PM2.5has become an increasing public concern recently because of its visibility reduction and severe health risks.For the whole year of 2013,hourly PM2.5data of 496 monitoring sites scattered in 74 cities of China are collected to analyze temporal and spatial variability of PM2.5concentration.Different temporal scales(seasonal variation,monthly variation and daily variation)and spatial scales(urban versus rural,typical areas and national scale)are discussed.Results show that PM2.5concentration changes significantly in both long-term and short-term scales.An apparent bimodal pattern exists in daily variation of PM2.5concentration and the daytime peak appears around 10:00 am while the lowest concentration appears around 16:00 pm.Spatial autocorrelation analysis and Ordinary Kriging are used to characterize spatial variability.Moran’s I of PM2.5concentration in three typical regions,the Beijing-Tianjin-Hebei region,the Yangtze River Delta region and the Pearl River Delta region,is 0.906,0.693,0.746,respectively,which indicates that PM2.5is strong spatial correlated.Spatial distribution of annual PM2.5concentration simulated by Ordinary Kriging shows that 7.94 million km2(83%)areas fail in meeting the requirement of China’s National Ambient Air Quality Standards Level-2(35?g/m3)and there are at least three concentrated highly polluted areas across the country. PM2.5has become an increasing public concern recently because of its visibility reduction and severe health risks. For the whole year of 2013, hourly PM2.5data of 496 monitoring sites scattered in 74 cities of China are collected to analyze temporal and spatial variability of PM2 .5concentration.Different temporal scales (seasonal variation, monthly variation and daily variation) and spatial scales (urban versus rural, typical areas and national scale) are discussed. Results show that PM2.5concentration changes significantly in both long-term and short-term scales. An apparent bimodal pattern exists in daily variation of PM2.5concentration and the daytime peak appears around 10:00 am while thehypocus point appears around 16:00 pm.Spatial autocorrelation analysis and Ordinary Kriging are used to characterize spatial variability. Moran’s I of PM2.5concentration in three typical regions, the Beijing-Tianjin-Hebei region, the Yangtze River Delta region and the Pearl River Delta region, is 0.906 , 0.693,0.746, respectively, which indicates that PM2.5is strong spatial correlated. Spatial distribution of annual PM2.5concentration simulated by Ordinary Kriging shows that 7.94 km km2 (83%) areas fail in meeting the requirement of China’s National Ambient Air Quality Standards Level-2 (35? G / m3) and there are at least three concentrated highly polluted areas across the country.