Evaluation of aquifer vulnerability using PCA technique and various clustering methods

One of the significant challenges in applying vulnerability methods is their dependence on the same weights and ranks for each parameter. Therefore, to modify these methods, a clustering technique independent of ranks and weights and based only on the nature of data was used in the first step in this study for vulnerability mapping. Moreover, Principal Component Analysis (PCA) was employed to consider all parameters and select the effective ones. After reviewing the literature on different vulnerability assessment methods, 14 parameters involved in determining the vulnerability of the Varamin Aquifer in Iran were selected. The principal component analysis revealed that seven of them were the major ones. In the next step, the vulnerability map of the aquifer was prepared by using the C-Means, the K-Medoids, and the K-Means clustering methods. It was found that the fuzzy C-Means clustering method outperformed others with a CI of 260.

Automated summary

The study initially utilized a clustering technique independent of ranks and weights, followed by employing Principal Component Analysis to select effective parameters, and ultimately prepared a vulnerability map of the Varamin Aquifer using the C-Means clustering method.