Vol. 5, Issue 5 (2017)
Modelling river suspended sediment load using artificial neural network and multiple linear regression: Vamsadhara River Basin, India
Author(s): Shreya Nivesh and Pravendra Kumar
Abstract: Accurate estimation of suspended sediment load carried by rivers is of utmost importance in the soil and water conservation practices in the watershed and also in large number of hydro-environmental issues such as planning, design and operations of reservoirs, dams and environmental impact assessment. This study explores the abilities of statistical models to improve the accuracy of rainfall-streamflow-suspended sediment relationships in daily suspended sediment estimation. In this study, a comparison was made between multiple linear regression and artificial neural networks (ANNs) for the Vamsadhara river catchment. Daily rainfall-runoff and suspended sediment data were used as inputs and outputs. The performance results based on three different types of indicators viz. root mean square error (RMSE), correlation coefficient (r) and coefficient of efficiency (CE) revealed that ANN (RMSE-110.15 kg/sec, r-0.97 and CE value 94.22 % ) can predict sediment load more efficiently than traditional models like multiple linear regression.
Pages: 337-344 | 1165 Views 102 Downloads
download (7207KB)
How to cite this article:
Shreya Nivesh, Pravendra Kumar. Modelling river suspended sediment load using artificial neural network and multiple linear regression: Vamsadhara River Basin, India. Int J Chem Stud 2017;5(5):337-344.
