Characterization of compound coastal flooding in the Keta basin

Low-lying coastal areas such as the Ghanaian coasts are increasingly exposed to multiple, interacting climate hazards that threaten ecosystems, livelihoods, and infrastructures. Compound flooding, where the combination or successive occurrence of two or more flood drivers leads to a greater impact, can exacerbate the adverse consequences of flooding, particularly in coastal–estuarine regions. This thesis delineates a reproducible framework for compound flooding characterisation in data-scarce environments, bridging the climate-informed data analysis to the planning and evaluation of Nature-based Solutions (NbS) for coastal risk reduction. To address this overarching goal, the research assesses and validates multi-source climate datasets, notably the GloFAS hydrological reanalysis and the JRC offshore water level dataset, and then characterises the statistical behaviour of extreme river discharge and water level through univariate and bivariate Extreme Value Analyses (EVA), combining percentile-based thresholding, Peaks-Over-Threshold fitting, and copula-based modelling to identify and quantify compound coastal–fluvial flood events. The results demonstrate that river discharge and coastal water level extremes in the Volta Delta region, where Keta basin is, exhibit strong but asynchronous seasonality, limited long-term trends, and weak tail dependence, indicating that compound events are infrequent and primarily driven by temporal coincidence rather than strong physical coupling. Lastly, a third phase of the framework is conceptually outlined as future work, proposing the generation of synthetic extreme events through the UNprecedented Simulated Extremes using ENsembles (UNSEEN) approach to enrich the historical characterisation of compound flooding, by expanding the observational records with physically plausible, yet historically unobserved, extreme events. This would enable the subsequent evaluation of NbS performance under plausible, non-observed scenarios.