Environmental Engineering Reference
In-Depth Information
and ecology. Results of watershed models are
used in resource management decisions, risk
and uncertainty analyses, and environmental
impact studies.
Among the more widely used watershed
models are SWAT (Arnold
et al
.,
1998
), PRMS
(Leavesley
et al
.,
1983
), and MIKE-SHE (Graham
et al
.,
2006
). Singh (
1995
) and Singh and Frevert
(
2006
) provided reviews of these and many
other watershed models. Watershed models
are all based on a water-budget equation, but
there are important differences among models
in terms of spatial and temporal scales of appli-
cability, processes that are accounted for (e.g.
interception, runoff, infiltration, evapotranspi-
ration, water movement through the unsatu-
rated zone, and groundwater flow), techniques
for representing those processes (e.g. physically
based vs. empirical), data input requirements,
and output options. Groundwater recharge and
discharge are components of a watershed water
budget; hence, watershed models can provide
estimates for these components. Applications of
watershed models to estimate recharge include
Weeks
e t a l
. (
1974
), Arnold and Allen (
1996
), Sami
and Hughes (
1996
), Bauer and Mastin (
1997
),
Desconnets
et al
. (
1997
), Laenen and Risley
(
1997
), Zhang
et al
. (
1999
), Arnold
et al
. (
2000
),
Steuer and Hunt (
2001
), Eckhardt and Ulbrich
(
2003
), and Cherkauer (
2004
).
A watershed model assumes a series of res-
ervoirs (
Figure 3.6
), representing water storage
on land surface, in the plant canopy, within the
unsaturated zone (soil zone and subsurface),
and in groundwater. These reservoirs fill and
drain in response to atmospheric phenomena
(e.g. precipitation, snowmelt, evaporation, and
transpiration) and to the current state of other
reservoirs. The complexity with which water
movement within a reservoir or among dif-
ferent reservoirs is represented varies among
models. In the simplest case, these reservoirs
can be thought of as tanks, buckets, or layers, as
depicted in
Figure 3.2
, that fill by precipitation
or drainage from other tanks. Drainage from
tanks can be at a constant rate, a rate depend-
ent on the height of water currently in the
tank, or a rate dependent on external param-
eters, such as air temperature and humidity.
More sophisticated representations of processes
include routing of surface-water flow over land
or through channels and surface reservoirs
and simulating water movement through the
unsaturated zone by an approach based on the
Richards equation.
Climatic and hydraulic data requirements
for watershed models are similar to those of soil
water-budget models (
Section 3.3
). Minimum
data requirements typically include average
daily precipitation and temperature (for esti-
mating potential evapotranspiration) and soil
water-storage capacity. Additional information
(such as a digital elevation map, storage and
conveyance properties of surface-water bod-
ies, depth to water table, hydraulic properties
of unsaturated and saturated zone materials,
and types of vegetation) can be used by many
models. Recent advances allow watershed mod-
els to work with GIS datasets, thus facilitating
a highly distributed representation of param-
eters within a watershed. Watershed models
have historically been calibrated by comparing
simulated streamflow with measured stream-
flow. It is not uncommon in current studies to
include measurements of evapotranspiration,
water storage in the unsaturated zone, and
groundwater levels in the calibration process.
An attractive aspect of using a watershed
model to estimate recharge is the power to
predict how future changes in climate and
land use may affect recharge patterns (Zhang
et al
.,
1999
; Steuer and Hunt,
2001
). Physically
based models can provide insight into mecha-
nisms that influence the hydrologic response
of a basin. Disadvantages to using watershed
models include the need for training in model
use and the large number of parameters that
may need to be determined or estimated. The
common assumption made in many applica-
tions of watershed models that groundwater-
flow boundaries coincide with surface-water
flow boundaries (it is often assumed that there
is no groundwater inflow to or outflow from the
watershed) should be fully evaluated. Because
watershed models are water-budget methods,
the caveats expressed in
Chapter 2
deserve
consideration. Most importantly, if the mag-
nitude of recharge is only a small fraction of
