[IDV #UNX-198008]: grib2 data appears to be missing, but it's not

Becky-

The problem is that the missing value for the u and v winds is
being defined as 0, so when one of the components is 0, it gets
set to NaN and is thus missing.   Can you check with Bob to see
what he is setting the missing value to in his grib files?  I'll
check with our GRIB person to see if there's an error in the
decoding.

Don Murray

> Never mind the bundle, I figured it out.
> 
> > We got a report from a McIDAS-V user that their loop of grib2 data
> > showed missing data, even though the file has no missing data.  Here's a
> > link to the loop he sent -
> > http://cimss.ssec.wisc.edu/model/realtime/cras45_NA/grib2/
> >
> > We were able to duplicate this in both McIDAS-V and the nightly build of
> > IDV.  Any idea what's going on?
> 
> the problem is that there are missing points in the u component of
> the wind field, so when absolute vorticity is calculated at those
> points, the values end up as missing since there is no u component.
> I'm not sure how McIDAS fills in the gaps at the grid points, but it
> seems to me that if you don't have wind values at a point, you can't
> calculate the vorticity.   I am surprised that there are vectors plotted
> at some of the points with a missing v, though.  Although that might
> be due to a weighted average sampling somewhere along the way.
> 
> Attached is a plot of the wind vectors over the absolute vorticity field
> overlain with the vector vield and value plots of the u (top) and v (bottom)
> components.  Notice that the holes line up with the areas where there
> are missing u values.
> 
> So, the question is why are there gaps in the  u wind field.  Is it in
> the data or a problem in the grib decoding layer?  I'll look a little
> closer at the data tomorrow, but if you have investigate on your end
> also, that would help.  It would be interesting to know what McIDAS
> does with missing data.
> 
> > Here are screen shots of McV, McX, and -X output of grid statistics from
> > the GRDINFO command.  If you'd like to grab it, I put the grib2 file in
> > ftp://ftp.ssec.wisc.edu/pub/incoming/20080712.00z.cras45_NA.grib2
> 
> Does McIDAS ignore missing values when it plots the statistics?
> 
> > I think that's everything, but let me know if you need any more
> > information from me.
> 
> Don
> 


Ticket Details
===================
Ticket ID: UNX-198008
Department: Support IDV
Priority: Normal
Status: Open