Use scipy.interpolate in message_ix.tools.add_year

About 500 of the ~1000 lines in `tools.add_year` are devoted to two functions, `interpolate_1d()` and `interpolate_2d()`, and one utility function, `intpol()` to do interpolation.

These can likely be slimmed down, and maybe made more performant, by fully using lower-level methods in pandas and scipy, e.g. [`scipy.interpolate.griddata()`](https://docs.scipy.org/doc/scipy/reference/tutorial/interpolate.html#multivariate-data-interpolation-griddata).

Start in message_ix.tools.add_year and read interpolate_1d(), interpolate_2d(), and intpol() to understand their current interpolation behavior. Compare the existing logic with the referenced pandas and scipy.interpolate.griddata() APIs. Done means the interpolation implementation is slimmer or more performant while preserving the existing behavior.

Written by the indexing model from the issue text.