I saw this answer in a similar thread in a different forum : "Any measure of calorie burn is an approximation, but that doesn't change that it's worth having at least some measurement. No one's asking for perfection, just questioning what assumptions are made in the calculations"
I guess what it means is that if you're a pedant, and not interested in anything which has a possibility of error, then don't look for answers to this sort of question. Myself, accepting that there will be a margin of error, I'm very happy to use the Garmin figures, knowing they are based on time, distance, elevation, and heart rate, and factoring in my age weight and height, and also 'learning' more about me over time.
I do not expect the figures to be totally accurate in an absolute sense, however they will be in a relative sense (i.e. I can measure one session against any other session). Furthermore, accepting the margin for error, I would rather have approximate data, than no data.
Looking at some of the early answers, I'm not surprised that the OP didn't return.