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

[CONDUIT #RGY-826441]: GFS (grib) forecast hours

Ryan,

I provided a software update for NCEP to install this morning, and the 18Z gfs
runs should have the correct forecast hour interpetation. I'm waiting to
see the 18Z gfs data come across to verify this! You should
be able to avoid having to rewrite your pqact lines.

The old tgsv32 server ran with the system clock set to UTC, so this bug in the
insertion code was never tripped. The new NCEP servers run with the clocks set 
to
the EDT5EST time zone so the bug in code suddenly appeared this week.
The GRIB2 processing code I wrote doesn't have the bug, but the
grib1 processing code uses the wgrib package to extract the time
which is "verification time" and then the subtraction of current time to get
the forecast hour. With the verification time after the daylight savings time,
the subtraction was off by one. I fixed this so that the program runs in UTC 
time now
regardless of the local time zone. Again, this bug was always in the code but
circumstances had prevented its appearance.

Steve Chiswell
Unidata User Support

> 
> Steve,
> 
> Thanks for the response and explanation....i guess that makes sense,
> though this hasn't happened previously.  what's your guess as to how
> long this adjustment to the feed would take and/or even possible for
> subsequent runs.  i am just trying to gauge whether it's worth it to
> creatively alter my scripts or not.  thanks again.
> 
> aloha,
> ryan
> 
> 
> On Oct 24, 2006, at 6:50 AM, Unidata CONDUIT Support wrote:
> 
> > Ryan,
> >
> > This is a daylight savings time change issue with the Linux time
> > call in the wgrib software. I'll see if there is anything that can
> > be done
> > during this time where the forecasts are for times past the time
> > change.
> >
> > Steve Chiswell
> > Unidata User Support
> >
> >
> >
> >
> >>
> >> I've noticed since the weekend that the forecast hours of the GFS
> >> grib file that come through the ldm  are now starting to drift off by
> >> an hour.  for example: previously the forecast hours were ever 3
> >> hours out to 180 hours and 12 hours after that.  now it continues to
> >> do that most of the way, but sometime in between 0-180 there is a 4
> >> hour jump.
> >>
> >> forecast hours: ...., 108, 111, 114, 118, 121, .....
> >>
> >> 2006102406 run had that jump at 114 hour
> >> 2006102400 run had that jump at 120 hour
> >> 2006102318 run had that jump at 126 hour
> >> 2006102312 run had that jump at 132 hour
> >> 2006102306 run had that jump at 138 hour
> >> 2006102300 run had that jump at 144 hour
> >>
> >> that continues back to the 20061019 runs and i think you see the
> >> pattern there.
> >>
> >> will this be a permanent change to the GFS grib feed once this jump
> >> gets down toward 0, or will this be fixed.  i would hate to change my
> >> scripts that look for specific forecast hours to process and/or
> >> display this data only for it to keep changing and/or change back.
> >>
> >> thanks for your help and i hope the above made sense.
> >>
> >> aloha,
> >> ryan lyman
> >>
> >>
> >> ---------------------------------------------------------------------
> >> ---
> >> -----
> >> Ryan Lyman                            Mauna Kea Weather Center
> >> Institute for Astronomy           Forecast Meteorologist
> >> 640 North A‘ohoku Place      Tel: (808) 932-2323
> >> Hilo, Hawaii 96720-2700      Fax: (808) 933-0737
> >> ---------------------------------------------------------------------
> >> ---
> >> -----
> >>
> >>
> >>
> >>
> >>
> >>
> >
> >
> > Ticket Details
> > ===================
> > Ticket ID: RGY-826441
> > Department: Support CONDUIT
> > Priority: Normal
> > Status: Closed
> >
> 
> ------------------------------------------------------------------------
> -----
> Ryan Lyman                            Mauna Kea Weather Center
> Institute for Astronomy           Forecast Meteorologist
> 640 North A‘ohoku Place      Tel: (808) 932-2323
> Hilo, Hawaii 96720-2700      Fax: (808) 933-0737
> ------------------------------------------------------------------------
> -----
> 
> 
> 
> 
> 
> 


Ticket Details
===================
Ticket ID: RGY-826441
Department: Support CONDUIT
Priority: Normal
Status: Closed