See I don’t have to have an answer for my ibservations to be correct. Not saying you are doing this but that question is basically a model hugger deflection
if you feel status quo is sufficient then that’s just very different from my thinking. Every occurrent snowstorm this season has been 50% higher or lower than just 24 hours in advance was depicting
I would venture that AI is going to formulate an analogue occurrent outcome data base from last 1-25 years and get into a more current and correct manner of evaluation, assessment and forecasts. Also ditch the 7+ day stuff, enhance the short term, destroy the floppy discs.