GFS model simulation of winter storm at 2 a.m. Tuesday morning

* Winter storm watch 7 p.m. Monday to 2 p.m. Tuesday *

A mere 1.4 inches of snow has fallen the entire winter in Washington. If everything comes together just right, we could increase that total by many multiples Monday night. But the further south and east you go in the region, the less certain significant snow becomes.

Model forecasts on Saturday evening generally suggested very significant snow will fall Monday night in the metro region, away from Southern Maryland, and could even reach double digits. However, nagging questions about how heavy the precipitation will be and where the rain-snow line will set up remain.

Based on the latest information, we think a general 5-10 inches seems like a reasonable first call for snowfall amounts in the immediate metro, with more to the north and northeast, and less to the south and southeast.

Below find the stream of updates from our evening live blog session, which informed this snowfall forecast.

12:30 a.m. update: Looking at all of the models (granted the European is still to come out), I am impressed with what the GFS, Canadian, and UKMet models are collectively conveying: A classic snowstorm for the D.C. area with a strong area of low pressure developing near Cape Hatteras and tracking to a position off the Delmarva Tuesday morning. The models all agree the storm will pass through relatively quickly but also produce heavy snow. I think some of the precipitation amounts they are simulating are probably overdone, but even cutting them by 25-50 percent would still produce substantial snowfall.

The NAM model, with its razor sharp rain-snow line cutting through the immediate metro area, is a concern. But I am mostly discounting it as it is not supported by the other models and also is not particularly good with forecasts beyond 36 hours. Having said that, especially as it’s March, we need to pay close attention to temperatures because if the other models are too cold and temperatures don’t fall as forecast with the snow, amounts will fall well short of their potential.

A March snowstorm is rare, but not unprecedented in the D.C. region. (Claritza Jimenez/The Washington Post)

A reasonable first forecast is for about 5-10 inches of snow in the immediate metro area and lesser amounts to the south and southeast and greater amounts to the north and northeast.

See our first accumulation map, which could well need altering Sunday, just above this update.

Consider that small changes in the forecast track on Sunday would mean major changes to these numbers.

We’ll have a fresh forecast up at 6 a.m. Sunday and then a detailed storm update out before noon after reviewing the next big set of model forecasts.

Remember to adjust your clocks forward to tonight.

Thanks for joining us!

12:05 a.m. update: I’m about to wrap things up for the night but have one more update in which I’ll try to synthesize what I’ve seen this evening. Look for it in about 20 minutes.

12:00 a.m. update (Sunday): The UKMet model is in which, while lesser known, is actually the second best-performer behind the famed European model. We don’t have great maps for this model, but the simulation we see looks more or less consistent with the GFS and Canadian in terms of the storm track – which is a favorable one for significant snow in Washington. At 8 a.m. Tuesday, the storm is centered off the Delmarva Coast and the model shows heavy precipitation having fallen, see image below.

UKMet model simulation of storm at 8 a.m. Tuesday morning. (Environment Canada)

11:48 p.m. update: The Canadian model snowfall totals from D.C. to Boston are obscene, in the 15-20 inch range, and I think probably overdone, at least in our region. For one, I don’t think the storm will intensify fast enough to deliver that amount of snow in Washington in under 12 hours. Secondly, that amount of snow has no historic precedent at this time of year. Third, it is generally the worst performing model. Finally, not all of the snow it simulates will stick.

Canadian model snowfall forecast for Monday night-Tuesday winter storm.

Having said all of that, it does have some support from the GFS model which, to me, means you at least have to consider the possibility of double digits totals in our area.

11:32 p.m. update: Below a visualization of the storm from the Canadian model. It’s eerily similar to the GFS. Snow begins just before 8 p.m. Monday in the metro region, quickly becomes and heavy and pastes the region overnight. It does not mix with rain or sleet in this model and is a classic track for heavy snow from Cape Hatteras to a position east of the Delmarva. The storm then exits quickly Tuesday morning.

Canadian model simulation of the Monday night winter storm.

11:24 p.m. update: The Canadian model forecast, for Monday night, just in, has a lot in common with the GFS: It keeps the rain-snow line well east of Interstate 95 and dumps a ton of snow in the immediate D.C. area. Graphics coming in a few.

11:18 p.m. update: The fun is not over yet. The Canadian model is in. Looking at it now and will have comments in a few minutes.

11:12 p.m. update: How much snow are we talking about if the GFS model is right? Assuming temperatures are at or below freezing for most of the storm and rain or sleet doesn’t mix with the snow, the numbers are mind-boggling:

GFS model snowfall forecast for Monday night-Tuesday storm. (

It suggests the immediate metro area would get 10-15 inches of snow, and our northern suburbs would see 15-20 inches – all in the course of about 12 hours. This would be a historic storm if that comes to pass.

A couple caveats: 1) This model sometimes over-predicts the amount of precipitation in these kinds of storms. 2) Some of the snow that fell initially would probably melt (or be wasted) due to temperatures above freezing for a time as well as above freezing ground temperature.

11:00 p.m. update: Below is a simulation of the storm from the GFS model. The storm moves into Washington around 8 p.m. Monday night and is on its way out by around 8 a.m. Tuesday morning. But, in that time, the model suggests it unloads A LOT of snow.

GFS model simulation of Monday night winter storm.

10:51 p.m. update: The GFS model is in and snow lovers in the Washington region will like it. It’s a cold, snowy forecast. Once the precipitation begins Monday evening, temperatures quickly fall to 32 degrees or lower and the rain-snow line is a good deal east of Interstate 95. The model produces over 1.4 inches of liquid equivalent in Washington which would convert to a lot of snow given the model’s temperature forecast. I’ll post some graphics shortly.

10:47 p.m. update: The GFS model is flowing in right now, so we’ll have some comments on it in the next few minutes.

While this storm will impact D.C., it could be a much bigger deal to our northeast. If you have travel plans in New England Tuesday into early Wednesday, expect there to be disruptions. Look at this forecast for snowfall amounts over the entire Northeastern U.S., it is impressive:

National Weather Service snowfall forecast for Northeast U.S. for early week winter storm. (

Widespread totals of at least a foot are expected from eastern Pennsylvania to eastern Massachusetts.

10:36 p.m. update: In addition to the two versions of the NAM model discussed below, there is actually a third one – run at even a higher resolution (3 kilometers), which we call the NAM 3K.  It has a forecast product which shows how much snow will fall each hour and its forecast of the snowfall in the hour ending at 4 a.m. Tuesday morning is fascinating:

Snow falling between 3 a.m. and 4 a.m. Tuesday from NAM 3K model. (

First of all, notice how sharply the snow cuts off just east of Interstate 95. But then notice how heavy the snow is west of Interstate 95 – ripping at the rate of 1 to 2 inches per hour.

This one graphic shows why the the location of the rain-snow line will be so important in this storm and how small deviations could have huge implications for the forecast.

10:19 p.m. update: For snow lovers depressed by the NAM model simulation just shown, I should point out an alternative version of this model is run at a higher resolution and is colder and snowier than its low-resolution version. Rather than the rain-snow line hovering west of Washington throughout this event, it straddles the Interstate 95 corridor in the high-resolution NAM and would give D.C. and its close-in suburbs more snow than low-resolution NAM forecasts.

Its forecast suggests at least 2 to 4 inches of snow in the region:

High resolution NAM snowfall forecast through 8 a.m. Tuesday morning. (

Compare that to the low-resolution NAM which only suggests 1-2 inches in the immediate metro, though a bit more to the north and west:

NAM model snowfall forecast through 8 a.m. Tuesday.

10:05 p.m. update: Below is a NAM model radar simulation of Monday night’s storm.  If you’re snow lover along and east of Interstate 95, you may not want to look as it would be your worst case scenario. Basically, it shows the rain-snow line west of the Beltway for a large part of the event which would mean mostly rain or a rain-snow mix near the city and points east. It even suggests our nearby western suburbs could have a good deal of mixing.

NAM model radar simulation of Monday night storm. (

This is not a good model forecast if you like snow except perhaps if you live north and west of Leesburg and Frederick.

But remember what I wrote in my 9:31 p.m. update: we’re outside the range at which this model is considered pretty reliable – so you can take it with a grain of salt. That said, if the GFS, running in about 30 minutes, is similar, then it’s time to worry.

9:50 p.m. update: The NAM model is in and produces a good deal of precipitation – around 1-1.4 inches liquid equivalent, which would convert to a healthy amount of snow if all the precipitation falls as snow and doesn’t mix with rain. But there are a few red flags for snow lovers in this run: 1) Precipitation starts off as rain in Washington and doesn’t change to snow until potentially after midnight. 2) It is a relatively short duration event, less than 12 hours. The precipitation starts around 8 p.m. and is all over between 6 and 8 a.m. 3) The model tends to have a wet bias and sometimes simulates too much precipitation.

I’ll show some graphics in my next update.

9:37 p.m. update: One thing this storm will have going for it, if you like snow, is a strong area of cold high pressure to the north and northwest. The presence of this high pressure is critical for supplying sufficient amounts of cold air for snow to fall.

GFS model simulation of high pressure over Canada on Monday night, which will feed cold air into this storm. (
GFS model simulation of high pressure over Canada on Monday night, which will feed cold air into this storm. (

Some of you may recall the March 6, 2013 “Snowquester” storm in which large amounts of snow were forecast but didn’t materialize. In that case, there was no cold high pressure system to the north – which should’ve been a big warning sign for forecasters.

9:31 p.m. update: As we await the NAM model, which is just starting to come in, we want to note that this model usually performs best within about 24-36 hours of a storm event. As we’re about 48 hours away from this still, we’re a little outside its best range – so take our reports on this model with a grain of salt.

9:27 p.m. update: In addition to producing a most likely snowfall forecast (see our first update from 8:50 p.m.), the National Weather Service now also issues forecasts for low-end and high-end possibilities. For this particular storm in Washington, its low-end forecast is 3 inches and its high-end forecast is 13 inches. That’s quite a range, but we think it’s a reasonable assessment given the complexity of the forecast:

9:25 p.m. update: Normally, at this time, the NAM model would be coming in so we could discuss it. But there appear to be some delays at the National Weather Service (which runs this model). We’ll hopefully be able to start writing about it shortly.

9:15 p.m. update: To illustrate how close D.C. may be to the rain-snow line Monday night, let’s look at this (admittedly crude) representation of it from the SREF model (described below). The animation below shows three snapshots of the rain-snow line Monday night into Tuesday morning at 11 p.m., 2 a.m, and 5 a.m.

SREF model simulation of rain snow line overnight Monday. (

Notice how the District goes from snow to rain to snow. The rain-snow line sets up right along Interstate 95. If you’re north and west of it, this group of models suggests you have a good chance to stay all snow. But along it and especially to the south and east, rain could mix in it times.

9:00 p.m. update: The first model we’ll look at this evening is actually an average of a group of models known as the Short-range Ensemble Forecast System or SREF. This group of models suggests the region will experience plenty of precipitation Monday night into Tuesday, the equivalent of 1 to 1.5 inches of rain.

SREF model total lquid equivalent precipitation for storm Monday night into Tuesday. (

Of course, we’re expecting the majority of this precipitation to fall as snow. Typically, an inch of rain equates to about ten inches of snow. But because Monday night’s snow will be heavy and wet, we might expect one inch of rain to equate to about seven inches of snow in the immediate metro area. Even with that assumption, when you do the math, this model is predicting the potential for a lot of snow – with double digit totals possible in some areas.

A limitation of this model group is that some of the individual models which comprise it sometimes produce precipitation totals that are unrealistically high. So it’s possible the amounts it is suggesting are slightly overdone.

8:50 p.m. update: To get us started, let’s take a look at the National Weather Service snow accumulation forecast.  The map below, which was released at 8:10 p.m., predicts widespread totals of 4 to 10 inches, with amounts increasing from southeast to northwest.  The highest amounts are north and west of the Beltway and a few spots in northern Maryland are predicted to exceed a foot.  The lowest totals are in southern Maryland, where one to four inches or less are expected.

National Weather Service snowfall forecast for Monday night and Tuesday, released at 8:10 p.m. Saturday night.