Winter Forecast -- Updated

October can be a very helpful month in determining what the winter might do. I looked at this October's 500hPa pattern and qualitatively came up with the following October analogs:
1948, 1948, 1950, 1953, 1955, 1956, 1957, 1958, 1962, 1962, 1963, 1963, 1964, 1967, 1977, 1979, 1980, 1989, 2005, 2007

Years that are listed twice were stronger analogs. Based on this list and the previous list of analogs I had already compiled, here is the updated "winter analog list" (for comparison, the old list is listed below)
1948; 1950; 1950; 1950; 1962; 1962; 1962; 1963; 1966; 1966; 1979; 2001
 (1950; 1950; 1950; 1960; 1962; 1962; 1966; 1966; 1966; 1979; 2001; 2001; 2009)

Again, the more a year is listed, the more it's weighted as an analog. Based on this "new" analog list, changes aren't actually that major... which gives me more confidence in the forecast. That being said, the November forecast seems to not jive with the two-week forecasts for the first half of November... so maybe the second half of November will overpower the first half? Anyway, here are the maps.

NOVEMBER

DECEMBER

JANUARY

FEBRUARY

MARCH

 

October 29th snow in State College - time lapse

Out my window:

Winter Forecast: 2011-2012

So, let's start by noting that winter forecasts are generally bunk (well, mine are, at least :P ). But I like coming up with a guess anyway.

My methodology? Very simplistic and probably worthless. I looked at northern hemisphere 500hPa anomalies for August and September and compared them with previous years. I found "analog" years that fit both months reasonably well, with more weighting given to years that fit better. It was completely qualitative and subject to my biases, whatever those may be.

I then checked the analog years for ENSO conditions. This winter will almost certainly be a Nina winter. So, I subtracted one "weight" for years that were Nino, left the "weight" the same for years that were Neutral, and added one "weight" for years that were Nina.

My final analog list was:
1950; 1950; 1950; 1960; 1962; 1962; 1966; 1966; 1966; 1979; 2001; 2001; 2009

Repeated years are years with more weight given them.


Given that list of analogs, I looked at the 500hPa pattern for each "extended winter" month (Nov - Mar), and drew a map of what I'd guess would be the corresponding temperature anomalies. Here they are:

NOVEMBER

DECEMBER

JANUARY 

FEBRUARY

MARCH

The key is...
Red: Torch
Orange: Warm
Yellow: Mild
White: Near normal
Blue: Cool
Purple: Cold
Lavender: Frigid

As for precipitation, I'd say overall winter precip would be well above normal in the Pacific Northwest, and below normal from Southern California through Western Texas. Elsewhere, it'll probably balance out to near normal overall. 

I'm in Pennsylvania now...

... for grad school (Penn State University). But I'll still update this blog for interesting weather events in both the Pac NW and central PA this winter.

I've been pretty lazy with it recently (no posts on the severe wx or hurricane? For shame). But I'll try to be better about it in the future!

Not that anyone reads it. :P

January 2011

What will January be like?

I've created an analog list based on a subjective look at the overall December pattern and came up with the following years:
1950**, 1952, 1955, 1961, 1964, 1965, 1968, 1981*, 2009

1950 is by far the best match, and 1981 is also a better match than the others on the list. The rest are only mediocre matches, but I have included them to smooth out wild solutions some.

Here was the December pattern we saw in 2010:

Here is the composite of my analog years:

Not perfect, but not a bad match, either. Here was that pattern rolled forward into January:

This is an interesting result. If we had a pattern like that in January, the Pacific Northwest would be very close to major cold and snow. However, there could be a sharp north-south temperature gradient in British Columbia or northern Washington State, with cool, moist Pacific flow south of that. Lots of borderline rain/snow showers in Portland would be my best guess based on that pattern on its own.

However, I'd like to note two major caveats. First, the pattern would obviously not look exactly the same all January, and if we average a pattern like the one in the above image, there would almost certainly be times where the cold/snow made it down to all of western Washington and Oregon. Second, you might notice a subtle difference between the composite pattern and the actual December pattern. It seems like many of the anomalies in the Western Hemisphere are slightly displaced towards the west in the composite compared to the actual. If that remains the case in January, then the actual pattern we'd see would be a slight shift east of what's shown above, which would be much colder for the Pacific Northwest.


On a final note, many of the models and ensembles are now hinting at a pattern not unlike that being shown in my January composite, so I have more confidence that this could show some semblance of reality.

Snow on the way

Of the five models I regularly look at, every single one of them now gives us here in Portland at least a little accumulating snow.

The GFS is the only solution that's further inland with the trough now, and even it gives us enough moisture to squeeze out up to an inch of accumulating snow. The UKMET, ECMWF, GGEM (Canadian), and NAM models are all further west with the trough, inducing a stronger surface low and much more moisture. The NAM, at face value, gives the Portland area at least 2-5" of snow, and the other models generally look like the NAM.

Afterwards it gets cold, with 850mb temperatures down to the -12 to -14C range. With snow on the ground, I wouldn't be surprised to see even the Portland Airport get down to the lower 20s, or even (*gasp*) the upper 10s.

Enjoy the snow and cold!

More Musings on Cold and Snow

What a change a day makes. Last night's 00z models were somewhat disappointing for us here in the Pacific Northwest, sliding the brunt of the cold (and all chances of snow) off to the east. The 12z Euro reinforced that idea.

But some magic has been happening in model land today, and the clear trend in the ensemble mean couldn't paint a better picture. In the four runs since and including the 00z GFS, the coldest 850mb temperatures for the ensemble mean in Portland have been -7C, -9C, -10C, and on the latest 18z ensembles, as shown above, -12C. That is an incredible ensemble mean for a time period less than 5 days out. Seattle, likewise, has trended, respectively, from -9C, -13C, -13C, -16C. Wow!

With the further west solution, the 18z GFS produces another area of low pressure along the coast after the initial cold surge Sunday night. As this drops southeastward, moisture associated with the low dumps quite a bit of snow in the northern and north-central Willamette Valley. This is just one run and specifics will change before all is said and done, but our chances for snow on Monday are now much higher than they seemed 24 hours ago.

Tuesday and Wednesday will likely be dry with a ridge building in, but should still be cold with east winds through the Gorge (and a remaining cold airmass over Seattle). There's no agreement on how or when we transition out of the cold. It could just be a gradual warm up with dry conditions, could be a messy overrunning snow or ice storm, or somewhere in between (moderating, then just rain?).



In the closer-range, it appears there is an outside chance that someone in Northwest Oregon or Southwest Washington could see a band of snow set up Saturday night. A band of moisture swinging around the low offshore could stall out in the area, and with heavy enough precipitation rates and a cold airmass, not to mention light offshore flow in the Portland area, we could see the same type of "surprise" we saw last night, where snow levels were a good 1,500 feet below predicted in the heavier bands of precipitation. Except this time, that would take the snow level all the way down below 500 feet... possibly near the surface. It would be a very borderline thing, but the potential is there.