[MITgcm-support] confusion about parameters diagstats
Jean-Michel Campin
jmc at ocean.mit.edu
Fri Aug 29 15:50:54 EDT 2008
Hi Martin,
The only thing I could think of to explain this:
> the first 15 regions have reasonable values and then everything in wrong:
would be to check that you really changed :
>> stat_region(1,1)= 0, 1, 2, 3, 4, 5, 6, 7, 8, 9,10,11,12,13,14,
If you did not change the list of selected regions you want,
then it goes back to our discussion (on the developpment list,
from July 30):
> (if I remember well, the problem I had was that all the regions
> and not only the selected ones were written in the mnc file).
and this would explain what you are getting.
Now, this is just a guess, but before I try to reproduce the problem,
generating a region mask with 4 sets of regions and a total of
382 regions, I would propose that you try 1rst to run a short test,
with at least 1 reccord in the diagstat output file (you can do 2
iterations and set stat_freq(1) equal to the time-step to get 2 reccords)
but without using mnc (just setting diagSt_mnc = .FALSE. should
do it). If it's still so "confusing" after this, then I will
try to reproduce the problem.
Cheers,
Jean-Michel
On Fri, Aug 29, 2008 at 04:40:54PM +0200, Martin Losch wrote:
> Hi Jean-Michel,
> I am making strange mistakes with the diagnostics package, in particular
> with the diagstats part:
>
> I am abusing the "regMsk" capability to save time averages at a lot of
> individual points (and small areas) for comparison with CTD
> sections/moorings etc. First I tried this (and it worked):
>> &DIAG_STATIS_PARMS
>> diagSt_mnc = .true.,
>> diagSt_regMaskFile='arcticNetMask.bin',
>> nSetRegMskFile= 4,
>> set_regMask = 1 , 1 , 2 , 2 , 3 , 1 , 2 , 3 , 4 , 1 , 1 , 1 , 1 , 2
>> ,
>> val_regMask = 1., 2., 3., 4., 5., 6., 7., 8.,
>> 9.,10.,11.,12.,13.,14.,
>> #---
>> stat_fields(1,1)= 'THETA ','SALT ','DRHODR ',
>> 'TRAC04 ','TRAC07 ','TRAC09 ',
>> 'TRAC05 ','TRAC08 ','TRAC10 ',
>> 'TRAC14 ','TRAC15 ',
>> 'TRAC01 ','TRAC06 ','TRAC13 ','TRAC16 ',
>> stat_fname(1) = 'ArcticNetMoorings',
>> stat_freq(1) = 86400.,
>> stat_region(1,1)= 0, 1, 2, 3, 4, 5, 6, 7, 8, 9,10,11,12,13,14,
>> &
> So obviously, I have a mask with 4 different levels (because there is
> overlap between the different regions). For each region I asign the
> level in set_regMask, the mask value in val_regMask and the actual
> regions to be used in output stream 1 in stat_region. The maskfile
> contains the appropriate masks with values 1:14. and I get a netcdf file
> with 15 regions.
>
> Now I tried to extend the number of regions, but still not using them
> with new mask file the following modifications:
>> nSetRegMskFile= 4,
>> set_regMask = 1 , 1 , 2 , 2 , 3 , 1 , 2 , 3 , 4 , 1 , 1 , 1 , 1 , 2
>> , 368*1,
>> val_regMask = 1., 2., 3., 4., 5., 6., 7., 8.,
>> 9.,10.,11.,12.,13.,14.,
>> 15., 16., 17., 18., 19., 20., [...], 382.,
>
> and now I get a netcdf file with 386 regions (which I don't understand),
> the first 15 regions have reasonable values and then everything in wrong:
> a lot of regions of the 386 that are completely zeros and some values
> that are obviously not from the regions I wanted, eg. too high
> temperatures.
>
> What am I doing wrong? Or is this a bug in the interplay of mnc and
> diagstats?
>
> Martin
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support
More information about the MITgcm-support
mailing list