[MITgcm-devel] Fwd: TEOS-10 and MITgcm

Jean-Michel Campin jmc at ocean.mit.edu
Thu Feb 21 09:25:01 EST 2013 

Hi Martin,

When Trevor was here at the Southern Ocean meeting, I advised him to contact
you (and I also told him that, in the code, there was no conversion when 
computing surface fluxes).

I think it makes sense to save in a 2-D array the surface in-situ temperature
(or potential temp; but at the surface they are equal). Don't know
what is better when it's also needed at depth (shelfice pkg).

I have the impression that the model only need 1 type of 
conversion, from conservative temp to in-situ temp (or pot. temp);
the other way consersion might just be needed when one wants to 
compare with obs.

Cheers,
Jean-Michel

On Thu, Feb 21, 2013 at 10:11:51AM +0100, Martin Losch wrote:
> Hi there,
> 
> for your interest: I had a short email contact with Trevor McDougall about the TEOS-10. Apparently, in order to complete "our" implementation, something needs to be done before applying the surface bulk formulae (also for sea-ice). Shouldn't be hard to do: there needs to be function/subroutine that converts theta (for TEOS-10 theta is interpreted as convservative temperature) to potential temperature. That routine can be part of model/src/seawater.F
> Maybe the best to have a global field that is computed once per time step and used for all surface interaction?
> 
> Further I should think about other places were all of this matters, e.g. in the shelfice package.
> 
> Martin
> 
> Begin forwarded message:
> 
> > From: Stephen Griffies - NOAA Federal <stephen.griffies at noaa.gov>
> > Subject: Re: TEOS-10 and MITgcm
> > Date: February 21, 2013 5:04:24 AM GMT+01:00
> > To: Trevor McDougall <trevor.mcdougall at unsw.edu.au>
> > Cc: Martin Losch <Martin.Losch at awi.de>, "Paul.Barker at csiro.au" <Paul.Barker at csiro.au>, John Marshall <johncmarshall54 at gmail.com>
> > 
> > Martin,
> > 
> > Just to support Trevor's recommendation about converting CT to PT at surface (potential temp = insitu temp at surface).  A comparison to bulk formulae uncertainties is not quite relevant, since they are presumably random.  In contrast,  difference between CT and PT is systematic.  
> > 
> > Best,
> >  Steve 
> > 
> > 
> > 
> > 
> > On Wed, Feb 20, 2013 at 8:20 PM, Trevor McDougall <trevor.mcdougall at unsw.edu.au> wrote:
> > Hello Martin,
> > 
> >     Thanks for this info.  I am pushing to get TEOS-10 into all ocean
> > models, mainly by simply encouragement of the principals, such as you.  So
> > it is great that you have already got the guts of TEOS-10 into MITgcm.
> > 
> >     A little team of me, Simon Marsland, Steve Griffies, Paul Barker, Russ
> > Fiedler and I are inserting TEOS-10 into MOM, and we will test the
> > importance of the change in salinity (by pursuing the two-salinity
> > variable approach that is formally recommended in the TEOS-10 Manual) as
> > well as the change in temperature.
> > 
> >     When we have done this we will have better definitive advice for you
> > and MITgcm.  For example, I would hope that we will be able to tell you
> > that yes indeed, having just one salinity variable, namely Absolute
> > Salinity, is fine for say 150 years into the future.  So I wold say that
> > what you have done for salinity is sensible.
> > 
> >     But I am sure that we will be telling you that at p = 0 (se surface)
> > MITgcm should convert the temperature from the model prognostic
> > temperature, Conservative Temperature, to potential temperature, so that
> > this can be used for air-sea flux bulk formulae.  This would be the main
> > thing I would advise to you right now.  If you do this, then I would be
> > happy that this TEOS-10 implementation in MITgcm has enough of my
> > "blessing" that it could be regarded as the default option of the MITgcm.
> > The GSW code to do this is GSW_pt_from_CT(SA,CT).  Conceptually, it is
> > easiest to think of the conversion from Conservative Temperature to
> > potential temperature as occurring only in the air-sea interaction module
> > of the model.
> > 
> >     Fabian Roquet of the University of Stockholm is planning to implement
> > TEOS-10 into NEMO in the way you have, but with the conversion to
> > potential temperature at the sea surface, as I suggest to you, above.
> > 
> >     As a summary of the expected changes under TEOS-10 compared with
> > EOS-80, (1) the horizontal density gradient (thermal wind) is improved by
> > 4% in the Southern Ocean and by 10% in the North Pacific (in both case,
> > for deeper than 1000m).  (2) the SST will be improved by 0.25C in the
> > equatorial Pacific (and up to 1C in some small locations that are very
> > fresh).
> > 
> >     Trevor
> > 
> > 
> > 
> > Trevor J McDougall
> > Scientia Professor of Ocean Physics
> > School of Mathematics and Statistics
> > University of New South Wales
> > NSW 2052, Australia
> > 
> > email: Trevor.McDougall at unsw.edu.au
> > tel: +61 2 9385 3498
> > fax: +61 2 9385 7123
> > mob: +61 407 518 183
> > 
> > 
> > 
> > 
> > 
> > 
> > On 20/02/13 7:20 PM, "Martin Losch" <Martin.Losch at awi.de> wrote:
> > 
> > >Hi Trevor,
> > >
> > >I added TEOS-10 to the MITgcm as an alternative to the other EOS's that
> > >we already have. I am not absolutely positive, but I think I started from
> > >the FORTRAN version, at least that's what it says in the head of the
> > >subroutine (o:, please have a look at
> > ><http://mitgcm.org/viewvc/MITgcm/MITgcm/model/src/find_rho.F?revision=1.41
> > >&view=markup> starting around line 720, if you are interested in the
> > >details.
> > >
> > >I am afraid, that I did things as sloppily as usual, and implementing the
> > >TEOS-10 was more of a fun exercise: I assume that once TEOS-10 is used
> > >the temperature and salinity variables of the model (THETA and SALT) are
> > >(interpreted as) conservative temperature and absolute salinity. Other
> > >than that, nothing changes, i.e. there is no conversion from conservative
> > >to potential temperature for the air-sea-interaction. The underlying
> > >assumption is that, the uncertainties in the bulk formulae are larger
> > >than any differences between the different hydrographic variables. Maybe
> > >you will disagree and I'll have to go back and read the TEOS-10
> > >publications more carefully. I have not yet done any serious tests on the
> > >effects of the new EOS on the solutions of an ocean circulation
> > >configuration, as I did not expect the effects to be large and
> > >significant. Wrong again?
> > >
> > >Martin
> > >
> > >On Feb 19, 2013, at 1:21 AM, Trevor McDougall
> > ><trevor.mcdougall at unsw.edu.au> wrote:
> > >
> > >> Hello Martin,
> > >>
> > >> I hear that you have put TEOS-10 into the MITgcm.
> > >>
> > >> Can you please tell me what you did with salinity?   I am assuming that
> > >>you have only one salinity variable, namely Absolute Salinity.   And I
> > >>assume that you treat Absolute Salinity as a conservative variable, yes?
> > >>
> > >> And I expect that your model temperature is Conservative Temperature,
> > >>and that in the air-sea interaction calculation, you calculate potential
> > >>temperature (from knowledge of Absolute Salinity and Conservative
> > >>Temperature there).
> > >>
> > >> Have I guessed what you have done correctly?
> > >>
> > >> By the way, which version of the GSW code did you start from? The
> > >>Matlab, the FORTRAN or the C version?
> > >>
> > >> I am in a small group that is implementing TEOS-10 into MOM (we are
> > >>keeping two different salinity variables, Preformed Salinity, and
> > >>another ratio term, F_delta), and any experience that you have found
> > >>that you want to share would be much appreciated.
> > >>
> > >> Also I hear that Fabien Roqet is implementing TEOS-10 into NEMO,
> > >>probably in the same manner that I am guessing you have done.
> > >>
> > >>    With very best wishes,
> > >>
> > >>       Trevor
> > >>
> > >>
> > >>
> > >> Trevor J McDougall
> > >> Scientia Professor of Ocean Physics
> > >> School of Mathematics and Statistics
> > >> University of New South Wales
> > >> NSW 2052, Australia
> > >>
> > >> email: Trevor.McDougall at unsw.edu.au
> > >> tel: +61 2 9385 3498
> > >> fax: +61 2 9385 7123
> > >> mob: +61 407 518 183
> > >>
> > >
> > 
> > 
> > 
> > 
> > -- 
> > Dr. Stephen M. Griffies
> > NOAA Geophysical Fluid Dynamics Lab
> > 201 Forrestal Road
> > Princeton, NJ 08542 
> > USA  
> > 
> 

> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel

More information about the MITgcm-devel mailing list