Does a cool summer equal a cold winter?

mean temps

For many, this summer of cooler than normal temperatures have been a welcome sight compared to the blazing hot summer of 2012. The cooler weather has had its downsides…some fruit trees are ripening early, ragweed is set for a brutal late summer/fall allergy season. One thing many are dreading is the assumption that this cool weather surely portends a cold winter, right?

After pouring over data since 1950 at Dayton International Airport (the official measuring site for Dayton), I can say that answer is no. There is no correlation between a cold summer and a cold winter or even a cold summer and mild winter. There is one primary reason for this…but first, let’s look at the stas.

I pulled monthly data for May, June and July to represent our cool summer thus far and winter will be represented by December, January and February, which is Meteorological winter. I did not include August because the month is not over, but it sure looks to be cooler than normal as well. What follows is the temperature data that NOAA keeps and what it all means. Graphs are made in Minitab, which is a tool I use at work. I like it so much that I put in on my personal computer and am putting it to use here.

Days in Summer with Highs above 90 compared to Days in Winter with lows below freezing.

The Stats:

Max 90 Min 32 DJF

Max 90 Min 32_1


Statistical Analysis:
What you see here is called a Negative Correlation. The Pearson Correlation here is small, at -0.225. The closer to -1.0 that number is, the stronger the correlation. The closer to 0, the weaker the corelation.  On top of this, the P-Value, which is an indicator of confidence in the data, is .077.  This is a very low confidence (.05 is considering meaningless).

In Layman’s terms: 

There is a very, very small relationship between the lack of high temps over 90 in this timeframe and a higher number of lows under freezing in winter (which is a fairly common occurrence to begin with). This is NOT conclusive by any means. For instance, in 2000 and 2004, we had zero days with high temps over 90! The corresponding winter of 2000 had 81 days with lows under freezing, while 2004-05 had only 68 days. The only other years to have no highs above 90 in May, June & July were 1958 & 2008. That’s right, 3 of the 4 years since 1950 have occurred since 2000! We had 9 days with high Temps over 90 in May, June & July of 2013…under the average of 10.19 days.

Days in Summer with Highs above 90 compared to Winter days with lows below zero.

The Stats:

Max 90 Subzero

Max 90 Subzero_2



Statistical Analysis:

While there is another negative correlation here, it is weaker than the first batch of statistics with a slightly higher P-value. Neither screams confidence and cannot be considered conclusive.

In Layman’s Terms:

There is a very small correlation between the lack of Highs above 90 and subzero lows the following winter. It is not nearly enough to consider cool summers = cold winters a truism.

Days in Summer with Highs above 90 compared to Winter days with Highs never get above freezing.

The Stats:

Max 90 Max Under 32


Statistical Analysis:

Once again, there is a negative correlation that is not strong or high in confidence.

In Layman’s Terms:

Like the two data sets preceding it, below freezing highs do not have a concrete correlation to a lack of very warm highs in May, June and July.

Mean Temperatures over May, June & July & Corresponding Winter:

The Stats:

Mean Temp

Mean Temp_2mean temps




Statistical Analysis:

When I first made the Scatterplot, I was excited to see a small positive correlation, but a sub .05 P-value renders any confidence meaningless.

In Layman’s Terms:

When I graphed out a Run Chart of the Mean Temps to compare Summer mean temperature versus winter mean temperature….two things stood out: 1. Summer Mean Temperature hasn’t trended higher or lower AT ALL for Dayton since 1950. It’s in a flat run with little volatility. 2. Winter temperatures swing wildly wear over year with very high volatility.

The Conclusion of the Matter:

While not having any major correlations with confidence may seem disappointing at first, it is not. Having no correlations across the board means there simply isn’t a link. For those of us in the field of meteorology, this really shouldn’t be a surprise. This is because our winters are driven by the jet stream…but our most brutal winters are due to the negative phase of the Arctic Oscillation (AO) aligning with a big trough in the jet stream (signaled by a Negative North Atlantic Oscillation). What feeds this negative AO with very cold air masses is the “Polar Vortex” above the Arctic Circle. In the Negative AO phase, it get dislodged from the Arctic and plunges into the mid-latitudes. Depending on the jet stream positioning, this can make for a very cold few weeks to a month. The Brutal cold of January of 2010 followed by two major snowstorms in February of 2010 comes to mind. What causes the Polar Vortex to be so strong and dislodge? Two key drivers are the Arctic Sea Ice and an event known as Sudden Stratospheric Warming. That can be a post for another day, because that is a very science-heavy topic to discuss.

If this seems like a long way to go to prove a point…your hunch is a good one. :) However, I’m not one to be satisfied with just a “yes or no” answer. I hope you learned something new and you didn’t hurt your brain (mine does now). But now, you can tell your friends the truth…better yet, share this article with them!

One Response to Does a cool summer equal a cold winter?

  1. Share e says:

    Where were you when I took my stats class?? This is great info.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>