Winter 2026-2027. Historic Potential

[Weather Will]

Source: WB JB post from latest Canadian seasonal.  Hope it is right!

[Weather Will]

WB Nov. Canadian seasonal

[Cobalt]

23 minutes ago, Weather Will said:

Source: WB JB post from latest Canadian seasonal.  Hope it is right!

I am almost certain that this is derived from JB asking an AI model to produce a snowfall output map based on the model's temperature and precip anomaly maps. AI like the one that made this map are only designed to interpret and create images, so it (and therefore this map) aren't grounded in any climatological understanding. Basically, it's AI slop wishcasting, and if you ask the same model to make the exact same map it will be different every time. 

[Cobalt]

2 examples that stand out. The AI model is creating a new light purple color code out of nowhere

Also its text does not understand how the color contours correspond to snowfall lol

So yeah, affirms my suspicion that he just fed the canSIPS output into an AI model that is good enough to make maps (naturally weather maps probably show up a lot in the training data so it has some stuff to work with), but he asked it to generate data that does not exist and can't be derived from just the precip and temp maps alone. 

[Elim72]