The role of beach state and the timing of pre-storm surveys in determining the accuracy of storm impact assessments

Dune erosion principally occurs when water level exceeds the elevation of the beach and predicting erosion is progressively becoming more important for management as coastal populations increase, sea level rises, and storms become more powerful. This study assesses storm impacts using a simple model from Stockdon et al. (2007) configured with oceanographic information from the ADCIRC + SWAN model and frequently collected beach profiles. We applied that model to barrier islands in North Carolina including: Core Banks with a more dissipative beach morphology and Shackleford Banks and Onslow Beach with intermediate beach morphologies. The study periods captured 10 events where wave collision with the dunes and/or overwash were either predicted or observed, including large multiple-day events caused by hurricanes and smaller events caused by onshore winds and high tide. Comparing model output with a time series of beach photographs shows the predictive power and sensitivity of the model was consistently high at the Core Banks Site with its wide and low-gradient beach, high-elevation dunes (2.58 m), and high resistance to overwash. Model predictive power and sensitivity was lowest at the Shackleford Banks Site because frequent and large changes to beach slope and intermediate dune elevation (0.54-1.25 m) caused small variations of modeled total water level to either overpredict or underpredict storm impacts. In addition, storm impacts were always overpredicted during hurricanes at the Shackleford Banks Site, which was likely due to storm waves decreasing the beach slope from what was measured prior to the event and used as model input. Like Shackleford Banks, the beach slope of the Onslow Beach Site was steep and variable, but the low-elevation dunes (0.24-0.28 m) made resistance to overwash low and the predictive power and sensitivity of the model higher than at the Shackleford Banks Site. Results suggest that storm impacts and the associated potential for dune erosion is predicted more accurately at beaches where the threshold for overwash is high or low because total water level during most events will commonly fall short of or exceed the overwash threshold, respectively. The accuracy of predicting the storm impact regime is sensitive to beach slope. The slope of intermediate beaches is more variable than dissipative beaches and requires frequent measurement if it is to be represented accurately in the model, but this can be impractical and costly even using the latest drone-surveying methods. To maximize the accuracy of predicting storm impacts, intermediate beach morphology should be constrained by surveying at seasonal or yearly time scales and used as input to numerical models that estimate beach slope over short time scales (hours during an event or daily), configured with the latest wave and water-level forecasts.