[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[GEMPAK #RGI-969779]: Gempak Decoders

Hi Flavio,

I'm sending this as a short guide for adding grids to NMAP2. The "big picture" 
is summed up in this article: 
https://www.unidata.ucar.edu/software/gempak/tutorial/adding_grids_to_nmap2.html
 

And thank you for sending along the grib files, that way I can see that dcgrib2 
works, and advise on what restore files to copy/edit.  Whatever additions you 
make to NMAP2 with these new models I would like to make part of the baseline 
GEMPAK release to the benefit of everybody.


ADDING GRIDS TO NMAP2

For Cosmo, I describe here how the NMAP_RESTORE files were chosen and edited.  
Because so many parameters are available, so you must determine with GDINFO 
what grids exist in the file, then choose from the available NMAP_RESTORE 
files, and make appropriate edits to match GLEVEL, GVCORD and GDPFUN for the 
new grid.

here is an example:  I started with a single Cosmo grid file, and I know that 
there are lots of HYBL level data, but I want to know what is at the surface, 
so I run GDINFO atGLEVEL = 0, GLEVEL=2 and GLEVEL=10

1) at GLEVEL = 0:

 GDFILE   = cosmo.gem
 GDATTIM  = all
 GLEVEL   = 0
 GVCORD   = all
 GFUNC    = all
 GEMPAK-GDINFO>

  NUM       TIME1              TIME2           LEVL1 LEVL2  VCORD PARM
  310     140514/0000F000                          0         NONE PRES
  311     140514/0000F000                          0         NONE LSNO
  312     140514/0000F000                          0         NONE CSNO
  313     140514/0000F000                          0         NONE TCLD
  314     140514/0000F000                          0         NONE CLDH
  315     140514/0000F000                          0         NONE CLDM
  316     140514/0000F000                          0         NONE CLDL
  317     140514/0000F000                          0         NONE ALBD
  318     140514/0000F000                          0         NONE SNOR
  319     140514/0000F000                          0         NONE PXXM
  320     140514/0000F000                          0         NONE URELFX
  321     140514/0000F000                          0         NONE VRELFX
  322     140514/0000F000                          0         NONE FXSH
  323     140514/0000F000                          0         NONE FXLH
  324     140514/0000F000                          0         NONE PMSL
  325     140514/0000F000                          0         NONE TMPK
  326     140514/0000F000                          0         NONE SPFH
  327     140514/0000F000                          0         NONE SNIR


2) at GLEVEL = 2

 GDFILE   = cosmo.gem
 GDATTIM  = all
 GLEVEL   = 2
 GVCORD   = all
 GFUNC    = all
 GEMPAK-GDINFO>

  NUM       TIME1              TIME2           LEVL1 LEVL2  VCORD PARM
  292     140514/0000F000                          2         HYBL VVEL
  305     140514/0000F000                          2         HGHT TMPK
  306     140514/0000F000                          2         HGHT DWPK
  307     140514/0000F000                          2         HGHT TMNK
  308     140514/0000F000                          2         HGHT TMXK
  309     140514/0000F000                          2         HGHT RELH


3) at GLEVEL =10

 GDFILE   = cosmo.gem
 GDATTIM  = all
 GLEVEL   = 10
 GVCORD   = all
 GFUNC    = all
 GEMPAK-GDINFO>

  NUM       TIME1              TIME2           LEVL1 LEVL2  VCORD PARM
  244     140514/0000F000                         10         HYBL VVEL
  303     140514/0000F000                         10         HGHT UREL
  304     140514/0000F000                         10         HGHT VREL



So here is what I know now:

available at 10m: (UREL and VREL @10 % HGHT)
available at 2m: TMPK, RELH, MIN/MAX TEMPS, Vertical Velocity)
available at 0m: SPFH, PRES, PMSL, TMPK, ALBD (albedo) and more.

It is from this set of known parameters / levels / vertical coordinates that 
you must construct new NMAP_RESTORE files.

You mention SPFH, so I looked for any existing restore files which use SPFH, 
and saw "rucs/sfc_moist_conv" 

cd $NMAP_RESTORE/modl/
grep -ir spfh * 
rucs/sfc_moist_conv:GDPFUN  sdiv(spfh;wnd)!dwpc!dwpc!dwpc ! kntv(wnd)

So I copied it to $NMAP_RESTORE/modl/cosmo/ and checked it such that SPFH used 
GLEVEL=0, GVCORD=NONE (it did already).

Another example: I looked through $NMAP_RESTORE for files called "sfc" or "bl". 
I copied the file $NMAP_RESTORE/modl/global/pmsl_and_bl_wind to 
$NMAP_RESTORE/modl/cosmo/pmsl_and_bl_wind directory and made edits.  The 
original file was:

GDATTIM fall
GLEVEL  9950!0
GVCORD  sgma!none
GDPFUN  mag(kntv(wnd))!sm5s(pmsl) ! kntv(wnd@9950%sgma)

Notice GLEVEL=9950, GVCORD=SGMA.  The cosmo grid does not have SGMA as a 
vertical level, so I changed each "wnd" reference to explicity be WND@10%HGHT.

GDATTIM fall
GLEVEL  0
GVCORD  none
GDPFUN  mag(kntv(wnd@10%hght))!sm5s(pmsl) ! kntv(wnd@10%hght)


"pmsl", you will notice, takes on the GLEVEL=0, GVCORD=none defintion.  
therefore, this new cosmo file pmsl_and_bl_wind will show surface pressure and 
10 meter winds.

It all starts with a grid field inventory using GDINFO, we must know what is 
inside the file in order to choose and edit NMAP_RESTORE files.

That is the bulk of the work, after this is complete, all you need are entries 
in mod_res.tbl and datatype.tbl, as you saw.  

You can see that it is quite manual, the checking of fields in the grid file 
against restore file syntax.  Hope this guides you well.

-Michael



> Michael,  I take a look at the files you sent me, the datatype.tbl and
> mod_res.tbl, looks like OK, but the files on the modl/cosmo folder is not
> so easy to work around, specially the GDPFUN parameter, I saw that this
> parameter receives the function used by Gempak to evaluate the data, so I
> need to know which functions this parameter support so the metereologists
> can write they own equations. Is there a doc about this parameter on the
> Gempak site, I did not find it no Gempak Parameter Help page.
> 
> The Navy will send me the samples of each GRIB and the variables they want
> to see, as soon I get it I will send you.
> 
> Thanks,
> 
> 
> 
> *Flavio TobiasDiretor de Tecnologia / CTO*
> 
> 
> *Itrax Sistemas e EquipamentosRua Astolfo Dutra nº 652- B. Vera Cruz, Belo
> Horizonte - MG*
> 
> *Tel.: +55 (31) 3358-0000 / Cel.: +55 (31) 9423-8019*
> 
> 
> 
> 2014-07-15 21:10 GMT-03:00 Flavio Tobias <address@hidden>:
> 
> > GREAT !!!!!!!!!! Worked !!!
> > Tomorrow the navy officers will be amazed.
> >
> > I will ask them to evaluate the data and tell if is that they a looking
> > for, I will let you know.
> >
> > However we have another two or three GRIB models and one satellite image
> > to display on nmap2. How is the Unidata process in this case? Do you teach
> > me how to do it or do I send you the files so you can create the TBL files?
> >
> > Thank you very much Michael.
> >
> >
> >
> >
> >
> >
> >
> >
> > *Flavio TobiasDiretor de Tecnologia / CTO*
> >
> >
> > *Itrax Sistemas e EquipamentosRua Astolfo Dutra nº 652- B. Vera Cruz, Belo
> > Horizonte - MG*
> >
> > *Tel.: +55 (31) 3358-0000 <%2B55%20%2831%29%203358-0000> / Cel.: +55 (31)
> > 9423-8019 <%2B55%20%2831%29%209423-8019>*
> >
> >
> >
> > address@hidden>:
> >
> > Hi Flavio,
> >>
> >> Attached is a tar.gz archive with that includes updates for datatype.tbl,
> >> mod_res.tbl, grid_cat.tbl and a small number of default NMAP restore files.
> >>
> >> Move the tar.gz file to GEMPAK7/gempak/ and un-compress (tar -xvzf
> >> gempak_nmap2_cosmo.tar.gz) and it will add/overwrite the new files to
> >> GEMPAK7/gempak/tables/ and GEMPAK7/gempak/nts/
> >>
> >> Be sure that the entry in datatype.tbl is exactly how the cosmo files are
> >> written to disk ($GEMDATA/model/YYYYMMDDHH_cosmo.gem).
> >>
> >> After that, re-start NMAP2 and navigate to GRID -> International_Models
> >> -> cosmo ... and you should see available dates and then product files.
> >>
> >> The product files will be in GEMPAK7/gempak/nts/nmap/restore/modl/cosmo/
> >> - that is where you can view, edit and copy to more.  Each new file
> >> requires an entry in the table $GEMTBL/nmap/mod_res.tbl
> >>
> >> Also, I was not able to put anything together using the hybrid layers
> >> (HYBL) in the Cosmo grid file.  From the Cosmo specification document it
> >> appears to be pressure/sigma levels but I'm not sure what that means in for
> >> dynamics and thermodynamics in these restore files.
> >>
> >> Hope this helps,
> >>
> >> -Michael
> >>
> >>
> >> > Hi Michael, I would like to know if you can give me a status regarding
> >> the
> >> > TBL file. Sorry for ask you that, but tomorrow I wil have a metting with
> >> > navy officers in order to give them some answers regargind the project
> >> > progress. As I told you we have another GRIBs here and some imagery
> >> files
> >> > that must be also opened on the NMAP2 and so far we have not succeed.
> >> > So if you can give some information I will appreciate.
> >> >
> >> > Many thanks,
> >> >
> >>
> >> Ticket Details
> >> ===================
> >> Ticket ID: RGI-969779
> >> Department: Support GEMPAK
> >> Priority: Normal
> >> Status: Open
> >
> >
> >
> 
> 

Ticket Details
===================
Ticket ID: RGI-969779
Department: Support GEMPAK
Priority: Normal
Status: Open