Some people swear by it, others swear at it, and many suggest meteorologists scrap their technology tools and use this instead. "The Old Farmer's Almanac" has been issuing forecasts for months ahead of time for every region of the United States since 1792.
The book says they derive their forecasts "from a secret formula that was devised by the founder of this Almanac, Robert B. Thomas, in 1792" that has been refined over the years as technology advanced.
The main ingredient to their forecasts are based on sun spots and solar activity and how similar activity in the past has influenced weather patterns. But the forecast also takes into account the region's climate and other meteorological factors.
I make sure I have a copy every year (got to keep tabs on the competition!) and the other day I was poking through it to see how it was doing this past fall and winter -- each year's edition begins in November and goes through the following October.
Their forecasts are two-pronged -- they do a general monthly/seasonal forecast, then also do a more specific forecast for 2-7 day periods within the months.
I'm intrigued that they can attempt a forecast for such a narrow date window so far in advance (Even a 10 day forecast for Portland using traditional computer models are iffy at best), so naturally, I always like to see if their forecasts nail the big weather events. And I was surprised to find that, actually, this year they did reasonably well. Not perfect, but then again...
Then I wondered: How did they do on all the other days? Can you really use the Almanac as a guide to, say, find a guaranteed sunny weekend and save some money and not order tents for a wedding 9 months in advance?
So I went through and compiled what the Almanac had from November 1st through the end of March for their Pacific Northwest Region, which is generally considered western Washington and western Oregon west of the Cascade crest, as well as the sliver of the northwestern California coast around Eureka.
I then compared this to Portland's actual weather from that time period, which of course, might not exactly match what happened in, say, Eureka the same day. I also compiled this data for Seattle and the Almanac did a little better in predicting Portland's weather than it did for Seattle. But since we live around Portland, I figured this is how the Almanac would work for someone here trying to use their forecast.
The Almanac generally breaks down weather forecasts in 2-7 day clumps and gives general tendencies such as "cool and rainy" or "occasional showers, mild." I gave it an admittedly arbitrary scoring system from 1-10 as such:
8-10: Totally nailed it!
4-7: Some facets correct
2-4: Missed badly on temp or rain
0-1: Totally blew it!
I also mentioned how their monthly forecasts shaped up versus what was observed:
|Date:||Farmer's Almanac Forecast||Actual Portland Weather||Score|
|1st-3rd||Ocn'l rain, cool||Cool 1st, rain on 2nd||8|
|4th-9th||Heavy rain, mild||Ocn'l light rain, cool||4|
|10th-16th||Rainy periods, turning cool||Rainy periods, turning cool||10|
|17th-19th||Showers, chilly||Showers, chilly||10|
|20th-30th||Rainy periods, cool||Cold on 20th, rainy periods, seasonal||7|
|Overall||1 deg warm / 3" wetter||2.2 deg cool / 0.8" drier||0|
|1st-6th||Rainy periods, mild||Dry and cold||0|
|7th-11th||Ocn'l rain, mild||Still dry, cold||0|
|12th-15th||Rain, then sunny, cool||Sunny, then rainy, cool||3|
|16th-18th||Rainy, mild||Sprinkles 18th, seasonable||5|
|19th-23rd||Sprinkles, cool, then mild||Sprinkles, cool, then mild||10|
|24th-27th||Rainy, mild||Rainy, seasonable||9|
|28th-31st||Rain to snow, then showers, cool||Rain to showers, warm, then cool||7|
|Overall||2 deg warm / 3" wetter||1.3 deg cool / 3.0" drier||0|
|1st-2nd||Sunny, mild||Mostly cloudy, mild||8|
|3rd-8th||Rainy periods, mild||Rainy periods, mild early, then cooler, dry||7|
|9th-10th||Sunny, cool||Rainy, cool||3|
|11th-18th||Stormy, mild, then Ocn'l rain||Dry thru 14th, then rainy, cold, stormy late||4|
|19th-22nd||Stormy, mild||Stormy, mild (cold on 20th)||8|
|23rd-27th||Sunny, cool, then rain periods, mild||Cold 23, than rainy, mild, cold 27th||7|
|28th-31st||Sunny, cool||Cloudy, cool, then seasonable||5|
|Overall||2 deg warm / 2" wetter||0.5 deg cool / 2" wetter||5|
|1st-6th||Rainy periods, seasonable||Sprinkles 1-2, then sunny, seasonable||4|
|7th-11th||Rain & wet snow, chilly||Rainy, seasonable||4|
|12th-16th||Sunny, seasonable||Rainy except 15th, cool||0|
|17th-23rd||Ocn'l rain and drizzle, cool||Ocn'l rain, drizzle, seasonable (Warm 21st)||5|
|24th-29th||Sunny, cool||Ocn'l rain & Cool (but partly sunny 27th)||2|
|Overall||2 deg cool / 2.5" drier||0.5 deg cooler / 1" drier||5|
|1st-4th||Rain and wet snow, cool||Rain, wet snow, cool (Mild 3rd/4th)||6|
|5th-11th||Sunny then rainy periods, mild||Rainy/cool, then sunny warm 8-9,||5|
|rainy, getting cool 10-11|
|12th-23rd||Rainy periods, cool||Rain 12-22, cool, wet snow at times||10|
|24th-27th||Sunny, nice||Nice on 24th, dry 25th , showers 26-27||4|
|28th-31st||Rain, then sunny, seasonable||Rainy, seasonable||7|
|Overall:||2 deg cool / 1" wetter||3 deg cool / 4.2" wetter||7|
If you average out my scores, not counting the overall scores, it comes out to about a 5.6 -- or I guess what you'd consider a little better than 50/50. The Almanac did do pretty well in predicting the early March snows as well as the mid November minor snow events. It didn't do so well catching our very dry December nor a rainy stretch in mid February.
So, I wouldn't go and plan a wedding on Sept. 19 and skip any rain contingencies just because the Almanac's forecast for this year in the Northwest is for "sunny, very warm" but I suppose you could hope the Almanac scores along the lines of its March performance than its December performance.
(I did also check it against some other notable national events. For example, it did accurately get the "sunny, warm" period for the Great Lakes with their record heat in mid-March. It also caught the very dry winter across Texas.)
Any better on the general monthly forecasts?
Overall, the Almanac claims about 80 percent accuracy based on their emphasis of temperature and precipitation deviances from averages and normals.The Almanac says they did well in the winter of 2010-11 where they claimed 90.6 percent accuracy in their winter and precipitation forecasts, but that number isn't really what you're thinking.
They said they got the departure from normal in temperature correct in all 16 of their regions, while they were correct in 13 of 16 regions on their precipitation forecasts, essentially scoring themselves as 39 out of 42 correct.
For this year so far, at least in our region, their scores aren't as sparkling. It predicted a warm and wet November and December, but we ended up generally cold and dry. It erroneously stuck with the warm theme into January, but at least got the rain right.
It finally got in sync with our actual weather in correctly predicting a cool and wet March, although it didn't quite grasp just how cool and wet it'd be-- wettest on record in Portland and 3rd wettest in Seattle.
For comparison, the NOAA/Nat'l Center for Environmental Prediction forecasts were generally correct in predicting cool winter -- mainly based I'm sure on the La Nina that had come around for a second helping. They were a bit too wet in November and December too but closer than the Almanac.
Hot summer on tap?
From here on out, the Almanac is predicting a cooler but drier than normal April and May. (NOAA agrees with the cool, but is still going wetter than normal.)
But to those of you looking ahead and ready for some warmth can only hope the Almanac is correct for this coming summer. It is predicting a warmer than normal summer with August a particularly a hot month this year at 5.5 degrees above average on the whole.
As for the hottest times? Early and late July, and early and mid-August, according to the Almanac.
(For the record, NOAA isn't predicting one way or the other this summer for our area, but does agree with a hotter and drier than normal summer across much of the Desert and Intermountain West. And NOAA has an advantage that they get to update their forecasts each month whereas the Almanac gets one shot in the late summer to generate their forecasts.)
The Almanac winds down summer with a warm but wet September and October. As for next winter? We'll have to wait until 2013's edition comes out.
In the meantime, I'd still have those tents on standby. It is Portland, after all.