Topographical Thresholds for Ephemeral Gully Identification in Watersheds with Highly Disturbed Terrain

Abstract. Topographic index models have been incorporated in planning tools for installation of structural best management practices to reduce soil erosion in watersheds. The performance of topographic indices may vary depending on geomorphology, climate, topography, and physiographic location of the watershed. In this study we assessed the optimal thresholds for a compound topographic index (CTI) model at prediction of ephemeral gullies (EG) in two paired watersheds (Running Turkey and Dry Turkey) in central Kansas with different levels of terrain disturbances. The Running Turkey watershed had 179 meters of grassed waterways and 525 meters of terraces, while Dry Turkey watershed had 379 meters of grassed waterways and 1900 meters of terraces. The EGs predicted by the model were compared with the observed EGs obtained using field reconnaissance, aerial imagery, and DEM processing. A range of optimum thresholds for the CTI model were obtained based on EG length, EG location, and catchment drainage density. The results indicated that watersheds experiencing higher disturbances in topography, such as Dry Turkey watershed, would have higher optimum topographical thresholds as compared to watersheds with lower disturbance, such as Running Turkey watershed. The optimum threshold ranges obtained through classification of head water (HW) and main stem (MS) EG catchments by using drainage density analysis within Running Turkey watershed were similar to those obtained using EG length and location analyses. However, within Dry Turkey watershed, there was a disagreement in the ranges of optimum thresholds for HW and MS catchments. The disagreement of threshold values for HW and MS catchments was attributed to the presence of high terrain disturbance in Dry Turkey watershed, mainly terraces and grassed waterways. This paired watershed study indicated that, in addition to other factors, the difference in optimum critical thresholds of the CTI model can be attributed to topographical alterations within the watershed.