Weather Blog

Hey, no peeking ahead at the long range forecast

Hey, no peeking ahead at the long range forecast
In the old days, weather forecasters could kind of peek at the long range forecast, but if it showed something interesting yet not exactly probable, they could keep it under wraps until confidence grew.

But in the days of blogging, twitter, and having all the forecasting models on the internet, more and more people are armchair weather forecasting, and word is easier to spread.

For example, if you go to some long range automated sites like or, you'll find they are predicting snow in Seattle around next Sunday.  Those are automated forecasts and just take whatever the computer is saying at face value, so they bite whenever the forecast models show an arctic outbreak, no matter how slim the actual chance may be. (More on that in a moment.)  But people see it, and blog/twitter about it, and the anticipation grows.

So with the cat easier to jump out of the proverbial bag these days, I figure I might as well join the revolution. :)  You're probably going to read about these events at other weather sites anyway, and we can use the web hits.

Many winter fans have already noticed some big time changes being advertised in the long range forecast, starting around next Friday through the following week (the week of Dec. 14) . I hinted to it in our forecast yesterday, but the forecast models from today are much more gung-ho about it. IF you were to take this at absolute face value, we could see snow levels flirting with the surface at times -- if nothing else, at least some colder air with some potential arctic-type winds coming out of the Fraser Valley in Whatcom County.

Now, going back to the second paragraph, there's several reasons why we don't hype these sort of events this far out. For one, the long range models run at a lower resolution in the longer range. (Read more about them in my earlier blog entry on how forecast models work.) Thus, they tend to overestimate the intrusion of cold air from Canada, because they don't factor in the blocking properties of the Canadian Rockies and Cascades.

(This is where those automated weather forecast sites fail with no human interaction to know of these limitations. I remember a forecasted arctic outbreak a few years ago that had a big computer error in it, and these web sites were forecasting highs in the single digits and lows below zero in Seattle, despite the fact that had never happened before. But we sure received a lot of e-mail about it!)

Second, the mere mention of snow sends 63.5% of the greater Puget Sound region residents into some sort of either panic or ecstatic tizzy. (The other 36.5% just start biting their nails.) So we want to have some better confidence before it we give it too much traction (no pun intended).

That said, there are some points in favor of at least getting pretty chilly. The first is simple: we're due. It's been a way warm October and November and first part of December.

Second, most "overdone" arctic outbreaks on a long range forecast model have the arctic high over Alberta or even Saskatchewan , and the arctic air coming in from the northeast -- thus ignoring those two big mountain ranges. This pattern instead has the arctic high over eastern B.C. and Alberta -- still a bit father east than I'd like, but closer to at least set up some arctic air where it could flow through the Fraser Valley.  Of also a benefit, we have some lower pressure occasionally passing to our south that would help to draw in some arctic air.

Third, the long range model has been kind of hinting at this for a while.  The details keep changing -- for one, this morning's model has it much colder but drier the week of the 14th, while the one from last night has it much wetter but a little warmer to where it's just a mountain snow event or possibly fringe foothill snow event.  So while the specifics of what the forecast models are showing are likely to keep changing wildly as we go along and approach next week, at least the general theme of colder weather has been somewhat consistent.

Armchair forecasting

Here are some examples of what I just wrote about. This is the forecast model run on Thursday evening, valid for the morning of Wednesday, Dec. 17:

That would portend a wet and breezy day as that low moves in from the northwest, dragging with it a juicy cold front that would bring solid rain to the lowlands and heavy snow to the mountains, and highs in the low 40s.

Now, here is the forecast model from Friday morning, valid for the same time:

This shows an arctic high better positioned in Alberta, with cold air damming up against the Canadian Cascades. This also shows that low coming in much farther to the south, but this scenario would then draw arctic air out of the Fraser Valley into Western Washington. This model scenario, at strict face value, would bring a possible rain/snow mix to southwestern Washington, depending on how much cold air comes out of the Fraser Valley.  So you can see the inconsistency, but also see what's in the realm.  (You can find these charts at this NOAA site)

[Fun side forecasting note for you armchair weather forecasters: The black lines on these models are isobars -- lines of constant pressure. But those dotted lines can be more interesting in the winter -- those are a measurement of atmospheric thickness in decameters.  What that is, is a measure of temperature. Cold air is more dense than warm air.  So what you can do is measure the thickness of air in a column between air at 1000 milibars pressure (near the surface) and 500 milibars pressure (about 18,000 feet.)  The lower the thickness the more dense the air, and the colder it is.

Think of it as a big plastic column, with warm air being like a bowling ball and cold air being like a golf ball. If you put in 20 bowling balls in one column and 20 golf balls in another. The bowling ball column will be filled higher and have a greater thickness.

The numbers on the dotted lines show the anticipated thickness. "540" would be 540 decameters, or 5400 meters, or about 18,000 feet. Note on this particular model, the lines greater than 540 are in red, and 540 and below are in blue. That's because over most of the nation, the 540 line is the "rain/snow" line -- i.e., if you're under 540, it'll snow, and over 540, it'll be rain.

That doesn't work in the Northwest, there the ocean air keeps us milder. For us, it's about the "snow level at 3,000" line, or put another way, the "Snow at Snoqualmie Pass" line. If we're in the blue lines, it's a good mountain snow event. Red lines, not so much.

For snow in Seattle, we have to get down to about 516, with 517-522 in the fringe area.]

But again, to stress, this time frame we're talking about is, what, 7-8 days out? Pretty far away. We've seen plenty an arctic air situation forecast 4-10 days out go poof as the days draw near. 

Still, I wanted to at least jump in the fray that next week's weather might at least be interesting.  Stay tuned!

(And for those wondering "Hey, where's my Friday time lapses?" I'll post them to go live on Saturday morning, so check back :) )