July 15, 2007, 4:45AM
Opinions vary on satellite's use as a tool for forecasting
Some at NOAA says it's vital for storm tracking, while others say that's a stretch
By ERIC BERGER
Copyright 2007 Houston Chronicle
Get smart with SciGuy Ousted National Hurricane Center director Bill Proenza never had a reputation for holding his tongue.
Last week, following a firestorm that erupted when he questioned his bosses' spending priorities and a subsequent staff revolt, he was placed on leave.
What may ultimately have cost him his job, however, was not a lack of political acumen but rather the science of hurricanes and satellites.
Shortly after becoming the hurricane center's director in January, Proenza began publicly beseeching his National Oceanic and Atmospheric Administration superiors for a replacement to the aging QuikSCAT satellite. If lost, he warned, his forecasters would lose a key tool to predict the path of land-falling storms.
Proenza gained support from politicians and editorial boards. But some scientists — including senior forecasters at the National Hurricane Center — said Proenza had vastly overstated the effect of the satellite's loss. One, Lixion Avila, compared it to driving a BMW with cloth rather than leather seats.
But did Proenza really stretch credibility? Are lives really at stake?
Scientific evidence supports arguments on both sides.
Two benefits
Launched in 1999, NASA's Quick Scatterometer satellite employs a specialized radar that takes 400,000 daily measurements of wind speeds and directions, covering 90 percent of the Earth's surface.
The satellite benefits hurricane forecasters in two ways: by improving the identification of tropical disturbances well out to sea as they begin to develop circular winds, and by adding data to sophisticated computer models that forecast hurricane tracks.
The first application becomes considerably less important to forecasters as a storm nears land. Combined with reconnaissance data, observations from ocean buoys, ships and Caribbean islands give a much clearer picture of a tropical system, said Chris Hebert, a hurricane specialist with Houston-based ImpactWeather, a private forecasting service.
"QuikSCAT data are generally ignored by operational forecasters when other, more reliable and more frequently updated observations become available," he said.
More debatable is the value of QuikSCAT data in several computer models upon which forecasters rely. There have been essentially three scientific inquiries into whether QuikSCAT data benefit the models:
• Robert Atlas, director of NOAA's Atlantic Oceanographic and Meteorological Laboratory, has published two peer-reviewed studies and concludes the loss of QuikSCAT data would harm hurricane track forecasting, in some cases severely. For example, he found the 60-hour forecast for Hurricane Cindy (1999) with QuikSCAT data was better than the 24-hour forecast without the satellite data.
• In an unpublished study of nearly 300 tropical systems, James Goerss and Timothy Hogan of the Naval Research Laboratory found that QuikSCAT improved the one-day forecast of the Navy's NOGAPS model by 2.5 percent but had no effect on longer forecasts.
• And in a study conducted by NOAA scientist James Jung and others, accepted for publication later this year in a peer-reviewed journal, QuikSCAT improved a key model's two-day forecast by 10 percent and the three-day forecast by 16 percent.
It was this last study that Proenza relied upon to say QuikSCAT is a life-saving tool for hurricane forecasters.
'A degradation' of accuracy
While standing behind his work, Jung acknowledged in an interview that his sample of 25 storms was relatively small, however, and that the effect of satellite data could vary from year to year. "In this study, there is no way to account for yearly variability, which I feel is significant," Jung said.
Goerss also defended his work but said the results applied to the NOGAPS model only and are not generalizable to other models used by the hurricane center.
Among the three researchers, only Atlas said the scientific results point toward a definitive conclusion that wholly backs Proenza. "The weight of the scientific evidence is that if we lose QuikSCAT, and we don't do anything to mitigate that loss, we will have a degradation of forecast track accuracy," Atlas said.
Hurricane center staffers haven't always dismissed QuikSCAT as a forecasting tool. As recently as early June, according to a summary of a workshop on satellite wind measurements, senior hurricane specialist Richard Knabb said of QuikSCAT:
"When QuikSCAT is gone, it will be like going back seven years in tropical cyclone analysis." He later added: "Losing QuikSCAT would be like losing a limb, especially for Tropical Analysis and Forecasting Branch."
However, a month later, Knabb's colleagues, including Avila, seemed to say something different when they held a news conference in Miami and minimized the importance that satellites play in their forecasting.
Later, in an e-mail, another staffer, James Franklin, disputed Atlas' favorable assessment of the satellite.
If Proenza feels betrayed by his staffers at the National Hurricane Center, he isn't saying. Attempts to interview him for this article were unsuccessful.
A plan in place
The QuikSCAT controversy is unlikely to go away soon. On Thursday, members of Congress who oversee NOAA sent its administrator a letter seeking a large number of records relating to the decision to oust Proenza.
Among them was the chairman of the subcommittee on energy and environment, Rep. Nick Lampson, D-Texas.
NOAA, the hurricane center's parent organization, has come down squarely in favor of Franklin and the senior forecasters. During a hearing before the Senate Commerce Committee last week, NOAA assistant administrator Mary Ellen Kicza essentially repeated Franklin's assessment.
She also said the agency has a plan if QuikSCAT fails. It could use two satellites, including one recently launched by Europe, to cover about 60 percent of QuikSCAT's area and increase the use of reconnaissance flights.
Proenza's supporters, including Atlas, noted that such a contingency plan wasn't developed until after the former hurricane center director began publicly agitating for a satellite replacement.
As for the politicians who initially backed Proenza's call for a new satellite, they no longer appear inclined to push for an immediate replacement. Instead, they seem likely to follow NOAA's plan of developing a version that can detect wind speeds above Category 1 hurricane status, which is the limit for the current satellite.
And, despite earlier fears among some forecasters, members of Congress will no longer consider raiding Hurricane Hunter funds for a satellite replacement.
"No, we are not going to rob Peter to pay Paul," said Rep. Ron Klein, D-Fla.
eric.berger@chron.com
