Road segment effect on soil loss prediction
Figure 5 illustrates the GeoWEPP “Flowpath” mode simulated onsite soil loss for the study area under the post-fire conditions. The paved road surface showed minimal soil loss and can be distinguished from other areas with the yellow or green deposition or zero erosion categories. In some locations, deposition (yellow pixels) was predicted to occur on the road surface when a flow path reached a road segment. Detailed deposition on the surface of road 1 was simulated by running 22 different hillslope profiles that intersected with Road 1 in WEPP Windows. From these runs, the average annual sediment deposition along the road segment within the fire was calculated as 40.4 kg m-2, or a sediment deposition depth averaging 4 cm on the Road 1 surface.
To evaluate the surface hydrology unit effect on soil loss due to road influence, the onsite soil loss was classified into 6 categories and overlaid with sub-catchments. Table 5 shows that sub-catchments that originated from the ridge resulted in a larger fraction of the high soil loss category (greater than 4 Mg ha-1yr-1) than the sub-catchments below roads. The average soil loss rate also showed that sub-catchments below the main ridgeline had the greatest onsite soil loss. Similarly, the sub-catchments below roads showed relatively more deposition and much lower average soil loss rates than those above roads. For sub-catchments below road 2, the average predicted onsite soil loss was negative, indicating that deposition was the dominant process.
To explore soil erosion variation with terrain positions in more detail, the onsite soil loss rate for different distances from ridge and road segments were measured by buffer analysis. The predicted onsite soil loss rate was extracted for each 10-m-width belt, and then the average value was calculated and plotted with distance (Figure 6). It can be seen that the average onsite soil loss rate below ridge increased exponentially with the downslope distance. On the other hand, the average soil loss rates below both road 1 and road 2 showed a logarithmically declining trend with the downslope distance. The first 10-m buffer belt below roads showed remarkably higher soil loss rate compared to belts further below. The average onsite soil loss within the first 10-m belt below road 1 is close to that of the last belt below the ridge. Similarly, the average onsite soil loss rate within the first belt below road 2 was similar to the last belt below road 1. When the flow path goes beyond 50 m downslope from road 2, deposition tends to occur. This is consistent with the earlier results that measured erosion by sub-catchment and observations in the field. Note that the average onsite soil loss rate 70 m below road 2 (Figure 6.c) showed an abnormally low value due to the high deposition rates predicted within the channels. This low value was not included when developing the regression equation shown on the graph.
The GeoWEPP Watershed analysis divides the study area into hillslope polygons, each with a representative hillslope profile (Flanagan et al., 2013). From the GeoWEPP watershed outputs, one of the hillslope polygons that passed through road 1 was selected and run on its own in WEPP Windows to generate a typical soil loss graph along the hillslope profile (Figure 7). Figure 7 shows that the erosion rate generally increased as the slope length increased from 0 to 140 m due to runoff accumulation. When reaching the road segment, the erosion rate dramatically decreased and deposition occurred on the road surface section. As runoff exited the road and continued flowing downslope, soil loss rate increased sharply within a short distance below road segment, likely due to the steep fillslope and clearer water. The loss reached a maximum value for the profile, and then gradually declined as the increase in runoff due to increased slope length was offset by a decrease in the hillslope steepness.