Daily suspended sediment yield estimation using soft-computing algorithms for hilly watersheds in a data-scarce situation: a case study of Bino watershed, Uttarakhand

Water erosion creates adverse impacts on agricultural production, infrastructure, and water quality across the world, especially in hilly areas. Regional-scale water erosion assessment is essential, but existing models could have been more efficient in predicting the suspended sediment load. Further, data scarcity is a common problem in predicting sediment load. Thus, the current study aimed at modeling the suspended sediment yield of a hilly watershed (i.e., Bino watershed, Uttarakhand-India) using machine learning (ML) algorithms for a data-scarce situation. For this purpose, the ML models, viz., adaptive neuro-fuzzy inference system (ANFIS) and fuzzy logic (FL) were developed using data from ten years (2000-2009) only. Further, runoff and suspended sediment concentration (SSC) were obtained as the primary influencing factors. Varying combinations of lagged SSC and runoff data were considered as model inputs. The ANFIS and FL models were compared with the conventional multiple linear regression (MLR) model. Results indicated that the ANFIS model performed better than the FL and MLR models. Thus, it was concluded that the ANFIS model could be used as a benchmark for sediment yield prediction in hilly terrain in data-scarce situations. The research work would help field investigators in selecting the proper tool for estimating suspended sediment yield/load and policymakers to make appropriate decisions to reduce the devastating impact of soil erosion in hilly terrains.