Leveraging the TabPFN algorithm for high-resolution mapping of groundwater bicarbonate and Its scaling risk across China

This dataset provides predicted groundwater bicarbonate (HCO₃⁻) concentrations across China at high spatial resolution. The data were generated using the Tabular Prior-data Fitted Network (TabPFN), which showed superior predictive performance under sparse sampling conditions compared with Random Forest, XGBoost, and Support Vector Machines. The dataset supports national-scale groundwater quality assessment, analysis of hydrogeochemical spatial patterns, and evaluation of scaling risks in underground pipelines and water conveyance systems. It offers a valuable resource for groundwater management, infrastructure protection, and sustainable water-use policy development.