Forecasting Hurricane Irma and its Storm Surge Using an MPAS-ADCIRC Ensemble

As Hurricane Irma and other storms during the 2017 hurricane season illustrated, it remains challenging to provide accurate location-specific forecasts of storm surge and other hurricane hazards at the lead times required for protective decision making. Ensemble prediction can help characterize forecast uncertainty, and it can help forecast the potential for different hurricane hazards at different locations. This presentation will examine an ensemble of Hurricane Irma storm surge forecasts produced by a one-way coupled ensemble featuring the Model for Prediction Across Scales (MPAS) as the meteorological driver of the Advanced Circulation Model (ADCIRC). By exploring the predictability of the hurricane and its resulting storm surge at varying lead times, this work will explore how uncertainty in an atmospheric model ensemble is translated by the storm surge model into uncertainty in surge inundation. Building on previous work, results will be analyzed using different metrics, including maximum water height at specific coastal locations and integrated, storm-following inundation volume. To help interpret the results, this ensemble of storm surge scenarios driven by output from a full meteorological model will be compared with an ensemble driven by an idealized storm and simple perturbations to its track, size, and intensity.