Methodology for estimating return intervals for storm demand and dune recession by clustered and non-clustered morphological events

Abstract

A new methodology is proposed to estimate storm demand and dune recession by clustered and non-clustered events, to determine if the morphological response to storm clusters results in greater beach erosion than that from individual storms that have the same average recurrence interval (ARI) or return period. The method is tested using a numerical morphodynamic model that combines both cross-shore and longshore beach profile evolution processes, forced by a 2D wave transformation model, and is applied as an example within a 20 km long coastal cell at an erosion hotspot at Old Bar, NSW mid-north coast, Australia. Wave and water level data hindcast in previous modelling (Davies et al., 2017) were used to provide two thousand different synthetic wave and tide records of 100 years duration for input to a nested nearshore 2D SWAN model that provides wave conditions at the 12 m depth contour. An open-source shoreline evolution model was used with these wave conditions to model cross-shore and longshore beach profile evolution, and was calibrated and verified against long-term dune recession observations. After a 50 year model spin up, 50 years of storm demand (change in sub-aerial beach volume) and dune toe position were simulated and ranked to form natural estimators for the 50, 25, 16, 12.5 and 10 year return period of individual events, together with confidence limits. The storm demand analysis was then repeated to find the return period of clustered and non-clustered morphological events. Morphological clusters are defined here by considering the response of the beach, rather than the forcing, with a sensitivity analysis of the influence of different recovery thresholds between storms also investigated. The new analysis approach provides storm demand versus return period curves for the combined population of clustered and non-clustered events, as well as a curve for the total population of individual events. In this approach, non-clustered events can be interpreted as the response to isolated storms. For clustered and non-clustered morphological events the expected storm demand for a 50-year return period is approximately 25% greater than that for individual events. Alternatively, for clustered and non-clustered events the magnitude of the storm demand that occurs at a return period of 17 years is the same as that which occurs at a return period of 50 years for individual events. However, further analysis shows that for a 50-year return period, the expected storm demand for the population of non-clustered events is similar to that of the clustered events, although the size of the population of the latter is much greater. Hence, isolated storms can generate the same storm demand as storm clusters, but there is a much higher probability that a given storm demand is generated by a morphologically clustered event.