[THREDDS #KNL-571421]: GFS Best Time Series not "best'

Hi Don,

I see the same thing. For example, if we compare the "best" for the GFS 1.0 
degree for

2013-09-08 18Z, 21Z

with the 2013-09-08 run using

18Z (analysis), 21Z (3 hour forecast)

the grids turn out to be different (I'm comparing MSLP_Eta_Model_Reduction),
even though they *should* be the same. Lansing pointed me to a jira ticket
that describes something similar, except it comes from an FMRC rather than
a GRIB feature collection:

https://bugtracking.unidata.ucar.edu/browse/TDS-470

I'm going to check a few other "best" collections to see if the same thing
is occurring elsewhere when the catalog is not partitioned by time.

Sean

> Hi Don,
> 
> I'll take a look in the logs and see if any error messages are in there.  
> I'll also float the runtime information feature.
> 
> Thanks,
> Lansing
> 
> > Hi-
> >
> > The GFS 0.5 degree Best Time series on thredds.ucar.edu seems to have
> > some missing timesteps of bad indexes.  If you look at the IDV bundle:
> >
> > http://www.esrl.noaa.gov/psd/repository/entry/get/arglobe.xidv?entryid=26dc3204-a783-49b6-9466-fd5c78e03c47
> >
> > you'll see that the animation jumps around a bunch. This is a case where
> > it would be helpful to have the correct runtime information for each
> > timestep.
> >
> > Don
> > --
> > Don Murray
> > NOAA/ESRL/PSD and CIRES
> > 303-497-3596
> > http://www.esrl.noaa.gov/psd/people/don.murray/
> >
> >
> 

Ticket Details
===================
Ticket ID: KNL-571421
Department: Support THREDDS
Priority: Normal
Status: Open