Coastal flood-depth maps at high spatial resolution are required for risk assessment and adaptation planning under rising sea levels and intensifying storms. Solving the two-dimensional flow equations at 30 m is too costly to run globally. We present a global coastal flood-depth model that propagates water inland from the coastline over a global 30 m coastal terrain model. The model takes cells in order of water-surface elevation and lowers that elevation at each step by a Manning friction loss, with the flow speed set by the local water depth. Neighbouring tiles exchange water across shared edges, and a tile is skipped once no new water reaches it. Five global iterations complete in 4.6 h on 70 parallel processes and give flood-depth maps at 18 return periods. The coastal water level sums the storm surge at each return period and the highest astronomical tide from a global tide atlas, checked against ten National Oceanic and Atmospheric Administration tide gauges (root-mean-square difference 0.69 m, mean bias +0.47 m). Against nine government flood polygons in Atlantic Canada, the US Atlantic and Gulf coasts, and Great Britain, each compared at its matched model return period, the model attains a mean Critical Success Index of 0.60 (median 0.56) over the coastal floodplain, on a scale where 1 is an exact match of modelled and mapped extent. Mean under- and over-prediction are 27% and 14% of the compared area; over-prediction concentrates in shallow upper estuaries and under-prediction where friction holds water back from broad, low-lying floodplains. Intersecting the flood extents with a global building-footprint dataset yields 7.7 million exposed footprints at the 100-year level; six large coastal-basin regions hold 77%. The model gives the extent water reaches once it stops spreading, suited to regional screening rather than building-level loss estimation; hydrograph timing, wave runup, overtopping momentum, and breaching lie outside scope.
No takes yet. Share an insight, caveat, or question.
Chaudhuri et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: