Modelling of Infiltration Rate in Different Soil Textures using Soft Computing Techniques in Kashkan Watershed, Lorestan Province

Infiltration is one of the most parameters of hydrology that plays a fundamental role in streamflow, groundwater recharge, subsurface flow, and surface and subsurface water quality and quantity. According to the importance of the mentioned subject, the infiltration changes and modeling were investigated in the different soil textures in Kashkan watershed, Lorestan province. In this study, the double-ring infiltrometer was used to measure the infiltration in the different soil textures. Also, Support vector regression (SVM), Gaussian Process (GP), Multi-Layer Perceptron (MLP), and Random Forest (RF) were used to Modeling of infiltration rate in different soil textures. Three statistical comparison criteria including root mean square error (RMSE), coefficient of correlation (C.C), and Nash Sutcliffe (NSE) were used to determine the best-performing infiltration model. The results showed that Random Forest model better estimated infiltration rate (C.C= 0.9912, NSE= 0.98622, and RMSE = 0.0177) compared to the other models. Thus, RF was found to be the most suitable for modeling infiltration in the study area. Also, sensitivity analysis concludes that the parameter time is the most effective parameter for the estimation of infiltration rate. The results indicated that the infiltration varies in the different soil textures, which should be considered in the management of groundwater recharge.