Influence of DEM Elaboration Methods on the USLE Model Topographical Factor Parameter on Steep Slopes

Abstract

Runoff erosion is an important theme in hydrological investigations. Models assessing soil erosion are based on various algorithms that determine the relief coefficient using rasterized digital elevation models (DEMs). For evaluation of soil loss, the most-used model worldwide is the USLE (Universal Soil Loss Equation), where the most essential part is the LS parameter, which is, in turn, generated from two parameters: L (slope length coefficient) and S (slope inclination). The most significant limitation of LS is the difficulty in obtaining the data needed to generate detailed DEMs. We investigated three popular data generation methods: aerial photographs (AP), aerial laser scanning (ALS), and terrestrial laser scanning (TLS) by assessing the quality and effect of DEMs generated from each method over an area of 40 m × 200 m in Silesia, Poland. Additionally, the relationship between particular LSUSLE  parameter components was carried out based on its final distribution. Our results show that resolution strongly influences DEMs and the LSUSLE  parameters. We found a strong relationship between the degree of height data resolution and the accuracy level of the calculated parameters. Based on our investigations we confirmed the highest influence on the LSUSLE  came from the S parameter. Additionally, we concluded that in examinations over large areas, terrestrial laser scanners are not ideal; the benefits of their additional accuracy are outweighed by the additional time and labor consumption; in addition, terrestrial-based scans are sometimes not possible due to ground obstacles the limited scope of most lasers. Aerial photographs or point clouds generated by aerial laser scanners are sufficient for most purposes connected with surface flow, and further developments can be based on the use of these techniques for obtaining ground information for modeling erosion processes.