An MIT hurricane expert says a simple account of the response of the ocean to a moving storm may be the key to predicting how intense a hurricane will be when it hits land.
While efforts to improve predictions of hurricane intensity have focused on the storm's interaction with the atmosphere, Kerry Emanuel, professor of meteorology in the Department of Earth, Atmospheric and Planetary Sciences, reported in the October 14 issue of Nature that the storm's intensity can best be predicted based on its initial intensity and the thermal structure of the atmosphere and ocean through which the storm moves.
The collection of thunderstorms that generally surround the eye of a hurricane are called the eye wall. By monitoring sea surface temperature changes under that part of the eye wall that lies along the storm track, researchers can predict how intense the storm will be.
"We now think we understand most of the physics that are critical to the intensity of hurricanes. The hurricane responds principally to sea surface temperature changes under its eye wall," Professor Emanuel said. Factors such as wind shear seem to matter most while the storms are still forming, but he says that the thermodynamic structure of the upper ocean is more critical when predicting intensity.
By using this very simple model that can be run quickly on an ordinary PC, "hurricane intensity can be predicted as far in advance as an accurate track prediction can be made," Professor Emanuel said. The model provides a significant improvement over the only tool now used to determine a hurricane's wind speed when it makes landfall: a statistical model that compares it to how similar storms acted in the past.
In the Nature paper, Professor Emanuel uses the model to simulate the intensity of several past hurricanes, such as Andrew, which hit Florida in 1992, and Hugo, which made landfall in South Carolina in 1989. He also simulated Hurricane Camille, a category 5 hurricane that walloped Biloxi, MS in 1969. Camille moved right along the axis of the Loop Current, an extension of the Gulf Stream that loops a deep, warm layer of water toward the coast of Alabama and Mississippi. His simulated predictions closely matched the reality of the storms' intensity.
A hurricane draws its energy from warm ocean water. But the hurricane winds churn through the relatively thin layer of warm water at the surface and bring up much cooler water from the depths.
Researchers have established that were the ocean under the eyewall to cool by 2.5ï¿½ï¿½ï¿½C, it shuts down its own source of energy and fizzles out. So even a one-degree change in ocean temperature is significant in determining how much punch a particular hurricane is packing.
"Cooling water under the eye wall will be detrimental to the hurricane," Professor Emanuel said. "It can't evaporate any more water."
Information about the thermal structure of the top 300 meters of the ocean ahead of the storm's expected path could provide meteorologists with information on where the deep, warm water is.
Measurements of ocean currents and temperature are not routinely made, but they could be gathered with a device that drops from an airplane a floating buoy that deploys a spool of wire. The wire measures temperature and radios the data back to the plane.
This information, in addition to helping make predictions about exisitng storms, would allow researchers to say (for example) that if ocean currents are aligned in such a way in a given year, Texas is unlikely to get a dangerous storm but New Orleans is in line for a whopper.
While the ability to predict a hurricane's track has improved dramatically in the past 30 years, there is still almost no skill at all in predicting intensity. "Beyond 12 hours, we have no predicting skills, and 12 hours is nothing," Professor Emanuel said. It's not enough time to make an informed decision on whether to evacuate, so many more evacuations take place than need to. In addition to being costly, unnecessary evacuations can cause people to think meteorologists are crying wolf. "They may not listen to you next time," he said.
A version of this article appeared in MIT Tech Talk on October 27, 1999.