Simultaneous comparison of modified-integrated water quality and entropy weighted indices: Implication for safe drinking water in the coastal region of Bangladesh

Abstract

• A new unbiased drinking water quality index has been developed. • Using the index, groundwater of coastal region of Bangladesh was classified. • A comparative assessment was done between newly developed index and entropy index. • Most of the coastal groundwater exhibited as unsuitable for drinking purposes. • The index sensitivity is very high as it considered below- and upper threshold levels. Drinking-water sources are subjected to vulnerability due to chemical contamination and health risk in the southern coastal region of Bangladesh. Different types of water quality indices are being used to express the water quality, though most of them have both advantages and limitations. In this study, we modified and proposed an integrated water quality index by principal component analysis (PCA), and compared with the entropy theory. Spatial autocorrelation index and fuzzy GIS analysis are applied to delineate the suitability of groundwater in the study area based on randomly selected 377 groundwater samples from various depths (15–300 m). This study determined 20 water quality parameters. To avoid the bias of parameter selection, PCA was performed to reduce 12 parameters loads (TDS, EC, Ca 2+ , Mg 2+ , Na + , K + , Cl − , HCO 3 − , NO 3 − , pH, PO₄ 2− , and F − ). Unlike traditional indices, both desired and tolerable limits were considered in the modified-integrated water quality index (MIWQI). The results of entropy analysis depict NO 3 − , Cl − , and Na + as the most contributing parameters influencing the groundwater quality and much emphasis should be paid on these parameters to stop further groundwater pollution. MIWQI and Entropy water quality index (EWQI) results classify 21.22% and 42.43% of water samples as good quality for drinking, respectively. The rest 79% and 58% are classified as medium to extremely poor-quality water in the study area. A significant spatial pattern was identified for both water quality indices which were positively correlated (Moran’s I > 0). The MIWQI classification system exhibited commendable outcomes compared to EWQI due to comprehensiveness. Fuzzy GIS map revealed the spatial heterogeneity of groundwater parameters showing weak spatial dependency.