Abstract
Sediment-Runoff process is highly variable and nonlinear in nature. In the present study an attempt has been made to develop a relationship between watershed stream length and Sediment Yield Index (SYI) and test it on Narmada watersheds, Madhya Pradesh, India. Area (A), Curve Number (CN) and stream length (SL) were utilized as input for model development. The three models (A model, CN model and simplified All India Soil and Land Use Survey (AISLUS) model including parameter SL) performed differently, with the coefficient of R-2 equal to 0.01, 0.02 and 0.71 (Shakkar watershed), 0.11, 0.23 and 0.91 (Bamhani watershed), 0.06, 0.001 and 0.80 (Manot watershed) and 0.40, 0.05 and 0.66 (Mohgaon watershed), respectively. The logarithmic simplified AISLUS model incorporating parameter SL resulted with the coefficient of R-2 as 0.76 (Shakkar watershed), 0.93 (Bamhani watershed), 0.84 (Manot watershed) and 0.66 (Mohgaon watershed). Therefore, the logarithm simplified AISLUS model was chosen as the best regression model for this study. It is observed that the simplified AISLUS model (logarithm form) incorporating parameter SL had a satisfactory efficiency as 76.35% (Shakkar watershed), 66.05% (Mohgaon watershed), 93.36% (Bamhani watershed), and 83.83% (Manot watershed) by Nash efficiency scale. The resulting higher Nash efficiency values support the versatility of the derived relationship and invoke assessment of SYI from the watershed stream length value. The prediction of SYI is important when adopting a suitable soil conservation measure in the watershed for minimizing soil erosion.