Erosion susceptibility mapping of sub-watersheds for management prioritization using MCDM-based ensemble approach

Abstract

Soil erosion (SE) creates several environmental problems in fragile tropical watersheds, such as channel degradation, reservoir sedimentation, and flash flood. Therefore, identification of erosion susceptible areas at the sub-watershed level is necessary for the sustainable management of available resources. Previous studies on sub-watershed prioritization are based on a particular method, which leads to uncertainty because other methods may produce different results. Therefore, in the present work, a novel MCDM (multi-criteria decision-making)–based ensemble approach has been used for the prioritization of sub-watersheds of the Dwarkeswar River basin. Five MCDM models were used in the present study, namely simple additive weighting (SAW), complex proportional assessment (COPRAS), additive ratio assessment (ARAS), technique for order preference by similarity to ideal solution (TOPSIS), and multi-objective optimization on the basis of ratio analysis (MOORA). The non-parametric Spearman rank correlation method was used to determine the best model and measure the degree of similarity among the results obtained from different models. Rank correlation indicated that the COPRAS model has the highest accuracy in the prediction of erosion susceptibility. But the final ranks were given to the sub-watersheds by averaging the ranks obtained from different MCDM models. The TOPSIS model was not included for averaging because TOPSIS shows a negative correlation with the SAW (− 0.082) and ARAS (− 0.179) models and a very low positive correlation with the COPRAS (0.181) model. This combined ensemble method placed sub-watersheds 10 and 11 in first and second ranks, respectively, on the basis of susceptibility to SE.