 Hydraulic conductivity tensor aligned with the z coordinate axes, in length per time
K() 68, 69a, 70, 71, 74, 75 Unsaturated vertical hydraulic conductivity as a function of water content and is equal to the
vertical flux, in length per time
K(z ) 71, 73, 76a Unsaturated vertical hydraulic conductivity at depth z1, in length per time
1
K(z ) 71, 73, 76a Unsaturated vertical hydraulic conductivity at depth z2, in length per time
2
KRj-1/2 102 Hydraulic conductivity between nodes i, j1, k and i, j, k, in length per time
L 46, 60, 66 Counter for the gravity reservoir number, dimensionless
lat 13a Latitude of the horizontal surface (basin centroid, HRU or equivalent-slope surface), positive
values are in the northern hemisphere, negative values are in the southern hemisphere, in
radians
lat'HRU 14b, 14c Latitude of the equivalent-slope surface of the HRU, in radians
lc 93 Counter used to refer to individual finite-difference cell faces in contact with the submerged
lakebed
LC 93 Number of finite-difference cell faces in contact with the submerged lakebed
Ldisti 5b Inverse distance between HRU centroid and measurement station i
Appendix 2 233

Lengthfdc 106 Length of the prism, in length
lengthstr 78, 89 Length of the stream reach, in length
LL 46, 60, 66 Total number of gravity reservoirs, dimensionless
long'HRU 14c Longitude offset between the equivalent-slope surface and the HRU, in radians
npstas 10a Number of precipitation stations, dimensionless
ntstas 4, 5a Number of air temperature measurement stations, dimensionless
out 114 Total outflow from the model, cubic length per time
outt 115b, 116d Total outflow at time t, in cubic length per time
m
P HRU
7, 9a, 10a, 22, 23 Precipitation at the HRU during time step m, in inches
m
PlakeHRU 91 Precipitation on the lake HRU for time step m, in inches
Pstam 7, 10a Measured precipitation at the station during time step m, in inches
pt 116c Volumetric flow rate of precipitation on modeled region at time t, in cubic length per time
P base 8 Mean monthly precipitation at the base station, in inches per day
P lapse 8 Mean monthly precipitation at the lapse station, in inches per day
P sta , month 10c, 10d Mean monthly precipitation at each measurement station, length
m
P stas 9c Mean precipitation of all stations during time step m, in inches
pwt 116a Volume of water in snowpack storage at time t, in cubic length

pwt-1 116b Volume of water in snowpack storage at time t-1, in cubic length
pafbase 8 Mean monthly factor used to adjust rain or snow lapse rate (usually 1.0), dimensionless
Pancoefmonth 21 Monthly pan-evaporation coefficient, dimensionless
m
Panevap sta
21 Pan evaporation at each measurement station for time step m, length
m
PETHRU 18, 20a, 21, 30, 32, 63e Potential evapotranspiration for the HRU during time step m, in inches
pmixafmonth 6 Monthly rain adjustment factor for a mixed precipitation event (usually 1.0), dimensionless
pptadjHRU 15b Precipitation-day adjustment factor, dimensionless
m
Pnet HRU
23, 31, 32, 33, 34b, Precipitation that reaches the ground during time step m, in inches
35, 37
m
PsHRU 24 Net snowfall rate, as a liquid water equivalent, on the HRU during time step m, in inches per day
m
Ptf HRU 22, 23 Precipitation throughfall on the HRU during time step m, in inches
m, n
qgd , pot 59 Potential gravity drainage per unit area for time step m and iteration n, in inches per day
m, n
qGVR , in
49, 51, 52, 56 Volumetric inflow rate per unit area to the gravity reservoir, in inches per day
m, n
qGVR , sif
49, 50, 56, 58 Slow interflow rate per unit area from the gravity reservoir, in inches per day
qim, sif, n 57 Slow interflow from gravity reservoir i for time step m, iteration n, in inches per day.
m, n
qgdc , net
61a, 61b, 62 Net gravity drainage to the connected finite-difference cell for time step m, and iteration n, in
length per time
m , n*
qgdc , net
62 Net gravity drainage to the saturated zone for time step m and iteration n, in length per time
m, n
qgdc , pot
60 Potential gravity drainage for time step m and iteration n from all gravity reservoirs connected to a
finite-difference cell, in length per time

m, n
q gd , L 60 Potential gravity drainage per unit area of gravity reservoir L for time step m and iteration n, in
inches per day
m , n 1
qgd ,L
60 Potential gravity drainage per unit area of gravity reservoir L for time step m and iteration n-1, in
inches per day
Qin 113 Total inflow to the modeled region during a time step, in cubic length per time
qi+1/2 98, 99 Volumetric flow rate between cells i and i+1, in cubic length per time
qi-1/2 98, 99 Volumetric flow rate between cells i-1 and i, in cubic length per time
qj+1/2 96, 98, 99 Volumetric flow rate between cells j and j+1, in cubic length per time
qj-1/2 96, 98, 99, 101 Volumetric flow rate between cells j-1 and j, in cubic length per time
qk+1/2 98, 99 Volumetric flow rate between cells k and k+1, in cubic length per time
qk-1/2 98, 99 Volumetric flow rate between cells k-1 and k, in cubic length per time
QL 99, 100 Volumetric flow rate into a cell (negative value is a volumetric flow out of cell), in cubic length
per time
m, n
QlakeHRUtolakemf 91 Volumetric flow rate from a lake HRU to a lake represented in MODFLOW-2005 for time step m
and iteration n, in cubic length per time
m, n
QlakeHRUtomf 93 Sum of precipitation, evaporation, surface runoff, and interflow calculated in PRMS, in cubic
length per time
m, n
Qlakeleak 92 Volumetric-flow rate across the lakebed to center of aquifer finite-difference cell for time step m
and iteration n and is outward or downward leakage when positive and ground-water discharge
when negative, in cubic length per time
m, n
Qlateral 77, 79, 84, 88 Volumetric flow rate into a stream reach from all connected HRUs for time step m and iteration n,
in cubic length per time
m, n
Qmdpt 84, 85, 86, 87 Volumetric-flow rate at the midpoint of a stream reach for time step m and iteration n, in cubic
length per time
Qout 113 Total outflow from the modeled region during a time step, in cubic length per time
m, n
Q mdpt
84 Volumetric-flow rate at the midpoint of a stream reach for time step m and iteration n, in cubic
length per time m
m, n
Qsrdvr 79, 84 Volumetric flow rate diverted from the end of the stream reach for time step m and iteration n, in
cubic length per time
m, n
Qsrevp 79, 84, 88 Specified evaporation rate from the stream surface multiplied by the wetted plan-view area for
time step m and iteration n, in cubic length per time
m
Qsrin 79, 84, 88 Specified volumetric flow rate at the beginning of a stream reach for time step m, in cubic length
per time
m, n
Qsrleak 78, 84, 89 Volumetric-flow rate across the streambed for time step m and iteration n and is downward
leakage when positive and ground-water discharge to the stream when negative, in cubic length
per time
m, n
Qsrout 79 Volumetric flow rate from the end of a stream reach for time step m and iteration n, in cubic length
per time
m, n
Qsrpp 79, 84, 88 Specified precipitation rate on the stream surface multiplied by the wetted plan-view area for time
step m and iteration n, in cubic length per time
m, n
Qsrup 79, 84, 88 Sum of the volumetric flow rate that enters the stream reach from outflow of upstream reaches for
time step m and iteration n, in cubic length per time
m, n
Qsrleak , mx 88 Total amount of water available for stream leakage to ground water for time step m and iteration n,
in cubic length per time
m, n
Qtotstrin 93 Sum of all tributary stream inflows for time step m and iteration n, in cubic length per time
m, n
Qtotstrout 93 Sum of lake outflows to streams for time step m and iteration n, in cubic length per time
Appendix 2 235

m
Qwd 93 Specified withdrawal from a lake (negative value adds water) for time step m, in cubic length per
time
msfr
Qdwn 83 Volumetric flow at the downstream end of the stream reach for the current streamflow-routing
time step msfr, in cubic length per time
msfr 1
Qdwn 83 Volumetric flow at the downstream end of the stream reach for the previous streamflow-routing
time step msfr-1, in cubic length per time
Qupmsfr 83 Volumetric flow at the upstream end of the stream reach for the current streamflow-routing time
step msfr, in cubic length per time
Qupmsfr 1 83 Volumetric flow at the upstream end of the stream reach for the previous streamflow-routing time
step msfr-1, in cubic length per time
qstr 80 Sum of all lateral inflows and outflows per unit length of stream listed on the right side of
equation 79, in length squared per time
QW 105 Volumetric flow rate added to variable-head cells that represents external sources and sinks to the
ground-water system, in cubic length per time
gift 116d Volumetric flow rate of interflow leaving modeled region at time t, in cubic length per time
m
qce HRU
30 Evaporation loss from interception storage for the HRU during time step m, in inches
m
qsi perv 37, 38a, 38b Soil infiltration over the pervious part of the HRU for time step m, in acre-inch
qsnmxHRU 37 Daily maximum snowmelt infiltration into the soil zone for the HRU, in inches
qsot 116d Volumetric flow rate of surface runoff leaving the modeled region at time t, in cubic length per
time
Rhydraulic 85 Hydraulic radius of the stream, which is equal to the stream area divided by the wetted perimeter,
in length
rad 11 Revolution speed of the Earth (~0.0172), radians per day
m
RahHRU 15a, 15b, 16b, 17 Measured or computed as the horizontal plane shortwave radiation on the HRU during time step
m, in calories per square centimeter per day
Rain HRU , month 10c Mean monthly rain on each HRU that can be obtained from National Weather Service's spatial
distribution of mean annual precipitation for the 19712000 climate normal period, length
m
RaswHRU 15b, 17, 20a Computed shortwave radiation on the HRU during time step m, in calories per square centimeter
per day
m, n
ROd HRU 65, 66 Volume of Dunnian runoff per unit area from an HRU for time step m and iteration n, in inches
ROd Jm , n 77, 90 Dunnian runoff per unit area from HRU J for time step m and iteration n, in inches
ROd Lm , n 66 Dunnian runoff from gravity reservoir L for time step m, iteration n, in inches
m, n
ROdupHRU 39 Volume of Dunnian runoff per unit area into the capillary reservoir from all contributing HRUs at
m
ROhHRU , dwn
36 Volume per unit area of Hortonian runoff to a downslope HRU or stream segment during time step
m, in inches
m
ROhimper 31, 32, 33 Hortonian runoff from the impervious part of the HRU per unit area during time step m, in inches
m
ROhimpr , up 36 Runoff per unit area from the impervious part of the upslope HRU for time step m, in inches
ROhJm , n 77, 90 Hortonian runoff per unit area from HRU J for time step m and iteration n, in inches
m
ROhperv 35, 37 Runoff per unit area from the pervious part of the HRU for time step m, in inches
m
ROhperv , up
36 Runoff per unit area from the pervious part of the upslope HRU for time step m, in inches
m
ROhupHRU 31, 32, 33, 35, 37 Sum of Hortonian runoff from all upslope contributing HRUs as a volume per unit area of the
HRU for time step m, in inches

Roughness 85, 86 Mannings roughness coefficient, dimensionless
m
RspHRU 14a, 15a, 16b Potential solar radiation on the HRU during time step m, in calories per square centimeter per day
SS 94, 97 Specific storage of the porous rock, per length

St 115a, 116a Storage at time t, in cubic length
St-1 115a, 116a Storage at the previous time step, in cubic length
Sy 61a Specific yield that is used to approximate s-r, dimensionless
sc m
14a 60-minute period solar constant for time step m, in calories per square centimeter per hour
sh m 13, 14b, 18 Daylight length on the HRU for time step m, hours
HRU
Slope0 81, 85, 86 Slope of the channel in the longitudinal profile, dimensionless
Slopef 81 Friction slope of the channel in the longitudinal profile, dimensionless
slopeHRU 15b, 17 Slope of the HRU, dimensionless
slwcoeflin 50, 51, 52, 53, 54, 56 Linear flow-routing coefficient for slow interflow, per day
slwcoefsq 50, 51, 52, 53, 54, 56 Nonlinear flow-routing coefficient for slow interflow, in day per inch
smt 116a Volume of water in capillary reservoirs of the soil zone at time t, in cubic length
smt-1 116b Volume of water in capillary reservoirs of the soil zone at time t-1, in cubic length
SmcHRU 34b Coefficient used to calculate decimal fraction of pervious surfaces, dimensionless
SmexHRU 34b Coefficient used to calculate the decimal fraction of pervious surfaces, per inch
m
SmidxHRU 34b Soil moisture index of the capillary reservoir for time step m, in inches
SnowHRU , month 10d Mean monthly snow on each HRU that can be obtained from National Weather Service's spatial
distribution of mean annual precipitation for the 19712000 climate normal period, length
m
SpcHRU 22 Storage in the plant canopy (summer or winter) on the HRU during time step m, in acre-inch
m
SpcaHRU 22 Available storage in the plant canopy of the HRU during time step m, in acre-inch
SpcmxHRU 22 Maximum storage in the plant canopy (snow, winter rain, or summer rain) of the HRU, in inches
srm 13b Hour angle of sunrise, measured from the local meridian of a horizontal surface (HRU or
equivalent-slope surface) for time step m, radians
m
srHRU 13c, 14c Hour angle of sunrise, measured from the local meridian of a horizontal surface (HRU or
equivalent-slope surface) for time step m, radians
SS 97, 98, 100, Volume averaged specific storage, per length
SSi,j,k 104, 111 Volume averaged specific storage for cell i, j, k, per length
ss m
13a, 13b Hour angle of sunset, measured from the local meridian of a horizontal surface (HRU or
equivalent-slope surface) for time step m, in radians
m
ssHRU 13c, 14c Hour angle of sunset on the sloped surface of the HRU for time step m, in radians
m
subHRU 30, 32 Sublimation from the HRU during time step m, in inches
t 68, 69a, 69b, 70, 71, 73, Time

80, 82, 94, 97
t*m 103, 111 Time at the end of time step m
t*m 1 103, 111 Time at the end of time step m-1
m
Tbase 2 Measured maximum (or minimum) daily temperature at the base station assigned to an HRU for
time step m, degrees Fahrenheit or Celsius
Appendix 2 237

m
THRU 1, 2, 3a, 5a Maximum (or minimum) daily temperature at each HRU for time step m, in degrees Fahrenheit or
Celsius
Ti m , Ti+1m 4, 5a Measured maximum (or minimum) temperature at measurement station i and i+1 for time step m,
degrees Fahrenheit or Celsius
m
Tlapse 4, 5a Basin-average maximum (or minimum) temperature lapse rate for time step m, degrees Fahrenheit
or Celsius per length
m
Tlower 25 Temperature of the snowpack lower layer during time step m, in degrees Celsius
Tstam 1 Measured maximum (or minimum) daily temperature at the station for time step m, in degrees
Fahrenheit or Celsius
m
Tsurf 25 Temperature of the snowpack surface layer during time step m, in degrees Celsius
THRU 19, 20a, 20b Mean daily temperature on the HRU, in degrees Celsius
m
T stas 3c Mean measured maximum or (minimum) daily temperature of all stations for time step m, in
degrees Fahrenheit or Celsius
tafHRU 1, 2, 3a, 5a Maximum (or minimum) daily HRU temperature adjustment factor, which is estimated on the
basis of slope and aspect, in degrees Fahrenheit or Celsius
thckCB 110 Thickness of the semi-confining unit, in length
thcki,j,k 109, 110 Thickness of cell i, j, k, in length
thcki,j,k+1 109, 110 Thickness of cell i, j, k+1, in length
thickstrbed 78, 89 Streambed thickness in the stream reach, in length
thickaq 92 Horizontal or vertical distance from lakebed to the center of the adjacent finite-difference cell, in
length
thicklkbd 92 Lakebed thickness, in length
Tmn mHRU 6, 16a Minimum air temperature assigned to the HRU for time step m, degrees Fahrenheit or Celsius
Tmx mHRU 6, 16a Maximum air temperature assigned to the HRU for time step m, degrees Fahrenheit or Celsius
TmxsnowHRU 6 Monthly maximum air temperature at which precipitation is all snow for the HRU, degrees
Fahrenheit or Celsius
Tr 106 Transmissivity of the cell, in length squared per time
us 73 Velocity of a sharp wetting front, in length per time
V 100, 104 Volume of the cell, in cubic length
m, n
V lakeHRU
90 Volume of surface runoff and interflow routed to a lake HRU from contributing HRUs for time
step m and iteration n, in acre-inch
v() 69a, 69c, 75 Characteristic velocity restricted to the downward (positive z) direction, in length per time
VKi,j,k 109, 110 Vertical hydraulic conductivity of cell i, j, k, in length per time
VKi,j,k+1 110 Vertical hydraulic conductivity of cell i, j, k+1, in length per time
VKCBi , j , k 110 Vertical hydraulic conductivity of the semi-confining unit beneath cell i, j,k, in length squared per
time
m
VsmHRU 29, 31, 32, 33, 37 Volume of snowpack melted during time step m, in acre-inch
w 60 Weighting factor used to average gravity drainage between iterations, set to 0.5
W 94, 98, 104 Volumetric flow rate per unit volume representing sources and/or sinks of water, W < 0.0 for flow
out of the ground-water system, and W > 0.0 for flow into the system, per time
m, n
wmdpt 87 Stream width at reach midpoint, in length
wstr 86 Steam width, in length


wetperstr 78, 89 Wetted perimeter of the streambed, in length
Widthfdc 106 Width of the prism perpendicular to the direction for which the conductance is defined, in length
x 94, 95, 96, 97 Horizontal coordinate axis normal to the y and z coordinate axis, in length
xchannel 80, 83 Distance along the stream channel, in length
XHRU 5b, 10b Longitude of the HRU centroid, length
X HRU 3b, 9b Normalized longitude of the HRU, dimensionless
X sta 3c, 9c Mean normalized longitude of all stations, dimensionless
Xppti 10b Longitude of each precipitation measurement station, length

Xtempi 5b Longitude of each air temperature measurement station, length
y 94, 95, 97 Horizontal coordinate axis normal to the x and z coordinate axis, in length
YHRU 3b, 9b Normalized longitude of the HRU, dimensionless
Y HRU 3b, 9b Normalized latitude of the HRU, dimensionless
Y sta 3c, 9c Mean normalized latitude of all stations, dimensionless
Yppti 10b Latitude of each precipitation measurement station, in length

Ytempi 5b Latitude of each air temperature measurement station, in length
z 68, 69a, 70, 71, 72, 94, Altitude in the vertical direction, in length
95, 97
z1 70, 71, 72, 76a Point above the wetting front, in length
z2 70, 71, 72, 76a Point below the wetting front, in length
Zbase 2, 8 Altitude of the base station, in length
zf 72, 73 Depth of the sharp front, in length
ZHRU 1, 2, 5a, 8 Mean land-surface altitude of the HRU, in length
Z tsta ,i , Z tsta ,i +1 4, 5a Altitude of measurement stations i and i+1, length
Zlapse 2, 8 Altitude of the lapse station, in length

3c, 9c Mean normalized altitude of all stations, dimensionless
Z tsta
ci 95, 96, 97, 98, 102, 109, Length of finite difference cell along the column direction, in length
110, 111
h
t* 97, 98, 100 Change in head over a specified time interval
him, j , k
103 Change in head at cell i, j, k at the end of time step m over a specified time interval
t*
rj 95, 97, 98, 109, 110, 111 Length of finite difference cell along the row direction, in length
rj-1/2 102 Distance between nodes i, j1, k and i, j, k, in length
st 115a, 115b Total storage change of the time step ending at time t, in cubic length
Stotal 113 Total storage change in all reservoirs from beginning to end of time step, in cubic length
storage 114 Total storage change, in cubic length
tsnow 25 Snow computation time step, in 43,200 seconds (half-day interval)
t 40, 42, 55, 56, 58, 67, 76b, GSFLOW time step, in one day
77, 91, 113, 114
tmf 93 MODFLOW-2005 time step, in time
Appendix 2 239

tsfr 82 Time-step for routing streamflow, in time
vk 95, 96, 97, 98, 102, 111 Length of finite difference cell along the layer (vertical) direction, in length
V 111 Rate of accumulation of water in the cell, in cubic length per time
t
x 83 Spatial increment for routing streamflow, equal to the reach length
z 108 Cell thickness, or, for unconfined conditions, the thickness of the saturated region of the cell, in
length
75 Change in water content between two adjacent locations along a trailing wave, volume of water
per volume of rock
61a, 74 Brooks-Corey exponent, dimensionless
68, 69a, 69b, 70, 71, 72, Volumetric water content, in volume of water per volume of rock
74
m,n 61a, 61b Volumetric water content at the top of the unsaturated zone for time step m, and iteration n,
dimensionless
z 72, 73 Volumetric water content at depth z1 in volume of water per volume of rock
1
z 72, 73 Volumetric water content at depth z2, in volume of water per volume of rock
2
76b Water content of trailing waves above the evapotranspiration extinction depth after time , in
volume of water per volume of rock
+t 76b Water content of trailing waves above the evapotranspiration extinction depth after time +, in
volume of water per volume of rock
r 61a, 74 Residual water content of the unsaturated zone, in volume of water per volume of rock
s 61b, 74 Saturated water content of the unsaturated zone, in volume of water per volume of rock
HRU 20a, 20b Latent heat of vaporization on the HRU for time step m, in calories per gram
13c, 14c, 25 Constant pi (~3.1415926535898), dimensionless
HRU 18, 19 Saturated water-vapor density (absolute humidity), in grams per cubic meter
'HRU 22, 23 Plant canopy density (winter or summer) as a decimal fraction of the HRU area, dimensionless
s 24 Density of new-fallen snow, as a decimal fraction, dimensionless
m
snow HRU 25 Volumetric density fraction of the snowpack for the HRU during time step m, dimensionless
"m 26, 27, 30 Volumetric density fraction of the snowpack for the HRU during time step m, dimensionless
HRU
sm 24 Average maximum snowpack density, as a decimal fraction of the liquid water equivelent,
dimensionless
83, 93 Time-weighting factor that ranges between 0.5 and 1, dimensionless
This page left intentionally blank

Markstrom and others GSFLOWFlow Model Based on the Integration of the PRMS and the MODFLOW-2005Techniques and Methods 6D1

GSFLOW

Uploaded by

Document Informationclick to expand document informationEs un modelo integrado de precipitación y escorrentía superficial con flujo de agua sunterránea

Document Informationclick to expand document information

Copyright:

Available Formats

GSFLOW

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

GSFLOW

Uploaded by

Copyright:

Available Formats

GSFLOWCoupled Ground-Water and Surface-Water Flow

Model Based on the Integration of the Precipitation-Runoff

Unsaturated Gravity drainage Precipitation

Techniques and Methods 6D1

U.S. Department of the Interior

By Steven L. Markstrom, Richard G. Niswonger, R. Steven Regan, David E. Prudic,

Techniques and Methods 6D1

U.S. Department of the Interior

U.S. Geological Survey

U.S. Geological Survey, Reston, Virginia: 2008

For product and ordering information:

Conversion Factors and Datums

Office of Ground Water

This page intentionally left blank.

Abstract estimate recharge using long-term average precipitation and

Unsaturated Gravity drainage Precipitation

Evaporation Sublimation Air temperature

Soil-Zone Reservoir Impervious-Zone Reservoir

Figure 3. Components of the snowpack energy balance, accumulation, snowmelt, and

Watershed DiscretizationHydrologic Response results in regular rectangles; or (3) a noncontiguous approach,

6 Hydrologic response unit and number

1 Stream segment and number

Streamflow gage at outlet

Time DiscretizationTime Steps utility software for optimization, sensitivity, forecasting,

low-permeability regions below land

MODFLOW-2005 (Harbaugh, 2005) is a new version

Water Well discharge

Subscript indicates the number of the row

Figure 7. A discretized hypothetical aquifer system (modified from Harbaugh, 2005).

Streams are divided into reaches and segments COLUMNS

Organization Processes are divided into packages. A package is defined

Allocate and Read

Read Boundary Conditions

Compute Flow and Storage

Save Results to File

Check Time Step;

Check Stress Period;

Region 1Plant canopy, Surface runoff

Ground-water discharge Ground-water discharge

Gravity drainage Leakage

Hydrologic or New or Model

Hydrologic or New or Model

Hydrologic or New or Model

Table 2. PRMS modules and MODFLOW-2005 packages not implemented in GSFLOW.

[HRU, hydrologic response unit]

A. Original PRMS soil-zone water B. Revised PRMS soil-zone water

Wilting threshold Saturation threshold

Field-capacity threshold Preferential-flow threshold

Field-capacity threshold Preferential-flow threshold

Ground-water reservoir Ground-water

Capillary reservoir Gravity Gravity

Ground-water reservoir Ground-water

Figure 13. Conceptual changes to the soil-zone, shallow-subsurface, and ground-

1 Land-based hydrologic response unit and parameter hru_type designation number

2 Lake hydrologic response unit and parameter hru_type designation number

Streamflow gage at outlet

Hydrologic response unit boundary

1 Stream reach number