Assessment of Pollution Risk Caused by Nitrate in the Azarshahr Plain Aquifer Using Artificial Intelligence Models

Azarshahr plain is one of the wide plains in Lake Urmia basin which has an important source of groundwater and it is at risk of nitrate pollution due to the agricultural prosperity in this plain. Therefore, it is necessary to protect these groundwater resources using the most suitable and inexpensive solutions by identifying the areas exposed to the risk of pollution. In this study, in order to investigate the inherent vulnerability and risk of contamination, geological data, hydrogeological information and geophysical data were used as well as the digital elevation map of SRTM satellite with a spatial accuracy of 30 meters. Also the concentration of nitrate contaminant sampled from 35 different sources with appropriate distribution was used for validation. The pollution risk of Azarshahr plain aquifer was investigated using "Source-Pathway- Receptor" method. In this method, after identifying the source of pollution, the aquifer vulnerability was considered as a pathway. Improved DRASTIC method by SFL, M5P and RS models were used to evaluate the aquifer vulnerability, then the vulnerability map to nitrate contaminant was obtained. Based on the highest correlation (r = 0.87) and the lowest error (RMSE = 0.06), the M5P model was selected as the best model for vulnerability assessment. Finally, the risk map of nitrate pollution was obtained by multiplying the vulnerability of the aquifer (based on the M5P model) and the velocity of the groundwater. The results showed that the pollution risk of aquifer to nitrate pollutant is high in the central part of the aquifer.