Now that our local weather has settled into the more routine summer pattern of sunny, hot days, clear, mild nights and the generally southwesterly flow that can bring the occasional higher-terrain thundershower, let’s pause to look at models weather forecasters use to make short and longer-term outlooks.
It might help you understand the unexpected ripples that visit us.
As with any predictive model, accuracy decreases as the forecast period lengthens. It is not hard to understand why a forecast for the next 24 to 48 hours would be inherently more accurate than one looking longer into the future. The atmosphere is wildly dynamic, and its constantly changing parameters of pressure, temperature, humidity and wind — among other factors — make for a shape-shifting, protean target that can prove extremely difficult to hit.
Many of weather models combine a snapshot of current conditions blended with the climatology of the forecast area and the outcomes of similar past scenarios to produce outlooks forecasters can hone for a specific location and time.
The likelihood of certain events occurring is normally quoted in a percentage probability — thus allowing the forecaster, much like any savvy politician of the day, some waffling and wiggle room in which a sudden about face can be performed if what is happening outside the weather office is embarrassingly at odds with what was predicted.
In the case of Walla Walla, the Blue Mountains add another factor that can seriously complicate what might ordinarily be a fairly straightforward forecast.
An excellent example of this occurred last winter during a capricious, protracted ice and snow storm that gave folks fits at the National Weather Service office in Pendleton. Subtle changes in temperature resulting from shifting winds over the mountains necessitated multiple forecast revisions in both precipitation type and amounts every couple of hours.
More recently, the severe thunderstorms that caused damage west of Walla Walla were a forecast challenge because their trajectory and intensity were greatly influenced by their interaction with our local topography.
There are dozens of weather models forecasters can choose, and much like a baseball manager who will keep a blazing hitter in the lineup until his streak cools, weather people will ride a hot model until others begin to outperform it.
Common models of reference are the European (ECMWF); Global Forecast System (GFS); United Kingdom, Canadian, North American (NAM); and Weather Research and Forecasting (WRF). The first two are used the most for short-term forecasting. There is also an extended 16-day GFS model that can be very helpful for planning agricultural operations — though the last six to eight days of that period are notorious for frequently being less than 100 percent accurate.
For truly long-range monthly and seasonal outlooks — a full year or more out! — folks at the Climate Prediction Center in Camp Springs, Md., whip up a monthly suite of forecasts on the third Thursday each month, when they are not using the supercomputer to track Washington Nationals baseball stats. They are certainly worth a look but, again, the more proximate the time frame, the higher the accuracy.
I wouldn’t recommend them to tell you if the caterer will need a tarp to protect the gefilte fish and chopped liver platters at your son’s backyard Bar Miztvah reception next May 19, or if your daughter’s $15,000 floral arrangement and 150 guests will suddenly have to be moved into your garage on her June 2013 garden wedding date.
But they can indicate trends corresponding to big climatic players like El Niño or La Niña, which can profoundly affect our weather for months in occasionally unpleasant ways, as was the case with our infamous “summer that never was” in 2011.