On Mar 4, 4:14 pm, David Fanning <n...@[EMAIL PROTECTED]
> wrote:
> Kenneth P. Bowman writes:
> > > Compute the "interpolation coordinates" from the original grid
> > > j = VALUE_LOCATE(y_original, y)
> > > yj = j + (y - y_original[j])/(y_original[j+1] - y_original[j])
>
> To tell you the truth, I can't get this to work at all. :-(
>
> IDL> lat = [-87.5, 50, 25, 0, 30, 45, 64, 87.5]
> IDL> y = Scale_Vector(findgen(7), -87.5, 87.5)
> IDL> j = Value_Locate(lat, y)
> IDL> yj = j + (y - lat[j])/(lat[j+1] - lat[j])
> % Program caused arithmetic error: Floating illegal operand
> IDL> print, yj
> 0.000000 0.212121 0.424242 3.00000 3.97222 5.70175 -NaN
>
Hi,
I think it comes a little clearer for me to leave VALUE_LOCATE out of
it. Instead simply normalize the output coords into the input
coordinate space. Something like this...
ilon = Scale_Vector(findgen(10), 10.0, 70.0)
ilat = [-87.5, 50, 25, 0, 30, 45, 64, 87.5]
olon = Scale_Vector(findgen(12), 0.0, 90.0)
olat = Scale_Vector(findgen(7), -87.5, 87.5)
ix = (olon - min(ilon)) / (max(ilon) - min(ilon))
iy = (olat - min(ilat)) / (max(ilat) - min(ilat))
idim = [n_elements(ilon), n_elements(ilat)]
iZ = HANNING(idim[0], idim[1])
oZ = INTERPOLATE(iZ, ix, iy, /GRID, MISSING = !VALUES.F_NAN)
It doesn't address the question about assuming a flat dataspce when
the data are really in spherical coordinates. You might have a good
solution to transform the coordinates to a projected space then
interpolate, but I really don't know for sure. Taking the data as
delivered, it is not projected at all - it just comes with two of the
three parts needed to specify spherical coords plus a data value. The
radius is whatever you want (in this case I was hoping to be on
earth).
Ben
|
