Monitoring ex-situ water parameters, namely heavy metals, needs time and laboratory work for water sampling and analytical processes, which can retard the response to ongoing pollution events. Previous studies have successfully applied fast modeling techniques such as artificial intelligence algorithms to predict heavy metals. However, neither low-cost feature predictability nor explainability assessments have been considered in the modeling process. This study proposes a reliable and explainable framework to find an effective model and feature set to predict heavy metals in groundwater. The integrated assessment framework has four steps: model selection uncertainty, feature selection uncertainty, predictive uncertainty, and model interpretability. The results show that Random Forest is the most suitable model, and quick-measure parameters can be used as predictors for arsenic (As), iron (Fe), and manganese (Mn). Although the model performance is auspicious, it likely produces significant uncertainties. The findings also demonstrate that arsenic is related to nutrients and spatial distribution, while Fe and Mn are affected by spatial distribution and salinity. Some limitations and suggestions are also discussed to improve the prediction accuracy and interpretability.

The Choushui River groundwater basin (CRGB) in Yunlin County, Taiwan, is a significant groundwater source for the western part of the region. However, increasing groundwater demand and human activities have triggered a potential crisis due to overexploitation. Therefore, groundwater potential zone (GWPZ) maps are crucial for mapping groundwater resources and water resource management. This study employs the normalized index–overlay method and fuzzy extended analytical hierarchy process (FE-AHP) to map GWPZs cost-effectively. The methodology objectively incorporates weightings from various thematic layers by normalizing and correlating parameters with observed groundwater availability (GA). Site-specific observations, including aquifer thickness, depth to the groundwater level, and porosity, inform GA calculations. Seven comprehensive layers derived from remote sensing (RS) data are processed to obtain weightings and ratings for the groundwater potential index (GWPI) in the CRGB. Selected parameters are categorized into hydrological processes, human interventions, geological, and surface profiles. Hydrological processes include precipitation, modified normalized difference water index (MNDWI), and drainage density. Human interventions consist of the enhanced vegetation index (EVI) and normalized difference building index (NDBI). Surface profiles encompass the terrain ruggedness index (TRI) and slope, enhancing the study’s multi-criteria approach. The observed GA validates the GWPZ accuracy, classifying zones into five categories. According to the GWPI of FE-AHP, about 59.56% of the CRGB area can be categorized as “moderate” to “very good” potential groundwater recharge zones. Pearson’s correlation coefficient between GWPI and GA, based on FE-AHP, outperforms the conventional AHP. This RS-based approach efficiently evaluates GA in aquifers with limited wells, highlighting crucial zones in CRGB’s proximal-fan and southeastern mid-fan for informed groundwater management strategies.

The increasing serious droughts recently might have significant impacts on socioeconomic development in the Red River basin (RRB). This study applied the variable infiltration capacity (VIC) model to investigate spatio-temporal dynamics of soil moisture in the northeast, northwest, and Red River Delta (RRD) regions of the RRB part belongs to territory of Vietnam. The soil moisture dataset simulated for 10 years (2005–2014) was utilized to establish the soil moisture anomaly percentage index (SMAPI) for assessing intensity of agricultural drought. Soil moisture appeared to co-vary with precipitation, air temperature, evapotranspiration, and various features of land cover, topography, and soil type in three regions of the RRB. SMAPI analysis revealed that more areas in the northeast experienced severe droughts compared to those in other regions, especially in the dry season and transitional months. Meanwhile, the northwest mainly suffered from mild drought and a slightly wet condition during the dry season. Different from that, the RRD mainly had moderately to very wet conditions throughout the year. The areas of both agricultural and forested lands associated with severe drought in the dry season were larger than those in the wet season. Generally, VIC-based soil moisture approach offered a feasible solution for improving soil moisture and agricultural drought monitoring capabilities at the regional scale.