To What Extent Can a Sediment Yield Model Be Trusted? A Case Study from the Passaúna Catchment, Brazil

Soil degradation and reservoir siltation are two of the major actual environmental, scientific, and engineering challenges. With the actual trend of world population increase, further pressure is expected on both water and soil systems around the world. Soil degradation and reservoir siltation are, however, strongly interlinked with the erosion processes that take place in the hydrological catchments, as both are consequences of these processes. Due to the spatial scale and duration of erosion events, the installation and operation of monitoring systems are rather cost- and time-consuming. Modeling is a feasible alternative for assessing the soil loss adequately. In this study, the possibility of adopting reservoir sediment stock as a validation measure for a monthly time-step sediment input model was investigated. For the assessment of sediment stock in the reservoir, the commercial free-fall penetrometer GraviProbe (GP) was used, while the calculation of sediment yield was calculated by combining a revised universal soil loss equation (RUSLE)-based model with a sediment delivery ratio model based on the connectivity approach. For the RUSLE factors, a combination of remote sensing, literature review, and conventional sampling was used. For calculation of the C Factor, satellite imagery from the Sentinel-2 platform was used. The C Factor was derived from an empirical approach by combining the normalized difference vegetation index (NDVI), the degree of soil sealing, and land-use/land-cover data. The key research objective of this study was to examine to what extent a reservoir can be used to validate a long-term erosion model, and to find out the limiting factors in this regard. Another focus was to assess the potential improvements in erosion modeling from the use of Sentinel-2 data. The use of such data showed good potential to improve the overall spatial and temporal performance of the model and also dictated further opportunities for using such types of model as reliable decision support systems for sustainable catchment management and reservoir protection measures.