Your Reading List

Fall frosts and the frost-free season

The number on your thermometer may not accurately reflect what’s happening on the ground

Unharvested soybeans on a cold winter morning.

Every year around this time the weather discussion starts to centre around the first fall frost. Some years, when we see an early-fall frost, the topic is at the forefront of conversations. This year, with only a couple of locations experiencing a fall frost as of Sept. 25, there hasn’t been much discussion.

The first question I’m usually asked about fall frost is when the different areas of agricultural Manitoba should expect to receive their first fall frost. To analyze this, we must first determine how frost is to be measured or recorded. The typical measurement we use to record whether or not frost has occurred is when the temperature recorded by a thermometer hits or falls below 0 C. As some of us have already had the unfortunate first-hand opportunity to realize, frost can occur even when the thermometer shows temperatures above the freezing mark. In fact, research has shown ground-level frost can occur at thermometer readings as high as +2 C, and in some cases, as high as +5 C! This can occur for a number of reasons, which largely depend on where the thermometer is located. As most thermometers are placed above the ground, they record air temperature several feet above the ground and may not accurately reflect actual ground temperature.

If you can remember back to previous discussions about frost, you may recall that air near the ground can cool to a greater degree than the air several feet above. The reason for this is that cold air is denser than warm air, so it tends to settle or flow to the lowest points. If the area is relatively flat, then the coldest air settles around the ground, resulting in ground-level temperatures that are cooler than the air several feet above. While this is the norm, there are occasions when temperatures measured above the ground, at the level of the thermometer, are actually cooler than those recorded at ground or crop level. Also, as some of us have seen this year, a frost with temperatures near the freezing mark may not severely damage or kill a crop. It will often take temperatures lower than -2 C to kill off most crops.

For these reasons we will look at a few different temperatures — namely, +2.0 C, 0.0 C, and -2.0 C — to determine when we may expect the first fall frost. Looking at the data for the three main stations we obtain the results shown in Table 1. These are the average dates that these temperatures may be anticipated, based on the entire record of climate data for each location.

Now, we need to remember that this is the average date and the standard deviation around these dates is somewhere around three to five days, depending on the location. This means that while most locations should not see any frost until early to mid-September, it would not be that unusual to see a frost in early September at most locations.

Table 1: Date of average first fall frost

So far this year, these three sites have not experienced a fall frost. In fact, looking at the data for all the agricultural sites I was only able to find six sites that have seen frost. Two of them (Moosehorn and Narcisse) are in the Interlake region, with the other four (Grandview, Minitonas, Rorketon and Swan Valley) in the northwest region. All these frosts occurred overnight from Sunday, Sept. 17 to Monday, Sept. 18.

The next piece of frost data to look at is the length of the frost-free season. Last year most parts of the Prairies saw frost-free seasons that were a little bit longer than the long-term average. Winnipeg ended up having the longest frost-free season, with a remarkable 144 days, which put it in the top 10 per cent. Listed in Table 2 is this year’s data, showing the dates for the last spring frost (LSF), first fall frost (FFF), and the length of the frost-free season (FFS). I have also included the 1981-2010 averages. The bold values indicate that the finals numbers are not yet in for this year.

Table 2: Frost-free seasons on the Prairies.

The final two columns give us a range of expected lengths for the frost-free season: the 90 per cent value means 90 per cent of the time, we should have an FFS no shorter than this; the 10 per cent value means 10 per cent of the time we could see an FFS that long.

Edmonton and Regina are the only two locations that had a shorter-than-average frost-free season, while Peace River has a much longer-than-average season thanks to an early LSF. For our region, all three major locations have now seen a longer-than-average frost-free season. The question is, just how long will it end up being?

About the author

Co-operator contributor

Daniel Bezte

Daniel Bezte is a teacher by profession with a BA (Hon.) in geography, specializing in climatology, from the U of W. He operates a computerized weather station near Birds Hill Park.



Stories from our other publications