Investigating an empirical approach to predict sediment yield for a design storm: a multi-site multi-variable study

Ishan Sharma,Surendra Kumar Mishra,Ashish Pandey,Henok Mekonnen Aragaw,Vijay P. Singh
DOI: https://doi.org/10.1007/s10668-024-04832-x
IF: 4.08
2024-04-11
Environment Development and Sustainability
Abstract:It is of common experience that the sizeable portion of sediment yield generated over a period in a catchment occurs largely due to only a few extreme storm events rather than the numerous small rain events, suggesting the vital role of rain duration besides its magnitude. This paper presents a novel empirical approach based on the Soil Conservation Service Curve Number equation and Universal Soil Loss Equation coupled with a sediment yield model to predict sediment yield resulting from a storm of a certain duration (t) and return period (T). To this end, an empirical relation relating potential erosion (A) with 'T' and 't' is proposed and calibrated/validated using historical rainfall-runoff-sediment series data from ten Indian and 12 United States (US) watersheds, respectively. It is calibrated with a low nRMSE, high NSE, and very less PBIAS values in all the sub-watersheds (mean nRMSE ≤ 0.285, NSE ≥ 0.84 and PBIAS < ± 10%). In validation, the proposed approach shows an excellent performance in 9 of the 10 sub-watersheds of India (0.77 ≤ NSE ≤ 0.98, 0.85 ≤ R 2 ≤ 0.99 and PBIAS ≤ ± 20%) a good performance in 8 out of 12 sub-watersheds of the USA (0.50 ≤ NSE ≤ 0.97, R 2 > 0.91 and PBIAS ≤ ± 30%). Finally, a correlation between the calibrated empirical parameters of the proposed relation and the catchment characteristics was ascertained, which showed that the size and stream length influence these parameters more in large sub-watersheds, while slope and stream length influence more in small sub-watersheds.
environmental sciences,green & sustainable science & technology
What problem does this paper attempt to address?