Over at Dot Earth, Andy Revkin has a smart story about the growing pressure to change how the Intergovernmental Panel on Climate Change (IPCC) operates, especially after the recent scandal over glaciers. (In the IPCC's 2007 report, there was a line about how glaciers in the Himalayas could vanish by 2035; it turns out that line had zero basis in peer-reviewed science, yet still got past reviewers.) This is a worthy topic: The IPCC has done terrific work over the years, and its reports are considered the best summaries of the state of knowledge about climate change, but they're not perfect and could stand to be improved.

For one, the safeguards obviously need to be strengthened. That glacier line should've never made it through the review process. Same thing, it seems, goes for this bit about rainforests. Better safeguards are especially crucial for the sections on potential impacts of global warming. This is an area of keen interest, but it also involves some of the murkiest research and can span multiple disciplines—sometimes even social science. (By contrast, the IPCC's Working Group 1 report, which deals with the physical basis for climate change—how we know that greenhouse gases are warming the planet, etc.—is more straightforward and gets the heaviest scrutiny from the physical-science community.)

Another criticism of the IPCC, raised by UC San Diego's David Victor, is that the panel is overly cautious and doesn't deal with outliers very well. Case in point: The IPCC's 2007 report projected that, at most, sea levels would rise 26 to 59 centimeters by 2100. Except the problem was the IPCC explicitly left out the full range of potential effects from melting ice sheets in Antarctica and Greenland, because that ice-sheet behavior can be difficult to model. Leaving out those sorts of uncertainties may make for a stronger "consensus," but it can also give a misleading picture of the risks we're actually facing. Since the 2007 report, plenty of research has suggested (e.g., 1, 2, 3) that the IPCC seriously low-balled its sea-level prediction.

Beyond that, there's another critique of the IPCC that Revkin doesn't mention, but which seems important. The IPCC moves very slowly, and it's usually a few steps behind the leading edge of climate research. The panel stops taking scientific input a year or two before the report comes out, and the lag often shows: Apart from sea levels and ice sheets, the 2007 IPCC reports badly underestimated the rate at which greenhouse gases were rising, because it didn't incorporate recent rapid growth in countries like China and India. So the emissions scenarios were obsolete almost as soon as they were published.

Looking ahead, there's another dilemma: The next IPCC reports won't be out until 2013 and 2014, even though a number of major policy decisions on greenhouse gases will likely need to be made before then. Now, it's understandable why the IPCC takes so long to assemble its big reports, since it needs to do a careful review of all the relevant science, survey a bunch of complex and often heated debates, and put everything in context. Plus, the scientists working on the IPCC are volunteering their time, and its a lot of work and a diversion from their own research. Still, the sluggish pace is out of sync with both the needs of policymakers and the rapidly advancing state of climate science.

(Flickr photo credit: kimberlyfaye)