If those opposed to action on climate change are like Ahab, the scientific consensus is their white whale. The reason is simple: as Frank Luntz’s famous memo pointed out, if they can convince the general public that the science of climate change is uncertain, they can drag the debate over policy to a grinding halt.

Consensus. Photo: iStockphoto

Thus, every so often, another argument emerges that purports to prove that scientific consensus on climate change does not exist.

This week, it’s a blast from the past: an analysis of the “Web of Science” that shows that no consensus exists and only a minority of scientists support the views of the Intergovernmental Panel on Climate Change (IPCC).

First, some background. For those who aren’t familiar, the Web of Science (WoS) is a massive database that includes the title and abstract of essentially every scientific paper published since the early ’90s. There’s also a ton of ancillary information in the database, such as how many times a paper has been cited. It’s an invaluable tool to the scientific community, one I use on an almost daily basis to find papers in the peer-reviewed literature.

Naomi Oreskes, a Professor of History and Science Studies at UC-San Diego, searched WoS for papers that include the phrase “global climate change” in the title or abstract and found that basically none of these papers explicitly reject the consensus position (i.e., the earth is warming, humans are very likely responsible for most of the recent warming, etc.). See Coby Beck’s writeup for more details.

A medical researcher, Dr. Klaus-Martin Schulte, has revisited Oreskes’ analysis. Oreskes looked at papers published between the mid-1990s and 2003, while Schulte looked at papers published after 2004. I have not actually seen a copy of this new paper, but I’ve reconstructed its salient points from a description of the analysis found here (PDF). The abstract of Dr. Schulte’s paper:

Fear of anthropogenic ‘global warming’ can adversely affect patients’ well-being. Accordingly, the state of the scientific consensus about climate change was studied by a review of the 539 papers on “global climate change” found on the Web of Science database from January 2004 to mid-February 2007, updating research by Oreskes (2004), who had reported that between 1993 and 2003 none of 928 scientific papers on “global climate change” had rejected the consensus that more than half of the warming of the past 50 years was likely to have been anthropogenic. In the present review, 32 papers (6% of the sample) explicitly or implicitly reject the consensus. Though Oreskes said that 75% of the papers in her sample endorsed the consensus, fewer than half now endorse it. Only 7% do so explicitly. Only one paper refers to “catastrophic” climate change, but without offering evidence. There appears to be little evidence in the learned journals to justify the climate-change alarm that now harms patients.

This analysis is rubbish. First, consider the following abstract, from a paper entitled, “An analysis of the regulation of tropical tropospheric water vapor“:

We use a simple trajectory model of mid- and upper-tropospheric H2O to investigate the mechanisms that regulate mid- and upper-tropospheric humidity. Our model advects water passively and contains no microphysics other than the requirement that water vapor is immediately removed so as to prevent the relative humidity (RH) from ever exceeding 100%. We demonstrate that our simple model accurately reproduces H2O measurements made by the Atmospheric Infrared Sounder onboard NASA Aqua satellite. Our results show that, given the large-scale circulation of the troposphere, detailed microphysical processes need not be included in order to accurately simulate H2O. We have also identified three preferred regions where air parcels in the mid- and upper troposphere experience their final dehydration. The first is in the equatorial upper troposphere, and is associated with convective outflow at the top of the tropical Hadley circulation. Final dehydration of air that detrains at potential temperature above ~340 K predominantly occurs here. The other two regions are found at lower altitudes in the mid-latitudes of both hemispheres, and are associated with dehydration during isentropic excursions to mid-latitudes. Final dehydration of air that detrains at potential temperatures below ~340 K predominantly occurs here. Finally, we analyze the water budget of the dry Eastern Pacific subtropics and find that dehydration in both the equatorial upper troposphere and the mid-latitudes contribute to the dryness there.

Can you tell me if this paper accepts or rejects the consensus view on climate change?

You can’t. It is impossible for anyone but an expert in this particular climate sub-field to be able to read this abstract and understand the implications for the theory of climate change. Most climate scientists could read the entire paper and understand the implications, but it takes a true expert in a particular field to be able to understand the implications just from the abstract.

Why? Because the abstract of the paper contains only what’s new in the paper. Thus, the implications of the paper to our wider knowledge can only be understood if one is familiar with everything that’s been previously published on the topic.

In this case, the paper strongly supports the IPCC view of climate science. I know, because I wrote it. But for non-experts like Dr. Schulte, it is just a bunch of gobbledy-gook.

The entire approach of determining whether a paper supports or contradicts the consensus view based on a reading of the abstract by a non-expert is flawed. Note that this criticism also applies to Oreskes’ original paper. While she got the right answer, I have always been uncomfortable with the methodology.

Another reason the analysis is rubbish is the keywords chosen: global climate change. I did a quick search of the WoS for this phrase and found 683 articles published since 2004, similar to the number obtained by Dr. Schulte. However, if I limit my search to the journals in which the vast majority of papers on the physics of climate are published …

  1. Journal of Geophysical Research
  2. Geophysical Research Letters
  3. Journal of Climate
  4. Journal of the Atmospheric Sciences

… I only get 18 articles. That’s surprising, because these journals have published nearly 8,000 papers over this time period, and a significant fraction are on climate change.

It turns out that most papers on global climate change do not actually use this term in the title or abstract. For example, my paper above was not captured because it does not use this term. And even if it were captured, I’m quite certain that it would be categorized as agnostic, even though an expert reading shows it strongly supports the consensus.

Bottom line: Shulte’s analysis is just another harpoon fired wildly at the white whale of consensus. Anyone who’s read Moby Dick knows how how that hunt ends (spoiler alert: the whale wins).

As I have argued in the past, if one wants to know what the peer reviewed literature says, read the reports of the IPCC. Those reports are written by experts, peer-reviewed by experts, and contain an expert synthesis of what the scientific community knows about climate change and how confidently we know it.

Update [2007-9-7 15:18:41 by Andrew Dessler]:After reading Oreskes’ reply to the Schulte work here, I think she makes a good defense and answers many of the critical comments I made. I suggest everyone interested in this question read this.