I dont think much influence, considering 1 day is <1% of the year. Perhaps think of it like a TMY file, where each of the 12 months are all from different years and get spliced together for 6 hours either side of the transition. Other than this, the day's weather either side of the transition are totally unrelated. Check the work by Marion and Urban (1995) here, specifically the section called Monthly Interface Smoothing.
