Good Friday Mid-Morning! Welcome to subscribers and visitors! This is post number 795. Author: Dr. Ryan Maue ( X @RyanMaue )
Lower 48 precipitation type + MSLP for the next 6-days | ECMWF 06z
California finally ditches its last cut-off low —> crosses the Rockies into the Plains and becomes a powerful storm system in Ontario/Quebec
That storm system well in Canada will drag a cold front through the Eastern U.S. —> colder but also quiet weather for Thanksgiving into Black Friday & Weekend! Main travel concern would be scattered Lake Effect Snow Showers.
The main polar vortex remains locked up in Arctic — but there currently isn’t much of a cold-pool at all. The ongoing Stratospheric Sudden Warming event will take 1-2 weeks to reconfigure the “ridge bridge” over Alaska to wipe out the very warm air over Yukon to Greenland.
Colder weather expected in early December — but ensembles aren’t committed to the magnitude nor timing of a major “Polar Vortex” blast.
Weather Trader is an entirely subscriber-supported publication from Dr. Ryan Maue. To receive all new posts, consider becoming a free or paid subscriber, and please LIKE and SHARE posts.
Lower 48 Temperature Anomaly was 6.7°F above normal
47 million below freezing or about 25% of the U.S. population
Atmospheric rivers landfall Pacific Northwest — giving California a break.
Northeast just misses a powerful storm system as the main low pressure center heads through Ontario —> Quebec.
Thought bubble: There are a bunch of options for global weather modeling in the 10-day range. Our eyeball can see from the 8 panels that there is a trough in the U.S. Southwest, ridge to the north over Gulf of Alaska, which connects to the Northeast. The upper-level patterns look highly correlated —> they are.
Things we should know: what is the expected skill of the upcoming pattern at a particular Day based upon the model choice?
Do we have an idea about which is most right ahead of time?
—> even if we did know, there is still significant error and being the “least worst” or the “best ensemble” is highly desirable
Can we combine these model fields into a Grand Ensemble and get more/better information that can help us determine?
—> yes, we can also perform elaborate sorting and clustering to extract composites of lower frequency signals. We can then say which model(s) are like each other — and are behaving similarly.
A.I. is perfect for this type of pattern recognition and matching. Probably not Grok as it is “full of it” but Gemini 3 has significant promise!
Loading comments…