The proposed subproject relates to work our Detection Group is undertaking regarding natural variability in CGCMs. We had decided to look at three CGCM runs (MPI, Hadley Center and GFDL) and compare the natural variability in their three-dimensional temperature fields. This has obvious implications for GHG signal detection. At the February 1997 AMS meeting in Long Beach we were told that CMIP has a far more extensive collection of models than we had originally planned to analyze. We felt that extending the analysis to this broader set would be highly desirable.
Given the above, I request, on behalf of the Group, that we be allowed to participate with CMIP in the analysis of the 3D temperature variability in the CGCM runs that CMIP has acquired. Our original plan was to do a fairly straightforward analysis of variance (projection onto a common EOF basis set, etc.) but we would be glad to help in any reasonable way we can. We do have limited resources so we need to stick to a sharply focused study that will carry over directly to our detection algorithms.
The analysis would be done by some combination of Santer, Hegerl and Barnett -- the mix depending on discussion with CMIP and our internal progress on the rest of the detection study.