The role of (in)dependence in flood risk assessments

Abstract

Flooding remains the most frequent natural hazard, affecting more people than any other weather-related disasters. The increase in frequency, duration and severity of floods in many parts of the world is projected to rise due to changes in climate, population and land use. Compound flooding, which occurs when multiple flood drivers interact to cause or exacerbate flooding, is a significant challenge in estuarine environments. However, the influence of flood drivers on compound flooding in estuaries remains poorly understood. To address this issue, the goal of this thesis is to improve flood risk assessments in estuarine environments by advancing our understanding and modelling of compound flood hazard. The thesis first identifies potential hotspots of compound flooding at the global scale by analyzing discharge and storm surge time series from state-of-the-art global models. The study is performed at 3,434 river mouths, encompassing all catchments along the coast with an upstream area larger than 1,000 km2. About 6% of the locations are classified as having the highest compound flooding potential. The results show that the dependence between river discharge and storm surge is significant in many areas, including the coasts of Madagascar, northern Morocco, northern Australia, Vietnam, and Taiwan. Furthermore, the study highlights the complex temporal dynamics between discharge and storm surge that cannot be captured by one indicator alone. Next, the thesis develops different probabilistic assessment methods to account for the dependence of flood drivers in flood risk assessment through two case studies. The first case study assesses the impact of the spatial dependence between river discharges and storm surge on the flood hazard quantification for a small catchment in Galveston Bay, Texas. The second case study expands the probabilistic assessment method into a computationally efficient probabilistic framework for calculating flood risk using Ho Chi Minh City as a case study. The framework allows for multiple dependence structures and is applied at a monthly time scale to account for seasonality. Lastly, the thesis moves beyond traditional expected annual damage (EAD) as the indicator of flood risk and discusses the many implications of spatial and temporal dynamics on disaster risk management. The development of plausible storylines and their impact on risk management decisions are explored within a serious game called Breaking the Silos. Breaking the Silos is a collaborative role-playing game designed to explore the impacts of these complexities on risk management decisions. Overall, the thesis provides important insights into the challenges of compound flooding in estuarine environments and proposes new methods to account for these challenges in flood risk assessments. The study highlights the need to include compound flooding in flood hazard assessments in locations where fluvial, pluvial, and coastal interactions are important. The thesis also emphasizes the need for disaster management measures to consider the spatial and temporal dynamics between hazards and their impact on residual risk.