NOAA gets superhelp to forecast season's hurricane strength

NOAA gets superhelp to forecast season's hurricane strength

BY PATRICIA DAUKANTAS | GCN STAFF

Better supercomputer simulations let the National Oceanic and Atmospheric Administration predict in advance that this year's Atlantic hurricane season will be of average strength.

Although meteorologists can't say exactly when or where a hurricane will strike land, they aren't clueless. Satellites are feeding increasingly detailed supercomputer models with data about sea surface temperatures, tropical rainfall and wind patterns. (See related stories, Page 26 and Page 37.)

In May, when the hurricane season began, NOAA predicted that the 2001 season, ending Nov. 30, would bring the usual eight to 11 tropical storms. Five to seven could become hurricanes.

One or two will likely make landfall in the United States, and one will likely strike land in the Caribbean.

The National Hurricane Center, part of NOAA's National Weather Service, will update the forecast next month.

The more dramatic 2000 hurricane season had 14 tropical storms big enough to be named. Eight of them turned into hurricanes.

This is the fourth season that NOAA has issued a long-range hurricane forecast, said Gerald Bell, a meteorologist with the National Centers for Environmental Prediction's Climate Prediction Center. NCEP is a division of the National Weather Service.


The Geostationary Operational Environmental Satellite 7 photographed Hurricane Andrew, one of North America's most destructive storms.


Although the duration and path of individual storms cannot be predicted more than 14 days in advance, it's possible to predict wind and pressure patterns earlier. The hurricane center posts graphical forecasts for wind and waves over various oceans several times a day for up to 72 hours in advance. There are 30-day seasonal predictions, too.

How much and where

The Tropical Rainfall Measuring Mission, or TRMM, satellite produces data about rainfall amounts and locations, said Naomi Surgi, advanced-project leader with NCEP's Environmental Modeling Center.
'We couldn't do any of this without models that assimilate satellite data,' Bell said.

The primary tool is the Princeton hurricane prediction model developed at NOAA's Geophysical Fluid Dynamics Laboratory in Princeton, N.J. For the first time this year, the Princeton model is coupling oceanic and atmospheric effects for greater realism. A hurricane grows stronger when it draws heat from a warm sea surface and weakens over colder stretches of ocean.

Bell's research looks at large-scale patterns such as the El Ni'o phenomenon, in which warm surface water appears in the eastern equatorial regions of the Pacific Ocean. That change affects global wind and rainfall patterns, and ultimately the number of tropical storms in a given season.

Arthur Wick, chief of the Network and Communications Branch of NCEP's Central Operations Division, said the hurricane models now run on an IBM SP supercomputer, which is configured as two independent systems that each rank among the world's 10 fastest. Each system theoretically can perform about 1.5 trillion floating-point operations per second, but it delivers only about 5 percent to 10 percent as much on real scientific problems.

A Cray C90 in Suitland, Md., ran the models until fire destroyed it in 1999. The weather models execute 42 times faster on the IBM SP, Wick said.
Even though NOAA predicted only an average tropical storm season this year, any one of them could wreak havoc, Surgi said.

The 1992 hurricane season started slow. But its first named storm'Hurricane Andrew'caused $26.5 billion in damage and killed 23 people in the United States.

inside gcn

  • data science (chombosan/Shutterstock.com)

    4 steps to excellence in data analysis

Reader Comments

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above