EPA-450/5-85-003
Regulatory Impact Analysis of the
National Ambient Air Quality Standards
for Nitrogen Dioxide
U.S. Environmental Protection Agency
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
June 1985
-------
A draft regulatory impact analysis (RIA) was reviewed by the Office
of Management and Budget (OMB) and by the Office of Air Quality Planning
and Standards, U.S. Environmental Protection Agency (EPA) and released to
the public at the'time of proposal. This final RIA was reviewed within
EPA and by OMB and is being published in conjunction with retention of
the nitrogen dioxide national ambient air quality standards. Please
address any questions on the RIA to Thomas McCurdy (919-541-5655) or
David McLamb (919-541-5611), MD-12, U.S. Environmental Protection Agency,
Research Triangle Park, N.C. 27711.
-------
TABLE OF CONTENTS
Page
I. INTRODUCTION 1
II. STATEMENT OF NEEDS AND CONSEQUENCES 3
A. Legislative Requirements Affecting Development
and Revision of NAAQS 3
B. Nature of the Nitrogen Dioxide Problem 5
C. Need for Regulatory Action 7
III. ALTERNATIVES EXAMINED 9
A. No Regulation 9
B. Other Regulatory Approaches 10
C. Regulatory Alternatives within the Scope of
Present Legislation 11
D. Market-Oriented Alternatives 12
IV. ASSESSING BENEFITS 15
A. Assessing Benefits of Air Pollution Control 15
B. Estimating Benefits Associated with Alternative
N02 NAAQS 17
1. Mechanisms of Toxicity and Nature of Effects 18
2. Sensitive Population Groups 23
V. COST ANALYSIS 26
A. Introduction 26
B. National Control Costs 30
C. Transactional Costs (Secondary Costs) •. . . . 40
1. Governmental Regulatory Costs 40
2. Adjustment Costs for Unemployed Resources 44
3. Adverse Effects on Market Structure 45
-------
Page
VI. EVALUATING BENEFITS AND COSTS 46
A. Introduction 46
B. Framework of the Incremental Cost Analysis 46
1. Cost Analysis Procedures and Assumptions 46
2. Concentrations of N02 as Effects Measure-
Procedures and Assumptions 47
3. Cases Analyzed 47
C. Incremental Cost Analysis 48
D. Qualifications of the ICA 51
1. Use of ICA in General 51
2. The Effects Measure of the ICA 53
3. Limitations in the Effects Measure of the ICA .... 53
VII. SUMMARY OF THE RATIONALE FOR SELECTING THE PROPOSED
STANDARDS 55
A. Health Effects and the Primary Standard 55
B. Welfare Effects and the Secondary Standard 61
VIII. STATUTORY AUTHORITY 66
REFERENCES 67
APPENDICES
A. TECHNICAL APPENDIX TO CHAPTER VI A-l
B. UPDATE OF NOX CONTROL COST ESTIMATES B-l
C. NOX CONTROL TECHNOLOGY FOR STATIONARY SOURCES C-l
-------
I. INTRODUCTION
This final Regulatory Impact Analysis (RIA) of the nitrogen dioxide
(N02) national ambient air quality standards (NAAQS) has been prepared in
accordance with Executive Order 12291, which requires that an RIA be done
for every rule that may result in an annual impact of $100 million or more
on the economy.
As provided for in sections 108 and 109 of the Clean Air Act as amended,
EPA has reviewed and revised the criteria upon which the existing primary
(to protect public health) and secondary (to protect public welfare) standards
are based. The existing primary and secondary standards for N02 are both
currently set at 0.053 ppm (100 ug/m3) as an annual arithmetic average.
As a result of the review and revision of the health and welfare criteria,
EPA proposed to reaffirm the existing annual average standards and
specifically requested comment on whether a separate short-term (less
than 3 hours) standard is needed to protect public health (49 FR 6866).
The Clean Air Act specifically requires that primary and secondary
NAAQS be based on scientific criteria relating to the level(s) that should be
attained to protect public health and welfare adequately. EPA and the
courts interpret the Act as precluding consideration of the cost or feasibility
of achieving such standards in determining the level of the ambient standards.
In response to Executive Order 12291, EPA has prepared a regulatory impact
analysis (RIA). However, EPA has not considered the results of this RIA in
selecting the final standards.
This RIA examines the benefits, costs, and other economic impacts of
alternative primary N02 standards on both the public and private sectors.
In specifying some aspects of the alternatives to be analyzed, EPA relied
heavily on the document Review of the National Ambient Air Quality Standards
for Nitrogen Oxides: Assessment of Scientific and Technical Information
-------
(N02 Staff Paper, EPA, 1982). The NOg Staff Paper interprets the relevant
scientific and technical information reviewed in the revised Air Quality
Criteria for Nitrogen Oxides (Criteria Document, EPA, 1983). The N0£ Staff
Paper, which has undergone careful review by the Clean Air Scientific
Advisory Committee (an independent advisory group) and the public, serves
to identify those conclusions and uncertainties in the available scientific
literature that should be considered in selecting the form, level, and
averaging times of the primary and secondary standards for
-------
II. STATEMENT OF NEEDS AND CONSEQUENCES
This chapter of the RIA summarizes the statutory requirements affecting
the development and revision of NAAQS and briefly describes the nature
of the ambient N0£ problem. The need for regulatory action and the
consequences of the regulation in terms of improving the functioning
of the market are also discussed.
A. LEGISLATIVE REQUIREMENTS AFFECTING DEVELOPMENT AND REVISION OF NAAQS
Two sections of the Clean Air Act govern the establishment and
revision of NAAQS. First, as a preliminary step in developing standards,
section 108 (42 U.S.C. § 7408) directs the Administrator to identify all
pollutants which may reasonably be anticipated to endanger public health
or welfare and to issue air quality criteria for such pollutants. Such
air quality criteria are to reflect the latest scientific information
useful in indicating the kind and extent of all identifiable effects on
public health or welfare that may be expected from the presence of the
pollutant in the ambient air. Section 109(d) of the Act (42 U.S.C.
§ 7409(d)) requires the Administrator to propose and promulgate primary
ambient air quality standards, based on the section 108 criteria and
allowing an adequate margin of safety.
Section 109(d) of the Act (42 U.S.C. § 7409(d)) requires periodic
review and, if appropriate, revision of existing criteria and standards.
If, in the Administrator's judgment, the Agency's review and revision of
criteria make appropriate the proposal of new or revised standards, such
standards are to be revised and promulgated in accordance with section
109(b). Alternatively, the Administrator may find that revision of the
standard is not appropriate and conclude the review by reaffirming the
existing standards. In addition, section 109(c) specifically requires
-------
the Administrator to promulgate a primary standard for NC>2 with an averaging
time of not more than 3 hours unless he finds no significant evidence that
such a short-term standard is required to protect public health.
Section 109(b)(l) defines the primary standard as that ambient air
quality standard the attainment and maintenance of which in the judgment of
the Administrator, based on the criteria and allowing an adequate margin of
safety, is requisite to protect the public health. The secondary standard,
as defined in section 109(b)(2), must specify a level of air quality the
attainment and maintenance of which in the judgment of the Administrator,
based on the criteria, is requisite to protect the public welfare from any
known or anticipated adverse effects associated with the presence of the
pollutant in the ambient air. Welfare effects are defined in section 302(h)
(42 U.S.C. § 7602(h)) and include effects on soils, water, crops, vegetation,
man-made materials, animals, weather, visibility, hazards to transportation,
economic values, personal comfort and well-being, and similar factors.
The Act, its legislative history, as well as recent judicial decisions
(Lead Industries Association, Inc. v. EPA, 1980; American Petroleum Institute
v. EPA, 1981) make clear that the costs and technological feasibility of
attainment are not to be considered in setting primary or secondary NAAQS.
Such factors can be considered to a degree in the development of state
plans to implement such standards. Under section 110 of the Act, the States
are to submit to EPA for approval State Implementation Plans (SIPs) that
provide for the attainment and maintenance of NAAQS by certain deadlines.
Finally, section 109(d) of the Act directs the Administrator to
periodically review all existing NAAQS and criteria and revise them as
necessary.
-------
B. NATURE OF THE NITROGEN DIOXIDE PROBLEM
N02 is an air pollutant generated by the oxidation of nitric
oxide (NO), which is emitted from both mobile and stationary sources.
At elevated concentrations N02 can adversely affect human health, vegetation,
materials, and visibility. Nitrogen oxide compounds (NOX) also contribute
to increased rates of acidic deposition. Typical long-term ambient
concentrations of NOg range from 0.001 ppm in isolated rural areas to a
maximum annual concentration of approximately 0.06 ppm in one of the
nation's most populated urban areas. Short-term hourly peak concentrations
rarely exceed 0.5 ppm.
A variety of respiratory system effects have been reported to be
associated with exposure to NOg concentrations less than 2.0 ppm in humans
and animals. The most frequent and significant N02-induced respiratory
effects reported in the scientific literature at the time the Criteria
Document and OAQPS Staff Paper were published include: (1) altered lung
function and symptomatic effects observed in controlled human exposure
studies and in community epidemiological studies, (2) increased prevalence
of acute respiratory illness and symptoms observed in outdoor community
epidemiological studies and in indoor community epidemiological studies
comparing residents of gas and electric stove homes, and (3) lung tissue
damage and increased susceptibility to infection observed in animal toxicology
studies. As the Criteria Document concludes, results from these several
kinds of studies collectively provide evidence indicating that certain
human health effects may occur as a result of exposures to N02 concentrations
at or approaching recorded ambient N02 levels.
-------
Only one urban area exceeded the NAAQS in 1982. By 1985, we expect
that all urban areas in the country will be able to maintain the existing
0.053 ppm N02 NAAQS except for this one urban area -- even without adding
new pollution control systems. Thus N02 pollution is not now, nor will it
be in the near future, an extensive problem. It should be noted, however,
that predictions of attainment status depend upon a number of factors and
analytic assumptions including the effectiveness of emission control systems,
particularly those used on private automobiles; the effectiveness of inspection
and maintenance (I&M) programs; the growth in vehicle miles travelled in
urban areas; the growth and mix of industrial sources of NOX emissions; and
the actual air quality levels monitored once all monitors meet EPA's siting,
instrumentation, and quality assurance criteria.
The net annualized cost, after energy savings are considered, of
implementing reasonably available NOX mobile and stationary source controls
to meet the existing 0.053 ppm annual standard in 1985 will amount to
$130-140 million (see Tables 5.1 and 5.3). These expenditures will, depending
on assumptions regarding federal motor vehicle control program (FMVCP)
standards and use of I&M programs, bring all but one urbanized area of
the country into attainment. These costs are in addition to approximately
$1,610-1,910 million per year required for the NOX portion of the FMVCP,
and in addition to almost $110 million per year incurred by industry to
meet NOX new source performance standards (NSPS). FMVCP and NSPS expenditures
are not directly related to a N02 NAAQS, so do not vary with the alternative
ambient standards investigated. (All cost estimates are in constant 1984
dollars.)
%
If attainment is delayed until 1990, annualized NOX control expenditures
directly related to the NOg annual NAAQS drop and all areas attain the
-------
7
0.053 ppm NAAQS. The net annualized cost of instituting reasonably
available controls in 1990 are estimated to range between $0-100 million
in constant 1984 dollars. However, the non-NAAQS related additional
FMVCP and NSPS costs are estimated to be higher because of additional
control equipment required for the NSPS and FMVCP programs for 1990.
Annualized 1990 costs for FMVCP and NSPS are estimated to be $2,640-4,470
and $250 million, respectively.
A draft environmental impact statement (EIS) indicates that controlling
NOX emissions may result in biological, ecosystem, and esthetic benefits
(EPA, 1982d). NOX controls will lower ambient nitrosamine concentrations
and reduce nitrate-based acid precipitation. In urban areas with a high
ratio of non-methane organic compounds-to-NOx concentrations, reducing
NOX emissions should reduce peak ozone levels. However, in urban areas
with a low ratio, reducing NOX emissions without simultaneously reducing
non-methane organic emissions may lead to higher peak ozone concentrations
downwind of major urban areas.
C. NEED FOR REGULATORY ACTION
To reiterate, N02 pollution imposes a cost on society in the form
of health effects and welfare losses. Unless polluters are forced to pay
the full cost of degrading the air, they will use more of the resource
(the assimilative capacity of the air) than is economically efficient.
In this respect, the market for air resources is said to fail.
As will be stated in the next section, non-regulatory approaches to
rectify this problem, while attractive in theory, are deficient in
practice. Voluntary solutions such as negotiations and subsidies do not
guarantee desired results and may in fact exacerbate the pollution
-------
8
problem. Other market-oriented schemes are not presently allowed by the
Clean Air Act.
The need for some control action has been demonstrated and the legal
authority for such action has been established by the Clean Air Act. It
is felt that only through government regulation will the appropriate
steps be taken.
-------
III. ALTERNATIVES EXAMINED
This chapter briefly presents potential alternatives to the proposed
retention of the current N(>2 NAAQS. The outline for the section is
adopted from Executive Order 12291 which requires that at a minimum the
following alternatives be examined:
a) No regulation
b) Regulations beyond the scope of present legislation
c) Alternate stringency levels and implementation schedules
d) Market oriented alternatives
Although Executive Order 12291 requires that all alternatives be examined,
only the most promising ones need be analyzed in detail.
A. NO REGULATION
The option of no regulation could be considered a baseline,
against which the incremental benefits and costs of all. abatement
actions could be compared. This option would essentially leave it up to
the damaged parties to negotiate or litigate a settlement for compensation or
greater pollution control. Realistically, however, there is little or
no incentive for a company to negotiate with individuals to reduce NOX
emissions since, without similar activity by its competitors, the
company would be put at a singular disadvantage. Likewise, litigation is
not a viable course for individuals due to the likely high costs of
proving damages and the reluctance of "free riders" to join class actions.
Given the unlikely success of these approaches in achieving a
socially desirable result, the option of no regulation is not considered
an acceptable option and therefore has not been analyzed further.
-------
10
B. OTHER REGULATORY APPROACHES
Other regulatory approaches include such options as technology
based emission standards and regional air quality standards. Technology
based emission standards are required by the present law in a variety of
forms (e.g. new source performance standards (NSPS) for new and modified
sources, motor vehicle standards, lowest achievable emission rate (LAER)
and reasonably available control technology (RACT) in non-attainment
areas. They typically specify allowable emission rates for specific
source categories, based on technological feasibility, and are designed
to accelerate attainment of the air quality standards. The LAER and
RACT provisions enable economic growth in non-attainment areas, while
encouraging progress toward attainment. NSPS and motor vehicle standards
help to reduce future pollution problems by controlling all new sources.
Although performance and technology based standards are useful in
achieving air quality goals, they cannot substitute for ambient standards
since they are not based on health criteria and cannot guarantee that
society's health objectives will be met. For example, these standards
do not account for local meteorology or the interaction of multiple
sources which could have dramatic effects on air quality.
Differences in terrain and meteorology as well as regional differences
in valuing clean air have been cited as reasons for adopting regional
air quality standards. This approach has not garnered serious attention
because it would require legislative change. Furthermore, there are
other objections. Differences in terrain and meteorology can and are
being considered in the setting of SIP limits. The grave problems of
long range pollutant transport may compromise the effectiveness and
-------
11
equity of a system of regional standards. Also, regional valuations of
air quality ignore the significance of the "existence" and "option"
values of environmental amenities. For example, a resident of New
York City might place value on knowing that air quality in the Grand
Canyon is being protected because it is an important part of the
Country's heritage or because he may opt to visit the Grand Canyon some
day and would like the air to be pristine.
Thus, because the use of these regulatory approaches alone is not
permitted by the present legislation and because they do not respond
to prevailing legislative interest, they are not considered viable
substitutes for the NAAQS and are not analyzed further.
C. REGULATORY ALTERNATIVES WITHIN THE SCOPE OF PRESENT LEGISLATION
EPA believes the cumulative evidence from animal, controlled human
exposure, and community indoor air pollution studies strongly suggests that
NC>2 may cause adverse health effects in sensitive population groups exposed
to NOg at or near existing ambient levels. However, given the uncertainties
existing in the available scientific data, no rigorous rationale can be
offered to support a specific N02 standard.
Two approaches to limit potential health effects associated with NOg
exposure in the ambient air have been considered. The first is to retain
an annual standard at a level between 0.05 ppm and 0.08 ppm to protect
against short-term peak and chronic long-term exposures. A 0.08 ppm standard
would be expected to limit the number of days with hourly peak concentrations
above 0.30 ppm to about ten per year based on an analysis-of existing ambient
air quality data. In most areas of the country attainment of an annual
-------
12
standard of 0.05 ppm should limit the occurrence of 0.30 ppm peaks and
limit hourly concentrations of greater than 0.15 ppm to a range of approximately
10-20 days (some southern California sites may exceed 0.15 ppm on as many
as 40 days). Results from several indoor epidemiology studies suggest that
exposure to repeated peaks up to 0.5-1.0 ppm and possibly as low as 0.15 to
0.30 may be of concern for children.
An annual standard in this range also would provide reasonable assurance
that 1-hour peak concentrations of N02 would not exceed 0.5 ppm, the level
at which mild symptomatic effects have been observed in asthmatics. An
alternative approach is to establish a new multiple exceedance 1-hour
average N02 standard at a level below 0.5 ppm. Such a standard acknowledges
medical evidence suggesting the importance of repeated peak exposures and
would incorporate an allowable rate of exceedance set at a value which
would depend on the standard level.
Either approach can provide a reasonable degree of protection against
repeated peak exposures in the range of 0.15 to 0.30 ppm. A long-term •
standard offers the practical advantage of not requiring formulation and
implementation of a new regulatory program. Establishing a new short-term
standard would require more significant changes in modeling and monitoring
procedures than retention of an annual standard.
D. MARKET-ORIENTED ALTERNATIVES
There are several market-oriented approaches which can be considered
as theoretical alternatives to achieve the NAAQS for N02, none of which
are contemplated by the Clean Air Act. These approaches include pollution
charges, marketable permits, and subsidies and are briefly discussed below.
-------
13
Charges. This policy places a charge (or tax) on each unit of a
pollutant emitted. Firms presumably would choose an abatement strategy
which minimizes their total control cost, including the pollution charge.
Since the charge is uniform, each polluter would control his emissions
until his marginal cost of abatement is equal to the pollution
charge. One difficulty with this strategy is choosing the appropriate
charge level that will lead to the desired air quality.
Permits. In this policy the regulatory agency places a limit
on the amount of emissions that may be released into a particular
region over a specific time period. The Agency then issues and
distributes discharge permits, allowing the holders to emit the amount
of pollution specified by the permits. (The sum of the discharges allowed
by the permits is equal to the emissions cap for the region.) These
permits would be fully transferable so that a market for these rights
would develop. Since the price of the discharge permit could fluctuate
fully, the permits would be alloted efficiently, with those with the
highest marginal control costs bidding the highest prices.
Subsidies. A subsidy system pays sources for each unit of pollution
that they do not emit. The subsidies can take various forms including
lump sum payments, tax and depreciation credits, and low interest loans.
In some minor cases, with a small number of participants, a subsidy
system may produce the desired environmental result. But, by its nature,
a subsidy system invites polluters to pollute heavily, or just threaten
to do so in order to enhance the subsidy. In addition, by making the
polluting activity more profitable and encouraging additional firms
to enter the industry, a subsidy system can actually increase total
emissions even though emissions per polluter may be decreased.
-------
14
Summary. Although the Clean Air Act allows States to use market-
oriented approaches in attaining air quality standards, they are not
legal alternatives to national ambient standards and therefore are not
analyzed further.
-------
15
IV. ASSESSING BENEFITS
A. ASSESSING BENEFITS OF AIR POLLUTION CONTROL
The benefits of air pollution control are the values (both known
and unknown) to society brought about by reducing the health and welfare
effects caused by pollution. The translation of impacts into dollar
values, where possible, allows the summation of impacts and direct
comparison with the costs of pollution control. Estimating the benefits
of air pollution control requires an understanding of how changes in
pollution emissions affect atmospheric conditions and thereby affect
people's health and welfare, and how these effects are valued. These
links are illustrated in Figure 4.1.
Step 1
Step 2
Step 3
Step 4
Changes in
Emissions
-»
Changes in
Ambient Air
Quality
I
— *
Changes in
Health and
Wei fare
\
j
Economic
Valuation
(changes in
well-being
__T
Figure 4.1
STEPS IN MEASURING BENEFITS
The Clean Air Act is designed to prevent adverse effects, such as;
1. Health
adverse impacts to human health—acute or chronic effects
that result in increased mortality or morbidity.
-------
16
2. Welfare
a. adverse impacts to materials—corrosion, soiling and
other damage to building materials, metals, fabrics,
equipment, etc.
b. adverse impacts to vegetation—reduction in productivity
or aesthetic appeal of domestic crops, ornamental plants
c. adverse impacts to animals—effects on health and pro-
ductivity of livestock, pets, and wildlife.
d. adverse impacts on climate—changes in temperatures
and/or precipitation.
e. adverse aesthetic impacts—reduced visibility or visual
discoloration of the air and objects viewed through it
and other aesthetic effects such as unpleasant odors.
These impacts on well-being, including the risks of incurring
them and the costs of avoiding or ameliorating them, determine the
economic value of the changes in air qualityj
Thus, the appropriate measure of the benefits of pollution control
will reflect the sum of the values assigned to improvements in air
quality by all individuals directly or indirectly affected. These
benefits are often classified into three categories—user, option, and
existence values. User values are the benefits of improved air quality
that the individual expects to enjoy directly. Option values result
from uncertainty concerning future demand for air quality. Existence
values result from the mere knowledge that certain environmental conditions
prevail even though the individual may never experience them himself.
The approach used to estimate environmental benefits is derived
from economic welfare theory. The benefits of a change in air quality
11n order to avoid confusion, the term "costs" will refer to the costs
of pollution control while the term "damages" will refer to the adverse
effects of pollution.
-------
17
are said to be equal to the change in well-being, or "utility", that results.
Such a change in utility is difficult to quantify. One technique is to
measure the maximum amount the individual is willing to pay (WTP) in order
to obtain a certain level of environmental quality. A slightly different
measure than WTP for the benefits of a change in air pollution is the
willingness to accept compensation (WTA) in return for incurring an increase
in air pollution or foregoing a decrease in air pollution. The WTP measure
presumes that an individual does not have a predetermined right to any
particular level of air quality, in which case the point of reference is
the lower level of air quality. The WTA measure presumes that the appropriate
point of reference is the improved level of air quality. Since Section 109
of the Clean Air Act authorizes EPA to protect the public from adverse
health and welfare effects of air pollution, it may be argued that WTA is
the appropriate benefit measure.
It is important to note that changes in air quality will affect people
over time. The benefits that accrue to these people must be discounted
to a standard time period to be validly compared with total pollution
control costs. The effects of differences in income levels on benefit
values and the distribution of benefits are non-trivial concerns that
can be incorporated in the benefits analysis.
B. ESTIMATING BENEFITS ASSOCIATED WITH ALTERNATIVE NO? NAAQS
As a result of project time constraints and uncertainties regarding
the effects of N02 exposure on health, vegetation, visibility and other
welfare measures (see Section II), EPA's approach to estimating the benefits
of N02 control focuses on the reductions in the concentration of N02 that
are obtained under different NO2 standards. This information is combined
-------
18
with cost data to estimate the incremental costs per PPM reduction. The
following sections discuss the mechanisms of toxicity, nature of the effects
and sensitive population groups.
1. Mechanisms of Toxicity and Nature of Effects
The mechanisms of toxicity responsible for effects caused by short-term
and long-term exposures to NC>2 are incompletely understood. The variety of
effects, such as (1) increased airway resistance and alterations in lung
hormone metabolism for short-term exposures and (2) increased susceptibility
to infection and morphological damage for long-term or multiple exposures,
may well be explained by related mechanisms of oxidative damage. Because
N0£ is relatively insoluble in water, some fraction normally penetrates
to the distal airways during inhalation. However, the reactivity of NOg
is sufficient to permit chemical interaction and absorption along the
entire tracheobronchial tree.
When NOg enters the lungs, most of the reacting N02 rapidly oxidizes
cellular lipids, although some slowly hydrolyzes to form HN02 and ^03.
The most destructive reaction involves oxidation of unsaturated lipids
of the cellular membrane and results in the formation of peroxidic
products. The disruption of the cellular membrane, which is essential
for maintaining cellular integrity and function, probably accounts for
many of the biological effects (e.g., hyperplasia, morphological damage,
pulmonary edema) which have been ascribed to N02-
Nature of Health Effects. The OAQPS Staff Paper (EPA, 1982a) presents
a detailed and comprehensive assessment by EPA staff of the key health
effect studies contained in the Criteria Document and other critical
scientific issues relevant to the review of the existing annual N02 standard
-------
19
and the need, if any, for a separate short-term (less than 3 hours)
standard. This assessment is summarized in the proposal preamble (49 FR
6866) and draft N02 RIA.
A variety of respiratory system effects have been reported to be
associated with exposure to short- and long-term N0£ concentrations less
than 2.0 ppm in humans and animals. The most frequent and significant
N02-induced respiratory effects reported in the scientific literature at
the time the Criteria Document and OAQPS Staff Paper were published include:
(1) altered lung function and symptomatic effects observed in controlled
human exposure studies and in community epidemiological studies, (2) increased
prevalence of acute respiratory illness and symptoms observed in outdoor
community epidemiological studies and in indoor community epidemiological
studies comparing residents of gas and electric stove homes, and (3) lung
tissue damage, development of emphysema-like lesions in the lung, and
increased susceptibility to infection observed in animal toxicology studies.
As the Criteria Document concludes, results from these several kinds of
studies collectively provide evidence indicating that certain human health
effects may occur as a result of exposures to N0£ concentrations at or
approaching recorded ambient NOg levels.
At the time of proposal, based on controlled human exposure studies,
EPA concluded that human pulmonary function effects of clear health concern
resulting from single, short-term exposures of less than 3 hours duration
have been unambiguously demonstrated only at concentrations (greater than
1.0 ppm) well in excess of ambient exposure levels typically encountered by
the public. More subtle health effects that were of uncertain health
significance, such as mild symptomatic effects, had been reported for some
asthmatics after a single 2-hour exposure to 0.5 ppm.
-------
20
The principal evidence reviewed in the OAQPS Staff Paper and proposal
on the effects of repeated short-term exposures came from a series of
cross-sectional epidemiological (community) studies, some ongoing, which
reported increased prevalence of acute respiratory illness and impaired
lung function in children living in homes with gas stoves (a source of N02)
as compared to children living in electric stove homes. Findings from
several animal studies demonstrating reduced resistance to infection due to
N02 exposures support the belief that N0£ exposures are probably related to
the effects observed in these indoor epidemiological studies. A limitation
of these studies with respect to setting an N02 NAAQS is that the investigators
did not measure short-term N0£ concentrations in the homes of the subjects
in the indoor epidemiology studies. Based on N02 monitoring data from
other gas stove homes, EPA staff estimated that the health effects observed
in gas stove homes, if due to N02 exposure, were likely to be associated
with frequent, repeated short-term peak exposures to N02 levels ranging up
to 0.5 to 1.0 ppm and.possibly as low as 0.15 to 0.30 ppm.
Findings from several animal studies, such as development of emphysema-
like lesions and increased susceptibility to infection, indicated at the
time of proposal that long-term exposures to elevated IMOg concentrations
can lead to serious adverse health effects in animals. A major limitation
in making quantitative use of these studies was the lack of satisfactory
methods for directly extrapolating the results to effect levels in humans.
Since proposal, EPA's ECAO has reviewed the scientific studies that have
become available since CASAC closure on the Criteria Document and OAQPS Staff
Paper and that were identified by EPA staff and/or in public comments on the
N02 proposal. This review was submitted to the CASAC and was discussed at a
meeting held on July 19-20, 1984; a revised document (Grant, 1984) reflecting
-------
21
CASAC and public comments has been placed in the public docket (OAQPS 78-9,
IV-A- ). It should be noted that a more complete scientific assessment of
these studies is not possible at this time because many of the studies have
yet to be published in the peer-reviewed scientific literature or appear
only as abstracts. The principal points from ECAO's review of the new
studies are summarized below.
(1) The more recent controlled human exposure studies (most of which
are presently in unpublished form) present mixed and conflicting results
concerning respiratory effects in asthmatics and healthy individuals at
concentrations in the range of 0.1 to 4.0 ppm N02. Some new studies have
reported an increased effect on airway resistance or lung function when
challenged by a bronchoconstricting agent and N0£ (Ahmed et al., 1982;
Kleinman et al., 1983; Bauer et al., 1984) while other recent studies have
reported no statistically significant effects from N02 alone or .with a
bronchoconstricting agent (Hazucha et al., 1983; Ahmed et al., 1983). It
is not possible, at this time, to evaluate the reasons for these mixed results
Only Kagawa and Tsuru (1979) have reported results possibly suggestive of
short-term N02 effects on pulmonary function without combined provocative
challenge by other agents (e.g., carbachol or cold air) for a group of 6
subjects exposed to 0.15 ppm N02. However, the small size of the decrements
reported (all less than 5 percent) in conjunction with questions regarding
the statistical analyses used suggest caution in accepting the reported
findings as demonstrating N02 effects on pulmonary function at 0.15 ppm,
especially in view of the lack of confirmatory findings by other investigators
at that exposure level.
(2) The most recent indoor epidemiological studies by the British and
Harvard groups indicate somewhat weaker findings of an association between
-------
22
N02 and respiratory effects than the original studies conducted by these
groups cited in the Criteria Document and proposal notice. For example, an
estimated odds ratio for respiratory illness before age 2 of 1.23 (p < 0.01)
previously reported by the Harvard group (Speizer et al., 1980), has been
reduced to 1.12 (p = .07) by the inclusion in the statistical analyses of
data from additional children enrolled in the study (Ware et al., 1984). The
association between residence in a gas stove home and respiratory illness
before age 2 is, therefore, no longer statistically significant. Nonetheless,
the Ware et al. study continued to find small statistically significant
decreases in pulmonary function when the data for this large sample of children
were analyzed.
The associations between use of gas stoves and increased respiratory
illness before age 2 and the use of gas stoves and decreases in lung function
levels in school age children were both reduced when the Harvard group
controlled for parental education (Ware et al., 1984). More specifically,
when an adjustment for parental education was included in the analysis, the
odds ratio for respiratory illness before age 2 was reduced further to 1.11
(p = 0.14) and the decreases in lung function were 30 percent smaller and no
longer statistically significant. Because level of parental education
is negatively associated with the use of gas stoves and positively associated
with respiratory illness and lung function level, the authors state that the
adjustment for parental education "may represent confounding but may also
represent overadjustment for a surrogate for gas stove use" (Ware et al., 1984),
Some other indoor epidemiological studies (with much smaller statistical
power) involving electric and gas stove homes have reported statistically
significant increased rates of symptoms and illness in residents of gas stove
homes (Comstock et al., 1981; Helsing et al., 1982; Lebowitz et al., 1982),
-------
23
while other studies have failed to find any statistically significant
associations with gas stove usage (Jones et al., 1982; Melia et al., 1982;
Melia et al., 1983). Unfortunately, none of the recent studies has provided
an assessment of short-term N0£ levels in the residences of the subjects
evaluated. Overall, then, the newly available data from indoor epidemiological
studies do not appear to resolve the mixed results reported in earlier
studies.
(3) The results from the more recent animal studies further substantiate
the N02 effects on immune function and increased susceptibility to infection.
However, the lack of an acceptable method at this time for quantitative
extrapolation of the animal data to man greatly limits their usefulness
beyond providing qualitative support for analogous effects plausibly being
associated with repeated, short-term high-level and chronic exposure to N02-
2. Sensitive Population Groups
On the basis of the health effects evidence reviewed in the Criteria
Document (EPA, 1983), EPA concludes that the following groups may be
particularly sensitive to low-level N0£ exposures: (1) young children,
(2) asthmatics, (3) chronic bronchitics and (4) individuals with emphysema
or other chronic respiratory diseases. In addition, there is reason to
believe that persons with cirrhosis of the liver or other liver, hormonal,
and blood disorders, or persons undergoing certain types of drug therapies
may also be more sensitive to N02 based on the findings from animal
studies showing increased systemic, hematological, and hormonal alterations
after exposure to N02. Due to the lack of human expertmental data for
these latter groups, however, EPA is considering the potential effects
on such persons only as a factor in providing an adequate margin of
safety.
-------
24
In EPA's judgment, the available health effects data identify young
children and asthmatics as the groups at greatest risk from low-level,
ambient exposures to N02- It has been estimated (U.S. Bureau of the Census,
1973) that in 1970, the total number of children under five years of age
was 17,163,000 and between five and 13 years of age was 36,575,000. Data
from the U.S. National Health Survey (HEW, 1973) for 1970 indicate that
there were 6,526,000 chronic bronchitics, 6,031,000 chronic bronchitics,
6,031,000 asthmatics, and 1,313,000 emphysematics at the time of the Survey.
Although there is overlap on the order of about one million persons for
these three categories, it is estimated that over twelve million persons
experienced these chronic respiratory conditions in the U.S. in 1970.
Table 4.1 presents estimates of the size of the population groups which are
most susceptible to presence of N02 in the atmosphere.
In regard to evidence for the secondary standard, NOX effects on man's
environment, personal comfort, and well-being include impacts on vegetation,
materials, visibility, rates of acidic deposition, and symptomatic effects
in humans. Because acidic deposition is an important and complex problem
associated with multi-pollutant interactions, it is being addressed in a
separate program by the Agency and has not been a specific element of the
N02 standard review. Please refer to the section, Welfare Effects and the
Secondary Standard in Chapter VII for further details.
-------
25
Asthmatics
Emphysematics
TABLE 4.1
SUMMARY OF POTENTIALLY SENSITIVE POPULATION GROUPS
Sensitive
Group
Children
Supporting
Evidence
Children under age 2 exhibit
increased prevalence of respiratory
infection when living in homes with
gas stoves. Children up to age 11
exhibited increased prevalence of
respiratory infections when living
in gas stove homes.
References for
Supporting
Evidence
Speizer et al ,
1980
Melia et al ,
1979
Population
Estimates
age 0-5
17.2 million3
age 5-13
36.6 million3
Asthmatics reacted to lower levels
of N02 than normal subjects in
controlled human exposure studies.
Kerr et al,
1979
Orehek et al,
1976
6.0 million
Chronic
Bronchi tics
Chronic bronchitics reacted to
low levels of NO- in controlled
human exposure studies.
Kerr et al ,
1979
Von Nieding et
al, 1971
Von Nieding et
al, 1970
6.5 million
Emphysematics have significantly
impaired respiratory systems.
Because studies have shown that
N02 impairs respiration by
increasing airway resistance, it
is reasonable to assume that
emphysematics may be sensitive
to N02.
Von Nieding et 1.3 million
al, 1971
Beil and Ulmer,
1976
Orehek et al,
1976
Persons with
Tuberculosis,
Pneumonia,
Pleurisy, Hay
Fever or Other
Allergies
Studies have shown that N02 increases
airway resistance. Persons who have
or have had these conditions may be
sufficiently impaired to be
sensitive to low levels of NO-
Von Nieding et
al, 1971
Beil and Ulmer,
1976
Orehek et al,
1976
unknown
Persons with
Liver, Blood
or Hormonal
Disorders
NOg induces changes in liver drug
metabolism, lung hormone metabolism,
and blood biochemistry.
Menzel, 1980
Miller et al,
1980
Posin et al,
1979
unknown
aU.S. Bureau of Census (1973).
b,
'U.S. Dept. of Health, Education, and Welfare (1973).
-------
V. COST ANALYSIS
A. INTRODUCTION
This chapter emphasizes the direct principal, or real-resource,
costs associated with controlling emission sources of nitrogen oxides
(NOX) air pollution in order to attain alternative nitrogen dioxide
(N02) national ambient air quality standards. The costs discussed are
for two years--!985 and 1990. The costs are in constant March 1984
dollars unless otherwise noted. Cost estimates are derived from an EPA
report entitled Cost and Economic Assessment of Regulatory Alternatives
for NO? NAAQS (DRAFT) (EPA, 1982c), as updated by Appendix B of this report.
The above report contains a number of detailed cost and economic analyses
that are not reproduced here. These include an investigation of the impacts
of setting alternative NOg standards on specific population income groups
(all income categories), industrial sectors, local governments, and small
businesses. Major findings from these specialized studies are:
1. no particular segment of the population is forced to pay a
disproportionate amount of the total cost of mobile source
control. Therefore, there are no significant adverse income
distribution impacts associated with any of the alternative
N02 standards. Repair costs resulting from an N02 inspection
and maintenance program, however, are a higher percentage of
income for low income groups than for middle and high income
groups.
2. while "distressed cities" are definitely affected by various
NOg control programs, the probable overall impact on local
government finances is considered to be negligible. The prime
-------
27
reason for this conclusion is that most of the direct costs of
N0£ control programs are borne by motor vehicle owners and
users, not by local governments. Inspection and maintenance
programs are self-supporting, paid for by user fees; transportation
control measures are largely paid for by the federal government
under a U.S. Department of Transportation program intended to
reduce energy consumption associated with motor vehicle use.
3. small businesses are probably not adversely affected by an
N02 control program. While a small number of NOX emitting
sources require controls to attain a 0.053 ppm NAAQS (172 out
of the 29,000 that were investigated), none of the source
categories affected contains many small businesses. The
affected categories are oil and gas extraction (SIC 13),
petroleum and coal products (SIC 29), primary metals.(SIC 33),
and public utilities (SIC 49).
4. financial capital is generally available for those relatively
large industrial sources that need to install NOX pollution
control equipment, in the sense that major shifts are not
anticipated in the firm's opportunity cost of capital. For
all but one industrial category in the country, NOX capital
expenditures are less than 1% of projected capital expenditures.
Even in that category, oil and natural gas extraction, NOX
capital expenditures are estimated to be only 1.5% of projected
expenditures.
5. for affected industries and plants, controls associated with
the alternative standards are judged to be economically
-------
28
affordable. No significant product output adjustments are
anticipated in response to product price changes, which should
be quite small for all alternative standards (less than 0.02%
at most).
The interested reader is referred to the detailed report (EPA, 1982c)
for more information on the items discussed above.
As discussed in the above reference, the methodology used to estimate
regulatory impacts associated with alternative N02 standards is highly
sensitive to the value used to represent air quality in the base year
(the so-called design value). A small change in this value can substantially
alter non-attainment area projections and the amount of control theoretically
needed in a particular location. The NOX control cost estimates are
also greatly affected by the nature and timing of the program assumed
for an area. In all cases, a cost effectiveness procedure was used to
determine what mix of NOX controls, such as inspection and maintenance
(I&M) programs and transportation control measures (TCM), would be used
in that area.
The control measures investigated in the N02 cost analysis were all
"reasonably available," meaning that the techniques are currently operational
and are expected to be commercially available in the time period analyzed.
For stationary sources, these included suppression of NOX formation and
removal of NOX from stack gas. Suppression of NOX formation is most
effective with large combustion sources. Suppression mechanisms include
alteration of operating conditions by modifying the fuel or air supply, by
lowering combustion intensity or temperature, or by combining these techniques.
Removal of NOX from combustion stack gases can be accomplished by ammonia
injection, which reduces NOX to elemental nitrogen. Removal of NOX from
-------
29
stack gases is most effective with noncombustion sources, chiefly chemical
manufacturing industries. Removal techniques include catalytic reduction
with wet chemical scrubbing extended and chilled absorption, and molecular
sieve adsorption. (See Appendix C for more information on stationary
source controls.)
NOX control measures used for mobile sources include an exhaust gas
recirculation (EGR) valve and associated equipment, an open-loop oxidation
catalyst system, and a three-way plus closed-loop oxidation catalyst
system, for analyses of 2.0, 1.5, and 1.0 gram per mile FMVCP emission
standards, respectively.
It should be recognized that NOX control programs developed in this
report, and the cost analyses upon which it is based, are hypothetical.
Actual control programs are developed by State and local air quality
management agencies in the state implementation planning process. Actual
national costs of NOX control programs will be the sum of many individual
state decisions, and of course may differ from those developed for this
regulatory impact analysis. In addition, this analysis does not "force"
attainment of any of the alternative NAAQS investigated. As mentioned, the
hypothetical national control program only uses reasonably available control
measures. If an area cannot attain an analyzed alternative NAAQS using
these measures, the fact is so noted. No attempt is made to develop a more
stringent program using NOX control techniques under development or being
proposed. Cost and efficiency data do not exist for these new techniques
in sufficient detail to incorporate them in the least-cost program used to
calculate national NOX control costs. Finally, in this national-level
analysis, it is frequently necessary to use national averages for important
factors such as emissions growth rate.
-------
30
B. NATIONAL CONTROL COSTS
National NOX control costs are presented for the current 0.053 ppm
annual average and for two alternatives: 0.06 and 0.07 ppm. The cost
estimates for reducing nitrogen oxides emissions to meet, or approach as
closely as possible, each alternative NAAQS standard by 1985 or 1990 are
presented below. A number of estimates are provided in order to cover
alternative motor vehicle emission standards currently being discussed by
Congress, and to account for whether or not NOX inspection and maintenance
programs will be installed in major urbanized areas.
Two methodologies are used in the analysis (EPA, 1982c) for estimating
nationwide control costs. They are based on current knowledge of N0£ formation,
especially in situations likely to cause high observed NOg concentrations.
The situations modeled reflect different types of N02 episodes. One
uses linear roll-back and includes only emissions from mobile and stationary
area sources in the emission inventory. This approach accounts for
suppressed mixing episodes where an inversion limits vertical mixing -and
where point sources with stack heights above the inversion layer do not
affect ground-level concentrations. Ambient concentrations of N02
during these episodes are likely to be recorded by the existing network
of N02 monitors. The second methodology accounts for situations where
emissions from point sources can cause high N0£ levels at the point
where the plume hits the ground. Ground-level N0£ concentrations for
each significant source in the National Emissions Data System (NEDS)
point source file of NOX emitters are estimated with a dispersion model,
considering source interaction within plants. A modified ozone-limiting
approach is used to translate NOX into N02 (Cole and Summerhays, 1979).
-------
31
In either case, the observed or calculated N02 annual average value
is compared with the alternative NAAQS under investigation to determine
how much reduction in NOX emissions is required to attain the NAAQS and
an integer linear program is used to determine the most cost effective
control stategy.
The least-cost set of control strategies is selected using estimated
annualized costs to industries or to owners of area and mobile NOX sources.
However, costs presented in EPA (1982c) and in this report are social
costs, or close proxies of them representing aggregrate before-tax costs.
EPA (1982c) presents in detail how the two cost indices are related.
Appendix B of this report describes how the control costs in EPA (1982c)
were updated using more recent NOX data, new mobile source emissions
estimates, and 1984 dollars.
There are other NOX control costs incurred by industry and the public
that do not vary with choice of a NC>2 ambient standard. These costs are
incurred to meet (1) the Federal Motor Vehicle Control Program (FMVCP) for
automobile and truck NOX controls, as governed by Title II of the Clean Air
Act, and (2) new source performance standards (NSPS) for certain new large
point sources of NOX emissions, as governed by section 111 of the Act. The
major categories of point, or stationary, sources affected by a NOX NSPS
are: utility boilers, coal-fired and oil- or gas-fired, industrial boilers,
stationary gas turbines, reciprocating internal-combustion engines, and
nitric acid plants.
For the sake of completeness, both EPA (1982c) and this report
present FMVCP and NSPS control costs even though they are not affected
by level of the N0£ standard. As will be seen, these costs—particularly
FMVCP costs—are much higher than the incremental cost incurred to just
-------
32
attain alternative ambient N0£ standards. Appendix B revises and adds
to the FMVCP costs in EPA (1982c).
Estimates of national NOX control costs appear in Tables 5.1-5.6.
Tables 5.1-5.3 are for 1985 annualized cost estimates, for the 1, 1.5, and
2 gpm FMVCP alternatives respectively. Tables 5.4-5.6 depict 1990 annualized
costs for the same FMVCP alternatives. The estimates are in constant March
1984 dollars rounded to the nearest ten million dollars. Finally, all
point source cost estimates are based on using a 0.01 ppm NOg background
value in the point source modeling analysis. EPA (1982c) includes results
using a zero N02 background value. If such a value is used, there would be
no point source costs in the Tables shown in this report, but all other
costs remain the same.
Also depicted in the Tables are the number of non-attainment urban
areas remaining after all reasonably available control measures have
been applied. As described in Appendix B, all areas in the country, can
attain the current 0.053 ppm standard (and all higher alternatives) in both
1985 and 1990, except for one area where the 0.053 ppm value cannot be
attained in 1985 without instituting a NOX I&M program.
As can be readily seen, the vast majority of NOX control costs are
due to the FMVCP and NSPS programs, which are not directly tied to the
ambient N0£ standard. The incremental costs associated with various air
standards, then, are those attributed to I&M, TCM, and stationary and
area source programs, which might be needed in addition to FMVCP and
NSPS controls. While the incremental costs of going from one standard to
another can be determined simply by subtracting total NOX control costs for
any consistent set of I&M and FMVCP conditions, the result is difficult
to interpret since the non-attainment situation differs for the various
-------
33
Table 5.1
NATIONAL ANNUALIZED 1985 RACT/RACM CONTROL
COSTS1 FOR THREE ALTERNATIVE N02 NAAQS,
ASSUMING A 1 GPM FMVCP
($106)a
Alternative Annual Average N02 NAAQS
ppm
0.053
Q.Q6Q
0.070
Control
Strategies
FMVCP
Additional Mobile/Area
Point Sources'3
NSPSC
Total NOX Control Costs
Number of Remaining
Non-Attainment Areas
With
I&M
1,910
120
10
110
2,150
0
No
I&M
1,910
180
10
110
2,210
1
With
I&M
1,910
0
_d
110
2,020
0
No
I&M
1,910
0
_d
110
2,020
0
With
I&M
1,910
0
0
110
2,020
0
No
I&M
1,910
0
0
110
2,020
0
Abbreviations: RACT: Reasonably Available Control Technology
RACM: Reasonably Available Control Measures
NO? : Nitrogen Dioxide
NAAQS: National Ambient Air Quality Standard
GPM : Gram per mile
FMVCP: Federal Motor Vehicle Control Program
PPM : Parts per million
I&M : Inspection and Maintenance Program
Notes: aAll costs are rounded to the nearest ten million dollars. The
estimates are in constant March 1984 dollars.
bThe point source figures are those obtained from using a N02
background value of 0.01 ppm in the modeling analysis. A 0.0 ppm
value was also used; point source costs are zero in all cases if a
zero background N02 level is assumed.
CA range of costs are presented in EPA (1982) for this item but have
been reduced to the mid-value for purposes of this presentation.
dLess than $5 x 106, which is rounded to zero.
Source: U.S. EPA, 1982c. Cost and Economic Assessment of Regulatory
Alternatives for NCU NAAQS (DRAFT). Research Triangle Park, N.C.,
and Appendix B.
-------
34
Table 5.2
NATIONAL ANNUALIZED 1985 RACT/RACM CONTROL
COSTS1 FOR THREE ALTERNATIVE N02 NAAQS,
ASSUMING A 1.5 GPM FMVCP
($106)a
Alternative Annual Average
ppm
0.053
Q.Q6Q
NAAQS
0.070
Control
Strategies
FMVCP
Additional Mobile/Area
Point Sourcesb
NSPSC -
Total NOV Control Costs
With
I.&.M.
1,640
140
10
110
1,900
No
I&M
1,640
180
10
110
1-940
With
I&M
1,640
30
_d
110
1,780
No
I&M
1,640
30
_d
110
1,780
With
I&M
1,640
0
0
110
1,750
No
I&M
1,640
0
0
110
1,750
Number of Remaining
Non-Attainment Areas
Abbreviations: RACT: Reasonably Available Control Technology
RACM: Reasonably Available Control Measures
N02 : Nitrogen Dioxide
NAAQS: National Ambient Air Quality Standard
GPM : Gram per mile
FMVCP: Federal Motor Vehicle Control Program
PPM : Parts per million
I&M : Inspection and Maintenance Program
Notes: aAll costs are rounded to the nearest ten million dollars. The
estimates are in constant March 1984 dollars.
bThe point source figures are those obtained from using a N02
background value of 0.01 ppm in the modeling analysis. A 0.0 ppm
value was also used; point source costs are zero in all cases if a
zero background N02 level is assumed.
CA range of costs are presented in EPA (1982) for this item but have
been reduced to the mid-value for purposes of this presentation.
dLess than $5 x 106, which is rounded to zero.
Source: U.S. EPA, 1982c. Cost and Economic Assessment of Regulatory
Alternatives for N02 NAAQS (DRAFT). Research Triangle Park, N.C.,
and Appendix B.
-------
35
Table 5.3
NATIONAL ANNUAL IZED 1985' RACT/RACM CONTROL
COSTS1 FOR THREE ALTERNATIVE N02 NAAQS,
ASSUMING A 2 GPM FMVCP
($106)a
Alternative Annual Average
ppm
O.Q53
0.060
NAAQS
O.Q7Q
Control
Strategies
FMVCP
Additional Mobi
Point Sourcesb
NSPSC
le/Area
With No
I&M I&M
1,610 1,610
140 180
10 10
110 110
Total NOX Control Costs 1,870 1,910
Number of Remai
Non-Attai nrnent
Abbreviations:
ning
Areas
RACT:
RACM:
N02 :
NAAQS:
GPM :
FMVCP:'
PPM :
I&M :
0 1
Reasonably Available
Reasonably Available
Nitrogen Dioxide
National Ambient Air
Gram per mile
With
I&M
1,610
30
_d
110
1,750
o
Control
Control
Quality
Federal Motor Vehicle Control
Parts per million
No
I&M
1,610
30
_d
110
1,750
o
Technology
Measures
Standard
Program
With
I&M
1,610
0
0
110
1,720
n
No
I&M
1,610
0
0
110
1,720
o
Inspection and Maintenance Program
Notes: aAll costs are rounded to the nearest ten million dollars.
estimates are in constant March 1984 dollars.
The
The point source figures are those obtained from using a N02
background value of 0.01 ppm in the modeling analysis. A 0.0 ppm
value was also used; point source costs are zero in all cases if a
zero background N02 level is assumed.
CA range of costs are presented in EPA (1982) for this item but have
been reduced to the mid-value for purposes of this presentation.
dLess than $5 x 106, which is rounded to zero.
Source: U.S. EPA, 1982c. Cost and Economic Assessment of Regulatory
Alternatives for NQ2 NAAQS (DRAFT). Research Triangle Park, N.C.,
and Appendix B.
-------
36
Table 5.4
NATIONAL ANNUALIZED 1990 RACT/RACM CONTROL
COSTS1 FOR THREE ALTERNATIVE N07 NAAQS,
ASSUMING A 1 GPM FMVCP
($106)a
Alternative Annual Average N02 NAAQS
ppm
0.053
0.060
Q.Q7Q
Control
Strategies
FMVCP
Additional Mobile/Area
Point Sources'3
NSPSC
Total NO Control Costs
With
I&M
4,470
0
_d
250
4,720
No
I&M
4,470
0
_d
250
4,720
With
I&M
4,470
0
_d
250
4,720
No
-I&M
4,470
0
_d
250
4,720
With
I&M
4,470
0
0
250
4,720
No
I&M
4,470
0
0
250
4,720
Number of Remaining
Non-Attainment Areas
Abbreviations: RACT: Reasonably Available Control Technology
RACM: Reasonably Available Control Measures
NOo : Nitrogen Dioxide
NAAQS: National Ambient Air Quality Standard
GPM : Gram per mile
FMVCP: Federal Motor Vehicle Control Program
PPM : Parts per million
I&M : Inspection and Maintenance Program
Notes: aAl1 costs are rounded to the nearest ten million dollars. The
estimates are in constant March 1984 dollars.
The point source figures are those obtained from using a N02
background value of 0.01 ppm in the modeling analysis. A 0.0 ppm
value was also used; point source costs are zero in all cases if a
zero background N02 level is assumed.
CA range of costs are presented in EPA (1982) for this item but have
been reduced to the mid-value for purposes of this presentation.
dLess than $5 x 10^, which is rounded to zero.
Source: U.S. EPA, 1982c. Cost and Economic Assessment of Regulatory
Alternatives for NOo NAAQS (DRAFT). Research Triangle Park, N.C.,
and Appendix B.
-------
37
Table 5.5
NATIONAL ANNUALIZED 1990 RACT/RACM CONTROL
COSTS1 FOR THREE ALTERNATIVE NOo NAAQS,
ASSUMING A 1.5 GPM FMVCP
($106)a
Alternative Annual Average N02 NAAQS
ppm
0.053
0.060
Q.Q7Q
Control
Strategies
FMVCP
Additional Mobile/Area
Point Sourcesb
NSPSC
Total NOX Control Costs
Number of Remaining
Non-Attainment Areas
With
I&M
2,850
0
_d
250
3,100
0
No
I&M
2,850
0
-d
250
3,100
0
With
I&M
2,850
0
_d
250
3,100
0
No
I&M
2,850
0
_d
250
3,100
0
With
I&M
2,850
0
0
250
3,100
0
No
I&M
2,850
0
0
250
3,100
0
Abbreviations: RACT: Reasonably Available Control Technology
RACM: Reasonably Available Control Measures
NOo : Nitrogen Dioxide
NAAQS: National Ambient Air Quality Standard
GPM : Gram per mile
FMVCP: Federal Motor Vehicle Control Program
PPM : Parts per million
I&M : Inspection and Maintenance Program
Notes: aAll costs are rounded to the nearest ten million dollars.
estimates are in constant March 1984 dollars.
The
The point source figures are those obtained from using a NOo
background value of 0.01 ppm in the modeling analysis. A 0.0 ppm
value was also used; point source costs are zero in all cases if a
zero background N0£ level is assumed.
CA range of costs are presented in EPA (1982) for this item but have
been reduced to the mid-value for purposes of this presentation.
dLess than $5 x 106, which is rounded to zero.
Source: U.S. EPA, 1982c. Cost and Economic Assessment of Regulatory
Alternatives for NOo NAAQS (DRAFT). Research Triangle Park, N.C.,
and Appendix B.
-------
33
Table 5.6
NATIONAL ANNUALIZED 1990 RACT/RACM CONTROL
COSTS1 FOR THREE ALTERNATIVE N02 NAAQS,
ASSUMING A 2 GPM FMVCP
($106)a
Alternative Annual Average N02 NAAQS
ppm
0^053
O.Q6Q
Q.Q7Q
Control
Strategies
FMVCP
Additional
Mobile/Area
Point Sourcesb
NSPSC
Total NOX
Control Costs
Number of Remaining
Non-Attainment Areas
Abbreviations: RACT:
RACM:
N02 :
NAAQS:
GPM :
FMVCP:
PPM :
I&M :
With No
I&M I&M
2,640 2,640
10 100
_d _d
250 250
2,900 2,990
With
I&M
2,640
0
_d
250
2,890
No
I&M
2,640
0
_d
250
2,890
0000
Reasonably Available Control Technology
Reasonably Available Control Measures
Nitrogen Dioxide
National Ambient Air Quality Standard
Gram per mile
Federal Motor Vehicle Control Program
Parts per million
Inspection and Maintenance Program
With
I&M
2,640
0
_d
250
2,890
0
No
I&M
2,640
0
_d
250
2,890
0
Notes: aAll costs are rounded to the nearest ten million dollars.
estimates are In constant March 1984 dollars.
The
DThe point source figures are those obtained from using a N0£
background value of 0.01 ppm 1n the modeling analysis. A 0.0 ppm
value was also used; point source costs are zero 1n all cases if a
zero background N02 level is assumed.
CA range of costs are presented in EPA (1982) for this item but have
been reduced to the mid-value for purposes of this presentation.
Less than $5 x 10^, which is rounded to zero.
Source: U.S. EPA, 1982c. Cost and Economic Assessment of Regulatory
Alternatives for NOo NAAQS (DRAFT). Research Triangle Park, N.C.,
and Appendix B.
-------
39
cases. Thus, there is no consistent basis for comparison. The incremental
control costs for 1985 and 1990 along with the number of remaining non-
attainment areas, are presented in Tables 5.7 and 5.8.
The present value (PV) of future expenditures made by individuals and
business to control NOx emissions can be approximated by applying the PV
formula to the cost estimates provided in Tables 5.1 to 5.6. The formula is:
PV = C (1 + r)-t
where: PV = present value of a future expenditure
C = cost of the future expenditure
r = annual interest rate
t = elapsed time in years
For our purposes, r=10% and t=l for 1985 and t=6 for 1990. The resultant
discount factors therefore are: 0.909 for 1985 and 0.565 for 1990. Applying
these factors to the cost estimates provided for 1985 and 1990 results in
the following present value cost estimates in millions of dollars for the
"with I&M" case.
Alternative NAAQS Standards
(ppm)
0.053 0.060 0.070
1 gpm FMVCP
1985
1990
1.5 gpm FMVCP
1985
1990
2.0 gpm FMVCP
1985
1990
1,950
2,670
1,730
1,750
1,700
1,640
1,840
2,670
1,620
1,750
1,590
1,630
1,840
2,670
1,590
1,750
1,560
1,630
Again, most of these present value expenditures are due to FMVCP and NSPS
costs that are not affected by a national ambient air quality standard.
-------
40
C. TRANSACTION COSTS (SECONDARY COSTS)
According to Office of Management and Budget guidance on performing
regulatory impact analyses, a complete cost analysis of major governmental
actions should contain an evaluation of certain secondary costs associated
with that action. These include governmental regulatory costs, adjustment
costs for unemployed resources, and costs associated with adverse effects
on market structure, innovation, and productivity. Most of these items
were only cursorily investigated in the cost and economic assessment
(EPA, 1982c) underlying this RIA. Consequently, they are only discussed
qualitatively below.
1. Governmental Regulatory Costs
Both federal and state agencies are involved in establishing,
implementing and enforcing the national ambient air quality standards.
EPA provides scientific and technical analyses in developing, reviewing,
and setting air standards. It also provides technical support in developing
and maintaining computerized national data systems that store ambient
air quality and emissions data. In addition, EPA provides technical
guidance to the states in operating and reporting ambient monitoring
networks, developing state implementation plans (SIPs), and undertaking
enforcement actions against certain violators of the state's own rules
and regulations.
It is extremely difficult to separate out that part of the
total costs for these functions attributable to N02, since it is only
one of six NAAQS and numerous other pollutants that are more or less
-------
41
oo
•P
c
ai
r— C
03 CD-r-
c: c (O
O T- 4->
•r- C +->
+-> T- =C
•r- rt3 I
-o E c
-a CD o
-
in
(O
OJ
£_
fO
oo
en
tO •
C '
o —•
T3 (/)
13 O
< O
o o
OO l£5
o o
00 CM
C
O
o
O)
(_
(T3
OJ
r- E
OO Q.
t— CM CD
OO O —
O Z
'O T3
UJ t_
_j z ro
O O T3
o: ai c
I— 2: uj «
z o IE 4-> uo
00:1 — - — - oo •
O U_ O to t-H
Z O Q.
Q CU << rH O
P-. UJ Z <^t >
• M i— ' O 2!
ID i— i iD 1 — «d- U_
_i z co
r- •=> 31 CT r-H
£i z o <:
rO Z =^ c-
1— at -T Z o
1— £_
I_O 1 — 1 1 IO
CO 3 < 21
cn rD —
r— O Z
UJ Z
_l 1—
•r- C +->
•t— (O t to
ID E C OJ
T3 O) O C_
< a: z <:
f.K
^j ^— ^
c *^
o ^-^
+J t/)
•r* +J
13 10
-o o
•i- C 4->
4-> T-
T3 t/1
"U O
^C. CJ
E
o
T3 -~~ t_
T3 Q_
-a o. cr>
C — ' C
fO *r-
40 O
OO CJ
O r-t O O
T3 E
O O O O t- ns
00 to 00 CM (O t_
*~"l r— < "O CD
C O
> "o
•r- 4O
i — C
1 t- r—
1 -r- 0
1 O t-H O O e -o c •>• -r-
•r- Ol r—
°B i—1 X -I— H) t- i—
O .Q ^— O *r*
) — t _C -r- E «p- 4J S
4J Q «5 E 0
O -r- S £-
z 3 c i— t_ ai
ai 2:
O < Q- 2: O.
Z Z CJ3 U_ 0-
O oo o oo
to LO to LO • •
O O O O
O O O O OJ
r*** VJD ^- vo 5—
O O O O J2
• • • • o
00 00 <
E
t_
CD
o
I_
Q-
O)
o
"3
c
o
ex
(/)
(3
H— I
2:
=3
1 — t
fO
E
•r-
4->
Ul
0)
ID
JC
1—
•
C/l
fO
r—
1— -
o
c
o
•f—
r—
r_
•r™
E
c
OJ
1J
(/>
0)
(O
O)
c
OJ
JC
4->
0
T3
0)
-a
c
3
O
I_
O)
£_
fO
4->
(/)
O
o
l~"
<
, — 1
O)
o
c
i *
o
o
u_
•
oo
13
C
fO
^
CM
«i
i-H
10
o
r-.
jQ
h^
(U
u
t_
3
O
oo
-------
42
C/l
J_
n3
C\J
i— C
(& t^ft *p~
C C (O
O -r- 4->
•r- C 4->
•r- to I IO
"O S C u">
o a: i— -^ oo •
o u_ o ^o I-M
Z 0 Q.
oo uj z v* >
• M H-> o s:
Lf) >— < CD 1— *3" it
_l Z CO
O> =£
r— 0 I CT .-H
J2 Z 0 <
fO 21 ^^ C~
i— < n: z o
I— s_
O — < _l (B
cn 3 < s
O^ ^3 -— *•
i— Q Z
UJ "ZZ.
—II — «C
< <
1 — •— 1
Z 0
uj o
SI 00
UJ OO
ac. «c
o
z
1— 1
o
•
t— 1
C *^"
o -~-
•r-
•i- -»->
T3 1/1
-o o
«=C 0
4-3
C
E
r^ C
IT3 Ol-r-
C C (0
O -i- 4-5
•I— C 4->
4-> T-
"^3 ^
•a o
(/)
•r- 4->
•^ ^
•o o
O
00 CD
O 0 O 0
0 <->
1— 1
o o o o
T3 S
O O O O t_ fO
(0 C-
•O 01 E
C O (13
> O Q-
4-> i-
•r- 4J Ol
i — co
(O O C
3 CJ (0
o- c
0) 0)
1 t_ r— 4->
1 T- <_> C
1 O O O O «=C -i— -r-
i Z J= c («
OJ 4-> -i-
aO >— i X -r- 0> «- i— C
O jQ f™" O T— rt3
i— ' -C -f- 6 -r- 4-> E
4-> O ^ S O C
0 -i- 2! t- 0
Z 3 C r- {_ 0) -r-
oj ITS
1 1 O^ C QL fO O
1 1 O O t- 1/1 (1)
oo oo t-'i-Eai+JCL
4-> 4-> *O "O t» 1/1
•r- IO t- O> (O C
Z Z CJ3 U_ Q- — •
oo a.
O" C_3
CM<: s > 2: s
O ^ CL Z Q- °3
Z Z CD U. Q. t-i
O oo o oo
^ •'A. ^
i I II •?
O O O O 0)
r*** to r^» "-O ^
o o o o _a
* • • • o
O 0 00 <
c:
o
o
c
*r~
OJ
c_
Ul
ai
4J
n3
g
1 ^
l/l
O)
QJ
g-
•
(/I
t_
r^
^
o
T3
p—
o
^.
r^
• r—
E
c
<1J
4->
1 %
1/1
OJ
t.
ns
O)
c
O)
.C
4->
o
[ *
-a
a>
•a
c
3
O
c_
O)
(13
4->
l/l
O
o
jm
<
i — i
4->
O
c
4->
O
O
U-
•
l*O
-a
03
««
Lf)
A
<=!-
to
QJ
c*i
(B
1 —
, ,
O)
0
(_
zs
o
oo
-------
43
handled simultaneously by EPA in performing the functions noted above.
However, EPA's costs to review the N02 NAAQS and promulgate a revised
standard can be estimated. The items covered to date include: (1) scientific
review of health literature and development of the N02 Criteria Document
and Staff Paper, (2) review of existing ambient data to determine potential
problem areas, (3) development of a set of control strategies, which are
then costed out, (4) analysis of the economic impacts associated with these
costs, (5) analysis of exposures to alternative N02 standards, (6) development
of the regulatory package for proposal, and (7) internal review of the
various regulatory documents. EPA's estimated total costs for N02 for
these items are:
1. 16 person years of staff time.
2. $1.5 million of contract funds.
3. $120,000 of computer time.
This comes to a total of approximately $2.5 million (in 1984 dollars). It
is a one-time cost* that occurs over 3 fiscal years. This estimate
does not include any costs of maintaining N02 air quality or emissions
inventory data. It also does not include any guidance activities costs.
The problem of costing these activities is difficult as data do not
exist to exactly apportion total costs of these activities to N02-
However, they are estimated to be an order-of-magnitude lower than the
cost incurred to review and set the standard.
States, and, if they so delegate, local governments, actually implement
Clean Air Act provisions directed toward attaining and maintaining NAAQS.
*Since section 109 of the Clean Air Act requires that all NAAQS be reviewed
every five years, it will be a repetitive one-time cost.
-------
44
Their major implementing functions include operating air monitors and
reporting ambient air quality data (required under sections 110 and 317 of
the Act), developing SIPs (required under section 110), and enforcing these
plans. For the most part, these are labor-intensive activities. The
Economic Impact Assessment of the lead NAAQS (EPA, 1978b) provides data
that can be used to estimate roughly the costs to state and local governments
in implementing the promulgated NOg NAAQS. That report suggests that these
annual costs would be on the order of 50 person-years of effort plus $1.5
million in monitoring equipment and other capital costs.1 State/local
control costs, then, would be around $6.8 million (in 1984 dollars) annually.
Total annual governmental N02 control costs are derived simply by
adding the annualized one-time standard-setting costs with other miscellaneous
federal costs (assumed to be $0.5 million) and adding this to the $6.8
million state/local estimate. Total annual costs obtained in this manner
for a five-year review cycle are approximately $7.8 million.
2. Adjustment Costs for Unemployed Resources
Three types of costs occur if a regulatory action results in
unemployment of labor or underuse of other productive resources. These
are: loss of value in the resources required to produce lost output,
resource real location costs (such as moving costs), and unemployment transfer
payments.
EPA does not believe any significant unemployment impacts are
associated with the N02 NAAQS being reviewed (see EPA, 1982c). In only one
major industrial sector does the estimated initial investment cost for N0£
control equipment exceed one percent of the usual annual capital budget of
iThese estimates apply to on-going activities and not to one-time events
associated with just implementing a new NAAQS.
-------
45
the industrial sector. Capital costs in that sector, extraction of natural
gas, amount to slightly less than 1.5% of annual capital outlays. The
effects on capital availability, product price, and import substitution
are also minor. In addition, no plant closings are predicted due to the
installation of N0£ controls to meet any of the annual average N02 NAAQS
investigated.
3. Adverse Effects on Market Structure
As just mentioned, none of the proposed N02 NAAQS has a significant
impact on product price, capital availability, or import substitution.
Likewise, their impacts on corporate debt and debt/equity ratios are small.
At most, a N02 NAAQS will add 0.020 percent to the cost of producing goods.
Also, in the U.S. census region containing an urban area where N02 controls
are theoretically needed to attain the 0.053 NAAQS, industrial area source
control costs represent less than 0.5 percent of value added by the manufacturing
sector.
Lacking a distribution of vehicle ownership, it is not possible to
estimate accurately the impact of I&M costs on the commercial and industrial
sectors. However, the average per vehicle cost of I&M for a failed vehicle
- approximately $23 - is not considered a significant increment in the
annual cost of vehicle operation for either the commercial or industrial
sectors. Compared to receipts for the commercial and industrial sectors,
the entire cost for I&M does not constitute even 0.1 percent of sales plus
value added in the area where costs are incurred. Thus, there will probably
be little or no impact on market structure in the industrial sectors affected
by NOX emission controls needed to attain alternative N02 NAAQS.
-------
46
VI. EVALUATING BENEFITS AND COSTS
A. INTRODUCTION
Since it was felt that a credible attempt could not be made to estimate
the benefits of various N02 ambient standards within project constraints,
and therefore a quantitative benefit-cost assessment could not be done,
analyses of the incremental costs to reduce potentially adverse concentrations
of N02 (hereafter referred to as ICAs) were performed instead. These
analyses concentrated on two of the alternative annual standards studied in
the preceding national cost analysis, namely 0.053 ppm and 0.060 ppm. The
object of the ICAs was to determine the incremental control costs and the
incremental N02 concentrations resulting from more stringent N0£ standards.
In so doing, estimates could be obtained of the incremental costs per
concentration reduction between standards.*
While this information does not reveal which standard would be best in
an economic sense—only the cost, and not the "benefit", of reducing adverse
levels of N02 has been estimated—it does provide a partial measure of the
relative impacts of the alternative N02 standards.
B. Framework and Scope of the Incremental Cost Analysis
This section discusses the procedures, assumptions, and scenarios that
were used in the ICAs.
1. Cost Analysis Procedures and Assumptions
The N02 control cost estimates were derived by analyzing only the
requirements each standard imposed on area and mobile sources since major
The original RIA used reduced people exposures as a measure of effect in
the ICA. This ICA could not use this preferred effects measure because of
time and budget constraints.
-------
47
stationary point sources were not considered a significant contributor to
monitored ambient air quality (See Chapter V).
Baseline NOg concentrations were calculated assuming that several
different N02 Federal Motor Vehicle Control Program (FMVCP) standards were in
effect. In areas where baseline concentrations exceeded the standard,
additional controls were applied in order of relative cost effectiveness
to bring the areas into compliance. The controls considered included area
source controls on commercial boilers and small industrial boilers, automobile
inspection and maintenance (I&M) programs, and transportation control
measures (TCM). Residential heating, though more cost effective than some
of these controls, was only applied as a last resort to achieve attainment.
2. Concentration of NOg as Effects Measure—Procedures and Assumptions
Given time and budget constraints, the only effects measure
available for use in the ICA was the design value level of forecast
air quality—the annual average N02 concentration in PPM—used in the cost
analysis under baseline and alternative regulations. The incremental change
in the average annual concentration of N02 was thus used as the measure of
effect when moving from a less to more stringent regulatory alternative.
3. Cases Analyzed
The I CA was limited to one urban area for six scenarios regarding
different combinations of FMVCP and the year of implementation and compliance.
Only Los Angeles was projected to need controls beyond FMVCP and NSPS to
meet ambient NOg standards as stringent as 0.060 ppm on an annual basis.
Analyses were conducted assuming three different 1-evels of FMVCP—1.0, 1.5
and 2.0 grams per mile (gpm). And finally, analyses were performed assuming
1985 and 1990 as alternative dates of implementation and compliance for the
N02 NAAQS.
-------
48
C. Incremental Cost Analysis
This section shows the estimates of costs per 0.001 ppm change in
annual N02 concentrations associated with incremental changes in the N02
ambient standard. Tables 6.1 and 6.2 show the costs and ambient concentrations
associated with alternative N02 standards. These tables were used to construct
the incremental costs per 0.001 ppm change in annual N02 concentrations for
moving from less to more stringent N02 standards.
Several noteworthy conclusions can be drawn from observing Tables 6.1
and 6.2. First of all, delay of the implementation/compliance date for
either the 0.060 or 0.053 ppm standard can significantly reduce the cost of
either alternative. For instance, under the 2.0 gpm FMVCP scenario, the
cost of the 0.053 ppm standard under a 1985 implementation date is 140 million
dollars. If the compliance date is extended to 1990, the annualized cost
drops to 10 million dollars. This decline can be explained by the general
decline in ambient N02 levels over time. More specifically, ambient N02
concentrations are falling as the old stock of automobiles are replaced by
new models with N02 controls. A second conclusion that can be drawn from
Tables 6.1 and 6.2 is that more stringent FMVCPs will reduce the costs of
controls associated with alternative N02 standards. In some cases, more
stringent FMVCPs will cause the ambient standards to be attained without any
additional controls above FMVCP being required. Thirdly, the tables show
that in the 1985 attainment case, the cost of the 0.053 standard is considerably
larger than the 0.060 standard. This observation is explained by the fact
that the baseline air quality estimate in Los Angeles is already near 0.060 ppm.
-------
49
Table 6.1
ANNUALIZEO COSTS INCURRED UNDER ALTERNATIVE NU2 NAAQS1
(Millions of March 1984 Dollars)
FMVCP Standard
1.0 gpm
1.5 gpm
2.0 gpm
1985
Alternative NO? NAAQS
Baseline
0
0
0
0.060 ppm
0
30
30
0.053 ppm
120
140
140
FMVCP Standard
1.0 gpm
1.5 gpm
2.0 gpm
1990
Alternative NO? NAAQS
Baseline
0
0
0
0.060 ppm
0
0
0
0.053 ppm
0
0
+10
Costs above and beyond those incurred by NSPS and FMVCP.
-------
50
Table 6.2
N02 CONCENTRATIONS ACHIEVED UNDER ALTERNATIVE NOo NAAQS1
(ANNUAL AVERAGES)
1985
Alternative NO? NAAQS
FMVCP Standard Baseline 0.060 ppm 0.053 ppm
1.0 gpm 0.060 0.060 0.053
1.5 gpm 0.061 0.055 0.053
2.0 gpm 0.061 0.055 0.053
FMVCP Standard
1.0 gpm
1.5 gpm
2.0 gpm
1990
Alternative
Baseline
0.052
0.053
0.057
NO? NAAQS
0.060 ppm
0.052
0.053
0.057
0.053 ppm
0.052
0.053
0.052
In some cases, the discrete characteristics of additional controls cause
the ambient N0 forecast to overshoot the actual ambient standard.
-------
51
Table 6.3 shows the results of the ICA analysis. The numbers show
the change in cost per change in 0.001 ppm annual average of N02- Changes
in 0.001 ppm were selected as opposed to one ppm changes because none of
the alternatives change the ambient N02 by more than a few one-thousandths
of a ppm. The table shows again that the incremental costs of moving to
more stringent standards are higher for 1985 than for 1990 compliance.
For example, the incremental cost of moving from the 0.060 to the 0.053 ppm
standard with an FMVCP of 1.0 gpm is 17.1 million dollars per 0.001 ppm
reduction in 1985 and zero in 1990. A second finding observable from
Table 6.3 is that in 1985, the incremental cost of the 0.053 ppm standard
is higher than the incremental cost of the 0.060 ppm standard. This occurs
because baseline air quality is already close to 0.060 ppm and hence less
costs are needed to improve air quality to the less stringent levels. By
1995, there is less difference in the incremental costs between the two
standards since no additional costs are needed to meet even the 0.053 ppm
standard except under the 2.0 gpm FMVCP.
D. QUALI FI CATI ONS
1. Use of ICA in General
ICA, in general, embodies the same methodology as cost effectiveness
analysis (CEA). The usefulness of such tools can allow one to identify the
least cost way to achieve a given goal. In addition, the methods can
eliminate inefficient ways of achieving given goals. The techniques used
above can in no way help identify the optimal goal in an economic efficiency
sense. Hence, the ICA can only be used to identify the opportunity costs
of implementing more stringent standards and cannot identify the socially
desired N02 NAAQS from an economic point of view.
-------
52
Table 6.3
INCREMENTAL COSTS PER PPM N02 CONCENTRATION REDUCTION1
(Millions of 1984 Dollars Per 0.001 ppm Reduction)
1985
Change in NO? NAAQS
FMVCP Standard Baseline to 0.060 ppm 0.060 to 0.053 ppm
1.0 gpm — 17.1
1.5 gpm 5.0 55.0
2.0 gpm 5.0 55.0
1990
Change in NO? NAAQS
FMVCP Standard Baseline to 0.060 ppm 0.060 to 0.053 ppm
1.0 gpm — —
1.5 gpm — —
2.0 gpm — 2.0
Annual average changes in N02 concentrations.
-------
53
2. The Effects Measure of the ICA
ICA--and for that matter CEA--can use a number of effects measures
in ascertaining the cost effectiveness of alternative actions. Since the
general purpose of the NAAQS is to reduce adverse effects of pollution on
sensitive populations, one measure of effectiveness for ICA could have
been the reduction in disease incidence. Insufficient scientific information
was available to use this effectiveness measure however. In fact, had the
information been available, a partial (other benefits not measured) benefit
and benefit cost analysis would have been performed in lieu of the ICA.
Two other measures of effects that could have been used include reduced
exposures above adverse concentration levels and/or reduced emissions of
NOg. The latter was not chosen because the ultimate goal of the NAAQS
decision-making process is to reduce pollution levels below those considered
harmful to public health and welfare and not to reduce emissions per se.
The former method—reductions in exposures—was originally used in the ICA.
However, updates in the cost data's benchmark air quality levels caused
inconsistencies between the old exposure analysis and the most recent cost
analysis. Due to limitations in time and budget the exposure analysis was
not updated.
3. Limitations in the Effects Measure of the ICA
To recapitulate, the effects measures used in the ICA are the
0.001 ppm reductions in ambient air quality of enforcing more stringent N02
NAAQS. This measure of ambient air quality is the same as used in the cost
analysis. The cost analysis used the design value (Chapter V) as a measure
of ambient air quality for urban areas to examine compliance with alternative
NAAQS and to determine appropriate control measures to ensure compliance.
-------
54
This measure may not fully represent the ambient air quality levels and
changes across alternative standards for regions as large as the urban
areas analyzed. Hence, the ICA results should be interpreted with caution,
-------
55
VII. SUMMARY OF THE RATIONALE FOR SELECTING THE PROPOSED STANDARDS
This Chapter summarizes the Agency's rationale for retention of the
current annual N02 primary and secondary NAAQS. A more complete discussion
of the rationale is contained in the proposal preamble.
A. HEALTH EFFECTS AND THE PRIMARY STANDARD
O
The current primary NAAQS for N02 is 0.053 ppm (100 yg/irr), averaged
over 1 year. In addition to requiring review of the existing criteria
and standards for N02, in 1977 Congress amended section 109(c) of the
Clean Air Act to require the Administrator to promulgate a short-term N02
primary standard with an averaging time of not more than 3 hours unless
he or she finds no significant evidence that such a standard is required to
protect public health (Section 109(c)).
Section 109(b)(l) of the Clean Air Act requires EPA to set primary
standards, based on air quality criteria and allowing an adequate margin
of safety, which in the Administrator's judgment are requisite to protect
the public health. The legislative history of the Act makes quite clear
the Congressional intent to protect sensitive persons who in the normal
course of daily activity are exposed to the ambient environment. Air
quality standards are to be established with reference to protecting the
health of a representative, statistically related, sample of persons
comprising the sensitive group rather than a single person in such a
group.
EPA's objective, in reviewing the adequacy of the existing annual
N02 primary standard and the need for a separate short-term primary
standard, therefore, is to set primary air quality standards which
accurately reflect consideration of the existing scientific evidence, an
adequate assessment of the uncertainties of this evidence, and a reasonable
-------
56
provision for scientific and medical knowledge yet to be acquired, so as
to protect sensitive population groups with an adequate margin of safety.
The Criteria Document (EPA, 1983) supports the conclusion that a clear
threshold of adverse health effects cannot be identified with certainty
for N02- There is rather a continuum consisting of N0£ levels at which
health effects are certain through levels at which scientists can generally
agree that health effects have been convincingly shown, and down to levels
at which the indications of health effects are less certain and harder to
identify. (This does not necessarily mean that there is no threshold,
other than zero, for NC^; it simply means no clear threshold can be
identified with certainty based on existing medical evidence.) Thus,
selecting a standard that takes into account the known continuum of
effects is a judgment of prudent health policy, and does not imply some
discrete 'or exact margin of safety appended to a known threshold.
Determinations Concerning the Averaging Time and Standard Level
As discussed previously, EPA is required both to review the adequacy
of the existing 0.053 ppm annual N0£ standard and to determine whether a
short-term (less than 3 hours) N0£ standard is required to protect public
health. Although the scientific literature supports the conclusion that
NOg does pose a risk to human health, there is no single study or group of
studies that clearly defines human exposure-response relationships at or
near current ambient N02 levels. This situation exists because of both
methodological limitations of health effects research and lack of
sufficient studies involving population groups suspected of being
particularly sensitive to N02. Based on the review of the health
effects evidence presented in the Criteria Document, however, both EPA
and the CASAC have concluded that studies have demonstrated the occurrence
-------
57
of health effects resulting from both short-term and long-term NC>2 exposures.
Unfortunately, the various uncertainties in the health effects data make it
impossible to specify at this time the lowest-level at which adverse health
effects are believed to occur in humans due to either short- or long-term
NOg exposures.
Annual Standard. In reviewing the scientific basis for an annual
standard, EPA finds that the evidence suggesting the most serious health
effects associated with N02 exposures (e.g., emphysematous alterations in
the lung and increased susceptibility to infection) comes from animal studies
conducted at concentrations well above those permitted in the ambient air
by the current annual standard. The major limitation of these studies for
standard-setting purposes is that currently there is no satisfactory method
for quantitatively extrapolating exposure-response results from these
animal studies directly to humans. However, the seriousness of these
effects coupled with the biological similarities between humans and test
animals suggests that there is some risk to human health from long-term
exposure to elevated NOg levels.
Other evidence suggesting health effects related to long-term, low-
level exposures, such as the community epidemiology and gas stove community
studies, provides some qualitative support for concluding that there may
be a relationship between long-term human exposure to near-ambient levels
of NOg and adverse health effects. However, various limitations in these
studies (e.g., unreliable or insufficient monitoring data and inadequate
treatment of potential confounding factors such as humidity and pollutants
other than NOg) preclude derivation of quantitative dose-response relationships,
-------
58
Given the uncertainty associated with the extrapolation from animal to
man, the seriousness of the observed effects, and the inability to determine
from the available data an effects level for humans, EPA believes it would
be prudent public health policy to maintain the current annual standard of
0.053 ppm. As discussed in the proposal notice, EPA is also concerned that
any relaxation of the current annual standard would allow a rise in the
frequency and severity of short-term ambient N02 concentrations. The
results of EPA's analysis of short-term ambient concentrations in areas
that meet the current 0.053 ppm annual standard and alternative annual
standards in the range 0.05 to 0.08 ppm are discussed in more detail in
McCurdy and Atherton (1983) and in the proposal preamble (49 FR 6873).
Despite the lack of a firm relationship between various averaging times, it
was observed that where the annual average is at or below the current 0.053
ppm standard, days with one-hour concentrations in excess of any specified level
(including levels in the range 0.15 to 0.30 ppm) tend to be fewer in number
than at locations where the current annual standard is exceeded.
While it is not possible currently to quantify the margin of safety
provided by the existing annual standard, two observations are relevant:
(1) a 0.053 ppm standard is consistent with CASAC's recommendation (Friedlander,
1982; Lippmann, 1984) to set the annual standard at the lower end of the
range (0.05 to 0.08 ppm) cited in the OAQPS Staff Paper to ensure an adequate
margin of safety against long-term effects and provide some measure of
protection against possible short-term health effects, and (2) a 0.053 ppm
standard would keep annual N02 concentrations considerably below the long-term
levels for which serious chronic effects have been observed in animals.
Maintaining the current annual primary standard is a prudent public health
-------
59
policy choice that will prevent any increased chronic health risk in large,
populated urban areas that are now attaining the standard. Consequently,
the Administrator has determined that retaining the current primary annual
standard of 0.053 ppm is both necessary and sufficiently prudent to protect
public health against chronic effects with an adequate margin of safety and
provides some measure of protection against possible short-term health
effects.
Need for a Short-term Standard
Section 109(c) of the Clean Air Act specifically requires the Administrator
to promulgate a primary N02 standard with an averaging time of not more
than 3 hours unless he or she finds no significant evidence that such a
short-term standard is required to protect public health. In conjunction
with the review of the annual standard, EPA also has carefully examined the
health effects data base to determine whether a separate short-term standard
is required to protect public health. As discussed in more detail in the
OAQPS Staff Paper and proposal preamble, there are considerable uncertainties
about whether short-term (less than 3 hours) exposures to N02 at levels
observed in the ambient air cause any adverse health effects in humans.
Citing these uncertainties, EPA did not propose to set a separate short-term
standard and solicited public comment on the need, if any, for a separate
short-term standard (49 FR 6866). EPA also requested that public comments
on this issue identify any scientific or technical evidence that would
support any particular standard level and other relevant elements of the
standard, such as averaging time, number of exceedances, and form of the
standard.
EPA's assessment of the health effects evidence relevant to any decision
on the need for a separate short-term standard and EPA's review of scientific
-------
60
studies that have become available since CASAC closure on the Criteria
Document and OAQPS Staff Paper, have been summarized earlier in this RIA
in the section, Mechanisms of Toxicity and Nature of Effects, in Chapter IV.
More detailed information about EPA's assessment of the scientific evidence
pertinent to the short-term standard issue can be found in the Criteria
Document, OAQPS Staff Paper, and ECAO's review of recent studies (Grant,
1984).
Public comments on the proposal generally argued for one of the following
positions: (1) EPA should propose a short-term primary standard, (2) EPA
should conclude that no short-term standard is needed at this time, or
(3) EPA should defer its decision on whether a separate short-term standard
is needed until results are available from a multi-year research program
focused on resolving or reducing the uncertainties surrounding the need for
a short-term standard. EPA staff discussed these three options and ECAO's
review of the newer scientific studies with the CASAC at the public meeting
held on July 19-20, 1984. A transcript of a meeting has been placed in the
docket (OAQPS 78-9).
The CASAC, as indicated in its October 18, 1984 letter to the Administrator
(Lippmann, 1984), concurred with the EPA staff that the available information
was insufficient to provide an adequate scientific basis for decisions on a
short-term standard level, averaging time, and number of allowable exceedances
which would be required to propose a separate short-term standard. At the
same time the CASAC stated that it could not rule out the possibility of adverse
health effects at ambient N02 levels given the large uncertainties in the
scientific data base. CASAC concluded that either of the remaining options,
which would not propose to set a short-term standard at this time, were
-------
61
functionally equivalent, i.e., EPA could aggressively pursue scientific
research to resolve or reduce the uncertainties about health effects related
to short-term N02 exposures under either option selected. CASAC recommended
that EPA "reaffirm the annual standard at the current level" and that EPA
"defer a decision on the short-term standard while pursuing an aggressive
research program on short-term effects of N02" (Lippmann, 1984).
Given (1) the language on the short-term standard in the Clean Air Act
which requires the Administrator to establish a short-term standard unless
he or she finds that there is no significant evidence that one is required
to protect public health and (2) the large scientific uncertainties remaining
about possible short-term'effects at ambient N02 levels, the Administrator
has concluded that it would be prudent to defer a decision on the need for
a short-term standard. The Agency is committed to carrying out a focused
research program designed to resolve or reduce the major uncertainties
associated with the question of whether short-term N02 exposures at ambient
levels adversely affect public health. "In the meantime, the Administrator
believes that attainment of the current 0.053 ppm annual standard will
provide some measure of protection against possible short-term health
effects.
B. WELFARE EFFECTS AND THE SECONDARY STANDARD
As indicated above, section 109(b) of the Clean Air Act mandates the
setting of secondary NAAQS to protect the public welfare from any known or
anticipated adverse effects associated with an air pollutant in the ambient
atmosphere. A variety of effects on public welfare have been attributed to
N02 and NOX compounds. These effects include increased rates of acidic
deposition, symptomatic effects in humans, vegetation effects, materials
damage, and visibility impairment. The OAQPS Staff Paper (OAQPS 78-9, II-A-7)
-------
62
describes in detail each of the welfare effects of concern. The following
discussion summarizes the welfare-related effects discussed in the OAQPS
Staff Paper, and CASAC's comments relating to the secondary NC>2 NAAQS.
The issue of acidic deposition was not directly assessed in the OAQPS
Staff Paper because EPA has followed the guidance which was given by CASAC
on this subject at its public meeting review of the draft document, "Air
Quality Criteria for Particulate Matter and Sulfur Oxides," which was held
on August 20-22, 1980. The CASAC concluded that acidic deposition is a
topic of extreme scientific complexity because of the difficulty in establishing
firm quantitative relationships between emissions of relevant pollutants,
formation of acidic wet and dry deposition products, and effects on terrestrial
and aquatic ecosystems. Secondly, acidic deposition involves, at a minimum,
the criteria pollutants of oxides of sulfur, oxides of nitrogen, and the
fine particulate fraction of suspended particulates. Finally, the Committee
felt that any document on this subject should address both wet and dry
deposition, since dry deposition is believed to account for at least one-half
of the total acid deposition problem. For these reasons, the Committee
felt that a significantly expanded and separate document should be prepared
prior to any consideration of using NAAQS as a regulatory mechanism for
control of acidic deposition. CASAC suggested that a discussion of acidic
deposition be included in the criteria documents for both NOX and particulate
matter and sulfur oxides, but that plans also be made for the development
of a separate, comprehensive document on acid deposition. In response to
these recommendations, EPA is in the process of developing an acidic deposition
document that will provide a more comprehensive treatment of this subject.
As defined in section 3Q2(h) of the Act, welfare effects include
effects on personal comfort and well being. Mild symptomatic effects
-------
63
were observed in 1 of 7 bronchitics and in 7 of 13 asthmatics during or
after exposure to 0.5 ppm N02 for 2 hours in the Kerr et al. (1979)
study. The authors indicate that the symptoms were mild and reversible
and included slight headache, nasal discharge, dizziness, chest tightness
and labored breathing during exercise. In EPA's judgment these mild symptomatic
effects affect personal comfort and well being and could be considered
adverse welfare effects in certain situations. CASAC generally agreed with
this judgment, but felt that because short-term peaks associated with these
effects are rarely observed in areas where the current annual standard of
0.053 ppm was met, the current annual standard is adequate to protect against
these effects.
Evidence in the Criteria Document and information provided by plant
physiologists (Heck, 1980; Tingey, 1980a; Tingey, 1980b) have indicated
that visible injury to vegetation due to N02 alone occurs at levels which
are above ambient concentrations generally occurring within the U.S.,
except around a few point sources. Several studies (Korth et al., 1964;
Haagen-Smit et al., 1952; Heck, 1964; Taylor et al., 1975; Thompson et al.,
1970) on the effects of N02 alone on vegetation have failed to show plant
injury at concentrations below 2 ppm for short-term exposures. For long-term
exposures, such as a growing season, the lowest concentration reported to
depress growth is approximately 0.25 ppm (Korth, 1964). The concentrations
which produced injury or impaired growth in these studies are higher than
those which would be expected to occur in the atmosphere for extended
periods of time in areas attaining a 0.053 ppm annual standard.
In regard to vegetation effects from N0£ in combination with other
pollutants, plant responses to pollutant mixtures appear to vary with
concentration, ratio(s) of pollutants, sequence of exposure, and other
-------
64
variables. Studies examining exposure to NOg and S02 as well as to 63 and
S02 (MacDowell and Cole, 1971; Tingey, 1973) have shown that the synergistic
response is most pronounced near the threshold doses of the gas combinations
tested and that, as concentrations increase beyond the threshold doses, the
synergistic response diminishes, often becoming additive, or in some cases,
antagonistic. Therefore, although the limited evidence available indicates
that low levels of N02 and S02 can have a synergistic effect, this type of
response is extremely variable and has not been sufficiently documented.
CASAC concurred with EPA's judgment that the data do not suggest significant
effects of NC>2 on vegetation at or below current ambient levels and that an
annual standard of 0.053 ppm would provide sufficient protection against
significant effects on vegetation.
In regard to visibility impairment due to N02, the scientific evidence
indicates that light scattering by particles is generally the primary cause
of degraded visual air quality and that aerosol optical effects alone can
impart a reddish brown color to a haze layer. Thus while it is clear that
both particles and N02 contribute to brown haze, the CASAC concurred with
EPA's judgment that the relationship between N02 concentrations and visibility
impairment has not been sufficiently established and that a separate secondary
standard to protect visibility is not warranted at this time. CASAC confirmed
this judgment at its public meeting held on July 19-20, 1984.
Finally, while N02 has been qualitatively associated with materials
damage, CASAC concurred with EPA's judgment that the available data do
not suggest major effects of N02 on materials for concentrations at or
below the current annual standard of 0.053 ppm.
Based on an evaluation of symptomatic effects, vegetation damage,
visibility impairment, and materials damage, and the levels at which these
-------
65
effects are observed, it is EPA's judgment that the current annual standard
provides adequate protection against both long- and short-term welfare
effects and that there is no need for a different secondary standard. For
these reasons, EPA is retaining the secondary standard at the same level as
the primary standard.
-------
66
VIII. STATUTORY AUTHORITY
The statutory authority for the retention of the annual NOg NAAQS is
contained in the Clean Air Act. Two sections of the Act govern the
establishment and revision of NAAQS. Section 108 (42 U.S.C. 7408) requires
EPA to document the most recent scientific information (criteria) on the
health and welfare effects of certain air pollutants. Section 109 provides
authority for establishing and revising primary (health based) and secondary
(welfare based) NAAQS. A more complete discussion of the legal authority
for this action is contained in the promulgation preamble.
-------
67
REFERENCES
1. American Petroleum Institute v. Costle 655 F. 2d 1176 (D.C. Cir.,
1981), cert, den. 102 S. Ct. 1737 (1982).
2. Ahmed, T., B. Marchette, I. Danta, S. Birch, R.L. Dougherty, R. Schreck,
and M.A. Sackner, 1982. "Effect of 0.1 ppm N02 on bronchial reactivity
in normals and subjects with bronchial asthma." Am. Rev. Respir. Pis.
125: 152, abstract.
3. Ahmed, T., I. Danta, R.L. Dougherty, R. Schreck, and M.A. Sackner, 1983.
"Effect of N02 (0.1 ppm) on specific bronchial reactivity to ragweed
antigen in subjects with allergic asthma." Am. Rev. Respir. Pis.
127: 160, abstract.
4. Aranyi, C., J. Fenters, R. Ehrlich, and P. Gardner, 1976. "Scanning
electron microscopy of alveolar macrophages after exposure to 02,
N02, and 03." Environ. Health Persp. 16: 180.
5. Bauer, M.A., M.J. Utell, P.E. Morrow, D.M. Speers, F.R. Gibb, 1984.
"0.30 ppm nitrogen dioxide inhalation potentiates exercise-induced
bronchospasm in asthmatics." Am. Rev. Respir. Pis. 129: A151, abstract.
6. Beil, M., and W. T. Ulmer, 1976. "Wirkung von N02 in MAP-Bereich
auf Atemmechanik and Acetylcholinempfindlichkeit bei Normalpersonen.
(Effect of N02 in workroom concentrations on respiratory mechanics and
bronchial susceptibility to acetylcholine in normal persons.)"
Int. Arch. Occup. Environ. Health. _38: 31-44.
7. Biller, W., T. Johnson, T. Feagans, G. Duggan, T. McCurdy, and H. Thomas,
1981. "A general model for estimating exposure associated with
alternative NAAQS." Presented at 74th Annual Meeting of the Air
Pollution Control Association. Philadelphia, PA, June 21-26, 1981.
8. Cole, H. S. and Summerhays, J.E. 1979. "A Review of Techniques
Available for Estimating Short-Term N02 Concentrations." J. Air.
Poll. Control Assoc. Volume 29, No. 8.
9. Comstock, G.W., M.B. Meyer, K.J. Helsing, and M.S. Tockman, 1981.
"Respiratory effects of household exposures to tobacco smoke and
gas cooking." Am. Rev. Respir. Pis. 124: 143.
10. Ehrlich, R., and M.C. Henry, 1968. "Chronic toxicity of nitrogen
dioxide, I. Effect on resistance to bacterial pneumonia." Arch.
Environ. Health 17: 860-865.
11. Ehrlich, R., J.C. Findlay, J.P. Fenters, and P.E. Gardner, 1977.
"Health effects of short-term exposures to inhalation of N02-03 mixtures,"
Environ. Res. 14: 223-231.
12. Florey, C. du V., R.J.W. Melia, S. Chinn, B.D. Goldstein, A.G.F.
Brooks, H.H. John, I.B. Craighead, and X. Webster, 1979. "The
relation between respiratory illness in primary school children
and the use of gas for cooking. Ill - Nitrogen dioxide, respiratory
illness and lung infection." Int. J. Epid. 8: 347.
-------
68
12. Friedlander, Sheldon K., Chairman, Clean Air Scientific Advisory
Committee (CASAC), 1982. Memorandum to EPA Administrator.
Subject: CASAC Review and Closure of the OAQPS Staff Paper for
Nitrogen Oxides. July 6, 1982.
13. Gardner, D.E., F.J. Miller, E.J. Blommer, and D.L. Coffin, 1977.
"Relationship between nitrogen dioxide concentration, time and level
of effect using an animal infectivity model." In: Proceedings
International Conference on Photochemical Qxidant Pollution and
Its Control, Vol. 1 EPA-600/3-77-001a. U.S. Environmental Protection
Agency, January, 1977. pp. 513-525.
14. Gardner, D.E., J.A. Graham, J.W. Illing, E.J. Blommer, and F.J. Miller,
1979. "Impact of Exposure Patterns on the lexicological Response
to N02 and Modifications by Added Stressors." In: Proceedings of
the U.S.-USSR 3rd Soviet Symposium on Problems of Environmental Health.
Suzdal, USSR, October, 1979. pp. 17-40. Report by NIEHS.
15. Goldstein, B.D., R.J.W. Melia, S. Chinn, C. du V. Florey, D. Clark,
and H.H. John, 1979. "The relation between respiratory illness in
primary school children and the use of gas for cooking. II - Factors
affecting nitrogen dioxide levels in the home." Int. J. of Epid.
8;. 339.
16. Grant, L., Director, Environmental Criteria and Assessment Office, 1984.
Memorandum to Gerald Emison, Director, Office of Air Quality Planning and
Standards. Subject: Review of Scientific Studies on NOX Available
Since CASAC Closure on the Criteria Document and OAQPS Staff Paper.
November 1984.
17. Haagen-Smit, A.J., E.F. Darley, M. Zaithlin, H. Hull, and W. Noble (1952).
Investigation of injury to plants from air pollution in the Los
Angeles area. Plant Physio!. 27: 18-34.
18. Harberger, A.C. 1971. "Three Basic Postulates for Applied Welfare
Economics." In: Benefit Cost Analysis 1971. Chicago and New York:
A!dine.
19. Hattori, S., R. Tateishi, T. Horai, and T. Nakajima, 1972. "Morphological
changes in the bronchial alveolar system of mice following continuous
exposure to N02 and CO." J. Jap. Soc. Chest Disease 10: 16-22.
(In Japanese).
20. Hattori, S, 1973. "Alterations of broncho-alveolar system by polluted
air: Experimental consideration," Clinician 219: 4-8. (In Japanese).
21. Hazucha, M.J., J.F. Ginsberg, W.F. McDonnell, E.D. Haak, Jr., R.L. Pimmel ,
S.A. Salaam, D.E. House, and P.A. Bromberg, 1983. "Effects of 0.1 ppm
nitrogen dioxide on airways of normal and asthmatic subjects." J. Appl .
Physio! . 54: 730.
-------
69
22. Heck, W. W. (1964). "Plant injury induced by photochemical reaction products
of propylene-nitrogen dioxide mixtures." J. Air Pollut. Control Assoc.
_H: 255-261.
23. Heck, W. W., North Carolina State University, Raleigh, N.C. (1980) Personal
Communication with P.M. Johnson, U.S. EPA, February.
24. Helsing, K.J., 6.W. Comstock, M.B. Meyer, and M.L. Tockman, 1982.
"Respiratory effects of household exposures to tobacco smoke and gas
cooking on nonsmokers." Environ. Int. 8: 365.
25. Ito, K, 1971. "Effect of nitrogen dioxide inhalation on influenza virus
infection in mice." Jap. J. Hygiene 26: 304-314. (In Japanese).
26. Jones, J.R., I.T.T. Higgins, M.W. Higgins, and J.B. Keller, 1983.
Effects of cooking fuels on lung function in nonsmoking women.
Arch. Environ. Health 38: 219.
27. Kagawa, J., and T. Toyama, 1975. "Photochemical air pollution: Its
effects on respiratory function of elementary school children,"
Arch. Environ. Health 30: 117-122.
28. Kagawa, J. and K. Tsuru, 1979. "Respiratory effects of 2 hour exposure to
ozone and nitrogen dioxide alone and in combination in normal subjects
performing intermittent exercise." Jpn. J. Thorac. Pis. 17: 765.
29. Kerr, H.D., T.J. Kulla, M.L. Mcllhany, and P. Swidersky, 1979. "Effect
of nitrogen dioxide on pulmonary function in human subjects: An
environmental chamber study." Environ. Res. 19: 392-404.
30. Kleinman, M.T., R.M. Bailey, W.S. Linn. K.R. Anderson, J.D. Whynot,
D.A. Shamoo, and J.D. Hackney, 1983. "Effects of 0.2 ppm nitrogen
dioxide on pulmonary function and response to bronchoprovocation in
asthmatics." J. Toxicol. Environ. Health 12: 815.
31. Korth, M.W., A.H. Rose, and R.C. Stahman (1964). Effects of hydrocarbon
to oxides of nitrogen ratios on irradiated auto exhaust, Part 1. J.
Air Pollut. Control Assoc. 14: 168-175.
32. Kosmider, S., A. Misiewicz, E. Felus, M. Drozdz, and K. Ludyga, 1973.
"Experimental and clinical studies on the effects of nitrogen oxides
on immunity." Intern. Arch Arbeitsmed. 31: 9-23, (English translation).
33. Lead Industries Association, Inc. v. EPA, 1980. 647 F. 2d 1130
(D.C. Cir., 1980), cert, den. 101 S. Ct. 621 (1980).
34. Lebowitz, M.D., M.K. O'Rourke, R. Dodge, C.J. Holberg G. Corman,
R.W. Hoshaw, J.L. Pinnas, R.A. Barbee, and M.R. Sneller, 1982. "The
adverse health effects of biological aerosols, other aerosols, and
indoor microclimate on asthmatics and nonasthmatics." Environ.
Int. 8: 375.
-------
70
35. Lippman, M., Chairman, Clean Air Scientific Advisory Committee (CASAC),
1984. Letter to EPA Administrator. October 18, 1984.
36. Love, 6.J., S. Lan, C.M. Shy, and W.B. Riggan, 1982. "Acute respiratory
illness in families exposed to nitrogen dioxide ambient air pollution
in Chattanooga, Tennessee." Arch. Env. Health 37(2): 75-80.
37. Melia, R.J.W., C. du V. Florey, D.S. Altman, and A.V. Swan, 1977.
"Association between gas cooling and respiratory disease in children."
Brit. Med. J. 2: 149-152.
38. Melia, R.J.W., C. du V. Florey, and S. Chinn, 1979. "The relation
between respiratory illness in primary school children and the use
of gas for cooking. I - Results from a national survey." Int. J.
Epid. 3: 333.
39. Melia, R.J.W., C. du Ve Florey, R.W. Morris, B.D. Goldstein, H.H. John,
D. Clark, I.B. Craighead, and J.C. Mackinlay, 1982. Childhood respiratory
illness and the home environment: II. Association between respiratory
illness and nitrogen dioxide, temperature, and relative humidity. Int. J.
Epidemic!. II: 164.
40. Melia, J., C. Florey, Y. Sittampalam, and C. Watkins, 1983. The relation
between respiratory illness in infants and gas cooking in the U.K.:
a preliminary report. In the Proceedings of the Vlth World Congress on
Air Quality, Vol. 2, pp. 263-269. International Union of Air Pollution
Prevention Association, Paris, May 16-20.
41. MacDowall, F.D.N., and A.F.W. Cole (1971). "Threshold and synergistic damage
to tobacco by ozone and sulfur dioxide." Atmos. Envir. 5: 553-559.
42. Menzel, D.B., M.D. Abou-donia, C.R. Roe, R. Ehrlich, D.E. Gardner,
and D.L. Coffin, 1977. "Biochemical indices of nitrogen dioxide
intoxication of guinea pigs following low-levels long-term exposure,"
In: Proceedings International Conference on Photochemical Oxidant
Pollution and Its Control. Vol. II, September, 1973, B. Dimitriades
(ed.).EPA-600/3-77-001b. U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, pp. 577-587.
43. Menzel, D. B., 1980. "Pharmacological Mechanisms in the Toxicity of
Nitrogen Dioxide and its Relation to Obstructive Respiratory Disease."
In: Nitrogen Oxides and Their Effect on Health. S.D. Lee, ed., Ann
Arbor Science, Ann Arbor, pp. 199-216.
44. Miller, F.J., J.A. Graham, J.W. Illing, and D.E. Gardner, 1980.
"Extrapulmonary mice." Tox. Lett. 6_, 267-274.
45. McCurdy, T., and R. Atherton, 1983. "Analysis of Peak Daily Maximum
N02 Values and Associated Annual Average in 1979-1981 Data." Research
Triangle Park, N.C.
-------
71
46. Nakajima, T., S. Kusumoto, C. Chen, and K. Okamoto, 1969. "Effects of
prolonged continuous exposure to nitrogen dioxide on the quantity of
reduced glutathione in lungs of mice and their histopathological
changes. Appendix: Effects of nitrite and nitrate on the glutathione
reductase." Osaka Pretectural Pub. Health Inst., Res. Reports, Labor
Sanitation Series. 7: 35-41. (In Japanese).
47. Orehek, J., J.P. Massari, P. Gayrard, C. Grimaud, and J. Charpin, 1976.
"Effect of short-term, low-level nitrogen dioxide exposure on bronchial
sensitivity of asthmatic patients." J. Clin. Invst. 57: 301-307.
48. Port, C.D., D.L. Coffin, and P. Kane, 1977. "A comparative study of
experimental and spontaneous emphysema." J. Toxicol. Environ. Health
2: 589-604.
49. Posin, C., K. Clark, M.P. Jones, J.V. Patterson, R.D. Buckley, and
J.D. Hackney, 1978. "Nitrogen dioxide and human blood biochemistry."
Arch. Environ. Health (Nov/Dec): 318-324.
50. Rowe, R.D. and L.G. Chestnut, 1981. Visibility Benefits Assessment
Guidebook. U.S. Environmental Protection Agency report EPA-450/5-81-001.
August.
51. Shy, C.M., J.P. Creason, M.E. Pearlman, K.E. McClain, F.B. Benson,
and M.M. Young, 1970a. "The Chattanooga school children study:
effects of community exposure of nitrogen dioxide. I. Methods,
description of pollutant exposure and results of ventilatory
function testing." J. Air Pollut. Control Assoc. 20(8): 539-545.
52. Shy, C.M., J.P. Creason, M.E. Pearlman, K.E. McClain, F.B. Benson,
and M.M. Young, 1970b. "The Chattanooga school study: effects of
community exposure to nitrogen dioxide. II. Incidence of acute
respiratory illness." J. Air Pollut. Control Assoc. 20(9): 582-588.
53. Speizer, F.E., B.G. Ferris, Jr., Y.M.M. Bishop, and J. Spengler, 1980.
"Respiratory disease rates and pulmonary function in children associated
with N02 exposure." Am. Rev. Resp. Pis. 121:3-10.
54. Spengler, J.D., B.G. Ferris, Jr., and D.W. Dockery, 1979. "Sulfur
dioxide and nitrogen dioxide levels inside and outside homes and
the implications on health effects research." Environ. Sci. Techn.
13_: 1276-1 271.
55. Suzuki, T., and K. Ishibawa, 1965. "Research on effect of smog on
human body." Research and Report on Air Pollution Prevention 2:
199-221, (In Japanese).
56. Taylor, O.C., C.R. Thompson, D.T. Tingey, and R.A. Reinert (1975).
Oxides of nitrogen. In: Responses of Plants to Air Pollution, J.B. Mudd
and T.T. Kozlowski, eds. Academic Press, Inc., New York, N.Y., pp. 121-139,
-------
72
57. Thompson, C.R., E.G. Hensel, G. Kats, and O.C. Taylor (1970). "Effects of
continuous exposure of navel oranges to NO?." Atmos. Environ. 4: 349-
355.
58. Tingey, D.T., U.S. EPA, Corvallis, OR. (1980a). Personal Communication with
P.M. Johnson, U.S. EPA, February.
59. Tingey, D.T. (1980b). Written correspondence to P.M. Johnson, U.S. EPA,
May 23.
60. Tingey, D.T., R.A. Reinert, C. Wickliff, and W.W. Heck (1973). " Foliar
injury responses of 11 plant species to ozone/ sulfur dioxide mixtures."
Atmos. Envir. 2: 201-208.
61. U.S. Bureau of the Census, 1973. 1970 Census of Population. Vol 1
Part 1. U.S. Dept. of Commerce, Washington, D.C.
62. U.S. Dept. of Health, Education, and Welfare (DHEW), 1973. Prevalence
of Selected Chronic Respiratory Conditions, United States, 1970. DHEW
Publication No. (HRA) 74-1511.Series 10, Number 84, Rockville, MD.
September.
63. U.S. Environmental Protection Agency, 1982a. Review of the National
Ambient Air Quality Standards for Nitrogen Oxides: Assessment of Scientific
and Technical Information (NO? Staff Paper).EPA report 450/5-82-002.
Office of Air Quality Planning and Standards, Research Triangle Park,
North Carolina.
64. U.S. Environmental Protection Agency, 1982c. Cost and Economic Assessment
of Regulatory Alternatives for NO? NAAQS (DRAFTTTOffice of Air Quality
Planning and Standards, Research Triangle Park, North Carolina.
65. U.S. Environmental Protection Agency, 1982d. NAAQS Environmental Impact
Statement for NO?. Office Of Air Quality Planning and Standards, Research
Triangle Park, North Carolina.
66. U.S. Environmental Protection Agency, 1983. Air Quality Criteria
for Oxides of Nitrogen (Criteria Document). EPA report 600/8-82-026F.
Environmental Criteria Assessment Office, Research Triangle Park,
North Carolina.
67. U.S. Environmental Protection Agency, 1983a. Control Techniques for
Nitrogen Oxides Emissions from Stationary Sources - Revised Second
EdrtTion. EPA report 450/3-83-002. Office of Air Quality Planning and
Standards, Research Triangle Park, North Carolina.
68. Von Nieding, G.H. Krekeler, R. Fuchs, H.M. Wagner, and K. Koppenhagen,
1973. "Studies of the acute effect of N02 on lung function:
influence on diffusion, perfusion, and ventilation in the lungs."
Int. Arch. Arbeitsmed. 31:61-72.
-------
73
69. Von Nieding, G., H.M. Wagner, H. Krekeler, U. Smidt, and K. Muysers, 1971
"Minimum concentrations of N0£ causing acute effects on the respiratory
gas exchange and airway resistance in patients with chronic bronchitis."
Int. Arch. Arbeitsmed. 27: 338-348. Translated from German by
Mundus Systems for Air Pollution Technical Information Center,
U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina.
70. Von Neiding, G., H.M. Wagner, H. Lollgen, and K. Krekeler, 1977.
"Acute effects of ozone on lung function of men." VDI-Ber. 270:
123-129.
71. Ware, J.H., D.W. Dockery, A. Spiro, II, F.E. Speizer, and
B.G. Ferris, Jr., 1984. "Passive smoking, gas cooking, and respiratory
health of children living in six cities." Am. Rev. Respir. Pis.
129: 366.
-------
TECHNICAL APPENDIX A
TO
CHAPTER VI
OF THE
N02 NAAQS REGULATORY IMPACT ANALYSIS
Winston Harrington
Alan J. Krupnick
October 1982
Resources for the Future, Inc.
1755 Massachusetts Avenue, N.S.
Washington, B.C. 20036
-------
A-l
Table of Contents
Part One
An Evaluation of Selected Studies
Deriving Damage Functions of NO,, Effects
I. Introduction
II. Human Health Dose-Response Functions
A. Outdoor Epidemiological Studies
1. Hickey et al. (1970)
2. Stebbings and Hayes (1975)
3- Chapman et al. (1973)
4. Linn et al. (1976)
5. Kagawa and Toyama (1975)
6. Cohen et al. (1976)
7- Petr and Schmidt (1966)
8. Kalpazanov et al. (1976)
9- Shy et al. (1970) and Pearlman (1971)
10. Shy and Love (1979)
11. Shy et al. (1973)
12. Conclusions
B. Indoor Epidemiological Studies
1. Keller (1979)
2. Lutz (1977)
3. Melia (1977), (1979a), and (1979b)
4. Speizer et al. (1980)
5. Speizer (forthcoming)
6. Comparing Melia et al. to Speizer et al.
C. Indoor Exposure Studies and Indoor Dose-Response Functions
1. Indoor NOg Measurement Studies
III. Materials Damages
A. Fabric Fading
B. Fabric Yellowing
C. Corrosion Damage
-------
A-2
IV. Vegetation Damages
V. Valuations
A. Textiles
3. Property Value Studies
-------
A-3
PAHT ONE
AN EVALUATION OF SELECTED' STUDIES
DERIVING DAMAGE FUNCTIONS OF NO^ EFFECTS
I. INTRODUCTION
This section examines studies found in the Criteria Document for their
possible use in constructing damage functions. By far the most emphasis is
placed on epidemiological studies of the human health effects of NOj exposures.
Both "outdoor" and "indoor" (or gas stove) studies are reviewed and, where
feasible, physical damage functions are derived from the published data. Less
detailed discussions follow of the materials and vegetation effects of NO,.
Finally, a section is included that reviews several valuation studies that do
not appear in the Criteria Document, including studies on textile damages from
NOg and several "property value" studies which relate housing price
differentials to variation in ambient NO- levels.
II. HUMAN HEALTH DOSE-RESPONSE FUNCTIONS
Controlled human exposure studies, community exposure studies, and animal
studies have been used to investigate health effects to humans arising from NO,
exposures. However, only the community exposure studies provide the kind of
information necessary to construct dose-response functions. Unless one is
prepared to quantify the relationship between animal and human response to a
given dose (or to equivalent doses adjusted for metabolic rates, body weight,
and other factors), the animal studies cannot be used to quantify human health
effects.
-------
A-4
It is also difficult to use controlled exposure studies in the Criteria
Document for this purpose, for several reasons. First, the studies measure
adverse effects in ways that do not readily translate into quantifiable health
effects. How does an increase in airway resistance, a reduction in the
diffusion capacity of the lung, an increase in sensitivity to a
bronchoconstrictor, or a decrease in pulmonary function, relate to how an
individual feela or how his or her daily life or life span has changed? While
contingent valuation and other behavioral response techniques can be used to
assess values of such effects, time and resource constraints precluded the use
of these techniques for this project.
The controlled experiment studies reviewed here also have flaws that
impair their credibility. Many ignore the standard laboratory procedure of
using control groups. The Orehek study (1976), for instance, failed to
establish a group of individuals that used the inhalation device but inhaled
"air" without NOg or with N02 at background levels. Other researchers exposed
their subjects to NOg concentrations far in excess of "normal" ambient
concentrations. While it is recognized that the use of such high
concentrations nay be necessary in the interest of saving research time and
money, any adverse effects discovered by this technique must be extrapolated to
ambient concentrations before likely "field" responses can be quantified. Such
extrapolations must be arbitrary to some degree.
Two types of community exposure studies are reviewed in the Criteria
Document—"outdoor" studies and "indoor" studies. Curiously, the outdoor
studies make no attempt to correct for variability in indoor exposure; while
the indoor studies usually fail to correct for differences in outdoor N02
concentrations.
-------
A-5
A. OUTDOOR SPIDEMIOLOGICAL STUDIES
All of the "outdoor" epidemiological studies listed in the Criteria
Document that attempt to link ambient N02 concentrations with various health
effects have been reviewed.
1. Hickey et al. (1970).
This study used aggregate data from 38 SMSAs to investigate possible
links between mortality rates in 1960 for various diseases (chiefly cancers and
heart disease) and ambient concentrations of NO-, S02» sulfates and twenty
trace metals in ambient air. Using multiple regression, this study found N02
concentrations to be a significant predictor of mortality rates for breast
cancer, respiratory system cancer, lung cancer, urinary organ cancer, stomach
cancer, and arterioschlerotic heart disease. NO- also entered the equations
for congenital malformations and total deaths under one year, but with the
wrong sign. No significant results for bronchitis were found.
Unfortunately, this study is afflicted with such severe methodological
difficulties that the results are questionable. First, almost all regressions
are stepwise, with variables being selected only on the basis of their
2
contribution to R . In addition, the only explanatory variables are the
pollution variables, even though variables such as population density, median
age, and per capita income would be likely to affect disease rates. The fact
that N02 significantly enters most of the cancer incidence equations tends to
support Hickey's contention that N02 is a carcinogen, but this fact could have
been accounted for by correlations among the cancer rates. Unfortunately,
Hickey et al. do not report the correlation matrix for the dependent variables.
2. Stebbinga and Hayes (1975).
Over 300 persons aged 60 living in three neighborhoods of New York City
participated in a study of the acute effects of air pollution on symptoms of
-------
A-6
heart and lung disease. The panel contained some healthy individuals as well
as some with reported cardiopulmonary symptoms. This study is unusable for
several reasons: (i) because N02 measurements were made with the
Jacobs-Hochheiser technique, N02 was dropped from the study, (ii) the number of
pollutants examined greatly exceeded the number of sites where pollution was
measured. Therefore, effects of individual pollutants could not be isolated.
3. Chapman et al. (1973).
This article is a survey of the CHESS studies. Only one of the studies
mentioned examined N02 exposure—a survey of 3,500 adults in Chattanooga in
1970. No significant differences in respiratory disease symptoms were
reported. However, the Jacobs-Hochheiser method was used to calculate N0_
concentrations; hence the result is called into question.
*• Linn et al. (1976).
Office workers in San Francisco and Los Angeles were examined for
differences in respiratory disease symptoms. Only minimal differences were
found. In any event, no difference could be ascribed to NO- because of other
pollutants, chiefly oxidants and particulates.
5. gagawa and Toyama (1975).
Lung function of 21 elementary school children was correlated with daily
concentrations of oxidant, ozone, hydrocarbon, NO, N02, S02, and particulates.
N02 varied between 0.01 and 0.08 ppm. Only pairwise correlations were
reported. N02 was significantly correlated with maximum expiratory flow rates
for only two children (p < .05). Since no multivariate analyses were
performed, the results are difficult to interpret. Nonetheless, it does not
appear that N02 concentrations had an acute effect on lung function, at least
in the range experienced. Besides, since the study does not relate lung
function to disease, a dose-response function that could lead to benefit
-------
A-7
valuations is difficult to derive without using contingent valuation or other
behavioral response techniques.
6. Cohen et al. (1972).
This study looked for differences in respiratory disease incidence among
Seventh Day Adventists in the San Gabriel Valley and San Diego. (Seventh Day
Adventists were chosen because very few of them smoke.) Mo significant
differences in respiratory symptoms were found. As in the Linn study, N02
concentrations are correlated with ozone and participates, making it difficult
to isolate effects of particular pollutants.
7. Petr and Schmidt (1966).
This study examined prevalence of disease symptoms (including enlarged
tonsils and lymph glands and lymphocyte count) and respiratory function of
children in two urban areas of Czechoslovakia. The .study areas included a
control area, a high N02-iow S02 area, and a high S02-low N02 area. Most of
the effects were reported for the high-N02 area. This study is unsuitable
because only pollutant ranges were reported, and these ranges varied by a
factor of up to ten.
8. Kalpazanov et al. (1976).
This study attempted to determine whether daily air pollution readings
were related to registered instances of influenza in Sofia, Bulgaria during an
epidemic in December 1974-February 1975. Unlike the other epidemiological
studies examined, this one used multivariate regression. Repeated illnesses
were regressed against meteorological and pollution variables with 0, 1, and
2-day lags. While the regressions were reported to have significant F-values,
the standard errors of the estimated coefficients were not given. Nor was the
correlation matrix. The mean values of the independent variables were given
(21 ug/m3 for N02), but not the standard deviations. Thus, the results are of
-------
A-8
questionable validity. Considerable doubt is thrown on their results by the
fact that oxidants entered the equations with the wrong sign.
9. Shy et al. (1970) and Pearlman (1971).
As part of the CHESS studies, groups of elementary school children and
infants in the Chattanooga area were examined in an effort to determine the
effects, if any, of long-term exposure to ambient concentrations of N02.
Chattanooga was selected for these studies because it is the site of a large
stationary source of N02 (a TNT plant); several neighborhoods in threa were
then selected to give an M02 exposure gradient. School children, infants, and
their parents were then tested for lung function and surveyed to determine
respiratory disease incidence. These studies have been severely criticized for
defects of analytical and chemical procedure. In particular, the NO-
concentration measurements were made by the Jacobs-Hochheiser method, which has
since been shown to be affected by other constituents of the air. Also,
concentration measurements were made for only a six-month period (Oct. 29, 1963
to April 26, 1969), yet these were matched with respiratory disease data
spanning a three year period. In addition, the statistical techniques used in
these studies were not appropriate to the estimation of does-response functions
(nor was that their purpose).
In Pearlman, et al. (1971), frequencies of bronchitis, croup, and
pneumonia among infants and school children in three neighborhoods were related
to ambient concentrations of NOg. For each disease, a table of illness
frequency (percent of children reporting one or more disease episodes) versus
average pollutant concentration and years of exposure was given. Chi square
tests were then used to find disease rates that were different from what would
be expected in the absence of a pollution gradient. Table 1 shows the
bronchitis frequencies reported by Pearlman, et al. They concluded that "one
-------
A-9
or more episodes of bronchitis were reported significantly more often by school
children residing for two (p = 0.02) and three (p = 0.01) years in the high and
intermediate N02 areas. This pattern was not consistent for infants."
Table 1
Percent of Children Reporting One or More Incidents
of All Lower Respiratory Illness
Concentration School Children Infants
(ppm) years of exposure years of exposure
123 123
0.083
0.063
0.043
20.9
31.6
25.1
34.7
45.5
20.3
32.2
31.2
23.2
33-3
26.2
21.1
37.5
29.5
34.0
46.8
50.5
36.3
The nine data points on the table allow a two- variable regression
equation to be constructed.
The model is
In p/(1-p) = a + bNQ2 C(N02) * bfct + u,
where p is the tabular entry (expressed as a fraction), C(NO-) the NO,
concentration in ppm, and t the time of exposure in years. For school
children, the results are
In p/(1-p) s - 0.96 + 5.47 C(N09) + 0.073 t
(6.44) * (0.129)
Neither .coefficient is significant. (Standard errors in parentheses.)
For infants, the results are as follows:
In p/(1-p) = - 1.60 + 8.47 C(NO?) + 0.42 t
(2.18) * (0.044)
-------
A-10
These results are significant for infants but not for school children, which is
the reverse reported by Pearlman on the basis of a chi-square test.
One problem with the study was the high correlation between participates
and NOg (r a 0.999). Moreover, the particulate concentrations varied over a
range in which a health response cannot be ruled out.
The second report from the Chattanooga study, Shy et al. (1970), allows
the relative importance of N02 and participates to be investigated. This
report includes illness from a fourth area, a "high particulate" area, with NO
concentrations of 0.055 ppm and particulate concentration of 99 ug/m . In each
area Shy reports the rate of respiratory illness (number of illnesses per 100
persons per week) for second graders, their siblings, mothers, and fathers. The
study uses statistical methods similar to Pearlaan et al., and concludes that
NQ£ significantly affects illness rates for some population groups. However,
when the illness rate is regressed on NOg and participates, while controlling
for the individual's status in the family, the following results are obtained:
Variable Parameter Standard
Estimate Error
Intercept 7.03 2.67
Second graders (0-1) 9.70 1.07
Siblings (0-1) 5.28 1.12
Mothers (0-1) 2.86 1.23
NO, (ppm) - -1.86 2.94
Particulate ug/mgj 0.058 0.030
The particulate variable is significant at the ten percent level and has the
right sign; N02 is insignificant and has the wrong sign.
Shy, et al also adjusted their illness rates to account for family size
and sibling order by regressing illness rates on these two variables and using
-------
A-11
the regression coefficients to adjust illness rates. If the adjusted illness
rates are regressed on the five variables above, the results are as follows:
Variable Parameter Standard
_ Estimate _ Error
Intercept 3-^3 2.50
Second graders 9.88 1.06
Siblings 6.63 1.06
Mothers 3.20 1.06
NO- 3.28 2.67
Particulates 0.060 0.029
N02 now has the right sign, but is still not significant.
In view of the inaccuracies of the Jacobs-Hochheiser method, these
results are probably not reliable. They are reported to illustrate how the use
of alternative and more appropriate statistical procedures can completely alter
the results of epidemiological studies.
10. Shy and Love (1979).
A new study of the link between NO- and respiratory disease was made in
1972 and 1973 by Shy and Love. From January to May in each of these years,
data on respiratory disease frequency was obtained from the same neighborhoods
in Chattanooga as in the earlier study. These data were matched against
pollution measures made during the same years. The Saltzman method was used
for measuring N02. Sufficient data are given in the report to use regression
analysis to relate disease frequency to pollution concentrations. Regressions
analysis of the data in fact showed that elevated M02 concentrations were asso-
ciated with increased acute respiratory disease incidence in preschool child-
ren. However, subsequent to the data analysis, errors were found in the data
reported by Shy and Love. The most serious is that personal illness rates of
one period were matched with N02 concentrations from another period. Thus, the
-------
A-12
regression results are invalid and are not appropriate for use as dose-response
functions.
11. Shy et al. (1973).
After the biases of the Jacobs-Hochheiser method became known, Shy et al.
attempted to retrieve the situation using an N02 data base compiled in
1967-1968 as part of an interstate air quality study for Chattanooga and
Rossville, Ga. The monitoring stations used for this study were not, of
course, located at the same sites used by the Shy study. Figure 1 shows the
relationship between the monitoring network of the interstate study and the
neighborhoods of the Shy study. In Figure 1, the circled numbers are
monitoring stations. H1, H2, H3 are the three elementary schools in the "high
N02" area, and C1, C2, and P are respectively, the "intermediate NOj," "low
N02" and "high particulate" areas. "VAAP" is the Volunter Aray Ammunition
Plant, the high-NOx source.
In their report (Shy, 1973)» the researchers attempted to use the 1967-1968
monitoring data to construct isopleths of N02 ambient concentration at the 90th
percentile. The report does not indicate the method used to generate the
isopleths. These isopleths are shown in Figure 2. Concentrations at each site
could then be estimated by their location relative to these isopleths, e.g.,
concentration in the high-particulate area is between 0.05 and 0.10 ppm. For a
point estimate one can take a linear interpolation based on the distance
between the isopleths.
Alternatively, concentrations in all the neighborhoods except C2 can be
estimated by averaging concentrations from appropriate monitoring stations.
This procedure allows estimation of the mean concentrations as well as the 90th
percentile hourly concentrations.
-------
A-13
Figure 1
5MAMICTC.N CCU.N7Y
-------
A-14
Figtire. 2
O SAMPLING
SiT£
A SCHOOL
CHATTANOOGA MQ2 ISCPLSTH.
-^ PSXGJITrLT HOURLY AVS5AGSS
-------
A-15
Table 2 compares the estimates of 90th percentile concentrations made
using the two procedures described above with the concentrations determined by
the Jacobs-Hochheiser method in the original study. For the Saltzman data,
isopleth interpolation and the averaging of nearest monitors give results that
agree closely. The observations based on Jacobs-Hochheiser, however, are
considerably less than the Saltzman data for H1, H2, H3» and C1 and slightly
less for P. No comparison is possible for C2, although Shy believes the
concentration in C2 to be less than 0.05 ppm (see Table 6 in Shy (1973a)). The
difference in concentrations for the two methods is attributable either to a
difference in measurement methods or to the fact that readings were taken in
different years. The difference in measurement periods could cause differences
in ambient N02 readings only if there was a difference in NOg emissions from
the plant or if there was a difference in meteorological conditions. However,
neither of these conditions seems to hold. Production at the VAAP in 1969 was
four percent less than in 1963 (Shy, 1973a, Table 3), not enough to account for
the difference in readings. Differences in meteorological conditions would
presumably affect other pollutants. However, concentrations of TSP, sulfate
and nitrates taken in downtown Chattanooga show no dramatic decrease from 1963
to 1969; in fact, nitrate concentrations increase substantially (Hasselblad,
1977).
Therefore, it can be concluded that ambient concentrations in 1963 and
1969 were roughly the same, and that reported differences are attributable to
the measurement methods. In view of the reported difficulties of the
Jacobs-Hochheiser method, it appears that the Saltzman readings are the more
accurate estimates of N02 concentrations in 1969.
The N02 concentrations given in Tables 2 and 3 can now be used to explain
incidence of respiratory disease among children in Chattanooga. The OLS
-------
A-16
tail. 2
Sstiaated Concentrations At Six Chattanooga Study Sitaa, 1968-9
Ccaparison of 9Ota Persentile Valuta
'•- ' '* Cppa)
Saapling
Period
Site
HI
32
H3
Cl
C2
a
Hoeaaeiaaer
10/63 - 4/69
0.242
0.141
0.093
0.096
0.069
0.037
Isopleth
Interpolation
0.55
0.23
0.15
0.13
0.05
0.09
Average Value
Nearest Moni
0.54
0.20
0.16
0.10
-
0.09
Monitors
20,201,19
15,17,16T
15
I6,l5l,5d
a. Data froa Shr «t al . (19706)
8. Data froa Shy (I973a). Coaput»d froa linear lnt*rpola:iaa of iaaplet^s en
e. Data froa Shy (1973a). irsraia of reported values froa aonitars icii.eat
d. Xo data siven for aoaliars 3,9, or * ; 5 used aa surrogate.
-------
A-17
T*bl» 3
Satimattd Cone«ntr»tiona.At Six Chattanooga Study 3it«a, T963-9
-, • -' Cospariaaa of Maaaa
(ppa)
Sanpliag
?«riod
Sits
HI
H2
H3
CT
C2
?
Jacobs
Socftaaiaaer1
10/63 - 4/69
0.019
0.078
0.062
0.063
0.0*3
0.055
S*itzaaa (9/6T - 11/68)
Art rig • Values Froa
H«ar«at Monitors
Cone»Rtra?ion
0.17
0.09
o.oa
0.05
0.04
Used
20,201,19
15,17,151
15
16,161,5
S«« 'aotaa at tad af TaBl* 7.
-------
A-18
technique is used, although uncertainty about the measurement of N02 suggests
that the use of the errors in variables model may be appropriate.
Variables are defined as follows:
ILLNESS: Rate of respiratory disease incidence, per 100 person-weeks
SECONDS: 1 if cell refers to second graders, 0 otherwise
SI3S: 1 if cell refers to siblings, 0 otherwise1
MOMS: 1 if cell refers to mothers, 0 otherwise
Mean hourly N02 concentration, ug/m
N0290« 90th percentile of hourly N02 concentrations, ug/m
PARTICM: Mean hourly particulate concentration, ug/ar
PARTIC90: 90th percentile of hourly particulate concentrations, ug/nr
SIBS contains data on all siblings of second graders, both older and
younger.
Tables 4 and 5 exhibit the data, means, and standard deviation of the
variables, and the correlation matrix. Table 6 gives the coefficients and
standard errors for various specifications of the model. In equation 1 both
mean and 90th percentile pollutant concentrations are included to see if the
effects of mean vs. peak concentrations can be disentangled. However, the high
correlations between the mean and 90th percentile concentrations of both
pollutants evidently prevented this disentanglement. In the other four
equations each NO- variable is entered with each particulate variable. In the
four equations the pollutant coefficients are fairly stable. The coefficient
for mean N02 i3 about four tiaes greater than the coefficient for the 90th
percentile. This is expected, for when N0290 is regressed against N02M alone,
the resulting equation is
-------
A-19
; J ' ! J : I !
21 I
Ifl
; !
I |
1
' '
OIO — O O
r1
c
r
1
i 1
!
I i
! t
1
,-^-., 1
ifi Cl— OC O
r S 1 1
V
>3-
0)
i— I
_KJ
e*<
C Oi O O
OO.'OOIOOJOOOOIOS
03 ** 4ap ^ o —
AJ *** * tr -c *»
i
1
•
j r 1
O —'O OIO OIO OlO — O O
Ml »"M . tftl
1
1
i i
— OIO OiO O
^J < f*»
(^ I«W
1 i
i ;
1
1
;
I
Ml*
1
|
•
i
j
:
O Oi— O'C O
1 if1
I M
|
i ;
l
; i
i 2
\
r* mim i«ic Ole ccic e:o cc
0 ,'o 0,0 £
O OI
<*• C O O>u^ CIO O
9- P~
i
'. '
i i
1 .
C c-CC eei^* f"!^ m>c C:O CI*S *N* N *Vne CIC 051
1
1
t 1
O Qi O !Z |O Ot^ Oim m
• ^
i ~*
'ac
l ^
a.
f
!
i
i
0 0
o a
2 °
O 0
O Iff"
a M
« ^
O l^
^y cc
10
^ CMC C"« 0
t
f
1 ,'
j
i
O O'O O:-y *•
1
j
(
1
i
i t
l
•" M|^ ^'ir -o
i
\ '
C C'l^ Ol— (SJ
1
•
'
1
^ rsiirs* 041 <> c*
P« ftfc.pfc ^-C* 9"
,
'
| '
j i
j !
i
ff> ^
c- c>
O O'
I
1
i
*• e
-------
A-20
!
: a a
' a a
a a
i o a
: e a
a a
2 §
m —
CD
• 1
CDi
i
ai a
ei e
ei e
. oi e
ei e
ei e
ei e
mi a
IM If
IM:
(•»'
. .
,
^i 4
»•- ^
i f •— i
cxi «a
j mi IM
i •a: *•
IM, +•
' . *
«•< a
j
'
(
!
!
,; !
1 !
' o, o
\ o1 e
! a a
: 8! O
: e e
1 if i e
~. u*
• « a
i —
1
1
i
j
i
!
I
1
8! O
iv rv
i
1
i WH (/!
> vn C
u iui Z
• Z! C
J -J W
: _1 LL
a
a
S
e
O
§
—
o
§
e
e
e
e
a
e
^
^
-o
«•
•a
IM
^
•
a
e
e
8
o
a
Q
if
o
o
~
£
*•
o
a
a
a
a
a
a
e
—
a
S
a
e
e
e
e
e
«••
^
^
••
IM
*•
^.
•
O
a
a
a
e
if
e
e
ts<
_
z
a
a
a
a
a
a
a
IM
m
a
a
a
e
a
a
a
a
if.
^*
e
e
e
e
a
e
e
e
+.
•r
IM
IM
c-
IM
•0
IM
a
a
.
0
e
§
e
c
a
IM
l£
e
IM
r
IM
i
a
a
a
a
a
a
e
a
o>
«•
a
••
a
a
a
a
a
a
e
8
a
'"•
e
a
e
e
e
e
e
f
a
*
CD
0*
IM
hO
•0
e
m
!cc
<9m
•
•r
r**
•
a
e
a
o
e
e
e-
IM
O
J
O
IV
a
IM
S
a
a
§
a
o
a
a
O"
9*
a
o
a
e
a
a
a
a
m
*
a
8
O
8
a
a
e
e
IM
tf
•0
S
e
IM
m
•c
Q
•
^
e
a
a
e
o
o
e
•o
•M
«C
a
IM
Z
w
gt
«
&
a
o
e
e
a
a
m
a
e
a
a
e
e
e
e
i
e
e
a
e
e
e
e
IN
If
C-
CE
m
^\
if-
A
—
»
•
8
e
a
c
e
e
e
c.
3
a
IMI
e
sr
a.
c.
a
a
a
a
a
8
§
M
f^j
f*
O
§
a
e
e
e
a
_
«•
*
o
r*
a
cc
s
•
^
9
a
e
a
^
a
a
in
e
*
o
IM
e
z
IM
X
IM
I
e
a
a
a
a
§
MB
rsj
m
O
§
a
a
e
S
a
m*
«•
ee
.
•
^
8
a
a
e
o
e
if
IT
e
•iP
O
r*
fl^
X
Z
i*
e
DM
M
z
x
N
e
jj
Of
a
9f
U*
C
§
o7
^
e
c
C£
0.
N,
VI
JJ
U
••
Ik
u.
—
u
z
e
*
u.
O£
er
c
u
*• w
CO
•"
1^
K
IM
C
z
o
z
IM
X
C
2
e
Of*
u
••
^»
ec
«
*•
z
u
~
oc
^
e.
e
IM
c
Z
z
IS,
C
2
VI
i
z
ts\
g
•H
VI
a
§
Sr
ir
r
mU
1
tO O: flD'*»
fin ^} i |^ i^
(*s Mi 'W'fW
CD (Nil SSirg
fy «i IN- •
a'a: o*:0
I 1 !
I '
— «) mi—
ee: »"'e
IM 8| O'O
a at tfiia
a ; a:
ce oi eia
e if : 818
^ •& i 8 '8
m if.' e:e
«. •! 81 •
• Ol •••»
o i ai
1 i
r» mi ese
if. if. 818
m tfN a,o
ff. r~' O.a
O m1 8:8
IM •' 81 •
. o •—
e i a.
1
m CM cva
S Si S!I\
•e f*j| JJ;—
. ei « miif.
IM mi mi»»
fM *' m* *
• ai ••a
a ! ai
j i
i i
a o, o o
•a ai e:a
IM Oi 8:8
1- •, 81 •
o°; j°
! 1
i
e ei i"*""*
e ei e'8
C O' ^ ' O
• e: " , IS"
V> 81
U. i Z'
z tri
•I i <->.
J U."
_ uv
, i
1
m —
f- o
a a
if a
CO •
• e
e
CO '-
IP If,
m IM
CO IM
• e
a
t
e e
e e
e o
e a
e •
e~
e a
e o
a a
8 8
8 •
• M
a
e e
e e
c a
e •
e
a e
s a
C 0
0 C
e •
8
m o
m —
m —
• 8
a
e a
e o
e e
2 °
~ a
m Q
S-
m •
c°
1
«r ir
^ '8>
* o^ ^ *
IM m m »m.
^ >M, f^.8
m ., f •
• 8! "O
a i e>
IT CO — ' —
* o , sc ' cr*
if, m> —p.
f" >M 88
• Q> »,C
6 , a
IM — , oie
e £' c'c
c c c e
e- ., a •
e — •
a e IM.—
C O , rf"» O
C 8 O.C
o c. s.o
c •• e- .
• 8, «'8
— a,
a o, a.o.
a a- a'8
ac- c c
88. 88
. — , .._
a e<
•
o e a-c
e c C'O
o o a e
a a. a.a
C ., 8' •
e i ai
I i
§ai eia
8! O'O
a a< 818
e ai aie
°.j' °.:-
0 ; 0|
1
miJ i«.'
m if' die
o <"i cr-*
• 81 "O
O 1 81
j 1
1
1 1
O'
r ! 0-"
(\( <%,
C 1 C'
z . z;
1
1
1
Q* *+ • O^ C
^ •• ^ ^
•^ *iO ^ O
f+ f^ ' £ f^
e •• C- •
* o «o
O O i
C* '*" O* O
r^ i** C>* O
>^ O "^ 'Q
c • o, •
* o «o
0 0,
rw — C C
f^ O • C C
e c c c
e o co
C* •! O •
• O ( • O
o , —.
O C rg,^
C O f^ iO
80. 00
88' CC 8
c •• a- •
• c ; •«
— c
— — • if e-
CC ff" IT 4"
ac ^ a
» • * •
• C * '8
a a
If C . IM i^
>£> m IM m
f~ M, •»• —
IM • - m •
• 8 • 8
a a
a s oo
o o so
88 O.C
C • ' 8 •
8 O
80 88
C S 88
8 Si 00
O O 8,O
e • 8 •
e a^"
8 C SO
oo a a
O Oi 88
e a; 88
8*1 8 •
C , C
C
C -O
e- c
a ••
• a
e
C* !*•
m — .
IM O'
a •'
• 8
a '
C r«
i£ r^
If- IV,
IN •
8
a 8'
A, «£
C IM
IN, •
• O
S ^""'
If, vf-
•^ "^
CE IM.
• O*:
Q
|
or r-.
IT u*
m IM
S rv.,
IM •
• 8
8
l
•
8 8,
u- 8i
e ••
i
e e
C C'
• c
c
c
z
•M
<£(
r*
^
Z
,
'
;
-------
A-21
«« «
«• i
u
U
••*
««
I.
.3
3
<3 -'
e
•* f -*
* —
i e ••
•W *• 3
a '3 A
"* ^ **
i a -
^ ^ ai
^ ^
- -j e
* *»
• a
e e
• o
a .
99
• a
3
=
• o
•C t
•J "«/^
TJ — -O
e *r«
O O*
O*
^ ^ o *»
-------
A-22
N02go a - 132 * 3.5 NOjM, R2 s 0.96.
Based on F values, adjusted R squares and the fact that equation 1 exhibits
evidence of multicollinearity, the model that seems to most closely fit the
data is equation 5. Here the N02 coefficient is positive and significant at
the ten percent level. However, these results must be viewed with caution
since the concentration values are from an independent air quality study
conducted earlier than the respiratory disease survey.
12. Conclusions
As was shown above, the 1968-69 Chattanooga survey data cannot be used to
develop dose-response functions because NO^ concentrations were measured using
the Jacobs-Hochheiser method. -While the reanalysis of the 1968-69 data using
interstate air quality survey data supports the Shy and Love contention that
higher levels of No2 are associated with increased respiratory disease
incidence, the dose-response functions are of questionable value since the air
quality data used are from different locations and time periods than the
respiratory disease data. Finally, the 1972-73 regression results are of no
use because of the reporting error in the study from which the data were taken.
However, further analysis of the 1972-73 data may lead to credible dose-
response functions for benefit analysis. Such an effort was not undertaken
here due to project time and resource constraints.
B. INDOOR EPIDEMIC-LOGICAL STUDIES
All of the available "gas stove" studies were reviewed for their
usefulness in the construction of dose-response functions. Two studies by
Melia et al. (1977 and 1979a) and two studies by Speizer et al. (1980 and
forthcoming) may be useful in this regard. As these studies only investigate
-------
A-23
the link between gas/electric stove use and health, studies that measured
exposures to N02 in homes with gas or electric stoves were also reviewed.
1. Keller (1979)
Keller (1979), in a study sponsored by the American Gas Association,
analyzes data from 441 families from a suburb of Columbus, Ohio, and 146 Long
Island households. Lutz (1977) extends this study somewhat. Mitchell (1974)
and Lutz et al. (1974)—cover the same material as Keller (1979) and Lutz
(1977). Keller finds no association between cooking fuel and respiratory
disease. His study is flawed because it uses symptom variables such as chronic
lung disease, presence of cough, and presence of asthsma as explanatory
variables, thereby obscuring any possible relationship between N02 and
respiratory disease. Further, the "children" subgroup used by Keller was "less
than 12 years old." Critical information is lost by this aggregation. Speizer
(1980) found effects of N02 only in children under 2, and Melia (1979a) found
larger N02 effects for children under 8 and, in a separate study, declining
effects as a child grows older. Thus, the Keller study offers only the weakest
contradictory evidence to Melia and Speizer.
However, one objection to Keller (1979) made by Speizer (1980) does not
hold up. Speizer states that Keller's sample size is too small to pick up any
N02 effects. This statement is merely a truism because any increase in sample
size narrows the sample variance. In fact, Keller found no difference (as
opposed to no statistically significant difference) between incidence rates in
gas and electric stove homes. Thus, increasing the sample size, assuming mean
incidence rates remained unchanged, would make no difference in Keller's
result. Only if bias in Keller's sampling technique could be shown could his
result be questioned on this statistical point.
-------
A-24
2. Lutz (1977)
Lutz (1977) repeats Keller's earlier work using 120 of the UM1 Columbus
households and addresses problems with the reporting of respiratory illness by
subjects in the original study. Lutz (1977) comes to the same conclusions as
Keller (1979) and validates the ability of subjects to report accurately on
respiratory illness.
3. Melia (1977), and (1979a), and (I979b).
From 1973-1978, Melia et al. conducted three studies of the relationship
between J^ exposures in the home and respiratory disease in children. Melia
(1977) reported on respiratory disease incidence from a sample of 5,758
children aged 6 to 11 years from 28 randomly selected areas of England and
Scotland. The presence of a gas or electric stove in the home served to
differentiate NOg exposures while respiratory disease rates were obtained from
a questionnaire filled out by parents.
Information about episodes of bronchitis and asthma and the experience of
respiratory symptoms over the previous 12 months was requested. What appears
to be Chi-square tests were performed on sub-populations divided on the basis
of age, sex, social class (occupation), latitude, and urbanization. Other
factors—number of siblings, overcrowding, smoke, and S02 levels—were also
used to explain respiratory rates but small sub-sample sizes led to more
inconclusive results.
Melia (1979a) reported on an analysis of respiratory disease incidence in
children from the 1973 study through 1977. Data on 2,408 children were
available for all five years. In addition, the 1973 study was replicated for
4,827 different children aged 5 to 10 from 27 randomly selected areas in
England and Scotland. It appears that these areas are identical to those in
-------
A-25
the 1973 study. The replication added variables on the presence of a gas water
heater, the use of pilot lights, smoking habits, and the number of bedrooms per
person.
Melia et al. (1979b) took a different tack by actually measuring mean
weekly N02 levels in the home (kitchen and bedroom) and relating this measure
not only to respiratory disease but also to lung function. Adding lung
function as a dependent variable is important because it is primarily through
changes in lung function that respiratory disease rates may be affected. Five
hundred twenty-seven children aged 6-7 years were included. The confounding
factors of age, sex, class, and smoking were included.
Data reporting was sufficient in Melia (1977) and Melia (1979a) to conduct
multivariate analysis by transforming incidence rates in each cell to log -
odds ratios. Regressing incidence rates on dummy variables for age, sex,
social class and presence of gas or electric stove yielded significant results
(at the 95 percent level) for all four variables in Melia (1979a), including
the gas/electric stove dummy variable. Variables in parentheses are standard
errors:
Illness = -1.169 * .213 N02 + .212 CLASS - .490 AGE + .258 SEX
(.077) (.069) (.069) (.069) (.069)
n = 16
r2 s .88, F = 20.8
where
Illness s in P/1-P, P a percent of sub-sample with respiratory disease
NOX s 1 for gas stove
AGE = 1 for 8 years old or older
CLASS s 1 for manual labor
SEX s 1 for male
-------
A-2 6
The other variables in Melia (1979a) are of expected sign with younger
children showing greater illness incidence than older children and children
with parents in manual labor showing greater incidence than those with parents
in non-manual labor.
The results for regressions using data in Melia (1977) tell a similar
story with regard to NOX and CLASS, but AGE and SEX are insignificant. The
coefficient on NOX in the 1977 study is nearly double that in Melia (1979a).
The regression results using data from Melia (1977) are:
ILLNESS = -1.249 + .331 NOX - .271 AGE * .394 CLASS - .023 SEX
(.145) (.129) (.129) (.129) (.129)
n « 16
r2 = .65, F - 5.05
Because both of the Melia studies found significant differences in health
effects between the sexes, separate regressions were run for boys and girls
using data from each study. The NOX relationship to illness is significant for
boys, not girls, a finding at variance with Melia (1977), where significant
effects for both sexes (larger for girls) were reported. The regression
results are troublesome because girls (at least the older ones) may spend more
time in the kitchen and, therefore, should be exposed to greater NOX levels
than boys.
The results for the Melia studies suggest that they may be useful for
constructing dose-response functions. However, many qualifications should be
emphasized. Neither Melia study distinguished between homes using coal (town)
gas, natural gas, or propane. The constituents of these gases may differ
-------
A-27
greatly. In 1973, the data collection year of the 1977 study, some children in
the study areas were exposed to both coal and natural gas. Melia notes that a
changeover from the former to the latter fuel was taking place in 1973 in most
study areas. As the 1979 study surveyed children from the same areas in 1977,
it may be presumed that this problem is less significant in the later study.
Another problem concerns the specification of the respiratory disease or
symptom variable in both studies. Parents were asked to recall chronic disease
symptoms over the previous 12 months. A "chronic" symptom was defined as one
that "usually" appeared, but the definition of "usually" was left to the
respondant. If no chronic symptoms were present, the variable took a value of
"1." It took a value of "2" when a chronic chest cold, wheeze, or cough was
observed, or when the subject had asthma, bronchitis, or other respiratory
illness. It took a value of "3" if two of these chronic symptoms were
observed, and a value of "4" if more than two were observed. In the
econometric work reported above, a dummy variable for presence or absence of
symptoms or disease was used because it is unclear that a person is more
incapacitated with, for instance, a chest cold and a wheeze than only a chest
cold.
There are additional problems with using the results of the regressions
reported above. To derive a dose-response function from this information, one
first would need to interpret how parents defined "usually," as in whether the
child "usually" coughed in the morning. Then, from this definition, some
quantitative measure of sickness—e.g., the number of sick days—would have to
be constructed. It is unclear how these tasks could be performed
non-arbitrarily.
-------
A-28
Moreover, the above regressions ignore many of the variables tested by
Melia as well as some important variables that were not introduced, e.g.,
ambient NOg, and humidity and temperature in the home. The indirect NOp
measure (i.e., gas vs. electric stove use) is extremely crude and, of course,
no actual measurements of N02 were taken. Further, as reported by Melia, the
significance of the NO^-disease relationship appears to fall off as more
variables are added.
In addition, the longitudinal study, Melia (1979a), casts some doubt on
the significance of childhood respiratory illness to health in later years:
the NO^ effect tended to disappear as the children grew older.
The third Melia study (1979b) also damages the credibility of results from
the other two. This better-constructed study found no association between
impaired lung function and measured N02 in homes. Siailarly, no association
was found between NOg in kitchens and respiratory disease symptoms. However,
the hypothesized association did arise with NO- measured in the bedroom.
At the end of this third paper, the authors note that "quantitative
conclusions about the exposure-response curve should not be drawn from the
statistically weak associations in this paper," (p. 352). However, no similar
comment accompanies the Melia (1977) or (1979a) papers.
-------
A-29
4. Speizer et al. (1980).
Speizer et al. (1980) reported on a study of respiratory disease symptoms,
lung function, and indoor N02 exposures using a sample of 8,120 white children
6-10 years old from six U.S. cities. Interactions between disease, age, sex,
city, social status (2-way), air conditioning, parental smoking, and cooking
fuel (gas vs. electric) were investigated using a variety of techniques,
including a multi-variate estimation procedure with unreported results.
Questionnaires on child health history were used to specify the respiratory
symptom variables: history of bronchitis, serious respiratory disease before
age 2, and respiratory illness in the past year. Lung function related to
cooking fuel, heating fuel, air conditioning, and smoking was analyzed, using
ANOVA.
Statistics on the likelihood of having a serious respiratory disease
before age 2 have been obtained from the authors of Speizer et al. (1980).
Only this illness variable showed significant association with type of stove.
Correcting for smoking, sex, and social class, a child in a gas stove home was
20 percent more likely to have a serious respiratory disease before age 2 than
one in an electric stove home.
In a separate portion of the study, which supports the significant results
reported above, lung function was found to be significantly affected by cooking
fuel, city-cohort, and home heating fuel. However, two anomalies appeared—the
sign on smoking was reversed and significant, and residents of Topeka had a
particularly low level of pulmonary function.
The results themselves are suspect for the usual reasons: no ambient NO-
or other pollutant measures, no in-house measures of N0_, no family disease
history, a crude social status variable, and crude symptom variables (although
less crude than in Melia).
-------
A-30
5« Speizer (forthcoming).
In response to criticisms about variable specifications in the Speizer
(1980) paper, certain refinements of that study were made. Of most importance,
the social class variable was further subdivided into six classes (from two):
three levels of educational status (junior high school, high school, college)
crossed with "father present/absent." The results of the study have not yet
been released but the authors have said that the disease frequency
differential, while still significant, is now only seven percent, down from 20
percent in the 1980 study.
6. Comparing Melia et al. to Speizer et al«
Unfortunately the Melia et al. and Speizer et al. results are not directly
comparable, primarily because Melia focused on respiratory disease or symptoms
in the previous year, while Speizer focused on parents' memories of respiratory
disease in their children before they were 2 years old. Speizer ran a
multivariate test on a variable identical to Melia's but found that only
smoking explained any of its variance. Melia's results could be made somewhat
more comparable to Speizer's by running the Melia regressions only with the
data for children under 8 years old. Unfortunately, Melia does not have an age
category for children under 2.
C. INDOOR EXPOSURE STUDIES AND "INDOOR" DOSE-RESPONSE FUNCTIONS
In this subsection, studies that estimate N02 emissions from gas stoves
and those that compare emissions from homes with gas or electric stoves are
reviewed. First, the use of such information to derive a dose-response
function from the Melia studies is discussed.
-------
A-31
The coefficient on the gas/electric dummy variable (NOX) in the Melia
regressions is the difference in frequency of respiratory disease associated
with living in a gas vs. an electric home. For Melia (1979a), the coefficient
of 0.21 implies that the frequency of respiratory disease is 55 percent higher
among children living in a house with a gas stove than in a house with an
electric stove. If the difference in N02 concentrations is Xpphm, then, for
every pphm change respiratory disease frequency increases by 55/X percent.
For the regressions on data in Melia (1977), where the coefficient on the
gas/electric dummy is 0.33, the frequency of respiratory disease and symptoms
would be 60 percent higher in gas stove homes, and 60/X percent higher per pphm
increase in N02 concentrations.
1. Indoor NO,, Measurement Studies.
A value for X should be estimated in the context of a multi-variate model
where differences in indoor N02 readings are explained by many variables, among
them whether the stove uses electricity or gas. Not surprisingly, no
researchers have made such a study, although Melia (1979c) mentions her
attempts to correct for confounding factors. One group of researchers (Traynor
and Hollowell, 1978), report a convincing inverse relationship between
ventilation (air exchange rates) and N02 concentrations in a test kitchen, but
consider only kitchens with gas stoves. Indeed, most studies examine the
N0—gas stove relationship while ignoring homes with electric stoves. Unless
1. Computed by taking J_ s B, and solving for 3p/3NOX.
3NOX
-------
A-32
it is assumed that NC^ in electric stove homes matches ambient levels (as
suggested by Speizer (1980)) and reliable ambient values for NO- concentrations
corresponding to the areas studied can be obtained, the studies that focus only
on N02 measurements in gas stove homes cannot be used.
The literature cited in the Criteria Document presents indoor NOg readings
for instantaneous peaks, hourly peaks, hourly averages, two hour averages, and
daily averages and peaks. The 1979 Criteria Document puts a range of 0.25-0.5
ppm on "usual" peak hourly readings in homes with gas stoves, while noting that
levels of 1.0 ppm could occasionally occur. Unfortunately, this figure appears
to include ambient levels, rather than the difference between gas and electric
stove homes, ceteris paribus. Measurements by other researchers range from
0.13 to 1.5 ppm for peak readings.
Researchers agree more closely on the peak hourly readings found in homes
with gas stoves: from 0.4 to 1.2 ppm (again, without correction for ambient
levels). Three of the five readings tallied results based on alternative air
exchange rates. Traynor and Hollowell suggest that an air exchange of 1.0
ppm/hour is "average." In that case, a reading of 0.8 ppm was observed.
Hourly average data, appearing in Melia (1978) for England and Palmes
(1977) for the U.S. allow a comparison between readings in homes with natural
gas and electric stoves:
1 hour averages of NO, (ppm)
Gas Electric Difference
Melia 0.072 0.0095 0.0625
Palmes 0.0491 0.0083 0.0408
-------
A-33
Measurement techniques in the two studies were similar, but Melia sampled
four homes, while Palmes sampled 19. Thus, the Palmes study is likely to be
more reliable. Note that his gas stove readings are only a bit more than
one-half of Melia's peak hourly readings, while his electric stove readings are
rather low relative to ambient conditions in most American cities.
The one 24-hour peak observed for gas stoves was 0.163 ppm (Lutz, 1974).
A 24-hour average of 0.074 ppm was measured by one study (Kollowell, et al.
1980) which can be compared to a reading of 0.053 ppm for gas stove homes and
0.018 ppm for homes with electric stoves reported by another study (Lutz,
1974). However, the Lutz study used the discredited Jacobs-Hochheiser method
to measure average concentrations of NO^.
Twenty-four hour annual average comparisons appear in Speizer, et
al.(1980), which tallied N02 .readings (from 5 to 11 homes per city averaged
over 5 cities) of 0.018 ppm for homes with gas stoves, 0.011 ppm for homes with
electric stoves, and average ambient concentrations of 0.015 ppm. For four of
the five cities the geometric mean of the N02 readings in gas stove homes was
significantly different from the mean in electric stove homes at the 99 percent
level. For one city, the level of significance was only 90 percent. Speizer
also found 95th percentile readings for NOg in electric and gas homes of 0.0309
ppm and 0.0437 ppm, respectively. The average of the differences in readings
over the five cities is, therefore, 0.0128 ppm.
Based on the discussion in this section, Palmes' value of 0.0408 ppm will
be uaed as the difference between one hour average NO^ readings in gas vs.
electric stove homes for deriving dose-response functions from the gas stove
studies. Based on Speizer et al.(1980), a 24-hour average difference of 0.007
ppm and a 95th percentile difference of 0.0128 ppm will be used. While this
method of assessing damage functions from the gas stove studies provides
-------
A-34
interesting results, using the functions for benefits analysis is not
considered prudent since the actual N02 concentrations in the homes of
respondents are highly uncertain.
2. Dose-Response Results.
The final dose-response results are given below.
Change in Respiratory Disease Frequency
for 1 pphm change in NO-
One Hour
Average
14.7
13.5
4.9
1.7
95*tile of
24 hr.
Average
46.9
43.0
15.6
5.5
24 Hour
Average
85.7
78.6
28.6
10.0
Melia (1977)
Melia (1979a)
Speizer (1980)
Speizer (forthcoming)
III. MATERIALS DAMAGES
Only one study (Upham et al. 1976) reports a dose-response function for
materials fading from NOg. Beloin estimated, but did not report, a number of
such functions. In addition, only one study reports a dose-response function
for corrosive effects of N02 on metals (Haynie, forthcoming).
A. FABRIC FADING
In the controlled experiment of Upham et al. (1976), three fabrics used
in draperies were exposed over 1,000 hours to varying concentrations of N02,
S02, and Oj, at different temperatures and relative humidities, in the presence
of light. The fabrics appeared to be vat dyed rather than colored with the
more pollution-sensitive dispersed dyes. From an analysis of variance (not
reported), N02 caused only the plum fabric (100$ cotton duck) to fade
-------
A-35
significantly. Fitting the fading data to a fading function yielded the
following result:
AE - 30 f i_3~(-75+-01H20+2-9xlo'5N02-H2°)t]
where AE = amount of fading, measured by a color difference meterj
30 is the maximum amount of .fading
H2o s moisture in mg/m3 at 25°C, 1ATM
N02 = concentration, ug/m3
t = years of exposure
Seventy-eight percent of the variability of fading in the plum fabric is
explained by this equation. Haynie then takes this equation and constructs an
equation for percentage life lost (PLL) as a function of N02 exposure only..
With correction factors appearing only in the unpublished original
version of the paper, Uphan's results check out. Using these results, the
change in PLL for alternative NO- standards is:
N02(PPM) Change in PLL
0.053
2.44
0.06
3.36
0.07
3.25
0.08
Thus, weakening the N02 standard from 0.06 to 0.07 PPM will reduce
product life (to reach any given level of fading) by 3.36J.
Beloin (1972) uses multiple regression analysis to relate ambient
°3» N02, temperature, and humidity to fading in 67 dye fabrics placed in
light-tight containers in four cities. N02 in combination with other
-------
A-36
pollutants was a significant cause of fading for red and blue cellulosic
acetate (disperse dye) used in apparel, blue cotton muslin used in apparel, and
blue wool flannel used in apparel, rugs, and upholstery. Interestingly, Beloin
did not find any draperies faded by NO-.
Beloin's study could have contributed 5 or 6 dose-response functions if
he had only reported his parameter estimates. As he did not, the study can
only be used to identify the types of fabrics "at risk."
Beloin (1973) also ran (light-tight) chamber tests on 20 of the 67
fabrics in the ambient exposure study, including four of the six fabrics
showing a significant N02 effect. This time these four fabrics showed high
sensitivity to N02 at 0.05 ppm even in the absence of other pollutants. In
addition, a blue nylon and a cotton vat-dyed material show high sensitivity to
N02. Beloin notes that the cotton fabric may be a similar material to that
used by Upham. Beloin also found several other materials fading from N02 in
high concentrations (0.53 ppm). In general, the amount of additional fading
decreased with time.
No studies were found that link fading to consumer behavior. Numerous
consumer complaints of fading from gas-fired clothes dryers were made in the
1930s and 1940s, but changes in dyes and the introduction of inhibitors have
largely eliminated such complaints. In the absence of studies about consumer
behavior with respect to fabric fading, it is not clear that a product lifetime
change of three percent resulting from a one ppm change in N02 concentrations
(from Upham's function) would even be noticeable, particularly for fashionable
apparel, where product lifetimes may be shorter than physical lifetimes because
of changes in fashion.
-------
A-37
B. FABRIC YELLOWING
In the course of this research, a DuPont official was queried about the
yellowing of fabrics, primarily those made with spandex (the generic name for
DuPont's Lycra). This discussion revealed that, because of DuPont's concern
about the yellowing of women's undergarments, in-house research is being
conducted on the problem.
C. CORROSION DAMAGE
Haynie (forthcoming) constructed dose-response functions by drawing on
data from field experiments in St. Louis covering corrosive action in metals
exposed to air pollutants, temperatures, and wind. He found that NO^ exposures
had a significant corrosive influence on zinc. NOg also was a significant
cause of steel corrosion, although co-linearity between temperature and NOj
prevented the construction of a dose-response function.
Haynie used multiple regression techniques and reported values of
coefficients and standard errors. The results of one of Haynie's zinc
equations is
Ns2974
R2s0.560
F=943.7
Independent
Variable
Coefficient
Standard
Deviation
F
RH
0.1000
0.0017
3462.2
E
-5.460
0.323
286.0
U SO,
Constant
-0.00150 0.00493 9.9242
0.00056 0.00051
7.0 93.5
where RH = relative humidity
inn ~[6'32(temperature-dewpoint)/(80-fteniperature) ]
-------
A-38
E s temperature factor
1000
ave. hr + 273«16
temperature
Og s ave. hr concentrations (ug/m )
U s Deposition velocity
0.3025 (V- 0.1875V0'9275) °'75A
where V s wind velocity at 30 m above ground level (m/sec)
The U variable is used because Haynie feels that pollutant flux represents
exposure better than concentrations.
Haynie measured galvanic corrosion by changes in the amount of electric
current measured between metal plates. The high sensitivity of this technique
to environmental changes allows for a wide range of conditions to be
incorporated into the data even though equipment is set up at only four sites.
Indeed, Haynie had over 2,000 observations of environmental conditions and
associated currents.
Haynie also selected data within a very narrow temperature range to
eliminate the influence of this variable, and still found a significant N02
effect. Unfortunately, he used the corrosion rate instead of its log as the
dependent variable, making the N02 coefficient not comparable to those from
several other specifications.
He also presented a relationship combining independently determined values
from stoicheometric reactions and the regression results. By this technique
N02 and S02 combined explained only 1.3 percent of the variability in
-------
A-39
corrosion. Humidity was most important. This relationship represents Haynie's
"best estimate."
CO
0.045 U-SO. + 0.0314 U-NCL + e
4 2
17.133-4.237E
-5.4(100-RH)/RH
where CO is the best estimate of the corrosion rate for specific con-
ditions of pollutants, temperature} and RH.
IV. VEGETATION DAMAGES
The Criteria Document was examined to identify candidate studies for the
derivation of dose-response functions. Then, attempts were made to obtain all
of these studies. Studies reviewed include: Thompson, et al. (1970),
Spierings (1971), Stone and Skelly (1974), Zahn (1975), Tingey et al. (1971),"
and Reinhert et al. (1975). None of the studies can support the construction
of dose-response functions. In this subsection, the reasons for this
conclusion will be discussed.
The response of vegetation to N02 has been analyzed at progressively
higher levels of biological organization (cellular, leaf injury, plant growth,
and yield). Normal cellular metabolism may be altered, leaves may turn brown,
yellow, or become spotted, growth may be stunted (or speeded), and yields may
fall. A willingness to pay to avoid damages can be associated with all but the
first of these effects. Assessing decreased marketable yield of fruit trees
(Thompson, et al., 1970), and some vegetable crops (Spierings, 1971) requires
dose-yield functions. Dose-growth functions could be used to assess damages to
trees producing lumber and other types of marketable crops (e.g., asparagus).
-------
A-40
Aesthetic damages to ornamental plants, trees, grass, etc., could possibly be
estimated from dose-leaf damage functions.
Because this study is concerned only with valuing damages within certain
cities, consideration of damages to agricultural crops and timber stands is
irrelevant. Perhaps some vendors of trees as well as ornamental and vegetable
plants might experience some effect. But such vendors are often located
outside of cities, where land is less expensive. Rather, lower yields in home
vegetable gardens and aesthetic damages to outdoor ornamental plants, trees,
lawns, and public parks constitute the only damages for consideration.
Unfortunately, dose-response functions cannot be constructed for either
aesthetics or vegetable gardens. Nowhere in the literature have aesthetic
damages of this type been valued. Instead, researchers have focused primarily
on valuing changes in visual range (although, even here, the effect of NOp on
visibility cannot be separated from the effects of other pollutants).
Estimates of the physical effects associated with changes in NC^ concentrations
are available, but only in relatively obscure terms, i.e., in terms of acres of
area affected, degree of color change or change in the density of plant cover.
Without additional information on how the public would react to such changes,
this information is of questionable value.
The problem of valuing damages to lawns and parks is further complicated
by the inhibiting effect of N02 on plant growth. Slower growth may be valuable
because of reduced park and lawn maintenance costs. Such benefits could offset
part or all of the aesthetic damages, assuming slower growth and leaf damage
are always related.
Literature on damages to garden vegetables was also reviewed. Only a
handful of studies focus on NO- damages to vegetable plants, and these usually
focus on leaf damage rather than the economically meaningful measures of growth
-------
A-41
or yield. The studies that do focus on these responses fail to construct
dose-response functions or design experiements to derive them. Spiering's
(1971) study illustrates this problem. This controlled study of the effects of
N02 on tomato plant growth and yield produced credible, if limited, results.
Of relevance here, tomato yields were found to be 22 percent lower after
exposure to 0.25 ppm of N02 for 128 days. Unfortunately, no attempt was made
in this or any other experiment to assess the effect on tomato yield of a
higher or lower dose for the same period of time (or any other period of time).
Credible dose-response functions cannot be constructed from one point.
Moreover, the yield change cannot be used directly because a continuous 0.25
ppm N02 concentration has not been observed in the target cities. Thus, this
study and the others reviewed cannot provide estimates of damages.
One additional problem with these studies is their failure to properly
account for the effect of other pollutants, particularly SO-, on yields.
Controlled studies of N02-S02 interaction convincingly point out their
synergistic effect. N02 concentrations at or below current annual ambient
standards can damage vegetation in the presence of S02 (at or below annual
ambient standards) (Reinert, et al., 1975). However, the Criteria Document
makes it clear that such pollutant interactions have not been integrated into
studies of changes in the yield or growth of vegetation.
-------
A-M2
V. VALUATIONS
In this section studies of dollar damages to textiles faded by N02
exposure and several property value studies are reviewed.
A. TEXTILES
Haynie (1980) used a method developed by Mueller and Stickney (1970) that
appears in the Ozone Criteria Document (1978) to value textile damage.
However, this method does not use any known N02 dose-response information, and
uses a ratio of preventative to direct costs pulled from a study on ozone and
rubber elastomers.
Salvin (1976) estimates the cost of fading and preventative measures but
does not give a date for which these costs apply and poorly documents his
sources. His ad hoc approach is more acceptable than Haynie's, however. Costs
are calculated by product. Acetate fabric costs are divided into use of more
expensive dyes in Class A dyeings (quality dyes), costs of inhibitors in Class
B dyeings (conventional dyes), costs of research and quality control, fading at
the manufacturing and retail level, and reduced wear life to consumers. For
cotton, costs for three common dyes (sulfur, direct, reactive) are calculated.
Vat dyes are not considered because "they are not vulnerable to fading from
NOX." This statement directly contradicts Upham's (1976) findings, although
Upham notes that the fabrics "appeared" to be vat dyed. This contradiction is
particularly unfortunate because the only fabric dose-response function
identified by this research relates to Upham's vat dyed material.
Finally, costs of fading to consumers of viscose rayon are calculated.
Much of this material is used in automobiles. However, losses for yellowing of
white cotton and fading of nylon carpets—"for which NOj effects are clearly
-------
established"—are not calculated in this paper, although yellowing costs were
calculated in Salvin's 1970 paper.
For the loss of product life calculations, Salvin used the formula:
Loss = Sales x (percent of goods used in polluted areas x percent of
products using sensitive dyes) x (percent yearly loss in wear life from N02).
Corrections are made for substitution of more expensive dyes. However,
the loss of wear life numbers look like guesses.
For preventative costs, Salvin obtained direct unit cost estimates for
expensive dyes and inhibitors, but also made guesses on research, quality
control, and other activities.
The estimates in Table 7 are costs for the nation of increasing NC^ from
zero, or possibly "background," to "ambient" levels, not for marginal changes
in ambient levels.
The discrepancy in total cost between Salvin (1976) and Barrett and
Waddell (1973) who quote Salvin (1970) is caused by a lower estimate for
acetate costs in the more recent report. The percentages of total costs for
consumer and manufacturing losses as well as preventative measures given by
Salvin (1970) and Barrett and Waddell (1973) bear no resemblance to those
calculated from Salvin (1976). Salvin states that the 1976 work supersedes his
earlier estimates.
B. PROPERTY VALUE STUDIES
The complexities of estimating the value of pollution damages by
constructing physical dose-response functions can be avoided if responses to
pollution leave their trace on markets. The housing market has received much
-------
A-4U
Table 7
Total Material3 Damages From N02
Category $ Millions
Acetate 52.8
Substitution of Expensive Dyes 17.6
Inhibitors 13.2
Costs of Research 0.5
Quality Control 0.5
Fading—Manufacturing/Wholesale 1.0
Reduced Wear Life 20.0
Cotton—Reduced Wear 22.0
Sulfur Dye 5.0
Direct Dye 15.0
Reactive Dye 2.0
Vat Dye 0
Viscose Rayon—Reduced Wear 21.6
Total 96.4
Yellowing (from 1970 Report) 6.0
Total 100.U
-------
A-45
attention in the attempt to value pollution damages. The idea is that people
will be willing to pay more for housing in an area with good air quality than
with poor air quality. These preferences should create price differentials
between houses located in areas with different levels of air quality, eeteris
paribus.
Harrison and Rubinfeld (1978) performed one of several studies that
investigated the effect of different ambient concentrations of N02 on housing
prices. They used 1970 census tract data from the Boston SMSA to obtain the
median value of owner-occupied homes and a number of independent variables
describing housing and population characteristics. Other sources were used to
define other neighborhood characteristics. Data on NOX and participates were
obtained from a computer simulation model. The authors did not explain what
they meant by nNOx" (as opposed to N02 or NO) or how the various nitrogen
oxides were aggregated into a single concentration number.
Harrison and Rubinfeld report regressions with either a NOX or a
particulate variable, but not both. With these variables so highly correlated
(r s 0.96) they felt that virtually no information would have been gained by
using both in their regressions; either pollutant could have been used as a
proxy for "air quality." In addition, when a regression with both variables
was estimated the sign of NOX reversed and became insignificant.2
If the collinearity problem is ignored and Harrison and Rubinfeld's "best
guess" .regression is used, then a one pphm increase in NO-, lowers median
housing prices by $1613 (in 1970 $'s), if all other variables take on mean
2. With both variables in the equation, their collinearity does not
introduce bias to the estimated coefficients. However, their standard errors
increase, which raises the possibility of wrongly accepting a null hypothesis.
Further, the estimates are extremely sensitive to particular sets of sample
data.
-------
A-46
values. Using the housing value regression relationships to estimate a demand
function for NOX reductions, Harrison and Rubinfeld find that a "household
earning $11,500 per year in 1970 would have been willing to pay (WTP) $800 for
a one pphm improvement when the NOX level was .03ppm, [and]... $2200 per year
when the NOX level was .09 ppm." (p. 89) Of course, what Harrison and
Rubinfeld actually measured was the WTP for "air quality improvements," a
concept that may not be easy to relate to changes in NOX and participates.
A study of the South Coast Air Basin (Brookshire, et al. 1979) presents
estimates of the willingness-to-pay for improvements in air quality using a
methodology similar to that used by Harrison and Rubinfeld. Here, again,
collinearity between N02 and participate concentrations precluded any estimate
of the independent effect of either pollutant on air quality. For comparison
with Harrison and Rubinfeld, Brookshire et al. find that a one pphm increase in
NO- (as a surrogate for air quality) reduces the average sales price of a home
in the South Coast Air Basin by $3,270 (in 1970 dollars). For the linear WTP
equation with air quality in the 0.06-0.09 ppm range, the WTP for a one pphm
improvement amounts to $2,719 per home (in 1970 dollars).
A follow-up study for San Francisco (Loehman, 1980) uses the same basic
methods as Brookshire, et al., and adds ozone variables. However, in the
Loehman study the NO, coefficient is positive and significant, a result the
author attributes to the fact that NOj standards are "rarely exceeded" in the
San Francisco area. Alternatively, it is possible that N02 concentrations are
correlated with some housing characteristic, such as housing location in feet
above sea level, which is positively associated with selling price but is
omitted from the regression.
-------
A-47
Bibliography
Barrett, L.B. and I.E. Waddell, 1973. Costs of Air Pollution Damage, USEPA,
AP-85, NTIS No. PB-220-040.
Beloin, N.J., 1972. "Fading of Dyed Fabrics by Air Pollution," Journal of
Textile Chemist and Colorist, Vol. 4, No. 3 (March).
, Journal of Textile Chemist and Colorist, 1973, Vol. 5, No. 7
(July).
Brookshire, D.S., R.C. d'Arge, W.D. Schulze, and M.A. Thayer, 1979. Methods
Development for Assessing Air Pollution Benefits; Vol. II Experiments in
Valuing Non-Market Goods: A Case Study of Alternative Benefit Measures of
Air Pollution Control in the South Coast Air Basin of Southern California,
for USEPA Offoce of Research and Development (February).
Chapman, R.S., et al., 1973. "Chronic Respiratory Disease in Military
Industries and Parents of School Children," Arch. Environ. Health, Vol. 27
(September).
Coffin, D.L., D.E. Gardner, and E.J. Blommer, 1976. "Time-Dose Response for
Nitrogen Dioxide Exposure in an Infectivity Model System," Environmental
Health Perspectives, vol. 13, pp. 11-15 (Feb.).
Coffin, D.L., et al., 1977. "Role of Time As a Factor in the Toxicity of
Chemical Compounds in Intermittant and Continuous Exposures. Part II.
Effects of Intermittant Exposure," Journal of Toxicology and Environmental
Health, Vol. 3, pp. 821-828.
Cohen, C.A., et al., 1972. "Respiratory Symptoms, Sperimetry and Oxidant Air
Pollution in Nonsmoking Adults," American Review Respiratory Disease,
Vol. 105, p. 251.
ECAO, 1978. Air Quality Criteria For Ozone and Other Photochemical Oxidants,
ORD, USEPA, Washington (April).
., 1979. Air Quality Criteria For Oxides of Nitrogen (Draft), USEPA,
(revised June).
, 1981. Air Quality Criteria for Oxides of Nitrogen (Draft), USEPA,
(revised November).
Harrison, David Jr. and D. Rubinfeld, 1978. "Hedonic Housing Prices and The
Demand for Clean Air," Journal of Environmental Economics and Management,
Vol. 5, No. 1 (March).
Hasselblad, V.,1977. "Lung Function in School Children: 1971-1972 Chattanooga
Study," USEPA, EPA-600/1-77-002.
Haynie, F., 1980. Effects of Nitrogen Oxides on Materials, (unpublished
paper).
-------
A-48
(forthcoming). Evaluation of the Effects of Microclimate Differences
on Corrosion.
Heck, W. W. and D. T. Tingey, 1979. "Nitrogen Dioxide: Time Concentration
Model to Predict Acute Foliar Injury," U.S. E.P.A., Corvallis, Publication
No. 600/3-79-057.
Hickey, et al., 1970. "Ecological Statistical Studies on Environmental
Pollution and Chronic Disease in Metropolitan Areas of the U.S.," RSRI
Discussion Paper #35, RSRI, Phila., Pa.
Hollowell, C.D., et al., 1980. "Indoor Air Quality in Residential Buildings,"
Lawrence Berkeley Laboratories, University of California, Berkeley.
Johnson, Ted and R. Paul, 1981. "The NAAQS Exposure Model (NEM) and Its
Application to Nitrogen Dioxide," prepared by PEDCO Environmental, Inc.,
for USEPA, OAQPS, SASD. (May).
Kagowa, J. and T. Tayama, 1975. "Photochemical Air Pollution: Its Effects on
respiratory Function of Elementary School Children," Arch. Environ.
Health. Vol. 30, p. 117 (March).
Kalpazanov, I., M. Stamenova, and G. Kurchatova, 1976. "Air Pollution and the
1974-1975 Influenza Epidemic in Sofia," Environmental Research. Vol. 12,
pp. 1-8.
Keller, M.D., et al., 1979. "Respiratory Illnesses in Households Using Gas and
Electricity For Cooking," Environmental Research, Vol. 19, pp. 495-515.
Keller, Martin, 1980. "Population Studies of the Health Effects of Atmospheric
NO- Exposures," Batelle Laboratories, (June 19).
Linn W., et al., 1976. "Respiratory Function and Symptoms in Urban Office
Workers in Relation to Oxidant Air Pollution Exposure," American Review
Respiratory Disease, Vol. 114, p. 477.
Loehman, E., D. Boldt, K. Chaikin, 1980. Methods Development for Assessing Air
Pollution Benefits: Vol. IV Measuring the Benefits of Air Quality
Improvements in the San Francisco Bay Area: Property Value and Contingent
Valuation Studies, for USEPA, Office of Research and Development.
Lutz, C. A., et al., 1974. "Indoor Epidemiology Study," Batelle, Columbus,
Catalog 0M50677.
, et al., 1977. "Respiratory Disease Symptom Study," Prepared for
American Gas Association by Batelle Columbus Laboratories. (February).
MacLean, D.C., 1975. "Stickstoffoxide als Phytotoxische Luftverunvenigungen,"
Staub. Reinhalt. Luft., Vol. 35, pp. 205-210.
Melia, R.J.W., et al., 1977. "Association Between Gas Cooking and Respiratory
Disease in Children," British Medical Journal. Vol. 2, pp. 149-152.
-------
A-49
, et al., 1978. "Differences in NO- Levels in Kitchens with Gas or
Electric Stoves," Atmospheric Environment, Vol. 12, pp. 1379-1381.
, C. Florey, and S. Chinn, 1979. "The Relation Between Respiratory
Illness in Primary Schoolchildren and the Use of Gas for Cooking," (a)
Results from a National Survey, (b) NO-, Respiratory Illness, and Lung
Infection, (c) "Factors Affecting NOg Levels in the Home," International
Journal of Epidemiology, Vol. 8, No. 4.
Mitchell, et al., 197*1. "Household Survey of the Incidence of Respiratory
Disease in Relation to Environmental Pollutants," WHO International
Symposium Proceedings, Paris, June 24-28.
Mueller, W.J. and P.B. Stickney, 1970. "A Survey and Economic Assessment of
the Effects of Air Pollution on Elastomers," NAPCA Contract No.
CPA-22-69-146, Batelle, Columbus (June).
Orehek, J. et al., 1976. "Effects of Short-term, Low Level NOp Exposure on
Bronchial Sensitivity of Asthmatic Patients," Journal of Clinical
Investigation. Vol. 57, pp. 301-307.
Palmers, E.D., Tomczyk C., and DiMallio, J., 1977. "Average NOp Concentration
in Dwellings with Gas or Electric Stoves," Atmospheric Environment, Vol.
11, 869-872.
Pearlman, M.E., et al., 1971. "Nitrogen Dioxide and Lower Respiratory
Illness," Pediatrics. Vol. 47, No. 2, pp. 391-398.,
Petr, B. and P. Schmidt, 1966. "Effect of Air Polluted with Sulfur Dioxide and
Nitrogen Oxides," Hygiene and Sanitation, Vol. 31, No. 7, pp. 111-121
(July).
Reinert, R.A., A.S. Heagle, and W.W. Keck, 1975* "Plant Responses to
Pollutant Combinations," in Responses of Plants to Air Pollution, J.B.
Mudd and T.T. Kozlowski, eds. Academic Press, Inc., New York.
Salvin, V.S., 1970. Survey and Economic Assessment of the Effects of Air
Pollution on Textile Fibers and Dyes, EPA Final Report (June).
, 1976. "Effects of Nitrogen Dioxide on Materials," Chapter 8 in
Nitrogen Oxides; Medical and Biologic Effects of Environmental
Pollutants, NAS, Washington, D.C.
Shy, C.M., et al., 1970. "The Chattanooga Study: Effects of Community
Exposure to Nitrogen Dioxide," Journal of the Air Pollution Control
Association, Vol. 20.
et. al, 1973. Reevaluation of the Chattanooga School Children
Studies and the Health Criteria for NO- Exposure, Inhouse Technical
report, Health Effects Research Laboratory, Environmental Research Center,
USEPA, RTF (March).
-------
A-50
., and G.J. Love, 1979. "Recent Evidence on the Human Health Effects
of Nitrogen Dioxide," Proceedings of the Symposium on Nitrogen Oxides,
Honolulu, Hawaii, April 4-5.
Speizer, F.E., B. Ferris, Y.M.M. Bishop, and J. Spengler, 1980. "Respiratory
Disease Rates and Pulmonary Function in Children Associated with N02
Exposure," American Review of Respiratory Disease, Vol. 121.
Spierings, F.H.F.G., 1971. "Influence of Fumigations with NO- on Growth and
Yield of Tomato Plants," Netherlands Journal of Plant Pathology, Vol. 77,
pp. 194-200.
Stone, L.L. and J.M. Skelly, 1974. "The Growth of Two Forest Tree Species
Adjacent to a Periodic Source of Air Pollution," Phytopathology, Vol.
64, No. 6, pp. 773-8.
Stebbings, J.H. and C.G. Hayes, 1976. "Cardiopulmonary Symptoms in Adults, New
York, 1971-1972," in Panel Studies of Acute Health Effects of Air
Pollution, Environmental Research, Vol. 11, pp. 89-111.
Thompson, C.R., E.G. Hensel, G. Kats, O.C. Taylor, 1970. "Effects of
Continuous Exposure of Navel Oranges to Nitrogen Dioxide," Atmospheric
Environment, Vol. 4, pp. 349-355.
Tingey, David T., R.A. Reinert, J.A. Dunning, and W.W. Heck. 1971.
"Vegetation Injury from the Interaction of Nitrogen Dioxide and Sulfur
Dioxide," Phytopathology, Vol. 61, pp. 1506-1511 (December).
Traynor, G.W.,1980. "Gas Stove Emissions," in Energy Efficient Buildings
Programs, DOE W-74-5-EN6-48.
and C. Hollowell, 1978. "Combustion Generated Indoor Air
Pollution," Lawrence Berkeley Lab (April).
Upham, J.B., Haynie, F., and Spence, J., 1976. "Fading of Selected Drapery
Fabrics by Air Pollutants," Journal of the Air Pollution Control
Association, Vol. 26, No. 8 (August).
Zahn, R., 1975. "Begasungsversuche mit N02 in Kleingewachshausern,"
Staub-Reinhalt. Luft.. Vol. 35 (May).
-------
TECHNICAL APPENDIX B
TO
CHAPTER V
OF THE
N02 NAAQS REGULATORY IMPACT ANALYSIS
OCTOBER 1984
GCA - Technology Division
213 Burlington Road
Bedford, Massachusetts 01730
-------
APPENDIX B
UPDATE OF NOX CONTROL COST ESTIMATES
This Appendix provides methods and results of a number of revisions
to control cost estimates 1n the 1982 draft Cost and Economic Assessment
of Regulatory Alternatives for N02 NAAQS. referred to hereafter as the
"CEA." (U.S. EPAf 1982). These revisions Include derivation of new
mobile and area source emission projections, review of recent ambient
data, and updating of NOX control costs for mobile and area sources, the
Federal Motor Vehicle Control Program, point sources, and new source
performance standards. In addition to the changes 1n Inputs and
methods, described 1n detail below, all costs were updated to constant
March 1984 dollars.
UPDATE OF MOBILE SOURCE EMISSION FACTORS
The draft CEA used composite NOX emission factors produced by EPA's
MOBILE2 Mobile Source Emissions Model to project NEDS mobile source
emission estimates from the 1978 baseline to 1985 and 1990. The MOBILE2
model was replaced by an updated MOBILES In mid-1984. The purpose of
the following discussion 1s to compare the originally-used MOBILE2
results with similar emission projections from MOBILE3. In addition,
projections based on a 1.5 gram/mile Federal Motor Vehicle Control
Program (FMVCP) NOX standard are added to the 1.0 and 2.0 gram/mile
standards considered 1n the CEA.
The 1982 CEA used MOBILE2 to produce the composite emission factors
1n Table 1, estimating mobile source emission reductions which would
occur due to the FMVCP. Two sets of factors were developed, one using
default MOBILE2 parameters to estimate effects of a 1.0 gram/mile FMVCP
standard for light duty vehicles (LDV) beginning 1n 1982, and the other
using modified MOBILE2 emission parameters to simulate a 2.0 gram/mile
B-l
-------
TABLE 1.
COMPOSITE NOX EMISSION FACTORS FROM JUNE 1982 CEA (grams/mile)
Year 1.0 gram/mile LDV Standard
1978 4.32
1985 3.18
1990 2.21
2.0 gram/mile LDV Standard
4.32
3.32
2.47
LDV standard. These factors are composites for a typical distribution
of vehicle types 1n 1978 Including LDV, trucks and heavy duty vehicles.
The MOBILE2 default parameters which result 1n the above emission
factors for the 1.0 gram/mile LDV standard for NOX are a base rate of
0.75 gram/mile for 1982 and later LDV, with a deterioration rate of 0.15
gram/mile per 10,000 miles travelled. The average base rate 1s less
than the 1.0 gram/mile LDV standard because most new cars will be well
below the standard. The 1982 CEA does not Indicate what parameters were
used to simulate the 2.0 gram/mile LDV NOX standard. Attempts to
recreate the CEA composite factors for this standard Indicate that a
base rate of 13 or 1.4 gram/mile and deterioration rate of 0.11 to 0.13
gram/mile per 10,000 miles were apparently used. The specific 2.0
gram/mile standard parameters used 1n the CEA are not critical, however,
since new MOBILES parameters for LDV, which were subject to a 2.0 gram/
mile standard 1n 1977 through 1980, can be adapted to simulate relaxa-
tion of the LDV standard 1n 1982.
Using MOBILES, the revised composite NOX emission factors 1n
Table.2 were generated for the default emission parameters which
represent the current 1.0 gram/mile NOX standard, and for a set of
altered emission parameters which approximate a 2.0 gram/mile standard.
The latter parameters, a base rate of 1.65 gram/mile and a deterioration
rate of 0.09 gram/mile per 10,000 miles, were based on MOBILES rates for
1977-1980 model year vehicles, which were actually subject to a 2.0
gram/mile NOX standard. In addition, a MOBILES run was made with an
B-2
-------
Intermediate set of emission parameters* to approximate a 1.5 gram/mile
NOV standard. These MOBILES composite emission factors are based on a
^\
1981 vehicle distribution and emission factors for individual vehicle
types which were modified somewhat from those in MOBILE2.
TABLE 2.
COMPOSITE NO* EMISSION FACTORS FROM MOBILES (grams/mile)
Base Emission Rate
Deterioration Rate
(per 10,000 miles)
Composite Emission Factor
1978
1985
1990
LDV NOX Standard (grams/n\1le)
2-fl
1.65
0.08
4.86
3.72
3.16
0.56
0.09
4.86
3.44
2.63
1.10
0.09
4.86
3.58
2.91
Tables 3 and 4 provide detailed information on the vehicle-type-
sped fie emission rates and the mix of vehicle miles traveled (VMT)
which were used in the MOBILE2 and MOBILES runs described above.
Attachment 1 provides more information on the difference between
vehicle-specific emission rates used in the two models. These data
indicate that the main reasons for differences between Tables 1 and 2
are the revision of emission factors for light duty gas trucks (LDGT)
and the change in the proportions of LDGV and LDGT 1n the VMT mix used.
The original MOBILE2 VMT mix, based on a 1978 "snapshot" of vehicle
registration and mileage rates, included a significantly lower propor-
tion of LDGT than the 1981 snapshot on which MOBILES is based. MOBILES
emission factors for LDGT are significantly higher than those in
MOBILE2, due to availability of additional information on emissions of
LDGT in service and to the use of an assumed 1987 1.2 gram/mile NOX
B-3
-------
TABLE 3.
MOBILE 2 EMISSION FACTORS AND VMT MIXES
Year LDGV LDGT HDGV
1.0 gm/m1le NOX Standard Emission
1978 3.23 3.99 10.05
1985 2.05 2.39 10.53
1990 1.80 1.46 5.96
2.0 gm/mlle NOX Standard Emission
1978 3.23 3.99 10.05
1985 2.23 2.39 10.53
1990 2.16 1.46 5.96
YMT Mix (percent of total fleet VMT)
1978 0.785 0.130 0.042
1985 0.750 0.125 0.042
1990 0.693 0.119 0.042
LDGV = Light duty gas vehicle
LDGT = Light duty gas truck
HDGY = Heavy duty gas vehicle
LDDV = Light duty dlesel vehicle
LDDT = Light duty dlesel truck
HDDV = Heavy duty dlesel vehicle
MC = Motorcycle
LDDV
Factors
1.60
1.14
1.01
Factors
1.60
1.14
1.01
0.001
0.036
0.092
LDDT
(gm/mlle)
1.75
1.96
1.15
(gm/m1le)
1.75
1.96
1.15
0.000
0.005
0.012
HDDV
25.64
25.64
12.95
25.64
25.64
12.95
0.033
0.033
0.033
MC
0.24
0.84
0.85
0.24
0.84
0.85
0.009
0.009
0.009
•
All
Vehicles
4.32
3.19
2.21
4.32
3.32
2.46
B-4
-------
TABLE 4.
MOBILE 3 EMISSION FACTORS AND VMT MIXES
All
Year LDGY LDGT HDGV LDDV LDDT HDDV MC Vehicles
1.0 gm/mlle NOX Standard Emission Factors (gm/mHe)
1.5
2.0
VMT
1978
1985
1990
gm/m1le
1978
1985
1990
gin/mile
1978
1985
1990
3.18
2.00
1.56
4.29
3.43
2.98
NOX Standard
3.18
2.22
2.00
4.29
3.43
2.98
NOX Standard
3.18
2.44
2.40
Mix (percent of
1978
1985
1990.
0.669
0.652
0.635
4.29
3.43
2.98
7.25
5.56
5.39
1.45
1.39
1.09
Emission Factors
7.25
5.56
5.39
Emission
7.25
5.56
5.38
1.45
1.39
1.09
Factors
1.45
1.39
1.09
1.97
1.65
1.28
(gm/m1le)
1.97
1.65
1.28
(gm/m1le)
1.97
1.65
1.28
24.
20.
15.
72
70
13
24.72
20.70
15.13
24
20
15
.72
.70
.13
0.24
0.84
0.85
0.
0.
0.
24
84
85
0.24
0.84
0.85
4.86
3.44
2.63
4.86
3.58
2.91
4.86
3.72
3.16
total fleet VMT)
0.222
0.215
0.201
0.040
0.040
0.041
0.001
0.023
0.046
0.000
0.008
0.021
0.
0.
0.
060
054
049
0
0
0
.007
.007
.007
LDGV = Light duty gas vehicle
LDGT = Light duty gas truck
HDGV = Heavy duty gas vehicle
LDDV = Light duty dlesel vehicle
LDDT = Light duty dlesel truck
HDDV = Heavy duty dlesel vehicle
MC = Motorcycle
B-5
-------
standard for LDGT 1n MOBILES Instead of the 0.9 gram/mile standard which
MOBILE2 assumed to be effective 1n 1984. Tables 3 and 4 also Indicate
that 1985 and 1990 LDGV emission factors are lower 1n the MOBILES simu-
lation of the 1.0 gram/mile standard, but higher for the 2.0 gram/mile
standard.
Two Important qualifications must be made with respect to the
composite emission factors presented above. First, the use of 1981 data
to construct the MOBILES VMT mix results In considerable Inaccuracy 1n
the MOBILES composite factor for 1978, since the relative proportions of
LDGV and LDGT changed substantially between 1978 and 1981. Thus, the
MOBILE2 composite emission factor for 1978 should be considered the
baseline for MOBILES as well as MOBILE2 projection. This approach 1s
also warranted by the fact that the NEDS mobile source emissions used as
the 1978 base!1ne Inventory are based on MOBILE2. Second, the
projections for related LDGV NOX standards 1n the preceding tables may
be quite conservative, since 1t 1s not clear that manufacturers would
actually abandon the post-1981 three-way catalyst/closed-loop control
systems and go back to the open-loop/exhaust gas redrculatlon systems
used to meet the 2.0 gram/mile NOX standard 1n 1977 through 1980.
Table 5 summarizes trends 1n mobile source NOX emission reductions
predicted by MOBILE2 and MOBILES, using the MOBILE2 1978 composite
emission factors 1n Table 1 as a baseline. Table 6 shows the net
changes 1n mobile source emission Inventories represented by the trends
1n Table 5.
TABLE 5.
NO MOBILE EMISSION REDUCTION TRENDS
PREDICTED BY MOBILE2 AND MOBILES
Composite emission reductions over
1978 MCBILE2 baseline (%)
MCBILE2 MOBILES
NO FMVCP Standard (gram/mile);
1985
1990
B-6
i-fl
26
49
2J1
23
43
m
20
39
lo£
17
33
2JI
14
27
-------
TABLE 6.
NET CHANGES IN MOBILE SOURCE EMISSION INVENTORY
BETWEEN MOBILE2 AND MOBILES PROJECTIONS (%)
FMVCP NO,, Standard; 1.0 gram/mile 2.0 gram/mile
' yv-——-i
1985 + 8.1 + 11.7
1990 + 19.6 + 28.1
MOBILE AND AREA SOURCE PROJECTIONS
Control requirements for mobile and area sources 1n the CEA are
determined by a linear rollback model, based on combined emissions for
these source catagorles 1n Individual count1es» recent air quality data,
and target ambient standards. Thus the effect of the changes 1n Table 6
on overall control requirements depends on the ratio of area source to
mobile source emissions. Since the CEA report does not provide this
Information for the specific areas 1n which controls may be required, a
survey of NOX emission Inventories 1n 1982 ozone SIP data bases was
undertaken. Table 7 shows that the relative magnitudes of area and
mobile sources do not vary greatly among the areas of Interest. Using
the Los Angeles distribution (80 percent mobile, 20 percent area) as the
extreme case, net changes 1n the overall emission Inventory used 1n
estimating non-point source control costs (mobile plus area sources)
that would occur with use of MOBILES are shown 1n Table 8.
Since the CEA report does not provide details on the baseline and
projected NOX emissions Inventories used In determining the need for
future controls, 1t 1s not possible to directly assess the potential
Impact of changes shown 1n Table 8. However, 1t 1s possible to estimate
B-7
-------
TABLE 7.
DISTRIBUTION OF AREA AND MOBILE SOURCE
EMISSIONS IN 1982 OZONE SIP
Distribution of Non-point Source NO..Emissions _(%).
Area Mobile Area Year
Los Angeles/South Coast 80 20 1979
Chicago 73 27 1980
Detroit 74 26 1980
TABLE 8.
NET CHANGES IN NON-POINT SOURCE EMISSION INVENTORY
BETWEEN MOBILE2 AND MOBILES
FMVCP NOX Standard: 1.0 gram/mile 2.0 gram/mile
1985 + 6.5 + 9.4
1990 +15.7 + 22.4
total change 1n the mobile/area source emission Inventory as predicted
by the CEA model, by using the mobile source emission trends shown 1n
Table 5, the CEA assumption of a one percent annual growth rate for
mobile and area emissions, and an assumed mobile/area source distribu-
tion. These estimates for the 1978 CEA baseline are presented 1n
Table 9, and were obtained as follows:
Total change » 1 - C 1 - 0.8 (Mobile reduction)] 1.01n
Where: Total change = Estimated total net change in mobile/area source
NOX emissions, shown 1n Table 9
Mobile reduction a Composite mobile NOV emission reduction from
Table 5
0.8 = Ratio of mobile to total mobile/area source emissions
(assumed)
1.01 = Annual CEA growth rate for mobile and area sources
(1 percent)
n = Years from 1978 baseline.
B-8
-------
TABLE 9.
ESTIMATED TOTAL CHANGE IN MOBILE AND AREA SOURCE NOX EMISSIONS
DUE TO FMVCP AND ONE PERCENT GROWTH RATE (£)
MOBILE2 MOBILES
FMVCP N0y Standard (gin/mile) LJI 2*0. 1*£ 1^1 £*£
'x
1978 Baseline
1985
1990
Interpolated 1982 Baseline
1985
1990
-15.1
-31.5
-5.7
-22.1
-12.5
-26.1
-4.7
-18.3
-9.9
-22.5
-3.7
-16.3
-6.4 -4.8
-17.1 -11.6
-2.4 -1.8
-13.1 -8.6
Thus, the "MOBILE2" columns for 1978 baseline 1n Table 9 provide
Independent estimates of the total projected change in mobile and area
source emissions. The "MOBILES" columns provide parallel projections
using the new EPA mobile source emissions model. To allow re-assessment
of future non-attainment based on 1982 ambient monitoring data, the
baseline was shifted to 1982 by assuming that changes from 1978 to 1985
were linear. The results of this adjustment are shown in the lower part
of Table 9. MOBILES results in a somewhat less dramatic overall reduc-
tion of mobile/area source emissions than MOBILE2 did. However, it
should be noted that in all cases there is still a net emission reduc-
tion for the mobile and area emission inventory due to the dominance of
the FMYCP mobile emission reductions.
AMBIENT NOX DATA UPDATE
Table 10 presents a summary of CEA base year N02 values for all
areas exceeding the 0.053 ppm standard, and more recent monitoring data
for those areas. The base year values are the highest annual average
for each area in 1976-1978, as selected specifically for use in the CEA.
B-9
-------
The more recent data are taken from EPA air quality trends reports, and
represent the highest annual average value for the SMSA. (U.S. EPA
1981, 1984). At least two areas not listed 1n Table 10 exceeded the
standard 1n 1979, and values of 0.053 were recorded 1n several areas 1n
1976-1978 and 1n one area 1n 1982. As Table 10 Indicates, however, only
the Los Angeles-Long Beach, California, SMSA exceeded the standard 1n
1982.
TABLE 10.
SUMMARY OF RECENT HIGHEST ANNUAL N02 AVERAGES
BY SMSA (ppm)
1976-19783 1979b 1980C 1981° 1982C
Los Angeles
- Long Beach, CA 0.081 0.078 0.071 0.071 0.062
Chicago, IL 0.078 0.078 0.060 0.050 0.052
Riverside -
San Bernadlno -
Ontario, CA 0.066 0.066 0.050 0.049 0.044
Detroit, MI 0.063 0.048 0.036 0.038 0.019
Anaheim - Santa Ana -
Orange Grove, CA 0.060 0.060 0.055 0.061 0.048
San Diego, CA 0.058 0.049 0.036 0.043 0.030
Philadelphia, PA 0.055 0.049 0.046 0.046 0.039
Indianapolis, IN 0.055 0.055 0.036 0.030 0.028
Denver, CO 0.054 0.051 0.050 0.047 0.039
aCEA base year annual average N02 values, the highest value for 1976-
1978.
Reference: U.S. EPA, 1981
Reference: U.S. EPA, 1984
B-10
-------
MOBILE AND AREA SOURCE CONTROL COST UPDATE
Table 9 Indicates that no area currently attaining the NOX standard
will exceed it In 1985 or 1990 under the assumed 1.0 percent mobile/area
source growth rate and 1.0, 1.5 or 2.0 gram/mile FMVCP standards. Thus,
according to our hypothetical analysis, the Los Angeles-Long Beach SMSA
1s the only area that needs to Implement a NOX emission control program
for any of the alternative NOX standards being considered (0.053, 0.060,
0.070). Tables 11 through 13 summarize an updated control cost analysis
for this area, based on previous CEA cost model estimates.
Table 11 provides the results of an analysis of ambient NOX
reductions needed to meet the 0.053 and 0.060 ppm NOX standards under
the revised 1982 baseline. The upper half of the table compares
estimated CEA ambient data projections to those derived using the new
Los Angeles-Long Beach high NOX value and revised emission projections.
Since the CEA used a linear rol Iback to project air quality, the changes
1n emissions from Table 9 can be used to project the available baseline
ambient NOX levels to 1985 and 1990. The bottom half of the table
Indicates the further emission reductions required to meet the
alternative standards.
Table 12 provides available control cost estimates for maximum
feasible control of mobile and area sources 1n the Los Angeles-
Long Beach SMSA, updated to March 1984 from cost breakdowns provided by
the'original CEA control cost model runs. As noted, a maximum control
level cost was not available for area source controls under a 1.0 gpm
FMVCP 1n 1990, and the cost of the highest available control 1s
presented. The data 1n Table 12 were used to generate the updated
control costs 1n Table 13. These costs parallel the CEA scenarios, 1n
which Inspection and maintenance (I/M) and transportation control
measures (TCM) were either assumed to be Implemented first, with area
source controls then applied using a least-cost algorithm, or In which
I/M and TCMs were not used at al1. Area source costs were estimated by
assuming a linear relationship between cost and ambient NOV reduction,
A
based on the values 1n Table 12. This probably results 1n some
B-ll
-------
TAB LE 11
ANBIENT AIR QUALITY ESTIMATES FOR
COST UPDATE
CEA (1978 Baseline)
Analysis
Los Angeles-Long Beach
SMSA High Value, ppm
FMVCP Standard, gm/mlle
Change from Baseline
(from Table 9). percent
1985
1990
Projected High Values
1985
1990
Amfrlent Reductions
Needed for Updated
Costs (1982 Baseline).
percent
Q.Q53 ppm NO.. NAAQS
^™^™^fc^fc^fc^~^^™**™""fc*^^
1985
1990
Q.06Q ppm NO NAAQS
- -T ' ^
1985
1990
0.081
2.0
-15.1 -12.5
-31.5 -25.1
0.069 0.071
0.055 0.060
1982 Baseline
Update
0.062
JLSL
FMVCP Standard
i.o
11.7
0
0
0
1.5
13.1
0
1.6
0
2.Q,
13.1
7.0
1.6
0
-3.7 -2.4 -1.8
-16.3 -13.1 -8.6
0.060 0.061 0.061
0.052 0.053 0.057
B-12
-------
overestimate on of costs for cases not requiring the maximum available
emission reduction, relative to the original least-cost solutions in the
CEA. Since there were no CEA costs for the 1.5 gpm FMYCP, the 2.0 gpm
costs in Table 12 were used. This should result in a further over-
estimate in this case, since the higher FMVCP costs result in more
expensive ambient level reductions.
FMVCP COST UPDATE
Table 14 contains updated costs for the Federal Motor Vehicle
Control Program (FMVCP). These costs were updated from costs contained
in Chapter 6 of the CEA. Costs and credits for FMVCP-rel ated changes in
fuel economy were eliminated from the original CEA estimates based on
guidance from EPA-Ann Arbor. (Hellman, 1984). This is because
potential changes in fuel economy due to FMVCP are impossible to isolate
from changes due to regulations specifically controlling fuel economy.
To assess potential cost impacts of instituting a 1.5 gpm NOX
FMVCP in 1984, CEA cost estimates were modified by assuming that all
light duty vehicles would use oxidation catalyst-based control systems
rather than the three-way catalyst systems assumed in the CEA. This is
a highly conservative assumption, since it is not likely that the entire
new-car fleet wil 1 revert to oxidation catalysts. According to recent
EPA estimates (Gray, 1983), the N0x-related capital cost for an oxida-
tion-catalyst open-loop control system is about one-quarter of the
capital cost of the dominant three-way catalyst systems in use in 1983
vehicles. The methodology in CEA Section 6.2.1 was applied to estimate
the cost of this assumed control technology for a 1.5 gpm NOX standard
beginning in 1985. The CEA analysis for a 2.0 gpm standard was also
revised to begin the relaxed standard in 1985 instead of 1981. In both
cases, the technology already in place from 1981 through 1984 dominates
the annualized costs shown in Table 14. Similarly, operating and main-
tenance costs shown in Table 14 for the two relaxed standards consist
entirely of costs for the 1981-1984 three-way catalyst technology, since
the 13 and 2.0 gpm technologies do not involve additional 0 & M costs.
B-13
-------
TABLE 12.
MAXIMUM CEA CONTROLS POSSIBLE
IN LOS ANGELES-LONG BEACH AREA
Annual1zed Cost
(millions of March 1984 dollars)
Ambient Reduction
(percent)3
1985
1.0 gpm FMVCP
I/M
TCM
Area sources
2.0 gpm FMVCP
I/M
TCM
Area sources
1990
1.0 gpm FMVCP
I/M
TCM
Area sources'"
2.0 9pm FMVCP
I/M
TCM
Area sources
25.3
176.5
25.3
176.5
10.2
105.1
10.2
190.9
4.3
1.4
11.5
5.6
1.4
9.8
8.6
Negligible
10.3
8.1
1.6
12.9
aAmb1ent reduction percentages subject to uncertain accuracy due to
rounding of predicted ambient concentrations by CEA cost model.
bTh1s entry represents the highest control level required by CEA scenarios;
however, maximum feasible control was never required 1n this case.
B-14
-------
TABLE 13.
ESTIMATED ANNUALIZED COST OF ATTAINING
ALTERNATIVE NOo NAAQSa
(millions of March 1984 dollars)
FMVCP Standard: 1.0 gm/mlle 1.5 gm/m1leb 2.0 gin/mile
I/M. TCM Used First
1985
1990
117.2
0
135.1
0
135.1
10.2
0.060 oom NO NAADS
I/Mf
Q
o
1985
1990
TCM Not Used
.055 ppm NO.. NAAQS
1985
1990
.060 ppm NO. NAAQS
1985
1990
0
0
176.5
0
0
0
25.3
0
176.5°
0
28.8
0
25.3
0
176.5°
103.8
28.8
0
aZero costs Indicate that no controls 1n addition to FMVCP and NSPS are
required to attain the NAAQS. A 0.070 ppm NAAQS would not be exceeded
under any CEA scenario.
Based on Table 12 costs for a 2.0 gpm FMVCP.
CNAAQS not attained with maximum available controls.
B-15
-------
TABLE 14.
TOTAL ANNUALIZED COST OF THE
FEDERAL MOTOR VEHICLE CONTROL PROGRAM
(Millions of March 1984 Dollars)
Vehicle type
Light-duty vehicles
Hardware
O&M
Unleaded fuel
1.0
1985
1,086
399
352
gpm
1990
2,003
790
747
FMVCP:
1.5 gom
1985
915
308
352
1990
944
225
747
2.0
1985
884
308
352
gpm
1990
735
225
747
Subtotal 1,837 3,540
Light-duty trucks 67 535
Heavy-duty vehicles 2 398
1,575 1,916
67 535
2 398
1,544 1,707
67 535
2 398
TOTAL
1,906
4,473 1,644 2,849 1,613 2,640
POINT SOURCE CONTROL COSTS
Table 15 contains updated annual1zed point source control cost
estimates for 1985 and 1990, taken from Tables 5-1 and 5-2 of the CEA,
There are no point source controls required for a 0.070 ppm NOX
standard.
NEW SOURCE PERFORMANCE STANDARD (NSPS) COSTS
Based on the mid-point of 1990 NSPS annual1zed costs presented 1n
Chapter 6 of the CEA, NOX NSPS control costs are updated to $110 million
1n 1985 and $250 million 1n 1990.
B-16
-------
TABLE 15.
POINT SOURCE CONTROL COST ESTIMATES
FOR ALTERNATIVE N02 STANDARDS
Number
of Sources
AnnualIzed
Cost
(millions of
Number of
Nonattalnment
Year
1985
1990
N02 Standard
0.053
0.060
0.053
0.060
Controlled
172
16
104
10
March 1984 $)
11.67
0.76
2.01
0.10
Plants
1
0
0
0
B-17
-------
THE RECENT NOX (AND PARTICIPATE) EMISSION STANDARDS PROPOSAL
EPA's Office of Mobile Sources recently proposed to reduce future NOx
emissions from light duty trucks and heavy duty engines (49 FR 40258). The
changes proposed are summmarized in Table 16. They do not apply to motor
vehicles sold in California because that State has its own vehicular
control program.
The newly proposed standards are more stringent than current FMVCP
standards, thus fewer emissions are to be expected from each regulated
vehicle on a per mile basis. However, as will be described below, these
changes do not impact upon most of the cost or non-attainment results
presented in this report.
Since the proposed changes in motor vehicle emission standards do
not take effect until the 1987 model year (beginning in late 1986) at the
earliest, the changes of course do not impact upon 1985 costs presented
in Tables 5.1-5.3 and in the updated cost tables of this Appendix (Tables
13-14). Since the new FMVCP standards do not affect emissions in the one
N02 non-attainment area that we have identified—Los Angeles—the changes
do not affect the "additional mobile/area" cost estimates contained in
Tables 5.4-5.6 or in the updated Table 12 of this Appendix. Thus, the
only impact of the proposed changes in NOx emission standards is to alter
FMVCP costs for 1990.
Table 14 of this Appendix presents data relating to FMVCP costs for
various vehicle types. Estimated costs presented in the Cost and Economic
Assessment (EPA, 1982) for the "light-duty trucks" and "heavy-duty vehicles,"
the two vehicle types most closely corresponding to those affected by the
proposed emission standard changes, are $933 million annualized 1990
B-18
-------
TABLE 16.
CHANGES IN MOTOR VEHICULAR NOx EMISSION STANDARD
PROPOSED IN 49 FR 40258
Light Duty Trucks
£6,000 GVW >6,000 GVW
(gpm) (gpm)
Heavy Duty Engines
(Gasoline and Diesel)
(g/BHP-hr)
Current FMVCP
1985 and on MYa
Proposal FMVCP
1987-1989 MY
1989 and on MY
2.3
1.2
1.2
2.3
1.7
1.7
10.7
6.0
4.0
Abbreviations:
NOx
GVW
gpm
g/BHP-hr
FMVCP
MY
nitrogen oxides
gross vehicle weight
grams per mile
grams per brake horsepower-hour
Federal Motor Vehicle Control Program
model year
Note: a. The current FMVCP emission standards do not distinguish
between the two weight classes of light duty trucks; thus
both classes have the same standard.
B-19
-------
costs in 1984 dollars. (This is approximately $527 million in present
value terms.) These costs do not vary by the FMVCP standards shown in
Table 14, since those standards apply only to light-duty vehicles.
The proposed NOx emission changes would significantly increase FMVCP
costs presented in Table 14. The increase in costs is due to the following
factors (Office of Mobile Sources, n.d.):
1. additional fixed or pre-production, costs assocaited with the research
development, and testing of new control technology to meet the
tighter proposed standards. (These costs include one-time certi-
fication activities that must occur prior to the manufacture of
control equipment.
2. variable costs associated with the manufacture and installation
of new control technology.
3. operating costs associated with use of the equipment. These costs
generally are comprised of fuel penalties and additional maintenance.
(However, it may be possible to obtain negative costs, i.e., a
savings, due to these items; see below.)
4. a 10% "contingency factor" that is added to Mobile Source's cost
estmates to account for uncertainty in the technological and
monetary assumptions used in their analysis.
While the Mobile Source Draft RIA presents cost (and economic) information
related to both NOx and particulate control, we are concerned here only with
NOx control cost estimates. In addition, for the reasons stated above, we are
concerned only with Mobile Sources' 1990 estimates. Said estimates appear in
Table 17. A brief discussion of the data appearing in the Table is in order.
B-20
-------
TABLE 17
ADDITIONAL 1990 FMVCP COSTS ASSOCIATED
WITH THE RECENTLY PROPOSED LOT
AND HDE EMISSION STANDARDS1
Vehicle
Class
LOT
HDGE
HDDE
Item
Fixed and variable costs
Fuel penalty/savings-^
Fixed and variable costs
Fixed and variable costs
Maintenance costs
Fuel penalty^
1990 Costs
($106)
254
+856/1%
103
4
0-452
Present Value
of the 1990 Costs2
(106)
144
+484/1%
1
58
2
0-255
Source: Office of Mobile Sources (n.d.)
Abbreviations: FMVCP = Federal motor vehicle control program
LOT = Light duty trucks (gasoline and diesel fueled)
HDE = Heavy duty engines (gasoline and diesel fueled)
HDGE = Heavy duty gasoline engines
HDDE = Heavy duty diesel engines
Notes: 1 Proposed in 49 FR 40258, October 15, 1984. Costs appearing in this
table should be added to those appearing in Table 14 of this RIA. All
costs are in 1984 dollars except where otherwise noted.
^Discounted back to 1984 using a 10% interest rate or discount factor.
3In 1990 discounted dollars, over the lifetime of the vehicle, assumed to
be 11 years.
4In 1990 discounted dollars for the time period 1990-1992.
B-21
-------
The increase in LOT costs due to control technology development and
testing (the "fixed" costs) and manufacture and installation ("variable"
costs) are $254 million in 1990. In addition, Mobile Sources states that
"it is reasonable to expect that some small increase could occur in the
fuel econoy of some LDGTs and some small decrease could occur in the fuel
economy of some IDDTs" (p. 3-22). Because it was difficult to estimate
(1) the number of trucks that would save fuel, or use more fuel, and (2)
the amount of fuel saved per vehicle (due to unknown sales figures and
unknown vehicle miles traveled per vehicle), Mobile Sources presented a
"contingent" fuel savings estimate for the LOT class of vehicles. Their
estimate is $856 million per one percent change in fuel economy over the
lifetime of the vehicle class (assumed to be 11 years). This figure is
in 1987 discounted dollars, so it is not comparable to the fixed and
variable costs estimate for LDTs. Thus, the 1990 LOT costs shown in
Table 17 could be higher or lower than $254 million by some undefined
amount due to changes in fuel consumption associated with NOx control
technology needed to achieve the 1.2/1.7 grams per mile emission standards
for LDTs.
The additional 1990 FMVCP costs for HDGEs are modest: only $2 million.
According to Mobile Sources, there are no additional maintenance or fuel
impacts associated with the 4.0 g/BHP-hour emission control technology for
heavy duty gasoline engines.
For heavy duty diesel engines, however, the proposed standard will have
a significant technological (and cost) impact. Fixed and variable costs
B-22
-------
are estimated by Mobile Sources to be $103 million in 1990 ($58 million PV).
To this is added (1) a maintenance expenditure of approximately $4 million
(due to extra maintenance of the NOx aftercooling system technology),* and
(2) a fuel penalty cost of between $0-452 million for 1990.
The wide range of the fuel penalty cost estimate is due to uncertainty
in the percentge fuel penalty associated with NOx control equipment. Mobile
Sources estimates that it could be between 0% and 2%, but only in the short-term
(2 years or so), because engine and vehicle improvements will be made that will
overcome any fuel penlty associated with the initial NOx control equipment.
As with the LOT fuel penalty/savings estimte contained in Table 17,
HDDE fuel penalty costs are not comparable with the other two cost categories
for the HDDE class of vehicles due to the form in which the fuel penalty costs
are presented in Office of Mobile Sources (n.d.).
In summary, the proposed changs in NOx emission standards contained in
49 FR 40258 would increase FMVCP costs appearing in this report by the amounts
shown in Table 17. However, as was explained above, the fuel penalty and
savings estimates contained in Table 17 cannot be used directly due to the
matter in which those estimates were calculated or presented in the Draft
Mobile Source RIA.
*Mobile Sources estimates that there is an overall maintenance savings
associated with HDDE control technology as a unit, but stated that the
NOx portion of the technology involved additional maintenance work (as
distinguished from the particulate portion, which saved money in equipment
repair).
B-23
-------
It should be noted that the methodologies used to estimate needed future
emission reductions differ In the Draft RIA for NOx emission standards
(Office of Mobile Sources, n.d.) and this Draft RIA for alternative NC>2
NAAQS. There are two important differences: the base year air quality
data used to obtain a design value and the annual growth rate in mobile
source vehicle miles traveled.
Mobile Sources' RIA used 1980-81 air quality data whereas OAQPS used
1981-1983 air quality data. This difference is significant since N0£ annual
averages are trending downwards more rapidly than the NOx emission inventory
changes. Thus, using an older air quality data base results in relatively
high design values vis-a-v"s the emissions base, leading to more control
hypothetically being needed to attain and maintain a given NAAQS level.
For growth in vehicle miles traveled, Mobile Sources used vehicle class
specific factors for light duty vehicles, heavy duty vehicles, light duty
trucks, and so forth. These factors vary between -0.3% per year for
heavy duty gas vehicles to +4.7% per year for light duty trucks. The
approximate average for all motor vehicles is on the order of 1.8-2.0%
per year. This is considerably higher than the NOg NAAQS analysis's use
of 1.0% per year for all motor vehicles as a class.* The 80-100% increase
in emissions growth rate directly results in the large number of NOx
*Due to increasing vehicle miles traveled, or VMT, in an area. This rate
varied between -0.5% and +1.5% per year for the 1979-1982 time period.
(Motor Vehicle Manufacturer's Association, 1983.).
B-24
-------
non-attainment areas that the Mobile Sources' Draft RIA predicts will
occur in 1995. (OAQPS did not carry its analysis out to 1995. It stopped
in 1990, because of the short planning horizon associated with the five-year
review cycle of all national ambient standards. For the sake of comparison,
Mobile Sources predicts 2 NOx non-attainment ares in 1990 whereas we
predict none.)
There are other analytic and procedural differences between the two
regulatory analyses, but none as important as those mentioned above. For
a complete understanding of the analyses, see Office of Mobile Sources
(n.d.) and reference 52 (p. 75).
B-25
-------
ATTACHMENT 1:
COMPARISON OF MOBILE 2 AND MOBILE 3
EMISSION FACTORS BY VEHICLE TYPE
The following graphs were prepared by EPA-Ann Arbor
staff for MOBILE 3 workshops. They are based on
MOBILE 3 default conditions and typical data values.
B-26
-------
en
a
en 2-
S
en
en
H
MOBILES VERSUS MOBILES
LDGV — LOW ALTITUDE
NOX EMISSIONS
—i 1 1 1 1 1 1
1970 1975 1980 1985 1990 1995 2000
CALENDAR YEAR
3-
i^^
S
is
en 2-
en
en
1-
MOBILE2 VERSUS MOBILES
LDGT1 =— LOW ALTITUDE
NOX EMISSIONS
1970 1975 1900 1905 1990 1995
CALENDAR YEAR
2000
B-27
-------
MOBILES VERSUS MOBILES
LDGT2 — LOW ALTITUDE
NOX EMISSIONS
6-
5-
CO
w
CO
CO
3 2-j
1-
Legend
A MOBILE2
X MOBILES
1970 1975 1980 1985 1990 1995 2000
CALENDAR YEAR
MOBILE2 VERSUS MOBILES
HDGV — LOW ALTITUDE
NOX EMISSIONS
12-1
10-
8-
or
CO 6
I
CO
CO
4-
2-
-X
1970 1975 1980 1985 1990 1995
CALENDAR YEAR
—i—
2000
B-28
-------
1.5-
s
>
co i-
o
>—4
CO
CO
MOBILES VERS JS MOBILES
LDDV — LOW ALTITUDE
NOX EMISSIONS
1970 1975 1980 1985 1990 1995 2000
CALENDAR YEAR
MOBILES VERSUS MOBILES
. LDDT — LOW ALTITUDE
NOX EMISSIONS
Legend
A MOB1LE2
X MOBILES
2.5-1
|
CO
2-
1.5-
O
co H
CO
0.5-1
X
1970 1975
1980 1905
CALENDAR YEAR
1990 1995
2000
B-29
-------
MOBILE2 VERSUS MOBILES
HDDV — LOW ALTITUDE
NOX EMISSIONS
en
25
S
GO
00
H
30-,
25-
20-
15-
10-
5-
1970 1975 1980 1985 1990 1995
CALENDAR YEAR
2000
MOBILE2 VERSUS MOBILES
ALL MOBILE SOURCES COMBINED — LOW ALTITUDE
NOX EMISSIONS
5-
^«1
GO 3
25
Q
GO
GO
U 2-
x^' \
1970 1975 1980 1985 1990 1995
CALENDAR YEAR
1
2000
B-30
-------
REFERENCES FOR APPENDIX B
Bellman, K.H., EPA-Ann Arbor. May 2, 1984. Memorandum to 8. Jordan,
EPA/OAQPS, on OMB Memo Regarding N02 RIA.
U.S. Environmental Protection Agency. April 1981. National Air Quality
and Emissions Trends Report, 1981. EPA-450/4-83-011.
U.S. Environmental Protection Agency. June 1982. Cost and Economic
Assessment of Regulatory Alternatives for N02 NAAQS. Draft Report.
U.S. Environmental Protection Agency. March 1984. National Air Quality
and Emissions Trends Report, 1982. EPA-450/4-84-002.
U.S. Environmental Protection Agency. October 15, 1984. "Control of
Air Pollution from New Motor Vehicles and New Motor Vehicle Engines."
49 Federal Register 40258; This is cited as 49 FR 40258 in the text.
U.S. Environmental Protection Agency, n.d., Office of Mobile Sources. Draft
Regulatory Impact Analysis and Oxides of Nitrogen Pollutant Specific Study.
Gray, C.L., EPA-Ann Arbor. November 28, 1983. Memorandum to B. Jordan,
EPA/OAQPS, on Updated Cost Estimates of Controlling HC Emissions from
Mobile Sources.
Motor Vehicle Manufacturers Association. MVMA Motor Vehicle Facts and
Figures '83. Detroit: MVMA, 1983.
B-31
-------
TECHNICAL APPENDIX C
OF THE
N02 NAAQS REGULATORY IMPACT ANALYSIS
OCTOBER 1984
GCA - Technology Division
213 Burlington Road
Bedford, Massachusetts 01730
-------
-------
APPENDIX C
NOV CONTROL TECHNOLOGY FOR STATIONARY SOURCES
X
The purpose of this Appendix is to review stationary source NO control
A
technologies relevant to the N02 NAAQS, and to provide background information
on control technologies for combustion processes used in the Cost and
Economic Assessment of Regulatory Alternatives for N00 NAAQS (CEA). The CEA
• " ~' (_
estimates of NO control costs for point and area stationary sources were
A
based on selection of the least-cost combination of controls which would
allow attainment of a given N02 standard. Controls were chosen from lists of
the types of control technologies applicable to each source type (CEA Tables 4-6
and 4-15). These controls include all known techniques believed to be suitable
for retrofit application to the sources of concern over the time frame of the
CEA analysis (i.e. through 1990).
For combustion sources, which account for over 90 percent of all stationary
source NO emissions, NO controls fall into two categories: those controls
A A
which modify the combustion process to suppress NO formation; and those
X
controls which remove NO from the flue gas after combustion has been completed.
X
Of the combustion controls used in the CEA, low excess air, staged combustion,
flue gas recirculation, reduced combustion intensity, and low NO burners are
A
combustion modification techniques, while selective catalytic reduction, and
ammonia injection (selective non-catalytic reduction) are flue gas treatments.
Detailed background information on each of these controls is presented later
in this Appendix. The CEA distinguishes between currently "commercially
demonstrated" controls, which are considered available in 1985, and "undemon-
strated" controls, which are applied only in estimating costs for 1990. The
latter category consists of the two flue gas treatment methods, ammonia
injection and selective catalytic reduction, which have been used in the U.S.
only on a limited pilot or demonstration project basis. Recent review of the
status of NO control technology development by EPA groups cognizant in the
X
field indicates that the control availability and cost information used in
the CEA remains valid. '
C-l
-------
There are a number of potential NO control techniques which were not
A
applied in the CEA analysis because their development is not expected to result
in significant control technology shifts by 1990. These include reburning
(fuel staging), precombustors, and stoker gas recirculation. Many specific
applications of these principles are in testing stages, but only a few are
commercially available on a very limited basis, or for control of pollutants
other than NO . Reburning is available only as a part of one low-NOv burner
A X
package. Stoker gas recirculation is being used on several stoker boilers
for particulate control, but has restricted applicability due to the small
2
number of stoker boilers. Other potential NO control approaches which are
A
actually modifications to basic combustion technologies, such as fluidized
bed combustion and combined cycle gasifiers, are not amenable to retrofit
situations and thus are not considered for control of existing NO sources
in the CEA.
The CEA controls discussed above apply mainly to industrial and utility
boilers, but some also apply to industrial process furnaces and reciprocating
internal combustion engines. Fine-tuning and changing air/fuel ratio or
increasing stack height are additional controls available for reciprocating
internal combustion engines. The other two CEA point source categories with
available controls are stationary gas turbines (water injection) and nitric
acid plants (chilled absorption). NO controls are not available for process
"
heaters or catalytic crackers in the petroleum industry, glass melting furnaces,
process-gas-fueled blast furnaces, and coke ovens.
LOW EXCESS AIR COMBUSTION (LEA)
Reducing the amount of excess air supplied for combustion has been shown
to be an effective method for reducing NO emissions from utility and industrial
A
boilers, residential and commercial furnaces, warm air furnaces, and process
furnaces. In this technique, the combustion air is reduced to the minimum
amount required for complete combustion, while maintaining acceptable furnace
cleanliness, (ie. no fouling or slagging) and steam temperature. Reducing
the amount of excess air also reduces the amount of both thermal and fuel NO
A
since there is less 02 available in the flame zone. Boiler efficiency also
is improved because the reduced airflow also lowers the quantity of the gas
4
released.
C-2
-------
Since boilers are not perfect combustion devices, excess air is always
supplied to ensure complete combustion. This includes complete reaction of
the fuel (HC) and oxidation of most of the resulting CO to COp. Too much
excess air causes increased NO formation due to an oxygen-rich combustion
/\
zone and reduced boiler efficiency due to increased dry gas loss up the
stack. Too little excess air causes some fuel to leave the combustion zone
unreacted and a large amount of CO to exit unoxidized. Both sets of conditions
4
are unacceptable from both a safety and efficiency standpoint.
Optimum conditions would require just enough excess air to avoid smoking
(caused by incomplete combustion), have zero unburned hydrocarbons, small CO
emissions (less than 100 ppm), and handle any process variations such as low
load operation, nonuniformity of air/fuel ratio, fuel and air control lags
during load swings, and fuel variations. Low excess air is a simple, feasible,
and effective technique to control NO emissions from utility and industrial
A
boilers and is presently considered a routine operating procedure in these
combustion processes. Most sources will require additional control methods,
in addition to LEA to'reduce NO levels to within statutory limits. Effective-
/\
ness of this technique may range from 5-25 percent NO reduction depending on
X
the initial level of excess air used.
STAGED COMBUSTION
Staged or off-stoichiometric combustion (OSC) involves initial combustion
in a fuel-rich zone followed by secondary combustion at lower temperatures in
a fuel-lean zone. The fuel-rich zone is essentially an oxygen-deficient zone
which inhibits the formation of thermal NOV and also reduces fuel NO formation
A X
by providing a time span for a portion of the fuel bound nitrogen to be
reduced to N^. If heat is transferred to the walls prior to secondary air
addition and the completion of combustion, then the temperatures in the fuel-
lean secondary zone will be lower. Lower temperatures inhibit thermal NO
5
formation.
The most commonly used staged combustion techniques are biased burner
firing (BBF), burners-out-of-service (BOOS), and overfire air injection
(OFA). The following gives a brief discussion of each technique.
C-3
-------
Biased Burner Firing (BBF)
Biased burner firing is achieved by creating fuel-rich and fuel-lean
regions in the primary combustion zone. Typically, this is done by firing
the lower rows of burners more fuel-rich than the upper rows of burners. The
additional air required for complete combustion enters through the upper rows
of burners which fire a fuel-lean mixture. This technique allows all of the
burners to remain in operation.
Burners Out of Service (BOOS)
The burners-out-of-service technique is the extreme of biasing in that
fuel flow is terminated at individual burners or rows of burners while maintaining
air flow and redistributing to the remaining burners in service. Thus,
fuel-rich conditions are created at the in-service burners while the remaining
air required for combustion is introduced through the out-of-service burners.
In some cases BOOS is not feasible because boiler load is decreased somewhat.
This load reduction may occur because burners may be unable to handle
4
increased fuel flow and also because some burners have been shut down.
Both BBF and BOOS are applicable to all fuels and are particularly
attractive as control methods for existing units, since few, if any, equipment
modifications are required. However, accurate flue gas monitoring equipment
and increased operator monitoring of furnace conditions are required with
these combustion modifications. Monitoring flue gas, excess air, and CO will
help to avoid boiler efficiency reduction through flue gas heat and unburned
combustible losses and to prevent unsafe operating conditions caused by
incomplete combustion.
The most effective BBF or BOOS burner configuration for a particular
unit is determined through field tests, although for BOOS, the out-of-service
burners are typically located in the upper portions of the furnace. Emissions
tests for BOOS firing on utility boilers have indicated average NOY reductions
/\
of 31 to 37 percent for coal, oil, and natural gas firing compared to earlier
baseline levels.
C-4
-------
Overfire Air Injection (OFA)
The overfire air technique involves operating the burners at near stoichiometric
conditions or fuel-rich with the remaining combustion air being supplied
through overfire air ports located above the rows of burners or through an
idle top row of burners.
Potential limitations of this technique include:
o Furnace tube wasteage due to local reducing conditions when firing
coal or high sulfur oil
o Tendency for slag accumulation in the furnace
o Additional excess air may be required to ensure complete combustion
resulting in a decrease in boiler efficiency.
However, these potential problems may be overcome with proper implementation
4
of staged combustion.
The major drawback with OFA is that additional duct work, furnace
penetration, and extra fan capacity may be required. Thus, OFA is more
attractive in original designs than in retrofit applications. In addition,
there may be physical obstructions outside of the boiler setting making
installation more costly. There may also be insufficient height between the
top row of burners and the furnace exit to permit the installation of overfire
4
air ports or to allow sufficient residence time for the completion of combustion.
Since staged combustion techniques serve to delay or prolong the combustion
process, furnace size (volume and wall surface) must necessarily be large
in order to ensure complete combustion. Thus, overfire air is more easily
implemented without large efficiency or cost penalties on large units than on
small ones. Staged combustion techniques are typically applied to utility
,., 6
size boilers.
Some emissions tests of OFA on utility boilers have reduced NOV emissions
A
25 to 60 percent for coal, oil, and natural gas firing compared to earlier
baseline levels.
C-5
-------
FLUE GAS RECIRCULATION (FGR)
Flue gas recirculation consists of extraction of a portion of flue gas
from the exhaust stream and subsequent re-introduction into the furnace. The
recirculated flue gas acts as a diluent in the primary combustion zone to
reduce peak temperatures and lower oxygen concentration. Since the bulk of
NO emissions from gas and most oil-fuel firing are thermal NOV, FGR is very
A X
effective on these fuel types. For a fuel oil with high nitrogen content and
coal combustion, FGR is less effective due to the large contribution of fuel
5
NO to the NO emissions. However, it can be used to reduce 09 level in
A A £
primary combustion air and thus reduce NO from coal-fired systems.
A
FGR is more attractive for new designs than as a retrofit application.
This is because a retrofit requires a fan, flues, dampers, controls and
possibly extra fan capacity. In addition, the FGR system itself may require
a substantial maintenance program due to the high temperature environment and
potential erosion from entrained ash.
As a new design feature, the furnace and convective surfaces can be
sized for the increase in mass flow and change in furnace temperatures. In
contrast, in retrofit applications the increased mass flow increases turbulence
and mixing in the burner zone and alters the boiler heat absorption profile.
Effectiveness of this technique may range from 0-30 percent NO reduction
X
depending on the type of boiler it is is applied to, and the fuel fired.
REDUCED COMBUSTION INTENSITY
Thermal NO formation can be controlled by reducing combustion intensity
X
through load reduction (or derating) in existing units, and by enlarging the
firebox in new units. The reduced heat release rate lowers the bulk gas
temperature which in turn reduces thermal NO formation.
X
Reduced firing rate often leads to several operating problems. Aside
from limiting capacity, low load operation usually requires higher levels of
excess air to maintain steam temperature and to control smoke and CO emissions.
The steam temperature control range also is reduced substantially. This will reduce
operating flexibility of the unit and its response to changes in load. The
combined results are reduced operating efficiency due to higher excess air
and a reduction in the system's load following capability due to a smaller
control range.
C-6
-------
When the unit is designed for a reduced heat release rate, problems
associated with derating are largely avoided. Use of an enlarged firebox
produces NO reductions similar to load reduction on existing units.
A
LOW N0¥ BURNERS
A
Burner or combustor modification for NO control is applicable to all
A
stationary combustion sources, except those without burners (spreader stoker
coal boilers, for example). The specific design and configuration of a
burner has an important bearing on the amount of NO formed. The main goal
A
of low NO burners is to reduce the amount of NO formed to a minimum while
A A
achieving acceptable combustion of the fuel. Low NO burners are widely used
A
on utility and industrial boilers, and employ LEA, OSC, or FGR principles.
Control of NO formation in most low NO burners for utility boilers is
j\ A
achieved by reducing flame turbulence, delaying fuel/air mixing, and establishing
fuel-rich zones where combustion initially takes place. The longer, less
intense flames produced with low NO burners result in lower flame temperatures
A
which reduce thermal NO generation. In addition, the reduced availability
A
of oxygen in the initial combustion zone inhibits conversion of fuel-bound
nitrogen to NO -.
A
Low NO burners typically are used as a NO emission control technique
A A
for process heaters, particularly on gas and oil fired refinery process
heaters. These are generally staged air burners, which use the concepts of
staged combustion to achieve NO reduction. The first stage incorporates
A
firing a fuel-rich combustion mixture. The burner is designed to inject air
after a sufficient time lag, to complete the combustion process. This technique
reduces thermal NO formation by limiting the amount of excess air and reduces
/\.
fuel NO by allowing enough time for reduction of fuel-bound nitrogen to N«
in the reducing zone of the flame. NOV reductions of around 55 percent are
6
typical for these devices. A number of designs based on the staged air
principle are available for pulverized coal, and similar units are available
for gas and oil boilers.
C-7
-------
Other types of low NO burners include the staged fuel burner, the
A
self-recirculating gasification (SR6) burner, and dual register burners.
Staged fuel burners involve initial combustion of the fuel with high excess
air followed by injection and combustion of the remaining fuel at low excess
air. This technique allows initial combustion to occur at a relatively low
temperature (1090°C, 2000°F), thus inhibiting thermal NO formation. As the
j\
combustion reaction goes to completion in the first zone, additional fuel is
injected. The second reaction begins with a reduced partial pressure of
oxygen which tends to limit formation of NO . The SRG burner involves use
A
of flue gas recirculation, two-staged combustion, gasification reactions, and
low excess air. The key design feature is creation of an exceptionally
strong recirculation eddy at the burner throat which draws combustion reaction
products from the furnace to gasify the fuel stream. Dual register burners
have been included in some new boilers. Babcock and Wilcox is currently
.installing a dual register pulverized coal-fired burner in all its new large
pulverized coal utility boilers. The B&W technique involves use of a limited
turbulence, controlled diffusion flame burner. This minimizes fuel and air
mixing at the burner to that required to obtain ignition and substain stable
combustion of the coal. A venturi mixing device, located in the coal nozzle,
provides a uniform coal/primary air mixture at the burner. Secondary air is
introduced through two concentric zones surrounding the control nozzle, each
of which is independently controlled by inner and outer air zone registers.
Reductions in NO emissions have ranged from 40 to 60 percent due to this
J\
improved design. Foster Wheeler Energy Corporation also has developed a dual
register coal burner in its new utility and industrial boilers. The new
burner also reduces turbulence and causes controlled, gradual mixing of fuel
and air at the burner. This is achieved using a dual throat with two registers
which splits the secondary air into two concentric streams with independently
variable swirl. Test results for this technique have realized NO reductions
6
of 40 to 50 percent.
Co
-o
-------
FLUE GAS TREATMENT
The following two methods of NO reduction are flue gas treatment (FGT)
X
techniques. FGT options may be either wet or dry processes. Wet processes
are those which remove a pollutant in a solution or slurry from a wet scrubbing
process. Dry processes are those which employ a spray of fine droplets of
absorbing solution which mixes with the flue gas. Subsequent absorption of
the pollutant occurs almost simultaneously with evaporation of the water in
the droplet. Evaporation occurs due to the heat of the flue gas thus creating
a dry powder. The FGT techniques considered here are dry processes, and both
utilize NHo as a reducing agent. NH^ is a selective reducing agent in that
it selectively reduces NOV without having to consume any excess 09 first.
X £
If NH3 is injected after the boiler economizer; where temperature of the
flue gas is about 370°C to 450°C, a catalyst is necessary. However, if NH3
is injected into the secondary superheater region of the boiler, where the
temperature of the flue gas approaches 980°C, a catalyst is not necessary.
NH3 reduction processes utilizing a catalyst are termed selective catalytic
reduction (SCR) processes while those NH3 reduction processes which operate
without a catalyst are called selective non-catalytic processes (SNR), or
simply ammonia injection.
Selective Catalytic Reduction (SCR)
The SCR method is the more available FGT technique at this time. It is
used commercially on over 60 gas- or oil-fired boilers and on two coal-fired
boilers in Japan. The use of this process has been limited in the United
States to a few pilot scale units on coal-fired boilers and a demonstration
scale unit under construction on an oil-fired utility boiler.
The chemical mechanisms used to reduce NO can be described by the
X
following gas phase reactions
(1) 4NO + 4NH3 + 02 4N2 + 6H20
(2) 2NO£ + 4NH3 + 02 3N2 + 6H20
Since most of the NO leaving the boiler is NO, the first reaction predominates.
/\
C-9
-------
Flue gas from the economizer is mixed with vaporized ammonia prior to
the reactor vessel. The gas then passes through the catalyst bed where the
reduction process takes place. The flue gas subsequently exits the reactor
and is sent to the air preheater and, if necessary, additional pollutant
control devices.
Problems associated with this technique primarily are involved with the
catalyst and with the formation of solid ammonium sulfate ((NH^pSO^) and
liquid ammonium bisulfate (NHLHSO*). For the combustion of sulfur containing
fuels, catalysts must be chosen which are resistant to SO poisoning.
J\
Catalysts which fall into this category are oxides of titanium and vanadium.
However, manufacturers will guarantee catalyst life only for one to two years
in applications with high SO and particulate loadings. Particulates may
A
"blind" the catalyst surface rendering it ineffective. (NH4)2$04 and NH4S04
form when the proper conditions of temperature and S03 and NH-, concentrations
exist. The formation of these compounds is difficult to avoid since part of
the NH3 passes through the reactor unconsumed, and some catalysts promote the
oxidation of S09 to SO,. This problem may be alleviated by reheating the
5
flue gas or by limiting NhL slip.
Reactor designs are dependent on the type of fuel used in the combustion
process. Gas-fired applications commonly use catalyst pellets in a fixed
bed. Since the flue gas from oil- and coal-fired applications contains
particulates, reactor designs usually incorporate honeycomb, pipe, or
parallel plate shaped catalysts which allow the flue gas to pass in parallel
along the catalyst surface. These configurations would limit the deposition
of particulates on the catalyst.
Extensive pilot plant and commercial plant data indicate that, for
90 percent NO removed, the ammonia to NO molar ratio is close to one.
X X
Ammonia feed control is essentially achieved by measuring the gas throughput
5
and the NO concentration.
J\
Ammonia Injection
Ammonia injection, also known as selective noncatalytic reduction, was developed
and patented by Exxon Research and Engineering Corporation under the trade
name Thermal DeNo . Injection of NH-, into the boiler at temperatures
X 0
c-io
-------
ranging from approximately 1070 K to 1290 K (1470°F to 1857°F) allows reduction
of NO to N? without the need for a catalyst. However, optimal NO reduction
/\ £
occurs within a narrow temperature range, around 1240 +_ 50 K (1770 +_ 90°F).
NO reductions on the order of 90 percent have been reported under well
A
controlled laboratory conditions. This technique has been used in full-
scale applications on gas- and oil-fired industrial boilers in Japan.
However, their use is limited to emergency use during a photochemical smog
5S1
4
alert or when total plant emissions exceed the regulation. This technique
has the following limitations:
o Performance is very sensitive to flue gas temperature, and is
maximized only within a 50K temperature gradient from the optimum
temperature of about 1240K. This temperature sensitivity may
require special procedures for load-following boilers.
o Performance is very sensitive to flue gas residence time at optimum
temperatures. High flue gas quench rates are expected to reduce
process performance.
o Costs of the process can be much higher than for other combustion
controls.
o - Successful retrofit application is highly dependent on the geometry
of convective section.
o Byproduct emissions such as ammonium bisulfate might cause operational
problems, especially in coal fired boilers.
C-ll
-------
REFERENCES
1. Wilson, J.H., Jr., et al. "Cost and Economic Assessment of Regulatory
Alternatives for N02 NAAQS-Draft Report" Energy and Environmental Analysis,
Inc., EPA Contract 68-02-3536, Work Assignment 15, for U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Research
Triangle Park, NC. June 1982.
2. Hall, R.E., Industrial Environmental Research Laboratory, U.S. Environ-
mental Protection Agency. Memorandum to P.M. Johnson, Strategies and
Air Standards Division, U.S. Environmental Protection Agency, on "Review
of GCA Draft Final Report Appendices on Additional NO Analyses",
September 6, 1984.
3. Sedman, C.B., Emission Standards and Engineering Division, U.S. Environ-
mental Protection Agency. Memorandum to P.M. Johnson, Strategies and
Air Standards Division, U.S. Environmental Protection Agency, on "Review
of Appendices-Additonal N02 Analysis", August 17, 1984.
4. Lim, K.J. et al_, "Technology Assessment Report for Industrial Boiler
'Applications: NO Combustion Modification", Acurex Corporation,
EPA-600/7-79-178f. U.S. Environmental Protection Agency, Industrial
Environmental Research Laboratory, Research Triangle Park, NC, December 1979.
5. Muzio, L.J., ert a_]_, "Control of Nitrogen Oxides: Assessment of Needs
and Options, Technical Support Document, Volume 5: Emissions Control
Technology for Combustion Sources," KVB Inc., EPRI Final Report EA-2048,
Volume 5, July 1983.
6. "Control Techniques for Nitrogen Oxides Emissions from Stationary
Sources - Revised Second Edition," EPA-450/3-83-002, U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Research
Triangle Park, NC, January 1983.
C-12
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
1. REPORT NO. ' 2.
EPA-450/5-85-003
4. TITLE AND SUBTITLE
Regulatory Impact Analysis of the National Ambient
Air Quality Standards for Nitrogen Dioxide
7. AUTHOR(S)
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Office of Air Quality Planning and Standards
Office of Air and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
12. SPONSORING AGENCY NAME AND ADDRESS
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
June 1985
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT N(
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
13. TYPE OF REPORT AND PERIOD COVEREC
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report examines the costs, benefits, and other economic impacts of
implementing alternative primary annual nitrogen dioxide (N02) ambient air quality
standards (NAAQSs). Three alternative NAAQSs are investigated: 0.053 ppm, the
current air standard, 0.060 ppm, and 0.070 ppm. Results are provided for two
mobile source inspection and maintenance scenarios and for three automotive air
pollution control standards. Also discussed is the health evidence and other
environmental effects evidence that support both a primary and secondary (welfare)
NAAQS being established at the same level.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Air Pollution
Nitrogen Oxides
Nitrogen Dioxide
18. DISTRIBUTION STATEMENT
Release to Public
b. IDENTIFIERS/OPEN ENDED TERMS
Air Quality Standards
Regulatory Impact Analys"
19. SECURITY CLASS (This Report)
Unclassified
20. SECURITV CLASS 'This page)
c. COSAT1 Field/Group
S
21. NO. OF PAGES
176
22. PRICE I
\
i
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the EPA report number as it appears on the cover of the publication.
2. LEAVE BLANK
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient.
TITLE AND SUBTITLE
^itle should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
ae or otherwise subordinate it to mam title. When a report is prepared in more than one volume, repeat the primary title, add volume
mber and include subtitle for the specific title.
5. REPORT DATE
Each report shall carry a date indicating at least month and year. Indicate the basis on which it was selected (e.g., date of issue, date of
Pi -oval, date of preparation, etc.).
6. PERFORMING ORGANIZATION CODE
Leave blank.
7. AUTHOR(S)
Give name(s) in conventional order (John R. Doe, J. Robert Doe, etc.). List author's affiliation if it differs from the performing organi
zation.
8. PERFORMING ORGANIZATION REPORT NUMBER
Insert if performing organization wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational htrearchy.
10. PROGRAM ELEMENT NUMBER
Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.
11. CONTRACT/G RANT NUMBE R
Insert contract or grant number under which report was prepared.
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code.
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable, dates covered.
14. SPONSORING AGENCY CODE
Insert appropriate code.
15. SUPPLEMENTARY NOTES
Enter information not included elsewhere but useful, such as: Prepared in cooperation with, Translation of, Presented'at conference of,
To be published in, Supersedes, Supplements, etc.
16. ABSTRACT
Include a brief (200 words or less) factual summary of the most significant information contained in the report. If the report Contains a
significant bibliography or literature survey, mention it here.
17. KEY WORDS AND DOCUMENT ANALYSIS
(a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authorized terms that identify the major
concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.
(b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc. Use open-
ended terms written in descriptor form for those subjects for which no descriptor exists.
(c) COSATI HELD GROUP - Field and group assignments are to be taken from the 1965 COSATI Subject Category List. Since the ma-
jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
endeavor, or type of physical object. The application(s) will be cross-referenced with secondary Field/Group assignments that will follov
the primary posting(s).
18. DISTRIBUTION STATEMENT
Denote releasability to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
the public, with address and price.
19. &20. SECURITY CLASSIFICATION
DO NOT submit classified reports to the National Technical Information service.
21. NUMBER OF PAGES
Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.
22. PRICE
Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
EPA Form 2220-1 (Rev. 4-77) (Reverse)
------- |