Close Window

Univ. of North Carolina: "Implementation of a high resolution coastal ocean model for operational forecasting"

Final Report

One of the routine responsibilities of NWS meteorologists at coastal offices is to issue marine forecasts that provide weather-related water level and flooding information along the coast. To accomplish this in a truly predictive mode requires accurate forecasts of meteorological variables (primarily wind speed and atmospheric pressure) and accurate forecasts of the resulting response of the coastal ocean. Within a few kilometers of shore, the coastal ocean is strongly influenced by the shape of the coastline and the bathymetry. These physical features are often extremely complicated due to the presence of rivers, sounds, inlets, barrier islands, capes, etc.

Extreme flooding along the North Carolina shore is associated with Atlantic hurricanes making landfall or tracking close to land. However, extratropical storms (which typically occur during the winter months) are more frequent producers of damaging floods and erosion. While these latter storms may be less intense than hurricanes, they are much larger in size and tend to affect a region for a much longer period of time than hurricanes.

The objective of this project was to provide NWSO Newport meteorologists (and others interested in the work) with a powerful tool for forecasting coastal storm surge primarily due to non-hurricane weather systems. The proposed emphasis was on extra-tropical weather systems, based on the assumption that coastal North Carolina typically experiences flooding from several of these storms each year. As it turned out, extra-tropical storm surge was relatively minor during the project year, while direct hits came from two hurricanes. Thus, the emphasis during the second half of the project was shifted to tropical storm surge forecasts, which were provided to NWSO meteorologists for both hurricanes.

The ADCIRC storm surge model was used with an improved grid having high resolution near shore. An extensive set of model runs was made to document the performance of the model using the new grid for purely tidal forcing. These runs also looked at the relative merit of several new global tidal models for providing the open water boundary condition. The outcome of this effort has been a significant improvement in the model's ability to predict tides throughout the domain and will translate into improved predictions of the combined tide and storm surge. An interface was developed to allow ADCIRC to use NWS gridded meteorological data in GRIB format, and the model was run with combined tide and meteorological forcing from early April to the end of June, 1996. Water level outputs were archived and compared to observed water level data at NOAA gauge stations along the entire U.S. East Coast for the purpose of model verification.

The ADCIRC model was run to forecast storm surge caused by hurricane Bertha and hurricane Fran. This information was transmitted to the NWSO Newport, NC and to the NWSO Wakefield, VA and was used by meteorologists for predicting storm surge associated with these storms. In each case, ADCIRC results showed superior accuracy to other storm surge information that was available at the time.