Modeling Storm Surge in a Small Tidal Two-Inlet System

Abstract

Model simulations using a depth-averaged ocean circulation model (ADCIRC) two-way coupled with a wave model (STWAVE) through the Coastal Storm Modeling System Coupling Framework (CSTORM-MS) are compared with observations made in the shallow, two-inlet tidal Katama Bay system on the Atlantic coast of Martha's Vineyard, Massachusetts, during Hurricane Irene. The CSTORM-MS framework integrates high-resolution bathymetric grids of this system with the North Atlantic Coast Comprehensive Study (NACCS) performed by the United States Army Corps of Engineers. The effects of bathymetric resolution and wave-flow coupling on the accuracy of modeled storm surge were investigated by comparing observations with the high bathymetric resolution, coupled model (CSTORM), a high-resolution uncoupled ADCIRC model, and a low bathymetric resolution, coupled model (NACCS). During the peak storm surge period, the coupled model using high-spatial resolution bathymetry reduced error in the study area by over 30% compared with the lower-resolution NACCS model, and by 16% compared with the high-resolution, uncoupled ADCIRC model. In addition, the high-resolution models indicate alongshore flows with magnitudes over 2.0 m/s along the southern coast of Martha's Vineyard, and a net northward circulation through Katama Bay and Edgartown Channel, which are not apparent in the lower-resolution simulations. Contrary to prior research suggesting small, if any setup in the Katama Bay system from wave forcing, in the extreme wave forcing event discussed here, the northward flux through Katama Inlet on the south side of the bay does not exit completely through Edgartown Channel on the north side of Katama Bay. Thus, the drainage path is not adequate to prevent increased water elevation in the bay, resulting in a setup within Katama Bay during the peak surge event, highlighting the need for adequate model resolution for local storm surge predictions.