Predicting Wave-Induced Flooding on Low-Lying Tropical Islands Using a Bayesian Network

Abstract

Low-lying tropical islands are highly vulnerable to the effects of sea-level rise and climate change. Most pressing is the threat posed to their fresh water supplies by wave-induced flooding. This thesis attempts to generalize previous site-specific studies of flooding on coral atolls and apply it in a framework that can be used for early warning systems or long-term climate change impact studies. To do so, a large synthetic database of representative reef properties and hydrodynamics was developed using the numerical wave model XBeach, and then analyzed using a Bayesian probabilistic network. The resulting tool allows us to make real-time flood predictions based on offshore wave or sea level conditions, and the unique characteristics of a given island (e.g. topography). Narrow, smooth reefs with steep fore reef slopes are the most vulnerable to runup. Extreme flooding is associated with anomalously high, resonant low frequency waves, which are more likely to occur on narrow, smooth reefs subjected to extreme swell waves (large wave heights with long periods). These extreme swells are "blue sky" events which originate from distant storms. Thus, they may thus arrive without warning, since they are completely independent of local weather conditions. This bolsters the need for more effective prediction tools. Validation of the results presented in this study is limited by the small number of field observations against which the model can be compared. Thus, there is a need to develop a comprehensive database of reef morphology and hydrodynamics. Offshore wave conditions, water levels, and reef width are the most essential variables for predicting runup, so future research efforts should be directed towards collecting those data. This model can also be used to ask questions like, "which islands will be most severely impacted by climate change?", or "can we increase flood resilience for a given island by restoring its coral reefs?". Hence, there is also potential for its use in guiding decision-makers to allocate limited funding in the places where it will have the most impact.