date: Wed May 16 17:06:01 2001
from: Tim Osborn <t.osbornatXYZxyz.ac.uk>
subject: Hadley model output and Tyndall projects
At the workshop I attended at NCAR last month, it was generally agreed by the modelling people there that it was necessary to have an AGCM with a resolution of about 1 degree in order to get the atmospheric circulation response to climate change correct. For the Tyndall project on storms/forestry/insurance, I raised this point and the implication that the high-resolution time slice runs with HadAM3 and the embedded regional runs with HadRM3 may give a *different* climate change signal for circulation and storms than does HadCM3. [Also HadCM3 probably has a poorer present-day circulation due to small errors in present-day climate (no flux adjustments), but that is a separate concern.] This is problematic because the assumption would be that HadAM3highres and HadRM3 would be "better" than the HadCM3 results, since the former have 1 degree or finer resolution.
I've checked this out with the Hadley Centre, and Richard Jones and Ruth McDonald have shown me some results indicating that the winter storminess change over the UK is different in HadCM3 and HadAM3highres.
This gives us something of a problem, given that our aim was to focus mainly on HadCM3 in the Tyndall "storms" project. If we continue with HadCM3 then are we (i) using results that we think are not the best available; and (ii) producing results that will be in conflict with the UKCIP02 scenarios which will be released in early 2002, prior to most of the output from the "storms" project.
It appears preferable to use HadAM3highres rather than HadCM3. We are constrained, however, by when the LINK project will get these data *and* by the release of these data by the Hadley Centre. David thought that he might have the HadAM3highres data by November, but that they might not be released till February. Nothing is certain though.
(1) What do you think about the choice of model?
(2) Would we be able to persuade the Hadley Centre to release the data as soon as the LINK project gets them, specifically for the "storms" project? Should we contact them and lobby them for early release?
You are, no doubt, keen for Tyndall-funded research to make use of best available information and data - so it's obviously appropriate for you to give us advice and also to do what you can to ensure access to the most appropriate data.