[trhway]

>Take the extremes; is cancer incidence markedly lower in perpetually sunny cities like Phoenix or Los Angeles compared to perpetually cloudy cities like Seattle or Portland?

you need to compare not the whole populations of geographies - you nee to look at people who isn't evolutionary adapted (in this particular case - skin color to manage the sunlight and vitamin D level in particular) to the geographies they are currently living at.

For example, children autism is much higher in the Somali immigrant population in Minnesota and Sweden (vs. Somalis in Somali) as this immigrant population has, for obvious reasons, extremely low, frequently almost non-detectable, vitamin D levels.

If cancer rate has negative correlation with vitamin D levels, we'll, unfortunately, see it for sure in like a decade or 2 in those immigrant populations.