[MITgcm-support] 9km and 4km Arctic configurations (from CS510)

[Menemenlis, Dimitris (3248)]

Hi Jiping, I have gathered all the comments from everyone and put them here:
http://ecco2.jpl.nasa.gov/data1/arctic/SetupInstructions/

For running the 9-km set-up, you can start from:
http://ecco2.jpl.nasa.gov/data1/arctic/code_9km/
http://ecco2.jpl.nasa.gov/data1/arctic/run_template_9km/
but replace SEAICE_OPTIONS.h and data.seaice
from the 4 or the 18-km versions that An sent you.

For completeness, I attach An's reply below, so that it is recorded
in the MITgcm-support archives so as to enlighten future
generations of Arctic MITgcmers!

Cheers

Dimitris Menemenlis

=========

Hello Jiping,

I have uploaded the required binaries + code + run_time_inputs here for the 18km and 4km:
https://www.dropbox.com/sh/i6l704bi0gesx0a/jU3tfCoMaA

I have not run the 9km set up in a long time so it would take a while for me to gather inputs (some time next week perhaps).  I'm also cc'ing Martin Losch who ran the 4km last December, perhaps he can tell us about the stability of 4km run with latest MITgcm code.

A few notes:

arctic18km setup:
- code_arctic18km_forBruno.tgz
- inputbinaries_arctic18km_forBruno.tgz
- namelists_arctic18km_forBruno.tgz

This one is most straight forward and has been verified to work with the latest MITgcm code.  The version you get here works with checkpoint c63r stably (as David pointed out).  check out using: cvs co -r checkpoint63r -P MITgcm_code

arctic4km setup:
inputbinaries_arctic4km_forMartin.tgz
namelists_arctic4km_forMartin.tgz
code4km_MITgcmc62k.tgz (produces mitgcmuv_4km_960p_oldcode.gz )
code4km_MITgcmc63r.tgz : successfully compiled but was not used

As you notice, there are two code directories.  The original one, checkpoint c62k, was used in 2009 for a complete run 1992-2010.  Last year, I wanted to extend the run to include 2011 but did not want to have potential discontinuities from latest updates in the sea-ice code in the MITgcm.  As a result, even though i compiled the code for checkpoint c63r, the old executable mitgcmuv_4km_960p_oldcode.gz was used instead.

Lastly, it's best not to use old MITgcm code but as you can see, because the code is constantly changing, we always have to make a decision regarding the checkpoints.  I'd be happy to help you with setting these config up if you want to use latest code (and Martin might also help here).  Next week would be best for me, have a couple of deadlines this week.

Regards,
An

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20130412/78d4eb30/attachment.htm>