Does Fall Weather Correlate to Winter Severity?



Republished with data through the 2019-2020 winter.

As we go into the winter season, it’s time to talk about how this one might end up. There’s a belief that fall weather is a good sign of how cold or warm winter will be. Let’s see if that holds true.

First, let’s just look at October temperatures.
The October normal mean temperature for Columbus is 55 degrees.

Between 1878 and 2019, there have been 47 Octobers that featured a mean temperature of 53.9 degrees or lower, what we’re considering a Cold October for the purposes of this comparison.
Of those 47 Octobers, 27 of the 47 had following winters that were colder than normal, or 57.4%, 13 had average temperature winters, or 27.7%, and the remaining 7 were warmer than normal, or 14.9%.
Interestingly, this category contains both the warmest winter on record- 1889-1890 and the coldest on record- 1976-1977- as shown by the chart below.

Next, we look at Normal Octobers, which are +/- 1 degree of the 1981-2010 Average of 55 degrees.
Between 1878 and 2019, there were 45 normal Octobers. Of those, 21 had colder than normal following winters, or 46.7%. 11 were followed by normal winters, or 24.4%, and 13 had warmer than normal winters, or 28.9%.

Finally, let’s look at warm Octobers, which are those with means of 56.1 degrees or higher. There were 49 Octobers with warmer than normal means since 1878. Of those, 18 featured following winters that were colder than normal, or 36.7%. Another 19, or 38.8%, were followed by average winters. The final 12 winters, or 24.5%, were warmer than normal. Here’s the graph.

So just based on the October mean temperature, Octobers that are colder than normal are almost twice as likely as normal Octobers and almost 4x as likely as warm Octobers to be followed by a cold winter. But what about Novembers? It is closer to winter itself, so does its weather matter even more?

Colder than normal Novembers- 43.3 degrees or lower- included 80 Novembers since 1878. Of those, 38 or 47.5% had colder than normal winters. 22 (27.5%) had normal winters and 19, or 23.8%, had warmer than normal winters.

With the 39 normal Novembers, 43.4 to 45.4 degrees, there were 18 that had colder than normal winters, or 46.2%, with 12 normal winters (30.8%) and 9 warmer than normal winters (23.1%).

Finally, there were 24 warmer than normal Novembers since 1878- 45.5 degrees or higher. Only 6, or 25%, were followed by cold winters. An additional 9 (37.5%) were normal, while the last 9 (37.5%) were warmer than normal.

To reiterate, here are the ranked percentages of cold winters by the preceding October or November.
1. Cold Octobers: 57.4%
2. Cold Novembers: 47.5%
3. Normal Octobers: 46.7%
4. Normal Novembers: 46.2%
5. Warm Octobers: 36.7%
6. Warm Novembers: 25.0%

It should be no surprise that cold Octobers and Novembers have a stronger correlation to the following winters also being colder, with colder winters becoming increasingly unlikely as those months warm. What is surprising is that October weather, across the board, was more predictive of cold winters than Novembers were.

Going further, though, what about bi-monthly combinations?

Rank of Bi-Monthly Combinations and the percentage of colder than normal following winters, along with total years in sample:
Normal October/Normal November: 87.5% 8 Years
Cold October/Warm November: 57.1% 7 Years
Cold October/Cold November: 53.8% 26 Years
Normal October/Cold November: 48.1% 27 Years
Warm October/Cold November: 40.7% 27 Years
Cold October/Normal November: 38.5% 13 Years
Warm October/Warm November: 28.6% 7 Years
Warm October/Normal November: 25.0% 16 Years
Normal October/Warm November: 0.0% 8 Years

So a normal fall is clearly the best, but the sample size is not particularly high. Normal to Warm is unanimously warm, but again, it has a small sample size.

So does fall weather ultimately have a correlation to winter severity? To some degree yes, as patterns established in the fall tend to appear in the winter as well, but October, rather than November, has a clear higher correlation. But as with everything, there are many other factors involved.

October 2020 looks to at least begin with below normal temperatures, but it remains to be seen how it ends up.

To see records for Columbus falls and winters going back to 1878, visit the following links.
Autumn Weather Records
Winter Season Records

And to see current local weather and forecasts, visit: Wilmington National Weather Service