Project: Flood inundation modelling in data sparse deltas
Supervisors: Professor Paul Bates and Dr Jeff Neal
Mega deltas are one the most susceptible to flood hazards due to their flat, low lying topography coupled with an immediate proximity to the coast, making them prone to both riverine and coastal flooding. Additional contributing factors stem from relative sea level rise, subsidence and reduced sediment delivery. Currently, an estimated 7% of global population live in these regions, with an increasing number in rapidly expanding mega cities. Future losses from flooding in deltaic cities are estimated to increase markedly, as a result of an expected increase in frequency and magnitude of hurricanes and cyclones, population increase and subsidence. Flood risk in these regions is growing rapidly in less developed countries as gross domestic product-enabled infrastructure and defences cannot be implemented to the same levels as wealthier countries. Yet, the rapid development of mega deltas presents an opportunity for stakeholders to plan and implement policies to reduce the flood risk.
At the highest level, this project aims to improve our capability to model flood inundation in data-sparse deltas. In practice, this has involved building a 2D regional scale hydrodynamic model of the data-sparse Mekong Delta with LISFLOOD-FP using freely available data and assessing what aspects of the flood model structure and data are most important for performance. At this scale, this is the first 2D model of the delta. Model results highlighted the importance of topography, yet the availability of topographic information for data-sparse areas remains limited. Consequently, we analysed spatial error structure of vertical heights for the SRTM DEM (Shuttle Radar Topography Mission) and the SRTM error reduced MERIT DEM (Yamazaki et al., 2017) by comparing to near ground truth LIDAR data for various floodplain locations across the world. This spatial error structure has in turn been taken to simulate candidates of the true DEM, creating a catalogue of multiple statistically plausible DEMs. Using these multiple DEMs in a flood inundation study in the Mekong Delta has demonstrated the variability of inundation extent, suggesting flood models should use multiple DEMs to explore the impact of topographic uncertainty on inundation extent. This workflow can be utilised to generate multiple DEMs quickly and efficiently for floodplains across the world, even in data-sparse locations. Ongoing work is focussing on exploring the spatial error structure further and developing this as an open source tool.
Correspondence: Laurence Hawker, School of Geographical Sciences, University of Bristol, Bristol, BS8 1SS, UK. <firstname.lastname@example.org>
Keywords: Mega Deltas; Hydrodynamic modelling; LISFLOOD FP; Flood; Floodplain Inundation; Topographic Uncertainty; Sea level Rise; Simulating DEMs