Incorporate these notes

[DaniJonesOcean]

From today's chat with @maikejulie:

• Interesting that the tSNE manifold and the PC manifolds are similar
• One wing is the ACC, the other wing is the gyre and "ventilated/gyre edge" class
• The transition class is between the two
• Crossing from the ACC, across the transition bridge, to the gyre, is the conversion of ACC/CDW to gyre water
• What is the probability of transitioning from the ACC wing to the gyre wing?
• This make the most sense in a barotropic framework. Less so if it's baroclinic
• But baroclinic structure would imply shear...which probably only happens in specific places
• Should we try agglomerative clustering methods?

[DaniJonesOcean]

Added to the discussion section, to be expanded on and turned into an actual paragraph