Fall Weather Correlation to Winter Severity?

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. How true is that for Columbus? Also, what might any correlation mean for the winter of 2017-2018?

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

Between 1878 and 2016, 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 2016, 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 46 Octobers with warmer than normal means since 1878. Of those, 18 featured following winters that were colder than normal, or 39.1%. Another 18, or 39.1%, were followed by average winters. The final 10 winters were warmer than normal. Here’s the graph.

So just based on the October mean temperature, Octobers that are colder than normal have a 47% higher chance of having a colder than normal winter than warmer than normal Octobers do. But is October a better indicator than November, a month that is closer to actual winter?

Colder than normal Novembers- 43.3 degrees or lower- included 78 Novembers since 1878. Of those, 38 or 48.7% had colder than normal winters. 21 (26.9%) had normal winters and 19 (24.4) had warmer than normal winters.

With the 38 normal Novembers, 43.4 to 45.4 degrees, there were 18 that had colder than normal winters, or 47.4%, with 11 normal winters (28.9%) and 9 warmer than normal winters (23.7%).

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: 48.7%
3. Normal Novembers: 47.4%
4. Normal Octobers: 46.7%
5. Warm Octobers: 39.1%
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. Cold Octobers have a higher correlation than Cold Novembers, as well as Warm Octobers, but Normal Novembers have a slight advantage over Normal Octobers. Based on this, October actually has a stronger correlation to the following winter’s temperature mean than does November.

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: 44.0% 25 Years
Cold October/Normal November: 38.5% 13 Years
Warm October/Warm November: 28.6% 7 Years
Warm October/Normal November: 26.7% 15 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.

October 2017 has been overwhelmingly warm. While this wouldn’t normally bode well for a cold winter, each year is influenced by a multitude of factors.

Historic Record of Early Season Cold

With temperatures predicted to fall to near freezing for the first time this week for the fall season, I thought it might be interesting to take a look at the incidence of early-season cold, and the average on when it tends to arrive.

Here are the earliest dates on record for the following:

Temperatures Below 40 and Temperature Range of Observed Dates
1. 8/29/1965: 39
2. 9/8/1951: 39
3. 9/9/1883: 39
4. 9/13/1964: 38
5. 9/14/1902, 9/14/1923, 9/14/1953, 19/14/1975: 38-39
6. 9/17/1959: 37
7. 9/20/1896, 9/20/1956, 9/20/1962: 37-39
8. 9/21/1889, 9/21/1897, 9/21/1991: 37-38
9. 9/22/1918, 9/22/1974, 9/22/1976, 9/22/1995: 37-38
10. 9/23/1885, 9/23/1913, 9/23/1963, 9/23/1967, 9/23/1981, 9/23/1989: 35-39

Average Date of First Under-40 Temp By Decade (1878-2014)
2010s: October 11th
2000s: October 8th
1990s: October 2nd
1980s: September 30th
1970s: October 1st
1960s: September 25th
1950s: September 25th
1940s: September 30th
1930s: October 11th
1920s: October 2nd
1910s: October 9th
1900s: October 5th
1890s: October 1st
1880s: October 2nd
1870s: October 1st

Highs Below 32
1. 10/30/1917: 32
2. 11/3/1951, 11/3/1966: 28-29
3. 11/4/1991: 27
4. 11/6/1967: 31
5. 11/7/1971: 31
6. 11/8/1976: 32
7. 11/10/1913: 27
8. 11/12/1920, 11/12/1921, 11/12/1932: 30-32
9. 11/13/1911, 11/13/1919, 11/13/1986, 11/13/1996: 25-32
10. 11/15/1880, 11/15/1893, 11/15/1916, 11/15/1933, 11/15/1940, 11/15/1969: 24-32

Average Date of First 32 or Below High By Decade
2010s: December 4th
2000s: December 2nd
1990s: December 7th
1980s: November 28th
1970s: November 29th
1960s: November 23rd
1950s: November 26th
1940s: December 2nd
1930s: November 27th
1920s: November 28th
1910s: November 22nd
1900s: November 30th
1890s: November 25th
1880s: November 30th
1870s: December 4th

Lows Below 32
1. 9/21/1962: 31
2. 9/29/1961: 32
3. 9/30/1888, 9/30/163: 31-32
4. 10/1/1899: 30
5. 10/2/1886, 10/2/1908, 10/2/1974: 31-32
6. 10/3/1975, 10/3/1981, 10/3/2003: 32
7. 10/4/1952, 10/4/1987: 29-32
8. 10/5/1965, 10/5/1968: 31-32
9. 10/6/1892, 10/6/1964, 10/6/1980, 10/6/1988: 30-31
10. 10/7/1889, 10/7/1935: 30-31

Average Date of First 32 or Below Low By Decade
2010s: October 24th
2000s: October 26th
1990s: October 22nd
1980s: October 17th
1970s: October 17th
1960s: October 8th
1950s: October 22nd
1940s: November 3rd
1930s: October 24th
1920s: October 28th
1910s: October 31st
1900s: October 24th
1890s: October 20th
1880s: October 20th
1870s: October 26th

Summer of Rain 2015

This summer has been wet… very wet. Rain has fallen on far more days than not, and cloud cover has hung thick and heavy in what are historically much sunnier months. To top it all off, temperatures have been somewhat below normal.

How wet has summer 2015 been up through today? Meteorological summer runs June 1st-August 31st, and today makes the midpoint of that period, so I thought it would be a good time to check in and answer that question.

Here are the top 10 wettest summers through yesterday.

1. 1958: 12.54″
2. 2008: 11.68″
3. 1990: 10.93″
4. 2015: 10.66″
5. 1949: 10.40″
6. 1973: 10.23″
7. 1917: 10.06″
8. 1928: 10.02″
9. 1937: 9.75″
10. 1969: 9.38″

So 2015 is running the 4th wettest. It looks to be dry the next few days before more thunderstorms and rain return, so 2015’s position may end up rising.

What are the wettest full summers of all time? A few of the above years went on to be in the top 10, but not all.

1. 1958: 22.02″
2. 1979: 20.92″
3. 1995: 19.85″
4. 2003: 19.39″
5. 1969: 19.09″
6. 1915: 18.92″
7. 1992: 18.44″
8. 1973: 17.81″
9. 1989: 17.45″
10. 1949: 16.54″

Summer 2015 needs about 6″ more to reach into the top 10 wettest summers, which at the current pace is more than possible.

Questions Answered: Columbus Snowfall By Season

I get *tons* of searches for Columbus snowfall history, and while I’m still building up individual monthly records in this regard, here are the season snowfall totals for every season since 1878-79.