Xiaoli Su (University of Bristol, United Kingdom); Laurence Hawker (University of Bristol, United Kingdom)

Southern Africa has recently been impacted by flooding from a series of devastating tropical cyclones, including Idai in 2019 and Freddy in 2023. Estimating the risk of flooding from these extreme weather events has been hampered by a lack of observation data in the region, especially with respect to river flows and water levels. We present research from the REPRESA project, funded by the UK and Canadian governments under their climate adaptation and resilience (CLARE) program, which aims to improve the simulation of present-day and future tropical cyclone flooding in Southern Africa.
Specifically, we present a flood inundation model driven by tropical cyclone simulations from kilometre-scale convection-permitting climate models, which has been calibrated to SWOT water height data.
To set up the model, weather reanalysis from ERA5 and a hydrological model were used to simulate river discharge in the region. Given these discharges, river bathymetry was estimated from all available SWOT River SP node heights, providing a means to calibrate a large-scale LISFLOOD-FP flood inundation model. To facilitate the simulation of complex river systems with multi-threaded rivers, we link SWOT River SP node data to the recently released Global River Topology (GRIT) river network and Fathom DEM. Compound flooding along the coastline was simulated by linking to an ocean tide and surge model (ADCIRC). The resulting model was validated against flood extents observed by Sentinel 1 from Tropical Cyclone Idai.