Completion of the GIS analysis for the Indian-McGregor Creek sub-watershed demonstrated the ease and efficiency spatial technologies allow users to evaluate a large area. Based off a 2.5m DEM derived from LiDAR point data, it became apparent this watershed’s erosion and stream power were highly influenced by the local topography. While resulting in a potential soil erosion map and SPI map, the predictive power of these tools were assessed against the local characteristics.
Due to the use of the revised RUSLE and the large-scale location in this assessment, the potential soil erosion map generated was dominated by slope. Using Equation (2), only the region’s rainfall and runoff factor (R), soil erodibility factor (K), and LS, the slope steepness and length factor were required to estimate erosion, and at this scale, R and K had little variation over the study site. This led to LS dominating the study location, with increases in slope, up to 62° in some locations along the main channel, resulting in increased erosion potential over this very flat watershed (Figure 5). It was discovered that the areas at highest risk of erosion are the sites closest to the creek and its tributaries, these areas had calculated values over 53000, much greater than the results of Miller (2018) where the same parameters gave a result of 107.2 for a field with a 4.3° slope, with the study site having an estimated erosion rate of 28.51 t/ha/yr after including all 5 parameters, which is above the acceptable rate. Much of the surrounding land at a very low risk due to its flat topography, whereas steeper channel walls and banks, and any other sudden changes in topography, are much more susceptible to erosion (Figure 5). There areas are also influenced by the SPI, with SPI being greatest in the same regions as the highest erosion rates, suggesting higher SPI results in higher erosion rates (Figure 6).While the produced potential soil erosion map cannot directly predict soil loss in tonnes per hectare per year, it serves as a baseline to identify high risk areas and can be effectively derived using programs such as ArcGIS.
Figure 5: Generated RUSLE Erosion Risk map (top) and associated large scale parameter maps for R, K, LS,
respectfully (top to bottom mini-maps), and large scale erosion risk map (bottom mini-map) for the same location.
SPI analysis was similarly influenced by slope, having been composed of the watershed’s slope and flow accumulation patterns. Although there is a noticeable pattern produced by the SPI analysis, the Indian-McGregor Creek watershed has very few locations with a high stream power index, with the only notably high values occurring within the main creek channels, which can be expected on such a flat landscape (Figure 6). The high SPI locations within the large channels also coincide with the identified areas with higher slopes, and therefore, more energy resulting in more erosion. These results ultimately suggest that while SPI analysis is possible to complete on level terrain, it may not be sensitive enough at this resolution to be effectively applied on a sub-watershed scale such as this study has done and may benefit from a finer resolution dataset.
Figure 6: SPI for the McGregor sub-watershed (top), determined using WhiteboxGAT’s D8 Method, and a large-scale excerpt for visibility (bottom).
The potential soil erosion map, and SPI analysis, were also able to accurately predict locations with higher erosion. When compared to imagery of the study site, erosion sites were clearly identified as locations with higher risk, relative to the locations slope and gradually decreasing as distance from the erosional feature, like a gully, increased. This can be expected as an erosional feature’s influence, such as increase slope, would be less with increased distance. The accurate placement of high erosion sites on recorded erosional features indicates that these methods can be confidently used to predict sites where erosion could become a hazard if not controlled.
