TRI Field Notes: Monitoring Urban Air Quality in Lanzhou, China
By Lan Jin, 2015 TRI Fellow in China
Air pollution levels are alarmingly high in China, which arouses great concerns of human health. Limited information is known about the impacts of air pollution on infant health, especially in developing countries . However, it is in these countries that the people have the heaviest burden of adverse birth outcomes [2-4]. Incomplete monitoring systems hamper the understanding of air pollution and prevention of disease in these areas. To address this challenge, I am conducting a series of rigorous monitoring campaigns to characterize the within-city variability of traffic pollution using Land Use Regression models (LUR) in Lanzhou, China. An important innovation of our study is that we will add a third dimension — building height — to the traditional LUR analysis. Traditional LUR has been widely used to characterize ground levels of air pollution in many countries of Europe and North America . Unlike these countries, Chinese urban areas have dense populations living in high-rise apartment buildings. A Korean study has shown a significant decrease in levels of Volatile Organic Compounds with increasing building height . Overlooking the variation of pollution at different heights might lead to exposure misclassification and inaccurate risk assessment. This summer, I conducted a pilot monitoring campaign to test our hypothesis, and build preliminary models.
Lanzhou is an important industrial city in China. It is the largest city and the capital of Gansu Province. With rapid economic development, emissions from factories and heavy traffic contributed to the city’s poor air quality . In my last project, I found positive associations of maternal exposure to PM10 (particulate matter with diameter <10 µm) and NO2 (nitrogen dioxide) with congenital heart defects in a Lanzhou birth cohort . Directed by Prof. Yawei Zhang at Yale School of Public Health, this cohort recruited more than 10,000 women who gave birth at Gansu Provincial Maternity and Child Care Hospital from 2010-2012. These women provided detailed information on demographics, pregnancy outcomes, medical history, and lifestyle choices. However, in the last project, we used only four regulatory monitors to estimate maternal exposures, which may not adequately capture the spatial variation in pollution. In this study, I am focusing on traffic pollution and its impact on birth outcomes in attempt to provide more support for future pollution reduction.
Over the summer of 2015, two monitoring campaigns were conducted from May 29th to July 9th. In one campaign, 135 Palmes tubes were deployed at 105 sites to measure NO2 (an indicator of traffic pollution) during a two-week period. I selected three kinds of sites: ground sites (41 sites), road sites (44 sites) and building sites (20 sites). Ground sites were selected using stratified random sampling followed by purposeful selection. In stratified random sampling, 24 sites were selected to capture the range of NO2 levels hypothesized to be influenced by the following predictors: road density, population density, land use types, distance to river, distance to green space and elevation. In purposeful selection, the distribution of the above 24 sites was evaluated and geographic gaps were identified to add 13 sites to complete the coverage. Four additional sites were selected to be collocated with regulatory monitors. For road sites, 44 sites were selected at four types of roads—primary, secondary, tertiary, and residential roads as classified by OpenStreetMap. One road perpendicular to the dominant wind direction was selected from each category of roads. For primary and secondary roads that have more traffic, a vector of 10 sites were placed at each side of the roads (upwind and downwind direction). Each vector of 10 sites was perpendicular to the selected roads. The 10 sites were selected at about 0, 20, 40, 60, 80, 100, 120, 140, 160, and 200m from roads. For tertiary and residential roads, one site was selected at each side of the roads. For the building sites, two tall buildings with 30 and 31 floors were selected.
According to local real estate agencies, many apartment building complexes have about 30 floors, and more communities like this are under construction. One of the selected buildings has hallway windows facing a busy road, and the other has windows facing the green space in the residential community. Each building has 10 sites arranged vertically at increasing floors. For the building with 31 floors, the floor 1, 4, 7, 10, 13, 16, 19, 23, 27, and 31 were selected. For the building with 30 floors, the floor 1, 4, 7, 10, 13, 16, 19, 22, 26, and 30 were selected. Then, we went into the field to find the exact locations nearest to the selected sites to install the tubes. Another unique feature of this city is that almost every street, big or small, is full of small shops, such as convenience or hardware stores. We installed most of the tubes at the front of these stores. These stores saved us time because we could easily change from one to another, if a store owner did not agree. In addition, the store owners were usually willing to help look after these tubes. We installed tubes at least 2m high on the façade of the buildings using zip ties or scotch tape. For building sites, we installed the tubes outside the hallway windows. After two weeks, we went to retrieve the tubes, and only two were missing. We also installed four video cameras at each of the four roads to record traffic volume from 9am to 9pm. The traffic volume will be counted later.
In the second campaign, mobile monitors were used to measure NO2 and BC (black carbon, indicator of diesel combustion) at the two buildings both inside and outside the windows during rush and non-rush hours for 6 days continuously. This campaign provides insights for the validity of using outdoor measurements to estimate indoor exposure. In addition, it has higher temporal resolution as we can distinguish the difference in pollution levels between rush and non-rush hours. I have two sets of monitors. One set was placed on a fixed floor as an “anchor,” and the other set was moved from one floor to another which are collocated with the monitoring tubes. At each floor, measurements were conducted for 5 minutes.
The next step is to analyze the concentrations of pollution from both the tubes and mobile monitors. The distribution of the NO2 concentrations will be analyzed to inform future monitoring campaigns in the following seasons. Potential spatial predictors of pollution at each site will be calculated using ArcGIS. These predictors along with the NO2 concentrations from the tubes will be used to build LUR models. Eventually, the LUR model will be used to estimate maternal exposure to traffic pollution and its health implications.
Understanding maternal exposure to air pollution, especially traffic-related pollution, may contribute to future prevention of adverse birth outcomes at both individual and community levels. I hope my work can help governments or the general public make more informative decisions in controlling traffic pollution in residential areas. For example, with knowledge of the distribution of traffic pollution and its related health impact, protective measures could be suggested for pregnant women living in certain areas with high traffic density or on certain floors of apartment buildings.
- Polichetti, G., et al., Effects of Ambient Air Pollution on Birth Outcomes: An Overview. Critical Reviews in Environmental Science and Technology, 2013. 43(12): p. 1223-1245.
- World Health Organization and UNICEF., Low birthweight : country, regional and global estimates. 2004, Geneva: World Health Organization. 31 p.
- van der Linde, D., et al., Birth prevalence of congenital heart disease worldwide: a systematic review and meta-analysis. J Am Coll Cardiol, 2011. 58(21): p. 2241-7.
- Beck, S., et al., The worldwide incidence of preterm birth: a systematic review of maternal mortality and morbidity. Bull World Health Organ, 2010. 88(1): p. 31-8.
- Hoek, G., et al., A review of land-use regression models to assess spatial variation of outdoor air pollution. Atmospheric Environment, 2008. 42(33): p. 7561-7578.
- Jo, W.K. and K.Y. Kim, Vertical variability of volatile organic compound (VOC) levels in ambient air of high-rise apartment buildings with and without occurrence of surface inversion. Atmospheric Environment, 2002. 36(36-37): p. 5645-5652.
- Zhang, Y.Q., et al., Air Quality in Lanzhou, a Major Industrial City in China: Characteristics of Air Pollution and Review of Existing Evidence from Air Pollution and Health Studies. Water Air and Soil Pollution, 2014. 225(11).
- Jin, L., et al., Ambient air pollution and congenital heart defects in Lanzhou, China. Environmental Research Letters, 2015. 10(7): p. 074005.
“TRI Field Notes” share the stories of TRI Fellows as they conduct independent summer research throughout the tropics. The Tropical Resources Institute (TRI) is a center at the Yale School of Forestry and Environmental Studies. For more information on research and fellowships visit http://environment.yale.edu/tri/