Canada Markets

Potential 2020 Canola Yields Based on Historical Trends

Cliff Jamieson
By  Cliff Jamieson , Canadian Grains Analyst
Connect with Cliff:
A range of prospective average Canadian canola yields for 2020 are arrived at based on the 10-year trend (black line), 20-year trend (grey line) and 30-year trend (brown line.) These trends point to an average 2020 yield ranging from 40.6 bushels per acre to 43.1 bpa. (DTN graphic by Cliff Jamieson)

Based on Statistics Canada's November estimates, Canada's average canola yield for 2019 was 40 bushels per acre, up 0.2 bpa from 2018 and above the five-year average of 39.44 bpa. The year-over-year change in yield in Manitoba resulted in a 1.3 bpa drop in average yield to 42 bpa, while the average yield in Saskatchewan was estimated unchanged at 39.3 bpa and the average estimated yield in Alberta was 1.5 bpa higher at 40.3 bpa.

When compared to the five-year average, the average yield in Manitoba for 2019 exceeded the province's five-year average by 2.2 bpa, the average yield in Saskatchewan exceeded its five-year average by 0.76 while the average Alberta yield fell short of its five-year average by 1.12 bpa.

As seen on the attached chart, this year's yield tracked the 30-year trend, as seen by the brown line. According to Excel's linear 30-year trend, yields are increasing by an amount to 0.7 bpa annually, while this trend was extended to arrive at an average yield of 40.6 bpa in 2020.

The grey line on the chart represents to 20-year trend, which is increasing at a rate of 0.9 bpa annually. Projecting this trend forward into 2020 would result in an average yield of 42.7 bpa for 2020, which would be a record national yield.

The 10-year trend, the steepest line of the three, is seen increasing at a rate of 1.06 bpa annually. This trend can be extended to project an average yield of 43.1 bpa in 2020, which would result in a record yield.

Of the three linear trend lines drawn, the highest R-squared or coefficient of determination is calculated for the 30-year trend at 81%, or in other words, the model explains 81% of the variability in the results for the long-term trend. The R-squared for the 20-year trend is close at 78%, while is only 50.6% for the 10-year trend.

Cliff Jamieson can be reached at

Follow him on Twitter @CliffJamieson



To comment, please Log In or Join our Community .