MULTIVARIATE DATA ANALYSIS TO ASSESS GROUNDWATER HYDROCHEMICAL CHARACTERIZATION IN RAWADANAU BASIN, BANTEN INDONESIA

Abstract

A multivariate statistical technique of principal component analysis (PCA) and hierarchical cluster analysis (HCA) has been applied to identify and classify the various water sources that comprise the Rawadanau Basin. The data collection includes 60 samples taken during the dry (29 samples) and the rainy season (31 samples) in tropical regions. Sources of sampled water include dug wells, rivers, cold springs, and hot springs. Water chemistry measurable variables include field data (T, pH, EC), major ions (Na+, K+, Ca2+, Mg2+, Cl-, HCO3 -, SO4 2-), SiO2 , Fetotal, Mn, and stable isotopes of water (δ2H, and δ18O). The correlation of the concentration of water chemistry shows changes in the rainy season to Fetotal and Mn. Interpretation based on HCA using the dendrogram based on the chemical elements of water produces two clusters. Cluster A reflects an unconfined aquifer and bicarbonate type. Meanwhile, cluster B is a chloride type from the confined aquifer and does not change in different seasons. The PCA results show that the three-component matrix accounts for 86.12% of the data structure describing the Rawadanau Basin water sources that volcanic rocks affect and strongly correlate with Na+, K+, Ca2+, and Mg2+. PC1 has a high positive value for hydrochemical composition, indicating that lithology influences the kind of water. PC2 has a positive value for the stable isotope (δ18O and δ2H), meaning it is the main water source in Rawadanau. PC3 has a positive value for elevation and a negative for longitude, indicating a recharge area influenced by geological factors and is correlated with geothermal influences and volcanic rocks. This multivariate analysis can identify components and clusters of hydrochemical variables that have not been determined in previous studies.