Non-urgent government operations are closed December 24 to January 1, reopening January 2. See the list of services available during this time.
Overview
Average dates of frost occurrence are often used to describe frost probability. However, they have limited value because they only denote a 50% risk level for a single frost temperature, usually defined as the freezing point (0 degrees C).
Average values for frost tells us that 50% of the time frost will occur before a specific date, and 50% of the time it will occur after that date. In reality, during the spring and fall frost, probabilities can change dramatically over a few days. This means average frost maps have limited use beyond giving us a general idea of how frost times vary over a wider area.
Using the Data Viewer
Rather than using average dates, the Alberta Climate Information Service (ACIS) estimates frost risk at 6 temperatures: 0, -1, -2, -3, -4 and -5 degrees C. Depending on the frost hardiness of the crop, the level of damaging frost can be selected by the user.
Figure 1. Weather station data viewer: generating frost probabilities
To start, go to the Current and Historical Alberta Weather Station Data Viewer:
- Select the Normal Tab.
- Select frost level(s).
- Select a station(s) location on the map; hovering over a location will make the station name appear.
- Select a date range.
- Click the Graph button to generate your graph, or click Download to download your data.
In the example shown in Figure 1, the graph in Figure 2 was generated.
Figure 2. Fall frost probabilities for 3 levels of frost severity
This type of graph provides detailed information on the risk of frost. For example, at Vegreville;
- The earliest a 0 degree C frost ever occurred was August 2 (red line).
- 4% of the time there has been a 0 degree C frost by August 9. (Roughly twice in the last 50 years).
- 10% of the time there has been a 0 degree C frost by August 20 (1 time in 10 years).
- The probability of a 0 degree C frost rises rapidly after September 2. In comparison, for a frost of ‑2 degrees C (blue line), the probability stays low until about September 8. For a -4 degrees C frost (green line) it is not until September 13 that we see a rapid rise in frost risk.
- 50% of the time there has been a 0 degree C frost event by September 13. For -2 degrees C that’s 6 days later (September 19); for -4 degrees C, a frost event does not occur until September 29.
Weather forecasts and historical frost probabilities
Weather forecasts are fairly reliable looking out 3 days or less. If on September 10 the local forecast is predicting overnight lows well above freezing over the 10-days, it’s likely that you will not see frost over the next day or 2, even if the historical probability is more than 40%. Keep in mind that the accuracy of forecasts begins to drop dramatically beyond about 3 to 5 days. We often see unpredicted weather events suddenly roll in, making the recent forecast incorrect.
This is one of the times that the frost probabilities in ACIS become useful (Figure 3). In this case, on September 10, the historical chance of having a frost of 0 degrees C on or before this date is 40%. But the frost risk rises rapidly after this date and by September 20, just 9 days later, the chance of experiencing frost has risen to more than 80%. This means that, on average, once in 4 years there has been a 0 degrees C frost by September 20. The risk then rises further to more than 95% by the 28 of September. In other words this means that, on average, only once in 20 years there will not be a frost of 0 degrees C by September 28.
Therefore it is wise to temper an overly optimistic long-range forecast with some data from the past. In this case, the frost probabilities tell us that there is a very good chance that frost will occur before the end of the month.
Figure 3. Plot of spring and fall frost probabilities
Contact
Connect with the Agro-meteorology unit:
Hours: 8:15 am to 4:30 pm (open Monday to Friday, closed statutory holidays)
Phone: 780-427-3556