Spatial-temporal variations and mineral dust fractions in particulate matter mass concentrations in an urban area of northwestern China.

PM10 and PM2.5 concentration data were collected from five air-quality monitoring sites in Lanzhou from October 2014 to October 2015, revealing the spatial-temporal behavior of local particulate matter (PM). The Hybrid Single Particle Lagrangian Integrated Trajectory Model (HYSPLIT) and the PM2.5-to-PM10 ratio model were used to investigate the primary transport path, potential source areas and contributions of the East Asian sandstorm to PM in Lanzhou. The analysis in three functional areas of the city indicated that the monthly variation in PM2.5 displayed a unimodal U pattern (the highest value was during the heating period), whereas that of PM10 displayed a bimodal pattern (the primary peak appeared in the spring, and the secondary peak appeared in the winter). These two patterns originated from different PM sources. The PM2.5 was primarily affected by human activities, and the PM10 was influenced by both natural and anthropogenic activities, but the relative contributions of these activities were associated with spatial-temporal variations. The daily PM10 and PM2.5 concentration variations displayed a bimodal pattern in the three functional areas: the peak values appeared at 11:00-13:00 and 22:00-1:00, respectively, and the lowest values appeared at 4:00-6:00 and 16:00-18:00, respectively. On the monthly, seasonal and daily scales, the PM concentrations exhibited similar patterns in the industrial, urban and rural areas, indicating that they were partly controlled by the regional natural environment. Meanwhile, due to anthropogenic factors, considerable PM amounts were discharged into the external environment, leading to maximum and minimum concentrations of PM appearing in the industrial and rural areas, respectively. The HYSPLIT model showed that dust storms from the northwest desert and Gobi regions affected Lanzhou three times in March 2015 and contributed 68% and 40% of the total mass of PM10 and PM2.5, respectively.