By September's end, AccuWeather already had their forecast for New Year's Eve out. In the District, they called for 0.29 inches of rain. No precipitation actually fell. (AccuWeather.com)

On Nov. 7, we published an article on the limits of long-range forecasting. The piece was in response to an Oct. 12 article by AccuWeather, which offered various specific forecasts well over a month in advance. Among them were predictions for Christmas and New Year’s Day.

Now that we’re beyond New Year’s Day, we wanted to take a look at how these long-range forecasts performed.

Peer-reviewed research has concluded the practical limit to forecasting specifics is on the order of eight to 10 days. So how did AccuWeather’s forecasts weeks and months into the future do? They were about half right.

For comparison, in our evaluation of AccuWeather’s forecasts below, we also present the predictions from the Old Farmer’s Almanac (beginning in November), which also publishes such long-range prognostications.

AccuWeather prediction 1: “Salt Lake City, Utah, is forecast to get its first measurable snow on November 18, with 0.42 of an inch.”

  • Temperatures were unseasonably warm, with a high of 60 degrees. The low was 32 degrees.
  • No precipitation occurred.
  • Salt Lake City’s first measurable snowfall occurred on Nov. 25.
  • Verdict? Did not verify.
  • Old Farmer’s Almanac forecast: “Periods of rain and snow north, sunny south; chilly.”

AccuWeather prediction 2: “There will be snow on Thanksgiving or the day after in Chicago …”

  • A trace of rain occurred on Thanksgiving, while a trace of snow occurred the day after.
  • Verdict? Verified (by a trace)
  • Old Farmer’s Almanac forecast: “[November] 26-29 Rain, then sunny, cold. [November] 30 Rain."

AccuWeather's article, posted Oct. 12, heralded Salt Lake City's first snowfall on Nov. 18. The article called for 0.42 inches of snow. (AccuWeather.com)

AccuWeather prediction 3: “There will be snow on Thanksgiving or the day after in Detroit …”

  • Temperatures never fell below freezing during this time period.
  • No precipitation occurred.
  • Verdict? Did not verify.
  • Old Farmer’s Almanac forecast: “[November] 26-29 Rain, then sunny, cold. [November] 30 Rain."

AccuWeather prediction 4: “There will be snow on Thanksgiving or the day after in Green Bay."

  • No snow fell on Thanksgiving, while 0.4 inches came down on Nov. 29, the day after Thanksgiving.
  • Verdict? Verified.
  • Old Farmer’s Almanac forecast: “[November] 26-30 Flurries, mild."

AccuWeather prediction 5: “A trace of snow is expected to fall on Christmas in Detroit."

  • No precipitation fell on Christmas in Detroit.
  • The high temperature was 54 degrees.
  • The next day set a record high at 58 degrees.
  • Verdict? Did not verify.
  • Old Farmer’s Almanac forecast: “[December] 24-27 Snowy periods, colder."

AccuWeather prediction 6: “And over the three-day New Year’s span of Dec. 30 to Jan. 1, measurable snowfall is forecast to occur in Boston …”

  • Measurable snowfall did not occur.
  • 1.29 inches of rain was observed on Dec. 30, along with only a trace of snow.
  • Verdict? Did not verify.
  • Old Farmer’s Almanac forecast: “[December] 28-31 Rain and wet snow, mild. [January] 1-5 rainy, mild."

AccuWeather prediction 7: “And over the three-day New Year’s span of Dec. 30 to Jan. 1, measurable snowfall is forecast to occur in … Minneapolis …”

  • 4.9 inches of snow fell on Dec. 30.
  • Verdict? Verified.
  • Old Farmer’s Almanac forecast: “[December] 25-31 Snow showers, mild. [January] 1-4 Snowstorm."

AccuWeather prediction 8: “And over the three-day New Year’s span of Dec. 30 to Jan. 1, measurable snowfall is forecast to occur in … Salt Lake City."

  • One inch of snow fell on Dec. 30.
  • 1.5 inches fell on Jan. 1.
  • Verdict? Verified.
  • Old Farmer’s Almanac forecast: “[December] 26-31 Snowy north, sunny south; turning milder. [January] 1-7 Snowstorm, then flurries, cold."

In the end, four out of AccuWeather’s eight forecasts were incorrect.

As for the Old Farmer’s Almanac, their forecasts were simply too broad-brushed to objectively evaluate. The predictions are so nonspecific, and incorporate so many possible outcomes, that it’s next to impossible to claim they’re right.