Apropos cherrypicking data, long before I started this blog, I happened to read a story in Germany’s Der Spiegel magazine back in 2005 with the alarming headline, “Number of Dangerous Hurricanes Doubles.” According to the blurb following the headline, “The strength of tropical cyclones has increased dramatically since the 70’s. The number of category 4 and 5 hurricanes has almost doubled. A possible result of the greenhouse effect?” I had just read Bjorn Lomborg’s “The Skeptical Environmentalist,” with its many accounts of environmentalists crying “wolf,” so the story had an eerily familiar ring to it. I decided to do some fact checking of my own.
The article on which the Spiegel account was based was entitled “Changes in Tropical Cyclone Number, Duration, and Intensity in a Warming Environment.” It had appeared in the prestigious journal “Science,” where its authors were listed as P. J. Webster, G. J. Holland, J. A. Curry, H.-R. Chang. Apparently the source of Spiegel’s “doubling of dangerous hurricanes” claim was included in the paragraph,
Examination of hurricane intensity (Fig. 4) shows a substantial change in the intensity distribution of hurricanes globally. The number of category 1 hurricanes has remained approximately constant (Fig. 4A) but has decreased monotonically as a percentage of the total number of hurricanes throughout the 35-year period (Fig. 4B). The trend of the sum of hurricane categories 2 and 3 is small also both in number and percentage. In contrast, hurricanes in the strongest categories (4 + 5) have almost doubled in number (50 per pentad in the 1970s to near 90 per pentad during the past decade) and in proportion (from around 20% to around 35% during the same period).
The articles basic conclusions can be found in the last two paragraphs:
We deliberately limited this study to the satellite era because of the known biases before this period, which means that a comprehensive analysis of longer-period oscillations and trends has not been attempted. There is evidence of a minimum of intense cyclones occurring in the 1970s, which could indicate that our observed trend toward more intense cyclones is a reflection of a long-period oscillation. However, the sustained increase over a period of 30 years in the proportion of category 4 and 5 hurricanes indicates that the related oscillation would have to be on a period substantially longer than that observed in previous studies.
We conclude that global data indicate a 30-year trend toward more frequent and intense hurricanes, corroborated by the results of the recent regional assessment. This trend is not inconsistent with recent climate model simulations that a doubling of CO2 may increase the frequency of the most intense cyclones, although attribution of the 30-year trends to global warming would require a longer global data record and, especially, a deeper understanding of the role of hurricanes in the general circulation of the atmosphere and ocean, even in the present climate state.
The authors have included all tropical hurricanes in their conclusions. However, as I have a day job and can’t do climatology full time, I will focus on the largest single region studied; the west Pacific Ocean. This seems reasonable to me, as it had 201 hurricanes during the period studied compared to 85 in the region with the second largest number of occurrences. Anyone who cares to do so is welcome to check the other regions as well. I will gladly post the results if they are substantially different from mine.
First, let’s consider the “doubling of powerful hurricanes” claim. It is based on a comparison of five year periods, or pentads, in the 1970’s versus the decade 1995 to 2005. And, sure enough, if one compares the 19 cat 4 and 5 hurricanes in the pentad 1974 to 1978 with the 44 in the pentad 2001 to 2005, we can go the Science article one better. The number has more than doubled!
But wait! The authors limited themselves to the “satellite era,” beginning in 1975, because, as they put it, “because of the known biases before this period.” The nature of these biases are made clear in a book by Jack Williams of USA Today, who writes,
Until weather satellites began “seeing” eastern Pacific hurricanes in the 1970s, meteorologists had underestimated how many occur because many storms never come near land and fewer ships sail the eastern Pacific than the Atlantic, Caribbean and Gulf of Mexico. The eastern Pacific hurricane season runs from May 15 through Nov. 30.
Assuming a similar bias applies to the western Pacific, then it resulted in under- and not overestimation of the frequency of storms. How very odd, then, that if we push our study back another fifteen years and include the decade of the 1960’s, the “pentads” tell an entirely different story. The largest number of storms in any five year period in that decade was 48, from 1961 to 1965! If pentads are really a compelling argument for a “doubling of powerful hurricanes,” as the authors and the editors of Spiegel claim, than they can all rest easy. The data from the 1960’s “proves” we have nothing to worry about.
According to the authors, they deliberately limited their study to the satellite era, which began in 1975, to avoid the bias referred to above. How very convenient, then, that according to a report of the Naval Pacific Meteorology and Oceanography Center, “1975 saw a sharp decrease in tropical cyclone activity from last season.” How very convenient as well that 2004, the last year cited in their study, had 12 powerful storms in the western Pacific, matching the highest number ever recorded in a given year. Can you say “cherrypicking?” Interestingly, since 2004, the numbers have begun to drop off. For the pentad 2005 to 2009 they are 9, 9, 8, 5 and 7. Presumably the authors, who put such faith in pentads, must be forced to conclude that their conclusions were wrong, and the trend for powerful storms is actually on the decline.
In fact, the pentad numbers that inspired Spiegel’s alarmist headline demonstrate nothing, one way or the other. As anyone can see who cares to actually check the data for the last 50 years, they are dominated by statistical noise.
Let’s take a look at the number of powerful storms in the western Pacific during those years, including the “under-respresented” ones before the satellite era. They are:
1960 to 1969: 7, 9, 8, 10, 9, 12, 5, 7, 8, 3
1970 to 1979: 8, 8, 6, 3, 2, 4, 8, 3, 2, 4
1980 to 1989: 4, 4, 6, 4, 7, 1, 4, 8, 6, 8
1990 to 1999: 7, 9, 10, 6, 11, 6, 8, 12, 4, 2
2000 to 2009: 7, 5, 9, 9, 12, 9, 9, 8, 5, 7
Pretty bumpy data, isn’t it? According to the authors, “There is evidence of a minimum of intense cyclones occurring in the 1970s.” That’s certainly an understatement, and one that the editors of Spiegel predictably didn’t even bother to mention. Include the data from the 1960’s and early 70’s, and the “long term trend” starts fading into the mist.
Tell me, who peer reviews stuff like this, and how does it get published in a journal like “Science?” Here’s what I think: It is likely that global warming is real, and it represents a significant threat. Under the circumstances, scientific integrity is essential to establish the credibility of the danger, and certainly outweighs the need of professors X, Y and Z to pad the list of publications in their CV’s with junk science like this. Surely the authors of the paper were aware of the data from the 1960’s. Was it too much to ask that they at least mention it? Was it too much to ask for them to give a more convincing reason than “the beginning of the satellite era” for cherrypicking a minimum and maximum for their starting and ending dates in accordance with what has now apparently become, as Voltaire put it, “a mere matter of tradition” among environmental scientists?
What can I say? Chalk up one more data point for Bjorn Lomborg. Let’s lighten this post up a bit with a bit of humor in the form of psychobabble from the pages of The Personality and Social Psychology Bulletin. Here’s the abstract:
Despite extensive evidence of climate change and environmental destruction, polls continue to reveal widespread denial and resistance to helping the environment. It is posited here that these responses are linked to the motivational tendency to defend and justify the societal status quo in the face of the threat posed by environmental problems. The present research finds that system justification tendencies are associated with greater denial of environmental realities and less commitment to pro-environmental action. Moreover, the effects of political conservatism, national identification, and gender on denial of environmental problems are explained by variability in system justification tendencies. However, this research finds that it is possible to eliminate the negative effect of system justification on environmentalism by encouraging people to regard pro-environmental change as patriotic and consistent with protecting the status quo (i.e., as a case of “system-sanctioned change”). Theoretical and practical implications of these findings are discussed.
There you go, Luboš. Any time you’re in the market for more psychoanalysis, just slip a nickel in my tip jar