Refined Assessment of Human PM2.5 Exposure in Chinese city by Incorporating Time-activity Data

Abstract:

Since urban residents tend to spend a majority of time indoors throughout a day, it has been widely discussed in recent years, whether fixed-site monitoring PM_2.5 ambient concentration is feasible as a surrogate of human PM_2.5 exposure. Comprehensive understanding of residents’ daily time-activity patterns (TAP) and possible indoor behavior are urgently needed to perform a more accurate estimate of human PM_2.5exposure, especially in China, where is experiencing rapid urbanization.

Field surveys of TAP were carried out in a Chinese city of Suzhou from 2014 to 2015 to evaluate PM_2.5 exposure in various micro-environments (ME, e.g., residence, outdoors and in-transit). We gathered and analyzed urban residents’ seasonal time-activity data using 24h retrospective time-location diaries, as well as diversified exposure-related indoor information (e.g. ventilation, environment tobacco smoke and cooking). PM_2.5exposure is calculated through the incorporation of ambient concentration data, modified indoor/outdoor empirical functions and TAP. The spatial distributions of TAP-based exposure and static-population based exposure are also compared.

Residents in Suzhou urban area spend over 65% of time at home and 90% indoors. There are significant temporal (season, day type) and socioeconomic differences (gender, age, education, living alone, having children at home, employment status, etc.) of time-activity distributions, which makes the sum of PM_2.5 ME exposure differs notably from static-population based ambient exposure. People prefer to spend more time at home both in winter (P<0.05) and on weekends (P<0.001), less time outdoors in winter but more on weekends (P<0.001). Gender, education and living alone are negative associated with time spent home, while age, children at home and employment status are positively related. On the other hand, due to lack of monitoring stations in unban Suzhou, the inverse distance squared weighting method is not ideally performed and may be less representative of the ambient PM_2.5characteristics than satellite data.