date: Mon May 16 08:21:51 2005 from: Phil Jones subject: Re: HadRM3 to: "Stephen Dorling" , "David Viner" , "Mike Hulme" Steve, Back from China and the IPCC meeting. Response a little brief as I've loads to go through before going to Exeter tomorrow. The first pdf is a paper assessing how good HadRM3 is. Over the UK and similar latitude regions of Europe it is pretty good. HadRM3 is too warm in summer in southern Europe (soil moisture dries out) and too cold in far northern Europe. The very warm temps in southern Europe appear to begin to occur in southern Britain in the future perturbed run, because the soil moisture dries out. The second paper takes this a step further looking at HadCM3 and comparing it to CET. Bottom line is that HadCM3 has the same very warm Tx values in the future as HadRM3. Whether they are real (possible) is the question, but the 2003 summer shows how the soil moisture can dry out and all the excess heat goes into sensible heat. I reckon they are likely in very dry years, particularly after 2040 (see the Brabson et al. paper). Hope they help and this isn't too late. Cheers Phil At 14:30 11/05/2005, Stephen Dorling wrote: Dave, Mike, Phil I'm trying to write a review report for DEFRA on climate change and air quality. One of my high profile co-authors is highly critical of HadRM3 in terms of its tendency, he says, to overestimate Tmax in the control climate, especially the extreme summer temperatures. Essentially, he was once provided, by Tom Holt, with daily gridpoint Tmax which, on analysis, showed occasional instances of 41C for central England in the control. On the basis of this, and other evidence such as disappointing results when HadRM3 was used to simulate climate change in India, he refuses to use HadRM3 claiming it is "much too warm" generally, prefering instead to stick with HadCM3. He wishes to make a strong point about this in the report. The main issue is that Tmax has a strong influence on surface ozone and natural emissions of volatile hydrocarbons. I've looked back through the UKCIP02 Scientific Report and also at the draft paper by Dave Rowell et al arising from the PRUDENCE RCM intercomparison. It is acknowledged that absolute simulations of Tmax are too high in the upper 1 percentile and the stress seems to be on working with scenario-baseline as a fairly reliable indicator. I wondered if you had any comments on this? I may need to go to the trouble of comparing HadRM3 and HadCM3 daily Tmax timeseries to really nail this but I feel sure someone has already done this. As ever it is difficult to find the appropriate report/paper. Just to add insult to injury, a separate study has recently used a 25km HadRM3 simulation, within the PRECIS system, to further investigate air quality impacts. Unfortunately only four individual years are being analysed in the control and the future scenario to draw conclusions. My "colleague" is rejecting this as bad practise (which it probably is) in terms of significance testing and dealing with natural variability. Very grateful if you can help at all. There is potential embarrasment within DEFRA on this of course. Steve Prof. Phil Jones Climatic Research Unit Telephone +44 (0) 1603 592090 School of Environmental Sciences Fax +44 (0) 1603 507784 University of East Anglia Norwich Email p.jones@uea.ac.uk NR4 7TJ UK ----------------------------------------------------------------------------