[MITgcm-support] Ice thickness category diagnostics

[Martin Losch]

Hi David,

did you try to keep the mnc output in the working directory ( mnc_use_outdir=.false., but, again, you need to either remove grid.*.nc or use debugLevel=-1). Do you think it is related the mnc_outdir_str? I run models with  mnc_use_outdir=.false., all the time and restarts are not a problem at all.

Martin

On Feb 8, 2013, at 2:07 PM, David Huard <david.huard at gmail.com> wrote:

> To clarify, the situation is the following:
> 
> I ran the model once from timestep0 to timestep N with results going into results_0. Pickup files are stored with mdsio, model outputs are stored in netCDF. All good. 
> 
> I'm then using a pickup file to restart the model at the timestep N, changing the output directory to results_N. 
> At this point, the directory results_N is empty. When I launch the model, nothing happens. No STDOUT, no STDERR, no output written in results_N. 
> 
> To get the model started, it seems like I need to launch the  model from timestep0 for one timestep, so it creates the grid files in results_N. Then I can set timestep0 to N and launch the real run.
> 
> 
> 
> 
> On Fri, Feb 8, 2013 at 3:44 AM, Martin Losch <Martin.Losch at awi.de> wrote:
> Hi David, and Manfredi,
> 
> I am not sure if I understand the problem, but ---as a feature--- mnc does not allow overwriting old netcdf files. In a "production-run" with many restarts, you can still have the output in the same (working directory) if you set the debugLevel = -1. This way the grid*.nc files are only written when nIter0 = 0. All other mnc-files are unique because they have the value of nIiter0. This works very well for me.
> 
> I have never seen the model wait for anything except for the file system, so I suspect that your problems are connected to that.
> 
> BTW, as a general rule, I prefer to not use mnc for the pickup files.
> 
> I can offer to put the noleap option into the repository, if it is general enough.
> 
> Martin
> 
> 
> On Feb 8, 2013, at 3:54 AM, David Huard <david.huard at gmail.com> wrote:
> 
> > Hi Manfredi,
> >
> > This is what I've done, but if the grid.xxx.nc files are not in this directory, the model just waits forever. This was on a model version checkpoint63r, so maybe it has been fixed.
> >
> >
> > On Thu, Feb 7, 2013 at 9:47 PM, Manfredi Manizza <mmanizza at ucsd.edu> wrote:
> > David,
> >
> > for the restart problem using MNC you have to mkdir another folder where to store
> > the new output and update the new name in data.mnc.
> >
> > This is what I do and at least this used work to me....
> >
> > Manfredi
> >
> >
> > On 02/07/2013 06:32 PM, David Huard wrote:
> >> Hi all,
> >>
> >> I'm doing a run with the MITgcm ocean+ice Arctic configuration driven by the GFDL GCM. Things have been pretty easy to set up, and I'd like to offer my kudos to the devs.
> >>
> >> I couldn't find in the docs how to output the sea ice concentration of the different sea ice thickness categories. How would I do that ?
> >>
> >> Also, I noticed that using MNC, if timestep is not zero but the output directory is empty, as would be the case for a restart, the model just hangs there forever without any STDOUT or STDERR files being created.
> >>
> >> Finally, I added a noleap option to the calendar package to support the GCM calendar. I could send the diff if there is interest.
> >>
> >> Cheers,
> >>
> >> David Huard, PhD
> >>
> >>
> >>
> >>
> >> _______________________________________________
> >> MITgcm-support mailing list
> >>
> >> MITgcm-support at mitgcm.org
> >> http://mitgcm.org/mailman/listinfo/mitgcm-support
> >
> >
> > --
> > Dr. Manfredi Manizza
> > Geosciences Research Division
> > Scripps Institution of Oceanography
> > University of California San Diego
> > 9500 Gilman Drive La Jolla, CA 92093-0244
> > email :
> > mmanizza at ucsd.edu
> >
> > tel   :
> > +1-858-534-7094
> >
> > web   :
> > http://bluemoon.ucsd.edu/mmanizza/
> >
> > _______________________________________________
> > MITgcm-support mailing list
> > MITgcm-support at mitgcm.org
> > http://mitgcm.org/mailman/listinfo/mitgcm-support
> >
> >
> > _______________________________________________
> > MITgcm-support mailing list
> > MITgcm-support at mitgcm.org
> > http://mitgcm.org/mailman/listinfo/mitgcm-support
> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support