The global-warming God: must it now be appeased?

FOR years, I have been concerned that a major hurricane strike on New Orleans could provoke legislation on global warming that will do absolutely nothing about tropical cyclones, but harm the U.S. economy for decades. We began seeing the shape of things to come when Robert F. Kennedy Jr. claimed that Katrina's severity was related to President Bush's reluctance to cap carbon-dioxide emissions, and Hillary Clinton declared she wanted to establish a commission to investigate the government's response to the hurricane.

Hurricane Katrina's magnitude was not changed by global warming. In fact, despite a hundred news stories to the contrary, it's not at all clear that any such warming will result in more frequent, let alone more intense, tropical cyclones. Take, for example, the recent rise in hurricane frequency in the Atlantic Basin; it's as if we have returned to the 1930s, '50s, and '60s, when storm activity reached a ferocious height before settling down for several decades.

Keep in mind, however, that there are some differences between now and then. Today's hurricanes tend to concentrate in Florida and the Gulf of Mexico, whereas in mid-century they repeatedly struck the Atlantic coast--all the way to Canada. This is worth noting because, in a warming world, simple reasoning predicts that activity should have moved north, not south. Obviously things are not so simple with hurricanes and climate change.

But the North Atlantic is just one of the world's many hurricane basins. In general, there are more tropical cyclones per year in the eastern North Pacific--off the Mexican coast--but few notice them because they rarely hit land as consequential cyclones. Instead, they peter out as they migrate westward into cooler waters. In fact, only one subtropical ocean, the South Atlantic, has virtually none of these storms.

So, rather than focusing exclusively on what's happening in our provincial part of the North Atlantic, we should be asking what is happening to hurricane activity around the planet. And the answer is ... nothing. More specifically, there has been no significant movement in either frequency or strength despite a warming trend since 1975, a cooling trend in mid-century, and a warming trend in the early 1900s that was similar to what we see today.

Yes, there is a much-cited paper in the scientific magazine Nature by Kerry Emanuel that claims that hurricanes have doubled in power in the last 30 years. But there are at least tour manuscripts in review at Nature that challenge this result. Indeed, people who normally stay out of the global-warming catfight, like the University of Colorado's Roger Pielke Jr., have been blogging that the peer-review process at Science--Nature's American competitor--has been badly compromised on the subject of global warming and weather-induced damages.

The Emanuel argument simply does not stand up to scrutiny. Put it this way: Since Atlantic hurricanes have a roughly equal chance of striking the U.S., and if they had doubled in power, then surely the amount of damage they have caused--after adjusting for population and property-value changes--would show an obvious upward trend. But, when allowing for these factors, as Pielke has, we cannot detect any such tendency. In any case, if hurricanes had actually doubled in power, then the insurance companies would already have been blown to smithereens. They're still here.

Another paper, by Peter Webster and colleagues, that argues along similar lines has just appeared in Science magazine. It attempts to show that while the global frequency of tropical systems shows no trend since the 1970s, the percentage of extremely strong hurricanes has been increasing relative to the weaker storm systems. This is an example of very selective data citation using post-1970 satellite observations of tropical systems. However, in the North Atlantic, hurricane-hunter reconnaissance flights have been flying since 1944. Guess what? Between 1944 and 1970, the percentage of extremely strong hurricanes declined relative to the weaker storms. Relative storm strength in the '40s and '50s is similar to what it is today. So, over the longer term, there has been no trend in hurricane intensity. How could Science's peer reviewers have missed this?

What of research published a year ago by Tom Knutson, which showed that hurricane winds will increase by 6 percent over the next 80 years as a result of planetary warming? Hurricane frequency varies tremendously from year to year and between decades, so such a small change couldn't be discerned for at least 50 years. Moreover, Knutson's work assumes that ocean-surface temperatures will continue to warm while nothing else changes. But this isn't realistic, for warming may increase the frequency of El Nino--poison to Atlantic hurricanes. Warming may also have an impact on the high-pressure systems that generate the trade winds, thereby placing hurricanes in less favorable environments, and consequently making them less able to inflict major damage.

Clearly, a lot of things can be overlooked in a computer simulation. For example, in the Knutson paper, over half of the simulated behavior of hurricanes is related to the warming of the sea surface. This is an eminently testable hypothesis, as we have a century's worth of data on Atlantic hurricanes and water temperature. When the statistical correspondence between the two is checked, just 10 percent of inter-annual hurricane behavior can be traced to temperature changes. That's right, 90 percent of hurricane fluctuation from year to year is related to factors other than ocean temperature.

How could this be? Let's define a perfect hurricane, like Katrina, as one whose surface barometric pressure approaches 26.5 inches. A storm this strong requires rare conditions. Warm water is just one: That's why there have been only four such storms in the Atlantic over the last 100 years. There must also be, for example, an efficient venting mechanism at the top of the storm, at around 40,000 feet, and there can be no dry air intruding into its circulation. In reality, there are often many bands of relatively dry air within a hurricane, owing to the rainstorms feeding into the eye that gather up moisture. When a significant dry band gets near the center, the winds will drop. Katrina was probably ready to undergo one of these wind-reduction cycles, but, unfortunately, New Orleans got in the way.

Yet still we hear that warmer oceans mean that storms will become more ferocious. We shouldn't. One of the peculiarities of hurricane behavior is that there is a threshold, at 28[degrees]C, where storms reach an intensity of Category 3 ("severe"). (In the last 50 years there have been just two such storms over cooler waters.) But once the temperature exceeds 28[degrees]C, there is no relationship between warmer water and intensity. At any temperature above the threshold, each storm has an equal chance of reaching Category 4 or 5.

Every August, the surface temperature of the Gulf of Mexico exceeds that threshold. This year it was very warm south of Louisiana: 31 [degrees]C--about as hot as it was in 1997 and 1998. And, given that the Gulf reaches the magic 28 [degrees]C threshold every year, no matter whether the planet is warming or cooling, there is no practicable, economical policy that can ever drive temperatures below this figure.

I don't expect this information to have any effect on policy. I'm just shouting into the hurricane.

Mr. Michaels is senior fellow in environment studies at the Cato Institute and professor of environmental sciences at the University of Virginia.