Quite a lot because of thermal expansion/contraction of air and differentials in incident sunlight heating the surface at different points at different times. A

Quite a lot because of thermal expansion/contraction of air and differentials in incident sunlight heating the surface at different points at different times. Albedo, thermal mass of the objects being heated, etc., all contribute to a complex system of air and energy transport via air currents. It would take quite a lot of planning and work but maybe there are ways to mitigate these losses by balancing out the effects between different processes. Typically nature finds such dynamic equilibria as a matter of course -- I suspect anything humans attempt to control will not be as resilient or sound.