In response to an inquiry to Scott Saleska, Dan Kirk-Davidoff, a prominent expert in the field, has sent the following suggestions:
Soden and Held 2006, An Assessment of Climate Feedbacks in Coupled Ocean—Atmosphere Models, J. Climate 19:3354, DOI: 10.1175/JCLI3799.1 reviews the relative role of various feedbacks in the IPCC AR4 runs.
Held and Soden, “Water Vapor Feedback and Global Warming” Annu. Rev. Energy Environ. 2000. 25:441—75 is a general review of the problem, including discussions of the different dynamics and physics involved in water vapor response to temperature changes in the boundary layer and free troposphere, as well as a discussion of the feedback on a global level. If you haven’t read it, I think you will find it very interesting and though-provoking.
Held and Soden 2000 is online here . Soden and Held was discussed last year in passing. What surprised me in the discussion was that Held’s expectation was that cloud feedback in the GCMs would be positive in some models and negative in others. He expressed surprise in a realclimate comment that it turned out that it was strongly positive in all models. [link] I must say that his surprise intrigued me.
Dan has sent in the following additional reference on convection and cloud modeling:
The problem has been a lack of detailed data. If you could accurately and precisely measure the temperature, water vapor content and cloud radiative forcing over the whole atmosphere at high space and time resolutions, you could go a long way towards testing and correcting the cloud and convection parameterization. Those data are finally becoming available; between the COSMIC GPS satellite network, which can produce accurate temperature profiles through most of the depth of the atmosphere 2500 times/daily, the various spectrally-resolved IR observing instruments (AIRS, IASI, eventually CrIS), and cloud-observing satellites like CloudSat, we are amassing data that will provide a serious capability to falsify model statistics at a wide range of time and space scales.
This is not to say that people haven’t been working on the problem with the data at hand. The ARM observation sites were designed expressly for this purpose- here’s a brief summary white paper on the topic:
The entire issue of water vapor feedback is very important in forming a view on whether increased CO2 content is a big or little problem. Both water vapor feedback and clouds have been identified as critical issues since at least as early as the Charney Report in 1979 and I suspect that much of the difficulty in making much (any) progress in reducing the uncertainties in climate models relates to the difficulty in reducing uncertainty in this area. If one were managing climate model engineering in the same way as space station engineering, I would really focus on this area. Given the sensitivity of GCM results to varying assumptions and parameterizations in this area, I would have liked it if AR4 had devoted entire chapters to each major issue in clouds and water vapor feedback, instead of a few cursory pages.