Using big data from air quality monitors to evaluate indoor PM_2.5 exposure in buildings: Case study in Beijing

Due to time- and expense- consuming of conventional indoor PM_2.5 (particulate matter with aerodynamic diameter of less than 2.5 μm) sampling, the sample size in previous studies was generally small, which leaded to high heterogeneity in indoor PM_2.5 exposure assessment. Based on 4403 indoor air monitors in Beijing, this study evaluated indoor PM_2.5 exposure from 15th March 2016 to 14th March 2017. Indoor PM_2.5 concentration in Beijing was estimated to be 38.6 ± 18.4 μg/m³. Specifically, the concentration in non-heating season was 34.9 ± 15.8 μg/m³, which was 24% lower than that in heating season (46.1 ± 21.2 μg/m³). A significant correlation between indoor and ambient PM_2.5 (p < 0.05) was evident with an infiltration factor of 0.21, and the ambient PM_2.5 contributed approximately 52% and 42% to indoor PM_2.5 for non-heating and heating seasons, respectively. Meanwhile, the mean indoor/outdoor (I/O) ratio was estimated to be 0.73 ± 0.54. Finally, the adjusted PM_2.5 exposure level integrating the indoor and outdoor impact was calculated to be 46.8 ± 27.4 μg/m³, which was approximately 42% lower than estimation only relied on ambient PM_2.5 concentration. This study is the first attempt to employ big data from commercial air monitors to evaluate indoor PM_2.5 exposure and risk in Beijing, which may be instrumental to indoor PM_2.5 pollution control.