Pollution Prevention and Abatement Handbook WORLD BANK GROUP Effective July 1998

Airborne Particulate Matter

Airborne particulate matter, which includes dust, haust, and secondary organic compounds formed dirt, soot, smoke, and liquid droplets emitted into by photochemistry). These species may be the most the air, is small enough to be suspended in the abundant fine particles after sulfates. Additionally, atmosphere. Airborne particulates may be a com- atmospheric reactions of nitrogen oxides produce plex mixture of organic and inorganic substances. nitric acid vapor (HNO3) that may accumulate as They can be characterized by their physical at- nitrate particles in both fine and coarse forms. The tributes, which influence their transport and most common combination of coarse particles con- deposition, and their chemical composition, sists of oxides of silicon, aluminum, calcium, and which influences their effect on health. iron. The physical attributes of airborne particulates include mass concentration and size distribution. Terms and Sampling Techniques Ambient levels of mass concentration are mea- sured in micrograms per cubic meter (µg/m3); size Several terms are used to describe particulates. attributes are usually measured in aerodynamic Generally, these terms are associated with the diameter. Particulate matter (PM) exceeding 2.5 sampling method: microns (µm) in aerodynamic diameter is gener- Total suspended particulates (TSP) includes par- ally defined as coarse particles, while particles ticles of various sizes. Some proportion of TSP smaller than 2.5 microns (PM2.5) are called fine par- consists of particles too large to enter the human ticles. The acid component of particulate matter, respiratory tract; therefore, TSP is not a good in- and most of its mutagenic activity, are generally dicator of health-related exposure. TSP is mea- contained in fine particles, although some coarse sured by a high-volume gravimetric sampler that acid droplets are also present in fog. Samples taken collects suspended particles on a glass-fiber fil- in the United States showed that about 30% of par- ter. The upper limit for TSP is 45 microns in diam- ticulate matter was in the fine fraction (Stern et al. eter in the United States and up to 160 microns in 1984). Europe. Particles interact with various substances in the TSP sampling and TSP-based standards were air to form organic or inorganic chemical com- used in the United States until 1987. Several coun- pounds. The most common combinations of fine tries in Central and Eastern Europe, Latin particles are those with sulfates. In the United America, and Asia still monitor and set standards States, sulfate ions account for about 40% of fine based on measurements of TSP. As monitoring particulates and may also be present in concen- methods and data analysis have become more trations exceeding 10 µg/m3 (USEPA 1982b). The sophisticated, the focus of attention has gradu- smaller particles contain the secondarily formed ally shifted to fine particulates. Recent evidence aerosols, combustion particles, and recondensed shows that fine particulates, which can reach the organic and metal vapors. The carbonaceous thoracic regions of the respiratory tract, or lower, component of fine particles—products of incom- are responsible for most of the excess mortality plete combustion—contains both elemental car- and morbidity associated with high levels of ex- bon (graphite and soot) and nonvolatile organic posure to particulates. Most sophisticated stud- carbon (hydrocarbons emitted in combustion ex- ies suggest that fine particulates are the sole factor

201 202 PROJECT GUIDELINES: POLLUTANTS accounting for this health damage, while exposure sult of combustion processes, including the burn- to coarse particulates has little or no independent ing of fossil fuels for steam generation, heating and effect. household cooking, agricultural field burning, die- The particles most likely to cause adverse health sel-fueled engine combustion, and various indus- effects are the fine particulates PM10 and PM2.5 — trial processes. Emissions from these particles smaller than 10 microns and 2.5 microns anthropogenic sources tend to be in fine fractions. in aerodynamic diameter, respectively. They are However, some industrial and other processes that sampled using (a) a high-volume sampler with a produce large amounts of dust, such as cement size-selective inlet using a quartz filter or (b) a di- manufacturing, mining, stone crushing, and flour chotomous sampler that operates at a slower flow milling, tend to generate particles larger than 1 rate, separating on a Teflon filter particles smaller micron and mostly larger than 2.5 microns. In than 2.5 microns and sizes between 2.5 microns and cold and temperate parts of the world, domestic 10 microns. No generally accepted conversion coal burning has been a major contributor to the method exists between TSP and PM10, which may particulate content of urban air. Traffic-related constitute between 40% and 70% of TSP emissions may make a substantial contribution to (USEPA1982b). the concentration of suspended particulates in ar- In 1987, the USEPA switched its air quality stan- eas close to traffic. Some agroindustrial processes dards from TSP to PM10. PM10 standards have also and road traffic represent additional anthropo- been adopted in, for example, Brazil, Japan, and genic sources of mostly coarse particulate emis- the Philippines. In light of the emerging evidence sions. on the health impacts of fine particulates, the The largest stationary sources of particulate USEPA has proposed that U.S. ambient standards emissions include fossil-fuel-based thermal for airborne particulates be defined in fine particu- power plants, metallurgical processes, and ce- lates. ment manufacturing. The physical and chemical Black smoke (BS) is a particulate measure that composition of particulate emissions is deter- typically contains at least 50% respirable particu- mined by the nature of pollution sources. Most lates smaller than 4.5 mm in aerodynamic diam- particles emitted by anthropogenic sources are eter, sampled by the British smokeshade (BS) less than 2.5 microns in diameter and include a method. The reflectance of light is measured by larger variety of toxic elements than particles emit- the darkness of the stain caused by particulates ted by natural sources. Fossil fuel combustion gen- on a white filter paper. The result of BS sampling erates metal and sulfur particulate emissions, depends on the density of the stain and the opti- depending on the chemical composition of the cal properties of the particulates. Because the fuel used. The USEPA (1982b) estimates that more method is based on reflectance from elemental than 90% of fine particulates emitted from sta- carbon, its use is recommended in areas where tionary combustion sources are combined with coal smoke from domestic fires is the dominant sulfur dioxide (SO2). Sulfates, however, do not component of ambient particulates (WHO and necessarily form the largest fraction of fine UNEP 1992). After reviewing the available data, particulates. In locations such as Bangkok, Ostro (1994) concluded that BS is roughly equiva- Chongqing (China), and São Paulo (Brazil), or- lent to PM10. However, there is no precise equiva- ganic carbon compounds account for a larger lence of the black smoke measurements with fraction of fine particulates, reflecting the role of other methods. The BS measure is most widely emissions from diesel and two-stroke vehicles or used in Great Britain and elsewhere in Europe. of smoke from burning coal and charcoal. Al- though sulfates represent a significant share (30– Sources of Particulates 40%) of fine particulates in these cases, care is required before making general assertions about Some particulates come from natural sources such the relationship between sulfates and fine par- as evaporated sea spray, windborne pollen, dust, ticulates, since the sources and species charac- and volcanic or other geothermal eruptions. Par- teristics of fine particulates may vary significantly ticulates from natural sources tend to be coarse. across locations. Combustion devices may emit Almost all fine particulates are generated as a re- particulates comprised of products of incomplete Airborne Particulate Matter 203 combustion (PICs, which may include toxic organ- Population-based cross-sectional and longitudi- ics) and toxic metals, which are present in the fuel nal studies (see, for example, Lipfert 1984; Dockery and in some cases may also be carcinogenic. Par- et al. 1993) have found an association between ticulates emitted by thermal power generation may long-term exposure and mortality. Using 14-to-16- contain lead, mercury, and other heavy metals. The year studies in six U.S. cities, and controlling for melting, pouring, and torch-cutting procedures of individual risk factors, including age, sex, smok- metallurgy emit metal particulates containing lead, ing, body-mass index, and occupational exposure, cadmium, and nickel. Particles emitted by the ce- Dockery et al. (1993) found a significant connec- ment industry are largely stone or clay-based par- tion between particulate air pollution and excess ticulates that may contain toxic metals such as mortality at average annual PM10 concentrations lead. as low as 18 µg/m3, well below the current U.S. ambient standard of 50 µg/m3. Studies on the ef- Impacts of Exposure fect of particulates on human health summarized by Ostro (1994) suggest an increase in human mor- Human Health Effects tality rates ranging from 0.3% to 1.6% for each 10 3 µg/m increase in average annual PM10 concentra- The respiratory system is the major route of entry tions. for airborne particulates. The deposition of particu- A study conducted on over a half million lates in different parts of the human respiratory people in 151 U.S. metropolitan areas during system depends on particle size, shape, density, and 1982–89 by Pope et al. (1995) found that death individual breathing patterns (mouth or nose rates in the areas most polluted with fine par- breathing). The effect on the human organism is ticulates were 17% higher than in the least pol- also influenced by the chemical composition of the luted areas, as a result of a 31% higher rate of particles, the duration of exposure, and individual death from heart and lung disease, even when susceptibility. While all particles smaller than 10 most cities complied with the U.S. federal stan- microns in diameter can reach the human lungs, dards for particulate pollution. Cities with aver- the retention rate is largest for the finer particles. age pollution that complied with federal standards Products of incomplete combustion, which form still had about a 5% higher death rate than the a significant portion of the fine particulates, may cleanest cities. enter deep into the lungs. PICs contribute signifi- In addition, relationships between morbidity cantly to health impacts associated with fine par- and short- and long-term exposure to particu- ticulates. late matter have been found in a number of stud- Clinical, epidemiologic, and toxicological ies. Schwartz et al. (1993) found a significant sources are used to estimate the mortality and increase in emergency room visits among people morbidity effects of short- and long-term expo- under the age of 65 in areas with daily average PM10 sure to various particulate concentration levels. concentrations that were less than 70% of the U.S. Several studies have found statistically signifi- air quality standard of 150 µg/m3. Several studies cant relationships between high short-term am- carried out in Canada, Germany, Switzerland, and bient particulate concentrations and excess the United States have found an association be- mortality in London and elsewhere. The esti- tween respiratory symptoms and exposure to long- mated 4,000 excess deaths in the London met- term ambient particulate concentrations of about ropolitan area in December 1952 were associated 30–35 µg/m3, without any evidence of a threshold with BS measurements equivalent to a 4,000 µg/ level below which health effects do not occur. (For m3 maximum daily average ambient concentration a summary, see Schwartz 1991/92.) Kane (1994) of particulates (Schwartz and Dockery 1992b). demonstrated an association between mineral Schwartz (1993b) has also found a significant as- dusts such as silica or asbestos fibers accumulat- sociation between daily average PM10 concentra- ing in the lungs and a characteristic spectrum of tions and mortality at concentrations below the diseases. Recently, the potential carcinogenic effect current U.S. standard of 150 µg/m3 for short-term of certain dust compounds has been analyzed, and

PM10 concentrations. in some cases (for example, for silica dust), 204 PROJECT GUIDELINES: POLLUTANTS limited evidence of carcinogenic effects has been the EU and USEPA standards also reflect the tech- found (see Ulm 1994). nological feasibility of meeting the standards. In Recent epidemiologic evidence (for example, the EU, a prolonged consultation and legislative Schwartz 1991/92; Schwartz and Dockery 1992b; decisionmaking process took into account the Ostro 1994) suggests that there may be no safe environmental conditions and the economic and threshold for fine particulate matter and that the social development of the various regions and effects are linearly related to concentration. countries and acknowledged a phased approach to compliance. A potential tradeoff was also rec- Other Effects ognized in the guidelines for the combined ef- fects of sulfur dioxide and particulate matter (see Vegetation exposed to wet and dry deposition of European Community 1992). particulates may be injured when particulates are combined with other pollutants. Coarse particles, Conclusions such as dust, directly deposited on leaf surfaces can reduce gas exchange and photosynthesis, The main objective of air quality guidelines and leading to reduced plant growth. Heavy metals standards is the protection of human health. Since that may be present in particulates, when depos- fine particulates (PM10) are more likely to cause ited on soil, inhibit the process in soil that makes adverse health effects than coarse particulates, nutrients available to plants. This, combined with guidelines and standards referring to fine par- the effects of particulates on leaves, may contrib- ticulate concentrations are preferred to those re- ute to reduction of plant growth and yields. In ferring to TSP, which includes coarse particulate addition, particulates contribute to the soiling concentrations. and erosion of buildings, materials, and paint, Scientific studies provide ample evidence of leading to increased cleaning and maintenance the relationship between exposure to short-term costs and to loss of utility. and long-term ambient particulate concentra- Particulate emissions have their greatest im- tions and human mortality and morbidity effects. pact on terrestrial ecosystems in the vicinity of However, the dose-response mechanism is not emissions sources. Ecological alterations may be yet fully understood. Furthermore, according to the result of particulate emissions that include WHO (1987), there is no safe threshold level toxic elements. Furthermore, the presence of fine below which health damage does not occur. particulates may cause light scattering, or atmo- Therefore, policymakers may have to consider ac- spheric haze, reducing visibility and adversely ceptable risk rather than try to achieve absolute affecting transport safety, property values, and safety when setting ambient particulate concen- aesthetics. tration standards. Furthermore, ambient guide- lines can become an effective part of the Ambient Standards and Guidelines environmental management system only if implementation is feasible and the enforcement The most frequently used reference guidelines for of other policy instruments ensures their attain- ambient particulate concentration are those of ment. Consideration should therefore be given WHO, the EU, and the USEPA. These guidelines to the technical feasibility and the costs of at- are based on clinical, toxicological, and epidemio- tainment. logic evidence and were established by determin- Another difficulty is that airborne particulates ing the concentrations with the lowest observed are rarely homogeneous: They vary greatly in size adverse effect (implicitly accepting the notion and shape, and their chemical composition is that a lower threshold exists under which no ad- determined by factors specific to the source and verse human health effects can be detected), ad- location of the emissions. The combined effects justed by an arbitrary margin of safety factor and interactions of various substances mixed to allow for uncertainties in extrapolation from with particulates have not yet been established animals to humans and from small groups of hu- (except for sulfur dioxide), but they are believed mans to larger populations.1 The WHO guide- to be significant, especially where long-term ex- lines are based on health considerations alone; posure occurs. Measurement techniques and their Airborne Particulate Matter 205 reliability may vary across regions and countries, appropriate to establish area-specific ambient and so may other factors, such as diet, lifestyle, standards case by case. and physical fitness, that influence the human Prior to carrying out an environmental assess- health effects of exposure to particulates. ment, a trigger value for annual average concen-

trations of PM10 should be agreed on by the country Recommendations and the World Bank. Countries may wish to adopt EU, USEPA, or WHO guidelines or standards as In the long term, countries should seek to ensure their trigger values. The trigger value should be that ambient exposure to particulates, especially equal to or lower than the country’s ambient stan- to PM10, do not exceed the WHO recommended dard. The trigger value is not an ambient air qual- guidelines (see Table 1). In the interim, countries ity standard but simply a threshold. If, as a result should set ambient standards for total particu- of the project, the trigger value is predicted to be lates, PM , or both that take into account (a) the 10 exceeded in the area affected by the project, the benefits to human health of reducing exposure EA assessment should seek mitigation alternatives to particulates; (b) the concentration levels achievable by pollution prevention and control on a regional or sectoral basis. In the absence of an measures; and (c) the costs involved in meeting agreed value, the World Bank Group will classify the standards. In adopting new ambient air qual- an airshed as moderately degraded if the annual ity standards, countries should set appropriate average concentration levels of particulates are phase-in periods during which districts or mu- above 50 µg/m3 or if the 98th percentile of 24-hour nicipalities that do not meet the new standards mean values over a one-year period is estimated 3 are expected to come into compliance and will to exceed 150 µg/m of PM10. In areas where PM10 be assisted to attain the standards. Where there measurements do not exist, the value of 80 µg/m3 are large differences between the costs and ben- for TSP will be used. Airsheds will be classified as hav- efits of meeting air quality standards, it may be ing poor air quality with respect to particulate mat-

Table 1. Reference Standards and Guidelines for Average Ambient Particulate Concentration (micrograms per cubic meter) Long-term (annual) Short-term (24 hours)

Standard or guideline PM10 BS TSP PM10 BS TSP

EU limit values 80a 150b 250c 300d EU guide values 40–60a 100–150e USEPA primary and secondary standards 50f 150g WHO guidelinesh 40–60 60–90 100–150 150–230 WHO guidelines for Europeg 50 70I 125 120

Notes: PM10, particulate matter less than 10 microns in aerodynamic diameter; BS, black smoke (converted to µg/m3 measure); TSP, total suspended particulates. a. Median of daily mean values. b. Arithmetic mean of daily mean values. c. 98th percentile of all daily mean values throughout the year. d. 95th percentile of all daily mean values throughout the year. e. Daily mean values. f. Arithmetic mean. g. Guideline values for combined exposure to sulfur dioxide and particulates. h. Not to be exceeded for more than one day per year. i. Guideline for thoracic particles. According to International Organization for Standardization standard ISO-TP, thoracic particle measurements are roughly equivalent to the sampling characteristics for particulate matter with a 50% cutoff point at 10 microns diameter. Values are to be regarded as tentative at this time, being based on a single study that also involved sulfur dioxide exposure. Sources: European Community 1992 (EU); United States, CFR (USEPA); WHO 1979 (WHO guidelines); WHO 1987 (WHO guide- lines for Europe). 206 PROJECT GUIDELINES: POLLUTANTS

ter if the annual mean value of PM10 is greater than Pope, C. A., M. Thun, M. Namboodiri, D. Dockery, 100 µg/m3 or if the 95th percentile of 24-hour mean J. Evans, F. Speizer, and C. Heath. 1995. “Particu- values of PM over a period of one year is esti- late Air Pollution as a Predictor of Mortality in a 10 Prospective Study of U.S. Adults.” American Jour- mated to exceed 150 µg/m3. nal of Respiratory and Critical Care Medicine 151: Good practice in airshed management should 669–74. encompass the establishment of an emergency response plan during industrial plant operation. Schwartz, Joel. 1991/92. “Particulate Air Pollution and It is recommended that this plan be put into ef- Daily Mortality: A Synthesis.” Public Health Reviews 19: 39–60. fect when levels of air pollution exceed one or more of the emergency trigger values (deter- ————. 1993a. “Particulate Air Pollution and Chronic mined for short-term concentrations of sulfur Respiratory Disease.” Environmental Research 62: 7–13. dioxide, nitrogen oxides, particulates, and ————. 1993b. “Air Pollution and Daily Mortality ozone). The recommended emergency trigger in Birmingham, Alabama.” American Journal of Epi- 3 value for PM10 is 150 µg/m for 24-hour average demiology 137(10). concentrations. Where PM10 measurements do not Schwartz, Joel, and D. W. Dockery. 1992a. “Increased exist, the value of 300 µg/m3 for TSP is recom- Mortality in Philadelphia Associated with Daily Air mended. Pollution Concentrations.” American Review of Res- piratory Disease 145: 600–604. Note ————. 1992b. “Particulate Air Pollution and Daily Mortality in Steubenville, Ohio.” American Review 1. Adverse effect is defined as “any effect result- of Respiratory Disease 135(1). ing in functional impairment and/or pathological lesions that may affect the performance of the whole Schwartz, Joel, et al. 1993. “Particulate Air Pollution and Hospital Emergency Room Visits for Asthma organism or which contributed to a reduced ability in Seattle.” American Review of Respiratory Disease to respond to an additional challenge” (see USEPA 147: 826–31. 1980). Stern, Arthur C., et al. 1984. Fundamentals of Air Pollu- References and Sources tion. Orlando, Fla.: Academic Press. Ulm, K. 1994. “Epidemiology of Chronic Dust Expo- Dockery, D., C. A. Pope, X. Xiping, J. Spengler, J. Ware, sure.” In D. L. Dungworth, J. C. Mauderly, and G. M. Fay, B. Ferris, and F. Speizer. 1993. “An Associa- Oberdorfer, eds., Toxic Carcinogenic Effects of Solid tion between Air Pollution and Mortality in Six U.S. Particles in the Respiratory Tract. Washington, D.C: Cities.” New England Journal of Medicine 329(24): International Life Sciences Institute. 1753–59. United States. CFR (Code of Federal Regulations). Wash- European Community. 1992. European Community ington, D.C.: Government Printing Office. Deskbook. Washington, D.C.: Environmental Law Institute. USEPA (United States Environmental Protection Agency). 1980. “Guidelines and Methodology Used Kane, A. B. 1994. “Particle- and Fiber-induced Lesions: in the Preparation of Health Effect Assessment An Overview.” In D. L. Dungworth, J. C. Mauderly, Chapters of the Consent Decree Water Quality Cri- and G. Oberdorfer, eds., Toxic and Carcinogenic Ef- teria.” Federal Register 45: 79347–57. fects of Solid Particles in the Respiratory Tract.. Wash- ington, D.C: International Life Sciences Institute. ————. 1982a. Air Quality Criteria for Particulate Mat- ter and Sulfur Oxides. Vol. I. EPA-600/8-82-029a. Lipfert, F. W. 1984. “Air Pollution and Mortality: Research Triangle Park, N.C. Specification Searches Using SMSA-Based Data.” Journal of Environmental Economics and Management ————. 1982b. Second Addendum to Air Quality Crite- 11: 208–43. ria for Particulate Matter and Sulfur Oxides (1982): Assessment of Newly Available Health Effects Informa- Ostro, Bart. 1994. “Estimating the Health Effects of Air tion. Research Triangle Park, N.C. Pollutants: A Method with an Application to Jakarta.” Policy Research Working Paper 1301. ————. 1990. Review of the National Ambient Air Quality World Bank, Policy Research Department, Washing- Standard for Particulate Matter: Assessment of Scientific and ton, D.C. Technical Information. Research Triangle Park, N.C. Airborne Particulate Matter 207

WHO (World Health Organization). 1979. “Sulfur WHO (World Health Organization) and UNEP Oxides and Suspended Particulate Matter.” Envi- (United Nations Environment Programme). 1992. ronmental Health Criteria 8. Geneva. Urban Air Pollution in Megacities of the World. Cam- bridge, Mass.: Blackwell Reference. ————. 1987. Air Quality Guidelines for Europe. Copenhagen: WHO Regional Office for Europe.