Characterization and source apportionment of volatile organic compounds based on 1-year of observational data in Tianjin, China

From November 2014 to October 2015, the concentrations of volatile organic compounds (VOCs), O3 and NOx were simultaneously monitored by using online instruments at the air monitoring station belonged to Tianjin Environmental Protection Bureau (TEPB). The results indicated that VOCs concentrations were higher in autumn and lower in spring, while O3 concentrations were higher in summer, and lower in winter. The diurnal variations of VOCs and NOx (NO2 plus NO) showed opposite tendency comparing to that of O3. The concentrations of alkanes were higher (the average of 18.2 ppbv) than that of aromatics (5.3 ppbv) and alkenes (5.2 ppbv), however, the alkenes and aromatics made larger contributions to ozone because of their high reactivity. Tianjin belonged to the VOC-limited region during most of seasons (except summer) according to the VOC/NOx ratios (the 8:1 threshold). The automobile exhaust, industrial emission, liquefied petroleum gas/natural gas (LPG/NG), combustion, gasoline evaporation, internal combustion engine emission and solvent usage were identified as major sources of VOCs by Positive Matrix Factorization (PMF) model in Tianjin, and the contributions to VOCs for the entire year were 23.1%, 19.9%, 18.6%, 10.6%, 8.7%, 5.4% and 4.7%, respectively. The conditional probability function (CPF) analysis indicated that the contributing directions of automobile exhaust and industrial emission were mainly affected by source distributions, and that of other sources might be mainly affected by wind direction. The backward trajectory analysis indicated that the trajectory of air mass originated from Mongolia, which reflected the features of large-scale and long-distance air transport, and that of beginning in Jiangsu, Shandong and Tianjin, which showed the features of small-scale and short-distance. Tianjin, Beijing, Hebei and Northwest of Shandong were identified as major potential source-areas of VOCs by using potential source contribution function (PSCF) and concentration-weighted trajectory (CWT) models.