[MITgcm-support] confusion about parameters diagstats

[Martin Losch]

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