[GEMPAK #WIN-831141]: bug in gdprof for soundings at a grid point

David,

After looking at you data, I can see that the problem with missing data values
is not that they are filtered, but that the adjacent grid point (103;67) has
a surface level of 923mb whereas the point in question (102;67) is 981mb. As
a result, the interpolated location with very small roundoff when calculating 
to 
lat/lon to create the new interpolated grid, and then back to the new grid 
point space 
so that the magnitude of vector components can be interpolated first before
if leading to missing values at the levels above the desired grid point.
This would be the expected result if you were asking for 102.0001;67 but
since you are specifying the exact grid point location, you don't want to lose 
those levels 
through the new grid interpolation scheme that I provided you before.

I'll have to re-design so that the precision of the transformation does not 
create this
problem.I'll let you know when I have a fix....but just fyi, the newly posted 
tarfile for 10.1
has the code as I sent you previously- so this problem will still exist.

Steve Chiswell
Unidata User Support





> David,
> 
> Thanks for the detailed information...I'll see whats up!
> 
> Steve Chiswell
> Unidata User Support
> 
> 
> > Steve,
> >
> > There is unfortunately still a bug in gdprof when using gpoint =
> > @x;y.  The winds that are displayed are filtering in the vertical,
> > leaving out a lot of low-level, very important winds.
> >
> > Using the NAM 104 grids, you cannot see this, but my high vertical
> > resolution MM5 grids show it distinctly.
> >
> > I have posted latlon.gif, xypoint.gif, latlon_at_x102y67.gif, and
> > mmout_d1.2007021312.f006.0000.gem at
> > http://www.atmos.washington.edu/~ovens/gdprof_bug/
> > You can loop the graphics using
> > http://www.atmos.washington.edu/~ovens/loops/wxloop.cgi?/home/disk/user_www/ovens/gdprof_bug+all
> >
> >
> > Here is how I produced latlon.gif:
> > GPOINT   = 47.69;-122.24
> > GDATTIM  = LAST
> > GVCORD   = pres
> > GFUNC    = tmpc
> > GVECT    = wnd
> > GDFILE   = mmout_d1.2007021312.f006.0000.gem
> > LINE     = 2/1/3/2/2
> > MARKER   =
> > BORDER   = 32;32;32/1;1/1;1
> > PTYPE    = stuve
> > SCALE    = 0
> > XAXIS    = -80/40/10/;1
> > YAXIS    = 1050/700/50/;1
> > WIND     = bk32/1.2/1.5/0112
> > REFVEC   = 10
> > WINPOS   = 2
> > FILTER   = no
> > TITLE    = 32/3/lat lon 47.69 -122.24
> > PANEL    = 0
> > CLEAR    = no
> > TEXT     = 1/21//111//l/hw
> > DEVICE   = gf|latlon.gif|900;700
> > OUTPUT   = y
> > THTALN   = 23/1/1
> > THTELN   = 8/3/1
> > MIXRLN   = 7/1/1
> >
> > and for xypoint.gif, same as above but with
> >
> > GPOINT = @102;67
> > DEVICE   = gf|xypoint.gif|900;700
> > TITLE    = 32/3/X Y point 102 67
> >
> > and for latlon_at_x102y67.gif, same as above but with
> >
> > GPOINT = 47.65601;-121.995087
> > DEVICE   = gf|latlon_at_x102y67.gif|900;700
> > TITLE    = 32/3/latlon 47.65601;-121.995087 = X Y exact point 102 67
> >
> >
> > Thanks,
> >
> > David
> > --
> > David Ovens          e-mail: address@hidden
> > Research Meteorologist    phone: (206) 685-8108
> > Dept of Atm. Sciences      plan: Real-time MM5 forecasting for the
> > Box 351640                        Pacific Northwest
> > University of Washington          http://www.atmos.washington.edu/mm5rt
> > Seattle, WA  98195               Weather Graphics and Loops
> > http://www.atmos.washington.edu/~ovens/loops
> >
> > On Thu, Feb 01, 2007 at 04:10:11PM -0700, Unidata GEMPAK Support wrote:
> > > David,
> > >
> > > Previously I modified the internal navigation to fix the grid 
> > > interpolation problem you were seeing
> > > for vectors between grid points by creating a new grid subregion centered 
> > > on the gpoint (eg so
> > > the gpoint location would be on a remapped grid point and the vectors 
> > > wouldn't be using interpolated
> > > components between grid points).
> > >
> > > In doing that, I created a new internal navigation, which is conflicting 
> > > with the later query of
> > > the user point back into the original navigation (now the new grid is a 
> > > 3x3 grid about the GPOINT
> > > rather than the original model domain, so your GPOINT=@x;y is not defined 
> > > in the new space).
> > > In fact, the output you are seeing for the GPOINT=lat;lon case should be 
> > > showing a pair of
> > > grid points in the new 3x3 domain now instead of the original grid if you 
> > > look at that too!
> > >
> > > I'll have to save the original navigation, and restore it after I do the 
> > > interpolation
> > > (I'll apply this to the gdpoint program as well since I made the same 
> > > change there).
> > >
> > > I'll get that done tomorrow (trying to beat the snow and subzero temps 
> > > home today ;-)
> > >
> > > Thanks for pointing this out.
> > >
> > > Steve Chiswell
> > > Unidata User SUpport
> > >
> > >
> > >
> > >
> > > > I'd like To follow up with some info on this bug.  It was introduced
> > > > in 5.9.1, 5.8.4 still worked fine.
> > > >
> > > > David
> > > >
> > > > On Wed, Jan 31, 2007 at 04:11:29PM -0700, Unidata GEMPAK Support wrote:
> > > > >
> > > > > David Ovens,
> > > > >
> > > > > Your Ticket has been received and a member of our staff will review 
> > > > > it and reply accordingly. Listed below are details of this Ticket. 
> > > > > Please make sure the Ticket ID remains in the Subject: line in all 
> > > > > future correspondence regarding this inquiry.
> > > > >
> > > > >       Ticket ID: WIN-831141
> > > > >       Subject: bug in gdprof for soundings at a grid point
> > > > >       Department: Support GEMPAK
> > > > >       Priority: Normal
> > > > >       Status: Open
> > > > >
> > > > >
> > > > >
> > > > > Please do let us know if we can assist you any further,
> > > > >
> > > > > Unidata User Support
> > > >
> > > >
> > >
> > >
> > > Ticket Details
> > > ===================
> > > Ticket ID: WIN-831141
> > > Department: Support GEMPAK
> > > Priority: Normal
> > > Status: On Hold
> >
> >
> 


Ticket Details
===================
Ticket ID: WIN-831141
Department: Support GEMPAK
Priority: Normal
Status: Closed