Spatio-temporal Clustering of Storm Surges along the Global Coastline

When storm surges often affect the same coastline stretches simultaneously (i.e., they cluster in space, leading to spatial compounding) or if they occur in close succession (i.e., they cluster in time, leading to temporal compounding), the impacts are often greatly amplified. Hurricanes Irma and Maria in 2017 in the eastern Caribbean and Hurricanes Ian and Nicole in Florida were recent reminders how back-to-back storm surges affecting long coastline stretches can cripple economies and societies which are still in recovery mode. This can be a significant burden for the (re-)insurance industry and government budgets, as has been shown for the case of river floods (Jongman et al., 2014). Despite many examples where spatial or temporal compounding effects worsened coastal flooding impacts, developing appropriate tools to incorporate such events into present-day and future coastal flood impact assessments and hazard mitigation planning is still at its infancy. This presentation will showcase a novel algorithm to identify independent storm surge events and preliminary results from applying it to a global tide gauge data set to detect hotspots of temporal storm surge clusters at different time scales and different levels of extremeness. Results from identifying spatial storm surge footprints along the global coast and associated non-stationarity (for selected coastline stretches) will also be presented. The latter will be linked to large-scale weather patterns causing shifts in the spatial footprints at seasonal to decadal time scales. The results can inform the development of flexible statistical models capable of capturing both spatial and temporal dependences to overcome existing limitations in flood risk assessments where this is typically ignored.