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

[IDV #PUE-499201]: IDV 2.4b3: Problem plotting 3D isobaric surface using GFS GRIB2

Hi Dave-

I thought I already answered this, but it is not showing up in the
system.  So, hopefully if you get two copies, they will have the same
answer.

> I'm running IDV 2.4b3 (an an Intel iMac running Leopard) and trying to
> make one of my favorite plots, a 3-D pressure surface, using a GFS
> GRIB2 model analysis as my data source. It's not working, though it
> works for GFS, NAM, and RUC model files in GRIB format that I've also
> got available to me locally.
> 
> In particular, I get a Mercator map background (I think), a display
> label ("3D Pressure Surface - Topography etc."), a legend entry ("3D
> Pressure Surface - Topography"), and a default color table, but no
> isosurface plot.
> 
> The obvious problem with the color table is that infinity is assigned
> to the left-most color and negative infinity is assigned to the right-
> most color. Upon closer examination of the color table, the pointers
> at each end of it are labeled "missing". Assigning reasonable values
> to these (such as 8,000 and 10,000 meters for the 300 mb pressure
> surface) changes nothing in the display.

The GridUtil.sliceAtLevel uses the native units of the grid's vertical
dimension to do the slicing.  For the GRIB2 data, the unit is Pa instead
of hPa as in Grib1.  So, you are trying to slice at 300 Pa which is
3 millibars.  Since there is no data at that level, you end up with a
grid of missing values.  If you use 30000, you'll get the display you
want.
 
> I've got the following entry in my Jython User's library:
> 
> ## Imported jython from bundle
> 
> #
> # A 3-D surface of pressure.
> #
> def threeDPressure(z, level):
> import ucar.unidata.data.grid.GridUtil as gu
> a = gu.sliceAtLevel(z, float(level))
> return gu.make2DGridFromSlice(a)
> #
> 
> I've attached a copy of the 3-D pressure surface formula
> (3D_PressureSurface.xml).
> 
> I've also attached a bundle that attempts to create the display. It
> uses a GFS GRIB2 analysis file for 12Z Dec 11, and it will hang around
> on our system only for about five days--after that, a more recent file
> will have to be used (with name of the form yymmddhh_gfs2_F00.grb,
> available in http://funnel.sfsu.edu/data/GRIB).
> 
> The formula:
> 
> The bundle:
> 
> (Note that the data source is a large file.)
> 
> In contrast, here's the same type of plot made (successfully) with a
> GFS GRIB file (K domain):

It's the value you use in relation to the vertical dimension.  If you were
slicing a grid with a height domain, you'd need to use a value in that
data range.

Another approach is to use the 3D Surface->Color Shaded Plan View over 
Topography
display.  When prompted for the topography field, select the geopotential height
again.  At present, you can't select one level, but you could reduce the 
download
time by selecting two contiguous levels (e.g. 300 hPa and 350 hPa).  Be sure
to select the same level range for the popup topography field selector or you'll
get an error.  When you use the sliceAtLevel, it has to download the entire grid
before slicing so you gain some speed/performance selecting a subset of the
levels.

Don Murray

Ticket Details
===================
Ticket ID: PUE-499201
Department: Support IDV
Priority: Normal
Status: Open