Impact of synoptic weather patterns on 24 h-average PM2.5 concentrations in the North China Plain during 2013-2017

North China Plain area (NCP) is one of the most densely populated and heavily polluted regions in the world. In the last five years, frequently happened fine particulate matter (PM2.5) serious pollution events were one of the top environmental concerns in China. As PM2.5 concentrations are highly influenced by synoptic flow patterns and local meteorological conditions, a two-stage hierarchical clustering method based on dynamic principal component analysis (DPCA) and standard k-means clustering algorithm was employed to classify synoptic wind fields into 6 patterns over the NCP area using the data of 5 PM2.5 seasons (Sept. 15th-Apr. 15th) from 2013 to 2017. Among the six identified synoptic patterns, pattern of uniform pressure field (U) and that of zonal high pressure (Z(H)) accounted for 78.21%, 65.55%, 63.56%, 57.11%, 59.13% and 58.27% studied heavy smog pollution events in Beijing, Tianjin, Tangshan, Baoding, Shijiazhuang and Xingtai city. The two particular patterns were associated with uniform pressure field and sparsely latitudinal isobar in 850 hPa level, respectively. They were also characterized by high relative humidity, low temperature, low-speed northerly wind in Tianjin and Tangshan, and southerly wind in the other cities. Under the continuous control of pattern Z(H), the values of 24h-average PM2.5 were found to increase at a rate of 31.78 mu g/m(3) per day. To evaluate the contribution of meteorological factors and precursors to PM2.5 levels, linear mixed-effects models (LMMs) were applied to establish relations among 24 h-average PM2.5 concentrations, concentrations of main precursors, local meteorological factors and synoptic patterns. Results show that the variations of precursors, local meteorological factors and synoptic flow patterns can explain 51.67%, 19.15% and 14.01% changes of the 24 h-average PM2.5 concentrations, respectively. This study illustrates that dense precursor emissions are still the main cause for heavy haze pollution events, although meteorological conditions play almost equal roles sometimes. (C) 2018 Elsevier B.V. All rights reserved.