Handle ensembles for cross-wavelet method

[khider]

@CommonClimate Computation-wise, it's feasible (albeit slow and processor consuming) but how could a human interpret this 4D data?

[CommonClimate]

I think we need to decide what should be done if 2 ensembles meet in cross-wavelet space... I'm certainly not advocating for brute force, for the reasons that you say. One approach would be to use a subset of series from ensemble 1, XWT it to a subset of series from ensemble 2, and have some sensible null underneath. But this is not a priority for the hackathon, so I suggest we punt for now.

[khider]

We also need to define a "pass/fail" test and a new visualization to see how many times the coherence matches across the age ensemble members.