On coastal flooding: from extreme offshore wave climate characterization to wave runup simulation with online-coupled numerical models

Coastal flooding is a topic of great relevance in the context of Coastal Engineering, particularly in the perspective of climate change and related sea level rise. Effective evaluation of coastal flooding at the hindcast, nowcast or forecast level, requires a high degree of interdisciplinary skills because of numerous aspects involved and different space and time scales usually considered. In particular, a peculiar perspective may be depicted by investigating the deep complementarity, as well as the close interrelation, coming up by the combined used of extreme value theory and advanced numerical modelling tools. As a matter of fact, numerical simulations are forced by offshore and/or nearshore boundary conditions, which, in the case of coastal flooding, are usually obtained by long-term statistical predictions of the sea state, related to the extreme regime of wave climate. The aim of this work is primarily focus on analysing and developing an operational methodology for the assessment of coastal flooding induced by extreme waves, with subsequent use of numerical modelling. In particular, the area of interest of wave climate analysis is related to the Sardinian coasts. It is evident how a reliable specification of boundary conditions is a fundamental component in numerical modelling. First of all, each dataset must formally fulfil correctness, representativeness, reliability and homogeneity characteristics. Unfortunately, even a rigorous assessment of the offshore boundary conditions is affected by an inherent uncertainty, due to the limited spatial and temporal extension which often characterizes a wave-climate dataset source. The sample size largely hinders the effectiveness of the asymptotical hypothesis, which is one of the cornerstones of the extreme value theory. On account of the above limitations, it is therefore still an open question in the literature whether upper limited (e.g. bounded Generalized Pareto) or upper unlimited (e.g. Weibull) distributions are most suitable for engineering purposes. Operationally, an investigation of the adequateness of the dataset at disposal is preliminary carried out. Next, several metrics are used to assess the effectiveness of most popular extreme wave distributions. Finally, the extreme, offshore sea states over the region of interest were computed. On the other hand, a plethora of numerical models have been implemented within the coastal community, with different levels of accuracy and complexity. A well-known and well-defined focused distinction is represented by the phase-averaged and phase-resolving approach. The main interest of this work is addressed to operational models, where outcomes’ accuracy criteria and non-prohibitive computational burden requirements should be well-balanced and properly considered. With this purpose in mind, an online coupling implementation between a spectral phase-averaged model and a non-hydrostatic phase-resolving model is provided. Specifically, two open-source models have been adopted as model components: the spectral SWAN model and the non-hydrostatic SWASH model. To sum up, the long term meteo-marine climate over the Sardinian coasts is evaluated first, by considering both mean and extreme characteristics of a heterogeneous dataset, previously deeply-corrected, ranging from waves to wind fields. From this, it is relatively straightforward to compute the offshore-boundary conditions. Second, a numerical coupled model involving two open source state-of-the-art models, was implemented, in order to enable a seamless and accurate chain simulation of an extreme sea state from offshore up to the shore.