[Wrf-users] Re: Availability of GFS data up to 1 mb

Mon Jan 12 15:26:30 MST 2009

Eric;

> but I have discovered that data is available up to 1 mb through the "OpenDAP" interface on nomads.ncep.noaa.gov (the data is also available on 47 pressure levels instead of the usual 26).
The quick answer is to see the following directory location on the 
NOMADS server:

http://nomads.ncep.noaa.gov/pub/data/nccf/com/gfs/prod/gfs.2009011206/master/

> data is available up to 1 mb through the "OpenDAP" interface on nomads.ncep.noaa.gov 

The OPeNDAP server can stream entire fields or area subsets which can be 
made into GRIB files, if one must:  see 
http://www.cpc.noaa.gov/products/wesley/
and check gribw and also http://www.nco.ncep.noaa.gov/pmb/codes/GRIB2/

Longer answer:

The data set exists but are located in a different storage location.  
NOMADS continues an early operational precedent that had all the levels 
and variables in one file.  Since NOMADS has applications that slice and 
dice files that are large it is not a problem.  Management decided that 
they would split many of their large operational model output files into 
two or more separate files for the public, keeping "native" files for 
internal use.   The GFS 0.5 degree was originally a single file with 47 
layers when NOMADS began its distribution, but it was changed to be two 
files "gfs.tCCz.pgrb2fXX" containing most of the levels and variables 
and "gfs.tCCz.pgrb2bfXX" for  the remainder of levels and variables. 
These are located on the ftpprd part of WOC and represents the 
operational files.  NOMADS combined the files into one.  The NOMADS 
combined file on the WOC server is "gfs.tCCz.mastergrb2fXX"  (CC is the 
cycle 00, 06, 12, 18 and XX is the forecast time, 00,03,06,...180) with 
47 levels to 1mb.  Other data sets in operations have split up the large 
files, like NAM tiles, ensembles, RUC, etc...

> 1. Are there any GRIB-2 files that include the GFS fields above 10 mb, up to 1 mb?
> 2. Are there any GRIB-2 files that have GFS data on all 47 pressure levels, rather than the subset of 26?
See above. 
I see that the "grib filter" is pointing to the gfs.tCCz.pgrb2fXX and is 
not including the gfs.tCCz.pgrb2bfXX or so called "b" file.    For the 
moment one can download these files with "http" (fast ftp) or partial 
ftp application.  see 
http://www.cpc.noaa.gov/products/wesley/fast_downloading_grib.html

> 3) If the files in the above two questions exist, is there any archive of them? I am interested in running cases from May through October 2008. At COLA, we have an archive of 0.5 degree GFS GRIB-2 data covering this period, but the files we receive only have data up to 10 mb on 26 pressure levels.

An archive of the 0.5 degree data is at the NCDC NOMADS archive component:

In short

http://nomads.ncdc.noaa.gov

has archives.  The TDS server contains past 0.5 degree GFS datasets.

http://nomads.ncep.noaa.gov

is a high availability server which has real time 0.5 degree GFS 
datasets and keeps almost a months worth in the past.

One can slice, dice and area subset results including DODS service.

> I vaguely recall a discussion some time ago about 'a', 'b' and 'master' GFS files, and I am under the impression that the 'a' files are actually the 'gfs.tHHz.pgrb2fFH' files (i.e. the files we receive at COLA), but what data is contained in the 'b' and 'master' files, and are these files archived anywhere?
>
>   

See above and ...

The "b" files have the levels above 10mb and a number of other less used 
variables levels:

For example, in the combined (master) file, for absolute vorticity, 
ABSV: a wgrib2 of the NOMADS (nomads.ncep.noaa.gov) 00z cycle 12-h 
forecast has 47 levels:

  wgrib2  gfs.t00z.mastergrb2f12|grep ABSV
4:480919:d=2009011200:ABSV:10 mb:12 hour fcst:
9:1226170:d=2009011200:ABSV:20 mb:12 hour fcst:
14:2069533:d=2009011200:ABSV:30 mb:12 hour fcst:
19:2949641:d=2009011200:ABSV:50 mb:12 hour fcst:
24:3887686:d=2009011200:ABSV:70 mb:12 hour fcst:
31:5134569:d=2009011200:ABSV:100 mb:12 hour fcst:
39:6428137:d=2009011200:ABSV:150 mb:12 hour fcst:
46:7418523:d=2009011200:ABSV:200 mb:12 hour fcst:
53:8475610:d=2009011200:ABSV:250 mb:12 hour fcst:
60:9595080:d=2009011200:ABSV:300 mb:12 hour fcst:
67:10634463:d=2009011200:ABSV:350 mb:12 hour fcst:
74:11693243:d=2009011200:ABSV:400 mb:12 hour fcst:
81:12938461:d=2009011200:ABSV:450 mb:12 hour fcst:
88:14211704:d=2009011200:ABSV:500 mb:12 hour fcst:
95:15481207:d=2009011200:ABSV:550 mb:12 hour fcst:
102:16757633:d=2009011200:ABSV:600 mb:12 hour fcst:
109:18053967:d=2009011200:ABSV:650 mb:12 hour fcst:
116:19378813:d=2009011200:ABSV:700 mb:12 hour fcst:
123:20723478:d=2009011200:ABSV:750 mb:12 hour fcst:
130:22072796:d=2009011200:ABSV:800 mb:12 hour fcst:
137:23419984:d=2009011200:ABSV:850 mb:12 hour fcst:
144:24835886:d=2009011200:ABSV:900 mb:12 hour fcst:
151:26240464:d=2009011200:ABSV:925 mb:12 hour fcst:
158:27619065:d=2009011200:ABSV:950 mb:12 hour fcst:
165:28950832:d=2009011200:ABSV:975 mb:12 hour fcst:
172:30224047:d=2009011200:ABSV:1000 mb:12 hour fcst:
299:49165450:d=2009011200:ABSV:1 mb:12 hour fcst:
304:49745865:d=2009011200:ABSV:2 mb:12 hour fcst:
309:50324664:d=2009011200:ABSV:3 mb:12 hour fcst:
314:51081531:d=2009011200:ABSV:5 mb:12 hour fcst:
319:51881414:d=2009011200:ABSV:7 mb:12 hour fcst:
328:53202423:d=2009011200:ABSV:125 mb:12 hour fcst:
337:54523835:d=2009011200:ABSV:175 mb:12 hour fcst:
346:55839684:d=2009011200:ABSV:225 mb:12 hour fcst:
355:57289714:d=2009011200:ABSV:275 mb:12 hour fcst:
364:58677398:d=2009011200:ABSV:325 mb:12 hour fcst:
373:60018135:d=2009011200:ABSV:375 mb:12 hour fcst:
382:61415386:d=2009011200:ABSV:425 mb:12 hour fcst:
391:63022376:d=2009011200:ABSV:475 mb:12 hour fcst:
400:64686858:d=2009011200:ABSV:525 mb:12 hour fcst:
409:66387651:d=2009011200:ABSV:575 mb:12 hour fcst:
418:68112672:d=2009011200:ABSV:625 mb:12 hour fcst:
427:69707219:d=2009011200:ABSV:675 mb:12 hour fcst:
436:71326059:d=2009011200:ABSV:725 mb:12 hour fcst:
445:72966738:d=2009011200:ABSV:775 mb:12 hour fcst:
454:74608150:d=2009011200:ABSV:825 mb:12 hour fcst:
463:76316562:d=2009011200:ABSV:875 mb:12 hour fcst:

wgrib2  gfs.t00z.mastergrb2f12|grep ABSV|wc -l
47

Eric Altshuler wrote:
> Dear Jordan,
>
> I am interested in running the WRF model with the model top at 1 mb, using 0.5 degree GFS initial and lateral boundary conditions. To be able to have the model top this high, GFS data is required up to 1 mb. As far as I'm aware, 0.5 degree GFS data is only available up to 10 mb in the standard 'gfs.tHHz.pgrb2fFH' GRIB-2 files, but I have discovered that data is available up to 1 mb through the "OpenDAP" interface on nomads.ncep.noaa.gov (the data is also available on 47 pressure levels instead of the usual 26). However, for running WRF, I need GRIB data. I have three questions regarding this:
>
> 1. Are there any GRIB-2 files that include the GFS fields above 10 mb, up to 1 mb?
>
> 2. Are there any GRIB-2 files that have GFS data on all 47 pressure levels, rather than the subset of 26?
>
> 3. If the files in the above two questions exist, is there any archive of them? I am interested in running cases from May through October 2008. At COLA, we have an archive of 0.5 degree GFS GRIB-2 data covering this period, but the files we receive only have data up to 10 mb on 26 pressure levels.
>
> I vaguely recall a discussion some time ago about 'a', 'b' and 'master' GFS files, and I am under the impression that the 'a' files are actually the 'gfs.tHHz.pgrb2fFH' files (i.e. the files we receive at COLA), but what data is contained in the 'b' and 'master' files, and are these files archived anywhere?
>
> Thanks in advance for your attention.
>
> Best regards,
>
> Eric L. Altshuler
> Assistant Research Scientist
> Center for Ocean-Land-Atmosphere Studies
> 4041 Powder Mill Road, Suite 302
> Calverton, MD 20705-3106
> USA
>
> E-mail: ela at cola.iges.org
> Phone: (301) 902-1257
> Fax: (301) 595-9793
>
> P.S. To the WRF user community, has anyone tried to run WRF with the model top as high as 1 mb? Is it advisable to run WRF with a vertical domain extending this far into the stratosphere? The default WRF model top is at 50 mb, and I've always wondered why the default model top is so low. Is there a specific reason for this?
>   

-- 
______________________________________________________________
Dr. Jordan C. Alpert, NCEP/Environmental Modeling Center, 
Room 204, 5200 Auth Road, Camp Springs, MD 20746
Jordan.Alpert at noaa.gov  301-763-8005x7205 fax:301-763-8545