STORM SERENDIPITY

Bad as it was, the 2005 Atlantic hurricane season could have been worse. Data from Earth-observing satellites originally launched to monitor long-term climate trends improved forecasters' predictions of the paths and intensities of the violent storms, and more improvements are coming from the same sources.

While weather satellites have long been used to track hurricanes, sophisticated sounding radars, passive radiometers and other instruments on scientific spacecraft have been pressed into service for predictions. With worldwide weather patterns apparently changing, the stakes for life and property are high.

In the U.S. alone, which gets hit by only about 1% of the big tropical cyclones worldwide, three Category 5 hurricanes approached landfall. One of them--Rita--threatened Galveston, Tex., which was blindsided by a powerful Cat. 4 hurricane on Sept. 8, 1900, that left more than 8,000 people dead.

This year, spaceborne systems gave about five days of warning that the storm was approaching, and sometimes indicated changes in its intensity almost as they happened. When Rita finally made landfall up the Gulf Cost in Louisiana, residents knew they had to move because they were in the path of a storm that, while weakening, still registered as a Cat. 4. Meanwhile, the barrier-island city of Galveston already had been evacuated across the single bridge linking it with the mainland.

In 1900, storm forecasters relied on reports from land-based weather stations and ships at sea, data that were spotty at best. Now the growing maturity of dedicated space-based weather sensors, inputs from the new scientific sensors and broader coverage with polar-orbiting satellites that cover every inch of the Earth's surface, give forecasters more data than ever for the computer models designed to predict how the atmosphere will behave.

"There's been a marked increase in forecasting skill," says Michael King, Earth Observing System (EOS) senior project scientist at NASA's Goddard Space Flight Center. "By skill, I mean the length of duration of which you have a reliable forecast. It used to be two days, and now it's out to four or five days [and] there's been comparable accuracy increasing over longer range."

In this year's storms, a joint U.S./Japanese scientific satellite that NASA almost killed to save money--the Tropical Rainfall Measurement Mission (TRMM)--provided invaluable storm-intensity data. Saved by an outcry from U.S. Gulf Coast lawmakers, TRMM is now set to operate until the end of the decade, when it will likely be replaced by a follow-on constellation of spacecraft called the Global Precipitation Measurement mission now in preliminary development by the two nations (see pp. 46, 48).

Forecasters also received a boost from a sensor mounted on two of NASA's big polar-orbiting EOS spacecraft--Terra and Aqua. The Moderate Resolution Imaging Spectroradiometer (Modis) on those platforms have provided unprecedented data on polar winds, which have generated unexpected improvements in accuracy when plugged into weather forecasting models. Crossing the poles 14 times a day, the Modis instrument allows scientists to calculate polar winds from changes in the water vapor profiles and clouds from pass to pass. The data are much higher resolution than is possible with the scant balloon and other ground-based data sampling in those remote areas.

"When those polar winds were inserted into medium-range weather forecast models . . . they have had a remarkable impact on forecasting, not just in polar regions but down to mid-latitudes," King says, adding that the polar-wind data were particularly valuable in situations where the models were missing severe weather events like snowfall. Forecasting centers in Europe, the U.S., Canada and Japan have started using the data, which were originally generated to serve climate-monitoring research.