Riding the wave: Enabling large-scale wave-resolving probabilistic coastal compound flood modeling

Extra-tropical storms and tropical cyclones significantly threaten coastal communities, affecting millions through loss of life, displacement, and damage from erosion, wind, and flooding. Historically, flood risks have been assessed by examining individual factors like storm surges and rainfall in isolation, limiting understanding of their compounding effects. Coastal flood managers need comprehensive information about these compound flood hazards to develop effective disaster risk reduction strategies, including flood protection and early warning systems. This thesis addresses three scientific challenges: 1) efficient modeling of compound flooding at large scales, 2) probabilistic boundary condition derivation in cyclone-prone areas, and 3) explicit inclusion of onshore wave processes. First, a new open-source reduced-complexity flood model called ‘Super-Fast INundation of CoastS’ (SFINCS) is introduced for the efficient modeling of compound flooding in coastal areas. The SFINCS model has been created a) to resolve the relevant marine, fluvial, and pluvial processes; and b) to perform at a much lower computational expense than traditional advanced process-based numerical models. It extends the local inertia equations to include advection and wind shear, allowing it to apply to coastal systems. Also, the option to force dynamic waves as boundary conditions is added. Results show, in various conceptual cases, that SFINCS can achieve similar accuracy relative to full-physics, advanced, process-based models. Second, we assess the contribution of non-marine flood drivers to coastal compound flooding by turning off flood drivers in SFINCS. We show there is a clear spatial dependence where non-marine flood drivers are relevant, but that none of the flood drivers can be neglected. Third, we assess coastal hazards from storm surge and offshore wave heights in the cyclone-prone Bay of Bengal using a probabilistic method. Our findings show that synthetic tracks generated by the open-source TCWiSE tool provide reliable estimates of cyclone properties, with significantly smaller 95% confidence intervals for extreme storm surge and wave heights compared to historical tracks. This approach enables a more robust sampling of the full parameter space for tropical cyclone tracks. Fourth, we present a new stationary wave energy solver that extends the SnapWave model to estimate the evolution of both incident and infragravity waves from offshore to nearshore. This solver incorporates an infragravity wave energy balance and resolves incoming infragravity wave energy. An infragravity wave source term is introduced and simplified using a parameterized shoaling parameter, allowing the solver to estimate infragravity wave transformation based on offshore wave heights and local bathymetry. The results demonstrate that this new solver can efficiently estimate nearshore infragravity wave conditions at large spatial scales. Finally, we demonstrate that an extended SFINCS model, after integrating the SnapWave model, allows to predict wave-driven flooding for Hurricane Florence (2018) at a 1,000 km scale. SnapWave efficiently computes wave forces that in SFINCS lead to spatially varying incident-wave-induced setup. It also estimates nearshore infragravity wave conditions, which at 5 m depth are forced as individual infragravity waves in SFINCS. This resolves wave runup and overtopping dynamically, with a finer resolution at the coast. We show that maximum offshore significant wave heights of up to 10 m during Hurricane Florence significantly impact total water levels through wave driven processes. The extended SFINCS model simulates 1,000 km of coastline 30 times faster than a traditional model covering only 7 km, and would be 3,700 times faster for the entire coastline. Caution is advised when using simpler methods for estimating wave setup, as they can lead to significant differences in predicted coastal flooding. With these results, it is shown that the newly developed methods in this thesis advance capabilities in modeling compound flood hazards for coastal systems at large spatial scales and including relevant wave processes.