[MITgcm-support] Regional statistics diag (region-mask)

[Jean-Michel Campin]

Hi Abbas,

it sounds like a problem in namelist "DIAG_STATIS_PARMS".
Could you try without the dot (since an Integer is expected)
for variable stat_region :
> stat_region(1,1)= 1,
instead of:
> stat_region(1,1)= 1.,

otherwise, if you want 4 differents profiles (one for each location)
instead of an average profile of the 4 locations, you need to
define 5 regions (the 4 locations + elsewhere) by setting:
> regMask(posaS1,poscS1,:)=1
> regMask(posaS1,posdS1,:)=2
> regMask(posbS1,poscS1,:)=3
> regMask(posbS1,posdS1,:)=4
in your matlab script.

Jean-Michel

On Thu, Apr 09, 2009 at 04:57:21PM -0400, Abbas Dorostkar wrote:
> Dear all,
> 
> I would like to use the "region-mask" capability to output temp profile at 4
> points in all vertical layers in a 3D model to compare with field data.
> When I run the model I get core_dumped and the last line of STDOUT gives me:
> (PID.TID 0002.0001) S/R DIAGNOSTICS_READPARMS, read namelist
> "DIAG_STATIS_PARMS": start
> 
> I defined two regions so that region_1 (with value of the mask=1) includes
> those 4 points in all layers and region_2 (with value of the mask= 0)
> includes points in the rest of domain. I have a mask with 1 level because
> there is no overlap between two different regions. I set the
> data.diagnostics as bellows:
> ---------------------------------------
> &DIAG_STATIS_PARMS
> diagSt_regMaskFile='regMask.bin',
> nSetRegMskFile=1,
> set_regMask(1)= 1,  1,
> val_regMask(1)= 0.,1.,
> stat_fields(1,1)='THETA',
> stat_region(1,1)= 1.,
> stat_fname(1)= 'dynStDiag',
> stat_freq(1)= -400.,
> stat_phase(1)= 0.,
> 
> I generated the regMask file using following script:
> -----------------------------------
> load strings_coord.mat
> nx=120;
> ny=48;
> nz=98;
> regMask=zeros(nx,ny,nz);
> regMask(posaS1,poscS1,:)=1
> regMask(posaS1,posdS1,:)=1
> regMask(posbS1,poscS1,:)=1
> regMask(posbS1,posdS1,:)=1
> namfil='regMask.bin';
> fid=fopen(namfil,'w','b'); fwrite(fid,regMask,'real*8'); fclose(fid);
> fprintf([' write mask to file: ',namfil,'\n']);
> return
> 
> I have no idea what I am doing wrong and I could not find any documentation
> about region-mask. Is something wrong with my data.diagnostics setup and/or
> regMask file? Should the regMask file be like regMask(nx,ny)?
> 
> Thanks very much for your help
> Abbas

> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support