Optimization of DRASTIC method using ANN to evaluating of vulnerability of multiple Varzqan aquifer

Due to population growth and agricultural development and mining activities in the plain Varzeqan where nitrate concentration exceeds from 5 times the standard World Health Organization (WHO). So, Evaluation of vulnerability and protection of groundwater resources are very important in this area. The DRASTIC method uses seven effective environmental parameters on assessment of aquifer vulnerability such as Depth to groundwater level, net Recharge, Aquifer media, Soil media, Topography, Impact of vadose zone and hydraulic Conductivity, as seven layers were prepared separately for unconfined and confined aquifer by corresponded the rate and weighting. The DRASTIC index value was evaluated for unconfined, confined aquifer 92-163 and 48-93 respectively. The artificial neural network model was used to optimize the DRASTIC method. In these model the DRASTIC parameters were considered as input, and conditioned DRASTIC index were used as output, and the data were divided into two categories of train and test. After model training, the model results were evaluated by the nitrate concentration through coefficient of determination (R2) and correlation index (CI) creteria. The results showed that artificial neural network model show high capability to improve the results of general DRASTIC and reduce subjectivity of model, especially in multiple aquifer.