Basically, it's something that's taken into account. The two main ways are calibrating urban stations relative to nearby rural stations, and by looking at the variation between windy and calm days, since the effects are larger on calm days (so if things are corrected well, there shouldn't be a difference)