Monitoring and modeling sediment transport in space in small loess catchments using UAV-SfM photogrammetry

Abstract

Sediment transport rates are widely used in studies of soil erosion, hydrological simulation, and regional ecological assessment. Although traditional field hydrological stations can provide a global sediment transport rate for observation areas, quantifying sediment transport rates in space is still challenging. This study presents a work frame for monitoring and modeling sediment transport in space using recently ongoing UAV-SfM photogrammetry. First, the DEMs of difference (DoDs) were obtained by repeated UAV surveys. Then, we used the root mean square error (RMSE) and precision map (spatially distributed errors) to detect topographic changes at different confidence levels and analyzed the relations between detected changes and the thresholding of confidence level. Finally, we compared different routing solutions for inferring the sediment transport rates in space based on the observed topographic changes. Results show that the thresholding of confidence level will lose observations and may underestimate global sediment transport rates. The sensitivity of observation loss to the confidence level by using the precision map is lower than that by using RMSE, suggesting that the use of the precision map is more suitable for topographic change detection. In terms of the spatially distributed sediment transport rates, the multiple flow routing based on slope exponent (MFD-se) shows superior performance than the multiple flow routing based on maximum downslope gradient (MFD-md) and eight-node (D-8) routing. The negative sediment transport in the study areas ranges from 2.53% to 7.85% by using MFD-se routing. Moreover, the use of thresholding of confidence level is beneficial to reduce negative transport. The proposed work frame can conveniently provide a detailed spatial distribution of sediment transport rates, which would provide a new perspective for studies of earth surface processes.