2009 Dec.: New Zealand & VA, U.S.A.: Climate change, tropospheric ozone and particulate matter, and health impacts
Ciência & Saúde Coletiva
Print version ISSN 1413-8123
Ciênc. saúde coletiva vol.14 no.6 Rio de Janeiro Dec. 2009
Climate change, tropospheric ozone and particulate matter, and health impacts*
Mudanças climáticas, ozônio troposférico e partículas suspensas: impactos na saúde
Kristie EbiI; Glenn McGregorII
IESS, LLC, Alexandria, Virginia, USA. 5249 Tancreti Lane, Alexandria, VA 22034. firstname.lastname@example.org
IIThe University of Auckland, Auckland, New Zealand
We review how climate change could affect future concentrations of tropospheric ozone and particulate matter (PM), and what changing concentrations could mean for population health, as well as studies projecting the impacts of climate change on air quality and the impacts of these changes on morbidity/mortality. Climate change could affect local to regional air quality through changes in chemical reaction rates, boundary layer heights that affect vertical mixing of pollutants, and changes in synoptic airflow patterns that govern pollutant transport. Sources of uncertainty are the degree of future climate change, future emissions of air pollutants and their precursors, and how population vulnerability may change in the future. Given the uncertainties, projections suggest that climate change will increase concentrations of tropospheric ozone, at least in high-income countries when precursor emissions are held constant, increasing morbidity/mortality. There are few projections for low- and middle-income countries. The evidence is less robust for PM, because few studies have been conducted. More research is needed to better understand the possible impacts of climate change on air pollution-related health impacts.
Key words: Air pollution, Climate change, Health impacts, Ozone, Particulate matter
Examinamos como as mudanças climáticas podem afetar o futuro das concentrações de ozônio troposférico e de partículas suspensas (PS), e o que uma mudança nas concentrações significaria para a saúde da população. Analisaram-se estudos que projetam os impactos das mudanças climáticas na qualidade do ar e morbidade/mortalidade. A mudança climática pode afetar a qualidade do ar local e regional com mudanças nas taxas de reações químicas, altura das camadas limite que afetam a mistura vertical de poluentes e mudanças nos padrões sinóticos de circulação atmosférica que gerenciam o transporte de poluentes. Dentre as fontes de incerteza, a mudança climática futura, emissões de poluentes do ar e seus precursores e como a vulnerabilidade da população poderá ser modificada no futuro. Devido às incertezas, as projeções sugerem que as mudanças climáticas aumentarão as concentrações de ozônio troposférico pelo menos nos países de alta renda, onde as emissões precursoras de ozônio são constantes, resultando em maior morbidade/mortalidade. Para os países de baixa e média renda, há poucas projeções. A evidência é menor para PS, pois há poucos estudos. É preciso mais pesquisa para compreender os possíveis impactos da mudança climática nos impactos à saúde relacionados à poluição do ar.
Palavras-chave: Poluição do ar, Mudanças climáticas, Impacto à saúde, Ozônio, Partículas em suspensão
Extensive literature documents the adverse health impacts of exposure to elevated concentrations of air pollutants, particularly ozone, particulate matter with aerodynamic diameters < 10 (PM10) and < 2.5 μm (PM2.5), sulfur dioxide, nitrogen dioxide, carbon monoxide, and lead. Worldwide in the year 2000, 0.8 million deaths and 7.9 million disability-adjusted life-years lost from respiratory problems, lung disease, and cancer were attributed to urban air pollution1. Because the state of the atmosphere at various scales determines the development, transport, dispersion, and deposition of air pollutants, there is concern that climate change could affect the burden of illness and mortality associated with these gases and fine particles. Therefore, we review studies projecting the impacts of climate change on air quality and studies projecting the impacts of these changes on morbidity and mortality, with a focus on studies published since 2000. We limited our review to the past several years because of significant advances in climate modeling2.
Meteorology and air pollution
Air pollution concentrations are the result of interactions among local weather patterns, atmospheric circulation features, wind, topography, human activities (i.e., transport and coal-fired electricity generation), human responses to weather changes (i.e., the onset of cold or warm spells may increase heating and cooling needs and therefore energy needs), and other factors. Some locations, because of their general climate and topographic setting, are predisposed to poor air quality because the climate is conducive to chemical reactions leading to the transformation of emissions, and the topography restricts the dispersion of pollutants3,4. Some air pollutants demonstrate clear seasonal cycles5-9. Certain weather situations provide the requisite meteorologic conditions for pollution episodes. Air pollution episodes are often associated with stationary or slowly migrating anticyclonic or high-pressure systems that reduce pollution dispersion, diffusion, and deposition10,11. The three-dimensional wind field, its related turbulence, and vertical temperature are important12,13. Meteorologic conditions also influence the chemical and physical processes involved in the formation of secondary pollutants such as ozone14,15. Airflow along the flanks of anticyclonic systems can transport ozone precursors, creating the conditions for an ozone event16-19. Large-scale airflows not necessarily related to anticyclonic systems can interact with local topography, sea/lake and land breezes, or mountain and valley winds to increase pollutant concentrations16,20-28. Distant weather systems such as tropical cyclones and low-pressure systems lying over coastal regions can lead to high pollution levels18,29,30. Climate change could affect local to regional air quality directly through changes in chemical reaction rates, boundary layer heights (i.e., the layer of air near the ground that is affected by diurnal heat, moisture, and momentum transfer to/from the surface) that affect vertical mixing of pollutants, and changes in synoptic airflow patterns that govern pollutant transport. The synoptic scale corresponds to the typical size of mid-latitude high and low pressure systems (approximately a horizontal length of 1,000 km or 620 miles). Indirect effects could result from increasing or decreasing anthropogenic emissions via changes in human behavior or from altering the levels of biogenic emissions because of higher temperatures and land cover change. Higher temperatures can increase emissions of isoprene, a volatile hydrocarbon and ozoneprecursor emitted by many woody plant species. However, establishing the scale (local, regional, global) and direction of change (improvement or deterioration) of air quality is challenging31-34. More is known about the potential impacts of climate change on ground-level ozone than on other air pollutants.
Ground-level ozone is a known pulmonary irritant that affects the respiratory mucous membranes, other lung tissues, and respiratory function. Exposure to elevated concentrations of ozone is associated with increased hospital admissions for pneumonia, chronic obstructive pulmonary disease, asthma, allergic rhinitis, and other respiratory diseases, and with premature mortality35-39. Outdoor ozone concentrations and activity patterns are the primary determinants of ozone exposure40. Although a considerable amount is known about the health effects of ozone in Europe and North America, few studies have been conducted in other regions. Ground-level ozone is both naturally occurring and, as the primary constituent of urban smog, a secondary pollutant formed through photochemical reactions involving nitrogen oxides and volatile organic compounds in the presence of bright sunshine with high temperatures41. Land use changes over the past century affect ozone concentrations by altering vegetation patterns affecting biogenic volatile organic compound emissions that influence ozone production2. In addition, urbanization leading to heat islands can influence the local production and dispersion of ozone. In urban areas, gasoline-burning engines are major sources of volatile organic compounds, and nitrogen oxides are produced whenever fossil fuels are burned41. Temperature, wind, solar radiation, atmospheric moisture, venting, and mixing affect both emissions of ozone precursors and production of ozone14,15,42. Because ozone formation depends on sunlight, concentrations are typically highest during the summer months, although not all cities have shown seasonality in ozone concentrations43. Observations show that trends in tropospheric ozone in the past few decades vary in sign, with increased or decreased ozone, and in magnitude in many locations, with significant upward trends at low latitudes2.
Projected changes in tropospheric concentrations of ozone associated with climate change
There are two major sources of uncertainty when assessing the health impacts of future changes in tropospheric ozone concentrations: the extent of future changes in emissions of ozone precursors, and the degree to which future weather conditions could increase ozone concentrations. Future emissions are, of course, uncertain and depend on assumptions of population growth, economic development, regulatory actions, and energy use44,45. Increased regulation of anthropogenic emissions of volatile organic compounds and nitrogen oxides from gasoline-powered engines means that biomass burning, including fires, will likely increase in importance as sources of ozone precursors. Assuming no change in the emissions of ozone precursors, the extent to which climate change affects the frequency of future ozone episodes will depend on the occurrence of the required meteorologic conditions46-52. Where climate change is projected to result in an increased frequency of stable anticyclonic conditions with little boundary layer ventilation and associated high temperatures, cloud-free conditions, and large solar radiation inputs, exceedance of current air quality standards will likely occur46,49,50,52. Future air quality, especially at the local to regional level, will depend partially on concentrations of pollutants at the global scale. Concentrations of ozone have risen since preindustrial times because of increasing emissions of methane, carbon monoxide, and nitrogen oxides, and this trend is expected to continue over the next 50 years, based on projections of annual mean maximum concentrations53. However, as many major cities propose to reduce vehiclebased emissions of pollutants, it is expected that urban concentrations of ozone will rise less rapidly or be reduced54,55. For example, it has been estimated that for the United States a 50% reduction of methane emissions would nearly halve the incidence of high ozone events56. Decreases in stratospheric ozone may also result in greater increases in ground level ozone in polluted regions because of an increase in ultraviolet radiation reaching the ground; ultraviolet radiation is involved in the formation of ozone57. Changes in concentrations of ground-level ozone driven by scenarios of future emissions and/or weather patterns have been projected worldwide, with most projections for Europe and North America36,48,50,51,58-71; these studies, which range from global to local levels, are summarized in subsequent paragraphs. Although these studies are inconsistent in approaches taken and factors considered, most project increased tropospheric ozone concentrations, with high variability across regions. The Intergovernmental Panel on Climate Change concluded that climate change would modify a variety of chemicals and processes that control air quality, and the net effects are likely to vary from one region to another2. On a global scale, if 1990 is treated as the reference period, then moderately high annual mean maximum ozone concentrations of 60 parts per billion (ppb) were projected for central Europe, China, Brazil, South Africa, and eastern North America during summertime58. By 2030, under a high emission scenario [Standardized Reference Emission Scenarios (SRES); see Appendix 1, A2 scenario], the area experiencing a background of 60 ppb was projected to expand significantly, especially in Europe and North America. By 2060, most of the populated continental areas would experience ozone concentrations of at least 60 ppb. By 2100, much of the Northern Hemisphere was projected to have annual mean maximum ozone levels of 60 ppb, as were most of the populated areas of the Southern Hemisphere58. However, Liao et al.48 and Racherla and Adams51, also using the SRES A2 scenario, projected that anthropogenic climate change could reduce the global ozone burden due to changes in atmospheric chemistry. Liao et al.48 also projected that surface ozone concentrations over or near populated and biomass-burning areas would increase. Ozone concentrations were projected for 10 world regions in 2030, using a coupled general circulation model with interactive chemistry (LMDz-INCA; http://aoc.ammainternational.org/researchProduct/aerosolchemistry/lmdz/index.en.php?current=20060930) driven by the SRES A2 scenario71. Modeled ozone concentrations for present conditions have been shown to reasonably agree with surface ozone measurements. The global average populationweighted 8-hr maximum ozone concentration was projected to increase by 9.4 parts per billion per volume (ppbv) compared with a simulation of the concentration in 2000, with the largest increases over South Asia (nearly 15 ppbv) and with large increases in the Middle East, Southeast Asia, Latin America, and East Asia. Forkel and Knoche72 projected ozone concentrations in Germany under the IS92a "business-as-usual" scenario for the 2030s compared with the 1990s. Both biogenic volatile organic compound emissions and soil nitrous oxide emissions were projected to increase as temperatures rise. Projected daily maximum ozone concentrations increased by between 2 and 6 ppb (6-10%) across the study region; the number of days in the 2030s when daily maximum ozone exceeded 90 ppb increased nearly 4-fold, from 99 to 384.
Murazaki and Hess50, using the SRES A1 scenario and a global chemical transport model [MOZART-2; Model of OZone And Related chemical Tracers, version 2 (gctm.acd.ucar.edu/Mozart)], projected that by the end of the twenty-first century anthropogenic climate change alone would decrease background ozone concentrations over the United States, while ozone produced internally would increase. Over the western United States, the two forces approximately equaled each other. The authors projected that over the eastern United States, up to 12 additional days annually would exceed 80 ppbv. Taha70 estimated increases in ozone concentrations by the end of the century in two large cities in California based on model results that linked output from two general circulation models to future emissions of nitrous oxides, volatile organic compounds, biogenic hydrocarbons, and sulfates, and air pollution models used to evaluate air quality compliance in these regions. Two frequently used attainmentdemonstration modeling episodes were selected for the study, the 2010 projection of the 26-28 August "1987" episode for Los Angeles, and the 2005 projection of the 11-13 July "1990" episode for Sacramento Valley. Ozone concentrations on the last day of each episode were modeled. Under assumptions of futureyear controlled emissions, the model suggested significant increases in ozone concentrations at the time of the base-case peak concentrations in the Los Angeles Basin (up to 26 ppb, an approximate 24% increase) and in the Sacramento Valley (up to 12 ppb, an approximate 10% increase). Aw and Kleeman59 simulated an episode of high air pollution in Southern California in 1996 with observed meteorology and then with higher temperatures. Ozone concentrations increased up to 16% with higher temperatures (+ 5°K); there was less consistency in PM2.5 response, depending on whether increased secondary particle formation or more evaporative losses from nitrate particles were more important. Steiner et al.68 reported variations across California in the sensitivity of ozone to changing temperatures, absolute humidity, biogenic volatile organic compound emissions, and pollution boundary conditions on a fine scale (4 km grid resolution). In a coarse-scale analysis of pollution over the continental United States, Mickley et al.49 projected that, because of climate change alone (SRES A1b scenario), air pollution (as estimated by including combustible carbon monoxide and black carbon as tracers of anthropogenic pollution) could increase in the upper Midwest because of decreases between 2000 and 2052 in the frequency of Canadian frontal passages that clear away stagnating air pollution. Leung and Gustafson67 used regional climate simulations for temperature, solar radiation, precipitation, and stagnation/ventilation, and projected worse air quality in Texas and better air quality in the Midwest in 2045-2055 compared with 1995-2005. Bell et al.36 showed greater sensitivity of ozone concentrations in the Mid-Atlantic to changes in biogenic than to changes in anthropogenic emissions. As part of the New York Climate and Health Project, Hogrefe and colleagues conducted local-scale analyses of air pollution impacts of future climate changes using integrated modeling46,61-64 to examine the potential impacts of climate and land use changes on heat- and ozone-related health impacts in the New York City metropolitan area36,73-75. Hourly meteorologic data from the 1990s through the 2080s were simulated based on the SRES A2 and B2 scenarios. The global climate outputs were downscaled to a 36-km grid over the eastern United States using the MM5 (Penn State/National Center for Atmospheric Research Mesoscale Model 5; www.mmm.ucar.edu/mm5/) regional climate model. The MM5 results were then used as inputs to the CMAQ (Community Multiscale Air Quality) regional-scale air quality model. Five summers (June-August) in each of four decades (1990s, 2020s, 2050s, and 2080s) were simulated at the 36-km scale. Pollution precursor emissions over the eastern United States were based on U.S. EPA estimates at the county level for 1996. Compared with observations from ozone monitoring stations, initial projections were consistent with ozone spatial and temporal patterns over the eastern United States in the 1990s46. Average daily maximum 8-hr concentrations were projected to increase by 2.7, 4.2, and 5.0 ppb in the 2020s, 2050s, and 2080s, respectively, because of climate change61. The influence of climate on mean ozone values was similar in magnitude to the influence of rising global background by the 2050s, but climate had a larger impact on extreme values. When biogenic volatile organic emissions were allowed to increase in response to warming, an additional increase in ozone concentrations was projected that was similar in magnitude to that of climate alone61. Climate change shifted the distribution of ozone concentrations toward higher values, with larger relative increases in future decades.
PM is well known to affect morbidity and mortality76-79, so increasing concentrations would have significant negative health impacts. Using a coupled climate-air pollution, three-dimensional model, Jacobson80 compared the health effects of preindustrial versus present-day atmospheric concentrations of carbon dioxide. The results suggest that increasing concentrations of CO2 increased tropospheric ozone and PM2.5, which increased mortality by about 1.1% per degree temperature increase over the baseline rate. Jacobson80 estimated that about 40% of the increase was attributable to ozone and the rest to PM. The estimated mortality increase was higher in locations with poorer air quality. In comparison with ozone, assessments of the impact of climate change on other pollutants are few. These emphasize the role of local abatement strategies in determining the future concentrations of pollutants such as PM and sulfur dioxide and tend to project the probability of air quality standards being exceeded instead of absolute concentrations81-84. The results vary by region. The severity and duration of summertime regional air pollution (combustion carbon monoxide and black carbon) episodes were projected to increase in the northeastern and midwestern United States by 2045-2052 because of climate change-induced decreases in the frequency of surface cyclones49. A U.K. study projected that climate change would result in a large decrease in days with high particulate concentrations due to changes in meteorologic conditions58. However, in the New York Climate and Health Project, PM2.5 concentrations were projected to increase with climate change, with the effects differing by component species, with sulfates and primary PM increasing markedly and with organic and nitrated components decreasing, mainly because of movement of these volatile species from the particulate to the gaseous phase64,65. Because transboundary transport of pollutants plays a significant role in determining local to regional air quality85-101, changing patterns of atmospheric circulation at the hemispheric to global level are likely to be equally important as regional patterns for future local air quality102,103.
Potential health effects
Table 1 summarizes projections of morbidity and mortality based on current exposure- mortality relationships applied to projected Climate change, air pollution, and health impacts ozone concentrations. An increase in ozone concentrations would affect the ability of regions to achieve air quality targets. There are few projections for cities in low- or middleincome countries, despite the heavier pollution burdens in these populations, presumably because of limited research funding.
The New York Climate and Health Project projected the potential health impacts of future ozone concentrations in the eastern United States36,75. Knowlton and colleagues computed absolute and percentage increases in ozone-related daily summer-season deaths in the New York City metropolitan region in the 2050s compared with the 1990s74,75. The availability of county-scale ozone projections made it possible to compare impacts in the urban core with those in outlying areas. Increases in ozone-related mortality due to climate change ranged from 0.4 to 7.0% across 31 counties. Bell et al.36 expanded the analysis to 50 eastern cities and examined both mortality and hospital admissions. Average ozone concentrations were projected to increase by 4.4 ppb (7.4%) in the 2050s; the range was 0.8-13.7%. In addition, ozone red-alert days could increase by 68%.
These changes were projected to result in an 0.11 to 0.27% increase in nonaccidental mortality and an average 0.31% increase in cardiovascular disease mortality. Ozone concentrations for Los Angeles in 2050 were projected under the SRES A2 emission scenario105. Using several estimates of the ozone exposure-response relationship, Hwang et al.105 projected that mortality and hospital admissions would together increase up to approximately 3.7%, with most of the projected increases < 1%, depending on the city and the health outcome. As part of a U.K. assessment of the potential impacts of climate change, Anderson et al.58 used projected daily meteorologic parameters for each day through December 2099, driven by the Intergovernmental Panel on Climate Change IS92a (business as usual) scenario, for a single grid point representing the British Isles. A global three-dimensional chemistry model was used to calculate the influence of the projected increases in emissions of methane, carbon monoxide, and nitrogen oxides from human activities on the global distribution of ozone through the year 2100. The impact of climate change on increases in the frequency and severity of the meteorologic conditions that lead to summertime ozone episodes was projected to be reduced by changes in European emissions of ozone precursor species. When the authors assumed thresholds for the health effects of ozone, the increase in health effects due to ozone was relatively small. If no threshold was assumed, then ozone was projected to increase premature deaths by 10, 20, and 40% for the years 2020, 2050, and 2080, respectively. Approximately 500,000 excess deaths were estimated for the year 2030 due to the impacts of changing ozone concentrations and population growth in 10 world regions under the SRES A2 scenario71. The daily acute mortality coefficient per parts per billion per volume ozone was taken from a study of 95 cities in the United States35. Assuming a low concentration threshold of 25 ppbv and taking into consideration recently enacted legislation to control zone precursors, an estimated 191,000 deaths would be avoided globally (0.2% of the projected total number of deaths in 2030). An estimated 458,000 deaths would be avoided (0.5% of the projected total number of deaths in 2030) if currently available emission control technologies were aggressively employed globally. Sensitivity analyses showed that the results were significantly affected by the threshold assumed and the daily acute ozone mortality coefficient used.
Poor air quality currently affects the health of millions of people. Climate change has the potential to increase harmful exposures to elevated concentrations of ozone and PM2.5 through changes in regional weather patterns. However, there is high uncertainty about future projections. Sources of uncertainty include not only future climate change but also future emissions of greenhouse gases, ozone precursors, and other pollutants, as well as how population vulnerability and activity patterns may differ in the future. Because of the high uncertainty of the extent and effectiveness of future emissions reductions, most studies that projected the impacts of climate change on air quality focused on future climate change alone and held precursor emissions constant over future decades. Therefore, the focus was on examining the sensitivity of ozone concentrations to alternative future climates rather than on attempting to project actual future ozone concentrations. On the basis of a limited number of modeling studies, climate change is likely to increase ozone concentrations in highincome countries when precursor emissions are held constant, leading to increased morbidity and mortality. There is less certainty of the possible impact of climate change on fine particulate concentrations. More stringent emissions controls for ozone, PM2.5, and other pollutants can be expected with the growing body of evidence of the adverse health impacts of these air pollutants. Therefore, the extent to which climate change affects air quality will depend partially on ongoing regulatory control of ozone and PM2.5. At the same time, population sensitivity may change because of medical advances and changes in risk factors. The main public health responses to the projected health impacts of climate change are mitigation and adaptation. Adaptation is not an effective risk management strategy for poor air quality, because physiologic mechanisms to decrease susceptibility to ozone and other air pollutants are limited. Therefore, if improved model experiments continue to project higher ozone concentrations under a changing climate, rapid reductions of emissions from fossil- fuel burning are needed to protect the health of current and future generations. Evidence suggests that reducing current tropospheric ozone concentrations reduces morbidity and mortality, with significant savings in medical care costs106. For relevant agencies and institutions to develop appropriate and timely responses, additional research is needed to reduce the uncertainties associated with projections of the health impacts of changing concentrations of ozone and PM due to climate change. Research is needed to better understand the impacts of future emissions pathways, climate change impacts on concentrations of fine particles and gases, how changing weather patterns could influence the frequency and severity of episodes of poor air quality, population sensitivity, and how these factors might interact. Increasing greenhouse gas emissions suggest that future air quality could decline without increased regulations and development and deployment of new technologies.
Appendix 1: SRES
SRES were developed by the Intergovernmental Panel on Climate Change as alternative images of how the future might unfold107. Four different narrative storylines were developed to describe the relationships between the driving forces of greenhouse gas emissions and their evolution. Probabilities or likelihood were not assigned to the individual scenarios. There is no single most likely, or best guess, scenario. None of the scenarios represents an estimate of a central tendency for all driving forces or emissions. Each SRES storyline assumes a distinctly different direction for future development, such that the four storylines differ in increasingly irreversible ways. The storylines were created along two dimensions: global versus regional development patterns and whether economic or environmental concerns would be primary. It is important to note that the scenarios do not cover all possible future worlds. For example, there is no SRES world in which absolute incomes are constant or falling. The A2 and B2 storylines are frequently used in modeling health impacts. The A2 storyline describes a heterogeneous world with an underlying theme of selfreliance and preservation of local identities. Fertility patterns across regions vary slowly, resulting in continuously increasing global population. Economic development is primarily region oriented, and per capita economic growth and technological change are fragmented and slower compared with the other scenarios. The B2 storyline describes a world in which the emphasis is on local solutions to economic, social, and environmental sustainability. It is a world with continuously increasing global population (at a rate slower than that of A2), intermediate levels of economic development, and less rapid and more diverse technological change. The total cumulative CO2 emissions are categorized as very high in the A2 scenario [> 1,800 GtC (gigatons)] and medium-low in the B2 scenario (1,100-1,450 GtC) in 2100.