Debris flows analysis through quantitative evaluation of soil depth distribution under limited data

Abstract

Soil depth is essential in studying natural disasters such as landslides and debris flow hazards. Despite the importance of soil depth in the mechanism of erosion-entrainment during the debris flow process, research on soil-depth data for analyzing debris flows is limited. Therefore, this study focused on the Gallam-ri area with a watershed of 0.9 km² to evaluate the soil depth mapping under limited data and significance of these maps for debris flow simulations. Based on the knocking pole test data, two-dimensional distribution soil depth maps were constructed using the S and Z models, the Kriging method, and a method that applies some values uniformly as the soil depth (U model). The accuracy of soil depth mapping methods was quantitatively evaluated using R² and root mean squared error analysis. Since soil depth demonstrated independent patterns with land-surface data, soil depth maps using S and Z models have structural limitations showing R² of 0.0003 and 0.002, respectively. The debris flows were analyzed through numerical model Deb2D, and the soil depth most significantly influenced the erosion volume and the damaged area. Analyses using S and U models showed 94 % and 98 % high similarity to the simulation results through the Kriging method, respectively. However, considering overall analyses, the S model was analyzed to be the most stable in constructing soil depth maps and simulating debris flows for ungauged basins.