Synoptic patterns identified by an automated procedure employing principal- component analysis and a two-stage cluster analysis,and backward trajectory analysis clustered by the HYSPLIT4.9 model were used to examine air quality patterns over (U)rümqi,China,one of the most heavily polluted cities in the world.Six synoptic patterns representing different atmospheric circulation patterns and air-mass characteristics were classified during the winter heating periods from 2001 to 2008,and seven trajectory clusters representing different paths of air masses arriving at (U)rümqi were calculated during the winter heating periods from 2005 to 2008.Then air quality was evaluated using these two approaches,and significant variations were found across both synoptic patterns and trajectory clusters.The heaviest air-pollution episodes occurred when (U)rümqi was either in an extremely cold,strong anticyclone or at the front of a migrating cyclone.Both conditions were characterized by with light winds,cold,wet surface air,and relatively dry upper air.(U)rümqi was predominately influenced by air masses from the southwest and from local areas.Air pollution index (API) levels were highest for air masses originating from the southwest with a longer path or for the local area,because of transport from semi-desert/desert regions by strong winds and because of local heavy pollution emissions,respectively.The interactions between these two analytical approaches showed that poor diffusion conditions,together with local circulation,enhanced air pollution,besides,regional air-mass transport caused by strong winds contributed to serious air quality under relatively good diffusion conditions.