* The major American forecast models — the North American Model and Global Forecast System — have each shifted the bulk of Wednesday’s snowfall north and east AWAY from Southwest Virginia. Amounts would be light in most areas if these models verify, with the heaviest snow north of Interstate 64.
* The European (ECMWF) and British (UKMET) forecast models, and to a lesser extent the Canadian, maintain a significant threat of heavy snow in our region, including some 10+ inch amounts, especially the Roanoke Valley north and east.
* Noting the wide disparity in forecast model solutions, the Hydrometeorological Prediction Center has put all of Virginia west of Richmond in a moderate risk of 4+ inches on Wednesday, and areas west/northwest of Roanoke in a high risk of 4+ inches and slight risk of 12+ inches.
* Much is contingent on the strength and location of a broad blocking low off the coast of Canada/Northeast U.S. and whether that blocking low can force an upper-level “vort max” far enough south for heavy snow to develop in our region. This accounts for much of the model discrepancy.
* My best shot at probabilities for Roanoke and Blacksburg now: 1 inch+, 70 percent; 4 inches+, 50 percent; 6 inches+, 40 percent; 8 inches+, 30 percent; 12 inches+, 20 percent. Chances will be better northward, and generally less to the south.
* This could truly be an “all or nothing” snowstorm with a large area of big snow totals giving way to small amounts within 30 miles or less.