Final Report
Assessing the Sources and Transport of Fine Sediment in Response to
Management Practices in the Tahoe Basin using the WEPP Model
Funding for this research was provided by the Bureau of Land Management through the sale of
public lands as authorized by the Southern Nevada Public Land Management Act (SNLPMA).
This Round 7 SNLPMA research grant was supported by an agreement with the USDA Forest
Service Pacific Southwest Research Station.
I. Project Team and Contact Information
Dr. Erin S. Brooks, Biological and Agricultural Engineering, University of Idaho, P.O. Box
442060, Moscow, ID 83844–2060; Phone: (208) 885–6562; fax: (208) 885–8923; email:
ebrooks@uidaho.edu
Dr. William Elliot USDA–FS Rocky Mountain Research Station, 1221 South Main, Moscow,
ID 83843; Phone: (208) 883 2338; fax: 208 883 2318; email: welliot@fs.fed.us
Dr. Jan Boll, Biological and Agricultural Engineering, University of Idaho, P.O. Box 442060,
Moscow, ID 83844–2060; Phone: (208) 885–7324; fax: (208) 885–8923; email:
jboll@uidaho.edu.
Dr. Joan Wu, Biological Systems Engineering, Washington State University, P.O. Box 646120,
Pullman, WA 99164–6120; Phone: (509) 335–5996; fax: (509) 335–2722; email: jwu@wsu.edu.
II. Executive Summary
The goal of this round 7 SNPLMA project was to test and develop the Water Erosion Prediction
Project (WEPP) as a watershed management tool for evaluating the impacts of specific
management practices on the generation of fine (< 20 micron) sediment transport at the hillslope
and watershed scales in the Lake Tahoe basin.
The project provided a Tahoe-specific database which includes all the necessary soil, climate,
and management input files to run the model at any location in the basin. The accuracy of the
model was tested against 18 years of observed snow water equivalent depth, streamflow and
sediment load data at the following relatively undisturbed watersheds: Blackwood Creek,
General Creek, Upper Truckee near Meyers, Logan House Creek, and Glenbrook Creek. Close
agreement between simulated and observed snow water equivalent, streamflow, and both fine
(<20 micron) and coarse (>20 micron) sediment load was achieved at each of the major
watersheds located in the high precipitation regions of the basin with minimal calibration.
Minimal sediment load was correctly simulated in the drier watersheds (i.e. Logan House and
Glenbrook) however annual streamflow was overpredicted. This overprediction was attributed
to unique hydrogeology on the east side of the basin.
The project also provided Tahoe-specific soil and management input files for representing the
impact of hillslope disturbances and subsequent restoration options on fine sediment loading.
1
These input files were created using existing rainfall simulation data collected in the basin over
the past 10 years. Major land cover treatments include: Mature forest, thin or young forest,
shrubs, good grass, poor grass, low severity fire, high severity fire, bare, mulch only, mulch and
till, low traffic road, high traffic road, and skid trail.
A high priority was placed on developing user-friendly tools and providing training. A simple
web-interface tool was developed and populated with the Tahoe-specific input files to allow end
users to easily predict the impact of disturbances on fine sediment loading from single hillslopes.
The input files were also modified to allow access to the Tahoe-specific input files in the WEPPwindows program. This free program, operating in the windows environment, provides
experienced users greater flexibility for simulating impacts of management practices from
complex hillslopes. For advanced users, algorithms were provided to simulate streamflow and
fine sediment loading from large complex watersheds. These watershed applications are
complex however and require users to be familiar GIS applications and basic programming
languages. A two day workshop was provided at the end of the project to train users on using
each of these projects. The feedback from this workshop was very positive.
III. Background
There is a great need in the Tahoe basin for objective, quantitative models for evaluating the
impacts of management practices on fine sediment loading to Lake Tahoe. Many of the models
currently used to evaluate the reduction in pollutant loading to Lake Tahoe are lumped
hydrologic pollutant loading models that do not route runoff, use constant event mean
concentrations for BMPs, and rely on specific source loading estimates and user-defined delivery
ratios which require subjective judgment and have considerable uncertainty (USACE, 2000).
Although it is often acknowledged that the more detailed, complex models may be able to better
describe hydrologic and pollutant transport mechanisms and overcome these limitations, processbased models have not been widely used in the basin due to fear that ―the level of complexity
may conflict with the objective of making the methodology practical for Lake Tahoe Basin
implementers,‖ USACE (2000). The Water Erosion Prediction Project (WEPP) model (Flanagan
and Nearing, 1995) is one example of a process-based hydrology and sediment transport model
that is able to provide objective, quantitative predictions of the impacts of forest management
practices on fine sediment loading; however at the onset of this project there were only a few
cases where WEPP had been used in the basin.
– The Water Erosion Prediction Project (WEPP) Model
The WEPP model is a process-based model which simulates soil detachment, deposition,
transport and delivery through hillslope, channel, and structural impoundment units (e.g.
sediment basins, culverts) within a watershed (Flanagan and Nearing, 1995). The model is based
fundamentally on infiltration theory, hydrology, soil physics, plant science, hydraulics, and
erosion mechanics. Each hillslope can be divided into multiple overland flow elements to
simulate flow from one land type to another (e.g., drainage from a disturbed upland area through
a grass buffer). The model accepts long term daily climate data or single storm event data. An
auxiliary climate generator program, CLIGEN (Nicks and Lane, 1989), creates long term climate
files if meteorological data are not available. A particular advantage of the model to the Tahoe
Basin‘s clarity reduction effort, which targets fine sediment (<20 microns), is its ability to predict
transport and delivery of specific size classes of sediment.
2
As in any complex process-based model, the challenge in applying WEPP is assigning
appropriate soil and vegetative input parameters. Early developers recognized this challenge and
invested much time and money conducting experiments on a wide range of soil types across the
country to develop parameter sets for specific soil, residue, and vegetative properties. Despite
this large database using local experimental data to create more appropriate site-specific input
files is encouraged.
The WEPP model can be applied to single hillslopes using a windows-based free software
program (WEPP-windows) or one of the forest service web-interface tools (Elliot, 2004) or it can
be applied to entire watersheds using the GeoWEPP model (Renschler, 2003). Stream channel
algorithms in the model are not appropriate for watershed areas exceeding ~1-2 sq. miles
however output from the model can be linked to detailed hydrodynamic models to account for
stream sediment transport in large rivers (Conroy et al., 2006). Alternatively, stream channel
algorithms can also be bypassed if the user is only interested in daily average flow and sediment
load from upland areas.
IV. Goal and Objectives
The primary goals of this project were to:
1.) Assess the accuracy of WEPP in the Lake Tahoe basin
a. Simulate streamflow and sediment loading in five, large, upland watersheds
2.) Develop Tahoe-specific input files and protocols which minimize the complexity
required to apply WEPP
a. Focus on specific management practices relevant to the Lake Tahoe Basin.
3.) Provide one-on-one training to encourage use of the model in the basin.
a. Encourage feedback and interaction with managers
V. Model Development and Assessment
The project was carried out in three phases. The first phase of the project focused on assessment
of the model at simulating snowmelt, streamflow and sediment load in large undisturbed
watersheds. Five watersheds were selected to represent the full range of climate, elevation, soil,
and geologic types found in the basin. The primary purpose for selecting ‗undisturbed‘
watersheds was to isolate the ability of the model to capture natural climate, soil, and vegetative
variability from the effects of anthropogenic disturbances (e.g. urban development) on output
hydrographs. The second phase of the project focused on developing the appropriate soil and
management input files to capture the effects of soil disturbances (e.g. roads, burns, harvested
sites) on erosion and sediment delivery. The third phase of the project was providing the final
watershed database and necessary protocols to potential end-users of the model. This included a
two day training session teaching participants how to use the web-interface tools, the windows
software program, and the GIS watershed analysis. Each of these Phases will be described
below.
PHASE I
3
- Development and Assessment of the WEPP-UI Approach for Large Undisturbed Watersheds
The accuracy of the WEPP model to predict streamflow and sediment loading was tested in five
upland watersheds in the basin: the headwaters of the Upper Truckee River near Meyers (14
mi2), Logan House Creek (2.1 mi2), Glenbrook creek (4.1 mi2), General Creek (7.4 mi2), and
Blackwood Creek (11 mi2). The entire Upper Truckee River watershed, including the disturbed
areas downstream of Meyers, is known to contribute the majority of the sediment to Lake Tahoe.
General Creek and Blackwood Creek watersheds are both located on the wetter, west side of the
basin. General Creek watershed is relatively undisturbed compared to Blackwood Creek
watershed, which has been impacted by roads, grazing, gravel mining, anthropogenic channel
modifications, and logging. The Logan House and Glenbrook Creek watersheds are located on
the dry eastern side of the lake and are small, undisturbed watersheds with relatively low
sediment loading. Each of these watersheds has minimal residential or urban development.
Hence, we were able to isolate upland responses from urban responses. Long term (> 15 years)
continuous daily streamflow and event-based sediment concentration data are publicly available
at the outlet of the watersheds through the US Geological Survey. These data sets include
measurements of the percent fine sediment (< 0.063 mm, silt and clay size) for water samples
having high sediment concentrations. Both of these data sets were used to assess the accuracy of
the model.
Since the WEPP model was not originally designed for application to large watersheds (i.e.
greater than 1-2 mi2) several modifications were necessary to capture the spatial variability in
climate, soil and canopy cover, as well as accounting for baseflow in the model. These
modifications will be described in detail below.
Undisturbed Soil Input Files
Although the WEPP model comes with specific soil parameters for many soils throughout the
US, many of the Tahoe soils were not in the database or did not have the complete information
required by recent beta versions of the model (e.g. soil porosity, field capacity moisture content,
wilting point moisture content, and saturated hydraulic conductivity for each soil layer). Rather
than relying on default soil properties calculated from soil texture in WEPP, we built WEPP soil
files using the measured soil properties listed for each of the soils listed in the 2007 Tahoe Basin
Soil Survey map. Other than base erodibility parameters which are not explicitly provided in the
soil survey, all soil properties were taken directly from the soil survey. Undisturbed erodibility
parameters were calculated using default algorithms in the WEPP model based on soil texture.
WEPP soil files representing undisturbed soil conditions were created for all 55 of the Tahoe
basin soils listed in the soil survey.
Climate Files
The WEPP model requires either a daily or a breakpoint-(e.g. 15 min, hourly) based climate file
which includes minimum and maximum air temperature, solar radiation, wind speed, wind
direction, dew point temperature, and a distribution of precipitation. In order to be compatible
with concurrent modeling efforts in the basin we acquired the hourly-cleaned weather data from
Northwest Hydraulic Consultants for each of the eight major weather stations/SNOTEL sites in
the basin which were originally developed by Tetra-Tech and incorporated into the Loading
Simulation Program C++ (LSPC) model when developing the TMDL for the Tahoe basin.
4
Northwest Hydraulic Consultants were using this data set with the Storm Water Management
Model (SWMM).
Precipitation data can be represented in climate files in WEPP using two different formats. For
long term return period analysis the climate generator algorithm (CLIGEN) can either generate
weather files with statistics calculated from existing climate data where precipitation on any
given day is unrelated to any actual observed data or it can take daily observed data to ensure
total precipitation for the day is conserved yet peak rainfall intensities and storm durations are
still generated and unrelated to observed values. Alternatively, the user also can create break point precipitation files where the actual distribution of the storm is explicitly provided. For the
Tahoe basin, there are advantages and disadvantages to both approaches. The best available
precipitation data in the Tahoe basin are recorded on an hourly basis. Since most of the
precipitation falls as snow it is important in order to represent the accumulation and melt of snow
that precipitation is explicitly distributed within a day. Therefore hourly weather files were
created using the hourly precipitation data which explicitly describe the distribution of
precipitation during the day. The disadvantage of these hourly weather files is that high intensity
storm bursts that often last less than an hour may not have been adequately represented with
hourly data. The daily storm generator algorithms can mimic these storm bursts however there
is no guarantee that the peak intensity generated by the storm matches reality. In the hydrologic
analysis we assessed the accuracy of the model using the hourly weather data, however, both the
hourly and daily generated weather files were used and discussed in the assessment simulated
sediment loading using WEPP.
Another modeling challenge when working in the Tahoe basin was representing the spatial
variability in climate. The Tahoe climate varies substantially both regionally and with elevation.
As a result it was necessary to develop a methodology for scaling the climate data within each
watershed. This protocol involved developing scaling factors using the mean monthly
precipitation and temperature at the particular hillslope and at the location of a nearby weather
station. The precipitation scaling factor is a ratio of the mean monthly precipitation at the
hillslope to the mean monthly precipitation at the weather station. The minimum and maximum
air temperature scaling factors are simply the differences between the mean monthly air
temperature at the hillslope and at the weather station. High spatial resolution (800 m) maps
which quantify the spatial variability in 30 year average monthly precipitation, minimum and
maximum temperature developed by the PRISM group at Oregon State University were used to
scale the climate data to each hillslope.
Forest Management files
Another key challenge in modeling the hydrology of the Tahoe basin was adequately
representing the vegetative canopy. Canopy cover can largely affect the rate at which snow
melts especially on steep south facing slopes where shortwave radiation drives most of the
snowmelt. The default forest file in WEPP assumes nearly 100% canopy cover. Although most
land use maps describe the Tahoe basin as mostly forested, it is apparent to anyone who has
visited there that the forest is quite open and variable. In order to capture this feature we
developed the necessary relationships to allow users to fix percent canopy cover in the model.
The average percent canopy cover for each hillslope was calculated for each hillslope using a 30
m resolution percent canopy cover maps for 2001 created using the National Land Cover Dataset
5
(http://www.mrlc.gov/). PERL scripts were used to automatically alter management files for each
hillslope when simulating the large undisturbed watersheds so that the variability in canopy
cover was properly represented in the model.
Hillslope files
The topography of each hillslope in a watershed is described in WEPP by a hillslope input file.
Hillslope files were created for each of the watersheds using the GeoWEPP model. Recent work
by Crabtree et al. (2006) demonstrated the importance of using multiple overland flow elements
(OFEs) in WEPP to represent variable source area hydrology. Since all hillslope files created in
GeoWEPP describe each hillslope with a single OFE it was necessary to modify the structure of
each hillslope created by WEPP to force multiple OFEs. A PERL script was written to
automatically convert these hillslope files.
Results of the Undisturbed Watershed Assessment (Phase I)
Statistics
The accuracy of the model was assessed using three statistics: the root mean square error
(RMSE), which is similar to the standard deviation of the error between model observations and
predictions; the Nash-Sutcliffe (NS-EFF) efficiency statistic (Nash and Sutcliffe, 1970), which
represents how well the model describes the observed variability relative to the mean observed
value for the selected time period; and the mean difference (MD) which describes the overall
bias of the model. The equations used to calculate the NS-EFF, RMSE and MD statistics are
given below.
N
vo N
NS
xi
yi
N
2
xi
i 1
1
vo N
yi
2
x
2
i 1
N
(1)
xi
i 1
N
xi
RMSE
i 1
N
yi
2
(2)
1 N
xi y i
(3)
N i1
where vo is the variance of observed values, N is the number of data points, xi is the observed
value, yi is the corresponding simulated value, and xi is the average observed value for the study
period. A NS-EFF value of 1.0 indicates perfect agreement between simulated and observed
data. A NS-EFF value of 0.0 indicates the model is no better than simply using the mean of the
observed data. It is possible to have negative NS-EFF values which indicate poor or no
agreement with observed data. Although calibrated hydrology models can often be fit to provide
NS-EFF near 0.70-0.90 for streamflow, especially when comparing monthly predictions, a NSEff. above 0.30 when determined using daily simulated data from an un-calibrated model, is a
good indication that the fundamental mechanics of the model are correct. Similarly, Foglia et al.
(2009) considered that a NS-Eff below 0.2 is insufficient, 0.2-0.4 is sufficient, 0.4-0.6 is good,
0.6-0.8 is very good, and greater than 0.8 is excellent. We used these relative criteria for rating
the accuracy of the WEPP model in this study.
MD
6
Snow Hydrology Assessment
The Lake Tahoe basin is a snow dominated landscape and therefore it is essential that any
hydrology model used in the basin can successfully simulate the accumulation and ablation of a
snowpack. There are eight SNOTEL stations located in the Tahoe Basin each of which has been
recording daily snow water equivalent (SWE) depth along with temperature and precipitation for
a minimum of 20 years. These SNOTEL sites provided the perfect data set for assessing the
snowmelt algorithms in WEPP. Table 1 summarizes the characteristics of each SNOTEL site
and provides a statistical summary of the agreement between simulated and observed daily SWE
depth. The SNOTEL sites range in elevation from essentially lake level at the Fallen Leaf
SNOTEL (1901 m a.s.l.) to over 700 m above the lake level at the Heavenly Valley SNOTEL
(2616 m a.s.l.). The peak annual SWE depth ranges from 218 mm at Fallen Leaf SNOTEL to
1145 mm at the Echo Peak SNOTEL, see Table 1. Figures 1-3 show predicted and observed
SWE at the Ward Creek site, which is located on the western side of the basin and has the largest
mean annual precipitation, the Marlette Lake site, which is located on the eastern, drier side of
the lake, and the Fallen Leaf site, respectively. Despite this variability the WEPP model was able
to closely simulate the observed accumulation and melt of the snow pack at each of these sites.
Assuming the Foglia et al. (2009) criteria applies to snowmelt, we can say the model agreement
ranges from ‗very good‘ at the sites with relatively shallow, transient snow packs (e.g. Fallen
Leaf and Hagan‘s Meadow) to ‗excellent‘ at the remaining sites which have deeper snowpacks.
The RMSE indicates that the overall error in simulated SWE is 106 mm and the MD statistic
indicates the error is typically a 37 mm underprediction in SWE. By dividing the RMSE by the
peak observed SWE we created a scaled percent error, see Table 1. The average percent error of
all the sites is 16%.
Streamflow and Water Yield Assessment
The following section is a summary of the accuracy achieved by the WEPP model in five
undisturbed watersheds in the Tahoe basin. Our goal was to demonstrate the variability in
accuracy that can be expected from WEPP in different regions of the basin. It is important to
note when interpreting the accuracy of model predictions that, other than an analysis explicitly
described below for Logan House and Glenbrook watersheds, no calibration was performed on
any of the WEPP simulations. Only the empirical method for baseflow generation after WEPP
simulations was calibrated in a few watersheds. The five watersheds included in this analysis are
Blackwood Creek and General Creek, both located on the wetter, west side of the basin; Upper
Truckee near Meyers (UTR nr. Meyers), the headwater watershed on the south end of the lake;
Logan House Creek and Glenbrook Creek, both located on the drier, east side of the basin. Note
that the Upper Truckee watershed is located upstream of Meyer and is not influenced by the city
of South Lake Tahoe.
Comparisons of simulated streamflow and water yield indicated that there may be a unique
feature to the east side watersheds that is not adequately represented by the WEPP model.
Because of these unique differences the assessment of simulated daily streamflow and average
annual water yield is broken down into two sections. The first describes the agreement with the
west and south side watersheds and the second describes the agreement for the drier east side
watersheds.
7
West and South Side Watersheds
Overall the agreement between simulated and observed streamflow was good to very good for
the west side and south side watersheds (i.e. Blackwood, General, UTR nr. Meyers). Table 2
provides the Nash-Sutcliffe efficiency (NS-Eff) (Nash and Sutcliffe, 1970) statistic between
simulated and observed daily streamflow for each watershed.
As shown in Table 2, the NS-Eff calculated for 17 years of daily simulated streamflow was 0.40,
0.60, and 0.52 for Blackwood, General, and UTR nr. Meyers, respectively. The NS-Eff
calculated for individual years reached as high as 0.80, 0.87, and 0.83 for Blackwood, General,
and UTR nr. Meyers, respectively, which would be considered ‗excellent‘ using the Foglia et al.
(2009) criteria. Simulated average annual water yield was also very good for each of these
watersheds, as shown in Table 2 and Figure 4.
Daily simulated and observed streamflow is provided for Blackwood, General, and UTR nr.
Meyers watersheds in Figures 5, 6, and 7, respectively. Although the agreement between
simulated and observed is very good most of the years, the large flood event on 1/1/1997 is
greatly under-predicted at all three locations by WEPP. This flood event was caused by a severe
rain and snow event. During these intense storms it is critical that distributed temperature and
precipitation data are well represented in the model. High elevation SNOTEL sites were used to
distribute temperature and precipitation at each of these sites. A low elevation SNOTEL site,
Fallen Leaf, is located below the UTR nr. Meyers watershed. By using this lower elevation input
data we were able to capture the extreme flood event in 1997, and only slightly overpredict the
peak flow (see Figure 8). Most likely this 1997 event was a very warm, wet storm which raised
air temperatures much higher at lower elevations near the lake than at higher elevations. Since
our modeling approach assumes that the temperature lapse rate can be characterized by static
mean monthly maps there is the possibility of great error in the extrapolated air temperature
depending upon which base weather station is used. The ability of the model to simulate the
extreme peak flow event using the lower elevation data provides confidence that the fundamental
hydrological processes in the model are able to represent these extreme events if properly
distributed weather data are used.
East Side Watersheds
Water yield and streamflow were overpredicted in the east side watersheds (i.e. Logan House,
Glenbrook), see Figure 4. Knowing that the Tahoe basin is located above consolidated bedrock,
which generally does not support large subsurface aquifers, we assumed in each of the west and
south side watershed simulations that no water bypasses the watershed outlet below the stream
gage. In other words we assumed that these watersheds were ―water tight‖. A plausible
explanation for the overpredicted streamflow in the east side watersheds is that water that drains
below the root zone either passes below the stream gage station through cracks in the bedrock
ultimately recharging the lake as groundwater flow or groundwater could possibly drain to the
east toward Carson City.
With a ―water tight‖ assumption, baseflow (i.e. that portion of total streamflow that is provided
by springs originating below the rooting depth) is simulated using the linear reservoir concept.
Water draining vertically below the root zone is assumed to feed an underground storage
reservoir. A linear reservoir coefficient derived from the recession portion of observed
8
hydrographs determines the fixed percentage of this storage amount that is discharged back to
the stream. Generally this coefficient does not vary within the same landscape. For the west and
south side watersheds we determined the recession coefficient to be 4%. This value was used for
all three watersheds.
To evaluate if a portion of the water potentially bypassed the stream gage station in the east side
watersheds a secondary reservoir coefficient was added. For each day, the primary linear
reservoir coefficient controls the portion of the groundwater reservoir that returns to the creek
and the secondary reservoir coefficient acts on the same reservoir but controls that amount of
water that bypasses the stream gage station.
For the east side watersheds the primary and secondary reservoir coefficients were calibrated so
that the simulated and measured total water yield at the stream gage matched. The simulated
amount of water that bypassed the stream gage is referred to as ―groundwater losses‖ in Table 2.
As seen in Table 2, the simulated amount of groundwater losses in Logan House and Glenbrook
watersheds is 177 mm and 81 mm, respectively.
Although we do not have conclusive data to quantify groundwater flow beneath the stream gage,
there is some compelling evidence that supports the idea that these watersheds have unique
hydrogeologic flow processes. Logan House stream gage station is located over 120 m above the
average annual water surface elevation of Lake Tahoe just upstream of a steep region of rock
outcroppings which have the potential for conducting flow through cracks and fissures. The
stream gage in the Glenbrook watershed, located just north of Logan House Creek, was installed
at essentially the same elevation as Lake Tahoe, and therefore is likely less susceptible to
subsurface groundwater losses beneath the station. If deep seepage losses were related to the
location of the stream gage station and the geology directly below the stream gage station then
we would expect the groundwater losses in Glenbrook Creek to be less than the losses in Logan
House Creek. As seen in Table 2 and shown in Figure 4, our modeling data supported this
hypothesis. The simulated ground water loss required to match average annual water yield is
over two times larger in Logan House Creek, 177 mm, than in Glenbrook Creek, 81 mm.
However since there still seem to be groundwater losses in Glenbrook Creek another geologic
feature may not be accounted for by the model.
Another interesting feature that points to unique groundwater flow and storage in these east side
watersheds can be seen in the long term summer streamflow recorded at both Logan House and
Glenbrook. From 1995-2000, the Tahoe region experienced above normal precipitation.
Interestingly this time period of higher precipitation had a large effect on summer streamflow
readings in Glenbrook and Logan House Creeks and had very little effect on west side streams.
As seen in Figure 9 the streamflow at the Logan House stream gage station remained well above
0.1 mm per day even during the driest summer months during these wet years. During the drier
years, however, this flow would often drop to below 0.01 mm per day. It is interesting to
compare Figure 9 to Figure 10 where there is no observable change in the summer streamflow at
General Creek throughout the entire 17 year flow record. Glenbrook streamflow data show the
same response as Logan House streamflow during these wet water years (data not shown). In
order for east side streams to maintain high flows even during hot summer months requires a
large source of subsurface spring flow. This would mean that the geology in this region must
9
have a large capacity to store water which slowly leaks out with time. As seen in Figure 9 we
were able to roughly mimic these long-term trends by using a very small linear reservoir
coefficient, 0.04%.
It is worth noting that regardless of whether the deep seepage losses are simulated or ignored,
generally, the WEPP model was able to capture the extreme differences in water yield between
the east side and west side (Figure 4). As seen in Table 5, when deep seepage losses were
simulated by the model the NS-Eff statistic was much better for Glenbrook watershed, 0.44, than
for Logan House Creek, -0.63. The low NS-Eff for Logan House was largely biased by a few
overpredicted runoff events (data not shown).
Identification of Sources of Fine Sediment
Although the primary goal of the watershed scale assessment was to validate the hydrologic
components of the WEPP model, these simulations also provided the opportunity to assess the
ability of WEPP to simulate sediment loading from upland sources. The existing stream channel
algorithms in WEPP were designed from small (less than 1 mi2) catchments and therefore were
not to be used to simulate sediment detachment, deposition, and delivery through the existing
channel network for the five watersheds in this analysis. As a result a direct comparison between
observed sediment loading at the watershed outlet should not be made because it neglects the
additional source-sink dynamics due to interaction with stream beds and banks. Despite the
missing sediment contribution from the streams, it is worth examining the relative magnitude and
distribution of sediment by the model compared with existing data. The following section
presents an analysis of simulated sediment transport and identifies the dominant sources of fine
sediment in each of the watersheds by soil type.
Total and Fine Sediment Loading
As discussed in the section on climate files, there are two types of climate files that can be used
in the WEPP model. The hourly files are preferable for ensuring that the snowpack is adequately
developed whereas the daily files allow for high intensity storm bursts that cannot be represented
by hourly data. In the following analysis we provide sediment load data using both types of
files. Table 3 provides a summary of 17 year average total and fine sediment loading delivered
to the stream network simulated by the model and observed total and fine sediment loads
measured at the watershed outlet. Since sediment load data are often skewed by a few extreme
events, both average and median sediment load information is provided in Table 3. Figure 11
graphically represents that simulated and observed average annual sediment loads from each of
the watersheds. Observed sediment load was calculated using the sediment rating curves
developed by Simon et al. (2003) from observed fine and coarse suspended sediment
concentration and streamflow data collected at each watershed outlet location by the USGS. As
seen in Table 3 and Figure 11, Blackwood Creek delivers considerably more sediment than each
of the other watersheds. The second largest contribution of sediment was from the UTR nr.
Meyers watershed which on average provides an order of magnitude less total sediment than
Blackwood Creek. These findings confirm the conclusions of the Simon et al. (2003) study
which indicated that most of the sediment in the Upper Truckee River was due to bank erosion in
flood plain regions downstream of Meyers near South Lake Tahoe. Total sediment load from
east side watersheds is minimal. Simulated sediment delivery for each of these watersheds
matched these trends fairly well. Simulated sediment loading using both the hourly and daily
10
models was less than observed at the watershed outlet which is to be expected considering that
contributions from stream banks and beds are not included in the simulated values. Simulated
sediment load using daily climate files are greater than simulated loads using hourly files,
emphasizing the potential for increased sediment load in the Tahoe basin from short duration
storms. Figure 12 compares the observed and simulated breakdown of total sediment load into
fine and coarse particle sizes for Blackwood Creek. Although the WEPP model was not used to
simulate stream bank erosion, the 40% underprediction in fine sediment loading for Blackwood
creek watershed suggests that 60% of the fine sediment loading was derived from upland sources
while the remaining 40% is due to stream bed and bank erosion. This distribution is similar to
the Simon et al. (2003) study which attributed 72%, 79%, and 49% of the fine sediment load to
upland sources in the General, Ward, and Upper Truckee River watersheds, respectively. In
contrast to the Simon et al. (2003) study the simulations using the WEPP model suggest that
nearly all the fine sediment load in General creek is derived from in-stream sources. The Simon
et al. (2003) estimates were based on predictions using the AnnAGNPS and CONCEPTS
models. Simon et al. (2003) acknowledged in the report that the AnnAGNPS predictions could
only be used reliably at monthly and yearly time scales due to an inability to represent the
change in climate across the watershed. It was suggested that the upland contributions of fine
sediment in the General creek watershed could be overestimated because of overprediction in
streamflow by AnnAGNPS during winter months and an inability of the AnnAGNPS model to
predict baseflow. Although no statistical analysis was provided in the Simon et al. (2003) report,
visually it appears the WEPP model was better captured the hydrology of the watersheds in the
basin.
Hydrograph Separation
One of the advantages of a physically based model is that multiple output features can be
examined to provide insight into the dominant hydrologic processes within a watershed. For
example, WEPP can be used to separate a storm hydrograph into three fundamental components:
1. surface runoff fed directly to the stream, 2. subsurface lateral flow or stormflow which is
shallow subsurface flow fed directly to the stream usually within the top 2 m of the soil, and 3.
baseflow, which is deep groundwater or spring flow generated through the parent material at
depths greater than 2 m below the soil surface. Figures 13 and 14 provide a breakdown of the
simulated streamflow into each of these fundamental components.
It is clear from Figures 13 and 14 that each of these watersheds is composed of unique
distributions of surface runoff, subsurface lateral flow, and baseflow. Even adjacent watersheds
can have very distinct hydrologic signatures. For example, Blackwood and General Creek,
located on the west side of the lake, have very similar climates yet hydrographs from Blackwood
Creek are composed mostly of lateral flow whereas General Creek is composed mostly of
baseflow. Surface runoff from Blackwood Creek is lower than from General Creek yet
simulated sediment load is greater for Blackwood Creek than for General Creek. The primary
controlling factors which dictate the proportion of surface runoff, subsurface lateral flow and
baseflow in the hydrograph are the soil properties and topographic attributes of the hillslopes
within the watershed. Strong agreement between predicted and observed streamflow, in addition
to matching overall trends with observed sediment load in adjacent watersheds, provides
evidence that WEPP is able to simulate drastically different dominant hydrologic processes with
minimal calibration.
11
Figures 15-22 provide a detailed breakdown of the simulated and observed streamflow and
sediment loading for the 2005 season at the four primary watersheds in the study. As seen at
each watershed the agreement between simulated and observed streamflow is good. According
to WEPP, Blackwood Creek was the only watershed with upland contributions of total sediment
(Figure 15). USGS streamflow records show that General Creek was carrying sediment during
this year; however, the WEPP simulations imply that the sediment carried during this time was
derived from stream beds and bank (Figure 17). Notice that despite a large portion of the total
streamflow from the UTR nr. Meyers watershed was composed of surface runoff there was very
little simulated and observed sediment loading from the watershed (Figures 18 and 19).
Although according to the NS-Eff statistic the agreement between simulated and observed
streamflow is poorest for Logan House Creek, the model was able to correctly describe the
overall magnitude of flow which is so small that the watershed typically never delivers sediment
(Figures 21-22).
Sources of Sediment
One of the primary goals of this project is to provide a tool that watershed managers can use to
identify sources of fine sediment. Since WEPP is a spatially explicit tool, output from the model
can be in the form of maps or can be tabulated by land use, topography, and/or soil type. Since
much of the exploratory work in the basin has focused on soil types, Tables 4-6 provide average
annual sediment loading, runoff, and subsurface lateral flow for each of the soil types in each
watershed.
Many of the relationships between dominant hydrologic processes and simulated sediment
transport can be explained through interpretation of results in Tables 4-6. For example, as
discussed in the previous section the total streamflow from the UTR nr. Meyers is composed
mostly of surface runoff yet the model predicts very little sediment delivery to the stream. The
explanation for this apparent contradiction is that the headwaters of the Upper Truckee River is
composed mostly of rock outcrops which have a low infiltration rate, a high runoff rate and a low
erodibility. Table 6 indicates that 38.3% of the entire UTR nr. Meyers watershed is composed of
rock outcrops. These rock outcrops generate 72.8% of the total surface runoff and deliver only
2.8% of the total sediment load.
Tables 4-6 also indicate that in order to minimize upland sediment loading from Blackwood
Creek, restoration efforts should focus on Melody soils. These volcanic soils make up 14.9% of
the total watershed yet deliver 65.8% of the total sediment load. These soils are often located in
the steep, high elevation, ‗Badland‘ regions of the Ward and Blackwood creek watersheds.
Stubblefield et al. (2009) estimated that the Ward creek badlands contributed 10-39% of the total
sediment load while only occupying 1.2% of the surface area of the watershed.
Summary of the Watershed Assessment (Phase I)
Overall we developed the necessary input files and algorithms to simulate the hydrology of
undisturbed watersheds in the Tahoe basin. We found excellent agreement between simulated
and observed snow water equivalent at all eight SNOTEL sites located in the Tahoe basin. We
found good agreement with streamflow and water yield in the Blackwood Creek, General Creek,
12
and Upper Truckee River near Meyers watersheds. The model overpredicted water yield and
streamflow in the drier eastern watersheds. Baseflow analysis revealed that this overprediction is
likely caused by local groundwater leakage/storage processes in the east side watersheds.
Despite this overprediction in streamflow, trends in simulated sediment loading agree well with
observations made at the outlet of all five of the watersheds investigated in this analysis.
PHASE II
Parameterization of the WEPP model for Disturbed Areas
After assessing the accuracy of the basin hydrologic functions of WEPP in undisturbed
watersheds the focus of the second phase of the project was developing the proper
parameterization for disturbed areas or areas impacted by a specific upland management practice
and developing the necessary protocols for applying the model to these disturbed sites.
The first step in this phase involved field visits and stakeholder meetings to identify the most
widespread disturbances in the basin and to create a list of key management practices most often
used to restore these disturbed sites. Much of this work has been described by Grismer and
Hogan (2004, 2005a, 2005b) and Hatchett et al. (2006).
Widespread disturbances
Through field visits, meetings with stakeholders, and literature reviews we developed the
following list of features and practices with potential for ground disturbance.
- Ski Runs
- Cut slopes
- Roads (bare and paved)
- Wildfire
- Prescribed burns
- Mechanical harvest (cut-to-length and whole tree)
- Landings
- Skid trails
Restoration Practices
The management practices used to restore these disturbed sites varied by the amount of time and
resources available. In some cases multiple management practices were applied to the same
disturbed site. The most common management practices are listed below (Grismer and Hogan,
2005b).
- Rock and large woody debris
- Pine needle mulch
- Hydroseed
- Revegetate with native grasses
- Fertilizer/compost addition (e.g. biosol)
- Wood chips
- Tillage to improve soil water storage
- Rebuild slope (e.g. road obliteration)
- Mastication
13
The second step in this phase was to select a few of the most common practices and acquire
relevant experimental data to parameterize the WEPP soil and management files. Through this
process we identified thirteen vegetative conditions/treatments that would be most applicable to
managers working in the Lake Tahoe Basin. For each of these treatments we developed WEPP
soil and management files from existing experimental datasets developed from experiments in
the Tahoe basin or from experiments conducted in similar climates, soils, and managements
(Copeland and Foltz, 2009; Elliot et al., 1989; Foltz et al., 2008; Grismer et al., 2008; Grismer
and Hogan, 2005b; Wagenbrenner et al., 2010). Since it was not possible to acquire information
on each specific soil type in the Tahoe basin we lumped soils into three major categories
following the approach of Gismer and Hogan (2004, 2005a, 2005b): volcanic, granitic, and
alluvial soils. Tables 7-9 identify the major treatments and baseline parameters used in the
study. Other than soil depth, each of the baseline values listed in Tables 7-9 were corrected
internally in WEPP by surface cover. Since surface cover will vary for each site, and it is
directly measureable in nearly all applications of the model, the user is prompted to enter percent
cover by the model. Depth of soil in the database is dependent on the treatment. Larger plants
were assumed to have deeper soils, which reflect both the deeper rooting depth and the increased
capacity of the plant to store water internally. The management file describes a fixed surface
cover along with the respective canopy, rather than a cover that varies throughout the year and is
affected by plant growth rates. This means the user can directly enter the cover that was
observed in the field without the need to calibrate as is required on the Disturbed WEPP
interface.
PHASE III
Training Users using the Lake Tahoe WEPP Database
The focus of the third phase of the project was on creating a Tahoe-specific WEPP database that
could be easily transferred to end-users and providing one-on-one training on using the database
with various WEPP based tools. The culmination of the project was a two-day training session
conducted in June 2010. Participants ranged from private consultants and Forest Service
hydrologists with an interest in learning how to apply the model to specific projects, to federal
regulatory agents who were more interested in learning about the accuracy, applicability, and
limitations of the model. The technological experience ranged from very limited modeling to
extensive programming and familiarity with geographic information systems. The feedback we
received back from the workshop was very positive.
To accommodate the wide background the two-day workshop provided an overview of the three
WEPP-based modeling tools, see Appendix A for the meeting agenda. These tools range in
complexity from a simple web-based tool, called the Tahoe Basin Sediment Model, to a
Windows-based model with additional input/output features that allows the model to be applied
to single hillslopes or simple watersheds called WEPP windows, to a GIS-based modeling tool
which can be applied to large watersheds called the WEPP-UI watershed approach. Each of
these tools will be summarized below. A detailed explanation of these tools is provided in
Appendix B and C.
Tahoe Basin Sediment Model
14
The simplest tool is a web-based program called the ―Tahoe Basin Sediment Model (TBSM)‖
(see http://forest.moscowfsl.wsu.edu/cgi-bin/fswepp/tahoe/tahoe.pl). This web-based tool is
part of a larger existing suite of web-based tools called the ―Forest Service WEPP Interfaces‖
(see http://forest.moscowfsl.wsu.edu/fswepp/). The TBSM is similar to the ‗Disturbed WEPP‘
interface however it is populated with the Tahoe specific soil and management input files
developed in Phase II of the project. Like the disturbed WEPP interface, users can simulate the
runoff, erosion and sediment yield from a simple hillslope composed of two linear segments or
overland flow elements (OFEs) using climate files generated by the ―Rock:Clime‖ weather
generator program. The Rock:Clime program was populated with the eight Tahoe-specific
climate files generated in Phase I of the project allowing users to capture the extreme variability
in climate across the basin. Training materials were developed for the workshop to demonstrate
setting up custom climate files and using TBSM to quantify the change in sediment load for
typical Tahoe basin management problems. These applications include: 1.) quantifying
sediment reductions from various road designs, 2.) comparing sediment load from natural
wildfires, to prescribed burn and thinning options, 3.) comparing options for road obliteration
and ski run restoration such as grass, mulch, and mulch-and-till revegetation treatments as well
as adding water bars, 4.) understanding how to extract fine sediment loading from model output.
See Appendix B for the TBSM training materials
WEPP Windows Model
The WEPP windows model provides users with more input and output options than the TBSM
tool and is useful for evaluating complex hillslopes, small watersheds, and various structural
impoundments (e.g. sediment basins and culverts). For the workshop we provided instructions
for importing the Tahoe-specific soil, management, and climate files developed in Phases I and II
of the project, see Appendix C. After importing these files users have the ability to break
hillslopes into multiple OFEs, link more than one hillslope together using channels and
impoundments, and create simple watersheds. The workshop handout provides step-by-step
instructions for the following applications: 1.) setting up a hillslope project, 2.) evaluating the
importance of using multiple OFEs on runoff amount and distribution for unique hillslope
shapes, 3.) assessing the importance of depth to bedrock on the amount of runoff and erosion for
various climates, 4.) using the model to predict the change in water yield, and 5.) examining the
importance of canopy cover and multiple OFEs for snowmelt simulation for various climates. In
the workshop we also demonstrated how the model can be used to design a sediment basin and
culvert.
The WEPP-UI Watershed Approach
For large watershed applications it is highly recommended that the users apply the WEPP-UI
Watershed approach to evaluate the impacts of management on upland runoff and sediment
yield. This approach was developed in Phase I of the project and applied to the Blackwood
Creek, General Creek, Upper Truckee near Meyers, Logan House, and Glenbrook Creek
watersheds. This is a complex approach which requires users to be familiar with GIS and
programming using PERL scripts, and to be comfortable working with large spreadsheets.
Detailed instructions were developed for the workshop. They are provided in Appendix C.
Users are first required to run GeoWEPP for defining the watershed, breaking out hillslopes, and
assigning soil and vegetative management attributes to each hillslope. These files are then
modified and formatted using spreadsheets and PERL scripts to set up large batch runs. Mean
15
monthly precipitation, maximum and minimum temperature, and canopy cover are calculated for
each hillslope using a GIS program and are arranged into a climate input file. The input files for
each hillslope in the watershed are then passed to the WEPP executable program and PERL
scripts are used to extract water balance, soil erosion, and particle size information from the
output files. These output files can be imported into a spreadsheet template which accumulates
output from each hillslope in the watershed and plots predicted versus observed data. Although
this watershed approach was successful at simulating streamflow and sediment load for large
watersheds in the Tahoe basin, it should still be considered as a beta-version until a more userfriendly computer interface can be developed. In the workshop materials we discuss the
limitations of using the GeoWEPP model as a stand-alone watershed program and why it was
necessary to develop the WEPP-UI watershed approach, see Appendix C. Each person at the
workshop was provided with a CD that included all the necessary GIS maps required for
applying the WEPP-UI approach to any watershed in the basin. We demonstrated how to use
GeoWEPP on the Ward Creek watershed and showed the necessary steps to calculate mean
monthly precipitation statistics for each hillslope using ArcGIS.
VI. Summary
Overall this project demonstrated that the WEPP model is well suited for representing the
extreme variability in climate, soils, vegetative conditions found in the Tahoe basin and can
successfully simulate streamflow and sediment transport in large upland watersheds. Through
meetings with local managers in the basin 13 potential WEPP applications were identified and
parameterized using existing datasets. These applications include wildfire, prescribed burning,
thinning, mulching, revegetating slopes, and road design. A digital database which included all
the necessary input files and maps necessary to run WEPP anywhere in the basin was created and
distributed to managers at a two-day training workshop.
One of the main concerns with using the WEPP model had been that it was too complex and
difficult to use. To address this concern a simple, web-based interface tool focused specifically
on Tahoe basin applications, called the Tahoe Basin Sediment Model, was developed and
populated with the input files created by the project. For managers who are interested in more
complex hillslope applications Tahoe-specific input files were created for the WEPP windows
program. Finally, protocols were developed for applying the model at the watershed scale using
the WEPP-UI approach. Step-by-step instructions on using these tools were provided and
demonstrated at a two-day training workshop at the end of the project.
VII. Future Work/Limitations/Data gaps
Although these tools have now greatly simplified the application of the WEPP model in the basin
and have increased confidence in the predictive ability of the model, there are some key data
gaps and process improvements that should be addressed by future projects. First, there is a
great wealth of experimental data quantifying infiltration rates and effective erodibility using
rainfall simulation experiments on 0.8 m x 0.8 m plots, however there are relatively few rill
erodibility experiments. Rill erodibility and critical shear stress can be calculated from
experiments where variable flows are added to a 4-6 m long rill. We used much of the rainfall
simulation experimental data to quantify relative effects of various revegetation treatments on
16
infiltration and erosion. Some of these data indicate, however, that the rainfall experiments were
not run long enough to reach equilibrium conditions. Future rainfall simulation experiments
should ensure that steady state conditions are reached.
Since the WEPP model does not simulate nutrient dynamics it cannot reliably predict whether a
vegetative cover will establish or how quickly existing vegetative cover will grow. Vegetative
growth in the model is dependent only upon temperature and moisture conditions and assumes
nutrients do not limit growth. WEPP is an ideal tool for predicting sediment transport for a
known surface cover and therefore we highly recommend direct, repeated measurements of
surface cover at application sites for direct use as input to the model. We also highly recommend
that continued emphasis be placed on adaptive management studies focused on developing the
most cost-effective techniques for establishing and maintaining vegetative cover for various soils
and climates in the basin.
At the present time, the WEPP model is not well suited for predicting stream sediment transport.
The algorithms in the model are only appropriate for small watersheds (i.e. less than 1 -2 sq
miles). In this project, daily watershed outlet predictions neglected scour and deposition in the
stream network. One option to account for stream sediment transport is to link hillslope ouput to
a hydro-dynamic stream sediment transport model such as the CONCEPTS model (Langendoen,
2000) which has been used extensively in the basin. There is also on-going work at Washington
State University towards improving/replacing the existing stream channel algorithms in the
WEPP model. This work should be continued in order to provide managers with a complete tool
for evaluating cumulative watershed effects.
A Round 10 SNPLMA proposal (Elliot at el., 2009) was funded to further develop the online
interface, essentially enhancing the output to make it more compatible with the PLRM model
developed for urban erosion. Activities associated with this proposal will include carrying out a
survey of current fuel management practices and modeling needs in the basin, and incorporation
of the future climate database under development from Round 9 projects. In addition, a survey
of potential data sets useful for model development or validation will be carried out, and the
online interface will receive further validation.
VIII. References
Conroy, W.J., R.H. Hotchkiss, and W.J. Elliot. 2006. A coupled upland–erosion and instream
hydrodynamic sediment transport model for evaluating sediment transport in forested
watersheds. Trans. ASABE 49: 1713–1722.
Copeland, N.S. and R.B. Foltz. 2009. Improving erosion modeling on forest roads in the Lake
Tahoe Basin: Small plot rainfall simulations to determine saturated hydraulic conductivity
and interrill erodiblity. Presented at the Annual International Meeting of the American Soc.
of Ag. & Bio. Engrs, 21-24 June, Reno, NV. St. Joseph, MI: ASABE. 11 p.
Crabtree, B., E.S. Brooks, K. Ostrowski, W.J. Elliot, and J. Boll. 2006. The Water Erosion
Prediction Project (WEPP) model for saturation excess conditions: application to an
agricultural and a forested watershed. Eos Trans. AGU, 87(52), Fall Meet. Suppl., Abstr.
H31C–1444.
17
Elliot, W. J., A. M. Liebenow, J. M. Laflen and K. D. Kohl. 1989. A compendium of soil
erodibility data from WEPP cropland soil field erodibility experiments 1987 & 88. Report
No. 3. USDA-ARS, National Soil Erosion Research Laboratory, W. Lafayette, IN. 316 p.
Elliot, W.J. 2004. WEPP Internet interfaces for forest erosion prediction. J. Am. Water Resour.
Assoc. 40: 299–309.
Elliot, W, E. Brooks and D. Traeumer. 2009. Development and validation of the Tahoe Project
Sediment Model. A research proposal submitted the Southern Nevada Public Land
Management Act Request for Proposals. Moscow, ID: U.S. Dept. of Agriculture Forest
Service, Rocky Mountain Research Station.
Flanagan, D.C., and M.A. Nearing. 1995. USDA-Water Erosion Prediction Project: Hillslope
Profile and Watershed Model Documentation. NSERL Rep. 10. USDA-ARS Natl. Soil
Erosion Res. Lab., West Lafayette, IN.
Foglia, L, Hill MC, Mehl SW, and Burlando P (2009) Sensitivity analysis, calibration, and
testing of a distributed hydrological model using error-based weighting and one objective
function. Water Resources Research, Vol. 45, W06427, doi:10.1029/2008WR007255
Foltz, R.B., H. Rhee and W.J. Elliot. 2008. Modeling changes in rill erodibility and critical
shear stress on native surface roads. Hydrol. Process. DOI: 10.1002/hyp.7092.
Grismer, M.E. and M.P. Hogan. 2005b. Simulated ranfall evaluation of revegetation/mulch
erosion control in the Lake Tahoe Basin - 3: Soil treatment effects. Land Degrad. &
Develop. 13:1-13.
Grismer, M.E. and M.P. Hogan. 2004. Evaluation of Revegetation/Mulch Erosion Control Using
Simulated Rainfall in the Lake Tahoe Basin: 1. Method Assessment. Land Degrad. & Dev.
13:573-588.
Grismer, M.E. and M.P. Hogan. 2005a. Evaluation of Revegetation/Mulch Erosion Control
Using Simulated Rainfall in the Lake Tahoe Basin: 2. Bare Soil Assessment. Land Degrad.
& Dev. 16:397-404.
Grismer, M.E. and M.P. Hogan. 2005b. Evaluation of Revegetation/Mulch Erosion Control
Using Simulated Rainfall in the Lake Tahoe Basin: 3. Treatment Assessment. Land Degrad.
& Dev. 16:489-501.
Grismer, M.E., A.L. Ellis and A. Fristensky. 2008. Runoff sediment particle sizes associated
with soil erosion in the Lake Tahoe Basin, USA. Land Degrad. Dev. 19: 331-350.
Hatchett, B., M. P. Hogan and M. E. Grismer. 2006. Mechanized Mastication Effects on Soil
Compaction and Runoff from Forests in the Western Lake Tahoe Basin. California Ag.
60(2):77-82.
Langendoen, E.J. 2000. CONCEPTS—Conservational channel evolution and pollutant transport
system: Stream corridor. ver. 1.0, Res. Rep. No. 16, USDA-ARS, Natl. Sediment. Lab.,
Oxford, MS.
Nicks, A.D., and L.J. Lane. 1989. Weather Generator. Chap. 2. In: L.J. Lane and M.A. Nearing
(eds.) USDA-Water Erosion Prediction Project. NSERL Rep. 2. USDA-ARS Natl. Soil
Erosion Res. Lab., Purdue University, West Lafayette, IN.
Renschler, C.S. 2003. Designing geo–spatial interfaces to scale process models: The GeoWEPP
approach. Hydrol. Process. 17: 1005–1017.
Simon, A., E. Langendoen, R. Bingner, R. Wells, A. Heins, N. Jokay, and I. Jaramillo, 2003.
Lake Tahoe Basin Framework Implementation Study: Sediment Loadings and Channel
Erosion. National Sedimentation Laboratory Technical Report 39, USDA-ARS National
Sedimentation Laboratory, Oxford, Mississippi, 320 pp.
18
Stubblefield, A.P., J.E. Reuter, and C.R. Goldman. 2009. Sediment budget for subalpine
watersheds, Lake Tahoe, California, USA. Catena, doi:10.1016/j.catena.2008.11.002.
USACE, 2006, Final Rep. on the Pollutant Load Reduction Calculation Methodology
http://www.spk.usace.army.mil/projects/civil/tahoe/documents.html, p. 84–87
Wagenbrenner, J.W.; Robichaud, P.R.; Elliot, W.J. 2010. Rill erosion in natural and disturbed
forests, part II: modeling approaches. In Press for Water Res. Research. Moscow, ID:
Rocky Mountain Research Station.
19
Table 1. Assessment of the snowmelt algorithms in WEPP at the eight SNOTEL sites in the Tahoe Basin. The Nash-Sutcliffe
Efficiency (NS-Eff), mean difference (MD), and root mean square error (RMSE) are calculated using daily simulated and observed
snow water equivalent (SWE).
Hagan's Meadow
Fallen Leaf
Heavenly Valley
Marlette Lake
Tahoe City Cross
Rubicon
Echo Peak
Ward Creek
2370
1901
2616
2402
2072
2344
2338
2028
Avg.
Annual
Precip
(mm)
776
825
844
854
879
1057
1497
1705
Average
2259
1055
SNOTEL Site
Elev
(m)
Avg.
Avg.
Annual Annual
Tmax
Tmin
(C)
(C)
13.4
-4.3
13.8
-2.6
9.6
-3.0
10.8
-1.3
12.8
-0.1
10.4
1.0
12.0
1.1
13.5
-2.6
12.0
-1.5
Avg.
Annual
Peak Obs.
SWE (mm)
472
218
709
650
404
781
1145
938
665
NS-Eff
MD
(mm)
RMSE
(mm)
RMSE/Pk Obs
SWE
0.72
0.73
0.90
0.88
0.91
0.81
0.87
0.82
-43
6
-51
-48
-16
-61
-94
14
101
40
92
95
47
144
170
159
21%
18%
13%
15%
12%
18%
15%
17%
0.83
-37
106
16%
20
Table 2. Hydrologic assessment of the WEPP model for the five Tahoe watersheds. Averages are provided for the years 1989-2005.
Streamflow simulations are assessed using the Nash-Sutcliffe Efficiency (NS Eff) statistic. Simulated total streamflow is divided into
three primary constituents: surface runoff, subsurface lateral flow, and baseflow. Simulated groundwater loss represents flow that
bypasses the stream gages station either recharging the lake directly or being stored in a groundwater aquifer.
Percentage of Total Streamflow
Watershed
NS Eff
Streamflow
Obs.
Water
Yield
(mm/yr)
Sim.
Water
Yield
(mm/yr)
Sim.
Pct.
Runoff
Sim. Pct.
Subsurface
Lateral
Sim. Pct.
Baseflow
Mean Annual
Precipitation
(mm/yr)
Sim.
Groundwater
Losses
(mm/yr)
Blackwood
0.40
1045
1062
12.8%
47.3%
39.9%
1620
0
General
0.60
738
753
28.2%
10.3%
61.5%
1281
0
UTR nr. Meyer
0.52
894
873
43.8%
25.3%
31.0%
1315
0
Logan House
-0.63
96
102
7.7%
9.3%
83.0%
807
177
Glenbrook
0.44
153
156
2.2%
23.1%
74.7%
716
81
21
Table 3. Simulated average annual total and fine sediment (<20 µm) yield from upland hillslopes and observed sediment load at the
watershed outlet.
Avg. Total Sediment
Load (Tonnes/yr)
Avg. Fine Sediment
Load (Tonnes/yr)
Median Total
Sediment Load
(Tonnes/yr)
Median Fine Sediment
Load (Tonnes/yr)
Watershed
Weather
Input
NS Eff.
Streamflow
Simulated
Observed
Simulated
Observed
Simulated
Observed
Simulated
Observed
Blackwood
Hourly
0.40
900
3098
487
1264
637
1990
356
771
General
Hourly
0.60
0
252
0
69
0
114
0
53
UTR nr. Meyer
Hourly
0.52
7
356
2
NA
0
287
0
NA
Logan House
Hourly
-0.63
0
4
0
NA
0
1
0
NA
Glenbrook
Hourly
0.44
0
5
0
NA
0
4
0
NA
Blackwood
Daily
0.37
2101
3098
1197
1264
794
1990
461
771
General
Daily
0.58
21
252
11
69
0
114
0
53
UTR nr. Meyer
Daily
0.50
41
356
24
NA
1
287
0
NA
Logan House
Daily
-4.96
191
4
167
NA
3
1
2
NA
Glenbrook
Daily
0.38
0
5
0
NA
0
4
0
NA
22
Table 4. Dominant hydrology and sediment loads for Blackwood Creek soil types.
Soil Type
Area
Subsurface
Lateral
Flow
Surface
Runoff
Sed.
Del.
Parent
Material
Waca
Sky
Ellispeak
Melody
Kneeridge
Rock
Tahoe
Tallac
Pits
Mountrose
Paige
Watah
Oxyaquic
31.3%
26.7%
15.7%
14.9%
6.5%
2.5%
1.3%
0.6%
0.2%
0.2%
0.1%
0.1%
0.0%
7.7%
21.2%
29.0%
38.7%
1.0%
0.6%
0.7%
0.1%
0.0%
0.0%
0.2%
0.7%
0.0%
6.4%
31.6%
12.2%
25.5%
1.6%
14.2%
1.9%
0.2%
6.0%
0.2%
0.0%
0.0%
0.2%
8.9%
0.7%
14.9%
65.8%
0.0%
9.2%
0.0%
0.3%
0.2%
0.1%
0.0%
0.0%
0.0%
Volcanic
Volcanic
Volcanic
Volcanic
Volcanic
Volcanic
Mixed
Granitic
Granitic
Volcanic
Volcanic
Organic
Mixed
Table 5. Dominant hydrology and sediment loads for General Creek soil types.
Soil Type
Area
Dagget
Meeks
Oxyaquic
Rock
Tallac
30.1%
28.9%
22.7%
17.9%
0.4%
Subsurface
Lateral
Flow
Surface
Runoff
62.4%
2.5%
30.7%
4.4%
0.0%
95.4%
0.4%
3.8%
0.4%
0.0%
Sed.
Del.
Parent
Material
86.2%
0.1%
9.1%
4.5%
0.0%
Granitic
Granitic
Mixed
Granitic
Granitic
23
Table 6. Dominant hydrology and sediment load for UTR nr. Meyers soil types.
Soil Type
Area
Subsurface
Lateral
Flow
Surface
Runoff
Sed.
Del.
Parent
Material
Rock
Lithnip
Dagget
Cagwin
Bidart
Meeks
Callat
Sky
Watah
Temo
Water
Tahoe
Oxyaquic
Cassenai
Tallac
38.3%
15.6%
9.3%
8.2%
7.8%
7.7%
7.5%
1.8%
1.1%
0.9%
0.8%
0.5%
0.2%
0.1%
0.1%
41.9%
29.2%
5.2%
0.9%
3.3%
1.2%
12.5%
1.7%
0.9%
2.3%
0.8%
0.1%
0.0%
0.0%
0.1%
72.8%
19.0%
0.4%
0.0%
0.3%
0.0%
3.6%
0.8%
0.0%
0.8%
2.0%
0.1%
0.1%
0.0%
0.0%
2.4%
42.8%
0.0%
0.0%
0.4%
0.3%
0.6%
0.0%
0.0%
52.8%
0.0%
0.0%
0.8%
0.0%
0.0%
Granitic
Volcanic
Granitic
Granitic
Mixed
Granitic
Volcanic
Volcanic
Organic
Granitic
Organic
Mixed
Mixed
Granitic
Granitic
Table 7. Baseline granitic soil parameters (90% sand, 8% silt, 2% clay).
Interrill
Rill
Critical
Keff
Treatment
Erodibility
Erodibility
Shear
(mm/hr)
(kg-s/m4)
(s/m)
(N/m2 )
Skid Trail
2700000
0.001
4
10
High Severity Fire
1800000
0.0005
4
15
Low Severity Fire
1000000
0.0003
4
20
Poor Grass
750000
0.0001
4
25
Good Grass
600000
0.00008
4
30
Shrubs
500000
0.00006
4
35
Thin or Young Forest
400000
0.00004
4
40
Mature Forest
250000
0.00003
4
45
Bare
300000
0.001
4
25
Mulch only
300000
0.001
4
30
Mulch and Till
300000
0.001
4
35
Low Traffic Road
225000
0.0013
4
10
High Traffic Road
900000
0.005
4
10
Soil
Depth
(mm)
300
300
300
350
400
500
600
800
300
400
500
200
200
24
Table 8. Baseline volcanic soil parameters (65% sand, 28% silt, 7% clay).
Interrill
Rill
Critical
Keff
Treatment
Erodibility
Erodibility
Shear
(mm/hr)
4
2
(kg-s/m )
(s/m)
(N/m )
Skid Trail
3000000
0.0008
1.5
8
High Severity Fire
2000000
0.0004
1.5
10
Low Severity Fire
1500000
0.0002
1.5
15
Poor Grass
1000000
0.00008
1.5
20
Good Grass
900000
0.00006
1.5
25
Shrubs
800000
0.00004
1.5
30
Thin or Young Forest
700000
0.00003
1.5
35
Mature Forest
600000
0.00002
1.5
40
Bare
750000
0.0008
1.5
20
Mulch only
750000
0.0008
1.5
25
Mulch and Till
750000
0.0008
1.5
30
Low Traffic Road
250000
0.001
1.5
8
High Traffic Road
1000000
0.004
1.5
8
Soil
Depth
(mm)
300
300
300
350
400
500
600
800
300
400
500
200
200
Table 9. Baseline alluvial soil parameters (60% sand, 30% silt, 10% clay).
Interrill
Rill
Critical
Keff
Treatment
Erodibility
Erodibility
Shear
(mm/hr)
4
2
(kg-s/m )
(s/m)
(N/m )
Skid Trail
2500000
0.0006
1
6
High Severity Fire
1500000
0.0003
1
8
Low Severity Fire
1000000
0.0002
1
10
Poor Grass
900000
0.00006
1
15
Good Grass
800000
0.00005
1
20
Shrubs
700000
0.00003
1
25
Thin or Young Forest
600000
0.00002
1
30
Mature Forest
500000
0.00001
1
35
Bare
600000
0.0006
1
15
Mulch only
600000
0.0006
1
20
Mulch and Till
600000
0.0006
1
25
Low Traffic Road
240000
0.0008
1
6
High Traffic Road
950000
0.003
1
6
Soil
Depth
(mm)
400
400
400
450
500
600
700
900
400
500
600
200
200
25
2000
Observed
Simulated
1800
Snow Water Equivalent (mm)
1600
1400
1200
1000
800
600
400
200
0
10/1/1989
6/27/1992
3/24/1995
12/18/1997
9/13/2000
6/10/2003
Figure 1. Simulated and observed snow water equivalent (SWE) depth at the Ward
Creek SNOTEL.
2000
Observed
Simulated
1800
Snow Water Equivalent (mm)
1600
1400
1200
1000
800
600
400
200
0
10/1/1989
6/27/1992
3/24/1995
12/18/1997
9/13/2000
6/10/2003
Figure 2. Simulated and observed snow water equivalent (SWE) depth at the Marlette
Lake SNOTEL.
26
2000
Observed
Simulated
1800
Snow Water Equivalent (mm)
1600
1400
1200
1000
800
600
400
200
0
10/1/1989
6/27/1992
3/24/1995
12/18/1997
9/13/2000
6/10/2003
Figure 3. Simulated and observed snow water equivalent (SWE) depth at the Fallen Leaf
SNOTEL.
27
1200
Observed
Simulated No Groundwater Losses
Average Water Yield (mm/yr)
1000
Simulated with Groundwater losses
800
600
400
200
0
Blackwood
General
UTR nr. Meyers Logan House
Glenbrook
Figure 4. Simulated and observed water yield from 17 year simulations (1990-2005).
180
160
Obs. streamflow
Sim. streamflow
140
Streamflow (mm)
120
100
80
60
40
20
0
1/1/89
9/28/91
6/24/94
3/20/97
12/15/99
9/10/02
6/6/05
Figure 5. Simulated and observed streamflow for Blackwood Creek.
28
80
Observed
Simulated
70
Streamflow (mm)
60
50
40
30
20
10
0
1/1/89
9/28/91
6/24/94
3/20/97 12/15/99 9/10/02
Figure 6. Simulated and observed streamflow for General Creek.
80
6/6/05
Obs. streamflow
Sim. streamflow
70
Streamflow (mm)
60
50
40
30
20
10
0
1/1/89
9/28/91
6/24/94
3/20/97
12/15/99
9/10/02
6/6/05
Figure 7. Simulated and observed streamflow for Upper Truckee River near Meyers
(UTR nr. Meyers).
29
120
Simulated using Heavenly SNOTEL data
Simulated using Fallen Leaf SNOTEL Data
Streamflow (mm)
100
Observed Streamflow
80
60
40
20
0
10/1/1996
11/30/1996
1/29/1997
3/30/1997
5/29/1997
7/28/1997
9/26/1997
Figure 8. UTR nr. Meyers observed streamflow versus simulated streamflow using a low
elevation base SNOTEL station, Fallen Leaf (1901 m a.s.l.), and using a high elevation
base SNOTEL station, Heavenly Valley (2616 m a.s.l.).
30
Obs. streamflow
Sim. streamflow
Streamflow (mm)
10
1
0.1
0.01
0.001
1/1/89
9/28/91
6/24/94
3/20/97
12/15/99
9/10/02
6/6/05
Figure 9. Simulated and observed streamflow plotted on a logarithmic scale for Logan
House Creek.
Observed
Simulated
Streamflow (mm)
10
1
0.1
0.01
0.001
1/1/89
9/28/91
6/24/94
3/20/97 12/15/99 9/10/02
6/6/05
Figure 10. Simulated and observed streamflow plotted on a logarithmic scale for General
Creek.
31
2500
Median Annual Sediment Yield (Tonnes/yr)
Observed
Daily Simulation
2000
Hourly Simulation
1500
1000
500
0
Blackwood
General
UTR nr. Meyers Logan House
Glenbrook
Figure 11. Observed sediment loading at the watershed outlet and simulated sediment loads
delivered to the stream network using both hourly and daily climate files.
2500
Median Annual Sediment Load (Tonnes/yr)
Coarse Sediment
Fine Sediment
2000
Blackwood Creek watershed
1500
1000
500
0
Observed
Simulated
Hourly Simulation
Figure 12. Breakdown of total sediment load delivered to Blackwood Creek into fine (< 20 μm)
and coarse (> 20 μm) particle sizes.
32
1200
Base Flow
Subsurface Lateral Flow
Surface Runoff
Composition of Streamflow (mm)
1000
800
600
400
200
0
Blackwood
General
UTR nr. Meyers Logan House
Glenbrook
Figure 13. Total simulated streamflow broken down into the major hydrologic components.
120%
Base Flow
Subsurface Lateral Flow
Surface Runoff
Percent of Total Streamflow
100%
80%
60%
40%
20%
0%
Blackwood
General
UTR nr. Meyers Logan House
Glenbrook
Figure 14. Percentage of total simulated streamflow provided by each of the major hydrologic
components.
33
Streamflow (mm)
70
4000
Obs. streamflow
Sim. streamflow
Obs Sed Load
Sim. Sed. Load
3500
60
3000
50
2500
40
2000
30
1500
20
1000
10
500
0
4/8/2005
0
5/8/2005
6/7/2005
Sediment Load (Tonnes/day)
80
7/7/2005
Figure 15. Simulated and observed streamflow and sediment load for Blackwood Creek.
80
Surface Runoff
Subsurface Lateral Flow
70
Baseflow
Streamflow (mm)
60
50
40
30
20
10
0
4/8/2005
5/8/2005
6/8/2005
7/8/2005
Figure 16. Simulated streamflow for Blackwood Creek broken down into the major hydrologic
components.
34
Streamflow (mm)
70
4000
Obs. streamflow
Sim. streamflow
Obs. Sed. Del.
Sim. Sed. Del.
3500
60
3000
50
2500
40
2000
30
1500
20
1000
10
500
0
4/8/2005
0
5/8/2005
6/7/2005
Sediment Load (Tonnes/day)
80
7/7/2005
Figure 17. Simulated and observed streamflow and sediment load for General Creek.
80
Surface Runoff
Subsurface Lateral Flow
70
Baseflow
Streamflow (mm)
60
50
40
30
20
10
0
4/8/2005
5/8/2005
6/8/2005
7/8/2005
Figure 18. Simulated streamflow for General Creek broken down into the major hydrologic
components.
35
Streamflow (mm)
70
4000
Obs. streamflow
Sim. streamflow
Obs. Sed. Del.
Sim. Sed. Del.
3500
60
3000
50
2500
40
2000
30
1500
20
1000
10
500
0
4/8/2005
0
5/8/2005
6/7/2005
Sediment Load (Tonnes/day)
80
7/7/2005
Figure 19. Simulated and observed streamflow and sediment load for UTR nr. Meyers.
80
Surface Runoff
Subsurface Lateral Flow
70
Baseflow
Streamflow (mm)
60
50
40
30
20
10
0
4/8/2005
5/8/2005
6/8/2005
7/8/2005
Figure 20. Simulated streamflow for UTR nr. Meyers broken down into the major hydrologic
components.
36
Streamflow (mm)
70
4000
Obs. streamflow
Sim. streamflow
Obs. Sed. Load
Sim. Sed. Load
3500
60
3000
50
2500
40
2000
30
1500
20
1000
10
500
0
4/8/2005
0
5/8/2005
6/7/2005
Sediment Load (Tonnes/day)
80
7/7/2005
Figure 21. Simulated and observed streamflow and sediment load for Logan House Creek.
80
Surface Runoff
Subsurface Lateral Flow
70
Baseflow
Streamflow (mm)
60
50
40
30
20
10
0
4/8/2005
5/8/2005
6/8/2005
7/8/2005
Figure 22. Simulated streamflow for Logan House Creek broken down into the major hydrologic
components.
37
APPENDIX A
Training Workshop Materials
Tahoe Basin Sediment Model
T P S
ahoe
roject
ediment
M
odel
Air Water and Aquatic Environments
& Department of Biological and Agricultural Engineering
Rocky Mountain Research Station,
University of Idaho
Moscow, Idaho 83843
June, 2010
Workshop Outline
June 16 – 17, 2010
Tahoe Center for Environmental Sciences
Sierra Nevada College
Incline Village, Nevada
Organizers and Presenters:
Bill Elliot, Rocky Mountain Research Station, Moscow, ID
Erin Brooks, University of Idaho, Moscow, ID
Drea Traeumer, Em Hydrology, Reno, NV
Wednesday Morning
Cross Drain
10:00
10:30
11:30
Tahoe Interface Climate Selection
Tahoe Interface for Roads
Tahoe Interface for Disturbed
Sites
Lunch at Cafeteria
Bill
Bill
Bill
12:30
Cross Drain
Bill
Erin
ing
pac
Introductions
Overview of WEPP
S
ain
8:45
9:00
r
ss D
Bill
Cro
8:30
Welcome, Housekeeping
Road
Fillslope
Forest
Buffer
Sediment Yield
to stream
Wednesday Afternoon
Tahoe Interface Cover
Calibrations
Tahoe interface for Fuel Mgt
Bill
Erin
4:30
WEPP Windows Tahoe
Climates, Soils & Vegetation
Select Topics for Day 2
5:00
Adjourn
1:30
2:00
3:30
Bill
Bill &
Erin
Possible Selected Topics For Day 2 (Will choose 4)
Water yield; Winter processes; GeoWEPP; Waterbars on ski slopes or trails; Future climates;
Wildfire mitigation; Flood flows; minimum flows; Sediment basins (hill slope or watershed); Others
…
WEPP Workshop, June 2010, Tahoe Center for Environmental Sciences, Incline Village, NV
Thursday Morning
8:30
Selected Topic 1
Bill or Erin
9:45
Selected Topic 2
Bill or Erin
11:00
Selected Topic 3
Bill or Erin
12:15
Lunch at Cafeteria
Thursday Afternoon
1:15
Selected Topic 4
Bill or Erin
2:45
Tahoe Project Sediment
Model Interface Needs
Drea
3:15
Workshop Evaluation
Erin
3:30
Adjourn
Software sites:
FSWEPP:
http://forest.moscowfsl.wsu.edu/fswepp/
WEPP software and documentation: http://www.ars.usda.gov/Research/docs.htm?docid=18084
GeoWEPP software & documentation site: http://www.geog.buffalo.edu/~rensch/geowepp/
Email addresses:
welliot@fs.fed.us
ebrooks@uidaho.edu
drea.em@gmail.com
GeoWEPP developer Chris S. Renschler - rensch@buffalo.edu
p2
WEPP Workshop, June 2010, Tahoe Center for Environmental Sciences, Incline Village, NV
p2
APPENDIX B
Training Workshop Materials
Tahoe Basin Sediment Model
T B S
ahoe
asin
ediment
M
odel
Air Water and Aquatic Environments
& Department of Biological and Agricultural Engineering
Rocky Mountain Research Station,
University of Idaho
Moscow, Idaho 83843
June, 2010
I.
Online Interface Workshop
Find the Web Site: http://forest.moscowfsl.wsu.edu/fswepp/ .
1. Select English units and enter your ―personality‖ code.
2. Select the Tahoe Basin sediment model
3. Ponder the interface.
II.
Select the Desired Climate
1.
2.
3.
4.
Click Custom Climate
Scroll to the bottom of the Region, select Tahoe Basin and Click Show Climates
Select Rubicon #2 Snotel and click Modify Climate
Note the Rubicon Climate (Seems a bit on the dry side), and in the upper right corner
click the PRISM box.
5. Select the PRISM grid cell 2.5 miles north and click Use Prism Values
6. Click □ Adjust Temperature for Elevation by Lapse Rate .
7. Name the Climate ― N of Rubicon CA ” and click Use These Values
8. Click Return to Input Screen
9. Select N of Rubicon CA from Climate List.
10. Click Run WEPP
11. At the bottom of the output screen, click Return to Input screen
III.
Tahoe Interface for Roads
Basic Approach: Model likely current condition and compare to benefits from reducing
traffic, outsloping the road, or paving.
A. Most Common surfaces, high traffic and rutted, insloped, or flat.
1. In Soil Texture Box, select granitic
2. For the upper element, select High traffic road
and for the lower element Thin or young forest
3. Specify the topography:
4%
4%
20%
15%
200 ft
50 ft
4. Click Run WEPP and fill in the first line in Table 1 on the next page.
B. Low Traffic: Change to Low Traffic Road and Run WEPP
Tahoe Basin Sediment Model Workshop, June, 2010
p2
C. Outsloped Road: Select High Traffic Road, change upper length to 30 feet, and
Run WEPP (total area remains unchanged).
D. Paved Road: Select Rock/Pavement in soil box, Change upper length to 200 ft ,
High Traffic Road in Treatment and Run WEPP
E. Add Waterbar: Select High Traffic Road, Granitic Soil, Upper length is 100 ft ,
(Assume area remains unchanged).
Table 1. Results of Road Erosion Runs
Climate Station:
Road Length:
ft
Buffer Length:
ft
Road Surface
Annual Precip:
in.
Road Width: 14
ft
Area:
Acres
Buffer Width: 14
ft
Area:
Acres
(43,560 Square feet = 1 Acre)
Road + Buffer Area:
Acres
Runoff (inches)
Road Erosion Road + Buffer Delivery from
Rate
Delivery Rate
buffer
Rain
Snow
Total
(tons/acre)
(Tons/acre)
Tons
Questions?
IV.
Tahoe Interface for Fuel Management
Basic Approach: Estimate ―background‖ sediment from undisturbed forest and wildfire;
Compare background to erosion associated with thinning and prescribed fire.
A. Undisturbed Forest Erosion:
1. Select climate N of Rubicon and soil granitic
2. Specify upper and lower treatments to be Mature Forest
3. Specify Slope to be:
20%
30%
30%
10%
300 ft
100 ft
4. Run WEPP and enter sediment delivery into Table 2 (Columns (1) and (5)).
B. Wildfire before treatment:
1. Change upper treatment to High severity fire 30 percent cover
and lower treatment to Low severity fire 50 percent cover.
2. Run WEPP and enter sediment delivery into table 2 (column (1)).
Tahoe Basin Sediment Model Workshop, June, 2010
p3
C. Calculate ―background‖ sediment budget:
1. Divide the erosion in column (1) by the return interval in column (2) and enter the
results in column (3).
2. Sum up the two average annual values in column (3) to get the background sediment
delivery rate.
D. Estimate the sediment generated by thinning and prescribed fire:
1. For thinning, Upper treatment: select Thin or young forest and set the cover to 85
percent and the Lower treatment to Mature forest leaving the default cover at 100
percent.
2. Run WEPP and enter the sediment delivery in Table 2, column 5.
3. For prescribed, Upper treatment: select Low severity fire and leave the default cover to
be 85 percent and the Lower treatment as Mature forest with 100 percent cover.
4. Run WEPP and enter the sediment deliveries in Table 2, column (5).
E. Estimate the sediment generated by wildfire following fuel treatment:
1. Set the upper treatment to High severity fire , 40 percent cover and
the Lower treatment to Low severity fire , 70 percent cover.
2. Run WEPP and enter the sediment delivery in Table 2, column (5).
F. Calculate ―Treated‖ sediment budget:
1. Divide the erosion in column (5) by the return interval in column (6) and enter the
results in column (7).
2. Sum up the four average annual values in column (7) to get the average annual treated
sediment delivery rate.
3. Discuss the background versus the treated sediment delivery. The extra sediment from
roads, if any, may also need to be considered (low traffic roads become high?)
Table 2. Summary of Erosion Analysis for Fuel Management
Column:
(1)
(2)
(3)
(4)
(5)
No
Sed
Return
Annual
Treated
Sed
Treatment
Delivery
Interval
Average
Delivery
(t/a)
(years)
(t/a/yr)
(t/a)
Forest
Wildfire
1
40
Background sediment delivery rate:
(6)
Return
Interval
(years)
Forest
1
Thinning
20
Rx Fire
20
Wildfire
50
Treated sediment delivery rate:
Questions?
(7)
Annual
Average
(t/a/yr)
Tahoe Basin Sediment Model Workshop, June, 2010
V.
p4
Tahoe Interface for Mitigating Bare Areas (Ski Slopes, road cuts, etc.)
Basic Approach: Estimate ―untreated‖ sediment from site with low cover.
Compare untreated to treatment scenarios (mulching, incorporate residue (tillage), and
shortening slope with a water bar).
A. Current condition:
1. Select climate N of Rubicon and soil volcanic .
2. Specify upper and lower treatments to be Bare , 20 percent cover.
3. Specify Slope to be
30%
30%
30%
20%
4.
B.
C.
D.
E.
50 ft
50 ft
Run WEPP and enter average runoff, erosion and sediment delivery, and 10-year
return period sediment delivery into table 3.
Mulching treatment:
1. Change upper and lower treatments to Mulching 85 percent cover.
2. Run WEPP and enter average runoff, erosion and sediment delivery, and 10-year
return period sediment delivery into table 3.
Mulch and tillage treatment:
1. Change upper and lower treatments to Mulch and till 85 percent cover.
2. Run WEPP and enter average runoff, erosion and sediment delivery, and 10-year
return period sediment delivery into table 3.
Waterbar only, no mulch:
1. Change upper treatment to Bare 20 percent cover
and lower treatment to Mature Forest 100 percent cover.
2. Run WEPP and enter average runoff, erosion and sediment delivery, and 10-year
return period sediment delivery into table 3.
Discuss results: costs, life of treatment, feasibility of treatment, local considerations…
Table 3. Erosion associated with mitigation bare areas
Runoff
Erosion
Sed
(in.)
Rate
Delivery
Treatment
(t/a)
(t/a)
Bare
Surface Mulched
Incorporate Mulch
Water Bar
Questions?
10-yr
Sediment
Deliver
Cost ($)
and/or Life of
Tmt
Tahoe Basin Sediment Model Workshop, June, 2010
VI.
p5
Estimating the amount of fine sediment delivered from a hillslope
Basic Approach: At the end of the run, look at the end of the extended output file to calculate
the amount of clay (or clay plus part of the silt) in the runoff. Clay size particles and
aggregates are assumed to be less than 4 microns in diameter. Silt is generally between 4 and
62.5 microns diameter. (1,000 microns = 1 mm)
1. Rerun the previous run of Bare + buffer.
2. Enter the delivered sediment into Table 4.
3. At the bottom of the output page, click extended .
4. At the bottom of the file that comes up, study the Sediment Characteristics
and Enrichment table.
5. Enter the information requested from the Sediment Characteristics and
Enrichment output into Table 4.
Table 4. Fraction of Delivered Sediment that is clay (less than 4 micron dia)
Column:
(1)
(2)
(3)
Class
Percent Clay
Fraction in Flow
Percent Clay in Class
Exiting
1
2
0.0
3
4
5
0.0
Total Clay in delivered sediment:
Percent
Sediment Delivered:
Tons/acre
Total Clay Delivered
Tons/acre
(Percent Clay x delivered sediment)
Lbs./acre
6. Multiply columns (1) and (2) to get the total percent of clay in each class.
7. Add up the clay percentage in the 5 classes and enter the total in Table 4.
8. Copy the delivered sediment from this run and enter it into Table 4.
9. Multiply the percent clay in the delivered sediment by the amount of sediment delivered
and enter the value in the last box in both tons/acre and pounds per acre
(1 ton has 2000 lbs).
10. The same can be done for the fraction of silt of interest, for example, if interested in silt
between 4 microns and 10 microns, this is (10-4)/(62.5-4) or about a tenth of the
delivered silt fraction. Additional data about the fraction of silt in this category may be
available from previous studies. (Beyond the scope of this project.)
Questions?
APPENDIX C
Workshop Materials
WEPP Windows and WEPP-UI
TAHOE WEPP WORKSHOP
June 16 – 17, 2010
Tahoe Center for Environmental Sciences
Sierra Nevada College
Incline Village, Nevada
Training Workshop Materials for the
Application of the WEPP Windows, GeoWEPP, and
WEPP-UI Watershed Model to the Tahoe Basin
Erin Brooks
Dept. Bio. and Ag. Eng
University of Idaho
Moscow, ID 83844
ebrooks@uidaho.edu
Table of Contents
Importing the Tahoe Database to WEPP Windows …………………………………………………….. 1
WEPP Windows Applications ……………………………………………………………………………………..
- Building a Hillslope Project ……………………………………………………………………………….
- Hillslope Shape Effects and the Importance of Multiple OFEs …………………………..
- Importance of Depth to Bedrock ………………………………………………………………………
- Water Yield, Lateral Flow, and Snow Cover ………………………………………………………
2
2
2
4
5
GeoWEPP Applications …………………………………………………………………………………………….. 8
- Creating GeoWEPP ASCII Input Files for Region of Interest Using ArcGIS ………… 8
- Setting up a GeoWEPP Project for Lake Tahoe: Ward Creek Example …………….. 10
WEPP-UI Watershed Model Applications……………………………………………………………………
- GeoWEPP Limitations ……………………………………………………………………………………..
- WEPP-UI Watershed Overview ………………………………………………………………………..
- WEPP-UI Watershed Directory Structure …………………………………………………………
- WEPP-UI Watershed Step-By-Step Procedures ………………………………………………..
14
14
15
16
17
Appendix A
- Generating PRISM Climate Statistics for WEPP Hillslopes ……………………………….. 22
ii
IMPORTING THE TAHOE DATABASE TO WEPP WINDOWS
Before importing the Tahoe Database for the WEPP windows program please download and install the
2010 version of the WEPP software and documentation from the following web site:
http://www.ars.usda.gov/Research/docs.htm?docid=18084
The installation package from the USDA-ARS web page includes a large database of soils, climate, and
vegetative management files for application to forestry, construction, and agricultural applications
throughout the US. However, this database only provides a single climate station for the entire Tahoe
basin and many of the Tahoe soils are not included in this database.
One of the primary goals of this project was to develop a WEPP database of Tahoe-specific soil, climate,
and management files. This database is included on the CD provided for you at the start of the
workshop. The following instructions describe how to import this Tahoe-specific database into the
WEPP windows program.
TAHOE WEPP CLIMATE FILES
Tahoe Specific Climate files are located in the following directory on the CD handed out at the
workshop.
E:\Tahoe_database\WEPP_Windows\Climates\Tahoe
Copy the “\Tahoe” directory and paste the directory into the following location on the local hard drive.
C:\Program Files\USDA-ARS\WEPP_2010\Data\climates\cligen\
TAHOE WEPP MANAGEMENT FILES
Tahoe Specific Management Files are located in the following directory on the CD handed out at the
workshop.
E:\Tahoe_database\WEPP_Windows\Managements\Tahoe
Copy the “\Tahoe” directory and paste the directory into the following location on the local hard drive.
C:\Program Files\USDA-ARS\WEPP_2010\Data\Managements\
TAHOE WEPP UserDB FILES
WEPP also needs UserDB files for each management practice. Copy all the files in the following
directory on the CD.
E:\Tahoe_database\WEPP_Windows\UserDB
Paste all the files to the following UserDB directory on your local hard drive.
C:\Program Files\USDA-ARS\WEPP_2010\Data\UserDB
TAHOE WEPP SOIL FILES
Tahoe Specific Soil Files are located in the following directory on the CD handed out at the workshop.
E:\Tahoe_database\WEPP_Windows\Soils\Tahoe
Copy the “\Tahoe” directory and paste the directory into the following location on the local hard drive.
C:\Program Files\USDA-ARS\WEPP_2010\Data\Soils\
1
TAHOE WEPP SLOPE FILES
There are a few specific slope files that we will be using in the workshop. These slope files are located
in the following directory on the CD handed out at the workshop.
E:\Tahoe_database\WEPP_Windows\Slopes\Tahoe
Copy the “\Tahoe” directory and paste the directory into the following location on the local hard drive.
C:\Program Files\USDA-ARS\WEPP_2010\Data\Slopes\
WEPP Windows Applications
The WEPP windows software program provides the users with more flexibility to deal with
more complicated applications than those addressed by the Tahoe Basin Sediment Model and
other FS-WEPP interfaces. However with this added flexibility comes added complexity. The
following examples will demonstrate some applications of the model that users in the Tahoe
Basin may be interested in.
I.
Building a hillslope project
a. Start the WEPP windows model
b. Check “Create a new hillslope project”
c. Under the “Select a State” drop down menu select “Tahoe” the under the
“Select a Station” drop down menu select “Rubicon #2 CA SNOTEL” then click
“Next”
d. In “Slope selection” window browse the different shapes then select “SShaped”, enter 500 ft for the length, 100 ft for the width, and 20% for the
steepness. Click the “Change” button then save this slope file as “S-Shaped20” in
the following directory (C:\Program Files\USDAARS\WEPP_2010\Data\Slopes\Tahoe\) then click “Next”.
e. Under the Management Selection drop down menu, enter the “Tahoe” directory
and select the “Tahoe_Bunch_Grass” management file, then click “Next”
f. Under the Soil Selection drop down menu, enter the “Tahoe” directory and
select the “Volcanic Bunch Grass” soil file, then click “Next”.
g. In the “Project Options” window, specify a 30 year simulation. Click the
“Browse” button and navigate to the following directory C:\Program Files\USDAARS\WEPP_2010\Data\Projects then create a directory called “Tahoe” and save
the project file as “S-Shaped Slope.prj”. Then click “Finish”.
II. Hillslope Shape Effects and the Importance of Multiple OFEs
Objective: Examine the importance of hillslope shape and the importance of using
multiple overland flow elements (OFEs).
2
a. This activity builds from the same WEPP project “S-Shaped Slope” built in the
proceeding steps. Note: this activity assumes WEPP is using English units. If
WEPP is currently using metric units then check off the “Use metric units” item
under the “Option” drop down menu at the top of the page.
b. Under the “Option” drop down menu select “Winter Processes”. Set the number
of Freeze/Thaw Layers in the top and bottom soil layers to 1 each. Also make
sure the “Apply water redistribution during freeze/thaw cycles” is not checked.
Set all thermal conductivity parameters to 0.1 and set the lower limit of
conductivity for frozen soil to 1.0. This ensures the simulations are not affected
by frozen soil. Frozen soil processes are not as important in forest soils as they
are in agricultural tilled soils. Click “ok” when finished.
c. Click the “Run options” button at the bottom of the page and select “Water
Output”, “Soil Output” and “Return Period Summary”. Enter 30 years for the
number of years to simulate. Then click “ok”
d. Click the “Run” button at the bottom of the page.
e. Enter the simulated precipitation, runoff, erosion rate, and sediment delivery
into Table 1.
f. Move mouse over the top layer of the hillslope profile on the screen. A window
should pop up saying “Tahoe Bunch Grass”. Right click this layer and select
“Insert break”. This tells WEPP to simulate the subsurface hydrology as two
distinct segments, rather than as a single plane.
g. Move the mouse over the uphill side of the vertical break line on this same top
layer. Right click this position and select “Change length”. Enter 300 ft for this
length. This moves the break line so that it is located a horizontal distance of
300 ft from the top of the hillslope.
h. Click the “run” button and write down the simulated values in the next line of
Table 1. Under the file menu click “save” to save the project at this point with
the 2 OFEs.
Question: Why did the simulated runoff change?
i.
On the left hand side of the window select the yellow folder under the “Slope
File” heading.
Question: What is the average percent slope for this hillslope? What is the steepest percent
slope for this hillslope?
j.
Create a Concave slope by editing the distance and slope fields such that the
slope(%) is 0% slope at a distance of 0 ft and the slope(%) is 40% at a distance of
500 ft. Note there should be only two rows of data.
k. Click the “SaveAs” button and save the slope file as “Concave20.slp” in the
“./slopes/tahoe/” directory. The click “Save”.
3
Question: What is the average percent slope for this hillslope? What is the steepest percent
slope for this hillslope?
l. Click the run button and enter the simulated values in table 1.
m. Move the mouse over the top OFE layer so that the “Tahoe Bunch Grass”
window is displayed again. Right click this layer and then select “delete” to
convert the hillslope back to 1 OFE. Select “Yes” when asked to “Delete
Management Section?”
n. Run the model again and enter the simulated values in table 1.
Question: How does the simulated values for this run compare to the 20% S-Shaped 1-OFE
run? What are the similarities and differences?
o. This time create a convex slope following steps j. and k. except this time the set
the slope(%) to 40% at a distance of 0 ft and set the slope(%) to 0% at a distance
of 500 ft. Save this slope under the name “Convex20.slp” in the
“./slopes/tahoe/” directory.
p. Run the WEPP model again with this new slope shape and write down the
simulated values for the Convex Slope 1-OFE simulation in Table 1.
q. Insert the upper OFE element again as in steps f and g then rerun the WEPP
model. Write down the simulated values for the Convex Slope 2-OFE in Table 1.
Table 1. WEPP simulated output for various hillslope shapes and multiple OFEs for a volcanic
bunch grass soil.
Management
Precip
Runoff
Erosion
Sed
(in.)
(in.)
Rate
Delivery
Slope
(t/a)
(t/a)
20% S-Shaped 1-OFE
Bunch Grass
43.03
10.79
0.036
0.034
20% S-Shaped 2-OFE
Bunch Grass
43.03
11.59
0.034
0.031
20% Concave 2-OFE
Bunch Grass
43.03
8.16
0.091
0.091
20% Concave 1-OFE
Bunch Grass
43.03
10.79
0.123
0.123
20% Convex 1-OFE
Bunch Grass
43.03
10.79
0.019
0.017
20% Convex 2-OFE
Bunch Grass
43.03
14.01
0.001
0.001
Question: How does using multiple OFEs affect runoff and erosion? How does hillslope shape
affect runoff and erosion?
III. Importance of Depth to Bedrock
Objectives: Learn how to modify soil properties. Explore the sensitivity of runoff and erosion
prediction to depth to bedrock.
4
a. Start this exercise using the same 20% S-Shaped 2-OFE Bunch Grass simulation
on a volcanic soil from the last exercise (the second simulation). You can reopen
the “S-Shaped Slope.prj” to get back to these settings. Enter the simulated
results from this simulation into the first line of Table 2.
b. Click the yellow folder under the “Soil Name” at the base of the screen. This
should open up the “Soil Database Layer”. Alternatively, you could right click the
soil layer on the screen, the lowest layer on the hillslope profile, and select
“View/edit”. Uncheck the “English Units” box on the lower left hand side of the
screen.
Question: What is the depth to restricting layer? What type of restricting layer is it? What are
the percent Sand and Clay?
c. Change the “Depth(mm)” for layer 1 to “800 mm”. Click “SaveAs” then create a
new directory called “revised” under the ./Soils/Tahoe/ and save the soil files as
“Volcanic Bunch Grass 800mm.sol”. Then click “save”. Run WEPP using this new
soil and enter the simulated results into Table 2.
d. Repeat step c. for a 1500 mm depth and then for a 200 mm depth. Write the
simulated results into Table 2.
e. For the last trial, change the soil depth back to 400 mm and then uncheck the
“Use Restrictive Layer” box. Enter the results of this simulation into Table 2.
Table 2. WEPP simulated output for a 500 ft long, 20% S-Shaped hillslope having 2-OFEs, in a
volcanic soil with bunch grass for vegetative cover.
Precip
Runoff
Erosion
Sed
Depth to Bedrock (mm)
(in.)
(in.)
Rate
Delivery
(t/a)
(t/a)
400
43.03
11.59
0.034
0.031
800
43.03
6.47
0.032
0.029
1500
43.03
1.82
0.004
0.003
200
43.03
15.85
0.057
0.045
No Restrictive Layer
43.03
0.29
0.004
0.004
Question: Why does the runoff and erosion drop with increasing soil depth? Why is the runoff
prediction sensitive and the erosion prediction relatively insensitive to soil depth at shallow soil
depths?
IV. Water Yield, Lateral Flow, and Snow Cover
Objectives: Become familiar with water output files and the impact of changing canopy cover
on snowmelt and evapotranspiration. Compare simulated snow water equivalent with
observed snow water equivalent at the Rubicon SNOTEL.
5
a. Start this application by re-opening the “S-Shaped Slope.prj” file.
b. Click the drop down menu under “Management” heading on the lower left hand
side of the screen and then select the “Tahoe_Old_Forest” management file. Do
this for both OFE segments.
c. Click the drop down menu under the “Soil Name” heading on the lower, middle
part of the screen and then select the “Volcanic Old Forest” soil file. Note: All
management files have an associated soil file. Therefore, whenever changing
management files you must also import the matching soil file.
d. Click on the yellow folder next to the management drop down window. This will
open up the management parameter file for the “Old Forest” file. Click on the
yellow folder next to the orange “initial conditions” box. Notice the third item
from the top is “Initial canopy cover (0-100%)”. By default this value is set to
90%. In the Tahoe management files neither canopy cover nor soil surface cover
(i.e initial interrill cover (row 7) and rill cover (row 11)) change throughout the
simulation. Changing these initial conditions will fix these conditions for the rest
of the simulation. We will come back and change the canopy cover later in this
exercise.
e. In this application we will be comparing simulated and observed snow water
equivalent (SWE) at the Rubicon SNOTEL. Rather than using a climate file
generated by the CLIGEN model we will use a climate file having the observed
data recorded at the Rubicon SNOTEL. Click the drop down menu under the
“Climate File” heading on the right panel of the screen and navigate to
“Rubicon_hourly” climate file under the
“climates\Tahoe\Observed_data\hourly_data\” directory. These data are from
1988 through 2006.
f. Click the “Run Options” button and change the “simulation years” to 17.
g. Run the simulation.
h. Click on the “Text Output” button then select “Water Output” from the WEPP
Output List and click the “View” button. This should open the WEPP water
output file in Notepad.
i. Open the “Rubicon_snotel_obs_data.xls” file provided on the workshop CD using
Microsoft excel.
j. Go to Notepad and select “Select all” under the “Edit” tab. The select “Copy”
under the “Edit” tab.
k. Go back to excel and paste the data into cell D1.
l. Highlight all the cells from D16 to D12438, then select “Text to Columns” button
under the “Data” Tab in Excel 2007. Note: there should be the same tool button
in earlier versions of excel.
m. Select the delimited option and check the “Space” box to delimit the columns.
n. You will notice each day has two output lines, one for each OFE. The upslope
OFE is OFE 1 and the downslope OFE is OFE 2.
6
o. Type in the following equation into cell H10
=SUMIF($E$21:$E$12438,1,H$21:H$12438)/17 This provides the average annual
precipitation for OFE 1.
p. Type in the following equation into cell H11
=SUMIF($E$21:$E$12438,2,H$21:H$12438)/17 This provides the average annual
precipitation for OFE 2.
q. Copy these cells and paste them into cells I10 through U11.
r. Paste the average annual values for each of the major water balance
components into Table 3. Compare these values to those calculated using the
same management file with an initial canopy cover of 25%.
Note: Here are a few definitions to help understand the WEPP water file.
Runoff Q: the total volume of runoff passing the outlet of the OFE divided by the entire
hillslope area upslope of the OFE.
Runoff QOFE: the total volume of runoff passing the outlet of the OFE divided by the area of
the OFE.
Lateral Subsurface Flow Latqcc: is the total volume of lateral flow passing the outlet of the OFE
divided by the area of the OFE.
Table 3. WEPP simulated water balance components for a 500 ft long, 20% S-Shaped hillslope
having 2-OFEs, in a volcanic soil with Old Forest for vegetative cover having 90% and 25%
canopy cover.
Canopy OFE
Precip
Runoff
EvapoPercolation
Lateral
Runoff
Cover
(P)
(Q)
Transpiration
(Dp)
Subsurface
QOFE
(mm/yr) (mm/yr)
(Ep+Es+Er)
(mm/yr)
Flow
(mm/yr)
(mm/yr)
(latqcc)
(mm/yr)
90%
1
1076
34.9
705.7
.7
308.4
34.9
90%
2
1076
104.7
775.7
1.0
525.0
261.8
25%
25%
1
2
1076
1076
38.3
103.8
599.7
680.9
1.1
1.5
412.5
777.1
38.3
259.5
Questions: What is the total water yield for this hillslope? Why does the largest amount of
runoff come off the flattest part of slope? What percentage of the total precipitation leaves the
hillslope as subsurface lateral flow? (hint: The second OFE represents 40% of the entire
hillslope area)
s. Highlight the simulated block of data in Excel (E21:U12438) then use the sort by
increase OFE (Column E). The simulated and observed SWE data can now be
plotted on the same chart for either OFE 1 or OFE 2.
7
Creating GeoWEPP ASCII Input Files for Region of Interest Using ArcGIS
Input:
10m_dem
Soils_2007
Landuse
Watersheds
Streams
DRG file (SID) Format (Download from http://tahoe.usgs.gov/ )
Output:
1.) dem.asc
2.) soilsmap.asc
3.) landcov.asc
4.) (optionally) DRG TIFF image
STEP A: CREATE DEM, SOILS, and LANDUSE ASCII FILES
1.) Open ArcGIS and “add” 10m_dem, soils_2007, landuse, DRG file (SID) format. Optionally add
in watersheds and streams
2.) Zoom in to (i.e. drop a box around) watershed/study area. Note: make sure to include entire
upslope area
3.) Under spatial analyst toolbar select “options”, then under the extent tab select “same as
display” for analysis extent and under “snap extent to” select “soils_2007”, then click ok.
4.) Select raster calculator under the spatial analyst toolbar.
5.) Double click on “10m_dem” and then click “Evaluate”.
6.) Rename the resulting map to “dem_ward” by first right clicking on the map named
“Calculation”, then click “Data” then “Make Permanent”. Create a directory called
“Ward_creek” then name the map “dem_ward”
7.) Repeat steps 4-6 for the soils and landuse maps. Name the resulting maps “soils_ward” and
“lu_ward”, respectively.
8.) Open ArcToolBox.
9.) Select “Conversion Tools”, then click “From Raster” then click “Raster to Ascii”.
10.)Navigate to the “ward_ck” directory then select “dem_ward” for input raster and save the
ouput as “dem.asc”.
11.)Repeat steps 9 and 10 for the “soils_ward” file and the “lu_ward” file saving them in the process
as “soilsmap.asc” and “landcov.asc”. Note: In earlier versions of GeoWEPP the filenames were
required to be “dem.asc”, “soilsmap.asc” and “landcov.asc”. This is not required in recent
versions but is used here for consistency.
STEP B: CREATE A DRG image in TIFF FORMAT FOR STUDY AREA.
The next step is to create a geo-referenced TIFF image of the DRG map. DRG maps are not required for
GeoWEPP but can be useful for interpreting output in GeoWEPP
1.) Right-click the drg.sid image, select data, then select export data.
2.) Select “Data frame (Current)” as the extent and select the “ward_ck” directory under location,
select TIFF under format, and provide the name “drg_ward.tif” as the map name.
NOTE on Soil and Landuse Association Files
8
GeoWEPP also requires the following four files
1.)
2.)
3.)
4.)
soilsmap.txt
soilsdb.txt
landcov.txt
landusedb.txt
These files relate the values in the soils and land use maps to existing WEPP soil and management input
files. For this project these files have been created for you. However it would be good for you to have
some understanding of these files in case you would like to add more soil or management files at some
later time.
The “soilsmap.txt” file relates the soil ID number (i.e. value) in the map to a specific soil type. The
“soilsdb.txt” file relates the soil types to a specific soil file in the WEPP directory.
Example of soilsmap.txt file:
"Value","Count","S_value"
-9999,,tallac
1652044,,beaches
1652103,,bidart
1652036,,burnlake
1652125,,cagwin
…
Example of the soilsdb.txt file:
callat|Tahoe\Volcanics\callat.sol
ellispeak|Tahoe\Volcanics\ellispeak.sol
kneeridge|Tahoe\Volcanics\kneeridge.sol
oxyaquic|Tahoe\Mixed\oxyaquic.sol
cagwin|Tahoe\Granitics\cagwin.sol
Likewise the ”landcov.txt” file relates the landuse ID number (i.e. value) in the map to a specific land
use. The “landusedb.txt associates the type of land cover to a specific management file in the WEPP
directory.
Example of landcov.txt file:
11 Open Water
12 Perennial Ice/Snow
21 Low Intensity Residential
22 High Intensity Residential
23 Commercial/Industrial/Transportation
…
Example of the landusedb.txt file:
High Intensity Residential|Tahoe\Tahoe_Skid_Trail.rot
Row Crops|Tahoe\Tahoe_Bunch_Grass.rot
grass|Tahoe\Tahoe_Sod_Grass.rot
Fallow|Tahoe\Tahoe_Bare_Slope.rot
9
Setting up a GeoWEPP Project for Lake Tahoe: Ward Creek Example
Required Input Files
1.)
2.)
3.)
4.)
5.)
6.)
7.)
8.)
Dem.asc
Soilsmap.asc
Landcov.asc
Soilsmap.txt
Soilsdb.txt
Landcov.txt
Landusedb.txt
(optionally) DRG TIFF image
Note: see instruction sheet entitled Creating “GeoWEPP ASCII Input Files for Region of Interest” if these
input files have not been created for you watershed
Step 1: Loading the Maps Into GeoWEPP
1.) The first step to begin an ArcGeoWEPP project is to double click the “ArcGeoWEPP.exe” file in
the “C:\ArcGeoWEPP” directory
2.) Click “OK” after reading the introduction window
3.) Since we have already provided the basic input data for Lake Tahoe we can Select “Use your
own GIS ASCII Data” then click “OK” after reading about the required map formatting.
4.) Create Workspace called “Ward_Ck” or whatever name relevant to the region of interest
5.) Navigate to the directory where the input files listed above are located for your region of
interest. In this workshop I have created these files for you and placed them in the
“gisascii_ward_ck” directory on the workshop CD. Then select “dem.asc”
6.) Click “yes” to add in the soilmap.asc file.
7.) Navigate to the “gisascii_ward_ck” directory then select “soilsmap.asc”
8.) Click “yes” to add soilsmap.txt file.
9.) Select the “soilsmap.txt” file
10.)Click “yes” to add the soilsdb.txt file then “add” this file.
11.)Click “yes” to add landuse file then select “landcov.asc”
12.)Click “yes”, then add landcov.txt file
13.)Click “yes”, then add landusedb.txt file
14.)Click “yes”, then add in the drg_ward.tif image
15.)Click “no” to indicate we don’t have another image file to add
16.)GeoWEPP will now load all the maps into ArcGIS and begin to create a stream network. Before
processing starts it will indicate that “This process may take some time”. After clicking “ok” the
program will begin calculating a stream network.
17.)Before displaying all the maps the program will ask if it should use the “USGS Classification”.
Click “yes”.
10
18.)Once all the maps are loaded, right click the display window and select “data frame properties”.
Under the “coordinate system” tab click “layers” then click “Topo Image”, then click “Nad27
UTM Zone 10N” then click “ok”
19.) After the display goes blank right-click the “Topo Image” map name under the list of layers on
the left side of the window, then select zoom to layer.
20.)All the maps should line up. At this time it would be a good idea to save the project. If you
move the mouse over the GeoWEPP toolbar you should find that the “Save Project” button is
second button to the right of the happy face!
STEP 2: Defining a stream network
GeoWEPP creates a stream network from the DEM using the ―critical source area‖ and
―minimum source length‖ criteria provided by the user. The default values are 5 ha and
100 m, respectively. The larger these values the less extensive the stream network. The
stream network is important because it is directly related to the number of hillslopes used
to simulate the hydrology and erosion of the watershed. The larger the stream network
the more hillslopes in the watershed. For a full discussion of these parameters see the ‖
GeoWEPP for ArcGIS 9 Manual.pdf‖ located in the ―ArcGeoWEPP‖ root directory.
In this example we will use a critical source area of 10 ha and a minimal source length 300 m.
1.) Find the “Modify Channel Network Delineation” button by moving the mouse over GeoWEPP
toolbar. Click this button then change the critical source area to 10 ha and change the minimal
source length to 300 m then click “Delineate New Network”.
2.) Zoom around the watershed to see how the delineated network changed. If satisfied then now
would be a good time to save the project again.
Step 3: Delineate Watershed
Once the stream network is defined the hillslopes and watershed should be delineated.
1.) For this workshop we will define the watershed outlet near the headwater of Ward creek,
specifically UTM coordinates 740470 m East and 4335530 m North (see figures 1 and 2). Select
the zoom in button on the GeoWEPP toolbar and zoom into this point. Note if you have the
DRG_ward.tif image active then this point is located right below the “a” in the word “Ward” in
the map.
2.) Find the “Select a Watershed Outlet Point (channel cell only)” button on the GeoWEPP toolbar.
Click this button then click on the map at the watershed outlet point. Note: you must click a
blue stream channel cell for the delineation to work.
3.) Click “Ok” to the window “Outlet Point Selected”
4.) Enter zone “10” at the prompt.
5.) After GeoWEPP processes this command zoom out to see the delineated hillslopes.
6.) Now would be a good time to save the project again.
Step 4: Simulate Erosion using Watershed Mode
11
Once the watershed is delineated and you are satisfied with the number of hillslopes a nd stream
network, you are now ready to run a simulation.
1.) Click the “Accept Watershed Delineation” button (i.e. the happy face) on the GeoWEPP toolbar .
2.) The program now asks for a climate file. Select the “Use Existing Climate File” button. Then
browse to the “Tahoe” climate directory and then select the “daily” directory and then select
“WardCreek_daily.cli” file. Note: If you cannot find a Tahoe directory in the climate file then
you haven’t loaded the Tahoe files into your computer. Please see the instruction sheet entitled
“Loading WEPP-Tahoe Base Files”.
3.) Select “yes” to the question “Use climate file previously selected?”
4.) A breakdown of the percent of the watershed in each landuse and soils will be displayed as well
as the type of stream bed material for each stream channel type. Click on the ”Channels” then
double-click the 01 order WEPP Channel Type entry. This should open a window that lists all the
channel types. Select the “Rock Channel” then click “OK”. Repeat this same process for all
stream orders. Note: this will minimize the stream channel erosion predicted by the model. The
stream channel algorithms in the GeoWEPP do not give reliable predictions for large
watersheds!
5.) Once all streams are converted to “OnRock” then click “Ok” to the main screen.
6.) Now you will see a window with a summary of all the hillslopes and stream channels. There
should be 43 Hillslopes and 17 channels. Change the number of years to “17” years. Note this is
the maximum amount of years in the climate files.
7.) Under simulation method select “Watershed” in the drop down menu.
8.) Then click the “Run WEPP button”. After this you should see GeoWEPP running through all the
hillslopes.
9.) Once WEPP has finished it will show a summary screen indicating how long it took to run the
simulation. Click “OK” to this screen.
10.)GeoWEPP will then ask for a map layer name for the “OffSite” sediment yield predictions. Click
“Ok” to the reminder window that all filenames should be less than 13 characters.
11.)Type in “OffSWdlyOrig” for the map name then click “Ok”. The model will create an output file
showing the total delivered sediment for each hillslope.
12.)Now would be a good time to save the project again.
Advanced Options
Once the model has been successfully imported and simulated, the user has several advanced
options for exploring management options, these include:
1.
Importing one of the hillslopes into the WEPP windows model to make detailed
comparisons of management options.
2. Re-run GeoWEPP using the “flow path” option. Under this option all flow paths, defined by
the width of a single cell in the DEM, are simulated separately by the model. Since there are
many more flow paths then hillslopes we recommend reselecting a watershed outlet point
for a much smaller watershed or simulating fewer years.
12
3. Another advanced option is to re-associate the management files of critical eroding slopes
to an alternative treatment, then re-run the model.
4. GeoWEPP can also be used in return period analyses.
We highly recommend reading through the GeoWEPP manual to further explore the advanced
options.
Watershed
Outlet
Figure 1. Watershed outlet point for Ward Creek example.
Watershed
Outlet
Figure 2. Close up view of watershed outlet point for Ward Creek example.
13
WEPP-UI Watershed Approach
GeoWEPP Limitations
Over the last 5 years the University of Idaho has been developing a technique to apply the
WEPP model to large watersheds which avoids some of the major assumptions associated with
GeoWEPP applications. Particularly the following five assumptions: uniform climate, fixed
vegetative management files, planar lateral subsurface flow, soil texture-based hydraulic
properties, and neglecting base flow. Each of these will be discussed below.
Uniform Climate Assumption
In GeoWEPP applications it is assumed that the climate does not vary across the watershed. In
mountainous regions, particularly those where precipitation falls primarily as snow, this is a
very poor assumption which can drastically affect simulated streamflow.
In the WEPP-UI watershed approach mean monthly PRISM precipitation, maximum
temperature, and minimum temperature are used to create unique climate files for each
hillslope.
Fixed Vegetative Management File Assumption
A second assumption in GeoWEPP is that vegetative attributes for each land type do not vary
across the watershed. For example, all old-growth forested regions will have the same canopy
cover as defined primarily by the initial condition. In the Tahoe region canopy cover can vary
greatly within a single land type.
In the WEPP-UI watershed approach mean percent canopy cover for each hillslope is used to
create unique management files for each hillslope.
Planar Lateral Subsurface Flow
A third limitation of GeoWEPP is related to basic algorithm used to simulate subsurface lateral
flow. In steep mountainous terrain, underlying bedrock often greatly limits downward vertical
movement of water leading to subsurface lateral flow, sometimes called interflow or
subsurface stormflow. Subsurface lateral flow on steep portions of a slope can lead to soil
saturation at flatter toe slope positions where subsurface lateral flow slows and converges.
Although GeoWEPP will simulate subsurface lateral flow, only the average slope steepness of
the hillslope is used simulate water flow and as a result no downslope subsurface convergence
is simulated. This lateral convergence can be the difference between water leaving the
hillslope as lateral flow versus leaving the hillslope as surface runoff and potentially
transporting sediment.
In the WEPP-UI watershed approach hillslopes are broken down into multiple overland flow
elements (OFEs) which allows for downslope lateral convergence. Note: to avoid numeric
problems on steep, short OFEs the WEPP model was rewritten so that calculations are made on
an hourly time step rather than a daily time step. The new executable program is provided on
the workshop CD.
14
Soil Texture Based Hydraulic Properties
A fourth limitation is that several of the key soil hydraulic characteristics related to soil water
storage (e.g. bulk density, field capacity, wilting point) and transport (e.g. saturated hydraulic
conductivity) are estimated based on soil texture. This may be unimportant when applying the
model to disturbed regions (i.e. roads, skid trails, ski runs) where surface hydraulic properties
largely control surface runoff, however having the correct water storage characteristics in
watershed applications can be important.
In the WEPP-UI watershed approach a soil input format is used which allows the user to provide
bulk density, saturated hydraulic conductivity, field capacity moisture content, and wilting point
moisture content for each soil layer. Values for these parameters were taken directly from the
2007 Tahoe Basin Soil Survey.
Neglecting Baseflow
Another limitation of GeoWEPP is that it does not simulate baseflow. Baseflow is that portion
of the streamflow that is supplied by bedrock springs. GeoWEPP, as with the original WEPP
watershed model, will only simulate surface runoff and subsurface lateral flow, not baseflow.
Direct comparison between observed and simulation streamflow can only be made after
baseflow has been manually separated out of the observed data using assumed separation
techniques.
In the WEPP-UI approach a simple linear reservoir approach is assumed where baseflow is
linearly related to the volume of water storage in subsurface aquifer. This linear reservoir
coefficient is a constant value which can be estimated from observed streamflow data. Flow
into the aquifer is assumed to be all water which percolates through the bedrock restrictive
layer as simulated by the WEPP model. Flow out is calculated as the linear reservoir coefficient
multiplied by the amount of water stored in the aquifer.
WEPP-UI Watershed Overview
The WEPP-UI Watershed Approach is a technique used to provide watershed outlet streamflow and
sediment load predictions. In-stream sediment erosion and deposition are not simulated. The input
files provided on the workshop CD are for undisturbed soils and reflect native conditions. As a result
this should be considered an approach for simulating background streamflow and sediment load for
Lake Tahoe watersheds. We recommend using the web interface or WEPP-Windows approach to
address hillslope scale disturbances and for assessing the impact of various management practices.
As with most watershed hydrology models, the primary challenge in using the WEPP-UI Watershed
approach is setting up the watershed database and getting all the input files in order. This is especially
challenging in the WEPP-UI Watershed approach since it is still under development. In the future we are
hoping many of the tedious steps will be automated to make this approach more widely used and
accepted. At this point, some of the steps require a fairly high level of familiarity with excel
spreadsheets, ArcGIS, and, in some cases, Arc/INFO. The approach also uses PERL scripts so familiarity
with this programming language is also useful.
In the following description we provide a description of the format of the major input files and
recommend different approaches for developing those input files. There are many ways to create these
15
files so use whatever approach you are most comfortable with. The input files provided on the CD are
applicable for undisturbed soils. We recommend
We have provided two example watershed applications using the WEPP-UI watershed approach on the
CD passed out at the workshop. One example is for the Blackwood creek watershed using observed
daily climate data from the Ward Creek SNOTEL, called “blackwood_ck_daily”. Another example is using
hourly observed data for the same watershed, see “blackwood_ck_hourly”. We also provided a
directory called “base_files” which is an empty directory which only includes the necessary executable
programs and PERL scripts which you can populate with your own watershed input files. Take some
time to explore through these directories so to become familiar with the file and directory structures.
WEPP-UI Watershed Directory Structure
There are five required directories and one optional directory in the WEPP watershed approach. Each
directory contains unique input files that must be populated before running the model. Each of these
directories are described below.
Cli directory: This is the Climate directory which stores the base climate file (*.cli) which is identical to
those used by the WEPP windows model. It also stores the “slope_monthly_prism_ward.txt”
file which has climate, canopy cover, and vegetative management properties of each hillslope.
See the detailed description of this file below.
Man directory: This is the management file directory which stores all the potential vegetative
management options for the watershed. In this directory you will see two different
management files. These are the hermada_*.man and the shrub_rangeland_*.man files. Note:
the hermada_*.man file is similar to the “Old Forest.rot” file used in the WEPP Windows
application. Each management file has 19 different variants. These numbers represent
different numbers of OFEs in each hillslope. The model will select a particular management file
that matches the number of OFEs in each hillslope.
Output directory: This is an optional directory. In the Blackwood creek examples provided on the
workshop CD you will find the excel spreadsheet used to compare simulated and observed
sediment load and streamflow at the watershed outlet.
Run directory: This directory stores the “run” files. Run files automatically communicate the userpreferred input and output options to the WEPP-UI executable program (BrLF3.exe). These files
make it possible to run many WEPP hillslopes without having to manually interact with the
program. As seen in the blackwood creek examples, each hillslope has its own run file. These
run files are created using the “runfile_creator.pl” PERL script and the “input2runfile_*.txt” file.
See the file descriptions below for details about how to apply and populate these files.
Slope directory: This directory stores the slope files for each hillslope. In the WEPP-UI approach
hillslopes are broken down into multiple OFEs. This is accomplished by running the
“slope_ofe_transformation.pl” PERL script in the GeoWEPP subdirectory. GeoWEPP hillslopes
must be pasted to this directory before running this PERL script. See the explanation below for
specific instructions.
16
Soil directory: This directory stores all the potential soil input files for the Tahoe basin. Like the
management files there are 19 variants of each soil type which are selected based on the
number of OFEs used in a particular hillslope. All the soil properties in these soil files are for
undisturbed soils and should not be assumed to be representative of disturbed soils.
The root directory contains the following three files:
BrLF3.exe This is the WEPP-UI executable program that has several algorithms that have not been
included on the latest public release versions of the WEPP program, the most important being
an hourly looping algorithm to avoid numeric problems with short, steep, OFE slopes.
wepp_UI_post_processing.pl This is a post processing Perl script used to create a single daily watershed
output file (totalwatsed.txt). See the detailed description below.
wepp_daily_weather.pl (also the wepp_hourly_weather.pl) This is main program used to direct the
WEPP processing for each individual hillslope. Once all the file directories are populated the
program will begins by running this file. See the detailed description of this file below.
WEPP-UI Watershed Step-By-Step Procedures
Software requirements: The WEPP-UI watershed approach requires that WEPP-Windows,
GeoWEPP, and PERL script software be installed on the computer. These can be downloaded at
the following web sites:
WEPP software and documentation:
http://www.ars.usda.gov/Research/docs.htm?docid=18084
GeoWEPP software & documentation site: http://www.geog.buffalo.edu/~rensch/geowepp/
PERL Programming: http://www.activestate.com/activeperl/downloads
The PERL scripts used in the WEPP-UI Watershed approach also require that the rm.bat and
ls.bat files are pasted into the c:\windows\system32 folder. These files can be found in the
E:\Tahoe_database\UI_WEPP\unix_batch_files\ directory on the workshop CD.
STEP A: Run GeoWEPP and Copy Base Files
1.) The first step in applying the WEPP-UI Watershed approach is to setup the watershed and run
the GeoWEPP model (See GeoWEPP instructions.docx file for instructions).
2.) Copy the “UI_WEPP\Base_Files” directory, including all sub-directories, from the CD provided in
the workshop to a location on your hard drive outside GeoWEPP to run the WEPP-UI Watershed
Model
a. You can rename this directory after the watershed you will be simulating (e.g.
Ward_ck_WEPP)
3.) Navigate to the ”.\ArcGeoWEPP\Projects\{project_name}\” directory and copy
a. topaz2wepp.txt
b. weppshed.txt
4.) Paste these files into the “Ward_ck_WEPP\run\” directory
17
5.) Navigate to the “.\USDA-ARS\WEPP_2010\runs\” directory and copy all the H*_slp.txt files
(where * is the WEPP hillslope ID number)
6.) Paste these files into the “\slope\geowepp\” directory
STEP B: Create Slope Files
1.) Navigate to the \slope\geowepp\ directory and double click the PERL script called
“slope_OFE_transformation.pl”. This PERL script takes the single OFE GeoWEPP slope files and
breaks then into multiple OFE slope files and renames them H*.slp. Note: if the PERL script
does not work make sure PERL software is installed on your computer and the rm.bat and ls.bat
files have been pasted into the c:\windows\system32 folder.
2.) Move all the H*.slp files to the “\slope\ directory.
3.) The slope_ofe_transformation.pl will also create a file called “slope_ofe.out“ The values in
each row from left to right are: WEPP_ID, Number_of_OFEs, hillslope width in meters, hillslope
length in meters, and hillsope area in square meters.
STEP C: Create Run Files
One the tedious necessary steps for applying the WEPP-UI approach is to create the input2runfile.txt file
from the topaz2wepp.txt, weppshed.txt, and slope_ofe.out files. There are undoubtedly numerous
ways to create this file. We suggest using an excel spreadsheet.
The input2runfile.txt file is a table having seven fields where each row describes the characteristics of a
unique hillslope. The seven fields from left to right are: Topaz_ID, WEPP_ID, OFEs, Area_cells, Soil_File,
Management_File, and Slope_File. A PERL script program called “runfile_creator.pl” uses this text file to
create “run files” for each hillslope. The WEPP model uses the “run file” to locate the primary input files
for each hillslope (e.g. slope, soil, management, climate files) and to indicate where output files will be
saved.
The input2runfile.txt file can be created by first merging both the topaz2wepp.txt and weppshed.txt file
in excel. The WEPP-UI approach does not simulate channel erosion and therefore all rows related to
channel characteristics in the topaz2wepp.txt and weppshed.txt files should be eliminated from these
files. Once imported into excel the topaz2wepp.txt and weppshed.txt files can be delimited into their
primary soil and management file names using the “text to columns” function. Then by using the “sort”
command the soil and management files can be grouped together. Make sure that the WEPP-ID
numbers are passed along in this step. Once all the slope and management files are sorted, then should
can organized into a table with Topaz_ID, WEPP_ID, Area_cells, Soil name and Management name. Next
the “slope_ofe.out” file located in the “\slope\geowepp\” directory can be imported into excel and used
to provide number of OFEs for each slope. Finally the “concatenate” function in excel can be used to
generate soil, management, and slope file names with the associated number of OFEs. Finally the file
can be saved as a tab-delimited text file called “input2runfile.txt” in the “\run\” directory.
Once the input2runfile.txt is created then double click the “runfile_creator.pl” file to generate run files
for each hillslope.
18
STEP D: Create the slope_monthly_prism_wshed.txt file
The slope_monthly_prism_wshed.txt file is called by the wepp_daily_weather.pl file and is used to
develop hillslope-specific climate files. The file is created manually in excel and saved as a text file. See
the section called “Generating PRISM Monthly Climate Statistics for WEPP Hillslopes” for using ArcGIS
to calculate mean monthly precipitation and temperature from 800 m resolution PRISM maps and for
calculating mean canopy cover for each hillslope. Temperature should be provided as degrees Celsius.
Precipitation should be provided in units of millimeters. Elevation should be in units of meters. A line
by line description of the file is provided below.
The first three lines describe the climate attributes of the base climate station
LINE 1: A one word name for the base climate station (no spaces) followed by mean annual
precipitation for each month starting in January and ending in December for the 800 x 800 m
PRISM grid cell where the base station is located.
LINE 2: Same as LINE 1 except for mean monthly maximum daily air temperature for the base station
according to the PRISM map instead of precipitation.
LINE 3: Same as LINE 1 except for mean monthly minimum daily air temperature for the base station
according to the PRISM map instead of precipitation.
LINE 4: A field heading line not used by the model
The remaining lines in this file describe the climate attributes for each WEPP hillslope. Each hillslope is
described by a block of 4 lines.
LINE 5: WEPP ID number for the hillslope, mean annual precipitation for the hillslope, mean annual
maximum daily air temperature for the hillslope, mean annual minimum daily air temperature
for the hillslope, mean elevation of the hillslope, mean fractional canopy cover for the hillslope,
name of the WEPP management file for the hillslope.
LINE 6: Mean monthly precipitation for the hillslope starting with January and ending with December.
LINE 7: Mean monthly maximum temperature for the hillslope starting with January and ending with
December.
LINE 8: Mean monthly maximum temperature for the hillslope starting with January and ending with
December.
The remaining lines in the file are a repeat of lines 5-8 for each unique WEPP hillslope in the watershed.
STEP E: Edit the wepp_daily_weather.pl (or the wepp_hourly_weather.pl) file
The final step before running the model is to edit the wepp_daily_weather.pl if using a daily
climate file or wepp_hourly_weather.pl if you are using an hourly climate file to make sure the
base climate file and the name of the slope_monthly_prism_wshed.txt file match the files in the
cli directory. This file can be edited using notepad, wordpad, or any other text editor program.
STEP F: Run the wepp_daily_weather.pl (or the wepp_hourly_weather.pl) file.
19
It is recommended that all PERL script applications be run using the “command prompt”
window rather than double-clicking the PERL script. This makes it easier to see if the program
has any problems. To run the program just type in wepp_daily_weather.pl on the command
line. This initiates the program and you will see *_wat.txt and *_pass.txt files being generated
in the root directory. This process can take several hours depending upon the number of
hillslopes and the processing speed of the computer. After the model completely finishes you
should find the final output file “totalwatsed.txt” in the root directory.
The only difference between the wepp_daily_weather.pl and the wepp_hourly_weather.pl
programs is the routine that corrects the precipitation. Note: if the PERL script does not
work make sure PERL software is installed on your computer and the rm.bat and ls.bat
files have been pasted into the c:\windows\system32 folder.
Step1: The program first calls the slope_monthly_prism_watershed.txt to acquire climate information
for the base climate station.
Step2: The program then iteratively steps through each hillslope, using the information in the
slope_monthly_prism_watershed.txt file for each hillslope to create modified climate and
management files. Precipitation for each day for a specific hillslope is set equal to the
precipitation recorded at the base climate station multiplied by the ratio of mean monthly
precipitation at each hillslope and the mean monthly precipitation at the base climate station
according to the PRISM map. Maximum daily air temperature for each hillslope for each day is
set equal to the difference between the mean monthly maximum daily air temperature at the
hillslope and at the base climate station added to the maximum air temperature measured at
the base station for the specific day. Minimum air temperature is treated similarly to maximum
air temperature.
Step 3: Management files are modified next to ensure that the forested canopy cover simulated in the
model matches observed values. It is assumed forest canopy cover does vary significantly with
time.
Step 4: The program calls the UI-WEPP program (i.e. BrLF3.exe) and iterative runs the model using “run”
files for each hillslope from the ./run/ directory.
Step 5: After all the hillslopes are simulated the program calls the post-processing PERL script
called “wepp_UI_post_processing.pl”. This program merges daily output files from
each hillslope (*_wat.txt, *_pass.txt) to create a single output file with daily output for
the outlet of the watershed called totalwatsed.txt.
STEP G: Import the watershed output file (totalwatsed.txt) into a watershed output
spreadsheet
The totalwadsed.txt file is the output text file generated from the wepp_UI_post_processing.pl perl
script for the watershed output. This file can imported into the final watershed spreadsheet to
make comparisons with observed data (see
UI_WEPP\blackwood_ck_daily\output\blackwood_daily_ward_output.xlsx for an example of
this spreadsheet).
Each line in the totalwatsed.txt file includes the following information:
20
Day of the year, year, watershed area in square meters, precipitation, rain, transpiration, evaporation,
percolation, runoff, lateral, soil water storage, detached total sediment, deposited total
sediment, delivered total sediment, delivered clay, delivered silt, delivered small aggregates,
delivered large aggregates, delivered sand, and streamflow
All water units are in cubic meters per day and all sediment units are in kilograms per day.
21
Appendix A.
Generating PRISM Climate Statistics for WEPP Hillslopes
This protocol can be used to acquire mean monthly precipitation, maximum temperature and
minimum temperature for each hillslope in watershed. The same approach can be used for
calculating the mean fractional canopy cover and mean elevation for each hillslope from
canopy cover and DEM maps of the watershed.
Input
1.) PRISM mean monthly maps for precipitation, maximum temperature and minimum
temperature http://www.prism.oregonstate.edu/
2.) A shape file or grid which has a unique ID number for each hillslope.
3.) PERL scripts called “precip_clip_resample.aml”, “tmx_clip_resample.aml”,
“tmn_clip_resample.aml”
4.) PERL scripts called tmin_stats.aml, tmax_stats.aml, precip_stats.aml
Output
1.) A table which provides mean monthly temperature and precipitation for each hillslope
Steps
1.) Download mean monthly precipitation, maximum temperature and minimum
temperature for the United States. In this example we will be using 800 m resolution
maps which provide thirty year averages for the 1971-2000 time period. All maps are
downloaded as *.gz files from the http://www.prism.oregonstate.edu/ web page.
Maps are in North American NAD83 world projection (GCS_North_American_1983)
where distances are measured by degrees.
2.) Extract the ascii grid files. Create three separate directories called “precip_grids”,
“tmax_grids” and “tmin_grids”. Extract the ascii files to their respective directories. An
example of the naming convention for January precipitation is “us_ppt_1971_2000.01”.
Tmax and Tmin maps for January are named similarly as “us_tmax_1971_2000.01” and
“us_tmin_1971_2000.01”, respectively.
3.) Reproject WEPP hillslope map into the same projection as the PRISM maps. WEPP
hillslope maps are created using the GEOWEPP program which will typically use UTM
coordinates where distances are measured in meters. This shape file can be converted
to the same project as the PRISM maps using the “project” command in ArcMap.
4.) Convert the WEPP hillslope map to a grid file. This can be achieved using “Feature to
Raster” command in ArcMap which allows the user to specify the “ouput cell size”. It is
22
5.)
6.)
7.)
8.)
9.)
important that the cell size be set to 0.0001041667 degrees. In this example we call the
grid file “pcw_hills”. Copy this grid file to each of the precip, tmax, and tmin grid
directories created in step 2.
Identify the spatial extent of the hillslope map. Open one of the PRISM maps and
overlay the hillslope map. Zoom into the map and identify the latitude and longitude
values which define the extent of the watershed. It is important that the boundaries be
taken at the edge of grid cells (i.e. where one grid cell ends and the next begins) so that
maps created using these boundaries do not split midway through a grid cell. Write
down the east, west, north, and south boundaries in degrees.
Enter boundaries into the “*_clip_resample.aml” files. Copy the
“*_clip_resample.aml” files to their appropriate directories (i.e.
precip_clip_resample.aml file should be copied to the “precip_grids” directory). Then
open up these files using an editor (i.e. word pad) and enter the boundaries of the
watershed acquired in step 5 into the first line of the file on the line starting with
“SETWINDOW”. Enter the boundaries in the following order: “Western_Longitude”,
“Southern_latitude”, “Eastern Longitude”, and “Northern Latitude”. All values should
be entered in decimal degrees.
Run the “*_clip_resample.aml” scripts to create clipped, resample grid maps. Open
ARC/INFO and navigate to the “precip_grids” directory. Enter into the “grid”
subprogram in ARC/INFO. Then type “&run precip_clip_resample.aml”. This program
will open each of the PRISM precipitation grid maps for the entire US, clips out only the
watershed area of interest, and resamples the grid to a cell size of 0.00010416667
degrees. The clipped maps will be called “pcp_{month}_clip” (i.e. pcp_jan_clip). An
annual average called “pcp_ann_clip” will also be created. Navigate to the “tmax_grids”
and “tmin_grids” directories and run the “tmax_clip_resample.aml” and
“tmin_clip_resample.aml” scripts, respectively.
Run the “*_stats.aml” scripts to generate output files (*.e00) with mean statistics for
each hillslope. Navigate to each of the directories and run the “*_stats.aml” scripts (i.e.
&run precip_stats.aml). In each directory *.e00 files will be created for each month (i.e.
precip_jan_pcw.e00).
Use *.e00 files to create a table which lists the monthly averages for each hillslope.
Open all the *.e00 files for a particular directory (i.e. precip_grids) in excel. These *.e00
files have 7 header lines then contain many rows with 4 values. The first value is the
hillslope ID number in the “pcw_hills” grid file. The second number represents the
number of grid cells that have the particular hillslope ID number in the “pcw_hills” grid
file. The third number represents the area in (square degrees!) for each hillslope. Most
importantly the fourth number represents the aerial weighted average PRISM value
within each hillslope for that particular month. Precipitation maps will have units of
23
mm*100. Tmax and Tmin maps will have units of Celsius * 100. Use the copy and paste
function to create a single table in excel with “Hillslope_ID” as the first column followed
by averages for each month starting in January ending with December. The last column
should be the annual average. Create separate precipitation, tmax, and tmin tables in
excel. Convert the monthly temperatures to Celsius by dividing to 100 and convert the
precipitation values to mm by dividing by 100. Save these tables as separate
worksheets in a spreadsheet called “PRISM_averages_by_hillslope.xls”.
Alternative Approach: Rather than using the *.aml scripts, it is also possible to generate these
statistics using “Zonal Statistics as Table” in ArcMap. With this command the user can select
the hillslope map as the “Input raster or feature zone data” and the PRISM map as the “Input
value raster”. The disadvantage of this approach is having to replicate this command for each
PRISM map. Since there are 39 different monthly maps this would require replicating this
command 39 different times. ArcMap will create separate output tables for each monthly map
which should be imported into spreadsheet and merged following the same approach as step 9
above.
24