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

19990625: code for GRIB to .gem for GARP use



John,

The thing that comes to mind is that you had GAREA set to the subset
region when you ran nagrib. Try rerunning nagrib with:
cpyfil=gds
garea=dset

The cpyfil=gds uses the grid definition section from the grib
message to define the area of the grid (instead of you having to
specify the GRDAREA, PROJ and KXKY). When you use cpyfil=gds, those
other 3 variables are ignored...but GAREA is still used to subset
the region you want to obtain the gridpoints within.

Running "phelp nagrib" at the GEMPAK-NAGRIB> prompt, will show 
the effect of GAREA:
     GAREA is used to subset the grid.  It may be specified as
     any valid input for GAREA which are equivalent to lat/lon bounds
     or as a single @ followed by the lower-left and upper-right grid
     point coordinates separated by semicolons.  In the former case,
     the locations are rounded to the nearest grid points.  The
     projection type and angles entered as described above along with
     the lat/lon coordinates of the corners of the subset define a
     new grid navigation different from that of the full grid.  The
     navigation entered as described in the preceding paragraph is
     always that of the full grid.  To get the full grid, set GAREA to
     blank, GRID or DSET.

Steve Chiswell
Unidata User Support




On Fri, 25 Jun 1999, John Knox wrote:

> Dear Steve,
> 
> I found my error [key omission of word 'source' at one 
> point] and got things to work--almost.
> 
> I was able to convert the data from GRIB to .gem 
> successfully, no problems there.
> 
> Now the data is readable by GEMPAK and looks great...
> ...except...
> 
> the only region for which there is data is the Eastern U.S. 
> I've looked at it carefully, and it's *not* a mapping 
> problem where all the data is squished down into one area.  
> Instead, it's a problem where the full data set is not 
> being displayed.  I know because the western edge of the 
> data terminates near the low pressure system I want to see, 
> and the pressure data fits exactly with what I expect to 
> see in that region, at that time.  [I'm used to mapping 
> errors from my Ph.D. work.]
> 
> I verified that I hadn't mucked things up overall. 
> I was able to visualize RUC data for the whole U.S. at the 
> same time that I could only see NCEP for that darned square 
> over the eastern U.S.
> 
> So I thought, 'somewhere there must be a 
> default for the range of data displayed that's messed up.'  
> I found such a thing in Garp_defaults, but at least once 
> when I changed it, it didn't make a difference.  
> 
> So, I'm almost there--just need to be able to see the whole 
> data set instead of a part of it.  If you can get me over 
> that hump, I'll be very grateful!
> 
> Thanks, 
> 
> John
> 
> 
> On Fri, 25 Jun 1999 16:03:57 -0600 Steve Chiswell 
> <address@hidden> wrote:
> 
> > 
> > 
> > John,
> > 
> > You can ftp the file to where you downloaded the distribution from:
> > ftp.unidata.ucar.edu
> > 
> > use the gbuddy login you were given and cd to the incoming directory.
> > 
> > If you are in a csh, try running the command:
> > source /home/gempak...../Gemenviron
> > 
> > Gemenviron is in the top directory of the Gempak tree- wherever it is on
> > your system. Once you have sourced the file, that should add the $GEMEXE
> > directory to your path and define a bunch of other table directories which
> > all the Gempak programs use. If Gempak can't find those tables, it is
> > very clueless.
> > 
> > The Gemenviron file defines the location of NAWIPS near the top
> > of the file. If your sysadmin had set up gempak, then this location
> > should be correct- otherwise you will need to edit the file.
> > 
> > Steve Chiswell
> > Unidata User Support
> > 
> > 
> >  On Fri, 25 Jun 1999, John Knox wrote:
> > 
> > > Dear Steve,
> > > 
> > > Thanks for the comments.  Our former sysadmin set us up in 
> > > a way that does seem to do Gemenviron, but I can't find a 
> > > place where "source Gemenviron" is actually stated.  We've 
> > > in .bash but we type "csh" and so far I can't determine 
> > > what that does exactly--clearly it puts us in a c-shell, 
> > > but I can't find what that command does exactly.  I'd guess 
> > > that we're not all hooked up Gemenviron-wise, given the 
> > > problem with getting nagrib to run outside of its directory.
> > > 
> > > Where exactly do you want me to ftp to [full address]?  
> > > I'll send you a GRIB file and the commands we used.
> > > 
> > > Thanks for your help, and again, sorry to be so slow about 
> > > these things!
> > > 
> > > John
> > > 
> > > On Fri, 25 Jun 1999 15:08:53 -0600 Steve Chiswell 
> > > <address@hidden> wrote:
> > > 
> > > > John,
> > > > 
> > > > Your environment should be configured so that your csh sources the 
> > > > $NAWIPS/Gemenviron file. This will add the location of the
> > > > nawips executable into your path. You should not need to
> > > > run within the $GEMEXE directory...and you may not have write
> > > > permission in that directory to create the output file.
> > > > If your environment does not find the nagrib program unless you are in 
> > > > the
> > > > GEMEXE directory, then it sounds like you have not sourced Gemenviron
> > > > into your working environment.
> > > > 
> > > > It would be of help to me to see your input commands to nagrib
> > > > and if possible, could you ftp sample of your grib file into
> > > > ~gbuddy/incoming 
> > > > 
> > > > To make garp see your gempak grid file, the resultant file will need to
> > > > be located in the directory pointed to by $HDS, and the
> > > > $GARPHOME/config/Garp_defaults file will need matching entries for
> > > > modellabels and modelkeys (eg strings that will be identify the data set
> > > > in the Garp model menu, and corresponding key string to match
> > > > when searching the $HDS directory). That is, Garp uses the modelkeys
> > > > list to identify files which match the patterns in the comma separated
> > > > list, and associates them with the model pull down menu lable which
> > > > uses the string in the corresponding entry in modellabels.
> > > > 
> > > > Steve Chiswell
> > > > Unidata User Support
> > > > 
> > > > 
> > > > 
> > > > On Fri, 25 Jun 1999, John Knox wrote:
> > > > 
> > > > > Dear Steve,
> > > > > 
> > > > > Thanks for the information... I'm new to GEMPAK, not a 
> > > > > computer wizard, and our systems person graduated (not to 
> > > > > be replaced) in May.  [Welcome to a small school.]
> > > > > 
> > > > > Quick follow-up questions:
> > > > > 
> > > > > 1) Am I right that to get NAGRIB working, you simply go to 
> > > > > the directory it's in and type "nagrib"?
> > > > > 2) We tried this and put in all the commands Bob Creasey 
> > > > > suggested plus the table command you suggested, but it gave 
> > > > > a [NAGRIB -12] error and didn't read anything or write 
> > > > > anything.  I thought this might be an error in specifying 
> > > > > where the GRIB data resides, but I placed the GRIB file in 
> > > > > the same directory as nagrib without any improvement.  
> > > > > What's wrong?
> > > > > 3) Assuming this hurdle can be overcome and the .gem files 
> > > > > are written, then one more question: how will GARP know 
> > > > > that this data is there?  Will I see "NCEP" or whatever 
> > > > > displayed as a choice on the top left of the GUI?
> > > > > 
> > > > > Thanks for your help, and sorry to be such a novice.
> > > > > 
> > > > > John
> > > > > 
> > > > > On Fri, 25 Jun 1999 12:50:33 -0600 Steve Chiswell 
> > > > > <address@hidden> wrote:
> > > > > 
> > > > > > John,
> > > > > > 
> > > > > > nagrib and dcgrib are both used for converting Grid format
> > > > > > data into gempak files. I provided a ncarncep reanalysis
> > > > > > parameter conversion table for mapping the grib parameters 
> > > > > > to gempak parameters (ncar/ncep reanalysis does not use the same
> > > > > > table as the ncep operational models). The table ncarncep1.tbl
> > > > > > should be located in the $GEMTBL/grid directory - you can
> > > > > > specify for nagrib to use this table by setting:
> > > > > > gbtbls = ncarncep1.tbl
> > > > > > 
> > > > > > Steve Chiswell
> > > > > > Unidata User Support
> > > > > > address@hidden
> > > > > > 
> > > > > > On Fri, 25 Jun 1999, John Knox wrote:
> > > > > > 
> > > > > > > 
> > > > > > > I have some data files in GRIB format from the NCEP 
> > > > > > > Reanalysis project that I would like to be able to view 
> > > > > > > using GEMPAK/GARP.  Does anyone have software that can 
> > > > > > > convert GRIB to .gem format?
> > > > > > > 
> > > > > > > Thanks!
> > > > > > > 
> > > > > > > John
> > > > > > > 
> > > > > > > ------------------------------------------------
> > > > > > > Dr. John A. Knox
> > > > > > > Assistant Professor of Geography and Meteorology
> > > > > > > Room 15, Mueller Hall
> > > > > > > Valparaiso University
> > > > > > > Valparaiso, IN 46383
> > > > > > > E-mail: address@hidden
> > > > > > > Office Phone: 219-464-6003
> > > > > > > 
> > > > > > > 
> > > > > > 
> > > > > 
> > > > > ------------------------------------------------
> > > > > Dr. John A. Knox
> > > > > Assistant Professor of Geography and Meteorology
> > > > > Room 15, Mueller Hall
> > > > > Valparaiso University
> > > > > Valparaiso, IN 46383
> > > > > E-mail: address@hidden
> > > > > Office Phone: 219-464-6003
> > > > > 
> > > > > 
> > > > 
> > > 
> > > ------------------------------------------------
> > > Dr. John A. Knox
> > > Assistant Professor of Geography and Meteorology
> > > Room 15, Mueller Hall
> > > Valparaiso University
> > > Valparaiso, IN 46383
> > > E-mail: address@hidden
> > > Office Phone: 219-464-6003
> > > 
> > > 
> > 
> 
> ------------------------------------------------
> Dr. John A. Knox
> Assistant Professor of Geography and Meteorology
> Room 15, Mueller Hall
> Valparaiso University
> Valparaiso, IN 46383
> E-mail: address@hidden
> Office Phone: 219-464-6003
> 
>