It is comically stubborn. I love that it initializes 3-5F warmer than the 18Z run at hour 0 and then INSTANTLY jumps to its old forecast at hour 1. "Nah, I was definitely right before"
If the code that accumulates snow is inside the model and driven by physics, then why are snow accumulations so wildly different on different websites? Presumably they all use the exact same data from the model.
I for one am hoping to see the UKIE/CMC get back in the game today. They were both crap yesterday. Slightly less interested in the waffling of GFS/EURO.