Modeling unsaturated hydraulic conductivity by hybrid soft computing techniques

Accurate prediction of the unsaturated hydraulic conductivity (K) is necessary to check the feasibility of the artificial and natural groundwater recharge. In this study, one artificial intelligence (AI), i.e., adaptive neuro-fuzzy inference system (ANFIS) technique, and two hybrid techniques (combination of traditional AI + optimization technique), i.e., ANFIS + firefly algorithms (ANFIS-FFA) and ANFIS + particle swarm optimization (ANFIS-PSO), are used to predict the K of the soil. The study area for this investigation is Ghaggar basin. For the present study, dataset (240 observations) was collected from field experiments using minidisk infiltrometer. Total dataset was segregated into two different parts. Larger part (170 data) was used for model development, and smaller part (70 data) was used to check the performance of developed models. Four popular statistical parameters were used to evaluate the performance of developed models. Results indicate that the performance of ANFIS-PSO and ANFIS-FFA was comparable with higher accuracy in prediction of K of the soil than traditional ANFIS model.