MIT Research :Seafoam hits fan over hurricane research

April 29, 2008; updated May 1, 2008


Kerry Emanuel, MIT professor of meteorology, ignited a storm of worldwide media attention in 2005, when he published a paper in Nature linking global warming with increased hurricane intensity. The paper appeared just three weeks before hurricane Katrina devastated New Orleans, and the reported connection with climate change spawned big headlines.

In the years since, a number of scientific papers have appeared on the subject, some supporting Emanuel's original findings, some disputing them, and a vigorous debate has raged on. So it's no surprise that when Emanuel published another paper on the subject earlier this month, it quickly attracted attention.

While the first paper looked at the record of actual hurricanes for the last 30 years, the new one used computer models to look both backward in time and ahead at what could happen two centuries from now. The results were a bit more complex than the earlier ones. We asked Emanuel about the new results, and their media coverage.

Q: How did the new results differ from your 2005 paper?

A: I got mixed results. On the one hand, they backed up the conclusions from what we had deduced earlier from the hurricane data: There was a big increase in storm power over the last 25 years. On the other hand, when the same technique was used go forward in time, the increase in hurricane power was much more modest. Although we get variations from one model to another [of the seven global circulation models used in this study], even the greatest increases were no more than what we've already seen. So there's a dichotomy -- what you see going into the future is not nearly as dramatic as what you would get if you just extrapolate from the past.

Q: How do you interpret that discrepancy?

A: First, it might be because the past had little to do with global warming. Second, it might be because of some unknown, systematic errors in the climate models. Or third, less likely, the climate may be different in a rapidly changing world than in one that has equilibrated. I don't have a good feeling as to which combination of these is responsible.

Q: What does this say about the connection between global warming and hurricane power?

A: Often these increases in hurricane power, thanks largely to my paper, are seen as directly related to global warming. But it should be seen as more complex than that.

Q: How much confidence do you have in the new method used in this paper?

A: The technique verifies very well when driven by past climate data-- it predicts hurricane activity that's pretty consonant with what we see. We get about the right number and magnitude of hurricanes.

Q: How do you feel about the media coverage of the new paper?

A: I thought [the Houston Chronicle story] wasn't bad except for the title [which was "Hurricane expert reconsiders global warming's impact"]. The actual content was okay. In other cases, the people you expect to put a spin on it, put a spin on it. Skeptics' blogs reported that I'd reversed my position. Other blogs latched onto the fact that we're still predicting a very substantial increase. There's a lot to spin.

How do you write a paper that you know could be spun both ways? You just do the best you can.

Q: Were you surprised by the reaction?

A: It is treacherous. Most of what I publish is not subject to public scrutiny, I'm writing for fellow scientists. But in this charged atmosphere [on global warming], most of what you write gets dissected by people outside the community.

[The Chronicle story] was clever to point out that people who are crowing [over the claim of a reversal] are in effect crowing over the same climate models that they spend most of their time criticizing.

When Fox News called me up, they started from the premise that I'd reversed myself. I said that's really not true, it's just that things are more complicated. It was a very short interview. I guess that's what happens when people don't say what's expected of them.

Q: Where does this research go from here?

A: The first thing we will do is to figure out why the simulation of the last 25 years did what it did. What specific physical factors caused the increase? Then, we perform the same analysis on future climate states. I think we will be able to figure out why future changes are so much less than what we have witnessed in the recent past.

Also, we can look at more regional effects. All we looked at was basin-wide effects. There could be differences -- for example, hurricanes could be more easily steered offshore, or more preferentially steered onshore. Even if the basin-wide effects are the same, you can get potentially important regional differences The technique can do that, but we haven't looked at that yet.