Specified drifts on universal kriging

[ghost]

I believe spec_vals.flatten()[j] should be instead of spec_vals.flatten()[i] on this line. Let me know if it is correct or not. If so, same line should be corrected on uk3d class.

https://github.com/bsmurphy/PyKrige/blob/db07202a4ff1cbe9a2211db5976a1511ab34fc21/pykrige/uk.py#L841

[rth]

Thanks for the suggestion, @landsite. I'm not very familiar with that part of the code. Is there a test we could write to validate this?

[MuellerSeb]

We could run a comparison against the GSTools Universal Kriging routine.