Modeling and Making Fairer Comparison of PM2.5 in Beijing

• Activity Number: 538
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Xuan Liang* and Song Xi Chen
• Companies: Peking University and Peking University/Iowa State University
• Keywords: PM2.5 ; Meteorological Condition ; Adjustment ; Environmental Assessment

Abstract

Severe air pollution is encountered in a substantial part of China, especially in the North China Plain. Since PM2.5 concentration is easily influenced by meteorological conditions, directly comparing raw monthly or yearly averages are subject to bias due to different meteorological conditions. Here we use partial linear and nonparametric approach to model the relation between PM2.5 and the meteorological variables, then adjust PM2.5 concentration to have fairer comparison by considering the weather conditions based on the nonparametric and semiparametric models. If we assume the effects of emission and weather on PM2.5 are additive and independent, we show our adjusted estimator of the monthly average is consistent and can be decomposed into two parts: one is due to the weather and the other due to the emission. We can apply the same framework to the percentiles. We use five years' PM2.5 readings at the US Embassy and corresponding meteorological data in Beijing for the case study. The results indicate our adjustment is necessary and can be applied in the environmental assessment.