Application of cubic spline in soil erosion modeling from Narmada Watersheds,India

Soil erosion by water is ubiquitous, exhibits spatio-temporal variability, and is fundamental to determining sediment yield which is key to proper watershed management. In this study, we propose a relationship between the curve number and sediment yield index (SYI) using cubic splines. Using field data from four watersheds, the relation between observed and computed SYI is found to have a coefficient of determination (R²) value from 0.63 to 0.88 suggesting that such a relation can be used to determine SYI from the available CN value. It is found that cubic splines perform satisfactorily with Nash-Sutcliff efficiency ranging from 60.18 to 64.01%, absolute prediction error from 1.35 to 5.56%, integral square error from 1.21 to 5.82%, coefficient of correlation from 79.32 to 93.78%, and degree of agreement from 0.87 to 0.99%.