Assessing the temporal variability in extreme storm-tide time series for coastal flood risk assessment

The probability of extreme storm-tide events has been extensively studied; however, the variability within the duration of such events and implications to flood risk is less well understood. This research quantifies such variability during extreme storm-tide events (the combined elevation of the tide, surge, and their interactions) at 44 national tide gauges around the UK. Extreme storm-tide events were sampled from water level measurements taken every 15 min between 1993 and 2012. At each site, the variability in elevation at each time step, relative to a given event peak, was quantified. The magnitude of this time series variability was influenced both by gauge location (and hence the tidal and nontidal residual characteristics) and the time relative to high water. The potential influence of this variability on coastal inundation was assessed across all UK gauge sites, followed by a detailed case study of Portsmouth. A two-dimensional hydrodynamic model of the Portsmouth region was used to demonstrate that given a current 1 in 200 year storm-tide event, the predicted number of buildings inundated differed by more than 30% when contrasting simulations forced with the upper and lower bounds of the observed time series variability. The results indicate that variability in the time series of the storm-tide event can have considerable influence upon overflow volumes, hence with implications for coastal flood risk assessments. Therefore, further evaluating and representing this uncertainty in future flood risk assessments is vital, while the envelopes of variability defined in this research provides a valuable tool for coastal flood modelers.