Monday, February 14, 2011

We are getting there

kw: weather forecasting, analysis

Not so many years ago, three-day weather forecasts were reasonably accurate, and the earliest forecasts you could get were five or six days. As computer models, particularly general circulation models, have improved, we have a bit more look-ahead, such that the Accuweather add-in to my browser presents fifteen-day forecasts. Getting ready for a trip, I decided to check about half that range (to tell the truth, I didn't think of it until I had eight days to go).

Anyway, I primarily looked at the day I'm leaving, which is Wednesday coming, or the 16th, and at the locations at both ends of the journey. At my end, then, we have the following; each line is the date of the forecast, the high, the low, and the outlook:
  • 2/08 - 54, 36, sunny
  • 2/10 - 49, 35, mostly sunny
  • 2/11 - 50, 35, mostly sunny
  • 2/13 - 51, 36, partly sunny
  • 2/14 - 49, 34, partly sunny
Now let's look at the other end, on the West coast near Portland, Oregon:
  • 2/08 - 41, 32, rain
  • 2/10 - 47, 36, a few showers
  • 2/11 - 47, 34, rain possible
  • 2/13 - 47, 34, clouds and showers
  • 2/14 - 47, 34, clouds and showers
A quick glance shows that the forecast in Oregon is steadier than the one here. The Delaware valley area is similar to New England in having volatile weather, what with 2-3 conflicting air masses in play on an ordinary day. During a typical cycle of two or three days' duration, a continental air mass arising in Canada or the northern plains (the Dakotas) bangs into a maritime air mass, and both are influenced by a warm air mass that arose in the Gulf of Mexico or off the coast of Georgia. The timing of the ensuing events gives weather forecasters fits.

On the other hand, in the northwest, the maritime weather is more stable, as the continental air masses that arise further north typically travel to the southeast and bypass the Seattle-Portland area. Now that I have a two-day forecast for both places that agrees with the three-day forecast, I can be pretty sure that I'll have good flying weather early in the trip, and at worst a little rain upon landing, depending on the timing of the showers.

In the meantime, I marvel that I can call up such data so easily, and test it for reliability. Having fifteen years' experience with the local weather, I could probably surmise the slight warming trend the forecasts describe, but I'd have no chance at guessing what lay at the other end. The people I'm visiting are newcomers to Oregon, though they have said that the weather patterns tend to run to longer cycles than those in this area. There was a time I'd have said that the weather problem is inherently unsolvable more than three days in advance, but clearly, I was wrong. I wonder if we'll one day have highly reliable two-week forecasts for everywhere.

No comments: