Friday, November 30, 2007

When a Tropical Storm Isn't a Tropical Storm

Today's Houston Chronicle has a fascinating article on a storm that's brewing (forgive the pun) in meteorology circles. Not only did hurricane forecasters again under-estimate the number of storms this year; some meteorologists now claim that at least a half-dozen storms may have been named prematurely. As the Chronicle's Eric Berger notes, the decision to "name" a tropical storm has significance that goes far beyond mere semantics:

The number of a season's named storms forms the foundation of historical records used to determine trends in hurricane activity. Insurance companies use these trends to set homeowners' rates. And such information is vital to scientists trying to determine whether global warming has had a measurable impact on hurricane activity.

According to Dr. Neil Frank, the long-time director of the National Hurricane Center and one of the nation's leading experts on tropical weather systems, "as many as six" of this year's 14 named storms might have failed in earlier decades to earn "named storm status."

"They seem to be naming storms a lot more than they used to," said Frank, who directed the hurricane center from 1974 to 1987 and is now chief meteorologist for KHOU-TV. "This year, I would put at least four storms in a very questionable category, and maybe even six."

Most of the storms in question briefly had tropical storm-force winds of at least 39 mph. But their central pressure — another measure of intensity — suggested they actually remained depressions or were non-tropical systems.

Meanwhile, the current director of the hurricane center says there's no inconsistency in the naming of storms:

"For at least the last two decades, I am certain most, if not all, the storms named this year would have also been named," said Bill Read, deputy director of the Miami-based center.

According to the Chronicle, everyone agrees that technological advances--most notably, an increase in the number of weather satellites--have made it easier to detect and track tropical systems. In fact, it's widely assumed that prior to the late 1970s, meteorologists missed 1-3 tropical storms a year that developed far from land and were short-lived.

What apparently has changed is a greater reliance on surface winds as an indicator of a tropical system. One of the newest weather satellites, dubbed QuikSCAT, was responsible for spotting Tropical Storm Chantal, which formed south of Nova Scotia in late July, and moved quickly out to sea. Some forecasters argue that Chantal was never a tropical system, because if formed off the Canadian coast. Others say that the decision to name the storm was influenced by QuikSCAT technology.

Dr. Frank says that the increased reliance on windspeed and satellite technology to designate tropical storms represents a break with the past:

In earlier years before widespread satellite coverage, the hurricane center placed more emphasis on measurements of central pressure than wind speeds in designating tropical storms and giving those systems names, Frank said. Central pressures and wind speed are related, but the relationship isn't absolute.

Frank said he prefers using central pressure, because it can be directly measured by aircraft dropping an instrument into a tropical system.

If a reconnaissance plane had measured a wind speed above 39 mph during Frank's tenure, the system would not automatically have been named. His forecasters might have waited a day to see if the central pressure fell, he said, to ensure that the system really was a tropical storm

Beyond the meteorological aspects of this argument, there are some rather obvious bureaucratic implications as well. Like other federal bureaucracies, the National Weather Service and its parent, the National Oceanic and Atmospheric Administration (NOAA) have invested heavily in satellite and computer technology in recent decades. The bureaucrats at both agencies are determined to show a return on that investment, and use the information to justify the next generation of satellites and forecasting tools.

Additionally, we believe that NOAA--and other government agencies--are still reeling from Hurricane Katrina and its aftermath. While no one can fault NOAA for its work during the storm, the destruction and havoc wreaked by Katrina has put more pressure on the bureaucracy to generate even better (and earlier) forecasts on tropical systems that might ultimately threaten the U.S. coastline.

Finally, it's also worth remembering that NOAA is heavily "invested" in global warming research and its potential impact on tropical storm intensity. While the agency hasn't officially linked global warming to increases in hurricanes and tropical storms, it does support the notion that greenhouse gases are responsible for the slight rise in surface temperatures, and a corresponding warming of the world's oceans. Amid the clamor to establish a link between global warming and hurricane activity, NOAA can, at the very least, use its data to gain more research dollars, hire additional staffers, and build new research tools.

But before NOAA moves further down the global warming road, we believe the agency should take a close look at Dr. Frank's argument. If he's correct, then hurricane forecasts for 2007 missed the mark badly; if the six "suspect" storms are excluded from the total, then there were only 8 hurricanes and tropical storms in the Atlantic basin this year. That's below the historical average for the past 60 years (10 storms a year) and barely half the original prediction for 2007 (17 named storms).

Despite Dr. Frank's criticism, it's a given that NOAA will not overhaul its criteria for "naming" tropical storms--too much money and ego involved. However, the agency chould take a page out of its own playbook, and provide a qualifier for systems that are borderline tropical storms, or in situations where data is limited. Using that approach, storms could be designated as "suspected" or "wind-indicated" tropical systems, just as the weather service issues warnings for "possible" or "Doppler-indicated" tornadoes. Once more accurate data is received--and tropical storm status is confirmed--the qualifier would be dropped, and the system would receive its designated name.

Obviously, this approach would be controversial, and in some cases, the hurricane center might still be required to make an early call--particularly when systems form quickly off the southeastern U.S., or in the northern Gulf of Mexico. But, with so much riding on tropical storm and hurricane predictions, NOAA owes it to the taxpayer--and the millions of us living along the coast--to provide forecasts that are accurate, not only in terms of storm track and intensity, but also in terms of the real number of tropical systems that form in a given year.

1 comment:

SMSgt Mac said...

Nice treatment of the subject. Even if there were no political or financial angles to the 'science' this would highlight a chronic problem with long term data collection and analyses: as the ability to measure the data changes it becomes increasingly more difficult to normalize the data over the timeframe and take measurement bias out. Perhaps a new standard should be tracked, such as % of named storms that actually become hurricanes or % of hurricanes that reach each level of intensity. That way the headlines could also read: "Fewer Storms Now Become Hurricanes" or "Fewer Hurricanes Reach Category X"