What's the deal with the snow forecast for Seattle?

As much of the nationwide media is now proclaiming, there is a potential for a major winter weather event in western Washington tonight and tomorrow.  Here's what we're looking at.

Snow showers continue over western Washington this afternoon as cool, moist air streams in off the Pacific.  Portland and the northern Willamette valley are really getting hammered.  Here's the latest composite radar image for the region.

However, the latest surface map shows that the temperatures around Portland are several degrees above freezing and most of what is falling is rain.

Showers will continue tonight, but the big snow event won't be until late tonight and into tomorrow as a relatively weak surface low approaches the coast.  Here's the 27-hour forecast from our 12Z initialized, 36 km WRF for tomorrow morning:

Note the really strong pressure and temperature gradients shaping up along the coast for tomorrow.  There's very cold air in the interior of British Columbia and relatively mild, moist air being brought up from the south behind the surface low.  So how does the interaction between these two air masses lead to a snow event?

As the surface low moves in, there will be a strong pressure gradient across the mountains--high pressure associated with the cold air inland and the lower pressure associated with this warm frontal wave off the coast.  This pressure gradient is going to draw down cold air near the surface out through gaps in the mountains and down into the lowlands.  Meanwhile, the warmer, moist air being brought up from the south will tend to lift over the cold air.  That lifting will cool the moist air, clouds will form, and precipitation will fall out into the cold air near the surface.  Here's a diagram that hopefully will help explain that setup:

That is a Google Earth perspective looking at the Olympics from the Cascade Mountains east of Seattle.  As the low pulls down a wedge of cold air out of the mountains to the north, warm moist air circulating around the low will move in from the south and rise over this low-level cold air.  The result is snow, and a lot of snow if the boundary between the leading edge of the cold air and the warm air to the south doesn't move much.

However, locating that boundary between the cold and warm air is critical for determining where the snow will fall.  If the warm air moves too far north, Seattle will be too warm for snow and we'll get a lot of rain.  If the cold air pushes too far south, the warmer, moist air will be precipitated out by the time it reaches Seattle and we'll see less snow.  This is all related to the placement of the surface low.  Here's two scenarios:

First, let's say the surface low stays further north--approaching the coast from the Strait of Juan de Fuca, perhaps.  In this case, warm, moist air will be brought up into western Washington.  However, because there is that nice gap between the Olympics and Vancouver Island (the Strait of Juan de Fuca), the surface low will draw most of the cold air straight west and very little will filter south into the Seattle area.  This would turn into an all rain event for Seattle.  I show this setup below.

Now let's consider another case.  If the low pressure center moves toward shore just a few hundred miles to the south, some things change.  We still will have warmer, moister air brought from the south around the low and into western Washington.  However, the cold air being drawn out of the mountains in Canada will be pulled more southwesterly out of the Cascades--and straight at the Olympics.  Some of the cold air will then get pulled down through the Puget Sound lowlands in addition to wrapping around the Pacific side.  This setup is much more favorable for snow in Seattle--we get that cold air near the surface that we need to support snow:

And it looks like our models so far are favoring a southern route.  Here's last night's 00Z forecast from our 4km WRF for where the surface low would be at 18Z on Wednesday (Wednesday morning).

Notice that the surface low was forecast to be right off the mouth of the Columbia river.  The pink horizontal line across the map marks the freezing line at 925mb (near the surface) anywhere north of that line should be cold enough to support snow.  Definitely a snow event for Seattle.  Here's the corresponding 24-hour snowfall accumulations that were forecast by tomorrow evening over western Washington:

This placed 6-8 inches of snow over the Seattle area--very impressive snowfall totals.

But, wait...this was last night's model.  What about this morning's 12Z run?

This paints a slightly different story.   Here's this morning's model surface forecast for 18Z Wednesday (the same time as the surface map above):

The low pressure center (and the freezing line with the cold air behind it) has been moved slightly further south--it's not much, but it's enough to make a dramatic difference in the final forecast snowfall amounts:

Now the model is only forecast 2-4 inches over the Seattle area, with a sharp decrease to the north to almost no snow north of Everett.  Since this morning's model run moved the surface low and cold air south, the most efficient precipitation production also moved to our south.  Thus, the model precipitation amounts have come down.

This change in precipitation amounts is very sensitive to the exact location and timing of the arrival of warm air and the low pressure center on the coast.  One thing we do know is that with this kind of model trend (moving the colder air further south), it should definitely be cold enough to support snow through the duration of this event.  We're not as concerned about the transition to rain at the end anymore.  The big question is still what final precipitation amounts will be.  The low may end up moving slightly further north and increase the snow amounts again--we'll have to wait and see what tonight's round of models have in store.

Even with the slightly diminished amounts, we're still looking at a major snow event late tonight and tomorrow in the Seattle area.  It looks like the only questions now are in the details.