Before I get to that, here is a BBC rebuttal to climate skeptics.

The statement that water vapour is “98% of the greenhouse effect” is simply false. In fact, it does about 50% of the work; clouds add another 25%, with CO2 and the other greenhouse gases contributing the remaining quarter. Water vapour concentrations are increasing in response to rising temperatures, and there is evidence that this is adding to warming, for example in Europe. The fact that water vapour is a feedback is included in all climate models.

This is to counter the alleged skeptic argument that “Water vapour is the most important greenhouse gas, accounting for about 98% of all warming.”

What I have read is that (a) water vapor is 98 percent of all greenhouse gases, not that (b) it accounts for 98 percent of all warming. So it is not clear to me whether they mean to refute (a) or (b). In principle, water vapor could be 98 percent of all greenhouse gases and CO2 emissions could account for 100 percent of all the warming.

My view of the water vapor issue is that it makes models of global warming highly sensitive to assumptions about how water vapor responds to an exogenous rise in CO2. If you assume a positive response, then you will get a lot more global warming in response to emissions than if you assume that water vapor remains constant.

Bruce Charlton points me to this analysis of contributors to the latest IPCC report.

Of the 51 UK contributors to the report, there were 5 economists, 3 epidemiologists, 5 who were either zoologists, entomologists, or biologists. 5 worked in civil engineering or risk management / insurance. 7 had specialisms in physical geography (we gave the benefit of the doubt to some academics whose profiles weren’t clear about whether they are physical or human geographers). And just 10 have specialisms in geophysics, climate science or modelling, or hydrology. But there were 15 who could only be described as social scientists. If we take the view that economics is a social science, that makes 20 social scientists.

…Of the 70 US contributors, there were 7 economists, 13 social scientists, 3 epidemiologists, 10 biologists/ecologists, 5 engineers, 2 modellers/statisticians, 1 full-time activist (and 1 part time), 5 were in public health and policy, and 4 were unknowns. 17 worked in earth/atmospheric sciences.

I’m thinking that my scientific qualifications to discuss global warming are about on par with those of the median contributor to the IPCC report. To be fair, there could be hundreds or even thousands of climate scientists who did not contribute to the IPCC report but who support its findings.

Actually, I don’t need a lot of scientists to convince me to buy into the IPCC report. It might only take just one. All I ask is that one climate scientist give me a list of the empirical findings that are persuasive that CO2 emissions are the cause of global warming. So far, all I am getting is the broad trend of higher CO2, the rise in temperatures since 1980, and the climate models that link the two by design.

UPDATE: A commenter finally points me to something helpful, namely

UPDATE 2: See below (more citations added).this history of climate science.

A final nail in the skeptics’ coffin came in 2005, when a team compared computer calculations with long-term measurements of temperatures in the world’s ocean basins (it was not in the air but the massive oceans, after all, that most of any heat added soon wound up). In each separate ocean basin, they showed a close match between observations of rising temperatures at particular depths, and calculations of where the greenhouse effect should appear. This was telling evidence that the computer models were on the right track. Nothing but greenhouse gases could produce the observed ocean warming — and other changes that were now showing up in many parts of the world, as predicted.

AHA! Something that sounds like evidence for a greenhouse effect, as opposed to solar activity or somesuch.

Another commenter recommends this chapter of the IPCC report. I wish the writing were punchier, as in the report described below.

this paper, recommended by another commenter, is an alternative reading of the state of climate science today, from the Fraser Institute, a Canadian think tank that leans libertarian. This excerpt from the alternative report captures my concern with much of the evidence as presented in the IPCC chapter cited above:

The term “attribution” means consistency with a climate model-generated scenario, rather than formal proof of causality. The same data could be consistent with contradictory hypotheses, including large or small greenhouse warming.

Typically, the IPCC chapter will show a graph comparing the actual results to model simulations with and without a CO2 effect, with the latter fitting the data better. That strikes me as closer to proof by assumption than proof by empirical evidence.

Here is an analogy: Suppose, hypothetically, that ability is what affects earnings, and once you control for ability education does nothing. But your model only includes education, which is also correlated with ability, which you don’t include. You can say that model simulations that incorporate education work better than simulations that don’t. Ergo, education matters. But that is a relationship you imposed on the data, by not including ability.

The IPCC attribution methodology only works if the models have included and properly specified all of the factors affecting temperature that have a positive trend in the last thirty years. It is unlikely that the models are so complete.

Below are some more excerpts from the contrarian report. I feel badly not giving equal time to the IPCC chapter, but as I said, the writing there is more turgid. Substantively, the IPCC chapter might be as good or better than the contrarian report. It would be interesting to see the issues debated.

studies have shown that the spatial pattern of warming trends over land correlate strongly with the distribution of industrial activity, even though such a correlation is not predicted by climate models (e.g., de Laat and Maurellis 2004, 2006).

…There is no significant warming in the tropical troposphere, which accounts for half the world’s lower atmosphere. This is where models that assume a strong influence of greenhouse gases forecast some of the most rapid warming should occur.

…The Third Assessment Report drew attention to the declining Diurnal Temperature Range (DTR) as evidence of global warming (Working Group 1 Summary for Policymakers, page 1). The decline in the DTR has now ceased, and appears to be growing in most places.

…There are differences in linear trends of tropospheric temperatures between the high latitudes of the Northern and Southern Hemispheres that are not consistent with computer model projections.

…At the global scale, some broad predictions made 30 years ago about the possible response to increased CO2 concentration in the atmosphere, namely increased average tropospheric temperature, decreased average stratospheric temperature and a more rapid hydrological cycle, are consistent with data that have emerged since then

…Models tuned to “perfectly” reproduce an observed mean climate state have nonetheless shown only a weak ability to predict subsequent climatic conditions. It is not possible to say which, if any, of today’s climate models are reliable for climate prediction and forecasting.

…Quantitatively, individual climate models are typically unable to reproduce the observed mean surface temperature to better than +/- 3 kelvin, with worse performance near the poles. They are also unable to reproduce the onset of ice ages. The margin of present-day error is similar to the size of the projected global warming trend over a century.