/* */

Issue Profile

Air Pollution

Air Pollution has become a hallmark of urban life, with more than 90 percent of the global population breathing unsafe air.1World Health Organization. (2016). WHO Global Urban Ambient Air Pollution Database (update 2016). Retrieved from: http://www.who.int/phe/health_topics/outdoorair/databases/cities/en/ (accessed July 2018).

 

Chapter Authors

Angel Hsu
Ryan Thomas
Saskia Comess

This issue category includes three indicators: Average Exposure to PM2.5 (fine particulate matter in micrograms per cubic meter (µg/m3); Average Exposure to Nitrogen Dioxide (ppm); and PM2.5 Exceedance (average percentage of the population exposed to PM2.5 levels at 10 µg/m3, 15 µg/m3, 25 µg/m3, and 35 µg/m3, which represent the World Health Organization’s (WHO) interim I, II, and III targets, respectively according to its air quality guidelines.2World Health Organization. (2005). WHO Air quality guidelines for particulate matter, ozone, nitrogen dioxide and sulfur dioxide. Retrieved from: http://apps.who.int/iris/bitstream/handle/10665/69477/WHO_SDE_PHE_OEH_06.02_eng.pdf?sequence=1.
Air pollution is a growing, worldwide problem that afflicts urban areas in both developed and developing countries. Although rural air pollution is a concern, particularly in households that still combust solid fuels indoors for cooking and heating, nearly every urban area in the world experiences unsafe air quality, whether episodically or on a continuous basis. Suspended air particles contribute to cardiovascular disease, stroke, acute lower respiratory infections and other diseases such as cancer. They can penetrate human lung and blood tissue, leading to higher incidences of cardiovascular and lung disease. Fine particulates, or PM2.5 (2.5 microns and smaller), lodge deep in blood and lung tissue and are particularly injurious to human health.

 

Nitrogen dioxide (NO2) is also harmful to humans both directly and when it reacts with other compounds like sunlight and volatile organic compounds (VOCs) to produce ozone or secondary particulate pollution.3World Health Organization. (2003). Health aspects of air pollution with particulate matter, ozone and nitrogen dioxide: report on a WHO working group, Bonn, Germany 13-15 January 2003. Available: http://www.euro.who.int/__data/assets/pdf_file/0005/112199/E79097.pdf. Last accessed: November 30, 2015. Strong associations between NO2 and mortality have been identified in multi-city studies around the world.4Geddes J.A., Martin R.V., Boys B.L., and van Donkelaar A. (2015). Long-Term Trends Worldwide in Ambient NO2 Concentrations Inferred from Satellite Observations. Environmental Health Perspectives. Available: http://dx.doi.org/10.1289/ehp.1409567. Last accessed: November 30, 2015. According to the U.S. Environmental Protection Agency, direct exposure to NO2, ranging from 30 minutes to 24 hours, can cause airway inflammation and negative respiratory effects for people with asthma.5US EPA. (2017). Health effects of Nitrogen Dioxide. Available: https://www.epa.gov/no2-pollution/basic-information-about-no2#Effects. Last accessed: December 28, 2017. Direct inhalation of both ozone and NO2 can aggravate the human respiratory system, especially in people who have respiratory illnesses such as asthma. Prolonged exposure to elevated concentrations of NO2 can also lead to asthma development and leave people more susceptible to respiratory infections.6U.S. Environmental Protection Agency (USEPA). (2017). Basic Information about NO2. Available: https://www.epa.gov/no2-pollution/basic-information-about-no2#What is NO2. Last accessed: November 30, 2017. NO2 can also serve as a robust indicator of many traffic- and combustion-related pollutants that are not always monitored routinely. 7Levy I, Mihele C, Lu G, Narayan J, Brook JR. 2014. Evaluating multipollutant exposure and urban air quality: pollutant interrelationships, neighborhood variability, and nitrogen dioxide as a proxy pollutant. Environ Health Perspect 122:65–72; doi:10.1289/ehp.1306518). 

Dalhousie University’s Atmospheric Composition Analysis Group provides the satellite-derived PM2.5 and NO2 data. Population data to measure the proportion of the population above various PM2.5 concentration thresholds are obtained from the Global Rural Urban Mapping Project, v.48 Center for International Earth Science Information Network – CIESIN – Columbia University. 2016. Gridded Population of the World, Version 4 (GPWv4): Administrative Unit Center Points with Population Estimates. Palisades, NY: NASA Socioeconomic Data and Applications Center (SEDAC).http://dx.doi.org/10.7927/H4F47M2C. at the NASA Socioeconomic Data and Applications Center, hosted by the Center for International Earth Science Information Network (CIESIN) at Columbia University.

For more information, see the UESI Metadata.

10 µg/m3 for Average Exposure to PM2.5 (fine particulate matter); 0 for PM2.5 Exceedance; 1st percentile for Average Concentration of NO2.

Description

Air pollution is the leading environmental risk factor for death worldwide and the fifth leading cause of death overall. 9Health Effects Institute. (2017). State of Global Air 2017. Special Report. Boston, MA:Health Effects Institute.It claims around 4.2 million lives a year, and more than 92 percent of the global population breathes unsafe air.10Health Effects Institute. (2017). State of Global Air 2017. Special Report. Boston, MA:Health Effects Institute.The primary culprit for air-pollution-related deaths is fine particulate pollution, made up of particles smaller than 2.5 microns in diameter (PM2.5),which are fine enough to lodge deep into human lung and blood tissue. A complex mixture of toxic particles, particulate matter places exposed populations at risk of cardiovascular and lung disease, ranging from stroke to chronic obstructive pulmonary disease, asthma, and lung cancer.11U.S. Environmental Protection Agency (USEPA). (2009). Integrated Science Assessment for Particulate Matter. National Center for Environmental Assessment, Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC. Elderly populations and young children are particularly vulnerable to health effects of PM2.5. The leading cause of mortality for children between the ages of one to five is pneumonia, and half of these cases are due to air pollution.12 UNICEF. (2016). Clear the air for children: The impact of air pollution on children. Available: https://www.unicef.org/publications/files/UNICEF_Clear_the_Air_for_Children_30_Oct_2016.pdf

Airborne particulates originate from a variety of natural and anthropogenic sources. PM2.5 is primarily the product of combustion, whether from human activity, such as burning coal or car emissions, or through forest fires and volcanoes. In parts of Asia, where coal combustion is the primary source of electricity generation, PM2.5 pollution has led some cities, like Beijing, to eliminate coal-fired power plants as part of their air pollution control plans.13South China Morning Post. (2017). Beijing shuts down its last coal-fired power plant as part of bid to clear air. Available: http://www.scmp.com/news/china/society/article/2080270/beijing-shuts-down-its-last-coal-fired-power-plant-part-bid-clear. In other parts of the world, wildfires, such as those that have plagued Western U.S. states like California, are a major source of PM.14Park, R. J., 2003: Sources of carbonaceous aerosols over the United States and implications for natural visibility. Journal of Geophysical Research, 108, D12, 4355. doi:10.1029/2002JD003190 PM2.5 can also be generated through secondary formation, when other precursor gases, such as sulfates from power plants and industrial facilities and nitrates from mobile sources and power plants, react in the atmosphere. During winter months, secondary PM formation can be particularly acute due to temperature inversions that trap warm air – and pollutants – below a layer of cold air. Temperature inversions and the secondary formation of PM2.5 during cold winter months are the primary cause of high levels of air pollution observed in cities like New Delhi. India’s capital city experienced air pollution levels 30 times higher than WHO recommended levels in November 2017. Its Chief Minister equated the city to a gas chamber.15Newman, S. (2017). New Delhi Chief Minister Calls India’s Smog-Choked Capital A ‘Gas Chamber’. NPR. Available: https://www.npr.org/sections/thetwo-way/2017/11/09/563008286/new-delhi-chief-minister-calls-indias-smog-choked-capital-a-gas-chamber.

 

 

Nitrogen dioxide (NO2) is derived from combustion processes similar to particulate matter formation. It forms from road traffic and power plants, and is a precursor to particulate matter and ozone, which also have significant human health effects. Ground-level ozone at high concentrations, particularly during summer months, is a major component of urban smog. It can have acute respiratory health effects and has been responsible for a million premature deaths each year.16Anenberg, S. C., and others, 2009: Intercontinental impacts of ozone pollution on human mortality. Environmental Science & Technology, 43, 6482-6287. doi:10.1021/es900518z; Malley, C. S., Henze, D. K., Kuylenstierna, J. C., Vallack, H. W., Davila, Y., Anenberg, S. C., … & Ashmore, M. R. (2017). Updated global estimates of respiratory mortality in adults>/= 30Years of age attributable to long-term ozone exposure. Environmental Health Perspectives, 2017, vol. 125, num. 8, p. 087021.

NO2 pollution has increased in many European countries where diesel fuel is subsidized, in some cases by as much as 15 percent more than other less-polluting fuels.17Willsher, K. (2015). “Paris chokes on pollution: City of Light becomes City of Haze.” Los Angeles Times. Available: http://www.latimes.com/world/europe/la-fg-france-paris-smog-20150323-story.html. Last accessed: December 22, 2015. Diesel cars are also not subject to the same strict emissions testing as other vehicles, such as heavy trucks and buses, meaning that modern diesel cars produce 10 times more air pollution than other kinds of vehicles.18 Carrington, D. (6 January 2017). “Diesel cars emit 10 times more toxic pollution than trucks and buses, data shows.” The Guardinan. Available: https://www.theguardian.com/environment/2017/jan/06/diesel-cars-are-10-times-more-toxic-than-trucks-and-buses-data-shows. This problem came under international scrutiny when “Dieselgate” revealed that Volkswagen had been circumventing NOx emissions controls on its vehicles, resulting in higher emissions than lab tests suggested.19Hotten, R. (2015). Volkswagen: The scandal explained. BBC. Available: http://www.bbc.com/news/business-34324772. A European Environment Agency report found that Italy had the highest number of NO2-related deaths in 2013, at 20,000, and London hosted the continent’s worst NO2 hotspot, at double the allowable EU limit.20 European Environment Agency. (2017). Air Quality in Europe 2017. Available: https://www.eea.europa.eu/publications/air-quality-in-europe-2017. Last accessed: December 28, 2017.

Box 1. Communicating Air Quality

As a way of communicating the risks associated with various levels of air pollution, governments around the world have developed indices that translate different exposures to easily understandable health risks. Many use the Air Quality Index (AQI), which normalizes air pollution concentrations to a scale from 0 to 500, with 0 representing good air quality that poses little to no health risk, and a score of 500 indicating hazardous air pollution that likely affects the entire population (see Figure 1). Within the United States, the AQI value for each major criteria air pollutant is determined from the station within a monitoring area that registers the highest concentration of that pollutant, among all monitors.21 US EPA. (2016). Technical Assistance Document for the Reporting of Daily Air Quality – the Air Quality Index (AQI). Available: https://www3.epa.gov/airnow/aqi-technical-assistance-document-may2016.pdf. The AQI communicated to the public is then the highest recorded AQI value among all of the pollutants. A color code, ranging from green (safe) to maroon (hazardous) communicates air quality and possibly health risks to the public (see Figure 1).

Governments utilize a diverse array of methods for communicating the AQI. The European Environment Agency began providing short-term (from 6 to 48 hours) AQI information from 2,000 monitoring stations across Europe in November 2017 (see Figure 2). Other governments use social media and other forms of online media to share AQI data. Shanghai’s Environmental Protection Bureau, for example, has utilized Sina Weibo, a microblogging platform similar to Twitter, to share air quality information through a character whose hair color reflects the current AQI value.22 Hsu A. (2013). Shanghai’s New Air Quality Mascot. Data-Driven Yale. Available: https://datadrivenlab.org/air-quality-2/shanghais-new-air-quality-mascot/. Some governments even issue warnings when AQI levels are high, shutting down schools and offices and cautioning people to stay indoors.23 Hatton, C. (2014). Beijing smog alert turns ‘orange’ as air pollution soars. BBC. Available: http://www.bbc.com/news/blogs-china-blog-26272905. Due to the high public health risks associated with air pollution, countries are experimenting with different forms of communication to inform residents when pollution levels may pose danger.

Figure 1. The Air Quality Index (AQI) utilized in many countries around the world, including the United States, ranges from 0 to 500 and communicates associated health risks. Image source: Santa Barbara County. Source: https://www.ourair.org/sbc/the-air-quality-index/?doing_wp_cron=1514486066.1177558898925781250000.

Figure 2. The European Environment Agency and the European Commission launched a new Air Quality Index communication platform in November 2017. This online dashboard reflects short-term air quality from more than 2,000 monitoring stations across Europe. Source: http://airindex.eea.europa.eu

Urbanization and Air Quality

Now that more than half of the global population lives in cities, air pollution has become a hallmark of urban life. Industrialization has concentrated economic activity in cities, which are responsible for some 80 percent of global gross domestic product (GDP).24 World Bank. (2017). Urban Development. Available: http://www.worldbank.org/en/topic/urbandevelopment/overview. This aggregation has increased population density in urban areas relative to rural surroundings, creating larger  demands on energy and natural resources while generating pollution and waste. The growth of personal car usage, building stock, and energy demand has led to severe air pollution crises in many cities around the world. London, for instance, was reported to have air pollution levels worse than Beijing in January 2017, reaching 197 on the Air Quality Index (AQI) (see Box 1, Communicating Air Quality).25Knapton, S. (2017). Air pollution in London passes levels in Beijing… and wood burners are making problem worse . The Telegraph. Available: http://www.telegraph.co.uk/science/2017/01/24/air-pollution-london-passes-levels-beijingand-wood-burners-making. Paris became so polluted in 2015 that the city government enacted emergency measures, such as restricting vehicles and subsidizing public transportation use.26Willsher, K. (2015). Paris chokes on pollution; City of Light becomes City of Haze. The Los Angeles Times, March 23. Available: http://www.latimes.com/world/europe/la-fg-france-paris-smog-20150323-story.html. Los Angeles has notoriously struggled with air pollution for decades, with smog levels persisting despite decreasing emissions in recent years.27 Barboza, T. (2017). Southern California smog worsens for second straight year despite reduced emissions. The Los Angeles Times. Available: http://www.latimes.com/local/lanow/la-me-ln-bad-air-days-20171115-story.html. Globally, about 95 percent of the world’s population breaths outdoor particulate matter concentrations in excess of the WHO’s Air Quality guideline (annual mean concentration less than 10 µg/m3). Nearly 60% of people live in areas where fine particulate matter exceeds the acute WHO air quality target (35 µg/m3).28 State of Global Air 2018. Available: https://www.stateofglobalair.org/air. The WHO notes that there is no threshold concentration for particulate matter below which no damage to health is observed, indicating that the vast majority of the global population is being exposed to air pollution with deleterious health effects.29 World Health Organization (2018). “Ambient (outdoor) air quality and health.” Available: http://www.who.int/en/news-room/fact-sheets/detail/ambient-(outdoor)-air-quality-and-health.

With worsening air pollution in urban areas, city managers have implemented a range of policies to address it. These efforts include tackling emissions sources, such as coal-fired power plants, to encouraging efficiency and cleaner fuel standards for motor vehicles. Since transportation is the number one contributor to air pollution in many urban areas around the world, many cities have made  reducing transit-related emissions a top priority. For nearly two decades, New Delhi, for instance, has mandated that all public transportation use compressed natural gas (CNG), which is lower in emissions than diesel, to lower the transport-related emissions that drive its pollution. Now that the number of private vehicles have far outpaced other transportation modes in the capital city, the government is looking to expand public transit options to address its growing air pollution problem.30 Jain, M. (2016). Delhi has relied on CNG to control its pollution in the past, but will it work this time?. Scroll.in. Available: https://scroll.in/article/807463/delhi-has-relied-on-cng-to-control-its-pollution-in-the-past-but-will-it-work-this-time. Other cities, like Singapore, are planning to restrict the total number of vehicles altogether, by outlawing future increases in private vehicle ownership.31 AFP. (2017). Singapore: no more cars allowed on the road, government says. The Guardian, Oct. 24. Available: https://www.theguardian.com/world/2017/oct/24/singapore-no-more-cars-allowed-on-the-road-government-says.

Globally, growing policy attention has focused on air quality as an urban issue, with air pollution inserted as a target in Sustainable Development Goal (SDG) 11 for cities. Goal 11, to “Make cities inclusive, safe, resilient and sustainable,” sets a target to “reduce the adverse per capita environmental impact of cities” with particular attention to air quality.”32United Nations, nd. Sustainable Development Goal 11. Available: http://www.un.org/sustainabledevelopment/cities/. Air is also included in the opening text of the Sustainable Development Goals (SDGs), cementing the issue as central to both sustainable development and human health.

Towards Improved Monitoring

Despite its known health impacts, global monitoring of air pollution is lagging, usually because of lack of capacity, resources, technology, or public demand. Monitoring gaps primarily occur in developing countries outside of North America and Western Europe, where air pollution is more severe and the number of air-pollution related deaths has increased dramatically over the last 15 years (see Figure 3).33 Engel-Cox, J., Kim Oanh, N. T., and van Donkelaar, A., et al. (2013). Toward the next generation of air quality monitoring: Particulate Matter. Atmospheric Environment, 80:584-590. Given the sparseness of ground-based monitors, satellite-derived estimates have been utilized for global comparability and applied in epidemiological studies.34 Forouzanfar, M. H., Afshin, A., Alexander, L. T., Anderson, H. R., Bhutta, Z. A., Biryukov, S., … & Cohen, A. J. (2016). Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. The Lancet, 388(10053), 1659-1724. Available: http://www.thelancet.com/gbd. Satellites develop “wall to wall” measures of aerosols and pollutants in the earth’s atmosphere, enabling consistent side-by-side comparisons between entities. While a potent proxy, these satellite data fail to measure ambient air conditions directly at the ground where people breathe. They can also miss short-term spikes in air pollution that can occur episodically, particularly if the sensor fails to pass through an area during these acute events. Satellite estimates are also averaged over long periods of time, tending to smooth out and present lower pollution concentration values than a city might experience on an hourly basis.

A growing number of bottom-up, citizen-driven and private efforts show promise to improve the landscape of global air quality monitoring  (see Box 2, Air quality data getting bigger). Improvements in technology, including low-cost air sensors, are critical in helping to fill air quality gaps and allow for the real time monitoring of health risks. These new data are helping citizens and governments understand new sources of air pollution, identify personal health exposure risks, and forecast future air pollution events. Combined with other sources of near real-time data, citizen-generated information is providing new, previously unexplored insights into the economic and social costs of air pollution (see Box 3, Hacking Data for Climate Action: Data Philanthropy Provides Real-Time Data).

Figure 3. A map juxtaposes the number of ground-based PM2.5 monitoring stations with the proportion of air pollution-related deaths within each country. Purple points represent the number of monitoring stations within each country (larger sizes indicate a higher number of monitors); the point size for 500 monitors is included in the legend for reference. The percentage of air-pollution related deaths relative to all causes of death in a country are shaded from blue to red. A shortage of ground-based monitoring sites are found in countries where air-pollution related deaths are highest. Data sources: IHME, 2017 and WHO, 2016. 1

Box 2. Air quality data is getting bigger

Mobile information communication technologies (ICTs) and growing concern over global air quality has led to a bloom of big data platforms boasting real-time data, often collected from a variety of low-cost sensors and crowdsourcing techniques. A proliferation of start-ups have capitalized on this new space, developing proprietary algorithms to forecast air pollution and fill gaps in existing public monitoring systems.

These organizations often integrate data collection with air pollution forecasting. Plume Labs, a France-based start-up, has designed its own algorithms to predict air pollution using 12,000 environmental monitoring stations across 60 countries. It has also designed a mobile sensor called The Flow to track indoor and outdoor air pollution for PM2.5 , NOx, Ozone, and Volatile Organic Compounds (VOCs). Another start-up, AirVisual, provides air pollution forecasts for 6,000 cities and has developed a map to display global air pollution from 8,000 monitoring stations in real time. Like Plume, AirVisual has developed a consumer-based sensor called The Node that allows users to send the data they collect back to the company’s air pollution models in a citizen science-like feedback mechanism.

Some of these air pollution data start-ups have strategic clientele for their data products. Breezometer, for instance, is an air quality analytics provider that targets cosmetics companies, who utilize its air pollution data to develop pollution-fighting skin-care products. Other platforms draw on the power of crowdsourcing to make air pollution data more transparent, centralized and easily accessible. For example, OpenAQ is an online community of scientists, researchers, and activists that has collected more than 76 million air quality measurements from 5,830 locations in 50 countries, drawn primarily from government and research-grade resources. Many of these companies develop their own proprietary Application Programmer Interfaces (APIs) to allow clients real-time access to air pollution data.

This convergence of technological innovation presents a prime opportunity for private companies and citizens to contribute to improve global monitoring of air pollution, particularly in areas that suffer from lack of monitoring capacity.

Results

Figure 5. Comparisons of NO2 (red) and PM2.5 (blue) proximity-to-target scores across cities. A score of 100 means a city has achieved the target, while 0 represents the low performance benchmark.

Cities in the UESI show a range of results for air pollution (Figure 5). A high score indicates better performance on that air pollution metric, while a low score indicates comparatively worse performance. Delhi, which includes India’s capital city New Delhi, scores the lowest on average exposure to PM2.5 pollution with a score of 0, reflecting the highest average annual concentrations of PM2.5. The average summer-time concentration of PM2.5 in Delhi is approximately 300 µg/m3, compared to the acceptable level of 60 µg/m3 set by Delhi’s Department of the Environment. The Department attributes these pollution levels primarily to airborne road dust, soil and ash as well as combustion-related carbons from energy generation, automobiles, and the burning of municipal waste.35 Department of Environment Government of National Capital Territory of Delhi and Delhi Pollution Control Committee. (2016). “Comprehensive Study on Air Pollution and Green House Gases (GHGs) in Delhi (Final Report: Air Pollution component).” Available: http://delhi.gov.in/DoIT/Environment/PDFs/Final_Report.pdf. Climatic conditions exacerbate the air pollution problems experienced in India. In general, cities in more developing countries perform poorly when it comes to PM2.5 pollution. Los Angeles and Singapore are notable exceptions to this trend, both having scores less than 75. PM2.5 production in these cities is primarily due to combustion, such as from burning coal, car emissions, and forest fires. The National Environmental Agency of Singapore attributes its air pollution levels to the same factors that affect most major cities, including industrial and motor vehicle emissions. However, the Agency notes that “Singapore enjoys an air quality better than many cities in Asia and is comparable with the air quality of US and European cities.”36 National Environmental Agency of Singapore. “Air Quality and Targets.” Available: http://www.nea.gov.sg/anti-pollution-radiation-protection/air-pollution-control/air-quality-and-targets

In contrast to PM2.5, NO2 performance appears worse in more developed countries. Primarily developed country cities like Seoul, Tokyo, New York, Paris, London, and Amsterdam all have scores below 50, while Los Angeles and Berlin score only moderately higher. These low scores likely reflect the use of diesel fuel in motorized vehicles as described above. Although regulations against polluting vehicles have strengthened in many of these cities over the past decades, pollution levels have continued to rise. Following the Volkswagen emissions scandal, studies by the German, French and British governments found that vehicle manufacturers routinely take advantage of loopholes in European Union regulations in order to produce vehicles that do not meet emissions standards.37Stack, Liam and Jack Ewing. “London Adds Charge for Older Diesel Vehicles to Fight Pollution.” The NYTimes. Available: https://www.nytimes.com/2017/10/23/business/london-diesel-congestion-charge.html These revelations have led many cities to propose new regulations regarding diesel-burning vehicles. In London, where approximately 40 percent of the city’s air pollution is due to diesel vehicles, a new regulation applies a daily monetary fine to vehicles that do not meet European Union emissions standards.38 Mayor of London. “London’s toxic air.” Available: https://www.london.gov.uk/sites/default/files/shorthand/clean_air/ The UESI results demonstrate the importance of tighter and more effective NO2 regulations.

Box 3. Hacking Data for Climate Action: Data Philanthropy Provides Real-Time Data

Greenpeace estimates an average of 25 million people are displaced each year due to natural disasters.39 Greenpeace. (2017). Climate Change, Migration, and Displacement:. The Underestimated Disaster. Berlin, Germany. Available: https://www.greenpeace.de/sites/www.greenpeace.de/files/20170524-greenpeace-studie-climate-change-migration-displacement-engl.pdf.  But what about short-term changes in behavior or movement, or micro-migrations? How does poor air quality affect individuals’ behavior? A team of Data-Driven Yale researchers sought to answer this question as part of the United Nations Data for Climate Action Hackathon, a competition that paired researchers with typically proprietary datasets from a range of private companies.40 Data-Driven Yale. (13 November 2017). Press Release: Data Driven Yale Wins Award in UN Data for Climate Action Contest. Retrieved from: https://datadrivenlab.org/big-data-2/data-driven-yale-wins-award-in-un-data-for-climate-action-contest/. In an example of “data philanthropy,” these companies, including BBVA, Orange, Planet Labs, and Earth Networks, among others, donated real-time datasets spanning traffic patterns, air quality, weather conditions and ultra high-resolution satellite imagery.

Data-Driven Yale researchers combined more than 150 million credit card transactions from BBVA, made between 2014 – 2015 in 12 Spanish provinces, along with weather data from EarthNetworks and air quality data from the European Environment Agency. They used this information to measure the economic impacts around short-term spikes in air pollution. The team found that a 10 percent increase in air pollution (fine particulate pollution or PM2.5 and ozone) could lead to economic losses of as much as 1.5 percent of Spain’s GDP. These findings suggest hazardous air pollution can wreak havoc on local economies when people opt to stay indoors, avoiding restaurants, shops and recreational spaces. In worst-case scenarios, people may permanently migrate elsewhere to escape poor environmental conditions.

This analysis, which garnered the UN Data for Climate Action Hackathon’s prize for the project that best addresses climate change and other Sustainable Development Goals, would not have been possible without data provided by private companies such as BBVA. The combination of credit card transaction data with publicly-available air pollution data allows for new insights and estimates of the societal and economic cost of air pollution. Its success illustrates the power of unconventional data sources to reveal new insights to policymakers, which would not have been visible otherwise.

Figure 4. A probability density distribution of consumer spending and air pollution shows that there are more highly-polluted days with less spending than lower-polluted days with more spending.

 

Inequity in Air Pollution Exposure

Recent policy attention has also focused on the unequal distribution of air pollution within an urban area (see Box 4: Environmental Injustice and Exposure to Air Pollution). Clark et al. (2014) found major differences in different demographic groups’ exposure to NO2 pollution within the United States. Nationally, non-white racial groups are exposed to NO2 concentrations 38 percent higher than white racial groups; in urban areas, NO2 pollution is higher for low-income groups than for high income groups.41Clark, L. P., Millet, D. B., & Marshall, J. D. (2014). National patterns in environmental injustice and inequality: outdoor NO2 air pollution in the United States. PloS one, 9(4), e94431. In some North American cities, however, opposite trends have been observed. Hajat et al. (2015) reviewed 37 studies from around the world (22 in North America, 10 in Europe, and 5 studies from Africa and the Asia Pacific) investigating air pollution exposure and socioeconomic status. This review found that larger cities, including New York, Toronto and Montreal, had opposite associations – areas with higher socioeconomic status had higher concentrations of ambient air pollution.42 Hajat, A., Hsia, C., & O’Neill, M. S. (2015). Socioeconomic disparities and air pollution exposure: a global review. Current environmental health reports, 2(4), 440-450. The authors suggested that these surprising findings could be the result of high socioeconomic status individuals clustering around busy roadways that may have better access to urban amenities.

Governments are calling attention to environmental inequities through visual mapping tools. The U.S. Environmental Protection Agency (EPA) has developed an environmental justice mapping tool (EJ Mapper)43 U.S. Environmental Protection Agency. EnviroMapper for Envirofacts. Retrieved from: https://www.epa.gov/emefdata/em4ef.home (accessed July 2018). to highlight which census blocks are more or less unequal with respect to the distribution of ozone, particulate pollution, and proximity to hazardous waste in major American cities, including Chicago, Boston and New York (see the Equity and Social Inclusion Issue Profile).44 US EPA. (2017). EJ Mapper. Available: https://www.google.com/search?q=us+epa+ej+mapper The state of California has developed a more granular tool, CalEnviroScreen,45 Office of Environmental Health Hazard Assessment (OEHHA). (2014). CalEnviroScreen, Version 2.0. Available: https://oehha.ca.gov/calenviroscreen/report/calenviroscreen-version-20 to reveal which communities are disproportionately burdened by environmental pollution, including their exposure to pollution sources such as traffic, diesel exhaust and toxic releases.

Click & Drag to Zoom In
Double-Click to Zoom Out

NO2 Equity Plot

Figure 9. NO Equity typology quadrant plot. The plot considers the Income Gini and NO Concentration Index to define four quadrants. The Income Gini Values represent the distribution of wealth across the population and range in value from 0 to 1. A Gini value of zero indicates a perfectly equal distribution of income across the population, while a high Gini value (out of a maximum of 1) suggests a highly unequal distribution of wealth. The Environmental Concentration Index (ECI) measures the variation of NO2 in response to income. Positive ECI values indicate that the environmental burden is allocated on the poorest citizens, while a negative ECI indicates that the environmental burden is allocated on the wealthier citizens. The size of the dots represents the extent of a city’s NO concentration (in ppb) (see the Equity and Social Inclusion Issue Profile for a more detailed description of this plot).

Click & Drag to Zoom In
Double-Click to Zoom Out

PM25 Equity Plot

Figure 11. PM2.5 Equity typology quadrant plot. The plot considers the Income Gini and PM2.5 Concentration Index to define four quadrants. The size of the dots represents the extent of a city’s PM2.5 concentration (in micrograms/m3) (see the Equity and Social Inclusion Issue Profile for a more detailed description of this plot).

Figure 10a.NO and income distribution curves for selected UESI cities. These plots show the concentration distributions of NO2 (e.g., the concentration curve) and income (e.g., the Lorenz curve) throughout neighborhoods in cities.

Figure 10b. PM2.5  and income distribution curves for selected UESI cities. These plots show the concentration distributions of PM2.5 (e.g., the concentration curve) and income (e.g., the Lorenz curve) throughout neighborhoods in cities.

The UESI’s pilot cities show fairly similar distributions with respect to socioeconomic status, as measured by average income and exposure to air pollution (PM2.5 and NO) (Figures 8 – 11). This result, however, is likely due to the nature of the satellite-derived air pollution data, which lacks the spatial resolution to distinguish small-scale differences between neighborhoods. In most cases, as illustrated in Figures 8 and 10, the air pollution curves appear to be more equitably distributed (i.e., closer to the 45-degree diagonal line representing perfectly equitable distribution) than income. Bangkok, Berlin, and Copenhagen are all examples where the air pollution and income distribution curves are nearly identical and on top of the line of perfect equity – suggesting that the distribution of income is likely not exacerbating air pollution exposure. In other cities, such as Johannesburg, income is much more unequally distributed than air pollution.

To better distinguish subtle differences in air pollution and income distribution in the UESI cities, the equity typology quadrant plots (Figures 9 and 11) provide a way of mathematically summarizing the relationship between air pollution and income distribution. In the top left quadrant (e.g., low Gini and negative Environmental Concentration Index or ECI), Los Angeles is located the furthest left for greatest inequality in NO exposure for the lowest income earners, while Vancouver has the most negative ECI for PM2.5 distribution. For both air pollutants, Beijing appears the right-most in the upper right-hand quadrant (low Gini, positive ECI), which suggests low income inequality has not aggravated the distribution of air pollution. Beijing, however, has among the highest in terms of absolute exposure to air pollution, suggesting that although the distribution may not be inequitable or burdening the poor more than wealthier populations, everyone regardless of income is exposed to poor air quality.

Box 4. Environmental Injustice and Expsoure to Air Pollution.

The global burden of environmental degradation is increasing, but not all people and populations share this burden equally. From air pollution to lead and polluted water, certain segments of the population are disproportionately exposed to environmental hazards. A recent report by the U.S. EPA, published in the American Journal of Public Health, quantified disparities in the location of particulate matter (PM)-emitting facilities based on the racial/ethnic composition and socio-demographic characteristics of the surrounding residential population. The study found that people living in poverty and non-Whites have higher exposure to PM, particularly PM2.5 (PM less than 2.5 micrometers in diameter). However, the most surprising finding was that disparities for Blacks are greater than the disparities observed on the basis of poverty status: considering PM2.5, those in poverty had a burden 1.35 times higher than the overall population and Blacks had a burden 1.54 times higher than the overall population (non-Whites, in general, had 1.28 times higher burden).46 Mikati, I., Benson, A. F., Luben, T. J., Sacks, J. D., & Richmond-Bryant, J. (2018). Disparities in distribution of particulate matter emission sources by race and poverty status. American journal of public health, 108(4), 480-485.

These results were consistent across national, state, and county scales. The researchers investigated whether individuals’ rural or urban status modified the relationship between race, socio-economic status and PM exposure. They found that high emissions in metropolitan (population of at least 50,000) and “micropolitan” (population between 10,000 and 50,000) cities, coupled with high representation of non-Whites in these population-dense centers, drove the national trends (Figure 6). 1 In other words, urban centers tend to have both high levels of PM2.5 pollution and higher proportions of non-White residents. Those living above the poverty line experience lower burdens than those below it within these urban areas; however, disparities in emissions are significantly larger when comparing Blacks and Whites (Figure 10). The authors conclude that this finding reinforces “the overall finding that racial disparities appear to be markedly higher than are poverty-based disparities.” 1

Figure 6. Absolute Burden of PM2.5 emissions from nearby facilities stratified by Rural-Urban Commuting Area (RUCA) code and sub-stratified by race/ethnicity and poverty (2009-2013). “High-commuting” is defined as greater than 30% flow to an urbanized area; “low-commuting” defined as 10 to 30% flow. Source: Mikati, et al., 2018.

Exposure to PM is not the only example of inequitable distribution of environmental risks. Disposal of wastewater from hydraulic-fracturing (“fracking”) occurs disproportionately in non-White and poor communities, according to a study performed in southern Texas. A small study using personal air samples collected through the National Health and Nutrition Examination Survey (NHANES) found that the levels of total volatile organic compound (VOC) exposure were 52 percent and 37 percent higher for Mexican Americans and non-Hispanic blacks, respectively, than for non-Hispanic whites, even after adjusting for socioeconomic status. Black children have the highest prevalence of elevated blood lead levels in the U.S. A recent study found that the disproportionate burden of lead exposure is transmitted from mother-to-child, such that Black children have elevated blood lead levels before birth (in utero) and into early childhood.

The life-long health effects of these exposures are well established. Exposure to PM is associated with respiratory and cardiovascular disease, premature mortality, and adverse birth outcomes.47 Mikati, I., Benson, A. F., Luben, T. J., Sacks, J. D., & Richmond-Bryant, J. (2018). Disparities in distribution of particulate matter emission sources by race and poverty status. American journal of public health, 108(4), 480-485. The International Agency for Research on Cancer (IARC) has designated PM in outdoor air pollution as carcinogenic to humans.48 Hamra, G. B., Guha, N., Cohen, A., Laden, F., Raaschou-Nielsen, O., Samet, J. M., … & Loomis, D. (2014). Outdoor particulate matter exposure and lung cancer: a systematic review and meta-analysis. Environmental health perspectives, 122(9), 906. Quantifying the potential harm from fracking has proven difficult because fracking companies classify the chemicals—and their concentrations—in fracking wastewater as classified “confidential business information.”49 Elliott, E. G., Ettinger, A. S., Leaderer, B. P., Bracken, M. B., & Deziel, N. C. (2017). A systematic evaluation of chemicals in hydraulic-fracturing fluids and wastewater for reproductive and developmental toxicity. Journal of Exposure Science and Environmental Epidemiology, 27(1), 90. Though the exact composition of fracking fluids remains unknown, over 1,000 substances have been identified in fluids and wastewater, including solvents, heavy metals, aromatic hydrocarbons, VOCs, and naturally-occurring radioactive materials.50 Elliott, E. G., Ettinger, A. S., Leaderer, B. P., Bracken, M. B., & Deziel, N. C. (2017). A systematic evaluation of chemicals in hydraulic-fracturing fluids and wastewater for reproductive and developmental toxicity. Journal of Exposure Science and Environmental Epidemiology, 27(1), 90. A systematic evaluation of 1021 chemicals in fracking fluids and wastewater found significant potential for reproductive and developmental health risks.51 Elliott, E. G., Ettinger, A. S., Leaderer, B. P., Bracken, M. B., & Deziel, N. C. (2017). A systematic evaluation of chemicals in hydraulic-fracturing fluids and wastewater for reproductive and developmental toxicity. Journal of Exposure Science and Environmental Epidemiology, 27(1), 90. Lead is a neurotoxin that affects IQ and development.52 Cassidy-Bushrow, Andrea, et al. “Burden of higher lead exposure in African-Americans starts in utero and persists into childhood.” Environment International. Vol. 108. Pp 221-227. November 2017. The health effects of disproportionate exposure to these environmental hazards suggest that poor communities and racial minorities are placed at greater risk for chronic and acute disease as well as developmental, neurologic and reproductive changes. Since the siting of polluting facilities and regulation of emissions is the result of a deliberate decision-making process, these disparities are hardly coincidental or benign; rather, they “may indicate underlying disparities in the power to influence that process.”53 Mikati, I., Benson, A. F., Luben, T. J., Sacks, J. D., & Richmond-Bryant, J. (2018). Disparities in distribution of particulate matter emission sources by race and poverty status. American journal of public health, 108(4), 480-485.

Figure 7. Changes in mortality attributable to ambient PM pollution by country, 1990-2015. Source: Cohen, et al., 2017.

Environmental injustice is not limited to the United States. Between and within countries, different populations are exposure to a range of different hazards. Although these disparities are seen across a range of environmental measures, they are best illustrated by air pollution. Mean annual exposure to PM2.5 air pollution in the U.S. was 8.4 μg/m3; that same year exposure measured at 106.2 μg/m3 in Saudi Arabia, 74.3 μg/m3 in India, and 58.4 μg/m3 in China.

The causes of these differences in air pollution include environmental conditions (e.g. dust storms in Saudi Arabia), pollution levels (e.g. in 2016, India had 22 of the 50 most polluted cities in the world) and industrial production (e.g. un-filtered coal burning power plants in China). In China, India, Bangladesh, and Japan, increases in exposure to PM combined with increases in population growth and an ageing population have led to net increases in mortality attributable to air pollution exposure (Figure 7).

This means that many countries have a growing and increasingly vulnerable population (elderly people, along with children, pregnant women, and people with pre-existing asthma, cardiovascular disease, and lung disease) being exposed to hazardous and injurious levels of air pollution; as a result, a greater number of deaths associated with air pollution exposure are occurring in these countries. Within countries, disparities in environmental exposure by socioeconomic status (SES) have been observed in North America, Asia and Africa; studies conducted on European cities have had more mixed results.54 Hajat, A., Hsia, C., & O’Neill, M. S. (2015). Socioeconomic disparities and air pollution exposure: a global review. Current environmental health reports, 2(4), 440-450. A meta-analysis of 37 studies regarding SES disparities and air pollution exposure included 22 North American studies, 10 European studies and 5 studies from New Zealand (3 studies), Asia (1 study in Hong Kong) and Africa (1 study in Ghana) (Hajat, et al., 2015); this distribution of studies itself suggests the need for greater geographic diversity in research on this topic. In New Zealand, low-income neighborhoods had higher concentrations of PM10 compared to higher SES areas.55 Hajat, A., Hsia, C., & O’Neill, M. S. (2015). Socioeconomic disparities and air pollution exposure: a global review. Current environmental health reports, 2(4), 440-450. The Ghana air pollution inequality study found that community SES was inversely related to both PM2.5 and PM10.56 Hajat, A., Hsia, C., & O’Neill, M. S. (2015). Socioeconomic disparities and air pollution exposure: a global review. Current environmental health reports, 2(4), 440-450. In Hong Kong, among those living in private housing, the lower SES population had higher exposure to PM10 and other air pollutants compared to the high SES population; among those living in public housing (low income families), no such inequalities were detected.57 Hajat, A., Hsia, C., & O’Neill, M. S. (2015). Socioeconomic disparities and air pollution exposure: a global review. Current environmental health reports, 2(4), 440-450. The authors hypothesize that the location of public housing (compared to siting of private housing) was an important factor in reducing residents’ exposure to traffic related air pollution.58 Hajat, A., Hsia, C., & O’Neill, M. S. (2015). Socioeconomic disparities and air pollution exposure: a global review. Current environmental health reports, 2(4), 440-450

Merely looking at overall air pollution trends globally, or even by country, belies the real burden that many populations face. Environmental exposures, whether by design or coincidence, disproportionately fall on already vulnerable populations. This inequitable burden should be a crucial piece of environmental conversations.

Box 5. Comparing Global and Local Air Pollution Analyses

The UESI develops standardized methods for measuring the distribution of environmental outcomes, to understand not only how a specific city performs, but how cities perform in relation to each other. To do so, it relies on global datasets and comparable approaches, that enable comparisons across different contexts. This strategy is particularly relevant in many developing cities, where data collection and monitoring systems are often still under construction, and decision makers have scarce data to inform urban interventions.

At the same time, we recognize that this approach may overlook more local data gathered in other urban areas. Many cities – especially those with a strong track record in environmental policy – have robust monitoring systems, reflecting years of data collection. In these cities, the UESI can leverage this more granular data to reveal patterns that global datasets might not be able to show, and shed light onto how environmental outcomes are distributed in a city.

To explore how the UESI can illuminate trends in cities with granular environmental data, we analyze two cities, using air pollution data at a higher – that is, a more detailed – resolution than the satellite-derived data source used to calculate the UESI’s air quality indicators. New York City’s Community Air Survey.59 The New York City Community Air Survey – Neighborhood Air Quality 2008 – 2016. Dataset available at https://data.cityofnewyork.us/Environment/NYCCAS-Air-Pollution-Rasters/q68s-8qxv. includes an extensive network of air quality monitors, and its results have allowed the New York Department of Health to create a spatially disaggregated air pollution dataset – using land use regression modeling – with a resolution of 300 meters. A recent published paper from Xu et al. (2018)60 Xu, H., Bechle, M. J., Wang, M., Szpiro, A. A., Vedal, S., Bai, Y., & Marshall, J. D. (2018). National PM2. 5 and NO2 exposure models for China based on land use regression, satellite measurements, and universal kriging. Science of The Total Environment. has developed national PM2.5 models for China – including for its capital city of Beijing – using land use regression and other statistical methods. 

Figure 1 shows the difference in the granularity of PM2.5 data between the New York Health Department Data (A) and the dataset used for the UESI (B). It demonstrates differences in the location of pockets of air pollution, as well as an overall difference in the concentration of PM2.5 : the New York Health Department data has an average city value of 7.83 micrograms/m3, while the UESI data shows an average value of 10.8 micrograms/m3. To the contrary, the distribution and average value for PM2.5 are similar for both sources of air pollution data in Beijing. These comparisons indicate there may be inconsistencies between the UESI61 van Donkelaar, A., R.V Martin, M.Brauer, N. C. Hsu, R. A. Kahn, R. C Levy, A. Lyapustin, A. M. Sayer, and D. M Winker, Global Estimates of Fine Particulate Matter using a Combined Geophysical-Statistical Method with Information from Satellites, Models, and Monitors, Environ. Sci. Technol, doi: 10.1021/acs.est.5b05833, 2016. data when compared to more disaggregated models. As a result, the UESI should not be considered a perfect substitute for more specific and local air pollutions models, but as a first source of data for places that don’t have specific and spatially disaggregated datasets yet – like many cities in the developing world. 

Figure 1. Comparison between raster dataset of PM2.5 data from the New York Health Department (A) and the dataset used for the UESI (van Donkelaar et al.) (B).

Using the methods developed by the UESI to analyze distributional equity, Figure 2 compares the results of the equity analysis shown using the UESI’s dataset (e.g. the results shown in the UESI report), and the results generated using the more granular datasets for both New York (A) and Beijing (B). Interestingly, and despite some of the differences seen in the overall PM2.5 values, the results of the equity analysis are not significantly different in either city, as the new Concentration curves for both New York and Beijing closely follow those generated using the UESI’s  dataset. For New York City, there is a slight difference starting around the 30 percent cumulative population value, where the concentration curves start to slightly diverge to opposites sides of the perfect equity line. This divergence means that the more granular dataset indicates that the concentration of PM2.5 might be slightly more allocated to wealthiest districts, while the UESI’s initial analysis indicates the concentration of PM2.5 might be slightly more allocated to less affluent districts. Overall, the results are still within very close proximity to each other and to the lines of equity, suggesting that overall the distribution is fairly equitable. Beijing’s results are similar, with both Concentration curves following a similar pattern and falling the same side of the line of perfect equity, indicating that the allocation of PM2.5 is indeed falling on the wealthiest districts of the city.

Figure 2. Comparison of the distributional equity analysis of both datasets for the city of Ney York – NY Health Department and van Donkeelar et.al. – (A) and dataset for the city of Beijing – Xu et.al and van Dankeelar et.al. (B).

The results of this comparative analysis reveal that higher resolution data is a next natural step for the UESI and that its methods are applicable to recently developed datasets, providing relevant results for cities that have the resources and capacity to generate more granular and spatially disaggregated data. There are other reasons to opt for more granular data: ground-based monitoring data reflects air quality at the level where people live and breathe, rather than satellites, which provide proxy measures of on-the-ground conditions. Although many of the datasets used by the UESI were indeed generated at a global scale, they are still relevant for cities that due to their own contexts and capacities don’t have access to specific and localized datasets, however, as cities and researchers generate more localized and granular data this new sources can be rapidly included in the UESI framework and used to generate relevant and consistent analysis on the distribution of environmental equity and its linkages with income.

Factory icon adapted from Freepik from www.flaticon.com licensed by CC 3.0 BY
css.php