Spatio-Temporal Optimization of Long-term Groundwater Monitoring Networks Using PSO Algorithm

Spatial and temporal variations of contamination in groundwater resources, necessitate long-term monitoring (LTM) at a given site. In this study, several groundwater quality parameters (EC, SAR, TH, TDS, pH, K, Na+, Ca2+, Mg2+, SO42-, HCO3-, and Cl-) for 113 samples sites clustered based on the particle swarm optimization (PSO) algorithm to significantly decrease cost and save time in LTM. The optimization of the clustering process was carried out according to the Silhouette index. For verification and validation of the results, Geology, soil order, land use, hydrological network, and TDS maps were used. According to the results, the best number of clusters was 5. An acceptable agreement was obtained between land conditions and clusters represented by the PSO algorithm. Consequently, it can be inferred that the clustering of the groundwater quality using the PSO algorithm and the Silhouette index optimizer could 70% decrease the number of spatio-temporal sampling in LTM.