date: Wed, 24 Jan 2001 12:01:26 -0700 from: Tom Wigley subject: Re: MAGICC tech. manual to: Mike Hulme Mike, see below.... Mike Hulme wrote: > > Tom, > > I have had UNDP (Bo Lim) asking me again about whether we can do any more > with the Technical Manual draft you prepared to accompany the scenario > Workbook. You will see the current draft is available at the following site: > > http://www.cru.uea.ac.uk/~mikeh/software/MAGICC_SCENGEN.htm > > It is still the draft, as of February 2000. Is this: > > a) something that could be published as it is *************** No -- too rough *************** > b) something that can easily be finished off and then published *************** Moderately easily (if I had the time), but see below *************** > c) something that now should not be published, given movement towards > MAGICC TAR Version 3 perhaps later this year. *************** This is the key issue. A lot of the material would survive, and I have already written up all the new stuff I did for the TAR. There may be some sticking points, however. What Sarah did for the TAR was calibrate a bunch of OAGCMs, so the user can select a model to emulate. I think she is planning a TAR version of MAGICC that will facilitate this. For general use (i.e., for UNDP types, or for pedagogical purposes) and in conjunction with SCENGEN I would NOT want to go this route. It is simply too messy and complicated. The code I produced for the TAR could probably be lifted straight into the UNDP version. Sarah modified this a lot for what she did for the TAR (although she has yet to explain to me just what she did). I am sure that *her* TAR version of MAGICC could not be slotted into the UNDP shell. It may seem less than ideal to have two new versions, but this may be necessary. One of these will be the one that gives Sarah's TAR results, which is up to her. The other would be a simpler (UNDP-ish) version, updating the present UNDP version only part way towards Sarah's TAR. The latter would give results consistent with the TAR, but it could not produce precise TAR numbers since it would not have the individual OAGCMs wired into it. I don't see this as a problem. Another issue related to the TAR is that the sea level bits would have to be changed. In my view, the TAR underestimates seal level rise, so one might have to tweak some new TAR-based algorithms a bit. Again, I don't see this as a problem given the audience. *************** > > UNDP have money to print and publish it if you wish (I know they are very > keen to do so since the technical details complement the Workbook > examples). By the way they tell me, the Workbook has been translated into > Spanish and French and they are considering another print run of 500 (so it > must be going down well). *************** Good. I don't have any money for this. However, the real issues are my time and just how we update things to be more TAR-like. Another issue here is that I was planning to try to get some money from NSF to produce a more educationally oriented version; i.e., directed to the college and university 'market' here in the USA (although NOT with a view to actually selling it). I will have to explain more about this later, but I think the two activities (USA education and UNDP et al.) can be mutually beneficial and coordinated. I don't see either of these activities clashing with what Sarah is doing, partly for reasons already stated and partly because Sarah is only working on the MAGICC part of MAGICC/SCENGEN. *************** *************** > > Mike