A new paper has been published today in Nature by Elsner et al. which documents trends in the strongest tropical cyclones in a satellite record. The analysis overcomes some of the challenges of other studies by relying on a single dataset from 1981-2006, thus avoiding the problems of inhomogeneities when datasets are combined, and thus adds some useful knowledge to a growing literature. The paper and its reception also indicate how dysfunctional the hurricane science-media interface is, particularly this time of year.

Consider the following examples of complete misinterpretations of the paper:

From the UK Telegraph:

Global warming will drive the number of major hurricanes to cause devastation up by one third in the coming century, says a new study that claims to have found the first conclusive evidence of the effect.

Actually, the paper offers absolutely no predictions about devastation in the coming century, and in fact, no predictions of the future at all.

From Live Science:

Strong hurricanes are getting stronger, likely thanks to global warming, a new study finds.

Actually, the paper says absolutely nothing about attribution, a point emphasized to me in an email from second author Jim Kossin.

From BBC:

Writing in the journal Nature, they say the number of weaker storms has not noticeably altered.

Maybe changes only matter if they go in one direction. If the total number of storms has not changed, and the number of strong storms has increased, then surely the number of weak storms must have decreased proportionately.

However, on the last point, the BBC’s confusion might be understandable. See if you can make sense of this passage from the paper:

Upward trends can be interpreted as an increase in the number of cyclones exceeding a threshold quantile. For example, at the 80th percentile, on average 17 cyclones globally exceed 49ms. With a 1 C rise in SST, the 80th percentile increases to 51ms. At this threshold level, on average 13 cyclones per year are observed. So the increase in SST of 1 C results in an increase in the global frequency of strong cyclones from 13 to 17 cyclones (31%) per year.

Either there is a transposition in there, or I just don’t get it. I’ve asked Jim Kossin about it but haven’t yet heard back. Any guesses as to how to interpret this passage are welcomed in the comments.

Finally, scientists commenting on the paper show their tribal affiliations all too readily. Climate scientist Judy Curry commented on the new trend analysis:

“It’ll be pretty hard now for anyone to claim that cyclone activity has not increased,” says Judith Curry, an atmospheric researcher at the Georgia Institute of Technology in Atlanta.

The only problem is that when another paper (PDF) came out a few years ago with opposite conclusions, Curry criticized it as having too short a data record:

35 years is marginally short to identify a statistically significant trend (people who criticized our study because the length of the data record is too short raised a legitimate point). 20 years is definitely too short. The reason for this is that both the atlantic and pacific have large multidecadal modes. if you pick a period that is too short, what you are seeing is one piece of the mode.

Perhaps detection of some trends requires a shorter time period than others Actually The Elsner et al. dataset shows no significant trends in the most intense storms for the period 1986-2006, according to a response to a query I made to Jim Kossin.

Given that Elsner et al. make no mention of what the IPCC calls “internal” or “natural” variability (PDF), the analysis will be little more than a Rorschach Test for its readers. As the IPCC notes (PDF, p. 667 ff.):

An identified change is ‘detected’ in observations if its likelihood of occurrence by chance due to internal variability alone is determined to be small.

Elsner et al. say nothing about how the observed trend might be interpreted in the context of variability. So, setting aside concerns about the data or methods and accepting their findings as reported, there is no basis for assess whether they have found one part of a lower frequency variation or have “detected” (in IPCC parlance) a trend.

What does all this nonsense, incoherence, and inconsistency say about the debate over hurricanes and climate change? The answers are not going to be clear or settled anytime soon. However, if you’ve already made up your mind, or prefer a certain outcome, there is plenty of available science out there to buttress those views.

The very public theater in which the hurricane-climate debate is playing out sure is entertaining, to paraphrase Richard Tol, but it has done very little for the practice of science.

Popularity: 1% [?]

This entry was posted on Wednesday, September 3rd, 2008 at 3:32 pm and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. Responses are currently closed, but you can trackback from your own site.