Understanding how groundwater and surface water interact in snow-dominated watersheds is essential for managing water supplies and ecosystem health. These exchanges—whether snowmelt flows quickly through near-surface pathways into streams or infiltrates deeply into aquifers—can vary from year to year. Such variability arises not only from changes in precipitation and temperature but also from how climate signals are filtered by a watershed’s topography and subsurface geology. In our study, we focused on two catchments with contrasting slopes and geological structures to see how differences in late‑summer and fall rainfall alter groundwater–surface water exchanges during the snowmelt season. We found that in steep landscapes with strong lateral contrasts in subsurface permeability, dry pre-snow conditions favor rapid near-surface flow into streams, whereas wet pre‑snow conditions encourage deeper infiltration and aquifer recharge. To test how general these findings are, we ran virtual experiments that systematically varied slope and subsurface heterogeneity. Across these simulations, catchments with high lateral variability consistently showed the largest shifts in GW–SW exchange, while more uniform systems remained relatively stable. Our results underscore the need to account for subsurface structure when predicting how climate variability will reshape water flow and storage in mountain regions.