Soil Erosion And Non-Point Source Pollution Regulations And Mechanisms In A Semi-Artificial Environment Based On Geo-Detector And Sediment Delivery Models

Soil erosion and Non-point(NSP)source pollution pose a major threat to water quality and contribute to land degradation worldwide.These also pose a threat to food security and quality water availability.Assessment of NSP and soil erosion processing is crucial especially for complex drainage basins so as to develop effective soil and water conservancy measures.Dongsheng basin and Bojiang watershed were used as study units due to the presence of soil erosion and NSP in these areas.Therefore,the purpose of this study was to investigate the space and time trend of soil erosion and NSP,identify key drivers of erosion and NSP,and identify suitable rainwater harvesting sites and structures as a way of abating NSP and soil erosion.We introduced a new approach involving the integration of the SDR of the InVEST and Geodetector models to explore the Spatio-temporal variability of erosion and investigate the key drivers contributing to soil erosion.The second innovative point of this study lies in integrating the SWAT model,random forest regression model,redundancy analysis,and correlation coefficient explain the contribution of landscape configuration on NSP and runoff.Lastly,we used the weighted linear combination and Boolean techniques including a depression depth layer to identify rainwater harvesting sites and suitable structures.Rainwater harvesting is widely known to reduce runoff and curb the rate of erosion and non-point source pollution.The results showed apparent LULC change in Dongsheng basin over the 30 years(1990-2020),and increasing trend in soil loss and sediment export,indicating a correlation between LULC and soil erosion,and land degradation.Cropland had the highest increase in average soil loss,from 26.53 t ha-lyr-1 in 1990 to 118.76 t ha-lyr-1 in 2020.This is attributed to a rise in cropland areas across the watershed.The primary causative factors to soil loss are vegetation,slope,precipitation,and aspect.The study revealed that TP,TN,and sediment annual trends are consistent with that of runoff and precipitation of the study area.NSP is highly distributed in the season around June to September.Over 90%of NSP occurs in cropland landcover,whereas forestland cover produces the least amount of pollution.The findings infer that the aggregation and connectivity of forest patches contribute to the decline in NSP load and vice versa for cropland cover.The largest patch index 24(LPI)shows a significant negative correlation with NSP,with an R2 of-0.58 for TP and TN and-0.62 for sediment load.The findings indicate that landscapes with larger patch sizes,a high number of patches,and aggregation of patches largely influence pollution distribution.Overall,the results suggest that the influence of landscape patterns on NSP outweighs that of runoff.NSP is temporally and spatially varying per watershed region and landscape patterns play a major role in the distribution.The results showed that combining the Boolean and the weighted linear combination techniques,and using a depression depth layer could effectively identify rainwater harvesting sites as well as appropriate structures in Bojiang drainage basin.The approaches and methodologies can be adopted by watershed managers to develop effective strategies regarding soil and water conservancy at watershed level.