Tuesday, December 25, 2007

White Christmas
All over town between noon and 1:30 mommies and daddies were shoving bundled-up kids out of doors to be models in their Portland Christmas Snow of 2007 photographs. I saw this throughout my ten-miler today. I heard one optimistic little guy shout, “When it gets really deep, we’ll make a snowman.”

First little white pellets then big wet gobs fell from the sky – first sporadically then in a frenzy – that slow-motion sort of frenzy that characterizes snow falling and, I don’t know, crashing on a bike. It was, yes, magical, even if it barely stuck, barely added up to a quarter-inch, barely will be recorded in meteorological history.

I had laced 'em up and gotten out at noon precisely, about 45 seconds into this little weather event. I was going to run no matter what but this, this was cool. It isn’t every day you get to run in the snow on Christmas Day. Even if this were, say, Buffalo, the odds would be slightly less than 1-in-365, right? I read somewhere that in here in Portland, there hadn’t been a trace on Christmas Day since ’90 and ’90 is further back in history than you or I want to think, my friend.

Tons of people were out, the parents and the kids, people and dogs, a guy in a T-shirt smoking a cigarette, woman runner in shorts, a bunch of runners in tights and wool hats ... me, in tights and my customary running cap. The temperature was around 33 or 34 and who knew how long the snow would last before the predicted changeover to rain? My house is between 50 and 100 feet above sea level, so after a jaunt around Laurelhurst I headed back toward and up Mount Tabor, 500 or so feet up. There the wind was swirling, blowing the big flakes around. They hurt when they hit an eyelash just so, refresh when they land in your mouth, and when one hit me on the nose I actually giggled.

Everyone was digging it, this Christmas snow. Where the trees didn’t catch the snow, it actually accumulated on the trails. It never stopped falling as I went around, down, up and down the mountain. It only turned to rain just as I rounded the corner onto my block, then it was snow again, then rain, all in the matter of a hundred yards. All rain since. Maybe more snow tonight, maybe enough to stick around into the morning. Maybe some on Thursday. Kind of a big deal in these parts. Nice around Christmas.


Dan Brekke said...

You must have made that Buffalo comment knowing that "someone" would chase that down. I can tell you that the NWS office in Buffalo has a chart that shows that falling snow has been recorded 46 times in the 67 Christmases from 1940 through 2006. There was no snow this year, so the tally is 46 "white" Xmases in 68 years. Even if you throw out the 10 occasions when snowfall was described as a trace, the odds would seem somewhat better than one in 365 that you'll see snow falling on Christmas in Buffalo.

Linkage: http://www.erh.noaa.gov/buf/xmasclimate.html

Pete said...

Oops, what I was trying to speak to wasn't precisely the probability of it snowing on Christmas, although that would necessarily be a part of the equation. I was talking about the probability that any random run would take place on a snowy Christmas Day -- "It isn’t every day you get to run in the snow on Christmas Day" vs. "It isn't every Christmas Day you get to run in the snow." And since there is just one Christmas Day each year, in calculating those chances I started with 1 in 356 -- the probability, first, that the day would be Christmas. I wasn't motivated to dig deeper, but since you have helpfully provided data, I will now!

OK, so about two-thirds of Christmases are white in Buffalo, So the probability of any randomly selected day in Buffalo being a snowy Christmas Day would be around 1 in 500. Let's use that -- .002.

Of course, this doesn't mean that if someone runs on 500 randomly selected days in Buffalo, they will for sure run on one snowy Christmas Day. Say I was a longtime Buffalo runner -- been at it for 20 years. And I run 200 times a year, always running randomly. Then we can get a meaningful standard deviation and get a better sense of what our "probability" means in the real world (the real world where runners don't look outside before deciding to run!).

OK, based on 4000 trials (20 years x 200 runs/year), with a probability of running on a snowy Christmas Day being .002 and the probability of running on a day that is not a snowy Christmas Day being .998, we come up with a standard deviation of 2.8 (square root of the sum of 4000 x 0.002 x 0.998). I think that gives us 2.8. So there's a 68% chance that the our veteran randomly running Buffaloan will have run 8 times on a snowy Christmas, plus or minus 2.8 days. Or, put another way, a 68% chance that he's run on somewhere between 5.2 and 10.8 snowy days.

I think.