It’s an odd day when I find myself agreeing with Benny Peiser on a climate change question, but his post Thursday to his CCNet list, in which he blasted media of coverage of the Nature paper on possible changes in the thermohaline circulation, seems on point.
In fact, as I’ll point out below, there’s been some good coverage – including good skepticism in the New York Times from one of the RealClimate crew. But lots of the coverage has been very bad, highlighting a problem with the system by which results get hyped by virtue of their presence in Nature.
James Annan, in the comments below, pointed to a post in the comments at RealClimate by ocean climate guy Martin Visbeck that gets to the heart of the issue about the way this issue played in Nature and therefore the international press.
This paper is a great case study of the issue I’ve been working through recently – how might the mainstream media most usefully report on these sorts of scientific controversies? So first, it’s worth reviewing some of the more, ahem, “interesting” media coverage. William helpfully linked to and dismembered the Guardian’s headline: Alarm over dramatic weakening of Gulf Stream:
The powerful ocean current that bathes Britain and northern Europe in warm waters from the tropics has weakened dramatically in recent years, a consequence of global warming that could trigger more severe winters and cooler summers across the region, scientists warn today.
Ignoring the bad “Gulf Stream” headline and looking instead at the substance of the paper, as Visbeck’s explanation makes clear, the result in the Nature paper says far less. It might be most accurately characterized as a tentative result suggesting a possibility, rather than a firm conclusion – the sort of thing that would be in GRL rather than Nature if the implications weren’t so sensational. As Richard Kerr points out in Science, paraphrasing the study’s author (this is something Benny seized upon), “The slowing, although sizable, is comparable to the estimated uncertainty of the observations.”
Here’s the problem, as Visbeck points out. This is a hard thing to measure, and other results don’t point in the same direction as Harry Bryden’s Nature paper. But, according to Visbeck, “unfortunately these updates usually does not make a Science or Nature paper.” The dilemma, therefore, is that the high public visiblity of the Nature press process, combined with Nature’s decision in this case to pick the outstandingly interesting result, has the potential to create a bias in the public’s understanding. Nature, in other words, acts as a megaphone, amplifying a result that should in fact be approached with caution.
You can’t entirely blame Nature. An accompanying News and Views piece by Detlef Quadfasel offers some appropriate caveats:
But how solid are these results? The findings are based on just five snapshots of the circulation, taken in 1957, 1981, 1992, 1998 and 2004, and along one latitudinal section. Higherfrequency variability (such as eddies or waves), at the ends of the section at the African coast and the Bahamas, may obscure the detection of long-term change. And the uncertainty of the estimates given is high, so the magnitude of the decline may well be smaller than suggested by the calculations.
It’s worth noting here how Andrew Revkin handled the dilemma in his New York Times piece:
Other scientists were more cautious. Gavin A. Schmidt, a climate modeler at the National Aeronautics and Space Administration Goddard Institute for Space Studies, said that the estimated decline in ocean circulation should have produced a perceptible decline in surface temperatures, but that no such dip had yet been measured.
Robert Dickson, of the British Center for Environment, Fisheries and Aquaculture Science, said that given the complexity and variability of the seas, much more data was needed to determine whether a slowdown was under way.
“However much statistical rigor is brought to bear, five transocean sections is still a small number on which to depend,” he said.
That’s about the right way to handle this.