NARRAGANSETT, R.I. – March 11, 2010 — Past efforts to develop the most accurate hurricane prediction models have been hampered by a lack of adequate computer power to run the models.
A new $20 million project funded by the National Oceanic and Atmospheric Administration aims to rectify that issue, and a University of Rhode Island scientist expects that one result will be a significant improvement in forecasters’ ability to predict the intensity of severe storms.
“The development of the next generation of hurricane prediction models will be fundamentally different from those in the past because no longer will we be faced with limited computational capabilities,” said Isaac Ginis, a professor at the URI Graduate School of Oceanography whose coupled air-sea models have been among the most accurate hurricane prediction models used by the National Hurricane Center over the last decade.
“Since the National Hurricane Center uses several models to create its forecasts, each model must be able to run on a computer in one hour for it to be useful. In the past this has been a major limitation,” he said.
Ginis said that a powerful new supercomputer recently became available at the National Center for Atmospheric Research in Boulder, Colo., that will enable scientists to develop more accurate models.
“In order to simulate a real hurricane, we must factor in wind speeds and air and sea temperatures and other data at certain locations, and the smaller the distance is between those locations the more accurate the model will be,” Ginis explained. “Currently, the spatial resolution is about 10 kilometers, but the optimal resolution is 1 to 2 kilometers or less, which would require a huge increase in computer resources.”
NOAA’s Hurricane Forecast Improvement Project will provide the necessary computer resources while also funding the development of new hurricane intensity models by the leading hurricane scientists in the nation. The project will provide Ginis with about $300,000 per year to continue his research.
Ginis was the first scientist to demonstrate the significant role the ocean plays in the formation, path and intensity of hurricanes. Using data collected from aircraft, satellites and ocean buoys, his models significantly improved hurricane predictions in the Atlantic Ocean and Gulf of Mexico in the last decade.
The new supercomputer will also enable Ginis and other hurricane scientists to factor in additional physical processes into their models, like the effect that sea surface waves and sea spray droplets have on storm intensity.
“Heat from the ocean makes a storm more intense, and friction from waves makes a storm weaker, so to predict hurricane intensity we must calculate the heat and the waves,” Ginis said “Current models frequently overestimate the intensity of weak hurricanes and underestimate the intensity of strong hurricanes because they don’t accurately predict waves and sea spray generated by wave breaking.”
In addition to funding from the Hurricane Forecast Improvement Project, Ginis and colleagues from 10 other institutions have been awarded a $2.1 million grant from the National Oceanographic Partnership Program to create an air-sea interface software module that will allow hurricane prediction models to be used by the U.S Navy, NOAA and private organizations with an interest in hurricane forecasting. The URI share of the grant is $610,000.
The program is a partnership of government agencies, academic institutions, and private organizations to support research that can be applied to commercial and government uses.