More Historic Cold, Dangerous Storm Possible in March

UPDATE 3/3: The cold weather has arrived. Here are record lows that are threatened on Monday night (more in this article). Richmond's all-time March record low is only 10 and we are forecasting 6 there tonight. At Philadelphia, it's 5 and we're forecasting 6. Scranton should trounce their daily record low of zero, as will Morgantown, WV.

UPDATE 2/27: Today's GFS has "moderated" to near-record temperatures in Missouri next week, but is still forecasting that the Philadelphia will be between 0 and -10 degrees F (suburbs) next Thursday (again, their coldest temperature ever in March is +5):

For Poughkeepsie, NY, the GFS says about -15 degrees; their daily record is +5 for Thursday but their March record is -13. Again, like with the Missouri example yesterday, this seems to be an anomaly not supported by the European model, but that model still predicts a low temp of near zero which would be a March record in Philly. It would also in Harrisburg, PA (March record +2) and -7 in Baltimore (record +5 for March).

We have our first official snow accumulation map published in this story (similar to what the 12Z models said yesterday and today). Meanwhile, record lows will be broken tonight in the Northeast:

ORIGINAL BLOG 2/26:

It's been a heck of a season, but winter is not over yet. Take a look at these (12Z) forecast model predictions for March 5th's minimum low temperatures. These would be daily, if not monthly records in many locations in the Midwest and Northeast.

Some stats: The GFS (the U.S. model) predicts -4 F in Philadelphia; the European model says around 0 F. The record low that day (and, in fact, for the entire month of March) is +5 in 1872. The GFS model also has a bizarre (read: probably not going to happen) area of temperatures less than -35 F in Missouri (see above)! It predicts that Des Moines, Iowa, gets down to -26* (the daily record is -15 in 1930). Springfield, Mo., drops to -10*; the record is +2 in 1908. And, as you would expect, this cold will help create a snowstorm. Here are the snowfall amounts from both models:

These are actually less than the models were predicting last night, but still show a nearly-country-wide snow storm with amounts over a foot in places. In fact, our story currently predicts "at least a 36-hour period" of snow for parts of the Midwest and Northeast, and our zip code forecasts show fairly steady snow, sometimes heavy, from Sunday to Tuesday in some towns! This is much different than your normal hit-and-miss quick storm; if this forecast is correct, this will slow travel and cancel schools over a large area of the country next week!

*These numbers were read off the high-res display in GREarth for the 2-meter temp, which may differ than what the graphics above show.

Report a Typo