Unfortunate to see the OP runs deteriorate. However, this was true, both at 12z, 18z, and now: OP runs should be considered pure theatre until ~72 hours before the first flakes would fall. It is best to tracks ensemble trends at H5
Someone ought to train an AI that takes deterministic and ensemble guidance 180 hrs out and predicts the likelihood of a significant event. This can be sort of our threat barometer
What would need to happen:
1) Define sig. event (average precipitation of .5 over a 48 hour period for some given area)
Parse all modeling data that preceded sig. events and train the AI to find model trends that correlated more with an event occurring than with non-sig weather.
Might be cool, Would be interested if someone here is an AI research or user to verify my methodology
If it works out, excitement -> real positive outcome.
If it doesn’t (probabilistically most likely outcome), we at least enjoyed the modeling theatre. Win/win
People are melting already? Give this a couple days to sort itself out. I mean, we are discussing a threat about 6 days from gracing the eastern seaboard.