Application of Artificial Intellegint Models for Suspended Sediment Load (SSL) Modeling in Watersheds With Management Operations (Case study: Ghaleh Gol Watershed, Lorestan province)

In recent years, extensive practices have been done on flood control, erosion and sediment in the fields of research and implementation of watershed management. The present study was carried out for the modeling of Suspended Sediment Load (SSL) by artificial intelligent models in two subwatersheds in Ghaleh Gol watershed in Lorestan province, Iran. In this research, the flow velocity was measured and the SSL was sampled directly from the beginning of the rainfall events until the end of them. In this study, four soft computing techniques, GP-PUK, GP-RBF, MLP and RF were used to predict the SSL in study area. Total data set consists of temperature, rain, discharge and suspended sediment load that 70 percent of the entire dataset was used in a training stage of the soft computing techniques and 30 percent was used for testing the models. Finally, the models’ accuracy was assessed using three performance evaluation parameters, which were Correlation Coefficient (C.C.), Root Mean Square Error (RMSE), Maximum Absolute Error (MAE) and Nash-Sutcliffe model efficiency (NSE). The obtained results suggest that the GP-RBF model (with C.C= 0.9509, RMSE= 0.067, MAE= 0.041 and NSE=0.924 in North watershed and with C.C= 0.966, RMSE= 0.048, MAE= 0.037 and NSE=0.932 in South watershed) is more accurate to estimate the SSL as compare to the GP-PUK, RF and MLP for the given study area. Thus, GP-RBF was found to be the most suitable model for modelling Suspended Sediment Load (SSL) in the study area. Also, sensitivity analysis concludes that the discharge is the most effective parameter for the estimation of Suspended Sediment Load (SSL).