Re: [conduit] 20070913: CONDUIT GRIB1 to GRIB2 transition

On Thu, 2007-09-13 at 12:22 -0500, Kevin W. Thomas wrote:
> >CONDUIT data recipients,
> >
> >Re:  GRIB1 to GRIB2 transition notification which has been provided in
> >numerous formats, such as:
> >
> >        All GRIB1 data currently residing on the NCEP & TOC FTP servers
> >        will be removed on 28 January 2008.  The schedule for the GRIB1
> >        to GRIB2 conversion for NOAAPORT data has not yet been
> >        established.  A schedule will be published as soon as it becomes
> >        available.  It is anticipated that no GRIB1 data will be removed
> >        from NOAAORT prior to January 2008
> >
> >We would like to begin a timetable to replace the remaining GRIB1 format
> >data sets
> >in the CONDUIT data stream with the GRIB2 versions that are available on
> >the NCEP ftp servers which are the source of the CONDUIT data stream.
> >
> >For the past 15 months, RUC data in both GRIB1 and GRIB2 format has been
> >delived
> >by CONDUIT. At this time, we would like to know if anyone still needs
> >the GRIB1 format data. By removing the redundant GRIB1 RUC data, we will
> >free up some room
> >to alternately add GRIB2 versions of NAM data sets for a brief period of
> >overlap (approximately  1 week in duration) and then remove the
> >redundant GRIB1 versions.
> 
> 1 week is a bit short.  2 weeks would be more desireable.
> 
> The CAPS code has recently had GRIB2 supported added, though it is only 
> lightly
> tested.  I suspect other realtime users might have similar issues, and may 
> need
> more time if "surprises" occur during the transition.  Note for CAPS this is
> grid #212.
> 

Kevin,

I think we can do at least 2 weeks, and probably start out with the grid
216 which
probably isn't used by most, then 104, then 212. The main thing is we
have to get
on to NCEP's schedule to set some dates and we don't want to run into
the holidays and then AMS in January. NCEP still has towork around
critical weather days if they pop up.


> >Current NAM data sets being delivered in GRIB1 are "awipak" grid #216,
> >"awip3d" grid #212, and "grbgrd" grid #104. The 12km NAM data is already
> >delivered in GRIB2 so will not be affected.
> 
> Is this the full 12km NAM tiled data set, including all 3D variables?  I'm
> currently ftp'ing both the GRIB1 and GRIB2 versions from NCEP.  Once tested,
> I'll transition the CAPS software to use GRIB2, which will save me a lot of
> disk space.

This is the composite grid 218 data set, and not the 54 tiles. It
contains approximately 177 fields per forecast hour file, whereas I
counted 429 fields in the tiles,
so it is not a complete replacement for the grid 212 data which have 616
fields.

Steve Chiswell
Unidata User Support

> 
> >After transitioning the NAM data sets, the GFS 1.0 degree (grid #003)
> >data set will
> >be considered. Options for this data set are:
> >   1) Transition the GRIB1 version to GRIB2
> >   or
> >   2) Remove this data set entirely as a lower resolution redundant data
> >set of the 
> >        0.5  degree GFS data already being delivered in GRIB2 for the
> >same levels,
> >       parameters, forecast hours.
> >
> >Other pending additions to CONDUIT include additional GEFS/TIGGE
> >ensemble parameter members (approximately 11GB per day).
> >
> >Your feedback, as always, is welcomed!
> >
> >-- 
> >Steve Chiswell <address@hidden>
> >Unidata User Support
> >_______________________________________________
> >conduit mailing list
> >address@hidden
> >For list information or to unsubscribe, visit: 
> >http://www.unidata.ucar.edu/mailing_lists/ 
> 
>       Kevin W. Thomas
>       Center for Analysis and Prediction of Storms
>       University of Oklahoma
>       Norman, Oklahoma
>       Email:  address@hidden
-- 
Steve Chiswell <address@hidden>
Unidata