With a rather brutal and gloomy start to fall across the Prairies, and conditions often feeling more like winter than summer, I’ve begun to hear more people talk about the return of unseasonable heat that’s called St. Martin’s summer in the U.K., second summer and other terms now outdated in North America. Some of the conversations are about just what it is, while other conversations are more about whether we’ll see one this year.
Let’s look at the first part and try to figure out just what defines that last taste of summer we’re accustomed to seeing mid-autumn. We have looked at this topic a couple of times over the past 10 or so years, but I figured we were due to revisit it. The most common response I hear or see is that second summer is a period of warm weather that occurs sometime in late September or October. After researching the details of the phenomenon, I found that this description is actually not that bad — but that’s the problem with the term. Just as there are many things to call it, there doesn’t seem to be a set definition. The definition appears to be set by the geographical region rather than by a definite set of criteria, with different areas of North America defining it a little bit differently.
For the area of the Great Plains and the Prairies of Canada, it is defined by a particular set of criteria that should help us to determine when it is occurring. It’s generally accepted that for the phenomenon to occur, the following conditions (as compiled by Keith C. Heidorn) must be met:
- There needs to have been a hard or killing frost;
- Mostly clear skies (or perhaps local fog at night);
- No precipitation;
- Light winds and generally calm nights;
- Daytime maximum temperatures greater than 18 C;
- Nighttime minimum temperatures staying above freezing; and
- Conditions lasting for at least three days.
The idea of a mid-autumn warm spell isn’t new. The earliest reference found in the literature dates back to January 17, 1778, when J. Hector St. John de Crèvecoeur wrote in a letter:
“Sometimes the rain is followed by an interval of calm and warmth… ; its characteristics are a tranquil atmosphere and general smokiness. Up to this epoch the approaches of winter are doubtful; it arrives about the middle of November, although snows and brief freezes often occur long before that date.”
This particular reference fits in nicely with what we have seen so far this fall across the Prairies, as we have definitely seen snows and brief freezes (maybe not that brief). Perhaps there is some reason to hope that we’ll still see second summer this year.
The meteorological conditions behind the phenomenon usually involve a ridge of high pressure building into our region in late September or October and sometimes, even November. A fairly strong ridge of high pressure in the upper atmosphere over western North America in mid- to late October, combined with a strong area of surface high pressure, can bring some incredible weather to much of the Prairies. The warmest October on record was back in 1963, when all the major reporting centres, with the exception of Calgary, broke their October heat records. The warmest November ever recorded occurred in 1917 for Saskatoon and Calgary, 1899 for Winnipeg and Regina, and 1987 for Edmonton. Edmonton doesn’t have as long a data record as the other stations do, so it’s possible that its warmest November could have also occurred in 1917 or 1899.
One of the reasons I brought up the warmest Octobers and Novembers on record is that one of the more dependable medium-range weather models has been fairly consistent with showing a return to well-above-average temperatures starting in the last week or so of October and peaking in intensity during the first half of November. The weather graphic in this article shows the forecast temperature anomalies for the first full week of November and it shows a large blob of warm air stretching from the Yukon south and eastward to cover pretty much all three Prairie provinces. Looking at the average temperatures for this time period, and adding on the forecasted anomalies, daytime highs would be in the 8 to 12 C range, with overnight lows around -4 C. Considering this would be the average, we would likely see some days much warmer than this with the odd colder day. Not too bad for November, if this weather model actually turns out to be right.