Ideas for leveraging xarray custom indexes once implemented

[jthielen]

A recent issue raised in xarray struck me as a particularly good example of where MetPy may be able to extend xarray's indexing functionality once its ongoing refactor is completed. This connects to a prior idea floated on the telecon about some form of "physically aware" vertical index that could aid layer-based operations.

While on the topic, geographic coordinate indexes for projected data (with grid factors and the like) would also be very useful in MetPy, though this kind of CRSIndex or GeoIndex serves a broader audience may be a better fit somewhere else/upstream (such as cf-xarray, xgcm, or geoxarray).

In any case, while the work on the indexes is ongoing upstream in xarray, I figure this discussion could serve as a place to log any ideas that come up on MetPy's end.