cc: Jerry Meehl <meehl@meeker.ucar.edu>, Timothy Carter <tim.carter@vyh.fi>, maureen.joseph@eci.ox.ac.uk, lindam@ucar.edu, m.hulme@uea.ac.uk, peter.whetton@dar.csiro.au, giorgi@ictp.trieste.it, tkarl@ncdc.noaa.gov, cubasch@dkrz.de, ckfolland@meto.gov.uk, hewitson@egs.uct.ac.za, "Stouffer, Ron" <rjs@gfdl.gov>
date: Wed, 23 Aug 2000 00:02:33 -0700 (PDT)
from: Stephen H Schneider <shs@stanford.edu>
subject: Re: THC collapse
to: Thomas Stocker <stocker@climate.unibe.ch>

Hello all. I appreciate the improvement in the table from WG 1,
particularly the inclusion of symmetrical confidence levels--but please
get rid of the ridiculous "inconclusive" for the .34 to .66 subjective
probability range. It will convey a completely differnt meaning to lay
persons--read decisionmakers--since that probability range represents
medium levels of confidence, not rare events. A phrase like "quite
possible" is closer to popular lexicon, but inconclusive applies as well
to very likely or very unlikely events and is undoubtedly going to be
misinterpreted on the outside. I also appreciate the addition of
increasing huricane intensities with warming moving out of the catch all
less than .66 category it was in the SOD. 
  I do have some concerns with the THC issue as dealt with here--echoing
the comments of Tim Carter and Thomas Stocker.  I fully agree that the
likelihood of a complete collapse in the THC by 2100 is very remote, but
to leave it at that is very misleading to policymakers given than there is
both empirical and modeling evidence that such events can be triggered by
phenomena in one century, but the occurrence of the event may be delayed
a century or two more. Given also that the likelihood of a collapse
depends on several uncertain parameters--CO2 stabilization level, CO2
buildup rate, climate sensitivity, hydrological sensitivity and initial
THC overturning rates, it is inconceivable to me that we could be 99% sure
of anything--implied by the "exceptionally unlikely" label--given the
plausibility of an unhappy combo of climate sensitivity, slower than
current A/OGCMs initial THC strength and more rapid CO2 increase
scenarios. Also, if 21st century actions could trigger 22nd century
irreversible consequences, it would be irresponsible of us to not mention
this possibility in a footnote at least, and not to simply let the matter
rest with a very low likelihood of a collapse wholly within the 21st
century.  So my view is to add a footnote to this effect and be sure to
convey the many paramenters that are uncertain which determine the
likelihood of this event.
  Thanks again for the good work on this improtant table. Cheers, Steve


On Wed, 23 Aug 2000, Thomas Stocker wrote:

> DEar Jerry, Tim and Ron et al
> 
> I agree that an abrupt collapse - abrupt meaning within less than a decade, say
> - has not been simulated by any climate model (3D and intermediate complexity)
> in response to increasing CO2. Some models do show for longer integrations a
> complete collapse that occurs within about 100-150 years. If you put that into
> context of the apparent stability of THC during the last 10,000 years or so,
> this is pretty "abrupt".
> 
> Following up on the discussion regarding THC collapse, I think the statement Ron
> apparently added to Ch9 needs to be made more specific. In order to keep Ch7 and
> Ch9 consistent, I propose to Ron the following revision:
> 
> "It seems that the likelihood of a collapse of the THC by year 2100 is less
> than previously thought in the SAR based on the AOGCM results to date."
> 
> There is really no model basis to extend this statement beyond 2100 as evidenced
> by the figures that we show in TAR. There are many models that now run up to
> 2060, some up to 2100, but very few longer.
> 
> Also I should add for your information, that we add to Ch7 a sentence:
> 
> "Models with reduced THC appear to be more susceptible for a
> shutdown."
> 
> Models indicate that the THC becomes more susceptible to collapse if previously
> reduced (GFDL results by Tziperman, Science 97 and JPO 99). This is important as
> "collapse unlikely by 2100" should not tempt people to conclude that THC
> collapse is hence not an issue. The contrary is true: reduction means
> destabilisation.
>  
> Best regards
> 
> thomas
> -- 
> ------------------------------------------------------------------
> Thomas Stocker                        
> Climate and Environmental Physics         stocker@climate.unibe.ch
> Physics Institute, University of Bern     phone:  +41 31 631 44 64
> Sidlerstrasse 5                      NEW    fax:  +41 31 631 87 42
> 3012 Bern, Switzerland        http://www.climate.unibe.ch/~stocker
> ------------------------------------------------------------------
> 

------
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

