cc: nleary@earth.usgcrp.gov, barrie.pittock@dar.csiro.au, lindam@ucar.edu, roger.jones@dar.csiro.au, m.hulme@uea.ac.uk, djgriggs@meto.gov.uk, meehl@ncar.ucar.edu, peter.whetton@dar.csiro.au, tkarl@ncdc.noaa.gov, m.manning@niwa.cri.nz date: Sun, 22 Oct 2000 08:54:04 -0700 (PDT) from: Stephen H Schneider subject: Re: Table 3-10: a third version and some other considerations to: Timothy Carter Thanks again TIm for a very fast and highly responsive third choice. I like it alot. Of course, some will say we "went beyond" WG1, but that is actually not true if we look in depth. The stars are based on the science--what is in the chapters in WG1 and the open literature. We simply translated it into the language necessary for most decision analytic methods to function--ergo, to do vulnerability analysis. Moreover, there are quite a few climate scientists in WG2--including several who were WG1 LAs last time or this time--and there is not extrapolation of likelihood here relative to that than can be inferred in WG1 text. We simply put the info in a form that is needed for our Working Group's charge from the IPCC--anyone who is unsure of this should consult the Policy Related Questions we must address in Synthesis, which would be very much tougher to do if this table were not there to serve as a handy summary and guide.Of course a few will polemicize, but many delegates will also demand just such a table--so lets go for what we think is right and let the Plenary chips fall where they may when it happens. As long as everybody is scientifically comfortable that we did our traslation consistently with the literature and the WG 1 text, then the fact that some scientists in WG 1 prefer to dwell in their "type 1 error paradigm" (see my attachment from last e-mail) and WG 2 in our type 2 is perfectly fine. We'll just have to do a good job of explaining that the difference is paradigmatic, not scientific. Thanks again for the umpteenth time Tim! CHeers, Steve On Sun, 22 Oct 2000, Timothy Carter wrote: > Dear extremes colleagues, > > Following Steve Schneider's rapid response yesterday, I have prepared a > third version of Table 3-10 that is even closer to the original. In this > case, I have merged old columns 1 and 2 (as in the WG I version and as Neil > suggested earlier), and have modified the footnotes and caption, but > retained all other features. > > In particular, note the part of the caption I have highlighted in bold (for > your consideration), which attempts to explain why WG II has a quantitative > scale and WG I does not (though the true reasons would require an academic > thesis!). I am repeating the key sentence below and then note a possible > alternative: > > In current version 3: > > "Entries in columns two and three are expressed in quantitative terms to > assist policy interpretation, based on judgements by WG I and WG II authors." > > Another way of presenting these reasons might be to stress that this type > of assessment is required by impact assessors in order to consider > vulnerability in terms of risks of impact. This way, we would keep the > reasoning "scientific", but from a WG II perspective, and would not need > to appeal to "policy relevance". > > In view of the three versions you have been sent, in deciding on what > course to take, I think we now have to consider seriously several points: > > (1) If we adopt version 2 (WG I-type descriptions), we are consistent with > WG I, though the table contains additional information invented by us, > using the same type of descriptive language as in the WG I Table. Here we > would (a) still have to defend our subjective interpretation of WG I > material for the categories not covered by WG I, (b) be inconsistent with > chapters in WG II (unless they are able to change their wording, as Leonard > Nurse indicated for small islands in his recent mail), and (c) go against > the principle of providing "policy useful" information that Steve expressed > (and see my alternatives above). > > (2) If we adopt versions 3 or 1 (likelihoods), we are taking a different > course from WG I, and we are open to the charge of being inconsistent with > WG I and providing additional information that does not appear (at least > prominently) in WG I. We could be accused of over-representing our > understanding and there will doubtless be an undercurrent of suspicion that > WG II authors are not qualified to make such judgements on climatological > matters. Are we prepared to defend ourselves against such (potential) charges? > > (3) If we use something like version 3 or version 1 (including > likelihoods), we must consider how this should be reported in the SPM. This > is what governments will comment on this time around (some, but I doubt > many, may refer to the revised chapters). We can try to revise the > simplified Table we have in the current draft SPM. Alternatively, we could > finesse the problem of consistency by NOT including a table at all in the > SPM, but rather use some appropriate (weasel?) wording to describe the > changes in extremes expected, which represents a compromise between the WG > I table and Table 3-10. Whatever the outcome, I need to draft some wording > anyway that is consistent with the final Table 3-10 for the Chapter 3 > executive summary. > > Best regards, > > Tim > ------ Stephen H. Schneider Dept. of Biological Sciences Stanford University Stanford, CA 94305-5020 U.S.A. Tel: (650)725-9978 Fax: (650)725-4387 shs@leland.stanford.edu