Interpolation subroutine

[erickjomp]

The interpolation subroutine has been fixed to prevent NaN values when the desired interpolation level matches a data level.

[kafitzgerald]

Thanks for the PR!

I'm a little surprised this hasn't come up as an issue in the past since this particular code came from back in NCL and has likely seen a good bit of use. Given that, I'd like to at least better understand what's going on here before getting this merged. I'll try to get to this soon though.