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

19991007: followup on MRF grids (2 vs 3)



Jim,

There have been no ensemble runs, and I have not seen the 0-168 hour
grid #2 since the Cray outage.

Your earlier statement that you wereable to get them from OSO made
me wonder if you had seen them today. If not, then this seems consistent.

Esentially I am sending everything to dcgrib and using the ### model grid
number template in the file name so that grid 2 products are written to
one file and 3 to another.

You can however selectively use the grid number field in the identifier
string in pqact to separate out just grid 2.

As I mentioned in the previous message, the general form of the
product identifiers will be:


/public/ncepb/mrf/model.YYMMDD/drfmr.THHZ.PGrbFff.opt 
!grib/ncep/SPEC_126_FNL/#00x/YYYYMMDDHHNN/Ffff/parm/level! #####

Where the "model" string is either mrf or ens. Since the model grid number
is provided in the !...! delineated information, a pattern for just grid 2
might look like:

NMC2    mrf/(mrf|ens).(......)/drfmr.T(..)Z.PGrbF(.*) !.*/#002/.*!

This should allow you to isolate just those products on grid #002.
Note that the string before the !...! is just the original file name
from the OSO computer. The information between the !...! is
information to identify the individual grib menber within the original
file as described at: http://www.unidata.ucar.edu/staff/chiz/ncep_idd/

Steve Chiswell
Unidata User Support


On Thu, 7 Oct 1999, W James Steenburgh wrote:

> Steve:
> 
>       Thanks for the response.  The complexities of the MRF processing
> are something else.  Essentially our problem is this.  We have just started
> receiving the MRF grids and have only been receiving hours 180 and up for
> the 2 grid.  From your reply, I suspect this is because the ensemble
> MRF runs are not being run so that the grid 2 data is unavailable at
> NCEP.  Is this the case, or do we possibly have a
> problem processing on our end?
> 
> Thanks,
> 
> Jim
> 
> 
> ----- Begin Included Message -----
> 
> >From address@hidden Thu Oct  7 14:46 MDT 1999
> Date: Thu, 7 Oct 1999 14:46:14 -0600
> From: Steve Chiswell <address@hidden>
> To: W James Steenburgh <address@hidden>
> cc: address@hidden
> Subject: 19991007: followup on MRF grids (2 vs 3)
> MIME-Version: 1.0
> 
> 
> 
> Jim,
> 
> Both grid 2 and 3 are available on NMC2 from 0-168 hours,
> and just grid 2 thereafter. Note also, that due to the fire
> on the Cray at NCEP, the current run for today was through 240 hours
> rather than 384 hours.
> 
> As you mention, the file naming changes for the grid 2 files from NCEP
> at the 184 hour point. Where from 0-168 they have the .2 identifier
> and thereafter it is omitted.
> 
> The .2 files were tagged as originating from the ensemble MRF runs,
> which hasn't been available since the fire- and I don't see those
> files in the OSO ftp directory at:
> ftp://140.90.6.103/ncepb/mrf/mrf.991007/
> 
> If you see the .2 files somewhere else today, then please let me know and
> I can contact NCEP to see why they aren't getting posted.
> 
> I tried to outline the MRF grid 2 vs 3 file naming at:
> http://www.unidata.ucar.edu/staff/chiz/ncep_idd/mrf_files.html
> 
> For 0-168 grid 2, you should see the headers:
> /public/ncepb/mrf/ens/YYMMDD/drfmr.THHZ.PGrbFff.2
> 
> then for 184-384:
> /public/ncepb/mrf/mrf.YYMMDD/drfmr.THHZ.PGrbFff
> 
> Steve Chiswell
> Unidata User Support
> 
> 
> On Thu, 7 Oct 1999, W James Steenburgh wrote:
> 
> > Steve:
> > 
> >     Oops, a couple of typos in that previous email concerning 
> > the MRF 2 grid.  The data is only available at and after forecast
> > hour 180, not 184.  As a quick note, on the OSO server, grib
> > files for all times before 180 have a ".2" suffix.  I am not sure
> > if this would have anything to do with why we are not getting the
> > data via the ldm feed, but perhaps it plays a role.
> > 
> > Thanks again,
> > 
> > Jim
> > 
> 
> 
> 
> ----- End Included Message -----
> 
> 
>