> If overnight a sizeable percentage of cars were no longer on the roads, would there be a noticeable increase in air quality?

We've already done this experiment a number of times. C.f. the hospitalization rates during the Atlanta Olympics.