This study considers night-time urban air and road surface temperature variations under different synoptic weather conditions. Five years of winter weather data from a single weather station were classified into groups with similar air-mass characteristics. Using a two-stage process of principal components analysis followed by cluster analysis, the resulting clusters were used to sort paired air and road-surface temperature observations recorded at 2100hr from the West Midlands road weather information network into days with similar meteorological controls. The six resulting clusters have good between-group variation and feature a dominant wind direction. The results show that the difference between air and road surface temperatures is significant except during strong westerly conditions. Days featuring clear and anticyclonic weather showed the greatest between-site variation and evidence of the urban heat island effect. Under these conditions, road surface temperatures were warmer than the air - the opposite of what one would normally expect in rural areas. Those days with wet and windy weather were found to have the least between-site variation, excepting that due to altitude. The presence of a high vapour flux in the urban environment literally dampens the temperature response.