SOM-DRASTIC: using self-organizing map for evaluating groundwater potential to pollution

Farshad Rezaei, PhD Candidate, Department of Civil Engineering, Isfahan University of Technology, Isfahan, Iran.

Mohammad R. Ahmadzadeh, Associate Prof., Department of Electrical and Computer Engineering, Isfahan University of Technology, Isfahan, Iran.

Hamid R. Safavi, Associate Prof., Department of Civil Engineering, Isfahan University of Technology, Isfahan, Iran.

Abstract

Groundwater is considered as the most important water resource, especially in arid and semi-arid regions, so it is crucial to impede this source of water to be contaminated. One of the most common methods to assess groundwater vulnerability is DRASTIC method. However, the subjectivity existing in defining DRASTIC weights and ratings as well as inadaptability of the parameters involved in this method with special geology, hydrogeology, land use and climatic conditions have urged researchers to modify this method. In this paper, a new method combining a special type of the neural networks called Self-Organizing Map (SOM) and the traditional DRASTIC model resulting in the hybrid SOM-DRASTIC model is applied to modify and improve DRASTIC Model. The traditional DRASTIC method holds a summation among all negative effects of different factors contributing to vulnerability, while the proposed hybrid method is able of classifying the groundwater vulnerability and deriving the real relation existing between the DRASTIC parameters as the inputs and the vulnerability class as the output of the method. The vulnerability assessment process was performed on the Zayandeh-Rud river basin aquifers in Iran. The SOM-DRASTIC identified the northern parts of the study area as the most vulnerable areas with a drastically fractured structure, while the traditional DRASTIC ranked the western parts as the most vulnerable regions with a high rate of net recharge. The results demonstrate that the proposed method can be used by managers and decision-makers as an alternative robust tool for vulnerability-based classification and land use planning.

doi.org/10.1007/s00477-016-1334-3