date: Mon, 30 Nov 1998 15:31:48 +1100
from: Trevor McDougall <Trevor.McDougallatXYZxyzine.csiro.au>
subject: Re: Thermal expansion
to: Jonathan Gregory <jmgregoryatXYZxyzo.gov.uk>, t.osbornatXYZxyz.ac.uk
A few comments on the thermal expansion issues. First, Tim suggests
that we may have prejudged the answer when we reduced the Greens function
in the lowest layer of our model. Without claiming that that model is the
last word, this is best answered by imagining the result we would have
obtained if we had not reduced the Greens function in that way. Then the
thermal expansion would have been overestimated with a model that had been
tuned to today's ocean circulation. So we had not prejudged the issue:- to
get the best fit we might have had to increase the Green's function rather
than decrease it. Of course, in our present and continuing work we are
adressing the issue of the slowing THC in a more comprehensive fashion by
allowing the response functions of all layers to be a function of the
globally averaged sea surface temperature. In some layers we fully expect
the ventillation rate to increase while in others it will decrease. In the
meantime, we intend to be more circumspect in the discussion of these
issues in the next draft of the present Jackett et al manuscript.
Then there is the issue about the existing THC giving up heat to the
atmosphere and warming Europe. When the THC slows there will be less heat
delivered in this way to the atmosphere which would tend to increase the
thermal expansion. But, if there is less heat delivered to the atmosphere
in this latitude band, then there will also be correspondingly less heat
taken from the atmosphere at low latitudes, so this argument is not
The argument about the Deacon Cell is very unsatisfying to a physical
oceanographer. The Deacon Cell is a phenomenon that occurs when averaging
in Cartesian coordinates and appears to be a diabatic circulation.
However, the Deacon Cell, and in particular, its diabatic nature, is simply
an artifact of the averaging. As Doos & Webb, and McIntosh & McDougall
have explained, when doing the zonal averaging in density coordinates, the
Deacon Cell disappears. In fact Doos & Webb have defined the Deacon Cell
in just this fashion as the cell that appears under z-coordinate averaging
but not under density-coordinate averaging. The Deacon Cell corresponds to
the Ferrel cells in the atmosphere (see Karoly et al, Q. J. R. Meteorol.
Soc., 1997, 123, 519-526). The Deacon Cell is primarily wind-driven and
since it is the result of a three-dimensional circulation that is
adiabatic, we would not want to model it in a diffusive way.
It is this desire to
(a) not prejudge the signs of the answers (of SST increase and thermal
expansion trend), and
(b) have a closer resemblence between the physical balance of the model
and of reality
that has driven us to embark on the present project. At the end of the
day, any simplified model will have an element of tuning involved in it as
it strives to mimick a fully coupled three-dimensional model. As physical
oceanographers, we are very uncomfortable with a model that is knowingly
constructed as diffusive when we all agree that this is not the primary
balance in the ocean or in the coupled models.
It is our intuition that the slowing circulation will lead the a
smaller-than-otherwise thermal expansion and probably a
higher-than-otherwise SST trend. However, we are prepared to be surprised
by the fully coupled model results. The work we are doing will answer this
question, and we will not be prejudging the answer.
Given the interest this discussion has generated in the above email
group, it is our intention to write up some notes about the work we are
embarked upon and to circulate it among the group for comment.
With best wishes,
Trevor & David.
From: Dr Trevor J McDougall [address for couriers]
CSIRO Marine Research CSIRO Marine Research
GPO Box 1538 Castray Esplanade
Hobart, TAS 7001 Hobart, TAS 7000