1
Executive Summary
1.1 Introduction
Air Quality Criteria for Ozone and Related Photochemical Oxidants evaluates the
latest scientific information useful in deriving criteria that form the scientific basis for U.S.
Environmental Protection Agency (EPA) decisions regarding the National Ambient Air Quality
Standards (NAAQS) for ozone (O3). This Executive Summary concisely summarizes key
conclusions from the document, which comprises nine chapters. Following this Executive
Summary is a brief Introduction (Chapter 2) containing information on the legislative and
regulatory background for review of the O3 NAAQS, as well as a brief discussion of the issues
presented and the format for their discussion in the document. Chapter 3 provides information
on the chemistry, sources, emissions, measurement, and transport of O3 and related
photochemical oxidants and their precursors, whereas Chapter 4 covers environmental
concentrations, patterns, and exposure estimates of O3 and oxidants. Chapter 5 deals with
environmental effects, and Chapters 6, 7, and 8 discuss animal toxicological studies, human
health effects, and extrapolation of animal toxicological data to humans, respectively. The last
chapter, Chapter 9, provides an integrative, interpretative characterization of health effects
associated with exposure to O3. The following sections conform to the chapter organization of
the criteria document.
1.2 Legislative and Regulatory Background
The photochemical oxidants found in ambient air in the highest concentrations are
O3 and nitrogen dioxide (NO2). Other oxidants, such as hydrogen peroxide (H2O2) and
peroxyacyl nitrates, also have been observed, but in lower and less certain concentrations. In
1971, EPA promulgated NAAQS to protect the public health and welfare from adverse effects
of photochemical oxidants, at that time, defined on the basis of commercially available
measurement methodology. After 1971, however, O3-specific commercial analytical methods
became available, as did information on the concentrations and effects of the related non-O3 photochemical oxidants. As a result, the chemical designation of the standards was
changed in 1979 from photochemical oxidants to O3.
The EPA is required under Sections 108 and 109 of the Clean Air Act to evaluate
periodically the air quality criteria that reflect the latest scientific information relevant to
review of the O3 NAAQS. These air quality criteria are useful for indicating the kind and
extent of all identifiable effects on public health or welfare that may be expected from the
presence of O3 and related photochemical oxidants in ambient air. The last O3 criteria
document was released in 1986, and a supplement was released in 1992. These documents
were the basis for a March 1993 decision by EPA that revision of the existing 1-h NAAQS for
O3 was not appropriate at that time. That decision, however, did not take into consideration
more recent scientific information that has been published since the last literature review in
early 1989. The purpose of this revised criteria document, therefore, is to summarize the
pertinent information contained in the previous O3 criteria document and to critically evaluate
and assess the more recent scientific data associated with exposure to O3 and, to a lesser
extent, to H2O2 and the peroxyacyl nitrates, particularly peroxyacetyl nitrate (PAN). This
document will be used by EPA's Office of Air Quality Planning and Standards to provide a
staff paper assessing the most significant scientific information and presenting staff
recommendations on whether revisions to the O3 NAAQS are appropriate.
1.3 Tropospheric Ozone and Its Precursors
Introduction
Ozone is found in the stratosphere, the "free" troposphere, and the planetary
boundary layer (PBL) of the earth's atmosphere. In the PBL, background O3 occurs as the
result of (1) the intrusions of stratospheric O3 into the "free" troposphere and downward
transport into the PBL, and (2) photochemical reactions of methane (CH4), carbon monoxide
(CO), and nitrogen oxides (NOx). These processes contribute to the background O3 near the
surface. The major source of O3 in the PBL is the photochemical process involving
anthropogenic and biogenic emissions of NOx with the many classes of volatile organic
compounds (VOCs).
The topics considered in this section of the document include: tropospheric
O3 chemistry; meteorological influences on O3 formation and transport; precursor VOC and
NOx emissions, ambient concentrations of VOCs and NOx, and source apportionment and
reconciliation of measured VOC ambient concentrations with emission inventories; O3 air
quality models; and analytical methods for oxidants and precursors.
Tropospheric Ozone Chemistry
Ozone occurs in the stratosphere as the result of chemical reactions initiated by
short-wavelength radiation from the sun. In the "free" troposphere, O3 occurs as the result of
incursions from the stratosphere; upward venting from the PBL (the layer next to the surface
of the earth) through certain cloud processes; and photochemical formation from precursors,
notably CH4, CO, and NOx.
The photochemical production of O3 and other oxidants found at the surface of the
earth (in the PBL, troposphere, or ambient air [used interchangeably in this summary]) is the
result of atmospheric physical processes and complex, nonlinear chemical processes involving
two classes of precursor pollutants: (1) reactive anthropogenic and biogenic VOCs and
(2) NOx. The only significant initiator of the photochemical production of O3 in the polluted
troposphere is the photolysis of NO2, yielding nitric oxide (NO) and a ground-state oxygen
atom that reacts with molecular oxygen to form O3. The O3 thus formed reacts with NO,
yielding oxygen and NO2. These cyclic reactions attain equilibrium in the absence of VOCs.
However, in the presence of VOCs, which are abundant in polluted ambient air, the
equilibrium is upset, resulting in a net increase in O3. Methane is the chief VOC found in the
free troposphere and in most "clean" areas of the PBL. The VOCs found in polluted ambient
air are much more complex and more reactive than CH4, but, as with CH4, their atmospheric
oxidative degradation is initiated through attack on the VOCs by hydroxyl (OH) radicals. As
in the CH4 oxidation cycle, the conversion of NO to NO2 during the oxidation of VOCs is
accompanied by the production of O3 and the efficient regeneration of the OH radical. The
O3, PAN, and higher homologues formed in polluted atmospheres increase with the NO2/NO
concentration ratio.
At night, in the absence of photolysis of reactants, the simultaneous presence of
O3 and NO2 results in the formation of the nitrate (NO3) radical. Reactions with NO3 radicals
appear to constitute major sinks for alkenes, cresols, and several other compounds, although
the chemistry is not well characterized.
Most inorganic gas-phase processes (i.e., the nitrogen cycle and its
interrelationships with O3 production) are well understood. The chemistry of the VOCs in
ambient air is not as well understood. It is well known, however, that the chemical loss
processes of gas-phase VOCs include reaction with OH and NO3 radicals and O3, and
photolysis. Reaction with the OH radical is the only important atmospheric reaction (loss
process) for alkanes, aromatic hydrocarbons, and the higher aldehydes and ketones that lack
>C=C< bonds; and the only atmospheric reaction of alcohols and ethers. Photolysis is the
major loss process for formaldehyde and acetone. Reactions with OH and NO3 radicals and
with O3 are all important loss processes for alkenes and for carbonyls containing >C=C<
bonds.
Uncertainties in the atmospheric chemistry of the VOCs can affect quantification of
the NO-to-NO2 conversion and of O3 yields, and can present difficulties in representation of
chemical mechanisms, products, and product yields in O3 air quality models. Major
uncertainties in understanding the atmospheric chemistry of the VOCs with NOx in both urban
and rural atmospheres include chemistry of alkyl nitrate formation, mechanisms and products
of >C4 n-alkanes and branched alkanes, mechanisms and products of alkene-O3 reactions, and
mechanisms and products of aromatic hydrocarbons.
It should be noted that the atmospheric chemical processes involved in the
photooxidation of certain higher molecular weight VOCs and in the formation of O3 also can
lead to the formation of particulate-phase organic compounds. The OH radicals produced not
only can oxidize VOCs to particulate-phase organic compounds but also can react with NO2
and sulfur dioxide (SO2) to form nitric acid (HNO3) and sulfuric acid (H2SO4), respectively,
portions of which become incorporated into aerosols as particulate nitrate and sulfate.
Meteorological Influences on Ozone Formation and Transport
The surface energy (radiation) budget of the earth strongly influences the dynamics
of the PBL. The redistribution of energy through the PBL creates thermodynamic conditions
that influence vertical mixing. Growing evidence indicates that the strict use of mixing heights
in modeling is an oversimplification of the complex processes by which pollutants are
redistributed within urban areas, and that it is necessary to treat the turbulent structure of the
atmosphere directly and acknowledge the vertical variations in mixing. Energy balances
therefore require study so that more realistic simulations can be made of the structure of the
PBL.
Day-to-day variability in O3 concentrations depends heavily on day-to-day
variations in meteorological conditions, including temperature, solar radiation, and the degree
of mixing that occurs between release of a pollutant or its precursors and their arrival at a
receptor; the occurrence of inversion layers (layers in which temperature increases with height
above ground level); and the transport of O3 left overnight in layers aloft and subsequent
downward mixing of that O3 to the surface.
The transport of O3 and its precursors beyond the urban scale (ó50 km) to
neighboring rural and urban areas has been well documented. Episodes of high
O3 concentrations in urban areas are often associated with high concentrations of O3 in the
surroundings. Areas of O3 accumulation usually are characterized by synoptic-scale
subsidence of air in the free troposphere, resulting in development of an elevated inversion
layer; relatively low wind speeds associated with the weak horizontal pressure gradient around
a surface high pressure system; a lack of cloudiness; and high temperatures.
Ultraviolet (UV) radiation from the sun plays a key role in initiating the
photochemical processes leading to O3 formation and affects individual photolytic reaction
steps. Still, there is little empirical evidence in the literature linking day-to-day variations in
observed UV radiation levels to variations in O3 levels. An association, however, between
tropospheric O3 concentrations and temperature has been demonstrated. Empirical data from
four urban areas, for example, show an apparent upper bound on O3 concentrations that
increases with temperature. A similar qualitative relationship exists at a number of rural
locations.
The relationship between wind speed and O3 buildup varies from one part of the
country to another.
Statistical techniques (e.g., regression techniques) can be used to help identify real
trends in O3 concentrations, both intra- and interannual, by normalizing meteorological
variability.
Precursors
Volatile Organic Compound Emissions
Hundreds of VOCs, usually containing from 2 to 12 carbon atoms, are emitted by
evaporative and combustion processes from a large number of source types. Total U.S.
anthropogenic VOC emissions in 1991 were estimated at 21.0 Tg; the two largest source
categories were (1) industrial processes (10.0 Tg) and (2) transportation (7.9 Tg). Emissions
of VOCs from highway vehicles accounted for almost 75% of the transportation-related
emissions; studies have shown that the majority of these VOC emissions come from about
20% of the automobiles in service, many, perhaps most, of which are older cars that are
poorly maintained. The accuracy of VOC emission estimates is difficult to determine for both
stationary and mobile sources.
Vegetation emits significant quantities of VOCs into the atmosphere, chiefly
monoterpenes and isoprene, but also oxygenated VOCs, according to recent studies. The most
recent biogenic VOC emissions estimate for the United States showed annual emissions of
29.1 Tg/year.
Although the biogenic VOC emission estimates exceed the anthropogenic estimates,
the biogenic emissions are more diffusely distributed than the anthropogenic emissions, which
tend to be concentrated in population centers. However, the large uncertainties in both
biogenic and anthropogenic VOC emission inventories prevent establishing the relative
contributions of these two categories.
Nitrogen Oxides Emissions
Anthropogenic NOx is associated with combustion processes. The primary
pollutant emitted is NO, formed at high combustion temperatures from nitrogen and oxygen in
the air and from nitrogen in the combustion fuel. Emissions of NOx in 1991 in the United
States totaled 21.39 Tg. The two largest single NOx emission sources are electric power
generating plants and highway vehicles. Because a large proportion of anthropogenic NOx
emissions come from distinct point sources, published annual estimates are thought to be much
more reliable than VOC estimates.
Natural NOx sources include stratospheric intrusion, oceans, lightning, soil, and
wildfires. Lightning and soil emissions are the only two significant natural sources of NOx in
the United States. It is estimated that combined natural sources contribute about 2.2 Tg of
NOx to the troposphere over the continental United States; however, uncertainties in natural
NOx emission inventories are much greater than those for anthropogenic NOx emissions.
Concentrations of Volatile Organic Compounds in Ambient Air
The VOCs most frequently analyzed in ambient air are the nonmethane
hydrocarbons (NMHCs). Morning (6:00 to 9:00 a.m.) concentrations most often have been
measured because of the use of morning data in the Empirical Kinetic Modeling Approach
(EKMA) and in air quality simulation models.
Concurrent measurements of anthropogenic and biogenic NMHCs have shown that
biogenic NMHCs usually constituted much less than 10% of the total NMHCs. For example,
average isoprene concentrations ranged from 0.001 to 0.020 ppm carbon (C) and terpenes
from 0.001 to 0.030 ppm C.
Concentrations of Nitrogen Oxides in Ambient Air
Measurements of NOx made in 22 and 19 U.S. cities in 1984 and 1985,
respectively, showed median 6:00-to-9:00 a.m. NOx concentrations ranging from 0.02 to 0.08
ppm in most of these cities. Nonurban NOx concentrations, reported as average seasonal or
annual NOx, range from <0.005 to 0.015 ppm.
Ratios of Concentrations of Nonmethane Organic Compounds to Nitrogen Oxides
Ratios of 6:00-to-9:00 a.m. nonmethane organic compounds (NMOC) to NOx are
higher in southeastern and southwestern U.S. cities than in northeastern and midwestern U.S.
cities, according to data from EPA's multicity studies conducted in 1984 and 1985. Rural
NMOC/NOx ratios tend to be higher than urban ratios. The NMOC/NOx ratios trended
downward to well below 10 in the South Coast Air Basin and in cities in the eastern United
States during the 1980s. Based on these low ratios, hydrocarbon control should be more
effective than NOx control within a number of cities. Morning (6:00-to-9:00 a.m.)
NMOC/NOx ratios are used in the EKMA type of trajectory model. The correlation of
NMOC/NOx ratios with maximum 1-h O3 concentrations, however, was weak in a recent
analysis.
Source Apportionment and Reconciliation
Source apportionment (regarded as synonymous with receptor modeling) refers to
determining the quantitative contributions of various sources of VOCs to ambient air pollutant
concentrations. Source reconciliation refers to the comparison of measured ambient VOC
concentrations with emissions inventory estimates of VOC source emission rates for the
purpose of validating the inventories.
Recent findings have shown that vehicle exhaust was the dominant contributor to
ambient VOCs in seven of eight U.S. cities studied. Whole gasoline contributions were
estimated to be equal to vehicle exhaust in one study and to 20% of vehicle exhaust in a
second study.
Estimates of biogenic VOCs at a downtown site in Atlanta, GA, in 1990 indicated a
lower limit of 2% (24-h average) for the biogenic percentage of total ambient VOCs at that
location (isoprene was used as the biogenic indicator species). The percentage varies during
the 24-h period because of the diurnal (e.g., temperature, light intensity) dependence of
isoprene concentrations.
Source reconciliation data have shown disparities between emission inventory
estimates and receptor-estimated contributions. For biogenics, emission estimates are greater
than receptor-estimated contributions. The reverse has been true for natural gas contributions
estimated for Los Angeles, CA; Columbus, OH; and Atlanta; and for refinery emissions in
Chicago, IL.
Ozone Air Quality Models
Models and Their Components
Photochemical air quality models are used to predict how O3 concentrations change
in response to prescribed changes in source emissions of NOx and VOCs. These models
operate on sets of input data that characterize the emissions, topography, and meteorology of a
region and produce outputs that describe air quality in that region.
Two kinds of photochemical models are recommended in guidelines issued by EPA:
(1) the use of EKMA is accepted under certain circumstances, and (2) the grid-based Urban
Airshed Model (UAM) is recommended for modeling O3 over urban areas. The 1990 Clean
Air Act Amendments mandate the use of three-dimensional (grid-based) air quality models
such as UAM in developing state implementation plans for areas designated as "extreme",
"severe", "serious", or "multistate moderate". General descriptions of EKMA and grid-based
models were given in the 1986 EPA criteria document for O3.
The EKMA-based method for determining O3 control strategies has limitations, the
most serious of which is that predicted emissions reductions are critically dependent on the
initial NMHC/NOx ratio used in the calculations. This ratio cannot be determined with any
certainty and is expected to be quite variable in time and space in an urban area.
Spatial and temporal characteristics of VOC and NOx emissions are major inputs to
a grid-based photochemical air quality model. Greater accuracy in emissions inventories is
needed for biogenics and for both mobile and stationary source components. Grid-based air
quality models also require as input the three-dimensional wind field for the photochemical
episode being simulated.
A chemical kinetic mechanism, representing the important chemical reactions that
occur in the atmosphere, is used in an air quality model to estimate the net rate of formation of
each pollutant simulated as a function of time.
Dry deposition is an important removal process for O3 on both urban and regional
scales and is included in all urban- and regional-scale models. Wet deposition is generally not
included in urban-scale photochemical models, because O3 episodes do not occur during
periods of significant clouds or rain.
Concentration fields of all species computed by the model must be specified at the
beginning of the simulation ("initial conditions"). These initial conditions are determined
mainly with ambient measurements, either from routinely collected data or from special
studies; but interpolation can be used to distribute the surface ambient measurements.
Use of Ozone Air Quality Models
Photochemical air quality models are used for control strategy evaluation by first
demonstrating that a past episode or episodes can be simulated adequately. The hydrocarbon
or NOx emissions or both are reduced in the model inputs, and the effects of these reductions
on O3 in the region are assessed. The adequacy of control strategies based on grid-based
models depends, in part, on the nature of input data for simulations and model validation, on
input emissions inventory data, and on the mismatch between the spacial output of the model
and the current form of the NAAQS for O3. Uncertainties in models obviously can affect their
outputs. Uncertainties exist in all components of grid-based O3 air quality models: emissions,
meteorological modules, chemical mechanisms, deposition rates, and determination of initial
conditions.
Grid-based models that have been widely used to evaluate control strategies for
O3 or acid deposition, or both, are the UAM, the California Institute of Technology/Carnegie
Institute of Technology model, the Regional Oxidant Model, the Acid Deposition and Oxidant
Model, and the Regional Acid Deposition Model. The UAM (Version IV) is the grid model
approved nationwide for control strategy development at this time.
Despite the many uncertainties in photochemical air quality modeling, including
emission inventories, these models are essential for regulatory analysis and solving the
O3 problem. Grid-based O3 air quality modeling is superior to the available alternatives for
O3 control planning, but the chances of its incorrect use must be minimized.
Analytical Methods for Oxidants and Their Precursors
Oxidants
Current methods used to measure O3 are chemiluminescence (CL); UV absorption
spectrometry; and newly developed spectroscopic and chemical approaches, including
chemical approaches applied to passive sampling devices (PSDs) for O3.
The CL method has been designated as the reference method by EPA. Detection
limits of 0.005 ppm and a response time of <30 s are typical of currently available
commercial instruments. A positive interference from atmospheric water vapor was reported
in the 1970s and recently has been confirmed. Proper calibration can minimize this source of
error.
Commercial UV photometers for measuring O3 have detection limits of about
0.005 ppm and a response time of <1 min. Because the measurement is absolute, UV
photometry is also used to calibrate O3 methods. A potential disadvantage of UV photometry
is that atmospheric constituents that absorb 254-nm radiation, the wavelength at which O3 is
measured, will cause a positive interference in O3 measurements. Interferences have been
reported in two recent studies, but assessment of the potential importance of such interferences
(e.g., toluene, styrene, cresols, nitrocresols) is hindered by lack of absorption spectra data in
the 250-nm range and by lack of aerometric data for the potentially interfering species. There
also can be some interference from water, possibly from the condensation of moisture in
sampling lines.
Calibration of O3 measurement methods (other than PSDs) is done by UV
spectrometry or by gas-phase titration (GPT) of O3 with NO. Ultraviolet photometry is the
reference calibration method approved by EPA. Ozone is unstable and must be generated in
situ at time of use to produce calibration mixtures.
Peroxyacetyl nitrate and the higher peroxyacyl nitrates normally are measured by
gas chromatography (GC) using an electron capture detector. Detection limits have been
extended to 1 to 5 ppt. The preparation of reliable calibration standards is difficult because
PAN is unstable, but several methods are available.
Volatile Organic Compounds
The method recommended by EPA for total NMOC measurement involves the
cryogenic preconcentration of NMOCs and the measurement of the revolatilized NMOCs using
flame ionization detection (FID). The primary technique for speciated NMOC/NMHC
measurements is cryogenic preconcentration followed by GC-FID. Systems for sampling and
analysis of VOCs have been developed that require no liquid cryogen for operation.
Stainless steel canisters have become the containers of choice for collection of
whole-air samples for NMHC/NMOC data. Calibration procedures for NMOC
instrumentation require the generation, by static or dynamic systems, of dilute mixtures at
concentrations expected to occur in ambient air.
Preferred methods for measuring carbonyl species (aldehydes and ketones) in
ambient air are spectroscopic methods; on-line colorimetric methods; and, the most common
method currently in use for measuring gas-phase carbonyl compounds in ambient air, the high-performance liquid chromatography method, which employs 2,4-dinitrophenylhydrazine
derivatization in a silica gel cartridge. Use of an O3 scrubber has been recommended to
prevent interference by O3 in this method in ambient air.
Oxides of Nitrogen
Nitric oxide and NO2 comprise the NOx compounds involved as precursors to
O3 and other photochemical oxidants.
The most common method of NO measurement is the gas-phase CL reaction with
O3, which is essentially specific for NO. Commercial NO monitors have detection limits of a
few parts per billion by volume (ppbv) in ambient air but may not have sensitivity sufficient
for surface measurements in rural or remote areas or for airborne measurements. Direct
spectroscopic methods for NO exist that have very high sensitivity and selectivity for NO, but
their complexity, size, and cost restrict these methods to research applications. No PSDs exist
for measurement of NO.
Chemiluminescence analyzers are the tools of choice for NO2 measurement, even
though they do not measure NO2 directly. Minimum detection levels for NO2 have been
reported to be 5 to 13 ppb, but more recent evaluations have indicated detection limits of
0.5 to 1 ppbv. Reduction of NO2 to NO is required for measurement. These analyzers
actually measure NOy (NOx + PAN + HNO3 + other reactive nitrogen species); however, for
most urban atmospheres, NOx is the predominant species measured diurnally.
Several spectroscopic approaches to NO2 detection have been developed but share
the drawbacks of spectroscopic NO methods. Passive samplers for NO2 exist but are still in
the developmental stage for ambient air monitoring.
Calibration of methods for NO measurement is done using standard cylinders of
NO in nitrogen. Calibration of methods for NO2 measurement include use of cylinders of NO2
in nitrogen or air, use of permeation tubes, and GPT.
1.4 Environmental Concentrations, Patterns, and Exposure
Estimates
Ozone is measured at concentrations above the minimum detectable level at all
monitoring locations in the world. In this section, hourly average concentration and exposure
information is summarized for urban, rural forested, and rural agricultural areas in the United
States.
Because O3 from urban area emissions is transported to rural downwind locations,
elevated O3 concentrations can occur at considerable distances from urban centers. Urban
O3 concentration values are often depressed because of titration by NO. Because of the
absence of chemical scavenging, O3 tends to persist longer in nonurban areas than in urban
areas, and nonurban exposures may be higher than those in urban locations.
Trends
Ozone hourly average concentrations have been recorded for many years by the
state and local air pollution agencies who report their data to EPA. The 10-year (1983 to
1992) composite average trend for the second highest daily maximum hourly average
concentration during the O3 season shows that the 1992 composite average for the trend sites
was 21% lower than the 1983 average. The 1992 value was the lowest composite average of
the 10-year period and was significantly less than each of the previous nine years, 1983 to
1991. The relatively high O3 concentrations in 1983 and 1988 likely were attributable, in part,
to hot, dry, stagnant conditions in some areas of the country, which were especially conducive
to O3 formation.
From 1991 to 1992, the composite mean of the second highest daily maximum
1-h O3 concentrations decreased 7%, and the composite average of the number of estimated
exceedances of the O3 standard decreased by 23%. Nationwide VOC emissions decreased 3%
from 1991 to 1992. The composite average of the second daily maximum concentrations
decreased in 8 of the 10 EPA regions from 1991 to 1992, and remained unchanged in Region
VII. Except for Region VII, the 1992 regional composite means were lower than the
corresponding 1990 levels. Although meteorological conditions in the east during 1993 were
more conducive to O3 than those in 1992, the composite mean level for 1993 was the second
lowest composite average of the decade, 1984 to 1993.
Surface Concentrations
Published data provide evidence showing the occurrence at some sites of multihour
periods within a day of O3 at levels of potential health effects. Although most of these
analyses were made using monitoring data collected from sites in or near nonattainment areas,
in one analysis of five sites (two in New York state, two in rural California, and one in rural
Oklahoma), none of which was in or near a nonattainment area, O3 concentrations showed
only moderate peaks but showed multihour levels above 0.1 ppm.
A small amount of the O3 concentration measured at a monitoring site is produced
by sources distant to the photochemical reactions occurring on an urban or regional scale.
Typical sources include stratospheric intrusions into the troposphere, photochemical
production by the CH4/CO/NOx cycle in the troposphere, and transport of very distant
anthropogenic or biogenic VOCs and NOx. The specific concentrations of this "background"
O3 vary with averaging times ranging from the daily 1-h maximum to daily, monthly,
seasonal, or annual values. The background concentrations also vary with geographical region
and with elevation of the monitoring site.
On the basis of O3 data from isolated monitoring sites, EPA has indicated that a
reasonable estimate of O3 background concentration near sea level in the United States is from
0.020 to 0.035 ppm for an annual average, 0.025 to 0.045 ppm for an 8-h daily summer
seasonal average, and from 0.03 to 0.05 ppm for the average summertime 1-h daily maximum.
This estimate includes a 0.005 to 0.015 ppm O3 contribution from stratospheric intrusions into
the troposphere.
Diurnal Variations
Diurnal patterns of O3 may be expected to vary with location, depending on the
balance among the many factors affecting O3 formation, transport, and destruction. Although
they vary with locality, diurnal patterns of O3 typically show a rise in concentration from low
levels, or levels near minimum detectable amounts, to an early afternoon peak. The diurnal
pattern of concentrations can be ascribed to three simultaneous processes: (1) downward
transport of O3 from layers aloft, (2) destruction of O3 through contact with surfaces and
through reaction with NO at ground level, and (3) in situ photochemical production of O3.
Seasonal Patterns
Seasonal variations in O3 concentrations in urban areas usually show the pattern of
high O3 in the late spring or in the summer and low levels in the winter; however, weather
conditions in a given year may be more favorable for the formation of O3 and other oxidants
than during the prior or following year.
Average O3 concentrations tend to be higher in the second versus the third quarter
of the year for many isolated rural sites. This observation has been attributed to either
stratospheric intrusions or an increasing frequency of slow-moving, high-pressure systems that
promote the formation of O3. However, for several clean rural sites, the highest exposures
have occurred in the third quarter rather than in the second. For rural O3 sites in the
southeastern United States, the daily maximum 1-h average concentration was found to peak
during the summer months.
Spatial Variations
Concentrations of O3 vary with altitude and with latitude. There appears to be no
consistent conclusion concerning the relationship between O3 exposure and elevation.
Indoor Ozone
Until the early 1970s, very little was known about the O3 concentrations
experienced inside buildings; to date, the database on this subject is not large, and a wide
range of indoor/outdoor O3 concentration relationships can be found in the literature (reported
indoor/outdoor values for O3 are highly variable). Indoor/outdoor O3 concentration ratios
generally fall in the range from 0.1 to 0.7 and indoor concentrations of O3 almost invariably
will be less than outdoors.
Estimating Exposure
Both fixed-site monitoring information and human exposure models are used to
estimate risks associated with O3 exposure. Because, for most cases, it is not possible to
estimate population exposure solely from fixed-station data, several human exposure models
have been developed. These models also contain submodels depicting the sources and
concentrations likely to be found in each microenvironment, including indoor, outdoor, and in-transit settings. Two distinct types of O3 exposure models exist: (1) those that focus narrowly
on predicting indoor O3 levels and (2) those that focus on predicting O3 exposures on a
community-wide basis. These latter models and their distinguishing features are:
1. pNEM/O3 based on the National Air Quality Standards Exposure Model (NEM)
series of models
Uses mass-balance approach and seasonal considerations for I/O ratio
estimation.
Variables affecting indoor exposure obtained by Monte Carlo sampling from
empirical distributions of measured data.
2. Systems Applications International (SAI)/NEM
More districts and microenvironments and more detailed mass-balance
model than pNEM/O3.
Human activity data outdated and inflexible.
3. Regional Human Exposure Model (REHEX)
More detailed geographic resolution than NEM.
Uses California-specific activity data and emphasizes in-transit and outdoor
microenvironments.
4. Event probability exposure model (EPEM)
Estimates probability that a randomly selected person will experience a
particular exposure regime.
Lacks multiday continuity.
Few data are available for individuals using personal exposure monitors. Results
from a pilot study demonstrated that fixed-site ambient measurements may not adequately
represent individual exposures. Models based on time-weighted indoor and outdoor
concentrations explained only 40% of the variability in personal exposures.
Peroxyacyl Nitrates
Peroxyacetyl nitrate and peroxypropionyl nitrate (PPN) are the most abundant of
the non-O3 oxidants in ambient air in the United States, other than the inorganic nitrogenous
oxidants such as NO2, and possibly HNO3. Most of the available data on concentrations of
PAN and PPN in ambient air are from urban areas. The levels to be found in nonurban areas
will be highly dependent on the transport of PAN and PPN or their precursors from urban
areas, because the concentrations of the NOx precursors to these compounds are considerably
lower in nonurban areas than in urban areas.
Co-occurrence
Studies of the joint occurrence of gaseous NO2/O3 and SO2/O3 at rural sites have
concluded that the periods of co-occurrence represent a small portion of the potential
plant-growing period. For human ambient exposure considerations, in most cases, the
simultaneous co-occurrence of NO2/O3 and SO2/O3 was infrequent. Some researchers have
reported the joint occurrence of O3, nitrogen, and sulfur in forested areas, combining
cumulative exposures of O3 with data on dry deposition of sulfur and nitrogen. One study
reported that several forest landscapes with the highest dry deposition loadings of sulfur and
nitrogen tended to experience the highest average O3 concentrations and largest cumulative
exposure. Although the authors concluded that the joint concentrations of multiple pollutants
in forest landscapes were important, nothing was mentioned about the hourly co-occurrences
of O3 and SO2 or O3 and NO2. Acid sulfates, which are usually composed of H2SO4,
ammonium bisulfate, and ammonium sulfate, have been measured at a number of locations in
North America. The potential for O3 and acidic sulfate aerosols to co-occur at some locations
in some form (i.e., simultaneously, sequentially, or complex-sequentially) is real and requires
further characterization. For human ambient exposures, the simultaneous co-occurrence of
NO2 and O3 was infrequent.
In one study, the relationship between O3 and hydrogen ions in precipitation was
explored using data from sites that monitored both O3 and wet deposition simultaneously and
within one minute latitude and longitude of each other. It was reported that individual sites
experienced years in which both hydrogen ion deposition and total O3 exposure were at least
moderately high. With data compiled from all sites, it was found that relatively acidic
precipitation occurred together with relatively high O3 levels approximately 20% of the time,
and highly acidic precipitation occurred together with a high O3 level approximately 6% of the
time. Sites most subject to relatively high levels of both hydrogen ions and O3 were located in
the eastern part of the United States, often in mountainous areas.
The co-occurrence of O3 and acidic cloudwater in high-elevation forests has been
characterized. The frequent O3-only and pH-only single-pollutant episodes, as well as the
simultaneous and sequential co-occurrences of O3 and acidic cloudwater, have been reported.
Both simultaneous and sequential co-occurrences were observed a few times each month above
cloud base.
1.5 Environmental Effects of Ozone and Related
Photochemical Oxidants
Ozone is the gaseous pollutant most injurious to agricultural crops, trees, and native
vegetation. Exposure of vegetation to O3 can inhibit photosynthesis, alter carbon
(carbohydrate) allocation, and interfere with mycorrhizal formation in tree roots. Disruption of
the important physiological processes of photosynthesis and carbon allocation can suppress the
growth of crops, trees, shrubs, and herbaceous vegetation by decreasing their capacity to form
the carbon (energy) compounds needed for growth and maintenance and their ability to absorb
the water and mineral nutrients that they require from the soil. In addition, loss of vigor
impairs the ability of trees and crops to reproduce and increases their susceptibility to insects
and pathogens. The following section summarizes key environmental effects associated with
O3 exposure.
Effects on Agroecosystems
Methodologies Used in Vegetation Research
Most of the knowledge concerning the effects of O3 on vegetation comes from the
exposure-response studies of important agricultural crop plants and some selected forest and
urban tree species, mostly as seedlings. A variety of methodologies have been used, ranging
from field exposures without chambers to open-top chambers and to exposures conducted in
chambers under highly controlled conditions. In general, the more controlled conditions are
most appropriate for investigating specific responses and for providing the scientific basis for
interpreting and extrapolating results. The greatest body of knowledge is from OTC studies.
Mode of Action
Leaves are important regulators of plant stress and function. Stress resulting from
exposure to O3 produces a leaf-mediated response. Effects expressed within cells in the leaf
(i.e., inhibition of photosynthesis) affect a plant's carbon (energy) budget. Plant processes are
impaired only by the O3 that enters the plant through the stomata (opening in the leaves). An
effect will occur only if sufficient O3 reaches sensitive sites within the leaf cells. The uptake
and movement of O3 to sensitive cellular sites within a leaf are subject to various biochemical
and physiological controls. Leaf injury will not be detected if the rate of uptake is small
enough for the plant to detoxify or metabolize O3 and its derivatives, or the plant is able to
repair or compensate for the impact at a rate equal to or greater than the rate of uptake.
Impairment of leaf cellular processes is the basis for all other plant effects. The diurnal
pattern of stomatal opening plays a critical role in O3 uptake, particularly at the canopy level.
Visible injury is usually the first observable indication of cellular response; injury
can occur, however, with no visible effects. Early senescence of leaves or needles is also a
result of cellular response. Impairment of cellular processes inhibits the rate of
photosynthesis, reduces carbon (sugars, carbohydrate) production, and alters carbon
allocation, causing a shift in growth pattern that favors shoots over roots. The reduced
allocation of carbon to leaf repair and new leaf formation limits the availability of carbon for
reproduction; stem and root growth; and, particularly, the formation of the mycorrhizae on
roots necessary for nutrient and water uptake. Reduction of plant vigor by O3 can result in
mortality, particularly when plant susceptibility to insects and pathogens is increased.
Factors That Modify Plant Response
Plant response to O3 exposure is influenced by a variety of biological, chemical,
and physical factors. When determining the impact of O3 exposure on plants, both the
influence of environmental factors on plant response and the effects of O3 on that response
must be considered. Biological factors within plants that affect their response to stresses
include, genetic composition, stage of development, and the diurnal pattern of stomatal
opening. Genotype significantly influences plant sensitivity to O3. Individuals, varieties, and
cultivars of a species are known to differ greatly in their responses to a given O3 exposure.
Genotype also influences the ability of plants to compete with one another for space, nutrients,
light, and water.
The magnitude of response of a particular species, variety, or cultivar depends on a
number of environmental factors. The plant's present and past environmental milieu, which
includes the temporal exposure pattern and stage of development, dictates the plant response.
The corollary is also true: exposure to O3 can modify plant response to other environmental
variables. Available light, temperature, atmospheric turbulence and moisture, in both the
atmosphere and soil; soil nutrition; and exposure to and interaction with other pollutants such
as agricultural chemical sprays also influence the magnitude of plant response.
Drought can reduce visible injury and the adverse effects of O3 on growth and yield
of crops. However, in the case of crops, drought, per se, much more adversely affects yield
than the effects of O3. Ozone, on the other hand, tends to reduce the water-use efficiency of
well-watered crops. In some plants, O3 exposure reduces cold/winter hardiness. Although
exposure to O3 tends to reduce attacks by obligate pathogens, susceptibility of plants to
facultative pests and pathogens increases.
Effects-Based Air Quality Exposure Indices
Environmental scientists for many years have attempted to characterize and
mathematically represent plant exposures to O3. A variety of averaging times have been used.
Although most studies have characterized exposure by using mean concentrations, such as
seasonal, monthly, weekly, daily, or peak hourly means, other studies have used cumulative
measures (e.g., the number of hours above selected concentrations). None of these statistics
completely characterizes the relationships among O3 concentration, exposure duration, interval
between exposures, and plant response.
The use of a mean concentration with long averaging times implies that all
concentrations of O3 are equally effective in causing plant responses and minimizes the
contributions of the peak concentrations to the response. Ozone effects are cumulative;
therefore, exposure duration should be included in any index if it is to be biologically relevant.
Present evidence suggests that cumulative effects of episodic exposures to either peak or mid-range concentrations, or both, can play an important role in producing growth responses. The
key to plant response is timing because peak and mid-range concentrations do not occur at the
same time. Potentially, the greatest effect of O3 on plants will occur when stomatal
conductance is greatest. When peaks occur at the time of greatest stomatal conductance, the
effect of mid-range concentrations will not be observable. Atmospheric conductivity also
strongly influences plant response because O3 must be in contact with the leaf surface if it is to
be taken up by a plant. Effects on vegetation appear when the amount of pollutant entering
exceeds the ability of the plant to repair or compensate for the impact. Increasing uptake of O3
will inhibit photosynthesis and result in increased reductions in biomass production.
An index of ambient exposures that relates well to plant response should
incorporate, directly or indirectly, environmental influences (e.g., temperature, humidity, soil-moisture status) and exposure dynamics. Peak indices (e.g., second highest daily maximum)
imply that a single high-concentration exposure (1- or 8-h concentration) during the course of
a 70- to 120-day growing season is related to eventual yield or growth reductions. On the
other hand, mean indices (e.g., 7-h seasonal mean) imply that duration of the exposure is not
important, and that all concentrations have equal effect on plants. Neither of these indices
relates ambient O3 concentrations to biological effects on plants because these indices do not
consider the duration of exposure. An index that cumulates all hourly concentration during the
season and gives greater weight to higher concentrations appears to be a more appropriate
index for relating ambient exposures to growth or yield effects.
No experimental studies have been designed specifically to evaluate the adequacy of
the various peak-weighted indices that have been proposed. In retrospective analyses in which
O3 is the primary source of variation in response, year-to-year variations in plant response are
minimized by peak-weighted, cumulative exposure indices. However, a number of different
forms of peak-weighted, cumulative indices have been examined for their ability to properly
order yield responses from the large number of studies of the National Crop Loss Assessment
Network (NCLAN) program. These exposure indices (i.e., SUM00, SUM06, SIGMOID,
W126) all performed equally well, and it is not possible to distinguish among them on the
basis of statistical fits of the data. The biological basis for these indices has not been
determined.
Exposure Response of Plant Species
The emphasis of experimental studies usually has been on the more economically
important crop plants and tree species, as seedlings. Crop species usually are monocultures
that are fertilized and, in many cases, watered. Therefore, because crop plants are usually
grown under optimal conditions, their sensitivity to O3 exposures can vary from that of native
trees, shrubs, and herbaceous vegetation.
The concept of limiting values was used in both the 1978 and 1986 criteria
documents to summarize visible foliar injury. Limiting values are defined as concentrations
and durations of exposure below which visible injury does not occur. The limit for visible
injury indicating reduced plant performance was an O3 exposure of 0.05 ppm for several hours
per day for more than 16 days. When the exposure period was decreased to 10 days, the
O3 concentration required to cause injury was increased to 0.10 ppm. A short, 6-day exposure
further increased the concentration to 0.30 ppm. These exposure and concentration periods
apply for those crops where appearance or aesthetic value (e.g, spinach, cabbage, lettuce) is
considered important. Limiting values for foliar injury to trees and shrubs range from 0.06 to
0.10 ppm for 4 h.
The following assertions can be made based on information from the 1986 criteria
document, its 1992 supplement, and literature published since 1986. Ambient
O3 concentrations in several regions of the country are high enough to impair growth and yield
of sensitive plant species. This clearly is indicated by comparison of data obtained from crop
yield in charcoal-filtered and unfiltered (ambient) exposures. These elevated levels are further
supported by data from studies using chemical protectants. These response data make possible
the extrapolation to plants not studied experimentally. Both approaches mentioned above
indicate that effects occur with only a few exposures above 0.08 ppm. Data from regression
studies conducted to develop an exposure-response function for estimating yield loss indicated
that at least 50% of the species and cultivars tested could be predicted to exhibit a 10% yield
loss at 7-h seasonal mean O3 concentrations of 0.05 ppm or less.
Effects on Natural Ecosystems
The responses of the San Bernardino mixed forest of Southern California to 50 or
more years of chronic ozone exposures based on many studies, present a classic example of
ecosystem response to severe stress. Data from an inventory conducted from 1968 through
1972 indicated that for 5 mo of each year, trees were exposed to O3 concentrations greater
than 0.08 ppm for more than 1,300 h. Concentrations rarely decreased below 0.05 ppm at
night near the crest of the mountain slope, approximately 5,500 ft. In addition, during the
years 1973 to 1978, average 24-h O3 concentrations ranged from a background of 0.03 to 0.04
ppm in the eastern part of the San Bernardino Mountains to a maximum of 0.10 to 0.12 ppm
in the western part during May through September.
Plants accumulate, store, and use the energy in carbon compounds (sugars)
produced during photosynthesis to build their structures and to maintain the physiological
processes necessary for life. The patterns of carbon allocation to roots, stems, and leaves
directly influence growth. The strategy for carbon allocation changes during the life of a
plant, as well as with environmental conditions. Mature trees have a higher ratio of
respiration to photosynthetic tissue. Impairment of photosynthesis shifts carbon allocation
from growth and maintenance to repair; increased respiration can result in resource
imbalances. The significant changes observed in the San Bernardino forest ecosystem were a
possible outcome of the combined influences of O3 on carbon, water, and nutrient allocation.
The biochemical changes within the leaves of ponderosa and Jeffrey pine in the
San Bernardino forest, expressed as visible foliar injury, premature needle senescence,
reduced photosynthesis, and reduced carbohydrate production and allocation, resulted in
reduced tree vigor, growth, and reproduction. Reduced vigor increases susceptibility of trees
to insect pests and fungal pathogens. Premature needle senescence alters microorganismal
succession on confer needles and changes the detritus-forming process and associated nutrient
cycling.
Altered carbon allocation is important in the formation of mycorrhizae (fungus
roots), which are an extremely important but unheralded component of all ecosystems; the
majority of all plants depend on them because they are integral in the uptake of mineral
nutrients and water from the soil. Carbon-containing exudates from the roots are necessary
for the formation of mycorrhizae. Reduced carbon allocation to plant roots affects
mycorrhizal formation and impacts plant growth. Exposure to ozone, therefore, affects plant
growth both above and below ground.
Small changes in photosynthesis or carbon allocation can alter profoundly the
structure of a forest. Ecosystem responses to stress begin with the response of the most
sensitive individuals of a population. Stresses, whose primary effects occur at the molecular
level (within the leaves), must be propagated progressively through more integrated levels of
organ physiology (e.g., leaf, branch, root) to whole plant physiology, then to populations
within the stand (community), and finally to the landscape level to produce ecosystem effects.
Only a small fraction of stresses at the molecular level become disturbances at the tree, stand,
or landscape level. The time required for a stress to be propagated from one level to the next
(it can take years) determines how soon the effects of the stress can be observed or measured.
The primary effect of O3 on ponderosa and Jeffrey pine, two of the more
susceptible members of the San Bernardino forest community, was that the trees were no
longer able to compete effectively for essential nutrients, water, light, and space. Decline in
the sensitive trees, a consequence of altered competitive conditions, permitted the enhanced
growth of more tolerant species. Removal of the ecosystem dominants at the population level
changed its structure and altered the processes of energy flow and nutrient cycling, returning
the ecosystem to a less complex stage.
The San Bernardino Mountains continue to experience exposure to O3; however,
there has been a gradual decline in concentrations and length of exposure. Ozone
concentrations of 0.06 ppm or higher of varying durations capable of causing injury to trees in
forest ecosystems have been observed during the past 5 years in the Sierra Nevada Mountains
and the Appalachian Mountains from Georgia to Maine. Visible injury to forest trees and
other vegetation in these areas has been observed.
Injury to sensitive trees from exposure to ozone concentrations 0.06 ppm or greater
in the Sierra Nevada Mountains and the Appalachian Mountains has never had the impact on
these ecosystems that it did on the San Bernardino forest. Forest stands differ greatly in age,
species composition, stability, and capacity to recover from disturbance. In addition, the
position in the stand or community of the most sensitive species is extremely important.
Ponderosa and Jeffrey pine were the dominant species in the San Bernardino forest. Removal
of populations of these trees altered both ecosystem structure and function. Both the Sierra
Nevada Mountains and the Appalachian Mountains are biologically more diverse. Removal of
sensitive individuals of eastern white pine and black cherry has not visibly altered the forest
ecosystems along the Appalachian Mountains, possibly because of the absence of population
changes in these species. Decline and dieback of trees on Mt. Mitchell, NC, and Camel's
Hump, VT, cannot be related solely to O3 injury.
Effects on Agriculture, Forestry, and Ecosystems: Economics
A number of economic assessments of the effects of O3 on agriculture have been
performed over the last decade. All use NCLAN response data to predict crop yield changes.
Although these studies employ somewhat different economic assessment methodologies, each
shows national-level economic losses to major crops in excess of $1 billion (1990 dollars)
from exposure to ambient concentrations of O3. These studies also evaluate the sensitivity of
the economic estimates to uncertainties in data, including the NCLAN response data. The
economic assessment models used could be adapted to future O3-crop yield response findings,
if available.
The plant science literature shows that O3 adversely influences physiological
performance of both urban and native tree species; the limited economic literature also
demonstrates that changes in growth have economic consequences. However, the natural
science and economic literature on the topic are not yet mature enough to conclude
unambiguously that ambient O3 is imposing economic costs. The economic effects of O3 on
ecosystems have not yet been addressed in the published literature. There is, however, an
emerging interest in applying economic concepts and methods to the management of
ecosystems.
Effects on Materials
Over four decades of research show that O3 damages certain materials such as
elastomers, textile fibers, and dyes. The amount of damage to actual in-use materials and the
economic consequences of that damage are poorly characterized.
Natural rubber and synthetic polymers of butadiene, isoprene, and styrene, used in
products like automobile tires and protective outdoor electrical coverings, account for most of
the elastomer production in the United States. The action of O3 on these compounds is well
known, and concentration-response relationships have been established and corroborated by
several studies. These relationships, however, must be correlated with adequate exposure
information based on product use. For these and other economically important materials,
protective measures have been formulated to reduce the rate of oxidative damage. When
antioxidants and other protective measures are incorporated in elastomer production, the
O3-induced damage is reduced considerably, although the extent of reduction differs widely
according to the material and the type and amount of protective measures used.
Both the type of dye and the material in which it is incorporated are important
factors in the resistance of a fabric to O3. Some dyed fabrics, such as royal blue and red
rayon-acetate and plum cotton are resistant to O3. On the other hand, anthraquinone dyes on
nylon fibers are sensitive to fading by O3. Field studies and laboratory work show a positive
association between O3 levels and dye fading of nylon materials. At present, the available
research is insufficient to quantify the amount of damaged materials attributable to O3 alone.
The degradation of fibers from exposure to O3 is poorly characterized. In general,
most synthetic fibers, such as modacrylic and polyester, are relatively resistant, whereas
cotton, nylon, and acrylic fibers have greater but varying sensitivities to O3. Ozone reduces
the breaking strength of these fibers, and the degree of strength reduction depends on the
amount of moisture present. The limited research in this area indicates that O3 in ambient air
may have a minimal effect on textile fibers, but additional research is needed to verify this
conclusion.
A number of artists' pigments and dyes are sensitive to O3 and other oxidants; in
particular, many organic pigments are subject to fading or other color changes when exposed
to O3. Although most, but not all, modern fine arts paints are more O3 resistant, many older
works of art are at risk of permanent damage due to O3-induced fading.
A great deal of work remains to be done to develop quantitative estimates of the
economic damage to materials from photochemical oxidants. Most of the available studies are
outdated in terms of O3 concentrations, technologies, and supply-demand relationships.
Additionally, little is known about the physical damage functions, so cost estimates have been
simplified to the point of not properly recognizing many of the scientific complexities of the
impact of O3.
1.6 Toxicological Effects of Ozone and Related
Photochemical Oxidants
Respiratory Tract Effects of Ozone
Biochemical Effects
Knowledge of molecular targets provides a basis for understanding mechanisms of
effects and strengthening animal-to-human extrapolations. Ozone reacts with polyunsaturated
fatty acids and sulfhydryl, amino, and some electron-rich compounds. These elements are
shared across species. Several types of reactions are involved, and free radicals may be
created. Based on this knowledge, it has been hypothesized that the O3 molecule is unlikely to
penetrate the liquid linings of the respiratory tract (RT) to reach the tissue, raising the
possibility that reaction products exert effects.
In acute and short-term exposure studies, a variety of lung lipid changes occur,
including an increase in arachidonic acid, the further metabolism of which produces a variety
of biologically active mediators that can affect host defenses, lung function, the immune
system, and other functions.
The level of lung antioxidant metabolism increases after O3 exposure, probably as a
result of the increase in the number of Type 2 cells, which are rich in antioxidant enzymes.
Collagen (the structural protein involved in fibrosis) increases in O3-exposed lungs
in a manner that has been correlated to structural changes (e.g., increased thickness of the
tissue between the air and blood after prolonged exposure). Some studies found that the
increased collagen persists after exposure ceases.
Generally, O3 enhances lung xenobiotic metabolism after both short- and long-term
exposure, possibly as a result of morphological changes (increased numbers of nonciliated
bronchiolar epithelial cells). The impact of this change is dependent on the xenobiotics
involved; for example, the metabolism of benzo[a]pyrene to active metabolites was enhanced
by O3.
Lung Inflammation and Permeability Changes
Elevated concentrations of O3 disrupt the barrier function of the lung, resulting in
the entry of compounds from the airspaces into the blood and the entry of serum components
(e.g., protein) and white blood cells (especially polymorphonuclear leukocytes [PMNs]) into
the airspaces and lung tissue. This latter impact reflects the initial stage of inflammation.
These cells can release biologically active mediators that are capable of a number of actions,
including damage to other cells in the lung. In lung tissue, this inflammation also can increase
the thickness of the air-blood barrier.
Increases in permeability and inflammation have been observed at levels as low as
0.1 ppm O3 (2 h/day, 6 days; rabbits). After acute exposures, the influence of the time of
exposure (from two to several hours) increases as the concentration of O3 increases.
Long-term exposure effects are discussed under lung morphology.
The impacts of these changes are not fully understood. At higher O3 concentrations
(e.g., 0.7 ppm, 28 days), the diffusion of oxygen into the blood decreases, possibly because
the air-blood barrier is thicker; cellular death may result from the enzymes released by the
inflammatory cells; and host defense functions may be altered by mediators.
Effects on Host Defense Mechanisms
Exposure to elevated concentrations of ozone results in alterations of all defense
mechanisms of the RT, including mucociliary and alveolobronchiolar clearance, functional and
biochemical activity of the alveolar macrophage (AM), and immunologic competence. These
effects can cause susceptibility to bacterial respiratory infections.
Mucociliary clearance, which removes particles and cellular debris from the
conducting airways, is slowed by acute, but not repeated exposures to O3. Ciliated epithelial
cells that move the mucous blanket are altered or destroyed by acute and chronic exposures.
Neonatal sheep exposed to O3 do not have normal development of the mucociliary system.
Such effects could prolong the retention of unwanted substances (e.g., inhaled particles) in the
lungs, allowing them to exert their toxicity for a longer period of time.
Alveolar clearance mechanisms, which center on the functioning of AMs, are
altered by O3. Short-term exposure to levels as low as 0.1 ppm O3 (2 h/day, 1 to 4 days;
rabbits) accelerates clearance, but longer exposures do not. Even so, after a 6-week exposure
of rats to an urban pattern of O3, the retention of asbestos fibers in a region protected by
alveolar clearance is prolonged.
Alveolar macrophages engulf and kill microbes, as well as clear the deeper regions
of the lungs of nonviable particles; AMs also participate in immunological responses, but little
is known about the effects of O3 on this function. Acute exposures of rabbits to levels as low
as 0.1 ppm O3 decrease the ability of AMs to ingest particles. This effect is displayed in
decreases in the ability of the lung to kill bacteria after acute exposure of mice to levels as low
as 0.4 ppm O3.
Both the pulmonary and systemic immune system are affected by O3, but in a
poorly understood way. It appears that the part of the immune system dependent on T-cell
function is more affected than is the part dependent on B-cell function.
Dysfunction of host defense systems results in enhanced susceptibility to bacterial
lung infections. For example, acute exposure to O3 concentrations as low as 0.08 ppm for
3 h can overcome the ability of mice to resist infection with streptococcal bacteria, resulting in
mortality. However, more prolonged exposures (weeks, months) do not cause greater effects
on infectivity.
Effects on antiviral defenses are more complex and less well understood. Only
high concentrations (1.0 ppm O3, 3 h/day, 5 days; mice) increase viral-induced mortality.
Apparently, O3 does not impact antiviral clearance mechanisms. Although O3 does not affect
acute lung injury from influenza virus infection, it does enhance later phases of the course of
an infection (i.e., postinfluenzal alveolitis).
Morphological Effects
Elevated concentrations of O3 cause similar types of alterations in lung structure in
all laboratory animal species studied, from rats to monkeys. In the lungs, the most affected
cells are the ciliated epithelial cells of the airways and Type 1 epithelial cells of the
gas-exchange region. In the nasal cavity, ciliated cells are also affected.
The centriacinar region (CAR; the junction of the conducting airways and gas-exchange regions) is the primary target, possibly because this area receives the greatest dose of
O3. The ciliated cells can be killed and replaced by nonciliated cells (i.e., cells not capable of
clearance functions that also have increased ability to metabolize some foreign compounds).
Mucous-secreting cells are affected, but to a lesser degree. Type 1 cells, across which gas
exchange occurs, can be killed; they are replaced by Type 2 cells, which are thicker and
produce more lipids. An inflammatory response also occurs in the tissue. The tissue is
thickened further in later stages when collagen (a structural protein increased in fibrosis) and
other elements accumulate. Although fibrotic changes have been observed in the CAR, they
have not been distributed throughout the whole lung.
The distal airway is remodeled; more specifically, bronchiolar epithelium replaces
the cells present in alveolar ducts. Concurrent inflammation may play a role. This effect has
been observed at 0.25 ppm O3 (8 h/day, 18 mo) in monkeys; at a higher concentration, this
remodeling persists after exposure stops.
The progression of effects during and after a chronic exposure is complex. Over
the first few days of exposure, inflammation peaks and then drops considerably, plateauing for
the remainder of exposure, after which it largely disappears. Epithelial hyperplasia increases
rapidly over the first few days and rises slowly or plateaus thereafter; when exposure ends, it
begins to return toward normal. In contrast, fibrotic changes in the tissue between the air and
blood increase very slowly over months of exposure, and, after exposure ceases, the changes
sometimes persist or increase.
The pattern of exposure can make a major difference in effects. Monkeys exposed
to 0.25 ppm O3 (8 h/day) every other month of an 18-mo period had equivalent changes in
lung structure, more fibrotic changes, and more of certain types of pulmonary function
changes than did monkeys exposed every day over the 18 mo. From this work and rat studies,
it appears that natural seasonal patterns may be of more concern than more continuous
exposures. Thus, long-term animal studies with uninterrupted exposures may underestimate
some of the effects of O3.
The morphologic lesions described in early publications on laboratory animals
exposed to O3 do not meet the current criteria for emphysema of the type seen in human lungs.
Effects on Pulmonary Function
Pulmonary function changes in animals resemble those observed in humans after
acute exposure.
During acute exposure, the most commonly observed alterations are increased
frequency of breathing and decreased tidal volume (i.e., rapid, shallow breathing). This has
been reported at exposures as low as 0.2 ppm O3 for 3 h (rats). Typically, higher
concentrations (around 1 ppm) are required to affect breathing mechanics (compliance and
resistance). Extended characterizations of pulmonary function show types of changes
generally seen in humans. For example, there are decreased lung volumes at levels ò0.5 ppm
O3 (a few hours; rats).
When rats are exposed to O3 for 2 h/day for 5 days, the pattern of attenuation of
pulmonary function responses is similar to that observed in humans. Other biochemical
indicators of lung injury did not return to control values by Day 5, and morphological changes
increased in severity over the period of exposure. Thus, attenuation did not result in
protection against all the effects of O3.
Long-term exposures have provided mixed results on pulmonary function, including
no or minimal effects, restrictive effects, and obstructive effects. When changes occurred and
postexposure examinations were performed, pulmonary function recovered.
Genotoxicity and Carcinogenicity of Ozone
The chemical reactivities of O3 give it the potential to be a genotoxic agent.
In vitro studies are difficult to interpret because the culture systems used allowed
the potential formation of artifacts, and high or very high concentrations of O3 often were
used. Generally, in these studies, O3 causes DNA strand breaks, sometimes is weakly
mutagenic, and causes cellular transformation and chromosomal breakage. The latter finding
has been investigated in vivo, with mixed results in animals.
The few earlier long-term carcinogenic studies in laboratory animals, with or
without coexposure to known carcinogens, are either negative or ambiguous.
The National Toxicology Program (NTP) completed chronic rat and mouse cancer
bioassays using commonly accepted experimental approaches and designs. Both male and
female rats and mice were studied. Animals were exposed for 2 years (6 h/day, 5 days/week)
to 0.12, 0.5, and 1.0 ppm O3 or for a lifetime to the same levels (except 0.12 ppm).
Following their standard procedures for determination of weight-of-evidence for
carcinogenicity, the NTP reported "no evidence" in rats, "equivocal evidence" in male mice,
and "some evidence" in female mice. The increases in adenomas and carcinomas were
observed only in the lungs. There was no concentration response. One of the reasons for the
designation of "some evidence" in female mice was that when the 2-year and lifetime exposure
studies were combined, there was a statistically significant increase in total tumors at 1.0 ppm.
Lung tumors from control and O3-exposed mice also were examined for the presence of
mutated Ha-ras oncogenes. Although the types of mutations found were similar in both
groups, a higher incidence of mutations was found in lung tumors from the O3-exposed mice.
At the present time, however, there is inadequate information to provide mechanistic support
for the finding in mice. Thus, the potential for animal carcinogenicity is uncertain.
In a companion NTP study, male rats were treated with a tobacco carcinogen and
exposed for 2 years to 0.5 ppm O3. Ozone did not affect the response and therefore had no
tumor promoting activity.
Systemic Effects of Ozone
Ozone causes a variety of effects on tissues and organs distant from the lung.
Because O3 itself is not thought to penetrate the lung, these systemic effects are either
secondary to lung alterations or result from reaction products of O3. Effects have been
observed on clinical chemistry, white blood cells, red blood cells, the circulatory system, the
liver, endocrine organs, and the central nervous system. Most of these effects cannot be
interpreted adequately at this time and have not been investigated in humans, but it is of
interest to note that O3 exposures causing effects on the RT of animals cause a wide array of
effects on other organs also.
Several behavior changes occur in response to O3. For example, 0.12 ppm O3 (6 h,
rats) decreases wheel-running activity, and 0.5 ppm (1 min) causes mice to avoid exposure.
These effects are not fully understood, but they may be related to lung irritation or decreased
ability to exercise.
Although cardiovascular effects, such as slowed heart rate and decreased blood
pressure, occur in O3-exposed rats, some observed interactions with thermoregulation prevent
qualitative extrapolation of these effects to humans at this time.
Developmental toxicity studies in pregnant rats summarized in the 1986 O3 criteria
document showed that levels up to about 2.0 ppm O3 did not cause birth defects. Rat pups
from females exposed to 1.0 ppm O3 during certain periods of gestation weighed less or had
delays in development of behaviors (e.g., righting, eye opening). No "classical" reproductive
assays with O3 were found.
Other studies have indicated that O3 can affect some endocrine organs (i.e.,
pituitary-thyroid-adrenal axis, parathyroid gland). It appears that the liver has less ability to
detoxify drugs after O3 exposure, but assays of liver enzymes involved in xenobiotic
metabolism are inconsistent.
Interactions of Ozone with Other Co-occurring Pollutants
Animal studies of the effects of O3 in combination with other air pollutants show
that antagonism, additivity, and synergism can result, depending on the animal species,
exposure regimen, and health endpoint. Thus, these studies clearly demonstrate the major
complexities and potential importance of interactions but do not provide a scientific basis for
predicting the results of interactions under untested ambient exposure scenarios.
1.7 Human Health Effects of Ozone and Related
Photochemical Oxidants
This section summarizes key effects associated with exposure to O3, the major
component of photochemical oxidant air pollution that is clearly of most concern to the health
of the human population. Another, often co-occurring photochemical oxidant component of
"smog" is PAN, but this compound has been demonstrated to be primarily responsible for
induction of smog-related eye irritation (stinging of eyes). Limited pulmonary function studies
have shown no effects of PAN at concentrations below 0.13 to 0.30 ppm, which are much
higher than the generally encountered ambient air levels in most cities.
Controlled Human Studies of Acute Ozone Effects
Effects on Lung Function
Controlled studies in healthy adult subjects have demonstrated O3-induced
decrements in pulmonary function, characterized by alterations in lung volumes and flow and
airway resistance and responsiveness. Respiratory symptoms, such as cough and pain on deep
inspiration, are associated with these changes in lung function.
Ozone-induced decreases in lung volume, specifically forced vital capacity (FVC)
and forced expiratory volume in 1 s (FEV1), largely can be attributed to decreases in
inspiratory capacity (the ability to take a deep breath), although at higher exposure
concentrations, there is clearly an additional component that is not volume dependent. Lung
volumes recover to a large extent within 2 to 6 h; normal baseline function typically is
reestablished within 24 h, but not fully with more severe exposures.
Ozone causes increased airway resistance and may cause reductions in expiratory
flow and the FEV1/FVC ratio.
Ozone causes an increase in airway responsiveness to nonallergenic stimuli (e.g.,
histamine, methacholine) in healthy and asthmatic subjects. There is no clear evidence of a
relationship between O3-induced lung volume changes and changes in airway responsiveness.
Inflammation and Host Defense Effects
Controlled studies in healthy adult subjects also indicate that O3 causes an
inflammatory response in the lungs characterized by elevated levels of PMNs, increased
epithelial permeability, and elevated levels of biologically active substances (e.g.,
prostaglandins, proinflammatory mediators, cytokines).
Inflammatory responses to O3 can be detected within 1 h after a single 1-h exposure
with exercise to concentrations ò0.3 ppm; the increased levels of some inflammatory cells and
mediators persist for at least 18 h. The temporal response profile is not defined adequately,
although it is clear that the time course of response varies for different mediators and cells.
Lung function and respiratory symptom responses to O3 do not seem to be
correlated with airway inflammation.
Ozone also causes inflammatory responses in the nose, marked by increased
numbers of PMNs and protein levels suggestive of increased permeability.
Alveolar macrophages removed from the lungs of human subjects after 6.6 h of
exposure to 0.08 and 0.10 ppm O3 have a decreased ability to ingest microorganisms,
indicating some impairment of host defense capability.
Ozone Exposure-Response Relationships
Functional, symptomatic, and inflammatory responses to O3 increase with
increasing exposure dose of O3. The major determinants of the exposure dose are
O3 concentration (C), exposure duration (T), and the amount of ventilation (E).
Exercise increases response to O3 by increasing E (greater mass delivered), tidal
volume, inspiratory flow (greater percentage delivery), and the intrapulmonary
O3 concentration.
Repeated daily exposures to relatively high levels of O3 doses (C T E)
causing substantial reductions in FEV1 (ò20% decrement) typically cause exacerbation of the
lung function and respiratory symptom responses on the second exposure day. However,
attenuation of these responses occurs with continued exposures for a few days. Most
inflammatory responses also attenuate; for example, the PMN influx is absent after five
consecutive exposures.
Multihour exposures (e.g., for up to 7 h) to O3 concentrations as low as 0.08 ppm
cause small but statistically significant decrements in lung function, increases in respiratory
symptoms, and increases in PMNs and protein levels. Ozone C is a more important factor
than exercise E or T in predicting responses to multihour low-level O3 exposure. There is
clear evidence of a response plateau in terms of lung volume response to prolonged
O3 exposure. This evidence suggests that for a given combination of exercise and
O3 concentration (i.e., dose rate), there is a response plateau; continued exposure (i.e,
increased T) at that dose rate will not increase response. Therefore, quantitative extrapolation
of responses to longer exposure durations is not valid.
Mechanisms of Acute Pulmonary Responses
The mechanisms leading to the observed pulmonary responses induced by O3 are
beginning to be better understood. The available descriptive data suggest a number of
mechanisms leading to the alterations in lung function and respiratory symptoms, including
O3 delivery to the tissue (i.e., the inhaled concentration, breathing pattern, airway geometry;
O3 reactions with the airway lining fluid and epithelial cell membranes; local tissue responses,
including injury and inflammation; and stimulation of neural afferents (bronchial C-fibers) and
the resulting reflex responses and symptoms. The cyclooxygenase inhibitors block production
of prostaglandin E2 and interleukin-6 as well as reduce lung volume responses; however, these
drugs do not reduce inflammation and levels of cell damage markers such as lactate
dehydrogenase.
Effects on Exercise Performance
Maximal oxygen uptake, a measure of peak exercise performance capacity, is
reduced in healthy young adults if preceded by O3 exposures sufficient to cause marked
changes in lung function (i.e., decreases of at least 20%) and increased subjective symptoms
of respiratory discomfort. Limitations in exercise performance may be related to increased
symptoms, especially those related to breathing discomfort.
Factors Modifying Responsiveness to Ozone
Many variables have the potential for influencing responsiveness to O3; however,
most are addressed inadequately in the available clinical data to make definitive conclusions.
Active smokers are less responsive to O3 exposure, which may reverse following
smoking cessation, but these results should be interpreted with caution.
The possibility of age-related differences in response to O3 has been explored,
although young adults historically have provided the subject population for controlled human
studies. Children and adolescents have lung volume responses to O3 similar to those of young
adults, but lack respiratory symptoms. Pulmonary function responsiveness in adults appears to
decrease with age, whereas symptom rates remain similar to young adults. Group mean lung
function responses of adults over 50 years of age are less than those of children, adolescents,
and young adults.
The available data have not demonstrated conclusively that men and women
respond differently to O3. Likewise, pulmonary function responses of women have been
compared during different phases of the menstrual cycle, but the results are conflicting.
If gender differences exist for lung function responsiveness to O3, they are not based on
hormonal changes, differences in lung volume, or the ratio of FVC to E.
There is no compelling evidence, to date, suggesting that any ethnic or racial
groups have a different distribution of responsiveness to O3.
Seasonal and ambient factors may vary responsiveness to O3, but further research is
needed to determine how they affect individual subjects. Individual sensitivity to O3 may vary
throughout the year, related to seasonal variations in ambient O3 concentrations.
The specific inhalation route appears to be of minor importance in exercising
adults. Exposure to O3 by oral breathing (i.e., mouthpiece) yields results similar to exposure
by oronasal breathing (i.e., chamber exposures).
Population Groups at Risk from Ozone Exposure
Population groups that have demonstrated increased responsiveness to ambient
concentrations of O3 consist of exercising healthy and asthmatic individuals, including
children, adolescents, and adults.
Available evidence from controlled human studies on subjects with preexisting
disease suggests that mild asthmatics have similar lung volume responses, but greater airway
resistance responses to O3 than nonasthmatics; and that moderate asthmatics may have, in
addition, greater lung volume responses than nonasthmatics.
Of all the other population groups studied, those with preexisting limitations in
pulmonary function and exercise capacity (e.g., chronic obstructive pulmonary disease,
chronic bronchitis, ischemic heart disease) would be of primary concern in evaluating the
health effects of O3. Unfortunately, limitations of subject selection, standardized methods of
subject characterization, and range of exposure hamper the ability to make definitive
conclusions regarding the relative responsiveness of most chronic disease subjects.
Effects of Ozone Mixed with Other Pollutants
No significant enhancement of respiratory effects has been demonstrated
consistently for simultaneous exposures of O3 mixed with SO2, NO2, H2SO4, HNO3, particulate
aerosols, or combinations of these pollutants. It is fairly well established that simultaneous
exposure of healthy adults and asthmatics to mixtures of O3 and other pollutants for short
periods of time (<2 h) induces pulmonary function responses not significantly different from
those following O3 alone when studies are conducted at the same O3 concentration. Exposure
to PAN has been reported to induce greater pulmonary function responses than exposure to
O3 alone, but at PAN concentrations (>0.27 ppm) much higher than ambient levels.
Unfortunately, only a limited number of pollutant combinations and exposure protocols have
been investigated, and subject groups are small and are representative of only small portions of
the general population. Thus, much is unknown about the relationships between O3 and the
complex mix of pollutants found in the ambient air.
Prior exposure to O3 in asthmatics may cause an increase in response to other
pollutant gases, especially SO2. Likewise, prior exposure to other pollutants can enhance
responses to O3 exposure.
Controlled Human Studies of Ambient Air Exposures
Mobile laboratory studies of lung function and respiratory symptoms in a local
subject population exposed to ambient photochemical oxidant pollution provide quantitative
information on exposure-response relationships for O3. A series of these studies from
Los Angeles has demonstrated pulmonary function decrements at mean ambient
O3 concentrations of 0.14 ppm in exercising healthy adolescents and increased respiratory
symptoms and pulmonary function decrements at 0.15 ppm in heavily exercising athletes and
at 0.17 ppm in lightly exercising healthy and asthmatic subjects. Comparison of the observed
effects in exercising athletes with controlled chamber studies at comparable O3 concentrations
showed no significant differences in lung function and symptoms, suggesting that coexisting
ambient pollutants have a minimal contribution to the measured responses under typical
summer ambient conditions in Southern California.
Field and Epidemiology Studies of Ambient Air Exposures
Individual-level field studies and aggregate-level time-series studies have addressed
the acute effects of O3 on lung function decrements and increased morbidity and mortality in
human populations exposed to real-world conditions of O3 exposure.
Camp and exercise studies of lung function provide quantitative information on
exposure-response relationships linking lung function declines with O3 exposure occurring in
ambient air. Combined statistical analysis of six recent camp studies in children yields an
average relationship between decrements in FEV1 and previous-hour O3 concentration of 0.50
mL/ppb. Two key studies of lung function measurements before and after well-defined
outdoor exercise events in adults have yielded exposure-response slopes of 0.40 and
1.35 mL/ppb. The magnitude of pulmonary function declines with O3 exposure is consistent
with the results of controlled human studies.
Daily life studies support a consistent relationship between O3 exposure and acute
respiratory morbidity in the population. Respiratory symptoms (or exacerbation of asthma)
and decrements in peak expiratory flow rate are associated with increasing ambient O3,
particularly in asthmatic children; however, concurrent temperature, particles, acidity
(hydrogen ions), aeroallergens, and asthma severity or medication status also may contribute
as independent or modifying factors. Aggregate results show greater responses in asthmatic
individuals than in nonasthmatics, indicating that asthmatics constitute a sensitive group in
epidemiologic studies of oxidant air pollution.
Summertime daily hospital admissions for respiratory causes in various locations of
eastern North America have consistently shown a relationship with ambient levels of O3,
accounting for approximately one to three excess respiratory hospital admissions per hundred
parts per billion O3 per million persons. This association has been shown to remain even after
statistically controlling for the possible confounding effects of temperature and copollutants
(e.g., hydrogen ions, sulfate, and particles less than 10 m), as well as when considering only
concentrations below 0.12 ppm O3.
Many of the time-series epidemiology studies looking for associations between
O3 exposure and daily human mortality have been difficult to interpret because of
methodological or statistical weaknesses, including the a failure to account for other pollutant
and environmental effects. One of the two most useful new studies on O3-mortality found a
small but statistically significant association in Los Angeles when peak 1-h maximum
O3 concentrations reached concentrations greater than 0.2 ppm during the study period.
A second study in regions with lower (ó0.15 ppm) maximum 1-h O3 concentrations (St. Louis,
MO, and Kingston-Harriman, TN) did not detect a significant O3 association with mortality.
Only suggestive epidemiologic evidence exists for health effects of chronic ambient
O3 exposure in the population. All of the available studies of chronic respiratory system
effects in exposed children and adults are limited by a simplistic assignment of exposure or by
their inability to isolate potential effects related to O3 from those of other pollutants, especially
particles.
1.8 Extrapolation of Animal Toxicological Data to Humans
There have been significant advances in O3 dosimetry since 1986 that better enable
quantitative extrapolation with marked reductions in uncertainty. Experiments and models
describing the uptake efficiency and delivered dose of O3 in the RT of animals and humans are
beginning to present a clearer picture than has existed previously.
The total RT uptake efficiency of rats at rest is approximately 50%. Within the RT
of the rat, 50% of the O3 taken up by the RT is removed in the head, 7% in the
larynx/trachea, and 43% in the lungs.
In humans at rest, the total RT uptake efficiency is between 80 and 95%. Total RT
uptake efficiency falls as flow increases. As tidal volume increases, uptake efficiency
increases and flow dependence lessens. Pulmonary function response data and O3 uptake
efficiency data in humans generally indicate that the mode of breathing (oral versus nasal
versus oronasal) has little effect on upper RT or on total RT uptake efficiency, although one
study suggests that the nose has a higher uptake efficiency than the mouth.
When all of the animal and human in vivo O3 uptake efficiency data are compared,
there is a good degree of consistency across data sets. This agreement raises the level of
confidence with which these data sets can be used to support dosimetric model formulations.
Several mathematical dosimetry models have been developed since 1986.
Generally, the models predict that net O3 dose to lung lining fluid plus tissue gradually
decreases distally from the trachea toward the end of the tracheobronchial region and then
rapidly decreases in the pulmonary region.
When the dose of O3 to lung tissue is computed theoretically, it is found to be very
low in the trachea; to increase to a maximum in the terminal bronchioles of the first generation
pulmonary region; and then to decrease rapidly, moving further into the pulmonary region.
The increased tidal volume and flow, associated with exercise in humans, shifts O3 dose
further into the periphery of the lung and causes a disproportionate increase in distal lung
dose.
Predictions of delivered dose have been used to investigate both acute and chronic
O3 responses in the context of intra- and interspecies comparisons. In the case of intraspecies
comparisons, for example, the distribution of predicted O3 tissue dose to a ventilatory unit in a
rat as a function of distance from the bronchoalveolar duct junction is very consistent with the
distribution of alveolar wall thickening. In the case of interspecies comparisons (using the
delivered O3 dose to the proximal alveolar regions), although the functional responses (e.g.,
rapid, shallow breathing) differ markedly between rats and humans, there is similarity of acute
dose-response patterns in inflammation (influx of cells and protein) among species, with
humans and guinea pigs more responsive than rats and rabbits, and similarity of chronic dose-response patterns for increased alveolar interstitial thickness in the CAR of the lung, with
monkeys being more responsive than humans and rats less responsive. In other words, the
quantitative relationship between animal and human responses is dependent on the animal
species and the endpoint.
In summary, there is an emerging consistency among a variety of O3 dosimetry data
sets and between the experimental data and theoretical predictions of O3 dose. The
convergence of experimental data with theoretical predictions lends a degree of confidence to
the use of theoretical models to predict total and regional O3 dose. The use of O3 dosimetry
data and models is beginning to provide a useful extrapolation of effects between animals and
humans. The data and models have thus far helped demonstrate that humans may be more
responsive to O3 than rats, but less responsive than monkeys with respect to acute and chronic
inflammatory responses. However, the monkey, with its similarity to the human in distal
airway structure, provides chronic effects data that may best reflect the degree to which a
comparably exposed human would respond. These findings, therefore, suggest that long-term
exposure to O3 could impart a chronic effect in humans.
1.9 Integrative Summary of Ozone Health Effects
This section summarizes the primary conclusions derived from an integration of the
known effects of O3 provided by animal toxicological, human clinical, and epidemiological
studies.
1. What are the effects of short-term (<8-h) exposures to ozone?
Recent epidemiology studies addressing the effects of short-term ambient exposure
to O3 in the population have yielded significant associations with a wide range of health
outcomes, including lung function decrements, aggravation of preexisting respiratory disease,
increases in daily hospital admissions and emergency department visits for respiratory causes,
and increased mortality. Results from lung function epidemiology studies generally are
consistent with the experimental studies in laboratory animals and humans.
Short-term O3 exposure of laboratory animals and humans causes changes in
pulmonary function, including tachypnea (rapid, shallow breathing), decreased lung volumes
and flows, and increased airway responsiveness to nonspecific stimuli. Increased airway
resistance occurs in both humans and laboratory animals, but typically at higher exposure
levels than other functional endpoints. In addition, adult human subjects experience
O3-induced symptoms of airway irritation such as cough or pain on deep inspiration. The
changes in pulmonary function and respiratory symptoms occur as a function of exposure
concentration, duration, and level of exercise. Adult human subjects with mild asthma have
responses in lung volume and airway responsiveness to bronchoconstrictor drugs that are
qualitatively similar to those of nonasthmatics. Respiratory symptoms are also similar, but
wheezing is a prevalent symptom in O3-exposed asthmatics in addition to the other
demonstrated symptoms of airway irritation. Airway resistance, however, increases relatively
more in asthmatics from an already higher baseline. Recovery from the effects of O3 on
pulmonary function and symptoms is usually complete within 24 h of the end of exposure,
although other responses may persist somewhat longer.
An association between daily mortality and O3 concentration for areas with high
O3 levels (e.g., Los Angeles) has been suggested, although the magnitude of
such an effect is unclear.
Increased O3 levels are associated with increased hospital admissions and
emergency department visits for respiratory causes. Analyses from data in the
northeastern United States suggest that O3 air pollution is associated with a
substantial portion (on the order of 10 to 20%) of all summertime respiratory
hospital visits and admissions.
Pulmonary function in children at summer camps in southern Ontario, Canada,
in the northeastern United States, and in Southern California is associated with
O3 concentration. Meta-analysis indicates that a 0.5-mL decrease in FEV1 is
associated with a 1-ppb increase in O3 concentration. For preadolescent
children exposed to 120 ppb (0.12 ppm) ambient O3, this amounts to an average
decrement of 2.4 to 3.0% in FEV1. Similar responses are reported for children
and adolescents exposed to O3 in ambient air or O3 in purified air for 1 to
2 h while exercising.
Pulmonary function decrements generally are observed in healthy subjects (8 to
45 years of age) after 1 to 3 h of exposure as a function of the level of exercise
performed and the O3 concentration inhaled during the exposure. Group mean
data from numerous controlled human exposure and field studies indicate that,
in general, statistically significant pulmonary function decrements beyond the
range of normal measurement variability (e.g., 3 to 5% for FEV1) occur
(1) at >0.50 ppm O3 when at rest,
(2) at >0.37 ppm O3 with light exercise (slow walking),
(3) at >0.30 ppm O3 with moderate exercise (brisk walking),
(4) at >0.18 ppm O3 with heavy exercise (easy jogging), and
(5) at >0.16 ppm O3 with very heavy exercise (running).
Smaller group mean changes (e.g., <5%) in FEV1 have been observed at lower
O3 concentrations than those listed above. For example, FEV1 decrements have
been shown to occur with very heavy exercise in healthy adults at 0.15 to 0.16
ppm O3, and such effects may occur in healthy young adults at levels as low as
0.12 ppm. Also, pulmonary function decrements have been observed in
children and adolescents at concentrations of 0.12 and 0.14 ppm O3 with heavy
exercise. Some individuals within a study may experience FEV1 decrements in
excess of 15% under these exposure conditions, even when the group mean
decrement is less than 5%.
For exposures of healthy subjects performing moderate exercise during longer
duration exposures (6 to 8 h), 5% group mean decrements in FEV1 were
observed at
(1) 0.08 ppm O3 after 5.6 h,
(2) 0.10 ppm O3 after 4.6 h, and
(3) 0.12 ppm O3 after 3 h.
For these same subjects, 10% group mean FEV1 decrements were observed at
0.12 ppm O3 after 5.6 and 6.6 h. As in the shorter duration studies, some
individuals experience changes larger than those represented by the group mean
changes.
An increase in the incidence of cough has been reported at O3 concentrations as
low as 0.12 ppm in healthy adults during 1 to 3 h of exposure with very heavy
exercise. Other respiratory symptoms, such as pain on deep inspiration,
shortness of breath, and lower respiratory scores (a combination of several
symptoms), have been observed at 0.16 to 0.18 ppm O3 with heavy and very
heavy exercise. Respiratory symptoms also have been observed following
exposure to 0.08, 0.10, and 0.12 ppm O3 for 6.6 h with moderate levels of
exercise.
Increases in nonspecific airway responsiveness in healthy adults have been
observed after 1 to 3 h of exposure to 0.40 but not 0.20 ppm O3 at rest and
have been observed at concentrations as low as 0.18 but not to 0.12 ppm
O3 during exposure with very heavy exercise. Increases in nonspecific airway
responsiveness during 6.6-h exposures with moderate levels of exercise have
been observed at 0.08, 0.10, and 0.12 ppm O3.
Short-term O3 exposure of laboratory animals and humans disrupts the barrier
function of the lung epithelium, permitting materials in the airspaces to enter lung tissue,
allowing cells and serum proteins to enter the airspaces (inflammation), and setting off a
cascade of responses.
Increased levels of PMNs and protein in lung lavage fluid have been observed
following exposure of healthy adults to 0.20, 0.30, and 0.40 ppm with very
heavy exercise and have not been studied at lower concentrations for 1- to 3-h
exposures. Increases in lung lavage protein and PMNs also have been observed
at 0.08 and 0.10 ppm O3 during 6.6-h exposures with moderate exercise; lower
concentrations have not been tested.
Short-term O3 exposure of laboratory animals and humans impairs AM clearance of
viable and nonviable particles from the lungs and decreases the effectiveness of host defenses
against bacterial lung infections in animals and perhaps in humans. The ability of AMs to
engulf microorganisms is decreased in humans exposed to 0.08 and 0.10 ppm O3 for
6.6 h with moderate exercise.
2. What are the effects of repeated, short-term exposures to ozone?
During repeated short-term exposures, some of the O3-induced responses are
partially or completely attenuated. Over a 5-day exposure, pulmonary function changes are
typically greatest on the second day, but return to control levels by the fifth day of exposure.
Most of the inflammatory markers (e.g., PMN influx) also attenuate by the fifth day of
exposure, but markers of cell damage (e.g., lactate dehydrogenase enzyme activity) do not
attenuate but continue to increase. Attenuation of lung function decrements is reversed
following 7 to 10 days without O3. Some inflammatory markers also are reversed during this
time period, but others still show attenuation even after 20 days without O3. The mechanisms
and impacts involved in attenuation are not known, although animal studies show that the
underlying cell damage continues throughout the attenuation process. In addition, attenuation
may alter the normal distribution of O3 within the lung, allowing more O3 to reach sensitive
regions, possibly affecting normal lung defenses (e.g., PMN influx in response to inhaled
microorganisms).
3. What are the effects of long-term exposures to ozone?
Available data indicate that exposure to O3 for months and years causes structural
changes in several regions of the RT, but effects may be of the greatest importance in the CAR
(where the alveoli and conducting airways meet); this region typically is affected in most
chronic airway diseases of the human lung. This information on O3 effects in the distal lung is
extrapolated from animal toxicological studies because, to date, comparable data are not
available from humans. The apparent lack of reversal of effects during periods of clean air
exposure raises concern that seasonal exposures may have a cumulative impact over many
years. The role of adaptive processes in this response is unknown but may be critically
dependent on the temporal frequency or profile of exposure. Furthermore, the interspecies
diversity in apparent sensitivity to the chronic effects of O3 is notable, with the rat representing
the lower limit of response, and the monkey the upper limit. Epidemiological studies
attempting to associate chronic health effects in humans with long-term O3 exposure provide
only suggestive evidence that such a linkage exists.
Long-term exposure of one strain of female mice to high O3 levels (1 ppm) caused a
small, but statistically significant increase in lung tumors. There was no concentration-
response relationship, and rats were not affected. Genotoxicity data are either negative or
weak. Given the nature of the database, the effects in one strain of mice cannot yet be
extrapolated qualitatively to humans. Ozone (0.5 ppm) did not show tumor-promoting activity
in a chronic rat study.
4. What are the effects of binary pollutant mixtures containing ozone?
Combined data from laboratory animal and controlled human exposure studies of
O3 support the hypothesis that coexposure to pollutants, each at low-effect levels, may result in
effects of significance. The data from human studies of O3 in combination with NO2, SO2,
H2SO4, HNO3, or CO show no more than an additive response on lung spirometry or
respiratory symptoms. The larger number of laboratory animal studies with O3 in mixture
with NO2 and H2SO4 show that effects can be additive, synergistic, or even antagonistic,
depending on the exposure regimen and the endpoint studied. This issue of exposure to
copollutants remains poorly understood, especially with regard to potential chronic effects.
5. What population groups are at risk as a result of exposure to ozone?
Identification of population groups that may show increased sensitivity to O3 is
based on their biological responses to O3, preexisting lung disease (e.g., asthma), activity
patterns, personal exposure history, and personal factors (e.g., age, nutritional status).
The predominant information on the health effects of O3 noted above comes from
clinical and field studies on healthy, nonsmoking, exercising subjects, 8 to 45 years of age.
These studies demonstrate that, among this group, there is a large variation in sensitivity and
responsiveness to O3, with at least a 10-fold difference between the most and least responsive
individuals. Individual sensitivity to O3 also may vary throughout the year, related to seasonal
variations in ambient O3 exposure. The specific factors that contribute to this large
intersubject variability, however, remain undefined. Although differences in response may be
due to the dosimetry of O3 in the RT, available data show little difference on O3 deposition in
the lungs for inhalation through the nose or mouth.
Daily life studies reporting an exacerbation of asthma and decrease in peak
expiratory flow rates, particularly in asthmatic children, appear to support the controlled
studies; however, those studies may be confounded by temperature, particle or aeroallergen
exposure, and asthma severity of the subjects or their medication use. In addition, field
studies of summertime daily hospital admissions for respiratory causes show a consistent
relationship between asthma and ambient levels of O3 in various locations in the northeastern
United States, even after controlling for independent contributing factors. Controlled studies
on mild asthmatics suggest that they have similar lung volume responses but greater airway
resistance changes to O3 than nonasthmatics. Furthermore, limited data from studies of
moderate asthmatics suggest that this group may have greater lung volume responses than
nonasthmatics.
Other population groups with preexisting limitations in pulmonary function and
exercise capacity (e.g., chronic obstructive pulmonary disease, chronic bronchitis, ischemic
heart disease) would be of primary concern in evaluating the health effects of O3.
Unfortunately, not enough is known about the responses of these individuals to make definitive
conclusions regarding their relative responsiveness to O3. Indeed, functional effects in these
individuals with reduced lung function may have greater clinical significance than comparable
changes in healthy individuals.
Currently available data follow on personal factors or personal exposure history
known or suspected of influencing responses to O3.
Human studies have identified a decrease in pulmonary function responsiveness
to O3 with increasing age, although symptom rates remain similar.
Toxicological studies are not easily interpreted but suggest that young animals
are not more responsive than adults.
Available toxicological and human data have not demonstrated conclusively that
males and females respond differently to O3. If gender differences exist for
lung function responsiveness to O3, they are not based on differences in baseline
pulmonary function.
Data are not adequate to determine whether any ethnic or racial group has a
different distribution of responsiveness to O3. In particular, the responses of
nonwhite asthmatics have not been investigated.
Information derived from O3 exposure of smokers is limited. The general trend
is that smokers are less responsive than nonsmokers. This reduced
responsiveness may wane after smoking cessation.
Although nutritional status (e.g., vitamin E deficiency) makes laboratory rats
more susceptible to O3-induced effects, it is not clear if vitamin E
supplementation has an effect in human populations. Such supplementation has
no or minimal effect in animals. The role of such antioxidant vitamins in
O3 responsiveness, especially their deficiency, has not been well studied.
Based on information presented in this document, the population groups that have
demonstrated increased responsiveness to ambient concentrations of O3 consist of exercising,
healthy and asthmatic individuals, including children, adolescents, and adults.
Please e-mail any comments or questions to the Clean Air Act Information Network.
Return to the Clean Air Act
Information Network Home Page