Document Type : Research Article
Professor, Department of Range and Watershed Management, College of Agriculture and Natural Resources of Darab, Shiraz University, Iran
PhD Student of Electrical Engineering, Shiraz University of Technology, Shiraz, Iran
Professor, Department of Range and Watershed Management, College of Agriculture, University of Fasa, Iran
Professor, Electrical and Electronics Engineering Department, Shiraz University of Technology, Shiraz, Iran
Due to the reduced rainfall and overuse of underground water, checking the water quality is one of the most important challenges discussed in various areas, such as Iran. Estimation of water quality using models such as neural network results in costs reduction and better management. The current study aims is to assess ground water quality using adaptive fuzzy neural network (ANFIS) in the west of Fars province during 2003 to 2013 period. Three methods including grid partitioning, sub-clustering and FCM with two models of Hybrid and back propagation were used to predict the quality of ground water for the study area. In this study, electrical conductivity (EC) and sodium adsorption ratio (SAR) were used to train the neural network. In addition, water quality class diagram Wilcox was used to train the network. In chemical pollution, according to Wilcox diagram, EC and SAR are the most important factors based on which waters can be classified in different classes such as very appropriate, suitable and unsuitable for agriculture. Results show that among various models provided to predict groundwater quality, Hybrid models in FCM method have the greatest accuracy for the prediction of water quality in the study area with a maximum R (0.99) and minimum error.