Canada's Forest Service hardiness zones are calculated
from more variables than the Unitd States Department of
Agriculture's hardiness zones. From the interesting
reference you gave, I see that they are using high and
low temperatures, amount of snowfall and rain and other
weather measurements.
If you have studied statistics you will recognize
"linear regression" (that is where they get the constants)
and "planar spline interpolation", which I have never done,
but it has to do with generating the nice smooth areas.
And if you realy know statistics you will exclaim:
"What, only 104 plants?"
USDA method is much simpler - average of the minimum
temperature for the previous 10 or 20 years.
After all that - "Plants don't read".