Seasonal groundwater–salinity dynamics and climate-driven saltwater intrusion in Coastal South Florida

Miami-Dade County, Florida. Coastal aquifers in South Florida face growing exposure to saltwater intrusion (SWI) driven by sea-level rise, climate variability, and groundwater withdrawal. This study presents a high-resolution, seasonally focused analysis of groundwater–salinity interactions under two climate scenarios (SSP2–4.5 and SSP5–8.5). Using statistically downscaled CMIP6 projections, a MODFLOW-based groundwater model, and explainable AI methods, we evaluate how precipitation patterns, groundwater levels, and SWI evolve over time and vary across three aquifer layers. Results show nonlinear and delayed seasonal feedbacks across aquifer depths. In the shallow layer, SWI increased by 10.6 % during fall under SSP5–8.5, compared with a 7.1 % increase during summer under SSP2–4.5, highlighting scenario-dependent seasonal shifts in intrusion peaks. Although wet-season precipitation anomalies exceeded 350 %, intense rainfall reduced infiltration efficiency and limited recharge, contributing to greater SWI penetration into deeper zones. Granger causality analysis indicated more than a 250 % increase in feedback strength between consecutive seasons under SSP5–8.5, reflecting stronger temporal coupling. The Random Forest model reached 77 % accuracy and 94 % recall for high-salinity conditions, and SHAP analysis identified hydraulic gradient, coastal distance, and structural proximity as the most influential predictors. Spatial mapping showed that the deepest layer experienced the widest inland salinization, occurring alongside groundwater drawdown. A structural change around 2020 marked accelerated SWI advancement, with annual expansion rates in shallow aquifers exceeding 50 %. Overall, this integrated framework improves understanding of seasonal SWI patterns and provides information to support adaptive groundwater management in climate-stressed coastal regions. Integrated framework showing climate downscaling, groundwater modeling, and machine learning analysis for seasonal SWI dynamics • Seasonal climate shifts intensify coastal saltwater intrusion risk. • CMIP6 data + MODFLOW + XAI reveal nonlinear groundwater responses. • High emissions trigger earlier and stronger aquifer salinization. • Recharge efficiency drops despite higher rainfall intensity. • Machine learning identifies hydraulic gradient as key SWI driver.