AN ACCURACY ASSESSMENT OF LIDAR-DERIVED HILLSLOPE BEDROCK EXPOSURE MAPPING TOOLS IN STEEP LANDSCAPES

Quantifying patterns of bare-bedrock exposure on hillslopes is important for understanding the controls on soil production and erosion and for predicting sediment delivery into river networks. Bare-Earth lidar topography highlights the contrast between soil and bare rock on hillslopes, and is less impacted by vegetation cover than imagery data. However, it is not clear which topographic metrics and parameters best resolve patterns of bedrock exposure in a landscape. This study aims to assess the accuracy of two topographic bedrock exposure mapping algorithms using three study areas in southeastern California: McElvoy Canyon in the Inyo Mountains, and the Sacatar Wilderness and Olancha Peak regions of the eastern Sierra Nevada. The first algorithm is slope based, where pixels with a slope greater than a given threshold are defined as bedrock. The second algorithm is roughness based, and assumes that surfaces that are rougher than a threshold value have more bedrock exposed. We used a variety of window sizes, roughness thresholds, and slope thresholds to generate binary maps of hillslope bedrock exposure, which were then compared to manual bedrock mapping based on satellite and ground-based imagery. Initial results from McElvoy Canyon show that the slope-based method with a slope threshold value of 42 degrees and a sampling radius of 5 meters provided the highest accuracy. These results are consistent with the setting for McElvoy Canyon, where mean hillslope angles range from 35-45 degrees and bedrock exposure appears to be due to soil instability. Our current work focuses on applying the same bedrock mapping methods to the more vegetated Sacatar and Olancha study areas to see how transferable the results from McElvoy Canyon are.