Will mine closures trigger low-flow crises in river systems? Insights from a machine learning approach for the Upper Vistula Basin, Poland

Study region Upper Silesian Coal Basin and Upper Vistula River system, southern Poland. Study focus This study assesses how the planned phase-out of mine drainage in the Upper Silesian Coal Basin (2020–2049) may interact with climate change to alter streamflow in the Upper Vistula River system, focusing on the Przemsza–Vistula corridor. Future streamflows were simulated using SWAT within the CHASE-PL framework, driven by nine bias-corrected EURO-CORDEX projections under RCP4.5 and RCP8.5. A machine-learning Conditional Weighted Regression approach dynamically combined the ensemble outputs into a Weighted Ensemble Streamflow projection. Mine-closure impacts were quantified by removing mine-water discharges according to the closure schedule. New hydrological insights Mine closures primarily affect low-flow conditions. In the Przemsza River system, the probability of flows below the historical 10th percentile increases markedly after closure. Relative to historical conditions, the combined effects of climate change and mine closure may reduce monthly 10th-percentile flows by up to 4.5 m³ /s (~35%) in the Przemsza River and 1.4 m³ /s (~15%) in the Vistula River. Projected flows increase in spring and summer but decrease in autumn, reflecting interacting climatic and anthropogenic pressures. The hybrid SWAT–machine-learning framework improves the robustness of climate-impact assessments and supports evaluation of the hydrological consequences of socio-economic transitions.