My main intention when I started responding was to point out the much larger winter heating differential in the north, in response to the discussion about heating versus cooling costs and efficiency. I am not sure why I got distracted about Miami versus Phoenix and inserted that before posting.

My experience living for a few years in a tropical country was that locals acclimate and do not cool nor dehumidify their living spaces nearly as much as many Americans seem to do. Much like in Florida, you will see people running around in jackets or even knit hats on slightly cooler days when someone from a colder climate would already think it is warm and time for shorts.

I think it is difficult to compare different regions with different climates and cultures. It seems impossible to me to choose a metric that isn't inherently biasing the analysis towards one arbitrary normative standard. Compare similar regions or one region year-to-year to evaluate the efficiency of local practices.

Edit: another issue is the daily cycle. You can use thermal mass to smooth out daily temperature extremes but that doesn't work when you spend weeks or months with temperature differentials that remain offset from comfortable.

Heat transfer, though, is based on temperature differentials - if Miami is +7degrees every day of the year, and Phoenix is +20degrees only 117 days per year, the energy demands of Phoenix for cooling could very easily be much higher (energy lost through walls, roof would be much higher, as would mass transfer of energy when doors open.) This is irrespective of local construction. Add to this, radiative gain (insolation through windows) is not captured in the degree-days statistic, nor is humidity related cooling (condensing water takes a lot of energy!)

No. No no.

Heat conduction rate is linear in response to the temperature difference [1], so 1 day of 20 degrees difference should give the same total heat loss as 2 days of 10 degrees difference. They also both give 20 degree days, so degree days is the correct measure and should correlate linearly with the energy needed for heating or cooling. The same goes for the heat content in the air that is exchanged due to draft, opening doors etc.

You might have a little bit of a point with sunlight, and humidity. But my guess is that they don't dominate, compared to the conductive and convective heat exchanges.

[1] https://en.wikipedia.org/wiki/Thermal_conduction#Fourier's_l...

Right, thank you for the correction. I was mistaken about the conduction. I'm not sure why I was thinking it was something other than linear. So degree days should account fairly closely for the conductive heat transfer through walls.

I think my comment about mass transfer when opening doors could maaaybe be defensible but I'll leave it alone :)