The effects of applying different DEM resolutions, DEM sources and flow tracing algorithms on LS factor and sediment yield estimation using USLE in Barajin river basin (BRB), Iran

Abstract

Accurate estimation of the spatial pattern of annual soil erosion and sediment yield is of paramount importance in watershed conservation measures, design of sedimentation dams, reduction of reservoir storage capacity, integrated watershed management, flood problems, and alteration of ecosystems’ locations. One of the most widely used empirical/ conceptual methods for estimating soil erosion and the sediment yield rate (SYR) is the universal soil loss equation (USLE) model. Topographic factor (LS) is one of the main inputs of RUSLE and can be calculated using digital elevation models (DEMs). This research studied the impacts of using different DEM resolutions (cell size ranges from 30 to 200 m), DEM sources (ALOS-30 m, ASTER-30 m, and SRTM-90 m), and various LS formulas on the USLE and estimated SYR at the Barajin river basin (BRB) outlet. Also, assessing the effect of distinct flow tracing algorithms (FTAs) on LS factor and RUSLE output is another purpose of this study. Findings show that using different LS equations without considering their limitations leads to a significant absolute relative error (ARE) in the estimation of SYR (ranging from 1.2 to 400%). Furthermore, Liu et al. (1994) and Moore and Burch (1986) equations, due to an ARE of lower than 10% in estimating SYR, are the best ones for high slope regions like BRB. Evaluating the effects of scale issues indicates that DEM resolution is more critical than DEM sources (or data resolution). Based on findings, the ARE of SYR estimation caused by DEM resolution varies between 1.5% and 58%, while for different DEM sources, it changes from 15.1% to 48.5%. Besides, ASTER-30 m and SRTM-90 m with the mean ARE value of 34.9 and 15.3%, respectively, are the worst and best DEM sources in BRB. Finally, investigating the impact of using different FTAs indicates that ignoring this issue can lead to significant errors ranging between 1.6 and 42.2%. Among all FTAs the MFD and Aspects are the best ones for applying in the USLE model. To summarize, in data-sparse regions or developing countries, the findings of this study can be considered as appropriate guidance.