A critically important component for accurate storm surge prediction is the atmospheric model. Since 1999, in collaboration with several National Weather Service (NWS) forecast offices; the Stony Brook MSRC has been running the MM5 meteorological model in real time, twice daily posting the results on the Web . A 17-member ensemble of MM5 and the newly available Weather Research and Forecasting model
(WRF)
is also available each day. The process is fully automated, so that MM5 forecasts are available 24/7. Local NWS forecasters use the information to announce forecasts affecting 10 million residents in the region. The MSRC MM5 forecasts are also used for energy-load forecasting in a project jointly supported by the Long Island Power Authority.
Storm Surges are created by surface winds and a drop in barometric pressure.
A representation of the meteorological conditions in hurricanes and severe nor 'easters is needed
to drive the hydrodynamics in ADCIRC. Fortunately, the Stony Brook Marine Sciences Research Center (MSRC)
has several years of experience in applying and refining the MM5 meteorological model for
the local weather forecasting.
( MM5- community model homepage)
MM5, the Penn State - National Center for Atmospheric Research Model, is a
terrain-following sigma-coordinate model designed to simulate and predict mesoscale and regional atmospheric
circulation and precipitation. Since September 1999, MSRC, in collaboration with the National Weather Service
forecast office in New York City, has been running MM5 in real time twice daily and posting the results on the web.
The National Weather Service and other forecasters use the MM5 predictions in their
daily operations to help forecast weather in the metropolitan region. These forecasters also provide feedback on
the model performance, which has been useful in fine-tuning the model. All real-time MM5 forecasts have been verified
using all conventional surface and upper air data.
An ensemble MM5 Forecast system has recently been developed, which uses various initializations and model physics to
help quantify forecast uncertainty.
The domain of the MSRC MM5 model centered in New York City, covers most of New Jersey
and Connecticut and extends out to sea to the edge of the continental shelf. The local MM5 model is driven by atmospheric
conditions simulated in the two larger National Weather Service models in which it is embedded, which have domains extending
across the country but with coarser resolution.
The database that has been retained for MM5 includes time-tested
simulations of two severe storms, Hurricane Floyd (downgraded in this region to a tropical storm) of September 1999 and
the December 2002 nor 'easter, which caused local storm surges. These were used to validate the Stony Brook Storm Surge
model.
