Use of machine learning and geographical information system to predict nitrate concentration in an unconfined aquifer in Iran

Increased nitrate concentration is one of the main groundwater quality problems today that needs to be measured and monitored. Water testing monitoring are time consuming costly. Therefore, new modeling methods such as machine learning algorithms can used an efficient solution for predicting concentration. In this study, three including deep neural network (DNN), extreme gradient boosting (EGB), multiple linear regression (MLR) were predict contamination in north Iran (Mazandaran plain) finally best method was selected mapping. The mean 250 piezometric wells considered output variable factors affecting (groundwater depth, transmissivity aquifers, precipitation, evaporation, distance from water resources Caspian Sea, industries residential centers, population density, topography, exploitation groundwater) input variables alluvial aquifer. same training data process methods. results stages showed EGB has highest performance due lowest error values correlation between predicted (training R-sqr = 0.98, Nash–Sutcliffe efficiency (NSE) test 0.86, NSE 0.84). Further, indicate industries, evaporation rates most important groundwater. Finally, tested model a geographic information system (GIS) tool prepare map pollution study area. Evaluating resulting by comparing indicated good accuracy (R-sqr 0.8).