For those looking at the ECM snow maps, note that the algorithm tends to overproduce snow near areas of mixing. Take NE NJ between 18z Sun-0z Mon — snow algorithm outputs nearly 4” of snow, but ptype is primarily sleet. would expect closer to 1-2” with such an evolution.

This point about the ECM snow algorithm applies to the ensembles too. I would similarly cut ensemble snow totals down from the EPS mean & median in the southern end of the storm. https://twitter.com/ryanmaue/status/1200437985076031488?s=21

My point here is that snow algorithms can be a useful first guess, but numerous caveats (counting sleet/ice as snow, snow ratios, mesoscale banding) always need to be kept in mind when using them to make a forecast.