Also, people often don't realize that the temperatures that matter thermodynamically are degrees Kelvin, not Celsius. 30 below 0 is 243K; 20 above (room temperature) is 293K. The coefficient of performance for a Carnot heat pump at those temperatures is 293 / (293 - 243) ~= 5.86; that means that one joule of work will transfer 5.86 joules of heat. This isn't great, but it's less of a difference than people would expect given our subjective experience of how cold -30C is. The coefficient of performance for a heat pump trying to heat 0C ambient temperatures up to room temperature (or alternatively, an air conditioner cooling 40C temperatures to room temperature) is 293 / (293 / 273) = 14.65, so your heat pump in Calgary will be maybe 2.5x less efficient than one in NYC. Drill below the frost line, like you suggest, and you'll get similar efficiencies regardless of where you are.

No it is not feasible. You are not digging up ground under then tens of thousands of houses that already exist in the city in any economically feasibly way.

Community "large loop" heat pumps should help with the economics, but there are still lots of details to fill in.

I can come up with napkin tier ideas too. But lets leave the city planning to the professionals

No need to dig, only to do some drilling. It is very common here in Sweden. You just need to check the map first so you don't drill into the subway.

Dude, you're arguing semantics. A drill is still digging up dirt.