date: Thu Sep 21 16:45:11 2006
from: Tim Osborn <t.osbornatXYZxyz.ac.uk>
subject: Re: use of CLIMGEN for IPCC: before we meet
to: "Rachel Warren" <R.WarrenatXYZxyz.ac.uk>
Rachel - we can talk some more when you come to see me, but Tom Wigley is not an author at
any level, as far as I know, on any chapter in IPCC WG1 for the AR4. So his approval or
disapproval may not be relevant (other than for scientific reasons of course!). In the way
that you are using CLIMGEN, I think that there really only two things that need to be
defended: (1) the GCM patterns that have been used; and (2) the pattern-scaling concept
itself. The reason why I think these are the only two issues is that you could probably do
the upscaling without using CLIMGEN at all, just by using these two things.
Both issues are defensible in general. With specific reference to IPCC WG1, however, you
should look to see what the draft says about regional climate change, and specifically
whether pattern-scaling is criticised or supported or even mentioned. CLIMGEN is not
reviewed or mentioned as far as I know, but provided the 2 concepts/issues mentioned above,
then this is fine. The GCM runs from which the patterns are derived are not all very new;
some of these runs are definitely still being used in AR4 WG1, but we can check if this
applies to them all. All are at least TAR or later.
At 16:32 21/09/2006, you wrote:
This is just to let you both know that I am using CLIMGEN to backwards-downscale (i.e.
upscale) regional to global temperature for a key table in an IPCC WG2 chapter on
climate impacts for ecosystems at different degrees of temperature rise using the range
of GCMs in CLIMGEN. The method for uncertainty analysis for the upscaling was agreed
with Martin Parry at the recent WG2 meeting I went to last week. I've already done a lot
of calculations but I am going to do more following last week's decision, to include
more uncertainty analysis for more of the entries in the table than I had previously
I am wondering, however, whether WGI, given their recent tendency to object to anything
that isn't in the WG1 report, may object to the use of CLIMGEN in this way. WG1 are
currently opposing the use of anything in WG2/3 on climate science that isn't covered in
their report, even if it's perfectly valid literature that they don't cover.
However, CLIMGEN has the same methods as SCENGEN which was used in the TAR. I am
wondering whether to talk to Tom Wigley about it - but given the SCENGEN/CLIMGEN tension
that I think may have existed, this may make matters worse. Is CLIMGEN reviewed in IPCC
WGI or not? Does Tom Wigley approve of CLIMGEN? If WG1 object to the use of CLIMGEN,
because it is not SCENGEN or Tom Wigley's latest MAGICC-SCENGEN combination, they may
argue to throw out the table, and I would lose all of my work on this subject and the
chapter would be much the poorer - the review editor thinks that my table will get more
attention than anything else in the chapter and some of it may end up in the SPM.
Hence I cannot emphasise the importance of making sure that WG1 cannot complain about my
using CLIMGEN to upscale! I'll be talking to you tomorrow to make sure that I fully
understand exactly which GCM datasets are used in CLIMGEN so that I and the CLA and
Martin Parry can defend how I have used it, and how it works, if the topic comes up in
discussion at any time. Tim, I need to check whether you are using the same updated GCM
patterns as the new version of SCENGEN (version 2.4 of 2003).
The best thing would be to get Tom Wigley's approval of my using it, prior to any of the
IPCC meetings ...
Dr Rachel Warren
Senior Research Fellow
University of East Anglia
Norwich NR4 7TJ
Telephone 01603 593912
Fax 01603 593901