Improving DRASTIC Model by Planning Gene Expression in Determining Aquifer Vulnerability to Nitrate(Case Study: Kermanshah Plain)

were prepared as seven raster layers, and after ranking and weighing, the obtained DRASTIC index ranged between 45 and 115. Yet, as far as the model's major problem is applying expert opinions in ranking and weighing the variables, the main purpose of this study is to improve the DRASTIC model by using the gene expression model, which as an intelligent model has shown a desirable performance. Also, in a mixed form, it can cope with other models to provide acceptable results. Thus, DRASTIC variables of a 20- year statistical period (1999-2009) were defined as the model input, and nitrate concentration was defined as its output. Data in GEP model were divided into two categories: training and experimentation. Moreover, using the statistical parameters (R2, RMSE, MAE and r), the simulation results of the gene expression model were evaluated. The results indicate the model's high ability in estimating nitrate concentration and its high capability in improving DRASTIC model. For validation and improvement of DRASTIC model, statistical parameters, R2 and r, were used, which were specified according to the error of the range model. Also, for each time combining the parameter with the GEP model, a score was gained during different stages and repeated performances of the weight ranking model using weighing rank model of each parameter. Finally, by removing two parameters, S and T, the modified formula of the DRASTIC index which was obtained based on weighing was 5D, 4R, 5A, 5I, and 4C.