Ensemble Hurricane Flood Forecasting with Atmosphere and Wind Stress Uncertainties

Here, we study ensemble hurricane storm tide forecasting that accounts for the dominant uncertainties in the coupled ocean-atmosphere system, using a retrospective forecast of Hurricane Sandy. Atmospheric inputs come from the Coupled Ocean-Atmosphere Mesoscale Prediction System (COAMPS-TC), for which a 21-member ensemble is created that accounts for all uncertainties in initial state. The retrospective forecasts are utilized with the New York Harbor Observing and Prediction System (NYHOPS) domain implementation of the Stevens Estuary and Coastal Ocean Model (sECOM). Uncertainty in the air-sea coupling is also included, through perturbations of the uncertain saturation value of the sea surface drag coefficient.

This research aims to provide societal value by improving coastal flood forecasting for a hurricane-vulnerable region that includes New York City. Forecast outputs are tailored to end-users, who are most interested in knowing if they will be inundated, to what depth, and at what time. These outputs include probabilistic maps of the 5%-chance flood exceedance (near-worst-case), 50%-chance flood exceedance, and 5%-chance earliest time-of-arrival of inundation. Lastly, as Category-3 hurricanes have made landfall in the region in 1788, 1821 and 1938, one of these will also be retrospectively forecasted with simplified parametric meteorological fields, to examine how the results might differ for stronger storms.