Full-Waveform Inversion–Informed Hydrologic Modeling Reveals Bedrock Heterogeneity Controls Flow Dynamics

Accurate characterization of the critical zone architecture is fundamental to physically based hydrologic modeling, yet resolving the complex geometry of subsurface boundaries remains a significant challenge. Critical zone seismic studies have predominantly used First-Arrival Traveltime (FAT) tomography, yet this method lacks the resolution to characterize heterogeneity at hydrologically relevant scales, leaving a gap in our understanding of how subsurface structure governs flow routing and storage. Full-Waveform Inversion (FWI) overcomes these limitations by utilizing the complete seismic wavefield to resolve fine-scale subsurface architecture. We compare 2D hydrologic modeling informed by subsurface structures from FAT and FWI. The FAT model yields smooth, layered velocity gradients, whereas FWI reveals pronounced heterogeneity, including depth-to-bedrock variations and steep low-velocity anomalies. We integrated both structures into ParFlow-CLM with consistent hydrologic properties and NLDAS meteorological forcing them to isolate the effects of subsurface geometry. Results show that while annual water budgets remain similar, reflecting comparable mean regolith and fractured bedrock depths, internal flow dynamics diverge markedly. The rugged bedrock topography resolved by FWI imposes geometric control on flow routing: infiltrating water fills deep bedrock troughs before lateral flow initiates, producing a "fill-and-spill" mechanism. These deep troughs act as subsurface reservoirs, temporarily storing water and extending drainage timescales. Consequently, the FWI-informed model buffers hydrologic response, generating a smoother hydrograph with attenuated peaks and sustained baseflow, whereas the FAT model exhibits rapid lateral drainage and flashier storm response. These findings demonstrate that smoothing subsurface heterogeneity in hydrologic models may mask critical storage dynamics and bias estimates of catchment response times.