Annual prediction of shoreline erosion and subsequent recovery

Abstract

Prediction of the potential impact of an extreme storm-sequence on coastal resilience and the subsequent rate of post-storm recovery is a fundamental goal for coastal scientists, engineers and managers alike. This paper presents a methodology for forecasting shoreline change over annual time-scales, including the prediction of the potential impact of storm sequences and the subsequent post-storm recovery. The methodology utilises an archive of measured or modelled wave data to produce multiple (≈103) synthetic hydrodynamic forcing time-series to drive an equilibrium shoreline model in a Monte Carlo simulation. A Generalised Extreme Value (GEV) analysis is conducted on the resulting shoreline predictions in order to extrapolate the magnitude of shoreline displacements for predefined return periods. Three shoreline displacement bands are defined in a ‘traffic light’ system, to aid the interpretation of results; a green (normal) band characterising shoreline displacements expected within the typical decade, an amber (high) band defining events with return periods outside the decadal recurrence threshold but within return periods <100 years, and a red (extreme) band designed to encompass the theoretical limit of the shoreline predictions. The methodology was tested on two field sites with distinctly contrasting wave climates and tidal regime. The first was Perranporth in the UK with a strong seasonal variability in both the wave climate and shoreline response. The second was Narrabeen, Australia, with a much smaller seasonal variability and more storm-dominated wave climate and shoreline response. In both cases an equilibrium shoreline model (ShoreFor) was calibrated using measured shoreline data and complementary wave data. The prediction methodology was found to be mildly sensitive to the temporal range of the wave data used, with at least 25-years of data required to achieve consistent classification of the magnitude of storm erosion and recovery. Two extreme storm sequences were targeted to test the methodology, the Pasha Bulker storm sequence recorded at Narrabeen in June 2007 and the extreme storm sequence which impacted the UK during the winter period of 2013/14. All wave and shoreline time-series recorded in this period were left unseen in model calibrations and subsequent predictions, in order to provide a rigorous test of the methodology. In all cases the methodology was able to predict both storm erosion and subsequent recovery and give a clear indication of the intensity of the shoreline displacement. The storm sequences studied forced shoreline displacements rated as high at Narrabeen and extreme at Perranporth and both sites displayed rapid post-storm recovery. The impact of extreme storms on shoreline recession and subsequent post storm recovery is predictable at these energetic cross-shore transport dominated sites, promising the potential for a new coastal management tool.