Sampling is incorrectly set when using DHRealGrid() class.
ij264 opened this issue · comments
I'm pretty sure that this is correct.
For DH grids with extend=False, nlat and nlon both are even. nlat is equal to N, which is equal to 2L+2. The size of the array is either NxN or Nx2N.
Since your input grid has nlat=91, this means that you are using grids with extend=True, which includes the redundant values at 360 E longitude and 90 S. For this case, nlat = N+1, and nlon is either N+1 or 2N+1. Thus the input nlon should be either 91 or 181.
Since your input grid doesn't conform to these rules, it isn't a valid grid that can be used with the Driscoll and Healy quadrature routine.