I'm always disappointed in how inaccurate the NESIS storm maps are. Not just for here, but for example, I can show you photos confirming about 4 feet of snow at around 500 feet elevation in Harriman much of it from the late February 2010 storm (you can also cross ref old PNS reports).
I think I understand the reason locally to be that they relied on one or two observations to cover central LI, with Setauket-Strongs neck having been one of them until fairly recently. That station didn't always measure even once in 24 hours, plus it is surrounded by water.
For Rockland County and eastern Orange, there was probably a low snowfall report (possibly accurate for its location) that was overweighted in developing that map.
The maps are useful in approximating a very broad picture of snowfall distribution, but have a lot of the details wrong. I think there is more data and better tools for abstracting the data into visual maps available than these are utilizing.