Thanks for your reply. I was not too clear on the
manifold dimension vs. domain dimension,
but I went back and read the Developer's guide
and now it all makes sense. Using manifold dimension
of 2 in the Gridded3D constructor fixes the problem.
Also, I am not clear on how this solution works:
> If you want the elevation to be part of your data, use the
> (elevation, ((radial,gate)->(x,y,z,gatevalue)))
> This is a Tuple combining a Real elevation with your
> original FlatField. Then include ScalarMaps
> radial->Longitude, gate->Radius and elevation->Latitude.
Is there something that you can point me to
that explains this? Specifically, I don't understand why
gate is radius and elevation is latitude other than
that the units and ranges match ...
Ideally, I would like to use this solution, because then
I don't really have to compute x,y and z ...
Also, since these are not really longitude and latitude,
will I be able to overlay another cone (at
a different location on the earth) and still have things
work correctly? The x,y,z solution works, because
my x,y,z is in a co-ordinate system with the center
of the earth as origin (not the apex of the cone).
> But use a Gridded3DSet with manfold dimension = 2.
> new Gridded3DSet( domainTuple, xyz, num_radials, num_gates )
this works, and I even understand why :)
However, I still don't know how to turn off interpolation
(i.e. do NEAREST_NEIGHBOR).