Realized I made a dumb mistake on the math and edited the post – probably should be more like 1"/mile/10 F temperature increase!
I think a substantial covariate in parts of the world that might lead to a bigger signal is that the hottest summer conditions come along with relatively weak large-scale winds, so temperature might appear to have a slightly larger role.
I also don’t know the answer to this, but does temperature have a substantial effect on rolling resistance?