Regarding T850s, I've kind of gone through an enlightenment journey with them from i) they are fantastic -6 ftw, ii) they are terrible, use soundings to iii) Actually they are great after all.
By soundings I mean the full vertical profile which you can view for GFS outputs, this is the best way of seeing whether the vertical profile supports snowfall, even better than things like partial thickness, freezing level or dewpoints; but the problem with it is that you only get a snapshot of one gridpoint at one time. Hence T850s are still quite useful, especially since the uncertainty in the long range GFS output means your soundings can also change.
So I do use T850s but a highly context independent assesment of them. You need to know the rules in full and they go something like this. These would be my snow threasholds
Is your precip frontal or convective?
If convective:
-6.5C should be your baseline. Subtract 1C if you live near the coast and your wind is being advected from the sea. Subtract 1-2C if you have warm air advection (i.e your T850s get warmer against the wind rathr than colder). Subtract 0.5C between the hours of 12pm and 3pm unless you are within 1 month of the winter solsdist. If cold air has been established for at least 2 days then you can add 0.5C-1C, and you can also do this if there is significant snow cover.
If frontal:
Base of -7C for cold fronts, -2C for warm or occluded fronts.
For warm or occluded fronts change your threashold based on how strong the precip is, how weak the wind is, and how established your cold air is. If all three are favourable (weak winds, coherent strong precip and cold air for several days) then use 0C rather than -2C, in extreme cases with low level CAA you can even use +1C but this is rare and the snow may mix with freezing rain and sleet. To be honest though, the better way of seeing where the snow/rain boundary is in these cases is looking for a strong T850 gradient. If you have a very strong T850 gradient then the cold side of where that gradient slackens off is where snow will be falling. If there is no strong gradient, then assume the front is less coherent and you would require colder T850s, probably at least -4C.
Originally Posted by: Quantum