Performance Six Intelligent Combined Methods in Groundwater quality modeling, Case study: Bafgh Plain

Assessment and controlling groundwater quality have an important role in planning and developing water resources. Therefore, the use of an efficient method can greatly increase accuracy and reduce costs in this field. In this study, 6 optimization algorithms Consist of Particle swarm optimization (PSO), Genetic algorithm (GA), Imperialist competitive algorithm (ICA), Fireflies algorithm(FA), Cultural algorithm(CA) and covariance matrix adaptation evolution strategy- Evolution strategies (CMA-ES) were used to train and optimize the parameters of the neural-fuzzy inference system model (ANFIS) to model the groundwater quality of Bafgh plain in Yazd province. At first, to select the best combination of input for an estimate of the electrical conductivity (EC), sodium adsorption (SAR) and total hardness (TH), Pearson and Spearman's methods were used to analyze the sensitivity and correlation of other parameters. then qualitative modeling is done with hybrid methods and the performance of the models was measured by correlation coefficients (R2), root mean square error (RMSE), and Nash-Sutcliffe Efficiency (NSE). The results showed that all six combined methods showed a very good performance in modeling groundwater parameters. Also, the ANFIS-FA model was one of the best models in all three modeling parts. So that the value of R2, RMSE and NSE for the test part in TH was, 0.99, 0.41, 0.99, for SAR, 0.98, 1.11,0.95 and for EC 0.99, 305.7 and 0.99. Other methods have also succeeded in modeling and predicting the desired parameters with proper precision. According to the accuracy of the calculations, these methods are suitable alternatives for the prediction of groundwater quality variables.