The average of the mean temps for all stations in the prairies of Canada shows no increase.

However, the extreme ends do show interesting trends.

The summer temps show an increase, while the winter temps show a decrease. Except, this is a meaningless plot. It is not a depiction of what is happening for the region. Because very few stations have a long range, this set of averages is skewed by those stations that dominate the early and late years of the dataset.

Thus, we cannot combine the dataset without “inventing” data to fill in the missing records. We already know this cannot be done. In science, you do not invent data to fill in the holes.

Thus we need to look at a specific station that has the longest recordset.

The station with the longest series of records is ID# 2205, Calgary Intl Airport. This is the yearly mean graph. It’s rising over all.

Except this increase in the average mean is again caused by an increase in winter temps not being so cold. Summer temps are flat.

Another station with a long dataset is ID#3080, Chaplin Saskatchewan. Here is the average of the mean temps:

The two spikes in the 1960’s is due to missing records. Notice the strong upward trend in this average of the mean temps. Want to guess what it causing this? Yep, a narrowing in the range:

This narrowing begs the question. When will the warmest and coldest join? That is at what time in the future and a what temp will the two regression lines intersect?

Equating the two -0.0173x + 36.088 = 0.0622x – 40.117 and solving for x gives some time in July of 2858 at a temp of 19.5C. Since it is impossible for the winter to be hotter than the summer, then the two lines cannot cross and continue on this trend. They either level off some time in the future, or turn in the direction of the slops (the most likely).

So again there is no global warming, and the averaging of all stations, which contain incomplete data, is skewed by the few stations that have the longest series of records. Hence using the average of the region gives a meaningless plot.