As intertidal ecosystems experience novel disturbance regimes due to climatic extremes, anticipating population trajectories becomes increasingly challenging. Size-dependence of disturbance effects further complicates recovery by altering population structures, yet size-dependent impacts of extreme storms on intertidal species, like slow-moving bivalves, remain poorly understood. To help unravel these impacts, we used flume experiments to quantify the relationship between erosion susceptibility and bivalve size, and applied the results to a simple age-structured model to scale up from individual to population impacts. In the flume, we examined whether bivalve erosion susceptibility depends on shell length, initial burying depth, and active (re-)burial behavior for juveniles and adults of two co-occurring tidal flat species in the Scheldt estuary: Cerastoderma edule and Macoma balthica. Despite species differences in living depths and shell shapes, shell length emerged as the key predictor of erosion susceptibility. Applying the length‑erosion probability curve to an age-structured population model revealed that the total impact of a single extreme storm can be much greater when storm mortality is size-dependent vs. size-independent. Our results suggest that increasing storminess may substantially weaken or even locally eradicate bivalve populations in erosion‑prone tidal flats, although landscape impacts may be buffered by high larval dispersal. More broadly, our study shows how empirically derived size–mortality relationships can be integrated into simple, data‑constrained models to better understand how extreme events compromise population resilience.