Vigilant differentiation between PDO the concept & PDO the measure is necessary to generate any hope of advancing discussion.
The problem isn’t so much with the concept as with the standard measure. The measure’s suboptimal and because of this it’s causing problems with human pattern recognition. The algorithmically built-in assumption that the thing in concept doesn’t change shape over time is the problem people aren’t facing. The basis doesn’t flex with reality.
Adaptive bases are needed. With them it will be possible to tune the aggregation to reveal the evolution of the attractor. This attractor doesn’t just sit still. It flexes.
Circulation is a function of both base background and gradients. Early in the 20th century, the base background was fundamentally different. Careful temporally windowed diagnostics (across a range of scales) on PDO EOF patterns should help explorers get a better handle on this.
The problem I see is that climate scientists know how to run canned statistical algorithms, but very, very few of them are well-trained in diagnostics.
After you’ve finished an analysis, the real work starts: diagnostics. Careful diagnostics are a lot of tedious work. Very few people have the intuition, patience, & judgement needed to do a good job at it.
Computers don’t automatically detect things they were never programmed to see. No matter how good algorithms get, there will always be an infinite number of things canned algorithms miss, so careful diagnostics are done manually.
PDO is a many-to-one function. Many climate states can map to the same index value. Some eras are systematically on one or the other side of the temporally-global (not to be confused with spatially-global) attractor and this is being ignored.
As the PDO metric is currently constructed, the same index value can represent significantly differing physical states. This limits the utility of the currently-favored PDO measure (it was a good & useful stepping stone on a path we should be moving along), even if the PDO concept is largely sound.
Remember: The measure is an approximation of the concept.
The measure can be improved.
Meanwhile: Vigilantly conscious differentiation between PDO the concept & PDO the measure is needed to facilitate any worthwhile discussion of PDO.