[Uncertainty analysis of emission inventory for volatile organic compounds from anthropogenic sources in China].

This study analyzed the uncertainties in the Chinese anthropogenic volatile organic compound (VOC) emission inventory. The uncertainty evaluation system for input information including activity data and emission factors was established to provide the probability density function of each input datum. Monte Carlo method was applied to propagate the uncertainties of emissions' input data and calculate the probability density function of total VOC emissions. The results indicated that the Chinese national anthropogenic VOC emissions in year 2005 had a log-normal distribution, with a relative standard deviation of 52%, and with 95% confidence interval of [- 51%, + 133%]. However, if the traditional method for error analysis was used based on the same input information, the uncertainty of the 2005 VOC emission inventory calculated was 40% lower than the former results. In addition, sensitivity analysis was conducted to identify the 20 most sensitive inputs influencing the uncertainty of emissions, which will be helpful to improve the accuracy of VOC emission inventory in the future.