Understanding Compound Flooding hazard in Estuaries: Insights and Implications

Estuaries are highly prone to compound flooding. These areas often face flooding prompted by fluvial discharge, coastal water levels, wind and pluvial (rainfall) conditions (Moftakhari et al. 2017). Flooding drivers, even if they are not extreme individually, can combine and generate extreme local impacts. Nevertheless, their dependence and co-occurrence are often ignored, leading to misinterpretation of flooding risk. In this regard, assessing multivariate extremes requires understanding their stochastic structure and interconnections. Sensitivity relies on modeling properties like tail dependence strength and symmetry (Hua and Joe 2011). Copulas enable the study of tail dependency, providing insights into the relative strength between the extremes (De Luca et al, 2023). Once these primary dependencies and interconnected relationships are appropriately captured and modelled, the next step involves translating them into potential impacts (Zscheischler 2020). Therefore, defining hazard scenarios establishes the connection between the dependence structure of multiple drivers and the associated impacts. The critical level or return period used in risk analysis and infrastructure design inherently represents a hazard scenario. It can be seen as upper sets encompassing all occurrences deemed hazardous, potentially leading to impacts and damages based on certain criteria. This definition implies a connection with the upper tails of variables, which depicts specific dangerous conditions. In contrast to univariate analysis, where critical events are defined by surpassing a specific threshold, the multivariate hazard scenario lacks a singular definition (Bernardi et al. 2018). Moreover, in an n-dimensional framework, this set collects all 'dangerous' values based on suitable criteria and consequently defines the (n-1) iso-hyper-surface that generates the 'dangerous region', known as the 'critical layer' (Salvadori et al, 2011). In higher dimensions, this critical layer possesses more of a mathematical than a graphical definition, entailing theoretical and computational challenges.This study aims to robustly characterize compound flooding in estuaries, employing high-dimensional analysis alongside multivariate statistical techniques and computational optimizations. Using a 100-year return level, critical events that compose the iso-hypersurface (critical layer) are identified. These design events capture variability, enabling the incorporation of uncertainty involved in predicting these dynamics.ReferencesBernardi, M., Durante, F., Jaworski, P., Petrella, L., & Salvadori, G. (2018). Conditional risk based on multivariate hazard scenarios. Stochastic Environmental Research and Risk Assessment, 32, 203-211.De Luca, G., Ruscone, M. N., & Amati, V. (2023). The use of conditional copula for studying the influence of economic sectors. Expert Systems with Applications, 120582.Hua, L., & Joe, H. (2011). Tail order and intermediate tail dependence of multivariate copulas. Journal of Multivariate Analysis, 102(10), 1454-1471.Moftakhari, H. R., Salvadori, G., AghaKouchak, A., Sanders, B. F., & Matthew, R. A. (2017). Compounding effects of sea level rise and fluvial flooding. Proceedings of the National Academy of Sciences, 114(37), 9785-9790.Salvadori, G., De Michele, C., & Durante, F. (2011). On the return period and design in a multivariate framework. Hydrology and Earth System Sciences, 15(11), 3293-3305.Zscheischler, J., Van Den Hurk, B., Ward, P. J., & Westra, S. (2020). Multivariate extremes and compound events. In Climate extremes and their implications for impact and risk assessment (pp. 59-76). Elsevier.