Asymmetric Interaction Between Terrain and Rainfall Data Resolution in Flood Inundation Mapping of Large Basins

Flood inundation maps derived from coupled hydrologic-hydraulic models are highly sensitive to input data resolution. However, the relative importance of rainfall temporal and terrain spatial resolutions—and their interaction—remains insufficiently quantified in large-basin flood inundation mapping. This study systematically evaluates these effects using a factorial experiment combining rainfall data at 1, 3, 6, and 12 h and digital elevation models (DEMs) at 25, 50, 100, and 250 m resolutions for the June 2024 extreme flood in the Hanjiang Basin, China (~30,000 km 2 ). With 1 h rainfall, coarsening DEM spatial resolution from 25 to 50, 100, and 250 m increases inundation area by 41, 82, and 277%, with expansion concentrated in deep-water regions. With the 25 m DEM, coarsening rainfall temporal resolution from 1 h to 3, 6, and 12 h reduces inundation area by approximately 4, 6, and 23%, also mainly in deep-water zones. Rainfall temporal resolution exhibits a saturation effect, with refinements finer than 6 h yielding only marginal improvements (~2%) in inundation extent. These results reveal an asymmetric influence of input data resolution on flood inundation mapping: terrain coarsening drives substantial expansion of inundation extent, whereas rainfall coarsening produces smaller reductions. Across the practical DEM range of 25–100 m, the impact of terrain data changes dominates over rainfall, and refining rainfall temporal resolution cannot compensate for geometric bias introduced by terrain data coarsening. Overall, accurate terrain representation is more critical than high-frequency rainfall inputs for reliable flood hazard mapping in large mountainous basins. The results provide insights for operational early warning applications.