[Mitgcm-support] RE: Dr Knudsen I presume

[mitgcm-support at dev.mitgcm.org]

Hi Tor,

 I hope EGS went well.

The 16 levels are certainly good for large scale
global problems. I think they should be pretty good for most places of the
ocean, however, for some anomalous local regions, however,
all our small scale convection is linear EOS stuff - energy conservation et.
al go to pieces for a non-linear EOS.

You might want to check them out

Bryan, K., and M. D. Cox, 1972: An approximate equation of state for
numerical models of ocean circulation. Journal of Physical Oceanography,
2(4), 510-514.

which has plots of the validity of the scheme.

You could easily customize Tmin, Tmax and Smin/Smax for your region
if you want to improve accuracy. We have never investigated
adding extra "sample" points in the vertical. My intuition would be
that if this makes a big difference, then you should probably
be using the full equation not the least-squares fit. Of course
we don't have any setups with the full equation integrated
into the code!!!!

 You might also want to worry about Knudsen versus the modern full EOS.
That could be significant in convection regions.

Chris

-----Original Message-----
From: ltorel at nersc.no [mailto:ltorel at nersc.no]On Behalf Of Tor Eldevik
Sent: Saturday, May 04, 2002 9:55 AM
To: Chris Hill
Cc: support at mitgcm.org
Subject: Dr Knudsen I presume


Chris,

I've started doing my experiments using the ``full''
(i.e., Knudsen) EOS.

What is the ideal, if any, relation between the
number of Tmin, Tmax etc (16 in the utils/knudsen2-file),
the 250m bands (why 250?), and the particular vertical
resolution (the dz's) of one's experiment.

Have a nice weekend,
tor
--
=======================================================
  Tor Eldevik, Research Scientist

  Nansen Environmental and Remote Sensing Center
  Edvard Griegs vei 3a, N-5059 Bergen, Norway

  +47 55 29 29 01 ext. 225 (direct)
  +47 55 29 72 88          (manual switch board)
  +47 55 20 00 50          (fax)

  tor.eldevik at nersc.no     http://www.nersc.no/~torel/
=======================================================