Soil Erosion and Conservation - Environmental Modelling: Finding Simplicity in Complexity - page 353

On the other hand, for the purposes stated above for
which an erosion model is used, the USLE has been, and
still can be, very successful. This issue is also discussed
below in more detail.
The USLE was upgraded to the Revised Universal Soil
Loss Equation (RUSLE) during the 1990s (Renard et al .,
1997). This is a hybridmodel. Its basic structure is themul-
tiplicative formof the USLE, but it also has many process-
based auxiliary components. It is computer based, and
has routines for calculating time-variable soil erodibility,
plant growth, residue management, residue decomposi-
tion, and soil surface roughness as a function of physical
and biological processes. The RUSLE also has updated val-
ues for erosivity ( R ), new relationships for L and S factors
which include ratios of rill and interrill erosion, and addi-
tional P factors for rangelands and subsurface drainage,
among other improvements. The RUSLE has the advan-
tage of being based on the same extensive database as is
the USLE, with some of the advantages of process-based
computations for time-varying environmental effects on
the erosional system. It still has the limitations, how-
ever, in model structure, which allows only for limited
interactions and interrelationships between the basicmul-
tiplicative factors of the USLE (Equation 22.1).
Various process-based erosion models have been
developed since the mid-1990s, including EUROSEM in
Europe (Morgan et al ., 1998), the GUEST model in Aus-
tralia (Misra and Rose, 1996), and the WEPP model in
the United States (Flanagan and Nearing, 1995). We will
focus here on the example of the WEPP model, largely
because it is the technology most familiar to the author.
The WEPP profile computer model includes seven
major components, including climate, infiltration, water
balance, plant growth and residue decomposition, surface
runoff, erosion, and channel routing for watersheds. The
climate component of the profile computermodel (Nicks,
1985) generates daily precipitation, daily maximum and
minimum temperature, and daily solar radiation based on
a statistical representation of weather data at a particular
location. The climate model has been tested for erosion
and well parameterized for the United States (Baffaut
et al ., 1996). The infiltration component of the hillslope
model is based on the Green and Ampt equation, as
modified by Mein and Larson (1973), with the ponding
time calculation for an unsteady rainfall (Chu, 1978).
The water balance and percolation component of the
profile model is based on the water balance compo-
nent of SWRRB (Simulator for Water Resources in Rural
Basins) (Williams and Nicks, 1985; Arnold et al ., 1990),
with some modifications for improving estimation of

percolation and soil evaporation parameters. The plant-
growth component of the model simulates plant growth
and residue decomposition for cropland and rangeland
conditions. The residue- and root-decomposition model
simulates decomposition of surface residue (both stand-
ing and flat), buried residue, and roots for the annual
crops specified in the WEPP User Requirements (Flana-
gan and Livingston, 1995) plus perennial crops of alfalfa
and grasses. Surface runoff is calculated using a kinematic
wave equation. Flow is partitioned into broad sheet flow
for interrill erosion calculations and concentrated flow for
rill erosion calculations. The erosion component of the
model uses a steady-state sediment continuity equation
that calculates net values of detachment or deposition
rates along the hillslope profile (Nearing et al ., 1989). The
erosion process is divided into rill and interrill compo-
nents where the interrill areas act as sediment feeds to the
rills, or small channel flows. The model is applicable to
hillslopes and small watersheds.
Because the model is based on all of the processes
described above, and more, it is possible with WEPP
to have an enormous array of possible system inter-
actions represented in the simulations. Just to name a
very few examples, slope-length and steepness effects are
functions of soil consolidation, surface sealing, ground
residue cover, canopy cover, soil water content, crop
type and many other factors. Ground residue cover is a
function of biomass production rates, tillage implement
types, residue type, soil moisture, temperature and solar
radiation, previous rainfall, and many other factors. Rill-
erosion rates are a function of soil-surface roughness,
ground cover, consolidation of the soil, soil physical and
chemical properties, organic matter, roots, interrill ero-
sion rates, slope, and runoff rates, among other factors.
The lists continue ad infinitum . These are interactions that
simply cannot be represented with an empirical model.
The WEPP is a very complex model in this sense.
The disadvantage of the process-based model is also
the complexity of the model. Data requirements are huge,
and with every new data element comes the opportunity
to introduce uncertainty, as a first-order error analysis
would clearly indicate. Model-structure interactions are
also enormous in number, and with every structural
interaction comes the opportunity for error, as well
(see also Chapter 15). In a sense, the goal in using the
process-based model is to capture the advantages of
the complexity of model interactions, while gaining the
accuracy and dependability associated with the simpler
empirically based model. This goal can be achieved, and
was achieved with the WEPP model, using a combination