Assessment of groundwater quality in a highly urbanized coastal city using water quality index model and bayesian model averaging

Prediction and assessment of water quality are important aspects resource management. To date, several index (WQI) models have been developed improved for effective However, the application these is limited because their inherent uncertainty. improve reliability WQI model quantify its uncertainty, we a WQI-Bayesian averaging (BMA) based on BMA method to merge different comprehensive groundwater assessment. This comprised two stages: i) stage, four traditional were used calculate values, ii) stage integrating results from multiple determine final status. In this study, machine learning method, namely, extreme gradient boosting algorithm was also adopted systematically assign weights sub-index functions aggregation function. It can avoid time consumption computational effort required find most parameters. The showed that status in study area mainly maintained fair good categories. values ranged 35.01 98.45 prediction area. Temporally, category exhibited seasonal fluctuations 2015 2020, with highest percentage lowest marginal category. Spatially, sites fell under fair-to-good category, few scattered areas falling indicating has well maintained. WQI-BMA relatively easy implement interpret, which significant implications regional