[MITgcm-support] confusion about parameters diagstats
Martin Losch
Martin.Losch at awi.de
Fri Aug 29 10:40:54 EDT 2008
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
More information about the MITgcm-support
mailing list