Interpretable machine learning for groundwater nitrate prediction using hydrochemical and isotopic indicators in China

Groundwater nitrate contamination poses significant environmental and public health concerns, necessitating reliable predictive approaches for effective water quality management. In this study, an explainable machine learning framework was developed to predict groundwater nitrate (NO3−) concentrations using hydrochemical, isotopic, and spatial variables compiled from groundwater studies conducted across diverse hydrogeological regions of China. Key variables, including pH, electrical conductivity (EC), total dissolved solids (TDS), dissolved oxygen (DO), temperature, dissolved organic carbon (DOC), and stable isotopes (δ15N-NO3 and δ18O-NO3), were integrated into Random Forest (RF) and Extreme Gradient Boosting (XGBoost) models. Statistical analysis, correlation assessment, and feature engineering were performed prior to model development. Hyperparameter optimization and 10-fold cross-validation were employed to improve model robustness and generalization. The optimized RF and XGBoost models exhibited moderate predictive performance, achieving R2 values of 0.489 and 0.477, respectively, reflecting the inherent complexity and spatial heterogeneity of groundwater nitrate contamination. Model predictions were interpreted using SHapley Additive exPlanations (SHAP), one and two dimensional Partial Dependence Plots (PDPs), feature importance, and sensitivity analysis, revealing that isotopic signatures, EC, TDS, and DO were the dominant factors influencing groundwater nitrate variability. By integrating hydrochemical and isotopic indicators with multiple complementary interpretability techniques, the framework provides quantitative insights into the individual and nonlinear contributions of key predictors to nitrate variability. The proposed explainable machine learning framework provides a transparent and scientifically interpretable approach for groundwater nitrate prediction and offers valuable insights into the hydrogeochemical factors governing nitrate contamination, thereby supporting groundwater quality assessment and sustainable water resource management.

Authors

Institutions

Publication Details

Journal
Journal of the Iranian Chemical Society
Published
2026-11-17
DOI
https://doi.org/10.1007/s13738-026-03650-4
Primary Topic
Water Quality and Pollution Assessment
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Interpretable machine learning for groundwater nitrate prediction using hydrochemical and isotopic indicators in China

Zhenyu Liu, Shaokang Wang, Danfeng Li
7 links
Journal of the Iranian Chemical Society
Water Quality and Pollution Assessment
article

Interpretable machine learning for groundwater nitrate prediction using hydrochemical and isotopic indicators in China

Zhenyu Liu, Shaokang Wang, Danfeng Li
article en

Abstract

Groundwater nitrate contamination poses significant environmental and public health concerns, necessitating reliable predictive approaches for effective water quality management. In this study, an explainable machine learning framework was developed to predict groundwater nitrate (NO3−) concentrations using hydrochemical, isotopic, and spatial variables compiled from groundwater studies conducted across diverse hydrogeological regions of China. Key variables, including pH, electrical conductivity (EC), total dissolved solids (TDS), dissolved oxygen (DO), temperature, dissolved organic carbon (DOC), and stable isotopes (δ15N-NO3 and δ18O-NO3), were integrated into Random Forest (RF) and Extreme Gradient Boosting (XGBoost) models. Statistical analysis, correlation assessment, and feature engineering were performed prior to model development. Hyperparameter optimization and 10-fold cross-validation were employed to improve model robustness and generalization. The optimized RF and XGBoost models exhibited moderate predictive performance, achieving R2 values of 0.489 and 0.477, respectively, reflecting the inherent complexity and spatial heterogeneity of groundwater nitrate contamination. Model predictions were interpreted using SHapley Additive exPlanations (SHAP), one and two dimensional Partial Dependence Plots (PDPs), feature importance, and sensitivity analysis, revealing that isotopic signatures, EC, TDS, and DO were the dominant factors influencing groundwater nitrate variability. By integrating hydrochemical and isotopic indicators with multiple complementary interpretability techniques, the framework provides quantitative insights into the individual and nonlinear contributions of key predictors to nitrate variability. The proposed explainable machine learning framework provides a transparent and scientifically interpretable approach for groundwater nitrate prediction and offers valuable insights into the hydrogeochemical factors governing nitrate contamination, thereby supporting groundwater quality assessment and sustainable water resource management.

Journal of the Iranian Chemical SocietyVol. 23(12)
Tianjin Tianhe Hospital (CN), Tianjin Medical University (CN)
Clean water and sanitation
Water Quality and Pollution Assessment
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon