Tuesday, December 13, 2011


From: Stefan Rahmstorf <regentage@gmx.de>
To: Eystein Jansen <eystein.jansen@geo.uib.no>
Subject: [Wg1-ar4-ch06] Ch6-Climate Sensitivity
Date: Fri, 01 Oct 2004 11:49:05 +0200
Reply-to: stefan@pik-potsdam.de
Cc: wg1-ar4-ch06@joss.ucar.edu

Hi co-authors,
here are some thoughts on what to say on climate sensitivity in our chapter - this is an
attempt to focus on the main, simple messages for policy makers. (I think we should try
retaining those important messages and not lose sight of them amidst all the details,
complexity and caveats.)
The main policy-relevant question could be phrased as follows: Does the past climate
history tell us how sensitive the climate system is to CO2?
I submit that the answers to this we get from different time periods are the following.
Deep Time:
Reconstructions are too uncertain (and boundary conditions too different, e.g. continents
in different places, different ocean circulation) to draw quantitative conclusions about
sensitivity to CO2, but there is clear evidence that times of high CO2 in Earth history
tend to be ice free (Royer et al. 2004). A second piece of evidence is the Late Paleocene
Thermal Maximum, which shows that the climate has responded by warming to a large carbon
release into the atmosphere. Just how large this carbon release was is not known, since
several origins of the carbon are possible, which have different isotope signature and
would thus imply different amounts. But the temperature response was large (6K), and if
anything this response would point to a high sensitivity.
Glacial-Interglacial Changes:
We have by now sufficiently good quantitative reconstructions of CO2 and other forcings as
well as temperatures in order to derive useful quantitative estimates of climate
sensitivity. LGM was the most recent time in history in which CO2 concentration differed
greatly from pre-industrial values, by as much as it does now. It is the closest test case
for response to CO2 changes that we have.
There are two basic methods to derive climate sensitivity:
(i) Based on data analysis - e.g. Lorius et al. 1991 (concluding sensitivity is 3-4 K).
This method has the caveat that this sensitivity applies to colder climate, which may
differ somewhat from that which applies in present climate as the strength of feedbacks is
expected to depend on the mean climate (e.g., stronger snow-albedo feedback in colder
(ii) Based on combining data and models - e.g. Schneider von Deimling et al. 2004. Does not
have the above caveat, but depends on models.
Lag of CO2 behind temperature does not imply a lack of CO2 effect on climate, since the lag
is small (centuries, not millennia).
Holocene, last millennium
Overall conclusions
Qualitatively, climate history is at least consistent with the accepted CO2 sensitivity.
There is no evidence for much lower or much higher CO2 sensitivity (note that CO2 is not
the only forcing). The more recent climate history (as far back as ice core data go) does
allow quantitative inferences. The results of these estimates all lie within the IPCC range
and provide strong support for this. Paleodata may even allow to reduce this range, since
at least one study argues that values above 4K are very likely inconsistent with the
reconstructed LGM climate: for high CO2 sensitivity, tropical cooling in the glacial should
have been larger.
_______________________________________________ Wg1-ar4-ch06 mailing list
Wg1-ar4-ch06@joss.ucar.edu http://www.joss.ucar.edu/mailman/listinfo/wg1-ar4-ch06

No comments:

Post a Comment