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

[IDV #CQH-644947]: Missing value is 1e15, beyond Java range



Greetings Brian,

Unfortunately, no ncml fix this time. I tried a sample dataset:

http://goldsmr3.sci.gsfc.nasa.gov/opendap/MERRA/MAI3CPASM.5.2.0/2012/09/MERRA300.prod.assim.inst3_3d_asm_Cp.20120904.hdf

and loaded in the field:

3D grid -> Air Temperature

Spatially subset over the US, stride in X and Y as every fourth point, and
created a Value Plot. As expected, Air Temperatures over Colorado for the
1000 mb pressure surface do not show up, indicating the IDV correctly 
interprets the missing value (it's a 32-bit floating point, so there isn't
anything to worry about in terms of resolution). However, if you go to the
display control for the Value Plot and change the level to, say, 500 mb, then
we do see values over Colorado.

In terms of the isosurface, I'm not sure how the IDV handles missing values
when trying to construct an isosurface. My initial though is that it does not
try to interpolate values to make a nice smooth surface, but rather create
empty holes in the isosurface...at least that's what it looks like when I
plot an isosurface and overlay it was the value plot.

> Hello IDV folks.
> 
> I use a gridded dataset and the missing values show up as huge numbers
> 
> MERRA uses 1e15 for missing data, including below ground pressure surfaces
> (they don't extrapolate data below ground).
> However, that's beyond the resolution of a float,
> so the netCDF-Java library doesn't pick that value up when comparing.
> 
> It messes up isosurface plots, since there is always an isosurface lying
> along the topography.
> 
> 
> Don said: "That might be worth a support-idv request." So here it is.
> Any ideas? An ncml fix?
> 
> 
> Thanks!
> Brian
> 
> Brian Mapes
> address@hidden<mailto:address@hidden>
> 
> 
> 
> 
> 


Ticket Details
===================
Ticket ID: CQH-644947
Department: Support IDV
Priority: Normal
Status: Open