Comparison between LogisticRegression, Modified DRASTIC and AHP-DRASTIC Methods in Groundwater Vulnerability Assessment

Natural aquifer vulnerability can be defined as reaching pollution to groundwater and the possibility to propogate in it after land contamination. This is a relative specification, with no dimension and cannot be measured; which it is not only depended on aquifer specification but also on geological and hydrological characteristics of the aquifer of the region. Varieties of methods have been developed for evaluating groundwater vulnerability and because of the ease of implementation index method and particular DRASTIC method are among of the widely used methods. Each parameter in the DRASTIC method that potentially affects the likelihood of contamination is categorized on the same scale and after applying coefficients of the parameters, vulnerability assessment is provided with a score. One important point in this method is that ranking and weighting of the parameters is highly depended on the personal decision and judgment which can reduce the quality of the results. Many suggestions for improving DRASTIC model have been developed by researchers. Most of the researchers suggested excluding less important parameters, adding affective parameters, modifying the model coefficients, and ranking parameters.In this research, in order to overcome the above mentioned drawbacks and selecting the appropriate model for assessing the vulnerability of aquifers; three

logistic regression, modified DRASTIC and AHP-DRASTIC are examined and compared, and after collecting input parameters, vulnerability is assessed based on the models. In order to select the most appropriate model and evaluating performance of the models, Spearman correlation coefficient for nitrate concentration and vulnerability class is used. The results indicate high performance of AHP-DRASTIC in comparison to the combined method.