Accurate evaluation of groundwater quality and identification of key characteristics are essential for maintaining groundwater resources. The purpose of this study is to strengthen water quality evaluation through the SHAP and XGBoost algorithms, analyze the key indicators affecting water quality in depth, and quantify their impact on groundwater quality through interpretable tools. The XGBoost algorithm shows that zinc (0.183), nitrate (0.159), and chloride (0.136) are the three indicators with the highest weight. The SHAP algorithm shows that zinc (34.62%), nitrate (17.65%), and chloride (16.98%) have higher contribution values, which explains the output results of XGBoost. According to the calculation scores and classification standards of the water quality model, 49% of the groundwater samples in the study area have excellent water quality, 33% of the samples are better, and 18% of the samples are polluted. The results of positive matrix factorization (PMF) show that natural conditions, metal processing, metal smelting and mining, and agricultural activities all cause pollution to groundwater. Zinc, chloride, nitrate, and manganese were the key variables determined by the SHAP algorithm to explain the vast majority of human health risk sources. These findings indicate that interpretable machine learning not only improves the correlation of water quality assessment but also quantifies the judgment basis of each sample and helps to track key pollution indicators.

The North China Plain faces severe water scarcity, and the efficient use of brackish water has become a crucial pathway for sustaining agricultural development. In this study, we combine scenario analysis with Data Envelopment Analysis to establish a multi-scenario efficiency evaluation framework. Focusing on six counties in Handan, Hebei Province, we employ an input-oriented Slack-Based Measure Data Envelopment Analysis (SBM-DEA) model to systematically evaluate brackish water irrigation efficiency (BWIE) across a baseline year (2020) and eight projected scenarios for 2030. The results show that the mean efficiency values across scenarios range from 0.646 to 0.909. Scenarios combining universal adoption of water-saving irrigation with normal hydrological conditions achieve the highest mean efficiency (>0.9), with minimal regional disparities and optimal system stability. The promotion of water-saving irrigation technologies is the primary driver of improved BWIE, whereas simply increasing brackish water application yields only limited marginal benefits. Redundancy analysis further indicates that water resource inputs are the main source of efficiency loss, with brackish water redundancy (42.3%) far exceeding that of land inputs (10.5%). These findings provide quantitative evidence and methodological support for optimizing regional water allocation and advancing sustainable agricultural development.

Groundwater nitrate (NO₃⁻) pollution is a critical environmental challenge with direct implications for human health. In this work, we propose a comprehensive analytical framework that integrates multi-model intercomparison, interpretable machine learning techniques, and quantitative health risk evaluation to tackle the pressing groundwater nitrate governance dilemmas in Handan City, a representative urban area in North China. Based on 157 groundwater samples and 17 hydrochemical parameters, comparative analysis of three state-of-the-art machine learning algorithms showed that the Light Gradient Boosting Machine (LightGBM) algorithm outperformed all counterparts, delivering the optimal predictive performance (R² = 0.753, RMSE = 3.67). SHapley Additive exPlanations (SHAP) analysis identified F⁻, Ca²⁺, Cl⁻, K⁺, total hardness, and Mg²⁺ as dominant factors influencing groundwater NO₃⁻ concentrations, reflecting the combined effects of carbonate dissolution, nitrification, and anthropogenic inputs. Subsequently, we performed a health risk assessment based on the standard methodological framework issued by the United States Environmental Protection Agency (USEPA), and the results indicated that children were the most vulnerable group, with hazard quotient (HQ, a non-carcinogenic risk indicator) values reaching 1.07 in the western mountainous region, exceeding the safety threshold (HQ > 1). These findings clarify the pollution mechanisms and spatial heterogeneity, and provide targeted policy guidance for groundwater protection as well as the safeguarding of public health.