date: Mon, 10 Nov 1997 15:34:56 -0700 (MST) from: Tom Wigley subject: Re: MMIA (NSF) proposal to: Mike Hulme Dear Mike, I appreciate your concerns regarding putting SUL (or the ECHAM3 equivalent) into SCENGEN. The reason I suggested it is in order to make this information more readily available to users -- the only alternative is through LINK. The demand for this is very large. David Viner can provide proof of this; but, also, I have used HadCM2 SUL in a series of state-by-state information sheets published by EPA. These are being distributed by the tens of thousands, so the potential demand for HadCM2 SUL results is much larger than currently. If we can supply this through SCENGEN, this would be a big boost for SCENGEN. Of course, one would have to add some additional information (caveats) for this specific model -- and put in a "lock" so it could not be used in conjunction with other models. Note that scaling by delta-T is still okay. The problem of realism and applicability is one that is common to all model data sets in SCENGEN (albeit in a slightly different way), so I don't think this is a serious issue. Carrying your argument to an extreme would mean abandoning SCENGEN completely! The bottom line of this is that I STRONGLY urge you to put HadCM2 SUL into SCENGEN. In my view, the benefits FAR outweigh any negative aspects. This is not just an "off the cuff" suggestion, but one that I have thought through and discussed with a number of people. John Mitchell should know about the ESO2 scenarios, although I may not have updated him on them recently. I've not talked to Dave Roberts directly. I assumed John was filling him in, but perhaps not. What we have done so far is develop time and spatially dependent ESO2 scenarios consistent with WRE550 (and others consistent with IS92a). It is a non-trivial task, but we are well aware of all the problems, pitfalls and complexities. We are, of course, also in touch with IIASA on this. Hugh Pitcher's involvement ensures consistency with what IPCC is doing. I have, in addition, also looked at other pathways to 550 ppmv accounting for protocols -- see Nov. 20 issue of _Nature_. I'll send you a report on the ESO2 work if you are interested. Cheers, Tom P.S. I don't agree with your statement that aerosols will "become of vanishing relevance" in Europe -- certainly not over the next 50 years, which is what people are most concerned about now. On Wed, 5 Nov 1997, Mike Hulme wrote: > Tom, > > SCENGEN currently has, in aging order of vintage: > > HADCM2 > ECHAM3-TR > CSIRO-Mk2 > UKTR > CSIRO > CCCEQ > BMRC > LLNL > ECHAM1-TR > UKHI > OSU > UKLO > GFDL > GISSEQ > > We have put Mike Schlesinger's aerosol patterns into SCENGEN using our > three regions. Mechanically, MAGICC/SCENGEN now works and allows variable > regional aerosol patterns to be generated. The questions however concern > additivity of T and P patterns and also S:N ratios re. precip. > > At present we only have the first realisation from HADCM2 (i.e., not the > ensemble mean) in SCENGEN and only the pattern from the 1% pa forcing rather > then the 0.5% pa forcing. These are matters we could change of course. > > About SUL runs from HADCM2 and ECHAM3 I'm not so sure. Yes, mechanically > we could put in the HADCM2 SUL patterns and scale them by a global dT, but > I don't think this is desirable. What would be a better solution would be > to difference the GHG and SUL patterns from HADCM2 and then treat this > difference > as the aerosol pattern and treat these fields like we have treated Mike > Schlesinger's results. The problem with this is that the aerosol patterns > (if one > could deconvolve them from the GHG patterns) in HADCM2 relate to one S > emissions > scenario with a particular geography. Unlike what Schlesinger did, for > some regions > the aerosol forcing in HADCM2 reduces to very small levels by 2100 > (therefore there > is only a very weak and probably meaningless aerosol signal remaining). For > precip. > I have showed this happening over Europe. > > Bottom line is that I am not convinced that we should use the HADCM2 and > ECHAM3 SUL > runs in SCENGEN. Much better to use Schlesinger's results and add them to the > GHG simulations from the good GCMs. > > About new O/AGCM runs, if I understand you correctly Hadley Centre (as well > as CSIRO and > NCAR) will be using new regionalised SO2 emissions scenarios in a > stablisation experiment > early next year. Does Dave Roberts (S transport modeller at Hadley) know > about this? > Last week he seemed not to. I guess I should talk to John Mitchell instead. > > The new SRES scenarios (4 or 6) are all supposed to be 'BAU' scenarios > (although the > distinction between BAU and policy intervention begins to blur) since that is > what IPCC have asked them to do. These different futures _do_ need to be > interpreted in > climate terms and the versatility of MAGICC/SCENGEN is excellent for this > without > calling upon more GCM runs. The FCCC only deals with CO2 stabilisation of > course and > says nothing about S emissions. There are a number of possible S futures > compatible with > FCCC and probably none of them will look like IS92a. And in most S > futures, the > relative importance of S forcing vs. GHG forcing will diminish over time > (for most > if not all large regions). I still think that given the problems with > modelling indirect > aerosol effects, and the vanishing relevance of S forcing vs. GHG (cf. > halocarbons and > GHG forcing), that the majority of impacts work should relate to > GHG-induced climate > scenarios. Impacts work with aerosol-included scenarios should be an > additional > sensitivity analysis, but not a central one, more relevant for some regions > than > for others. > > For example over Europe and N.America, our (or anyone else's!) baseline > 1961-90 climate > has got a large aerosol signal embedded within it. Getting aerosol effects > right for > this region in scenario terms, means successfully being able to _remove_ > the aerosol signal > from the baseline climate, not _adding_ any aerosol signal. But clearly > what will > dominate future climate patterns for this region is GHG-induced change. > Aerosols > become of vanishing relevance. > > Mike > > At 12:20 04/11/97 -0700, you wrote: > >Dear Mike, > > > >Can you give me an update on SCENGEN please. > > > >(1) What GCMs are currently in the database? > > > >(2) Could you put HadCM2 SUL in? In other words, produce normalized > >changes for 2080-99 minus 1980-99 for the SUL run, and treat these like > >any other model. > > > >(3) Could you do (2), but use SUL (2080-99 minus 1980-99) minus CONTROL > >(2080-99 minus 1980-99) as a way of looking at the possible effects of > >drift in the model? > > > >(4) Have you composited Michael Schlesinger's runs into our three regions, > >North Atlantic, Asia, ROW? > > > >(5) Have you put these into SCENGEN yet? > > > >Cheers, > >Tom > ----------------------------------------------------------------------------- > Dr Mike Hulme tel: +44 1603 593162 > Climatic Research Unit fax: +44 1603 507784 > School of Environmental Sciences email: m.hulme@uea.ac.uk > University of East Anglia web site: http://www.cru.uea.ac.uk/~mikeh/ > Norwich NR4 7TJ > **************************************************************************** > ********************************************************** *Tom M.L. Wigley * *Senior Scientist * *National Center for Atmospheric Research * *P.O. Box 3000 * *Boulder, CO 80307-3000 * *USA * *Phone: 303-497-2690 * *Fax: 303-497-2699 * *E-mail: wigley@ucar.edu * **********************************************************