Geology Reference
In-Depth Information
approach is more appropriate. For example, one
interpretation of the data in Table 16.2 is that
the average erosion rate in the year following a
wildfire is 16.8 Mg ha
−1
. A more meaningful
interpretation of these data might be that on one
of these plots there is a 1 in 4 chance that the
total erosion will exceed 23.9 Mg ha
−1
from the
four large storms in the year following the wild-
fire (Plot A total), and a 1 in 4 chance that ero-
sion will exceed 20.9 Mg ha
−1
from the largest
single erosion event in the year following the
wildfire (21 July on Plot A).
With risk-based erosion modelling, the
modeller must estimate the probability distribu-
tion for a given set of conditions, and from that
distribution, determine the probability of a given
erosion rate occurring. Probability distributions
should account for climate, soil properties and
distribution of disturbance.
these plots. In order to generate a climate for this
remote area, the climate statistics of a nearby
low-elevation weather station were modified
with an online interface (Scheele
et al
., 2001;
Elliot, 2004). Monthly precipitation amounts
were modified with data from a nearby high-
elevation snow monitoring station, and the
number of wet days was increased by half the pro-
portionate increase in precipitation. The monthly
maximum and minimum temperatures were
decreased from the valley station by the adiabatic
lapse rate (Scheele
et al
., 2001). Observed erosion
rates from Robichaud
et al
. (2008a) are presented
for comparison with values predicted by the
examples (Tables 16.1 and 16.2).
16.3 WEPP Windows
The weather file that drives the WEPP model
contains daily data, and so WEPP predicts runoff
and erosion on a storm-by-storm basis. All runoff
events predicted by WEPP are stored in a single
file. WEPP Windows accesses this file and deter-
mines the probability of exceeding a given amount
of daily precipitation, daily runoff, peak runoff
rate, or daily sediment delivery using a Weibull
plotting formula (WEPP Help screen).
16.2 Risk-based Approach
This chapter will consider four different tools to
use for risk-based erosion modelling, using inter-
faces developed for the Water Erosion Prediction
Project (WEPP) model (Flanagan & Livingston,
1995). The interfaces are the Windows Interface
(Flanagan
et al
., 1998), the online Disturbed
WEPP Interface (Elliot, 2004), the online Erosion
Risk Management Tool (ERMiT; Robichaud
et
al
., 2007), and the GeoWEPP GIS wizard
(Renschler, 2003).
An example application for each of these inter-
faces will be given to assist in understanding the
technology. All examples will apply to an analy-
sis of erosion following a high severity wildfire
that occurred in forested mountains of the
Bitterroot Valley in Western Montana, US, in July
2000. The soils in this area are gravelly sandy
loam over granitic colluvium, with slopes typi-
cally from 20% to 50% (Robichaud
et al
., 2008a).
For the hillslope examples, a horizontal slope
length of 20 m with a maximum steepness of 61%
will be used, similar to the silt fence plots
installed by Robichaud
et al
. (2008a). The ground
cover is assumed to be 5%, as was observed on
Example 16.1
To model the Bitterroot Valley
site in the WEPP Windows interface, the follow-
ing were selected from the downloaded databases
and menus: a 50-yr stochastic weather file; a
sandy loam, high-severity fire soil; the described
topography; and the Return Period Analysis
option. The management file was calibrated to
ensure approximately 5% ground cover for every
year of simulation.
The 'Return Period Analysis' output screen
from this model run (Fig. 16.2) shows an estimate
that for any given storm there is a 10% probability
that erosion will exceed 2.2 Mg ha
−1
. As a
comparison, the 'Average' value predicted by this
WEPP run was 0.55 Mg ha
−1
.
In this example, the 10-year sediment delivery
may or may not have been associated with the
10-year rainfall or 10-year peak runoff rate or
