Suitability of revision to MUSLE for estimating sediment yield in the Loess Plateau of China

This study evaluated the effect of a combination of revisions to the cover and management factor ( C ), support practice factor ( P ), and topographic factor ( LS ), and to determine the best use of each combination for modified universal soil loss equation (MUSLE) being applied in the Loess Plateau of China. Eight combinations were built based on common local methods of revision to factors C , P , and LS in three ways (single factor revised, two factors revised, and three factors revised, respectively). By validating the effect of these combinations on sediment yield estimation using the Nash–Sutcliffe coefficient of efficiency (NSE), normalized root-mean-square error (NRMSE), correlation coefficient (R), and angle between liner trend line and 1:1 line, the optimal combination was selected. Results show that: (1) the original method significantly overestimates the sediment yield; (2) when revising a single factor, the LS factor revision has the greatest positive effect; (3) when revising two factors, the combination of a simultaneous revision to C and to P has the smallest angle between the liner trend line and the 1:1 line (1.37°); if the estimating accuracy of sediment peak is required, this combination may thus be a useful approach; and (4) when revising C , P , and LS simultaneously, the combination has the highest NSE (0.76), lowest NRMSE (1.19), and the highest correlation coefficient (0.760), and may be used to estimate sediment yield. The suitability of MUSLE is enhanced using the common local revision to factors C , P , and LS . Under the assumption that level terrace spatial information can obtain effectively (or that there are few level terraces), these findings can be applied to the Loess Plateau for the design of soil and water conservation measures and for flood disaster assessment, and applied to future model applications in the Loess Plateau.