The decade since the First IPCC Assessment Report (IPCC, 1990) has seen progressive evolution in sophistication of thinking about water vapour feedback. Concern about the role of upper-tropospheric humidity has stimulated much theoretical, model diagnostic and observational study. The period since the SAR has seen continued improvement in the analysis of observations of water vapour from sondes and satellite instrumentation. Theoretical understanding of the atmospheric hydrological cycle has also increased. As a result, observational tests of how well models represent the processes governing water vapour content have become more sophisticated and more meaningful. Since the SAR, appraisal of the confidence in simulated water vapour feedback has shifted from a diffuse concern about upper-tropospheric humidity to a more focused concern about the role of microphysical processes in the convection parametrizations, and particularly those affecting tropical deep convection. Further progress will almost certainly require abandoning the artificial diagnostic separation between water vapour and cloud feedbacks.
In the SAR, a crude distinction was made between the effect of "upper-tropospheric" and "lower-tropospheric" water vapour, and it was implied that lower-tropospheric water vapour feedback was a straightforward consequence of the Clausius-Clapeyron relation. It is now appreciated that it is only in the boundary layer that the control of water vapour by Clausius-Clapeyron can be regarded as straightforward, so that "lower-tropospheric" feedback is no less subtle than "upper-tropospheric". It is more meaningful to separate the problem instead into "boundary layer" and "free tropospheric" water vapour, with the former contributing little to the feedback.
The successes of the current models lend some confidence to their results. For a challenge to the current view of water vapour feedback to succeed, relevant processes would have to be incorporated into a GCM, and it would have to be shown that the resulting GCM accounted for observations at least as well as the current generation. A challenge that meets this test has not yet emerged. Therefore, the balance of evidence favours a positive clear-sky water vapour feedback of magnitude comparable to that found in simulations.
Other reports in this collection