Geology Reference
In-Depth Information
16
Risk-Based Erosion Assessment:
Application to Forest Watershed
Management and Planning
W.J. ELLIOT
1
AND P.R. ROBICHAUD
1
1
USDA Forest Service, Rocky Mountain Research Station, Moscow, ID, USA
This chapter discusses conditions where risk-based
erosion modelling may be appropriate in fores-
ted watersheds. It then describes four modelling
approaches for risk-based erosion modelling using
WEPP-based erosion technology.
relatively low erosion risk within a few years.
Wildfires can cause soils to become water
repellent for a few years, increasing the risk of
runoff and erosion immediately after fire. The
repellency, however, dissipates in subsequent
years on many soils (Doerr
et al
., 2000; Robichaud,
2000). For example, Table 16.1 shows three
different studies in which erosion rapidly declined
during the three years following wildfires in all
but one year in one study.
One characteristic of these highly disturbed
conditions is a high spatial variability of the
disturbance. The disturbance rather than soil
properties dominates the erodibility of the soils
(Robichaud
et al
., 1993). The distribution of the
disturbance following wildfires is seldom uni-
form or predictable (Robichaud
et al
., 2007) as is
the case with agricultural conditions. There will
be sites following wildfire where the fire burned
at a higher severity, leading to a complete loss of
surface cover and most likely the generation or
augmentation of a water repellent soil condition.
There will be other sites where the fire burned
very little, or not at all, resulting in an area of
minimal erosion risk. There is often considerable
spatial variability in erodibility on a hillslope. For
example, in a study of hillslope erosion after a
wildfire in the Bitterroot Valley, Montana, US, the
four plots were within a 100-m wide hillside, yet
sediment delivered from the 15 July 2001 storm
ranged from 0.13 to 18 Mg ha
−1
(Table 16.2).
The weather in the years following the
disturbance is crucial in determining the erosion
16.1 Background
In many applications of erosion modelling, the
vegetation condition to be modelled is relatively
similar year after year, as in continuous agricul-
ture or grazing lands. In other cases, a known
sequence of surface conditions occurs over a
period of several years, like agricultural systems
with fixed crop rotations or short rotation for-
estry. In such cases, describing erosion with an
average annual value is usually an adequate
approach to conservation planning.
In some conditions, however, erosion processes
are dominated by extreme disturbance events
followed by a prolonged period of minimal distur-
bance, like unmanaged forests or rangelands in
fire-driven ecosystems (Fig. 16.1), or managed
forests that experience a major harvesting or
thinning operation only once every few decades.
In these cases, erosion is minimal prior to the
disturbance, potentially high immediately
following the disturbance, and then returns to a