NOTE: The decoders mailing list is no longer active. The list archives are made available for historical reasons.
Brent, Could you run gribdump on the raw file and show me the headers for the different parameters? Are the parameter names correct in gribdump? If not could you match the parameters to the correct names so I can make a table for gribtonc then you will get the correct names in the output. Also have you contacted UCAR about this problem, they might already have a table. RObb... On Thu, 22 Apr 2004, Brent A McDaniel wrote:
Hi Robb, Thanks for the reply. Here's a sample of the output from gribtonc: ... Apr 12 18:33:04 gribtonc[13809]: 27203: wrote reserved(107,11,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27204: wrote thnd_prob(107,10,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27205: wrote Pm(107,10,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27206: wrote Psl_et(107,10,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27207: wrote LI(107,10,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27208: wrote LI4(107,10,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27209: wrote Kind(107,10,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27210: wrote mois_div(107,10,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27211: wrote bvf_2(107,10,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27212: wrote gflux(107,10,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27213: wrote cape(107,10,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27214: wrote reserved(107,10,*,*) to output.nc Apr 12 18:33:04 gribtonc[13809]: 27215: wrote thnd_prob(107,9,*,*) to output.nc ... As far as sending a sample, unforunately the grib file is quite large (~666mb) and I don't know how to subset that. Is there an easy way to do that? Thanks for the help! Cheers, Brent On Tue, 20 Apr 2004, Robb Kambic wrote: > Brent, > > If nobody has the proper cdl file then could you send me a sample file? > My speculation is that a table is being used to convert variables from > ECMWF variables to the NCEP variable names. Reason for the weird names. > Also, what was the log messages from gribtonc like? > > Robb... > > On Tue, 20 Apr 2004, Brent A McDaniel wrote: > > > > > I've just begun to use the ECMWF ERA40 data that is available from the > > NCAR data archives. I'm using the 4x daily upper air pressure level data > > that has 11 variables: geopotential, temperature, u wind component, v wind > > component, w wind component, relative humidity, specific humidity, > > vorticity, divergence, potential vorticity, ozone mass mixing ratio. The > > data is archived in grib format for 1 month. > > > > Running wgrib -v e4oper.pl.195801 |more results in: > > 1:0:D=1958010100:var60:1 mb:kpds=60,100,1:anl:ECMWF stream?:"undefined > > 2:21240:D=1958010100:Z:1 mb:kpds=129,100,1:anl:ECMWF stream?:"Geopotential (at the surface = orography) m**2 s**-2 > > 3:42480:D=1958010100:T:1 mb:kpds=130,100,1:anl:ECMWF stream?:"Temperature K > > 4:63720:D=1958010100:U:1 mb:kpds=131,100,1:anl:ECMWF stream?:"U-velocity m s**-1 > > 5:84960:D=1958010100:V:1 mb:kpds=132,100,1:anl:ECMWF stream?:"V-velocity m s**-1 > > 6:106200:D=1958010100:Q:1 mb:kpds=133,100,1:anl:ECMWF stream?:"Specific humidity kg kg**-1 > > 7:127440:D=1958010100:W:1 mb:kpds=135,100,1:anl:ECMWF stream?:"Vertical velocity Pa s**-1 > > 8:148680:D=1958010100:VO:1 mb:kpds=138,100,1:anl:ECMWF stream?:"Vorticity (relative) s**-1 > > 9:169920:D=1958010100:D:1 mb:kpds=155,100,1:anl:ECMWF stream?:"Divergence s**-1 > > 10:191160:D=1958010100:R:1 mb:kpds=157,100,1:anl:ECMWF stream?:"Relative humidity % > > 11:212400:D=1958010100:O3:1 mb:kpds=203,100,1:anl:ECMWF stream?:"Ozone mass mixing ratio kg kg**-1 > > 12:233640:D=1958010100:var60:2 mb:kpds=60,100,2:anl:ECMWF stream?:"undefined > > ... > > > > which appears to be the variables that are supposed to be there although I > > don't see potential vorticity (is that the undefined?). > > > > I then run gribtocdl to generate the .cdl file: > > gribtocdl -v e4oper.pl.195801 > era40.jan.1958.cdl > > > > and then run gribtonc to convert to netcdf: > > gribtonc -v -l- era40.jan.1958.cdl output.nc < e4oper.pl.195801 > > > > This works properly and I get a netcdf file. When I look at it though, > > all the variable names are wierd. Looking at the .cdl file I see that in > > there the variable names aren't Z,T, U etc that were in the grib file > > they are things like thund_prob, K index at isobaric levels, and other odd > > things. The time is also off, showing the initial time as time since 1994 > > (data being tested was 1958). I did check one thing though. I plotted > > the 3rd variable in the created netcdf file (which should be temperature) > > and compared it the ncep reanalysis temperature field for jan 1958 and > > they match up nicely so the data is somewhat correct. > > > > So, can someome help me out with this? Does anyone have a .cdl template > > for the ERA40 data set already? I'm most concerned with the time > > variables as I'm not sure exactly how grib keeps track of time. > > > > Thanks for any and all help. > > > > Cheers, > > > > Brent > > > > -- > > Brent A. McDaniel > > > > Dept of Earth and Atmospheric Sciences > > Georgia Institute of Technology > > Atlanta, Ga. USA > >
============================================================================== Robb Kambic Unidata Program Center Software Engineer III Univ. Corp for Atmospheric Research rkambic@xxxxxxxxxxxxxxxx WWW: http://www.unidata.ucar.edu/ ==============================================================================
decoders archives: