date: Thu, 01 Jul 2004 08:35:15 -0600 from: Tom Wigley subject: subtitle to: Sarah Raper By the way, I would expunge the alternative title. The fact that we can use MAGICC with a standard (and constant) sensitivity and get excellent results demonstrates that it is a very useful concept. Further, my work with volcanoes and PCM (and all the other forcing cases we have done with PCM that I have considered) shows that it works for different forcings and for forcings operating on very different timescales. Are you trying to get IPCC to reject MAGICC? This is what it looks like with that subtitle. I don't think we understand variable sensitivity, but this is a secondary and relatively minor problem. Constant sensitivity usually works well. If some models appear to have a changing sensitivity then so what -- is this just a model artifact, or is it something that may apply to the real world? Another issue is that, because sensitivity varies spatially, if the pattern of forcing varies in time then the global senstivity will appear to change in time -- unless one is careful in how the global sensitivity is extracted from the AOGCM results. I would urge caution in what you say. Tom.