Comparative Analysis between Morphometry and Geo-Environmental Factor Based Soil Erosion Risk Assessment Using Weight of Evidence Model: a Study on Jainti River Basin, Eastern India

Abstract

Assessment of spatial soil erosion risk is a viable effort signifying the needs of conservation measures due to the deterioration of land as well as soil quality degradation at various scales. Among several non-quantitative approaches regarding erosion risk prediction, watershed morphometry and other geo-environmental parameter based assessments were performed largely and separately which showed varied results. In the present work, using 15 morphometric and 13 geo-environmental parameters, spatial soil erosion risk was modelled in order to inspect the performances and consistency of both approaches in predicting Spatial Soil Erosion Risk (SSER). Field site erosion patch inventory (a total of 164 erosion patches), google earth imagery and a probabilistic model, i.e., Weight of Evidence (WoE) enabled the analysis. Training patches (115 patches) were used to model the SSER while validation patches (49 patches) were used to assess the consistency of model output. Both approaches quantify 25.41% and 20.18% of the area to high to very high susceptibility class, separately. The contribution of each factor of both parameter groups in risk predicting was analysed through Map Removal Sensitivity Analysis (MRSA). Further, the results of performance were evaluated through Repetitive Operator Choice (ROC) curve (success rate and prediction rate curves) measuring Area Under Curve (AUC). The success and prediction rate curves show that when considering morphometric parameters, the AUC is 0.775 and 0.729, respectively, whereas in the case of geo-environmental parameters, AUC = 0.892 and 0.878 accordingly. This reveals the better consistency of geo-environmental parameters in context with the spatial erosion risk zoning in the present scenario.