External Review Draft No. 1
April 1980
Draft
Do Not Quote or Cite
Air Quality Criteria
for Particulate Matter
and Sulfur Oxides
Volume I
Summary and Conclusions
NOTICE
This document is a preliminary draft. It has not been formally released by EPA and should not at this stage be
cons oied to represent Agency policy. It is being circulated for comment on its technical accuracy and
polity implications.
Environmental Criteria and Assessment Office
Office of Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
-------
Draft
Do Not Quote or Cite
External Review Draft No. 1
April 1980
Air Quality Criteria
for Participate Matter
and Sulfur Oxides
Volume I
Summary and Conclusions
NOTICE
This document is a preliminary draft. It has not been formally released by EPA and should not at this stage be
construed to represent Agency policy. It is being circulated for comment on its technical accuracy and policy
implications.
Environmental Criteria and Assessment Office
Office of Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
-------
PREFACE
AIR QUALITY CRITERIA FOR PARTICIPATE MATTER AND SULFUR OXIDES
This criteria document has been prepared in response to Sections
108 and 109 of the 1977 Clean Air Act Amendments. The 1977 Amendments
modified Section 109 of the 1970 Clean Air Act by adding the following
subsection:
"Not later than December 31, 1980, and at five-year intervals
thereafter, the Administrator shall complete a thorough review
of the criteria published under section 108 and the national
ambient air quality standards promulgated under this section
and shall make such revisions in such criteria and standards
and promulgate such new standards as may be appropriate in
accordance with section 108 and subsection (b) of this section...."
The goal of the Clean Air Act is to protect public health and
welfare and enhance the quality of the nation's air. Under the Clean
Air Act, the Environmental Protection Agency is responsible for reviewing
air quality criteria and establishing, on a nationwide basis, ambient
air quality standards protective of health (national primary ambient air
quality standards) and welfare (national secondary ambient air quality
standards). To meet the national ambient air quality standards, the
States and Territories are responsible for developing pollutant emission
limiting regulations and strategies for controlling particular sources.
The first step in carrying out the nation's air quality management
program is to identify specific pollutants which, in the words of the
Clean Air Act, "may reasonably be anticipated to endanger the public
health or welfare." Particulate matter and sulfur oxides have been
identified as such pollutants since 1971. Once a pollutant is "listed"
-------
under Section 108(a) of the Clean Air Act, EPA publishes an air quality
criteria document which forms the scientific basis for the national
ambient air quality standard. The Act requires the criteria document to
contain the "latest scientific knowledge useful in indicating the kind
and extent of all identifiable effects on public health or welfare."
Separate criteria documents for sulfur oxides and particulate
matter were published in 1969. These documents, Air Quality Criteria
for Sulfur Oxides (AP-50) and Air Quality Criteria for Particulate
Matter (AP-49), were the bases for the national ambient air quality
standards for sulfur oxides (SO ) and total suspended particulate matter
/\
(TSP) promulgated in 1971. In this present review of criteria, EPA has
combined information on sulfur oxides and particulate matter into a
single document. Sulfur oxides and particulate matter have many common
sources, such as combustion of fossil fuels, and they often exert a
combined action which adversely affects man and his environment.
Manmade emissions of sulfur oxides come primarily from burning oil
and coal, while airborne particulate matter emanates from virtually all
of man's industrial activities. Sulfur oxides exist in the atmosphere as
gaseous sulfur dioxide (S02) and as other sulfur compounds. Much of the
sulfur from the combustion of fossil fuels enters the atmosphere as SOp.
About two-thirds of these emissions are deposited in the region of their
origin. Of the S02 which enters the atmosphere, it has been suggested
that approximately one-third of it is converted to sulfates.
The primary objective of this air quality criteria document is to
identify the effects of particulate matter and sulfur oxides on the
public health and welfare and provide a sound scientific basis for the
IV
-------
consideration of National Ambient Air Quality Standards for these pollutants.
In addition to updating and reviewing the scientific evidence for health
and welfare effects, the document discusses sources, emissions and exposures
for these pollutants.
This document is presented in four volumes. Volume I, "Summary and
Conclusions" provides an overview of the document and its major findings.
Volume II focuses on measurement methods; sources and emissions; concentrations
and exposures, and transport, transformation, and removal mechanisms. Volume
III examines the impact on the public welfare, covering effects on ecosystems
and vegetation, acidic precipitation, visibility impairment, climatological
effects, and effects in materials. This information is intended to identify
criteria for development of National Secondary Ambient Air Quality Standards.
Volume IV presents the effects on human health by examing deposition, animal
and human experimental studies and epidemiological studies. This volume
provides EPA's Administrator with information to propose and promulgate
National Primary Ambient Air Quality Standards.
EPA gratefully acknowledges the efforts and contributions of all
persons and groups who have participated in the preparation of this
document. The Environmental Protection Agency assumes full responsibility
for its content.
-------
CONTENTS
INTRODUCTION, EXECUTIVE SUMMARY, AND CONCLUSIONS 1-1
1.1 INTRODUCTION 1-1
1. 2 HISTORICAL PERSPECTIVE 1-3
1.3 SO AND PM AIR QUALITY ASPECTS 1-16
1.5.1 Chemistry and Analytical Methods 1-17
1.3.2 Air Quality Measurement Applications 1-19
1.3.2.1 British Approaches 1-20
1. 3. 2. 2 American Approaches 1-30
1.3.2.3 BS-TSP Comparison Studies 1-39
1.3.3 Sources and Emissions 1-43
1.3.4 Environmental Concentrations and Exposure 1-45
1.3.5 Transmission Through the Atmosphere 1-47
1.4 WELFARE EFFECTS ASPECTS 1-49
1.4.1 Effects on Vegetation 1^49
1.4.2 Acidic Precipitation 1-77
1.4.3 Effects on Visibility and Climate 1-84
1.4.4 Materials Damage and Soiling Effects 1-87
1.5 HEALTH EFFECTS ASPECTS 1-90
1.5.1 Respiratory Tract Deposition and Biological Fate 1-91
1. 5. 2 Animal Toxicology Studies 1-93
1.5.2.1 Effects of Acute and Chronic Exposure to
Particles or SOp 1-95
1.5.2.2 Effects of Exposures to Combinations of SO
and Particles .... 1-103
1.5.3 Studies on the Oncogenic Properties of SO and PM 1-107
1.5.4 Experimental Investigations of Human Subjects 1-108
1.5.5 Community Health Observational Studies 1-113
1.5.5.1 Overview Summary of Chapter 14 Contents 1-115
1.5.5.2 Methodological Factors Impacting
Interpretation of Results 1-133
1.5.5.3 Quantitative Dose-Response Relationships
Defined by Community Health Studies 1-136
-------
LIST OF TABLES
Jable Page
1-1 Summary of evaluation of sources, magnitudes, and directional
biases of errors associated with British SCL measurements 1-22
1-2 Summary of evaluation of sources, magnitudes, and directional
biases of errors associated with American S0? measurements 1-27
1-3 Summary of evaluation of sources, magnitudes, and directional
biases of errors associated with British smoke (particulate)
measurements 1-32
1-4 Summary of evaluation of sources, magnitudes, and directional
biases of errors associated with American total suspended
particulate (TSP) measurements 1-36
1-5 Dose-response information summarized from literature pertaining
to cultivated agronomic crops as related to folilar, yield, and
specific effects induced by increasing S09 dose 1-52
c.
1-6 Dose-response information summarized from literature pertaining
to forest tree species as related to foliar, yield, and specific
effects induced by increasing S0? dose 1-63
1-7 Dose-response information summarized from literature
pertaining to native plants as related to foliar, yield
and specific effects induced by increasing S02 dose 1-69
1-8 Summary of effects of acute exposure to < mg/m particles3
in animals 1-96
1-9 Summary of effects of chronic exposure to < mg/m particles3
in animals 1-98
3
1-10 Summary of effects of exposure to < 13.1 mg/m (5 ppm)
sulfur dioxide in animals 1-99
1-11 Summary of effects of combinations of particles3 (< mg/m ) and
sulfur dioxide (13.1 mg/m , 5 ppm) in animals 1-104
1-12 Pulmonary effects of aerosols 1-109
1-13 Effects of S02 1-110
1-14 Pulmonary effects of SO,, and other air pollutants 1-111
VII
-------
l-14a Qualitative association of geographic differences in mortality
with residence in areas of heavy air pollution 1-117
1-15 Qualitative studies of air pollution and acute respiratory
disease 1-120
1-16 Summary table - acute exposure effects 1-123
1-17 Qualitative studies of air pollution and prevalence of chronic
respiratory symptoms and pulmonary function declines 1-127
1-18 Summary table - chronic exposure effects 1-131
1-19 Summary of various reviewers' evaluations of quantitative
dose-response relationships derived from studies of
mortality effects associated with acute exposures to
S0? and parti culate matter 1-140
1-20 Summary of various reviewers' evaluations of quantitative
dose-response relationships derived from studies of
morbidity effects associated with acute exposures to
SO™ and parti cul ate matter 1-141
1-21 Summary of various reviewers' evaluations of quantitative
dose-response relationships derived from studies of
morbidity effects associated with chronic exposures to
S02 and particulate matter 1-142
1-22 Expected effects of air pollutants on health in selected
segments of the population: effects of short-term exposures 1-150
1-23 Expected effects of air pollutants on health in selected
segments of the population: effects of long-term exposures 1-151
1-24 World Health Organization Guidelines for exposure limits
consistent with the protection of public health 1-152
-------
LIST OF FIGURES
Number Page
1-1 Comparison of smoke calibration curves for Eel reflectometer,
Whatman No. 1 paper and a 1-in diameter filter. From WSL (1967).
The computer followed curve D during 1961-64 instead of the
correct curve(s) B and C 1-29
1-2 Representative examples of BS/TSP relationships defined by linear
regression analyses employed to fit BS/TSP comparison data points
as described in published reports 1-40
1-3 Measurements of British Smoke vs Hi-vol TSP, showing a consistent
relation between these measures over the entire range of reported
observations. Most points shown are annual mean values; see text
for discussion 1-42
1-4 Comparison of interpretations of studies evaluated by Holland
et al. (1979), WHO (1979), and other reviews such as those
in the NRC/NAS documents and the present chapter 1-149
IX
-------
CONTENTS
VOLUMES I, II, III, AND IV
Volume I. Summary and Conclusions
Chapter 1. Introduction, Executive Summary, and Conclusions 1-1
Volume II. Air Quality
Chapter 2. Physical and Chemical Properties of Sulfur Oxides and
Particulate Matter and Analytical Techniques for Their
Measurement . 2-1
Chapter 3. Critical Assessment of Practical Applications of Sulfur
Oxides and Particulate Matter Measurement Techniques... 3-1
Chapter 4. Sources and Emissions 4-1
Chapter 5. Environmental Concentration and Exposure 5-1
Chapter 6. Transmission Through the Atmosphere 6-1
Volume III. Welfare Effects
Chapter 7. Effects on Vegetation 7-1
Chapter 8. Acidic Precipitation 8-1
Chapter 9. Effects on Visibility and Climate 9-1
Chapter 10. Effects on Materials 10-1
Volume IV. Health Effects
Chapter 11. Respiratory Deposition and Biological Fate of Inhaled
Aerosol s and SO- 11-1
Chapter 12. Toxicological Studies 12-1
Chapter 13. Controlled Human Studies 13-1
Chapter 14. Epidemiological Studies of the Effects of Atmospheric
Concentrations of Sulfur Dioxide and Particulate Matter
on Human Health 14-1
-------
CONTRIBUTORS AND REVIEWERS
Mr. John Acquavella
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Roy E. Albert
Institute of Environmental Medicine
New York University Medical Center
New York, New York 10016
Dr. Martin Alexander
Department of Agronomy
Cornell University
Ithaca, New York 14850
Dr. A. P. Altshuller
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. David S. Anthony
Department of Botany
University of Florida
Gainesville, Florida 32611
Mr. John D. Bachmann
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Allen C. Basala
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Neil Berg
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Michael A. Berry
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
XI
-------
Mr. Francis M. Black
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Joseph Blair
Environmental Division
U. S. Department of Energy
Washington, D.C. 20545
Dr. Edward Bobalek
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Ms. F. Vandiver P. Bradow
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Ronald L. Bradow
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Robert Bruce
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Angelo Capparella
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North'Carolina 27711
Dr. Robert Chapman
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Robert J. Charlson
Department of Environmental Medicine
University of Washington
Seattle, Washington 98195
Dr. Peter Coffey
New York State Department of Environmental Conservation
Division of Air Resources
Albany, New York 12233
Mr. Chatten Cowherd
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
xii
-------
Dr. Ellis B. Cowling
School of Forest Resources
North Carolina State University
Raleigh, North Carolina 27650
Mr. William M. Cox
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. T. Timothy Crocker
Department of Community and Environmental Medicine
Irvine, California 92664
Mr. Stanley T. Cuffe
Emission Standards and Engineering Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Thomas C. Curran
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Michael Davis
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Gerrold A. Demarrais
National Oceanic and Atmospheric Administration
U. S. Department Of Commerce
Dr. Jerrold L. Dodd
Natural Resources Ecology Laboratory
Colorado State University
Fort Collins, Colorado 80523
Dr. Thomas G. Dzubay
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Thomas G. Ellestad
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. John Evans
School of Public Health
Harvard University
Boston, Massachusetts 02115
-------
Dr. Lance Evans
Department of Energy and Environment
Brookhaven National Laboratory
Upton, New York 11973
Mr. Douglas Fennel!
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Benjamin G. Ferris, Jr.
School of Public Health
Harvard University
Boston, Massachusetts 02115
Mr. Patrick Festa
New York Department of Environmental Conservation
Division of Fish and Wildlife
Albany, New York 12233
Mr. Terrence Fitz-Simmons
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Christopher R. Fortune
Northrop Services, Inc.-Environmental Sciences
P. 0. Box 12313
Research Triangle Park, North Carolina 27709
Dr. Robert Frank
Department of Environmental Health
University of Washington
Seattle, Washington 98195
Dr. Warren Galke
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Phil Galvin
New York Department of Environmental Conservation
Division of Air Resources
Albany, New York 12233
Dr. Donald Gardner
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. J.H.B. Garner
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xiv
-------
Dr. Donald Gillette
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Judy Graham
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Lester D. Grant
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Armin Gropp
Department of Chemistry
University of Miami
Miami, Florida 33124
Dr. Jack Hackney
Rancho Los Amigos Hospital
Downey. California 90242
Mr. Bertil Hagerhall
Ministry of Agriculture
Pack
S-163 20 Stockholm
Sweden
Dr. Douglas Hammer
2910 Wycliff Road
Raleigh, North Carolina 27607
Mr. R. P. Hangebrauck
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Thomas A. Hartlage
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Victor Hasselblad
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Thomas R. Hauser
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xv
-------
Dr. Car] Hayes
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Fred H. Haynie
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Walter Heck
Department of Botany
North Carolina State University
Raleigh, North Carolina 27650
Dr. Howard Heggestad
USDA-SAE
The Plant Stress Laboratory
Plant Physiology Institute
BeltsviTle, Maryland 20705
Dr. George R. Hendrey
Department of Energy and Environment
Brookhaven National Laboratory
Upton, New York 11973
Dr. Ian Higgins
Department of Epidemiology
School of Public Health
University of Michigan
Ann Arbor, Michigan 48109
Mrs. Patricia Hodgson
Editorial Associates
Chapel Hill, North Carolina 27514
Mr. George C. Holzworth
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Robert Horton
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Steven M. Horvath
Institute of Environmental Stress
University of California
Santa Barbara, California 93106
XV7
-------
Dr. F. Gordon Hueter
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Janja Husar
CAPITA
Washington University
St. Louis, Missouri 63130
Dr. Rudolf Husar
Department of Mechanical Engineering
Washington University
St. Louis, Missouri 63130
Dr. William T. Ingram
Consulting Engineer
7 North Drive
Whitestone, New York 11357
Dr. Patricia M. Irving
Argonne National Laboratory
9700 South Cass Avenue
Argonne, Illinois 60439
Dr. Jay Jacobson
Boyce Thompson Institute
Cornell University
Ithaca, New York 14850
Mr. James Kawecki
Biospherics, Inc.
4928 Wyaconda Road
Rockville, Maryland 20852
Dr. Sagar V. Krupa
Department of Plant Pathology
University of Minnesota
St. Paul, Minnesota 55108
Dr. Edmund J. LaVoie
Section of Metabolic Biochemistry
American Health Foundation
Dana Road
Valhalla, New York 10592
Dr. Michael D. Lebowitz
Arizona Health Sciences Center
1501 North Campbell
Tucson, Arizona 85724
xvii
-------
Dr. Robert E. Lee
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Allan H. Legge
Environmental Science Center
University of Calgary
Calgary, Alberta, Canada T2N 1N4
Ms. Peggy Le Sueur
Atmospheric Environment Service
Downsview, Ontario, Canada M3H5T4
Dr. Morton Lippmann
Institute of Environmental Medicine
New York University
New York, New York 10016
Dr. James P. Lodge
385 Broadway
Boulder, Colorado 80903
Dr. Gory J. Love
Institute of Environmental Studies
University of North Carolina
Chapel Hill, North Carolina 27514
Dr. David T. Mage
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Delbert McCune
Boyce Thompson Institute
Cornell University
Ithaca, New York 14850
Mr. Frank F. McElroy
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. David J. McKee
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Thomas McMullen
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xvm
-------
Dr. Daniel B. Menzel
Department of Pharmacology
Duke University Medical Center
Durham, North Carolina 27710
Dr. Edwin L. Meyer
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Fred Miller
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. John 0. Mil liken
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Jarvis Moyers
Department of Chemistry
University of Arizona
Tucson, Arizona 85721
Dr. Thaddeus J. Murawski
New York State Department of Health
Empire State Plaza
Albany New York 12337
Dr. David S. Natusch
Department of Chemistry
Colorado State University
Fort Collins, Colorado 80523
Dr. Stephen A. Nielsen
Environmental Affairs
Joyce Environmental Consultants
414 Live Oak Boulevard
Casselberry, Florida 32707
Dr. Kenneth Noll
Department of Environmental Engineering
Illinois Institute of Technology
Chicago, Illinois 60616
Mr. John R. O'Connor
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xix
-------
Mr. Thompson G. Pace
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Jean Parker
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Nancy Pate
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Thomas W. Peterson
Department of Chemical Engineering
University of Arizona
Tucson, Arizona 85721
Mr. Martin Pfeiffer
New York State Department of Environmental Conservation
Bureau of Fisheries
Raybrook, New York 12977
Dr. Marlene Phillips
Atmospheric Chemistry Division
Environment Canada
Downsview, Ontario, Canada M3H5T4
Dr. Charles Powers
Environmental Research Laboratory
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
Mr. Larry J. Purdue
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. John C. Puzak
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xx
-------
Dr. Otto Raabe
Radiobiology Laboratory
University of California
Davis, California 95616
Mr. Danny Rambo
Environmental Research Laboratory
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
Mr. Kenneth A. Rehme
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Elmer Robinson
Department of Chemical Engineering
Washington State University
Pullman, Washington 99163
Mr. Charles E. Rodes
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Douglas R. Roeck
GCA Corporation
Technology Division
Burlington Road
Bedford, Massachusetts 01730
Mr. J. C. Romanovsky
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. August Rossano
University of Washington
Seattle, Washington 98195
Mr. Joseph D. Sableski
Control Programs Development Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Dallas Safriet
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
xxi
-------
Dr. Victor S. Salvin
University of North Carolina at Greensboro
Greensboro, North Carolina 27408
Dr. Shahbeg Sandhu
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. Joseph P. Santodonato
Life and Material Sciences Division
Syracuse Research Corporation
Merrill Lane
Syracuse, New York 13210
Dr. Herbert Schimmel
Neurology Department
Albert Einstein Medical College
26 Usonia Road
Pleasantville, New York 10570
Dr. Carl L. Schofield
Department of Natural Resources
Cornell University
Ithaca, New York 14850
Dr. David Shriner
Environmental Sciences Division
Oak Ridge National Laboratory
Ms. Donna Sivulka
Environmental Criteria and Assessment Office
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. John M. Skelly
Department of Plant Pathology and Physiology
Virginia Polytechnic Institute
Blacksburg, Virginia 24061
Mr. Scott Smith
Biospherics, Inc.
4928 Wyaconda Road
Rockviell. Maryland 20852
Ms. Elaine Smolko
Department of Pharmacology
Duke University Medical Center
Durham, North Carolina
xxn
-------
Dr. Frank Speizer
School of Public Health
Harvard University
Boston, Massachusetts 02115
Dr. John D. Spengler
School of Public Health
Harvard University
Boston, Massachusetts 02115
Mr. Robert K. Stevens
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. George E. Taylor, Jr.
Environmental Sciences Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37830
Dr. Larry Thibodeau
School of Public Health
Harvard University
Boston, Massachusetts 02115
Dr. W. Gene Tucker
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. D. Bruce Turner
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. James B. Upham
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. Robert Waller
Toxicology Unit
St. Bartholomew's Hospital
London, England
Mr. Stanley Wall in
Warren Spring Laboratory
Department of Industry
Stevenage, Hertfordshire SGI 2BX
England
xxm
-------
Dr. Joseph F. Walling
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. James Ware
School of Public Health
Harvard University
Boston, Massachusetts 02115
Dr. David Weber
Office of Air, Land, and Water Use
U.S. Environmental Protection Agency
Washington, D. C. 20460
Dr. Jean Weister
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Mr. R. Murray Wells
Radian Corporation
8500 Shaol Creek Boulevard
Austin, Texas 78766
Dr. Kenneth T. Whitby
Mechanical Engineering Department
University of Minnesota
Minneapolis, Minnesota 55455
Dr. Warren White
CAPITA
Washington, University
St. Louis, Missouri 63130
Dr. Raymond Wilhour
Environmental Research Laboratory
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
Dr. William E. Wilson
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. John W. Winchester
Department of Oceanography
Florida State University
Tallahassee, Florida 32306
XXIV
-------
Mr. Larry Zaragoza
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dr. William H. Zoller
Chemistry Department
University of Maryland
College Park, Maryland 20742
We wish to thank everyone who contributed their efforts to the preparation of
this document, including the following staff members of the Environmental
Criteria and Assessment Office, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina:
Mrs. Dela Bates
Ms. Hope Brown
Ms. Diane Chappell
Ms. Deborah Doerr
Ms. Mary El ing
Ms. Bettie Haley
Mr. Allen Hoyt
Ms. Susan Nobs
Ms. Evelynne Rash
Ms. Connie van Oosten
Ms. Donna Wicker
The final draft of this document will cite the many persons outside of the
Environmental Criteria and Assessment Office who have assisted in its pre-
paration.
xxv
-------
1. INTRODUCTION, EXECUTIVE SUMMARY, AND CONCLUSIONS
1.1 INTRODUCTION
Section 109(d) of the Clean Air Act as amended in 1977 requires the
Environmental Protection Agency (EPA) to review the air quality criteria for
sulfur oxides (S0x) and particulate matter (PM) and to revise related standards,
as appropriate, by December 31, 1980, and at 5-year intervals thereafter. In
addition, the Administrator may from time to time review and, where appropriate,
modify pertinent criteria under the authority of Section 108(c) of the Clean
Air Act. Pursuant to the above provisions of the Clean Air Act, the present
document represents an important step in the preparation of the revised
criteria documents for SO and PM.
/\
The decision to issue a single document for these two pollutants was
based on several considerations. These include primarily the recognition
that: (1) both SO and PM originate from many common emission sources such
as the burning of fossil fuels; (2) SO and PM levels frequently covary in
ambient air; (3) SO and PM appear to likely act together to adversely affect
/^
both man and his environment; (4), in community health studies and other
types of studies where combined exposure to SO and PM is a factor, it is
difficult to isolate the effects of one pollutant from those of the other;
and (5) the combination of sulfur oxides and particulate matter has been
linked with acidic precipitation, which damages aquatic and terrestrial
ecosystems.
This document critically reviews scientific information bearing on
health and welfare criteria that will form the bases for National Ambient Air
Quality Standards (NAAQS) for S0x and PM.
1-1
-------
The information reviewed concerns: (1) pertinent air quality Information;
(2) welfare effects associated with S0x and PM; and (3) health effects associated
with exposure to those air pollutants. The present volume (Volume I) includes
the general introduction, executive summary and conclusions for the entire
document. The second volume (Volume II) on air quality aspects, Includes
chapters discussing: physical properties of SOX and PM and air quality
measurement techniques for each; critical appraisal of practical applications
of such measurement methods; sources of SO and PM emissions; related
/\
atmsopheric transport and transformation processes; and ambient air
concentrations and exposure levels. Volume III, concerning welfare effects,
contains document chapters on: SO and PM effects on vegetation and natural
ecosystems; acidic precipitation formation and effects; effects on visibility
and climate; and materials damage effects. The last volume (Volume IV),
dealing with health effects of SO and PM contains chapters on: the uptake,
)\.
deposition, and absorption of SO and PM health effects; human clinical
(experimental) studies of SO and PM health effects; and pertinent community
f\
health (epidemiology) studies.
In the present volume, there is first provided a brief historical review
of important events which have (1) contributed to concern about the health
and welfare effects of SO and PM; (2) stimulated extensive and highly
varied related research efforts; and (3) helped to shape social/ legal
actions taken to eliminate or ameliorate such effects as major public health
and welfare problems. It is hoped that such a historical review will provide
background material useful in assisting readers, especially those not well
familiar with the present subject matter, to develop a better Informed vantage
point or perspective from which to view the information subsequently presented
in this document. The major points addressed by such information and important
1-2
-------
conclusions regarding them, as discussed in each of the remaining three
volumes of the document, are then summarized following the historical
perspective discussion presented below.
Before proceeding with the historical review, it should be mentioned
that a number of other important reviews and commentaries concerning the
health and welfare effects of SO and PM have appeared during the past decade.
In that regard, the reader is referred to certain review articles and documents
appearing in the mid-to late 1970s for additional information on the present
subject matter. Such materials include critical reviews and commentaries
written by Rail (1974), Higgins et al. (1974), Goldsmith and Friberg (1977),
Ferris (1978), and Waller (1978). They also include the following evaluative
documents appearing in 1978: an American Thoraic Society (ATS) review of
Health Effects of Air Pollution (1978); a National Research Council/National
Academy of Science (NRS/NAS) document on Airborne Particles (1978); and an
NRC/NAS document on Sulfur Oxides (1978). More recent such reviews and
commentary appearing in 1979 include: the 1979 World Health Organization
(WHO) document, Environmental Health (8): Sulfur Oxides and Suspended
Particulate Matter; a report by Holland et al. (1979) written for the
American Iron and Steel Institute and appearing in the American Journal of
Epidemiology; and a reply to that report in the same journal by Shy (1979).
1.2 HISTORICAL PERSPECTIVE
There is little question that severe air pollution generated by anthro-
pogenic activities has long exerted significant, even lethal, adverse effects
on the health and welfare of many industrial societies. It has only been
1-3
-------
within the past 50 years or so, however, that sufficient recognition of suet
effects as major public problems has stimulated both (1) extensive research to
better understand such problems and (2) strong governmental and private
sector actions to control or eliminate them.
It is not surprising that the air pollution incidents most often cited
as signal events precipitating great concern about air pollution problems and
strong actions in response to them occurred among the most heavily industrialized
societies then extant, i.e., in Western Europe, Great Britain and the United
States. It is also easy to understand the urgency associated with taking
strong actions to control such problems in light of the magnitude and seriousness
of the effects experienced during those incidents.
For example, in describing the effects of a thick fog that covered the
industrial Meuse Valley in Belgium during early December 1930, Firket (1931)
noted that several hundred people were afflicted by suddenly appearing acute
respiratory symptoms, complicated in many instances by serious cardiovascular
failure. Firket (1931) further noted: "More than sixty died on the 4th and
5th of December after only a few hours of sickness. A sizeable number of
livestock had to be slaughtered." Also, taking into account that mortality
rates were more than 10 times normal, Firket projected that over 3,000 deaths
would occur if a similar fog were to occur in a city the size of London.
Twenty-two years later, such an event did occur in London and more than
4,000 deaths appeared to be attributable to the four-day London Fog of December
1952, according to Logan (1953). Logan further noted: "The incident was a
catastrophe of the first magnitude in which, for a few days, death rates
£.
attained a level that has been exceeded only rarely during the past hundred
1-4
-------
years - for example, at the height of the cholera epidemic of 1854 and of the
influenza epidemic of 1918-19." Indeed, the death rate rivaled or exceeded
that on many of the worst days of other catastrophic events afflicting London
injts more recent past, e.g., the Battle of Britain during World War II.
Similar catastrophic air pollution incidents also occurred around the
same time in the United States. Almost half of the residents of Donora,
Pennsylvania, for example, were afflicted with respiratory symptoms as the
result of a "smog" covering the coke- and steel-producing Monongahela River
Valley during October, 1948 (Schrenk et al., 1949). Twenty people in the
small town of about 10,000 population died during the final week of the
"Donora Smog Episode," in comparison to the 2 or 3 deaths normally expected
for the same period. Over the next few years, catastrophic air pollution
incidents also affected other United States communities. One such case was a
dramatic increase in infant mortality in Detroit, Mich., attributed to a
"pollution incident" in September 1952 (Int. Joint Commission, 1960). This
was followed by marked increases in respiratory distress cases and fatalities
attributable (Greenburg et al., 1962) to air pollution occuring on an even
larger scale during a "Thanksgiving Day" pollution episode in New York City
in November, 1953—an experience later to be repeated several more times in
New York City during the early 1960s.
It was clear from the above incidents and others occurring elsewhere in
the industrialized world that air pollution, especially under certain weather
conditions leading to stagnant masses of pollutant-laden "fog" or "smog", was
capable of causing incidents rivaling natural disease epidemics or man-made
1-5
-------
wartime disasters that constituted national emergencies for the affected
societies. Equally clear was the urgent need to take immediate action to
avert or reduce in severity future air pollution disasters. Among the first
to_act were two of the most severely affected countries, Great Britain and
the United States.
Many parallel historical threads can be discerned regarding the paths
taken by the two nations in trying to cope with air pollution problems,
including a number of similar mistakes as well as successes. In each country,
for example,, there occurred extensive expansion of epidemiological and
toxicological research aimed at identifying components of the killer smogs or
fogs responsible for the observed lethal effects and increased morbidity.
Also, in each country, although occurring within different specific time
frames and at different specific paces or rates, there ensued the upgrading
and expansion of air quality monitoring networks. This included the expansion
of monitoring networks capable of indexing high levels of various industrial
pollutants implicated by the historical data as being associated with increased
incidences of both morbidity and mortality. The latter invariably included
oxides of sulfur and particulate matter present in the killer fogs or smogs
of the 1940s and 1950s, although the relative contributions of each could not
be precisely linked to observed health effects.
It is important to note in regard to the air quality data that very
precise accuracy and specificity were not then demanded; rather, "benchmark"
ranges of estimates of air pollutant concentrations were acceptable, especially
for indexing the rather high levels of pollution associated with severe
1-6
-------
morbidity or mortality effects. It mattered little if the exact amounts of
S02, for example, were 143 or 781 or 1408 ug/m plus or minus 5 or 10 percent,
when one was concerned that people would begin to experience severe morbidity
or_even die when S02 or particulate matter exceeded several hundred ug/m .
Being able to establish that increased mortality or severe morbidity occurred
at, say, a range of 550 to 600 or even 500 to 1000 ug/m of SOp or particulate
matter in comparison to the incidence of such effects observed at, say, 70 to
3
100 or even 50 to 200 ug/m of the same pollutants was, understandably,
sufficient scientific evidence to support political and social measures
needed to avert the worst air pollution episodes.
Nor was it particularly important whether or not SOp, specifically, or
particulate matter of whatever specific size-range or chemical composition
could be precisely implicated as the "culprit" causing one or another very
specific health effect. Rather, it was sufficient to recognize that those
substances might not be any more than representative indicators of the total
mix or some other potentially lethal subfraction of pollutants typically
present during the dangerous past pollution episodes. Basically, the main
objective, regardless of specific fine details associated with various pollution
situations, was to obtain sufficient information: (1) to allow for reasonable
conclusions to be drawn regarding ranges of air levels of various pollutants
(or indices) empirically linked to the occurrence of severe health effects
and (2) to help serve as a guide in directing pollution control efforts
toward sources emitting such pollutants (or mixes containing them).
In addition to the above developments, certain regulatory actions were
L
taken in response to the severe air pollution situations in Britain and the
1-7
-------
United States in the 1950s. The British acted strongly on a national scale
by passing the Clean Air Act of 1956 which forced implementation of very
stringent controls on emissions from coal-fired combustion sources, essentially
eliminating the use of coal for home-heating and many industrial purposes in
heavily industrialized and congested urban areas. The effectiveness of those
measures was reflected in the resulting declines seen by the early 1960s in
both atmospheric sulfur oxides and particulate matter and associated mortality
effects (WSL, 1967).
As for the United States, action initially tended to be taken mostly on
more restricted geographic bases and consisted mainly of local or state air
pollution control ordinances being passed, often with the cooperation of
local industrial leaders, to reduce air pollution. Among the more notable
examples were early actions taken to control air pollution in one of the most
heavily industrialized regions of the country, the Pittsburgh area. There,
and in other areas of Pennsylvania, extensive steel-making and coking operations
and the burning of locally produced high-sulfur coal contributed to widespread
elevations in both particuate matter and sulfur dioxide air concentrations.
Notable improvements in air quality in various American regions were attained
as the result of such initial actions. However, simultaneous deterioration
in air quality in many other communities or states lacking effective pollution
control laws, and the growing recognition of air pollution as a multi-state,
regional or national problem eventually led to the passage in the United
States of the Clean Air Act of 1970 and consequent promulgation of National
Ambient Air Quality Standards for sulfur oxides, particulate matter, and
other air pollutants of widespread concern.
Subsequent to many of the above actions taken in the 1950s and early
1960s, questions began to be raised regarding what less severe, but important,
1-8
-------
effects on human health and welfare might be caused by lower air levels of
oxides of sulfur, particulate matter, and other air pollutants, especially
Under conditions of prolonged periods (months or years) of exposure. This
included increased recognition of and research on such problems as SO and PM
_ „ A
effects on vegetation, visibility, and climate, as well as related materials
damage effects and acidic precipitation formation and effects. Again, fairly
similar paths were followed in both Great Britain and the United States in
trying to deal with such questions.
In both countries, epidemiologists began to design studies to either
retrospectively or prospectively define in more precise terms both qualitative
and quantitative relationships between elevations of various pollutants in
the ambient air and specific sublethal health effects, such as acute or
chronic bronchitis, respiratory infections, temporary decrements in pulmonary
function, and asthma attacks. Air quality monitoring networks set up or
expanded earlier to provide at least representative, but not necessarily
thorough, coverage of geographic areas having varying pollution levels were
often looked to by the epidemiologists to provide requisite, albeit less than
perfect, quantitative estimates of air quality to help define quantitative
air pollution-effect relationships. Only in relatively rare circumstances
were sufficient funds or other resources available to allow for more thorough
monitoring coverage to be arranged specifically for collection of community
air quality data to be coupled with health endpoint measurements. Similarly,
expanded demands were placed on air monitoring capabilities in terms of their
being needed to provide requisite air quality data for use in evaluating
t
various welfare effects and related atmospheric transport and transformation
phenomena.
1-9
-------
At the same time as the above expansion of research efforts, more precise
estimates of air pollutant levels were being demanded or expected from existing
or expanding air monitoring networks to help index progress in reductions of
air pollutants in Britain and the United States. Bearing on this point, the
following was later noted in Her Majesty's Report on the Investigation of
Atmospheric Pollution 1958-1966 (Warren Spring Laboratory, 1967):
The industrial provisions of the Clean Air Act 1956 had come
into force by June 1958, and the first smoke control areas were
declared under the domestic provisions in April 1958. This timetable
added urgency to the view that the existing survey of air pollution
throughout the United Kingdom was too blunt an instrument either to
assess the benefits accruing from the Act or to guide its future
application, and that a scientifically planned National Survey was
necessary. Such a survey was designed and the co-operation of the
local authorities concerned was obtained where measurements were
required in addition to those already being made. Observations were
started in the winter of 1962-63. Not only has the whole pattern of
co-operative observations been transformed in this way, but so also
has the basis of the co-operation: the local authorities and other
organizations are now making measurements as required to conform to
an overall statistical plan.
A growing need was also felt for establishing or maintaining monitoring
systems using sufficiently uniform and reliable measurement approaches to
allow for comparability of air quality data from disparate geographic sites.
This need became perhaps most acutely felt in the United States in the late
1960s in light of the growing prospect of having to attain national ambient
air quality standards at lower air concentrations than were earlier
envisaged. Pressure to meet such needs and demonstrate the benefits of air
pollution control in the United States intensified greatly with the passage
of the Clean Air Act of 1970.
f. Both the above needs and the anticipated use of air quality data in
future epidemiology and welfare effects studies increased the necessity for
1-10
-------
better air quality data in terms of their specificity, sensitivity, accuracy,
precision, and reliability. Thus, it became increasingly more important to
be able to distinguish with confidence between much closer absolute levels of
afr pollutants even at relatively low ambient air concentrations (e.g., at
levels around or lower than 100 to 200 pg/m3 of either S02 or particulate
matter). In other words, the emerging demands required practical applications
of available or newly developed measurement techniques to meet previously
unheard-of levels of accuracy. At times expectations may have exceeded what
could realistically be achieved with field applications of available technology,
especially in comparison to theoretical limits of what could be achieved with
particular measurement techniques under ideal laboratory conditions.
In responding to the above demands, several similar historical parallels
can again be discerned between the British and American experiences, as well
as some quite significant differences. In Britain, steps were taken to
assure that a high degree of uniformity in measurements of pollutants was
maintained or further enhanced; this included the establishment of the National
Air Pollution Survey alluded to above. As part of this effort, the British
Smoke (BS) filter method, widely used in Britain since the early 1900s for
the measurement of black suspended particulate matter, was officially adopted
to monitor air quality across the United Kingdom. The daily smoke filter has
since been the standard air particulate matter measurement instrument used in
the United Kingdom, and responsibility for quality assurance for the use of
the instrument throughout the entire United Kingdom was assigned to the
Warren Spring Laboratory (WSL) of the Department of Industry, where central
coordination and evaluation of uniformity, accuracy, and reliability of all
National Survey air quality measurement have been carried out since the late
1-11
-------
1950s. Based largely on WSL evaluations, the daily smoke filter was later
adopted as a standard also by the British Standards Institute and became one
of OECD's recommended procedures for measuring suspended particulate matter.
The WSL was also instrumental in selecting S02 measurement methods for
the British National Survey air monitoring program. Under WSL guidance, the
hydrogen peroxide method was adopted in the early 1960s as the standard
approach for measuring SOp. This replaced the lead dioxide gauge still in
use at many sites in Britain (in 1962 there were about 1200 gauges in operation)
The decision to adopt the hydrogen peroxide method was based on recommendations
of a WSL-organized "Work Party" that considered both the deliberations of an
OECD Working Party and detailed comparisons of the two methods under field
conditions by WSL.
Turning to concomitant developments in the United States during the past
twenty years, one can discern a lage in the development of a standardized
nationwide approach to air pollution monitoring and control. During the
1950s and early 1960s numerous air monitoring systems established by different
governmental units sprang up around the country, often to meet needs associated
with enforcement of newly enacted air pollution control ordinances. It was
not until the mid-to-late 1960s that effective procedures were implemented to
feed data obtained from the multiplicity of air monitoring sites operated by
city, county, state, and federal agencies into a central data bank as part of
a National Air Survey Network (NASN). Even then, those data were still often
derived from different measurement methods used by various agencies for
monitoring of a given pollutant. Generally, the high-volume TSP measurement
method developed in the United States in the 1950s was used for assessing
airborne particulate levels; but, at times, other methods such as coefficient
of haze (CoH) measurements were also employed to assess particulate levels.
1-12
-------
Similarly, several approaches, e.g., the West-Gaeke or sulfation methods,
were used to measure SOp.
Considerably greater uniformity in monitoring approaches has, however,
beeii achieved over the past decade or so through the publication of "Federal
Reference Methods." Additional efforts were undertaken in the late 1960s and
early 1970s by EPA and its predecessor Federal agencies to establish a new
nationwide air monitoring network using uniform measurement methods. That
network was established as part of what became known as EPA's Community
Health and Environment Surveillance System (CHESS) Program. The "CHESS"
monitoring network, set up in addition to other Federal air sampling stations
used for monitoring compliance with air regulations, included monitoring
sites dispersed in widespread urban and semi-rural areas of the United States
to provide air quality data representative of pollutant exposures experienced
by surrounding population groups. Various health endpoints were evaluated
for those population groups as part of CHESS Program epidemiology studies.
Thus, the CHESS monitoring network, including sampling sites often situated
near or along side local or state monitoring sites, was designed to provide
air quality data from a nationwide network using uniform measuring methods
that supplemented other data entered into the NASN data bank. The hi-volume
TSP sampling procedure was usually employed in the CHESS Program to monitor
atmospheric particulate matter levels and the West-Gaeke method was generally
employed for SO,, measurements, along with additional procedures for estimating
suspended sulfates discussed later. The series of major air pollution/health
effects studies carried out between 1969 and 1975 as part of the CHESS Program,
«.
and coupling such aerometric measurements with community health surveys, have
been considered by many experts as the most comprehensive of their kind.
1-13
-------
It should be noted, however, that a number of methodological problems
were engendered by attempts to rapidly deploy air monitoring stations at
widespread sites across the United States and to bring them up to full operational
status in time to collect air quality data to be coupled with health surveys
as part of the CHESS Program. Of particular concern are problems which were
detected regarding the collection of health endpoint data and certain errors
in air quality data generated from CHESS network sampling sites—several
types of errors which were either not detected at all during the CHESS health
endpoint data collection period (1969-75) or were only detected and corrected
through improved quality control procedures implemented and applied in the
last few years of the Program (i.e. 1972 or 1973 onward).* Improvements in
the air quality data obtained during the last few years of the CHESS Program
and, also, in other EPA air monitoring efforts,, were accomplished via a
substantially expanded in-house EPA quality assurance program. That program,
conducted by EPA's Environmental Monitoring Systems Laboratory in Research
Triangle Park, N.C., has since provided quality assurance backup for all of
EPA's research and nationwide enforcement air monitoring activities.
: s discussed in more detail later in Chapters 3 and 14, the matter of errors
in air quality data collected as part of the CHESS Program (together with
other concerns regarding the collection of health endpoint data in CHESS
; jdies) contributed to considerable controversy regarding the validity and
accuracy of results of early CHESS studies, as interpreted and reported in a
1974 EPA monograph entitled "Health Consequences of Sulfur Oxides: A Report
from CHESS" 1970-71, U.S.EPA Document No. EPA-650/1-74-004 (May 1974). The
controversy eventually led to the 1974 "CHESS Monograph" becoming the subject
of U.S. Congressional oversight hearings in 1976. Subcommittees of the U.S.
Ho&se of Representatives Committee on Science and Technology produced a report
on the Monograph, other aspects of the CHESS Program, and EPA's air pollution
research programs generally--a report entitled "The Environmental Protection
Agency's Research Program with Primary Emphasis on the Community Health and
Surveillance System (CHESS): An Investigative Report." Of primary importance
for the present discussion, that report, widely referred to either as the
1-14
-------
From the foregoing, it can be seen that similar needs and demands as
those felt by the British in conjunction with the passage of their Clean Air
Act of 1956 were later experienced by the Americans with the passage of their
Clean Air Act of 1970. Remarkably similar paths were also followed in both
countries in responding to those needs; that is, in both cases intensive
efforts were carried out to rapidly expand and improve air monitoring
capabilities, including introduction of more uniformity in measurement
approaches across geographic regions along with increased quality control
efforts to help assure the validity of the aerometry data collected. Also,
in each case it was contemplated that such improved and uniformly obtained
aerometry data from throughout the two nations could serve as useful data
pools to be coupled with welfare effects studies and community health
epidemiology studies aimed at both (1) improving knowledge of quantitative
air pollution/effects relationships and (2) demonstrating the benefits to
*(continued)
"Brown Committee Report" or the "Investigative Report" (IR), contained various
comments regarding sources of error in CHESS Program air quality and health
data and
quality control problems associated with the data collection and analysis.
The I.R. also contained various recommendations to be implemented by the
Administrator of EPA pursuant to Section 10 of the Environmental Research,
Development, and Demonstration Authorization Act of 1978 ("ERDDAA," P.L.
95-155, 91 Stat. 1257, November 8, 1977). ERDDAA also requires that EPA and
the Agency's Science Advisory Board report to Congress on the implementation
of the IR recommendations.
One recommendation of the IR was that an addendum to the sulfur oxides
monograph be published, to be used in part to qualify the usefulness of the
CHESS studies, and to apprise the public of the controversy surrounding
CHESS. An addendum has been published, and is available from EPA, as
announced in the Federal Register of April 2, 1980, 45 F.R. 21702. The addendum
isHncorporated by reference in this document in partial qualification of the
CHESS studies cited herein, and is part of the public file (or docket) established
for revision of this criteria document. The addendum contains the full text of
the IR, reports to Congress by EPA on its implementation of the IR recommendations,
and a report to Congress by EPA's Science Advisory Board on the same subject.
See also Appendix A of of Chapter 14 of the present draft criteria document.
1-15
-------
be accrued from implementation of the respective Clean Air Acts. In addition,
as will become apparent below, remarkably similar problems were encountered
and responded to with roughly comparable degrees of success in the course of
practical applications of air pollution measurement techniques undertaken to
achieve the above objectives.
In addition to the above developments in the United Kingdom and the
United States, many analogous steps have been taken by numerous other
industrialized countries over the past 30-40 years to cope with air pollution
problems. In that regard, again many parallels (and dissimilarities) in the
historical evolution of their air monitoring programs, epidemiologic research
efforts, and political/legal regulatory control activities could be noted in
comparison to developments in Britain and the United States. However, a
historical review of the evolution of such activities or analysis below of
results obtained with practical applications of sulfur oxides or particulate
matter measurement approaches or with community health studies outside
Britain and the United States is beyond the scope of present purposes.
1.3 SO AND PM AIR QUALITY ASPECTS
/\
Turning now to the major points addressed in the remainder of this
document, much of the key information contained in Volume II, on air quality
aspects, is summarized below under Subsections 1.3.1 to 1.3.7. Particular
emphasis is placed here on the critical assessment of practical applications
of air quality measurement techniques of crucial later importance in
evaluating community health studies. Of additional special interest below
are the discussions on transport and transformation phenomena and ambient air
<.
concentration and human exposure data.
1-16
-------
1.3.1 Chemistry and Analytical Methods
The chemical and physical properties of SO and PM determine their
A
behavior in the atmosphere and biologic toxicologic activity. Increasing
sophistication of analytical instrumentation and methodology has enhanced the
understanding of the relationship between atmospheric processes and sulfur
oxides and particulate matter.
Sulfur dioxide dissolves readily in water, such as atmospheric moisture,
to form sulfurous acid which oxidizes to sulfuric acid and subsequently to a
variety of particulate sulfates. Metallic oxides such as manganese or iron
catalyze the reaction. Particles may vary in size, shape, molecular composition,
and optical properties. Thus, analytical methods for particulate matter vary
according to the specific parameters measured.
Methods used to measure sulfur dioxide can be classified as integrated
or continuous. Most of the methods are based on techniques involving absorption
and stabilization on a substrate. The analysis of the collected sample is
commonly based on colorimetric, titrimetric, turbidimetric, gravimetric, and
ion chromatographic measurement principles. The most widely used integrated
method to determine atmospheric sulfur dioxide is an improved version of the
colorimetric method developed by West and Gaeke and adopted as the EPA reference
method in 1971. Sulfation methods, based on the reaction of SO with lead
J\
peroxide paste to form lead sulfate, have commonly been used to estimate
ambient SO concentration over extended time periods. Continuous methods for
/\
the measurement of ambient levels of sulfur dioxide have gained widespread
use in the air monitoring community. Continuous sulfur dioxide analyzers
«.
using the techniques of flame photometric detection, fluorescence, and second
-------
derivative spectrometry have been developed over the past 10 years and are
commercially available.
Methods for determining soluble sulfates, total sulfates, and specific
sulfate species involve the collection of participate matter and its subsequent
analysis by direct or indirect methods. For trace soluble sulfate determinations,
a commercial method based on ion exchange chromatography is specific, exceptionally
accurate, and sensitive. Methods of analysis for total sulfur include x-ray
fluorescence, electron spectroscopy, and flame photometry. X-ray fluorescence
methods are nondestructive and applicable to large numbers of ambient aerosol
samples. Procedures for determining specific sulfate species include thermal
volatilization and solvent extraction techniques, gas phase ammonia titration,
infrared and visible spectrometry, flame photometry, and electron microscopy.
Sulfate species may be estimated quantitatively by gas phase ammonia titration
methods and infrared spectroscopy.
The majority of atmospheric particle samples are collected by either
filtration or impaction devices. The collected samples are usually analyzed
gravimetrically to provide a direct measure of mass concentration. The most
sidely used method for gravimetric analysis of mass concentrations is the
high volume sampler, the current EPA reference method for particles. Other
methods include the British Smoke Shade and AISI tape samplers, multi-stage
cascade impactors, cy&lone samplers, and the dichotomous sampler. Chemical
analyses consist of manual wet-chemistry, atomic absorption, x-ray fluorescence,
optical emission spectrometry„ spark source mass spectrometry, neutron activation
analysis, and thin-layer chromatography with fluorescence detection. Methods
i.
for continuous or HI situ monitoring or atmospheric particles are also available.
They are not as closely related to mass concentration as are the integrated
1-18
-------
methods, but they provide useful information for studying particulate sources,
transport, episodes, and effects such as visibility. The integrating nephelometer
measures light scattering by atmospheric particles and is commonly used as an
indicator of visibility.
It should be noted that the most widely used particulate matter measurement
device in the United States, the high-volume sampler, usually has a relatively
high cut point, so that not only fine particles (^2.5 urn MMD) but also coarse
mode (2.5 to 15.0 urn) and large particles (mainly up to 30 urn MMD) are sampled.
In contrast, the other most frequently used method, the British smoke (BS)
instrument tends to sample a much more restricted range of particles sizes,
these typically being less than 5.0 ug/m (mainly <3.0 urn MMD) and chiefly in
the fine mode particle range.
1.3.2 Air Quality Measurement Applications
The critical assessment contained in Chapter 3 regarding practical
applications of measurement approaches employed in Great Britain and the
United States for determinations of air concentrations of suflur oxides (SO )
x\
and particulate matter (PM) is concisely summarized here. Information presented
concerns published evaluations of the relative specificity, sensitivity,
accuracy, precision, and reliability of the methods when used under optimum
conditions in the hands of technically-expert analysts. Much more emphasis,
however, is placed on evaluation of results obtained with practical applica-
tions of the measurement methods, often by less technically-skilled personnel.
That evaluation draws mainly upon published commentary on quality control
assessments for the different applications. Also, the major focus here is on
British and American air measurement approaches most widely used in acquiring
S02 and particulate matter data utilized in quantitative community health
studies discussed later.
1-19
-------
1.3.2.1 British Approaches
British S0p Measurements—As noted earlier, the lead dioxide gauge was used
extensively in Britain during the years prior to 1960. However, use of the
hydrogen peroxide method was gradually interspersed with the lead dioxide
gauge during the course of the 1950s, often being coupled in tandem, as it
were, with the apparatus for smoke measurements. Much of the early (1950s)
British epidemiology data discussed later has been related to SO,, measure-
ments obtained either by the hydrogen peroxide method, especially where 24-hr
S02 values are used, or the lead dioxide sulfation rate method in some
cases where long-term (days to a month) data were acquired.
In 1962, as part of the establishment of the British National Air
Pollution Survey, a working party was set up to compare the lead dioxide
gauge with the hydrogen peroxide method, which was then chosen as the standard
method for use in the Survey. As quoted in Atmospheric Pollution, 1958-1966
(WSL, 1967):
The hydrogen peroxide method is subject to the limitation that
its reaction is not confined to sulphur compounds; the lead dioxide
method has the limitation that the extent of the reaction can be
substantially influenced by weather conditions. Despite limitations,
both methods estimate pollution by sulphur compounds; the hydrogen
peroxide method is somewhat more complicated, but has the outstanding
advantage that it can measure concentrations of pollution over short
periods; the lead dioxide method is simple in operation, but it is
incapable of measuring concentrations over short periods.
Even so, it was considered desirable to compare the results
from the two types of instrument under controlled conditions. A
statistical analysis was made by Warren Spring Laboratory of results
from a group of 20 sites at which both lead dioxide and hydrogen
peroxide instruments had been operated over a period of 48 months.
The 20 sites selected were those with a reasonably complete set of
results from March 1957 to February 1961 at which the two instruments
were not more than 100 feet apart.
The correlation between 829 pairs of results from the 20 sites
over a period of four years was highly significant, showing that
both instruments were predominantly affected by the same pollutant,
sulphur dioxide.
1-20
-------
The WSL (1967) report presented a plot of the comparison data showing that
o
the ± 2o confidence limits correspond to ± 1.8 mg SO-/100 cm/day for a given
hydrogen peroxide reading and ± 0.18 mg S02/m3 for a given lead dioxide
reading.
In other words, estimates of SOp levels derived from lead dioxide sulfation
rate measurements, especially 24-hr estimates, can only be roughly compared
with SOp estimates obtained by the hydrogen peroxide method at other geographic
sites or at later times at the same location(s). Also, comparisons between
sulfation rate readings may only be meaningful when such readings differ by
o
the equivalent of about 180 ug/m of SOp. Some of the types and magnitudes
of errors encountered in the British application of lead dioxide gauges to
measure SOp levels are summarized in Table 1-1. As shown in Table 1-1,
several problems (e.g., humidity and temperature effects) result in the lead
dioxide method being essentially useless for 24-hr, measurements and in their
otherwise having a rather large (±180 ug/m 2a) error band associated with
them.
Based on some of the above problems, when the National Survey began in
1961 it was recognized that the lead dioxide method could not provide the
24-hour SOp measurements necessary for correlation with mortality and morbidity
effects investigated by epidemiology studies. The hydrogen peroxide method
for S09 was, therefore, adopted as being more valid than the old lead dioxide
gauge sulfation method. Because many of the staff making the measurements
would be the same people who had been servicing particle deposit gauges and
the lead candles without detailed technical knowledge of the analyses, however,
E.
an Instruction Manual (IM) issued by WSL in 1966 had to be quite detailed and
clearly readable by people with no training in analytical techniques.
1-21
-------
TABLE 1-1. SUMMARY OF EVALUATION OF SOURCES .^MAGNITUDES. AND DIRECTIONAL BIASES OF ERRORS
ASSOCIATED WITH BRITISH S0; WASUHWIlfS
Tl«*
•wrloel
evthod
Reported tevrc*
of error
Direction and eagnltcd* ef
reported error
likely general <*a>act en
Brltiih SOj date
Pro-19C1
1W1-1*00
(Brlttih Hot1«««1
Air P»l. Swvey)
194* •luldt Mualdlty (RM)
Te*»>*rature (T)
Wind speed (W)
(Overall erron)
Siting ef Step I* Line
Intake:
•ere»ld*
r\>
Reaction rat* Increaiet with RH.
((•action rale Incraatat 2S
per S* rite.
Reaction rat« IncrcaMt
• I Hi VS.
a. toe mar beller chlemeyi §4) - 100 MO/* ov*rMtlMt1ofl.
B. toe n»«r »«ff»Ut)or. Sffi - 70 p«rc«nt und*r«it1«t1on.
Saaclt Una Adtorptten: ,
a. Good cara & claantng 10 )>e/a> SO ufl/*
annuity Bean.
Tltratlon frror:
a. Normal-iharp color
change of Indicator at
pH 4 S
k. Cradual color ch*ng» of
Indicator at pH 45
C. Rounding off to 0.1 •!
ef alkali voluea
(vaewratlan of reagent:
Teaperature and Pretture:
a. Correctloni - norael
b. large a/ at filter
thly
n lig/a1 ondereatlMtla*) an 1M ef
•ua»er iee|>l*> In urftan areas.
unc>r.
Varlasle poattlv* »1*». molar hlfh win* ami.
CM to «•> U t UO iif/e1 (ID).
Occatienal (pro*, rare) aawtttv* (la*.
Occatlonal (prob. rare) negative blat.
»*»>1»1* fcneral 10 MS/"1 «*«atlf« kla
Occailonal 40-MS negative blai.
Likely rare SO-MS negative b1e>.
Negligible la
of data.
ct. Pr*
-SZ negative blai en nigh SO.-tS deys.
»S» potlttve bias en ton SO^'BS dayi.
Negligible le»*ct. r
of data.
Likely occn tonal 9-10
d tZX prwltlon
ative klm.
negligible li»>*ct. ,
Occatlonal 5-10 po/» negetlv* kin.
Occailonal negatlv* klat ef up te W
>» uc/a *«f. klM OB IIS of timur
~ia*pUi In urb*n are«i.
Occailonal nea. blei In eewntrv ar«a«-
«p t* 80 ug/B1 dally data & up U IMS
aonthly lean In tuner.
t %
aa «f d»U.
ActMl i 10 Mt/"3 rreclslm level.'
Added t S pf/to1 precltlm «rror.c
»-10G> pot. kill for SO, **U <1N .
7.S-1SX pot. klat for SO, of 100-700 »>«/•••
1.2S-7.SX pot. blat for ». of 700-400 uo/» .
<}.2SX pot. blai for SO. dlta >400 po/oi •
General 5S neg. blat In SO. iaU.
Occailonal - tlOS negative blai In SO.
*Data fro*) 196S-1968 eait clearly lapected.
bO«ta fro* 1966-1967 eait clearly lapacted.
£At MK In errer.
-------
As mentioned above, the lead peroxide method was selected because its
sensitivity, reliability and precision were demonstrated to be isuch better
than that obtained in comparison to the lead dioxide method. More
specifically, the British Standard for sulfur dioxide determination by the
hydrogen peroxide method states that replicate determinations can be expected
to be within ±20 ug/m for concentrations up to 500 mg/m and within ±4 percent
o /^T^yf7"/" V
for concentrations above 500 ug/m ; and an OECD Working Parting stated the
accuracy of the method to be ±10% at levels >100 ug/m . However, as summarized
in Table 1-1, numerous sources of errors have been encountered in the practical
application of the method in collecting data for the British National Survey
over the past 15-20 years.
Certain of the sources of error listed in Table 1-1, it can be seen,
resulted in relatively small errors, whereas others produced errors ranging up
to 50-100% in magnitude. Also, some errors appear to have been restricted to
affecting data from only limited locations (usually unspecified as to specific
names of localities) or during only limited time periods. Many of these types
of errors appear to have been detected fairly quickly and steps taken to
successfully correct or minimize them. Still other sources of errors exist
(e.g., those from reagent evaporation), which have likely affected essentially
all British National Survey SOp data. Some of these appear to remain uncorrected
to this date, in some cases more than 10 or 15 years after they were first
detected and brought to the attention of Warren Spring Laboratory officials
responsible for overseeing quality control for the entire National Air Pollution
Survey. See Chapter 3 for a more detailed discussion of each type of error.
Taking the above information into account for present purposes, it would
be extremely difficult to determine precisely which errors affected particular
1-23
-------
National Survey data sets employed in British epidemiology and other studies
discussed later in this document. That would likely require a thorough examina-
tion, on a time- and site-specific basis, of records detailing Information on
how each pertinent data set was collected and WSL quality control assessment
rep"orts for the data sets. Alternatively, in later evaluations of British
epidemiology studies one could accept the following overall evaluation and set
of conclusions by the WSL (1975) regarding British National Survey air pollution
data (emphases added):
The actual degree of accuracy attained in the Survey is not known.
Input data are scrutinized by WSL staff, and subjected to computer
checks, and any reflectances, titres, or air flows which are abnor-
mally high or low or show unusually abrupt changes from one day to
the next are queried and data known to be invalid are excluded from
the annual summary tables. Such checks can however eliminate only
some of the gross errors. More information will become available on
accuracy when current (1974) plans to institute additional quality
control, e.g., on reagent solutions, are put into operation.
However, although the accuracy of the Survey data cannot at present
be quantified, many of the errors discussed in the previous para-
graphs will cancel out when data are averaged over periods of a few
months or a year, or for groups of sites. The remainder tend to
show up as anomalies when data are compared with past or subsequent
data at the same site or with data from other sites; anomalies of
this kind have been commented upon throughout the Reports. Members
of Warren Spring Laboratory staff have devoted a large effort over
the years to site visiting and checking on procedures. It is their
experience that the vast majority of the instruments are maintained
and operated with reasonable care and accuracy. The Laboratory is
therefore confident that the accuracy is sufficient for the type of
data analyses carried out in the present series Of reports.
Presumably, it is the opinion of the WSL and British epidemiologists that the
accuracy of the survey data is also sufficient to meet the original objectives
of the Survey, ie. to assess the benefits accruing from the Clean Air Act of
1956, which requires use of the survey air quality data along with community
health endpoint evaluations in order to define quanititative air pollution/health
1-24
-------
effects relationships. That this presumption is likely correct is further
attested to by the long history of reliance on these data by British epidemio-
logists, such as in the making of statements regarding such quantitative relation-
ships in innumerable journal articles and reviews appearing during the past
twenty years, up to and including the very recent review by Holland et al. (1979).
Daily Smoke Measurements of the United Kingdom National Survey--The
general technique for the British Smoke shade (BS) measurement is described in
detail in Chapter 2, and a detailed critical assessment of the measurement
procedure is provided in Chapter 3 to allow for evaluation of the precision,
accuracy, and reliability of the measurements. Also, details of the BS
measurements are provided by an Instruction Manual (IM) issued by Warren
Spring Laboratory in 1966. At the start of the National Survey in 1961 (WSL,
1961) it was recognized: "The daily instrument, while comparatively simple in
design and operation gives reliable results jm good hands* and seemed the best
choice for the National Survey." WSL circulated the specifications of the
apparatus and methods to all the cooperating organizations as careful, uniform
work was essential if the results from the different sites throughout the
country were to be comparable. However, WSL found that detailed instructions
were necessary as most of the Local Authority staff making the measurements
had no training in analytical techniques. These methods were reviewed by an
O.E.C.D. Working Party and a report "Methods of Measuring Air Pollution"
(OECD, 1964) was prepared, which was accepted into the British Standards
Specification 1747, Parts 2 and 3. The Manual of Instruction (WSL, 1966)
incorporated the improvements in techniques, "but apparatus and procedures are
1-25
-------
specified in much greater detail to assist operation by observers with no
technical knowledge."*
Partly due to the lack of analytical training of survey monitoring site
operators, and other factors as well, various errors were encountered in
carrying out BS measurements for the National Survey.
Table 1-? summarizes information discussed in Chapter 3 on the sources,
magnitudes and directional biases of errors associated with British smoke
measurements during the past 30 to 40 years. For example, prior to 1961, the
use of weights for sealing purposes led to highly variable errors in BS
measurements due to leakage at filter clamps, and steps were taken to require
screw-down clamps as standard procedure as part of the later British National
Survey work implemented after 1961. It is not clear to what extent any specific
British BS data sets from the 1950s may have been affected by the clamp leakage
problem, but one must assume that such errors could not have often been very
large or serious and that the WSL took appropriate steps to eliminate or
invalidate any data in gross error as they were detected via their quality
control efforts in the late 1950s. Analogously, there is evidence that WSL
did take steps to inform users of pre-1961 BS data of errors arising from (1)
comparing reflectance on filters to photographs of painted stains and (2) use
of reflectance readings below 25 percent, where the stain was too dark to use
the Clark-Owens DSIR curve. However, it also appears that only a few investigators
(e.g., Commins and Waller, 1970) took steps to go back and correct published
reports based on the affected pre-1961 data and to publish revised analyses
taking into account corrections for the pre-1961 data errors.
'Underline added for present emphasis.
1-26
-------
TABLE 1-JF. SUMMARY OF EVALUATION OF SOURCES. MAGNITUDES, AND DIRECTIONAL BIASES OF ERRORS
t
ASSOCIATED WITH AMERICAN S02 MEASUREMENTS
Time
period
Measurement r
method
Reported source
of error
Direction and Magnitude
of reported error
Likely general Impact on American SO^ data
1944-1968
Lead dioxide.
1969-1975
(EPA CHESS
PROGRAM)
West-Gaeke
Pararosanallne.
Humidity (RH).
Temperature (T).
Wlndspeed (WS).
Saturation of Reagent
(sulfatlon plate mainly).
(Overall Errors).
Spillage of reagent
during shipment.
Tine delay for reagent-
SO- complex.
Concentration dependence
of sampling method.
Low flow correction.
Bubbler train leakage.
(Overall errors).
Reaction rate Increases with RH.
Reaction rate Increases 2X per 5°
rise.
Reaction rate Increases with WS.
Variable underestimation beyond
pt. where 15X of PbO. on plate
reacted.
18% of total volume SOX of time;
occasional total loss
SO, losses of 1.0, 5, 25. and
75X at 20. 30, 40, and 50°C,
respectively.
Underestimation of unspecified
magnitude at dally SO. >200
ug/m3. Z
±10X to SOX variable error.
Small underestimation error of
unspecified magnitude.
"November, 1970. to April, 1973. CHESS Program data Impacted before error corrected.
Applies to CHESS Program SO, data from all years 1970-1975
cAs summarized by Congressional Investigative Report (IR, 1976).
Variable positive bias, especially In s
Variable positive bias, especially in s
Variable positive bias, especially in summer.
Possible large negative bias, especially for 30-
day samples for summer monthly readings.
Generally wide ± error band associated with data.
Possible negative bias up to >100X. mainly in
summer, with 30-day reading.
Half of SO
mean of
0, data likely negatively biased by
17X; some up to 100X.
Usually small (<5X) negative bias, but consistent
negativebsummer bias up to 25X at 40°C temp,
extreme.
Probable general negative blas.ln daily,
monthly, and yearly S02 data.
Usually error of < 110X; occasionally up to
t SOX in daily, but dampened statistically In
annual mean.
Slight negative bias suspected.
From Nov., 1970, to Dec., 1971, data biased
low by 50-100X. From Nov.. 1971. to
conclusion of CHESS Prograa In 1975, fall-
winter data appear valid but summer data biased
low by maximum of 60-80X. From 1972 to1975
annual average data approximately 15-2.0X low.
Dally data highly random, not useful.
-------
Probably of much greater concern than the pre-1961 BS measurement errors
are those encountered after the establishment and initial implementation of
the British National Survey in 1961. These include certain errors, e.g., the
"computer error of 1961-1964," (see Figure 1-1), which were eventually detected
by WSL and resulted in steps being taken to correct affected BS data in
National Survey data banks. It is clear, however, that whereas users of the
affected data may have been informed of such errors by WSL, virtually none of
them have taken steps to (1) alert recipients of publications containing
analyses based on the affected data of the likely inaccuracies or ranges of
error involved; (2) to reanalyze the study results based on the affected data
sets; or (3) to reissue or publish anew any revised analyses. In fact, even
some Warren Spring Laboratory quality control literature prepared and published
during the 1960s or 1970s and still in use may contain incorrect information
and recommended standard procedures for BS measurements based on analyses
"contaminated" by computer errors or other problems summarized in Table 1-2
and discussed in more detail in Chapter 3. One such example of this relates
to ambiguities in the use of certain correction factors in calculating BS
values, which have been questioned by Ellison (1968).
In regard to determining which British BS data sets and related epide-
miology studies are affected by different post-1961 National Survey errors,
it is again presently very difficult, as was the case with British SO- measure-
ments, to specify with any confidence the nature and magnitude of specific
errors impacting particular studies. This would probably require thorough
examination of records and WSL quality control reports concerning each of the
pertinent data sets. On the other hand one can project that certain data
sets and British epidemiology studies were almost certainly affected by some
subset of BS measurement errors and these are taken into account in evaluating
1-28
-------
500
A - BRITISH STANDARD CURVE
B - DSIR INTERIM CURVE
C — DSIR -CLARK • OWENS CURVE
O- 1961 TO 19&4 NATIONAL SURVEY CURVE
DARKNESS INDEX
Figure 1-1.
Comparison of smoke calibration curves for Eel reflactometer,
Whatman No. 1 paper and a 1-in diameter filter. From WSL (1967)
The computer followed curve D during 1961-64 instead of the
correct curve(s) B and C.
1-29
-------
such studies later this Chapter. For example, published reports of the
"Ministry of Pensions" (1965) and Douglas and Waller (1966) studies contain
specific reference to usage of National Survey data from the 1961-64 period
and, therefore, the results of those studies should be reevaluated in light
of measurement errors reported by the WSL for that period.
1.3.2.2 American Approaches
American S00 Measurements—Turning to American measurement approaches, different
types of measurement methods for a given pollutant were adopted by various
local, state, and federal agencies in establishing or expanding air quality
monitoring systems that proliferated across the United States during the
1950s and 1960s. Rather than discuss methods used for S0? measurements by
all of the different American air monitoring systems, main emphasis is placed
here on the discussion of only certain key American applications of measurement
methods for SO that are of crucial importance for later discussions of
quantitative relationships between health effects and atmospheric levels of
sulfur dioxide. These include mainly applications of S0_ measurement methods
as employed in the EPA "CHESS Program" as the single largest attempt to
define quantitative relationships between air pollution and health effects.
In regard to sulfur oxides measurement approaches used in the United States,
lead dioxide or other "sulfation rate" measurement methods were, as in Britain,
widely employed prior to the early 1960s for assessing S0_ air levels.
However, probably to a somewhat greater extent than in Britain, sulfation
rate measurement techniques continued to be used later into the mid or late
1960s by some monitoring programs in the United States or in connection with
certain community health epidemiology studies, as discussed later in this
1-30
-------
chapter. As shortcomings of the "sulfation" methods became more widely
recognized, however, their use was generally abandoned and more specific
methods for the measurement of SO- or other sulfur oxide compounds were
adopted, as was done in Britain. The hydrogen peroxide acidimetric method
(see OECD, 1965) selected for use in the British National Air Pollution
Survey, however, was not very widely adopted in the United States for S02
measurements. Rather, versions of West-Gaeke (1956) colorimetric procedures
were much more widely used in the USA. Conductivity measurements for SO-
(Adams et al., 1971), based on an acidimetric method adaptation often used in
automatic instruments and most suitable for measuring periods of around 24
hours, later began to be applied in the operation of some American air monitor-
ing networks in the 1970s.
The West-Gaeke method was the method mainly employed in the EPA "CHESS
Program" for determining SO- air levels for inclusion in analyses of community
health end point data in "CHESS" epidemiology studies. The application of
that method in the CHESS Program was accordingly most thoroughly discussed in
Chapter 3. The types of errors in measurement associated with CHESS SO- data
*j
are summarized in Table l-^Sy, along with notation of some factors affecting
earlier sulfation methods. Much of the information on the former subject is
derived from a 1976 Congressional Investigative Report (IR) which contained a
thorough evaluation of EPA CHESS Program air quality measurements and other
aspects of the Program.
Looking at the types of errors associated with earlier American use of
sulfation rate lead dioxide methods, similar effects of temperature, humidity,
etc., as affected analogous British S02 methods are seen to apply here to
American data as well.
1-31
-------
TABLE 1-3. SUMMARY OF EVALUATION OF SOURCES, MAGNITUDES, AND DIRECTIONAL BIASES OF
ERRORS ASSOCIATED WITH BRITISH SMOKE (PARTICULATE) MEASUREMENTS
Tim
pa Hod
Measurement
method
Reported source
of error
Direction and magnitude
of reported error
Likely general impact
on published 6S data
1944-1950*
Pre-1961
Smoke filter
co
ro
1961-1964
1964-1980
Leakage st clasp.
Weights used to make the
seal.
Highly variable under-
estimation of BS levels.
Depending upon observer
and value of 8.
Comparing reflectance to
photographs of painted
standard stains.
Reflectance (R) below 25X, 50-100% underestimation.
stain too dark with use
of Clark-Owens DSIR curve.
Computer not following <80S underestimation at low
proper calibration curve. R if not corrected by WSL
(See Moulds,1961) and
discussion of clamp size
correction factor.
Clamp correction factor
for other than 1-inch
clasp.
Flow rate - normal 1 day.
Flow by 8-port with 1
reading per week.
Variability of reading
reflectance.
Averaging reflectance
instead.of averaging
mass/cm .
Use of coarse side of
filler facing upstrean.
Uncertain; derivation
cannot be verified.
Possible +20X.
»3X variation.
-10% underestimation.
+10X overestination.
+2 units of R
Highly variable under-
estimation due to non-
linearity of R.
6-15X underestimation.
Probable widespread highly
variable negative bias.
Probable widespread relatively
smalI negative bias.
Occasional 50-100% negative
bias in some data sets.
Negligible for BS <~1QO pg/«3.
Increasing negative bias up to BOX
as BS values increase over 100
Possible underestimate for 2-inch
and 4- inch clamps
Possible overestimate for 1/2-Inch
and 10 cm clamps.
Presumed t 3X precision level.
10X negative bias on high BS days.
10X positive bias on low 85 days.
Error increases with BS level
at SO ug/M up to 120X at 400 ug/m
Probable snail negative bias at low
BS levels, could be large at high BS
Occasional negative bias of 6-1SX.
-------
TABLE 1-3. (continued)
Time
period
Measurement
method
Reported source
of error
Direction and magnitude
of reported error
Likely general Impact
on published BS data
Reading of wrong side of
stained filter.
Leakage at filter clamp
a. Normal, with good care
b. With inadequate care.
c. Careless loading where
uneven stains are
produced.
Use of'wrong clamp size
a. Stain too light R>90X.
b. Stain too dark R<25%.
SO-75X underestimation.
1-2% underestimation.
2-BX underestimation.
10-20% underestimation.
Highly variable over-
estimation.
Highly variable under-
estimation.
Occasional negative bias
of 50-75X.
General 1-2X negative bias.
Occasional 2-OX negative bias.
Occasional 10-20% negative bias.
Data usage not recommended.
Data usage not recommended.
t
oo
CO
-------
Turning to American applications of SOp measurements since the wide-
spread abandonment of sulfur dioxide sulfation rate methods in the mid to
late 1960s, several different types of errors were identified as being associated
with EPA CHESS Program S02 measurements via a thorough evaluation of the
CHESS Program, as reported in the IR (1976). As can be seen, the magnitudes
of some errors in CHESS SO^ measurements spanned about the same range as
those seen for British National Survey S02 measurements and, at times, derived
from analogous sources of error, e.g., evaporation or other loss of reagents.
In the case of the American CHESS Program data, however, the specific overall
impact of the various detected errors on particular CHESS data sets appears
to have been more definitively defined by the work of the IR (1976); more
specifically, it appears that the CHESS data generally tended to be somewhat
negatively biased in comparison to other local or state SOp data from monitoring
sites proximal to the CHESS sites, with the local and state data judged by
the IR (1976) to be reasonably accurate and reliable. The specific magnitude
of the negative bias for particular years of CHESS data is summarized in
Table 14-3, and appears to have been around 30-40% in some circumstances and
up to around 100% in other cases.
American High-Volume TSP Sampling Measurements—As discussed earlier, the
hi-volume TSP sampler, since its development in the early 1950s, has been the
instrument most commonly used in United States for measurement of atmospheric
particulate matter; and high-volume TSP readings have most typically been
used in American epidemiology study evaluations of associated air pollution-health
effects relationships. In contrast, other particulate matter measurement
approaches (e.g., the coefficient of haze method) saw only relatively limited
application during the 1950s and early 1960s in certain American locations
1-34
-------
and were infrequently used in estimating quantitative relationships between
airborne particulate matter and health or welfare effects. Accordingly,
major emphasis is placed below on the critical appraisal of certain key
appjications of hi-volume TSP measurements in the United States. As before,
in discussing American applications for measurement of oxides of sulfur, the
present summarization focuses most heavily on evaluation of applications of
TSP measurement methods employed as part of the EPA "CHESS Program," as the
single most extensive and comprehensive use of such methods as part of American
community health epidemiology studies. Much of the information is derived
from the 1976 Congressional Investigative Report (IR), which included a
thorough analysis of EPA CHESS Program TSP measurements and comments regarding
certain local or state TSP measurements.
The main sources, directions and magnitudes of errors identified as
possibly affecting American TSP measurements are summarized in Table 1-4. In
addition to various sources of minor errors inherent to the basic TSP sampling
method, certain other nuances of procedures included in the Federal Reference
Method (40 CFR 50, Appendix B) may have resulted in the introduction of an
additional slight negative bias in TSP data obtained by American researchers.
This, more specifically, pertains to the manner in which flow rate calculations
are made upon which final TSP concentration determinations are based.
The Federal Reference procedure calls for the averaging of the initial
and final recorded airflow rates. However, as described in Appendix 3-A of
Chapter 3, the uncontrolled flow rate drops more rapidly at the start of the
run than at the end of the run. Therefore, a linear approximation leads to an
t.
overestimate of the flow rate, which will reduce the measured value. Consequently,
1-35
-------
TABLE 1-4. SUMMARY OF EVALUATION OF SOURCES, MAGNITUDES, AND DIRECTIONAL BIASES OF ERRORS
ASSOCIATED WITH AMERICAN TOTAL SUSPENDED PARTICULATE (TSP) MEASUREMENTS
Tine
period
Measurement
method
Reported source
of error
Direction and magnitude
of reported error
Likely general Impact
on published BS data
1954-1980 High-Volume TSP
Time Off (Due to power
failure).
Weighing error.
Flow measurement (with
control).
Flow measurement (without
control)
a. Constant TSP—Average
of flows.
1. Low TSP level.
2. High TSP level.
b. Rising TSP-Average of
flows.
c. Falling TSP-Average of
flows.
Aerosol evaporation on
standing.
Condensation of water vapor.
Foreign bodies on filter
(Insects).
Windblown dust into filter
during off-mode.
Wind speed effect on pene-
tration of dust Into the
Hi-Vol shelter.
Wind direction effect due to
H1-Vol Asymmetry
Artifact formation, NO,
sor. 3
Variable underestimation.
±2% random variation.
±2% random variation.
2% underestimation.
5-10% underestimation.
10-20% underestimation
10-20% overestimatlon.
1-2% underestimation.
5% overestimatlon.
Generally small over-
estimation.
Generally small over-
estimation.
Less penetration at high
windspeed.
Higher penetration when
normal to sides.
5-10 ug/m overestimate.
Negligible impact, rare negative bias.
Negligible Impact.
Negligible Impact.
Negligible Impact.
Possible 510% negative bias.
Possible 10-20% negative bias.
Possible 10-20% possible bias.
Probable negligible Impact.
Possible 5% positive bias.
Possible 5% positive bias.
Occasional (rare) positive bias.
Occasional (rare) negative bias.
Probable Increase 1n random (±) error.
Occasional positive bias.
-------
TABLE 1-4.(continued).
Time
period
Measurement
method
Reported source
of error
Direction and magnitude
of reported error
Likely general Impact
on published BS data
1969-1975
(EPA CHESS
Program).
Fed. Reference
Method Standard
HI-Vol Sampler
Loss of sampling material
1n field.
Loss of sampling material
In mailing.
Evaporation of organic sub-
stances.
Wlndflow velocity and
asymmetry.
(Overall errors).
No specific estimate of
magnitude of error; but
would be underestimation.
Reported 4-25X apparent
loss; max. likely due to
crustal (sand, etc.)
fall-off from selected
Utah sampling sites.
No specific estimate of error
magnitude, but not likely to
exceed 5% underestimation.
No specific estimate of error
magnitude; but most likely to
increase random variation of
small underestimation.
Probable slight negative bias
1n Utah winter data. No known impact
on other CHESS TSP data.
Probable general small <10X negative bias;
occasional 25% negative bias.
Probable slight negative bias
of <5X for TSP data from urban/
industrial areas.
Negligible Impact or slight
negative bias.
Generally <10% negative bias;
occasional 10 to 30% negative bias.
-------
all TSP data computed in this manner have a slight negative bias which is
likely usually of the order of 5 percent; on occassion, however, under
circumstances where the flow rate may have fallen below 40 ft /min, larger
errors (up to approximately 15 percent) may have been introduced. Assuming
that monitoring site operators in the United States adhere to the recommended
Federal Reference Method procedures, then this type of bias is likely
inherent in essentially all American TSP data collected without flow rate
control or recording. Despite such problems, it can be seen that the maximum
range of uncertainty derived from the various errors associated with American
TSP measurements is generally less than 20 percent in either a positive or
negative direction on a random (±) basis.
Errors in addition to general TSP measurement errors reported by the
1976 Congressional Committee Investigative Report (IR, 1976) to affect CHESS
Program TSP measurements during 1969-1975 are broken out and listed
seperately in Table 1-4. Some of those errors (e.g., loss of sample
materials in filter removal from the field monitoring apparatus) were
reported by the IR (1976) as likely affecting only very restricted CHESS data
sets. Others, e.g., errors due to loss of sample in mailing, appear to have
been more widespread and presumably impacted on many CHESS data sets. It is
interesting to note, however, that the IR (1976) concluded that the net
effect of all of the errors was to introduce, in general, a slight negative
bias of 10 to 30 percent into CHESS TSP data, which is not much beyond the
range of different types of errors (e.g., linear flow corrections) more
generally associated with American applications of TSP measurements. Section
IV C 3 of the IR (1976) further concluded that:
"...the TSP data were by far the best quality data taken in the
CHESS monitoring program. Differences measured between High and Low
sites are probably reasonable estimates of the differences of TSP
exposures as received by populations in these areas."
1-38
-------
It appears reasonable to concur with the IR (1976) and, jccordingly, to
accept CHESS TSP measurements as reasonable estimates of TSP exposures of
CHESS Program community health study populations, taking into account that
such data may be biased low by no more than 10 or,^r£ mo^fe-, 30 percent.
1.3.2.3 BS-TSP Comparison Studies—In Chapter 3, information is reviewed on
BS-TSP comparisons derived from a number of studies published over the past
twenty years and reporting results obtained from the sampling of air in many
disparate geographic areas (Britain, other European countries, and the United
States) and varying time periods (from the early 1950s to the mid 1970s). It
was earlier hinted at and is now clear from a present perspective, with the
advantage of viewing the various BS-TSP comparison data sets together in
relation to each other, that a nonlinear relationship exists between BS and
TSP measurements obtained with colocated samplers. That is, regardless of
where or when such readings were obtained, TSP values were usually found to
be two or more times higher than corresponding BS readings up to BS levels of
about 100 |jg/m . At higher levels, however, the TSP and BS readings tend to
converge toward each other, such that TSP/BS approaches unity at BS levels of
3 3
500 jjg/m or more. Thus, above 500 pg/m or so, BS and TSP readings from
colocated samplers are essentially identical.
In carrying out analyses of BS-TSP comparison data sets, various investiga-
tors in the past generally employed linear regression analyses in an effort
to define straight lines that best fit data points obtained by them over a
limited range of BS-TSP values (see Figure 1-2). They also often extrapolated
the thusly defined straight lines to BS-TSP values beyond the range of their
emperical observations and found inconsistencies between the BS-TSP relationships
implied by their line(s) and those defined by linear analyses of different
1-39
-------
500
400
s
uT
§!
V) \
tfi I
! 300 —
200
100
BS - TSP COMPARISON STUDIES
COMMINS-'WALLER!
(LONDON, 1955-63)
O LEE, ET AL (LONDON, 1970)_
OBALL& HUMEj
(LONDON, SUMMER 1975)
1001 200 300 400 500
AMERICAN HI-VOL TSP. Jig/m3,
600
700
Figure 1-2.
Representative examples of BS/TSP relationships defined by linear
regression analyses employed to fit BS/TSP comparison data points
as described in published reports pum«
1-40
-------
BS-TSP comparison data sets obtained at other times or locations. Such
apparent inconsistencies between BS-TSP relationships, arising from linear
regression analyses of various data sets obtained over limited and often
different ranges of BS-TSP values, have contributed to and reinforced the
view that no consistent relationships exist between BS and TSP measurements
obtained at different locations or even at different times at the same location.
Paradoxically, this view has gained widespread credence despite the concurrent
realization that corresponding BS-TSP readings are nonlinearly related.
Recognition of the acknowledged nonlinearity of BS-TSP relationships and
the necessity to meet certain boundary conditions defined by emperical observations
and certain theoretical considerations as sine-qua-non starting points led to
the formation of a "bounded nonlinear model" (BNLM) proposed in Chapter 3 as
a unifying concept or means by which to interconvert monthly or annual average
BS and TSP values obtained under vastly different circumstances. The BNLM
(see Figure 1-3) essentially employs a power function defining a nonlinear
relationship that meets the boundary conditions of (1) BS -* 0, where TSP •* 0
and (2) BS -> TSP, as BS -»• °°; or, in other words, when there is no particulate
matter in the air, both TSP and BS readings must be zero or tend toward zero
and, also, they must tend to converge toward each other as BS values become
very large as observed in empirical BS-TSP comparison studies. In addition,
the particular BNLM model chosen defines a curve which fits other empirically-
derived observations to the effect that TSP/BS = 2 at 100 ug/m3 BS and
TSP/BS = 4/3 at 250 ug/m3 BS. Plotting of corresponding BS-TSP values from
numerous published BS-TSP comparison studies reveals that the BNLM model fits
well virtually all presently available BS-TSP comparison data, in the mean.
1-41
-------
it
O
c
D
O HOLLAND «t (I (1979) • LONDON
A COMMINS and WALLER (1967) - LONDON
D LEE n tl (19721 ENGLAND
^ FERRIS tt «l (19731 - BERLIN. NH
O-DALAGER (1975] • ALBORG. DK
9DALAGER(1975) - K0BENHAVN
•OBALL »nd HUME (1977) • LONDON
VKRETZSCHMAR (1975) ANTWERP
AMERICAN HHVOLTSP.
Figure 1-3.
Measurements of British Smoke vs Hi-vol TSP, showing a consistent
relation between these measures over the entire range of reported
observations. Most points shown are annual mean values; see text for
discussion.
1-42
-------
Only the reported results from one recently obtained data set comparing
BS and TSP values from colocated samplers at 16 locations in the United States
appear, at first sight, to be greatly inconsistent with the BNLM model (see
AISI line in Figure 1-3). Closer inspection of the study, however, reveals
that numerous methodological errors were made in conducting the study, including
the failure to follow published standard procedures employed in the collection
of BS data earlier used in British epidemiology studies on the health effects
of particulate matter and, also, in most other BS-TSP comparison studies.
Nevertheless, when the particular methodological errors and other deficiencies
are evaluated, it becomes apparent that some of the basic observations reported
may not be as inconsistent with other published results and the BNLM model as
initially seems to be the case. See Chapter 3 for a more detailed discussion
of this matter.
1.3.3 Sources and Emissions
Sulfur oxides and particulate matter are emitted into the atmosphere from
a variety of sources, both natural and man-made. Natural particulate matter
emissions include terrestrial dust (windblown soil and rock particles), radioactive
particles, sea spray, volcanoes, biosphere emanations (products of biological
processes), and biomass burning. Most natural particle mass is greater than 1 urn
in diameter. On a global scale, natural emissions of sulfur oxides and particulate
matter into the atmosphere greatly exceed man-made emissions. The latter,
however, tend to be much more concentrated on a local or regional scale.
In the United States, anthropogenic air pollutants, that is, those pollutants
associated with human activities, are primarily caused by the combustion of
fossil fuels and the smelting of metals. The marked increase in energy consumption
since the middle of the twentieth century has led to a generally upward trend
in emissions of particulate matter and sulfur oxides.
1-43
-------
Projections indicate that secondary PM will account for an increasingly
greater proportion of the total emissions over the next decade. (Secondary
particles are formed by atmospheric conversion of gases.) Most of this increase
wiJl occur in the fine fraction (less that 2-3 urn), and its impact will be
felt over increasingly greater geographic areas. For example, the use of
higher smoke stacks will result in a longer range transport of SO^- This
long-range transport creates more secondary fine particulate matter and greater
geographical distribution of particulate pollutants.
Man-made emissions of sulfur oxides from stationary sources in the
United States are currently estimated at 29 million metric tons per year.
Transportation sources contribute less than 1 million metric tons of sulfur
oxides, or not quite 3 percent of the total emissions. Of all stationary
sources, utility plants burning coal and oil are the primary contributor
(about 62 percent), while other industrial processes account for about 32
percent, and residential and commercial use of coal and oil accounts for less
than 3 percent. Emissions vary considerably from one region to another; over
75 percent of the total national emissions come from the eastern half of the
United States. Annual emissions of sulfur oxides have increased from about 23
million metric tons in 1940 to as high as 35 million metric tons in 1973.
Although the past few years have seen a slight decline in emissions, an increase
to nearly 40 million metric tons is projected by 1990 unless reduced by implementatio
of air regulations.
Over 90 percent of the national sulfur oxide emissions occur in the form
of sulfur dioxide; the balance consists of sulfates in various forms. The
quantity and composition of these emissions vary from source to source and
depend on factors such as fuel characteristics, operating conditions, and
emissions-control equipment.
1-44
-------
The characteristics of participate matter, like those of sulfur oxides,
vary according to source, emissions-control equipment, and other factors.
Man-made emissions of particulate matter amounted to about 12.5 million metric
tons in 1975 in the United States. The distribution of sources varies geographi-
cally: the eastern half of the United States accounts for 66 percent of the
national total. Forty percent of the emissions come from stationary fuel
combustion sources, principally in electric generation and industry and 25
percent from mineral processing. Additional emissions come from primary metal
production, land vehicles, food and agricultural processes, solid waste disposal,
and other sources. In addition, there are considerable emissions of fugitive
particulate matter from both industrial (3.3 million metric tons/yr) and
non-industrial (4.9 million tons/yr) sources.
1.3.4 Environmental Concentrations and Exposure
To protect the public health from air pollution, it is important to know
how many people are likely at risk from pollutant concentrations capable of
inducing adverse effects. For financial and technical reasons it is impractical
to obtain direct measurements of pollutant doses incurred by individuals. It
is necessary, therefore, to rely on fixed-point air monitoring to estimate
pollutant exposures in representative environments. Such monitoring is conducted
predominantly outdoors—in the ambient air, within the first 100 feet above
the surface. A meaningful evaluation of population exposure to sulfur oxides
and particulate matter must include measurements of specific local and area-wide
concentrations, data on particle size and composition, and knowledge of the
number of people in specific areas and their activity patterns.
Fixed-point ambient air monitoring conducted at several thousand sites in
the United States over the last three decades shows a general trend of declining
pollutant concentrations. For example, in the late 1950s, monitors in five
1-45
-------
heavily industrialized cities were recording annual arithmetic mean concentrations
of total suspended particles in the range of 130 to 195 ug/m . By the early
1970s, those sites were recording annual means in the range of 70 to 115
ug/m3. By 1978, only 17 percent of reporting stations in the nation were
3
exceeding annual average particulate levels of 75 ug/m and six percent were
exceeding the levels of 260 ug/m .
Studies of urban aerosols indicate that while the mass balance of total
TSP is similar from city to city, its chemical composition varies. Many of
these differences arise from the type of fossil fuel used and the type of
industries dominant. For example, soot produced from the combustion of fuel
oil can vary from 1 to 13 percent of the TSP mass in a particular area. Coal
soot, on the other hand, may account for up to 30 percent of the TSP mass in
some urban areas.
Sulfur dioxide concentrations have also decreased. In 1964, a represent-
ative group of 32 SO- monitors across the nation averaged about 50 ug/m ,
with the peak station reporting an annual average of some 200 ug/m . By 1971,
their composite average had decreased to about 25 ug/m , and the peak station
3
reported less than 50 ug/m . In 1978, only one percent of reporting stations
o
were exceeding the primary annual standard for SO- (80 ug/m ); and two percent
were exceeding the primary 24-hour standard (365 ug/m ). Sulfur dioxide
levels in the ambient air generally rise and fall according to the amount of
coal or other fossil fuels burned other than natural gas.
Analyses of some specific constituents of air pollution, notably sulfates,
nitrates, and fine particles (<2.5 ug/m MMD) indicate that these constituents
are not declining at the same rate as the gross particulate concentrations.
Sulfate, ammonium, and nitrate ions dominate the fine particle fractions
in urban and rural areas. Generally, toxic elements such as arsenic and lead
1-46
-------
are associated with the fine fractions, while less toxic elements such as cal-
cium and iron occur more often in the coarse fraction.
At least 50 percent of atmospheric particulate sulfur occurs in the fine
fraction as sulfates and often accounts for 40 percent or more of the fine
fraction mass. Most of the sulfate aerosols occur as ammonium salts rather
than as sulfuric acid. High sulfate levels occur more often in summer than in
winter.
There are regional differences in trace metal concentrations. Vanadium
and nickel, for example, are correlated with the use of fuel oil. Their con-
centrations are, therefore, highest during the winter months. But in areas
requiring low sulfur fuels, vanadium and nickel levels are quite low, since
these elements are removed during fuel desulfurization processes. Where coal
is used, titanium tends to exist at somewhat higher concentrations.
1.3.5 Transmission Through the Atmosphere
Pollutants emitted into the atmosphere are transported vertically and
horizontally, transformed physically and chemically, and deposited by dry and
wet removal mechanisms. Since each of these processes is a function of numerous
physicochemical and meteorological variables, source-receptor relationships
are necessarily complex. Despite the difficulty in analyzing atmospheric
transmission, certain findings have been substantiated.
Atmospheric particulate mass is distributed bimodally in relation to
particle size: fine particles are smaller than 2.5 micrometers (|jm) and
coarse particles are larger than 2.5 urn. Comparable masses of fine and coarse
particles have been measured within urban areas. Outside of urban areas, the
fine particulate mass tends to exceed the coarse particulate mass, especially
in the Eastern United States.
1-47
-------
Many fine particles are formed in the atmosphere from precursor gases.
Most coarse particles are emitted directly from combustion or industrial pro-
cesses, or are of natural origin, and become suspended in the atmosphere
through human activity.
A portion of the fine particulate mass is the result of the atmospheric
conversion of sulfur dioxide, nitrogen oxides, and higher molecular weight
hydrocarbons (Cfi+) to particulate sulfate, nitrate, and organic aerosols,
respectively. The remainder of the fine particulate mass consists of com-
bustion-derived sulfates and carbonaceous particles, lead, small amounts of
other metal-containing particles from combustion and industrial processes, and
small amounts of finely divided dusts. The coarse particulate mass consists
of substances emitted directly from industrial sources and from combustion
processes, along with substantial amounts of suspended or resuspended dusts of
various types.
Because their deposition occurs over several days, fine particles can be
transported for distances up to 1000 kilometers or more from their origins or
the origins of their precursor gases. Finely divided particles formed con-
tinuously by secondary atmospheric reactions are especially likely to undergo
long-range transport and to exist at appreciable concentration levels. Since
urban-area sources and large point sources contribute precursors of secondary
particles, concentrations of such particles may become superimposed upon those
of fine particles during transport.
Gas-to-particle conversion processes depend on a variety of environmental
parameters, such as solar radiation, concentrations of oxidizing radicals, and
humidity. Therefore, precursor gas emission rates alone are of very limited
utility in estimating the mass of secondarily formed fine particles.
1-48
-------
Although it has been known that many factors contribute to sulfate levels
in the atmosphere, past results suggest that heterogeneous reactions may play
an important role in determining those levels. More detailed studies are
necessary to assess the role of these reactions. Past work has shown the
importance of temperature, oxygen, water, and other parameters in removal
rates.
1.4 WELFARE EFFECTS ASPECTS
The third volume of this document discusses welfare effects of sulfur
oxides and particulate matter. This includes information summarized below on:
effects on vegetation (Section 1.4.1); acidic precipitation formation and
effects (1.4.2); effects on visibility and climate (1.4.3); and materials
damage effects (1.4.4). Major emphasis is placed in the discussion of lengthy
materials contained in Chapters 7 and 8 on vegetation effects and acidic
precipitation.
1.4.1 Effects on Vegetation
Terrestrial vegetation is normally exposed to a variety of substances
from the atmosphere. Among these substances are sulfur oxides, particulate
matter, and other phytotoxic pollutants. More is known concerning the effects
of sulfur oxides, mainly sulfur dioxide, than about the effects of particulate
matter.
The sensitivity to sulfur dioxide of plant species and varieties differs
because of the genetic composition of the plants and the influence of environ-
mental conditions on their response. Temperature, light, humidity, other air
pollutants, soil conditions, and the stage of plant growth all interact in
affecting the sensitivity of plants to injury from sulfur dioxide and particulate
matter. Since the ambient air contains many substances other than sulfur
dioxide and particulate matter, interaction with these other substances must
be considered when analyzing the effects of sulfur oxides or particulate
1-49
-------
matter on vegetation. Mixtures of the substances in the ambient air, some of
which may be nitrogen oxides or ozone, may have adverse growth or foliar
effects of a greater magnitude than the effects from sulfur dioxide alone.
i
- The response of a given variety or species of plants to a specific air
pollutant cannot be predetermined on the basis of the known response of related
plants to the same pollutant. Neither can the response be predetermined by a
given response of a plant to similar doses of different pollutants. The
interplay of genetic sensitivity and environmental influences must be considered
for each plant and pollutant.
In general, sulfur dioxide must enter a plant to induce toxic effects,
although some toxic actions on plant surface coatings can be induced by surface
contact,alone. It is generally accepted that entry of sulfur dioxide into a
plant is through leaf openings termed stomata. Environmental conditions
(e.g., light, temperature, and humidity) that favor open stomata at the time
of exposure permit the assimilation of S0~. In the plant, S0? is converted to
sulfite or bisulfite and eventually to sulfate. Sulfite and bisulfite are
approximately 30 times more toxic than sulfate.
Air pollutants, as do all plant stress-inducing agents, initiate changes
within plant metabolic systems. Changes in metabolic pathways may lead to
extensive physiological dysfunctions. If physiological dysfunctions are
severe, visible symptoms may be manifested. Two basic forms of injury occur:
(1) injury not visibly expressed that may or may not result in reduction in
growth or loss in yield; and (2) injury visibly expressed as clinical symptoms
that may be observed, recorded, and evaluated.
As long as the rate of absorption in plants does not exceed the rate of
conversion of SO- to sulfate, the only effects of exposure to S0? may be
changes in stomatal opening or closing, or biochemical or physiological changes.
If SQ^ exposure concentrations are reduced, abatement of effects may occur.
1-50
-------
Reduced growth and yield and/or predisposition to ot.ier biotic or abiotic
stress-inducing agents may occur if alterations in plant metabolism or phy-
siology persist for a period of time. Significant reductions 1n growth and
yield of major forest tree species and agronomic crops have been reported
without the presence of visible symptoms.
Visible symptoms result from both acute and chronic injury. Acute injury
symptoms include necrosis or death of cells, tissues, organs, or the entire
plant. Chronic injury symptoms include plant responses that usually involve
chlorophyll disruption, followed by induction of chlorosis or yellowing of
tissues. Pigmentation changes resulting in stippling or general discoloration
characterize this type of injury. Chronic injury results from either high-dose
or low-dose exposure; high-dose exposure may lead rapidly to acute injury.
These terms refer to plant response rather than to the exposure conditions and
the dose received. Dose-response relationships relate variations in the
length of exposure and pollutant concentration to variations in plant responses
as mediated by the plant response system and the environment. Dose is defined
as the concentration of pollutant multiplied by the length of the period of
exposure.
Variations in exposure regimens and response measurements frequently make
it difficult to compare the results of different studies. The conclusions
listed below are based on a synthesis of the data discussed in Chapter 7 and
summarized in Tables 1-5, 1-6, 1-7. These conclusions were formed by summarizing
the dose-response data without considering the confounding environmental
variables: (1) Significant suppression in yield of economically important
agricultural species by SO- concentrations in the range from 0.05 to 0.06 ppm
can occur if the period of exposure is at least two weeks in length. (2) Both
1-51
-------
TABLE 1-5. DOSE-RESPONSE INFORMATION SUMMARIZED FROM LITERATURE PERTAINING TO CULTIVATED AGRONOMIC
CROPS AS RELATED TO FOLIAR, YIELD, AND SPECIFIC EFFECTS INDUCED BY INCREASING S02 DOSE
Ul
ro
Cone,
ppm
0.01
0.01
0.015
0.02
0.035
0.02
0.05
0.10
0.03-
0.10
0.3-
0.06
0.10
0.15
0.20
0.035
1.75
0.05
Exposure
time
10 mln
Growing season
Growing season
10 m1n
3 hr for 8 exp,
growing season
Growing season
(2 year)
72 hr/wk for
growing season
24 hr/day
for 30
days
8 hr
10 mln
Exposure.
condition
EC
EC
F/CC
F-ZAP
F/CC
F/CC
EC/SO
EC
Effects on
P1antc Foliar Yield
Corn
Oats X
Wheat X
Bean
Wheat
Winter wheat
Prairie June grass
Barley X
Durham wheat
Spring wheat
Winter wheat X
ev. Yamhin
ev. Ilyslop X
Broadbean
Broadbean
Species Effect0
i I
Stomata open wider
Light leaf Injury
15X decrease 1n grain weight yield
Stomata open wider
No effect on apparent photosynthesis, no
effect on the average head length or
number of grains per head
S content Increased with Increase In S0«
concentration; digestibility of dry
matter was reduced by 2 years of treat-
ment; crude protein content In winter
wheat decreased significantly
Ho effect on yield
0.03 ppm Increased yield 271
0.06-0.15 ppm. no effect
0.06 ppm decreased yield 221)
0.20 ppm decreased yield 70J
Depressed photosynthesis
Stomata open wider; threshold 0.02 pom for
10, mln
Reference
448
159
159
448
394
473
473
473
40
37
-------
TABLE 1-5. (continued)
Cone,
ppm
0.05
0.05
0.10
0.25
Exposure
time
5 hr/day; 5 *ay/wk
for 4 wk
4 hr
Exposure.,
condition0
EC/SO
EC/SO
Plantc
Alfalfa
Tobacco
Oats
Radish
Soybean
Tobacco
Effects on
Foliar
Yield
X
X
X
Species Effect*
261 decrease In foliage dry weight
491 decrease In root dry weight
22% decrease 1n leaf dry weight
No foliar Injury
Reference
430
431
0.05
0.20
0.06
0.065
0.13
0.26
0.10
0.10
0.10
8 hr/day
5 days/wk
for 18 days
68 days
1-55 days
EC/SO
EC/SO
GC
20 m1n
1 hr
8 hr
EC
EC
GH
Soybean
Alfalfa
Cabbage
Chinese cabbage
Cucumber
Eggplant
Lettuce
Spinach
Bean
Corn
Bean
Tobacco
No effect on top fresh or dry weight; root
fresh or dry weight; plant height; shoot/
root fresh or dry weight ratio
281 decrease 1n foliage stubble, 451 decrease
root dry weight
211 decrease In total protein content, amlno
add content, total nonstructured carbohy-
drates, symblotlcally fixed nitrogen 327
Foliage Injury threshold 0.13x27 days 140
Foliage Injury threshold 0.26x55 days
Foliage Injury threshold 0.13x26 days
Foliage Injury threshold 0.13x11 days
Foliage Injury threshold 0.13x15 days
Foliage Injury threshold 0.13x1 day
Other parameters measured such as plant height,
number of leaves, top fresh weight, number of
flowers, fresh weight vs. dry weight of roots
were not found to be significantly different
from controls
Stomata open wider, effect also shown to occur 448
In dark
Stomata open wider, water-stressed plant had 448
wider opening of stomata compared with
controls
Foliar Injury threshold for development of 283
fleck-like lesions
-------
en
Cone,
ppm
0.10
0.10
0.10
0.125-
1.0
0.1S
0.1S
0.25
0.11
Exposure
time
6 hr/day
for 133 days
18 days
6 hr/day
43 days
92 days
133 days
1-3 hr
18 days
72 hr/wk for
growing season
18 days
103.5 hr/wk
for 20 wk
Exposure.
condition
F/CC
GC
Plant0
Soybean
Pea
Effects on
YTeT?
Foliar
F/OT
EC/SO
GC
Soybean
Oats
Radish
Sweet pea
Swiss chard
Pea
F/CC
GC
Barley
Durham wheat
Spring wheat
Pea
GC
Cocksfoot
Meadowgrass
Italian ryegrass
Timothy
Species Effect6 Reference
No significant effect on growth or yield; 169
92nd-day defoliation was \2l greater;
135th-day seed weight was 1% reduced fnw
control
31 decrease In fresh weight of shoot i i 204
51 decrease 1n dry weight of shoot
41 decrease 1n tojal nitrogen
301 decrease In H (buffer capacity)
101 Increase glutamate dehydrogenase activity
1101 Increase In Inorganic sulfur content
No significant effect on foliar Injury, defolla- 171
tlon fresh weights, seeds/plant, or weight
of seeds/plant
No foliar Injury 32
31 decrease 1n fresh weight of shoot 204
81 decrease In dry weight of shoot
21 decrease 1n tojal nitrogen
35S decrease In H (buffer capacity)
321 Increase 1n glutamate dehydrogenase activity
140X Increase In Inorganic sulfur content
421 decreased yield In Durham wheat; 441 473
decreased yield 1n barley; no effect on
spring wheat
32X decrease In fresh Height of shoot 204
26! decrease In dry weight of shoot
241 decrease In to.ta1 nitrogen
421 decrease In H (buffer capacity)
BOt Increase In glutamate dehydrogenase activity
1501 Increase In Inorganic sulfur content
401 decrease In total dry weight 13,14
281 decrease In total dry weight
Nonsignificant
511 decrease In total dry weight
(Yield reductions were related to decrease in
leaf areas)
-------
TABLE 1-5. (continued)
en
01
Cone.
PP»
0.15
0.15-
0.30
0.15
0.17
0.20
0.25
0.25
0.20
0.30
0.40
0.50
0.60
0.70
0.20
0.20
0.218
Exposure
tine
24 hr
7 days
14 days
2 hr
1 hr
2 hr
2 hr
15 days
To maturation
4.5 hr/day
for 4 days
Exposure,,
condition0
CC/SD
EC/SO
GH
EC/SO
EC/SO
GC
GC
EC/SO
F-ZAP
P1antc
Corn
Rice
Barley
Bean
Com
Celery
Big plantain
Bean
Big mallow
Broadbean
Alfalfa
Alfalfa
Barley
Oats
Alfalfa
Barley
Tomato
Kidney bean
Soybean
Effects an
Foliar
YfeTo*
Species Effect6 Reference
Absorbed SO. remained In water-soluble 490
form and very difficult to assimilate to
protein
Severe foliar Injury 288
Ho Injury | (
Severe foliar Injury
Increased peroxldase activity, caused 356
chlorosis of leaves
Increased peroxldase activity, decreased
buffering capacity of cells, caused
necrotlc leaf Injury
Caused necrotlc leaf Injury
Caused necrotlc leaf Injury
Decreased photosynthetlc rate, decreased 39
stomatal resistance If RH > 401. Increased
stomatal resistance If RH < 401
No effect on photosynthesis 471
Threshold dose for Inhibition of photosynthesis 30
No effects at these doses
30
Threshold dose for Initial symptom of tissue
death, decrease or change In vitamin B..
Bg, and nlcotlnlc acid content
151 decrease In total yield
No visible damage
201 decrease 1n yield
449
33
201
-------
TABLE 1-5. (continued)
I
on
Concj
ppm
0.23
Exposure
time
14 days
Exposure.
condition
Plant0
GH Buckwheat
Lucerne
Red clover
Little stinging nettle
Ryegrass
Effects on
foliar
0.25
0.25
0.25
0.40
o.eo
1.20
0.25
0.25
0.25
0.25
0.25
0.40
0.80
1.20
0.25
0.25
0.29
1 hr
2 hr
Once every wk
(3 hr) to once
1n ^wk (3 hr)
4 hr
8 hr
24 hr
1 hr
3 hr every
2 wk for
growing season
Unknown
Unknown
15 days
F/CC
EC/SO
EC/SO
GC
EC
F/CC
EC
EC
GC
Broadbean
Broadbean
Alfalfa
Barley
Durham wheat
Spring wheat
Broccoli
Tobacco, Bel B,
Tobacco
Broadbean
Begonia
Alfalfa
Barley
Durham wheat
Spring wheat
Pea
Sunflower
Morning glory
Corn
Sorghum
Lettuce
X
X
X
X
Species Effect*" Reference
Caused necrotlc leaf Injury 356
Caused necrotlc leaf Injury
Caused necrotlc leaf Injury
Increased peroxldase activity
Increased S content In leaves i I
Slight swelling of stroma thylakolds of 464
chloroplast, effect reversible
Stroma thylakolds spread to granum
thylakolds, effect reversible
Ho effect on yield 473
It leaf Injury 432
6t leaf Injury
Chlorophyll a decreased more sharply than
chlorophyll b 40
Stomata opened faster and wider In light 286
condition; stomata opened longer In darkness
Foliar Injury 431
No effect on yield 473
501 decrease In net photosynthesis 62
lOt decrease In photosynthetlc rate 450
20-30X decrease 1n photosynthetlc rate
No effect
No effect
Foliar Injury, 30t decrease In thlamlne content 449
-------
TABLE 1-5. (continued)
I
en
Cone,
ppm
0.30
0.35
0.35
Exposure
time
5 hr/day
6 day/wk
12 days
26 days
1 hr
21 days
Exposure.
condition13
EC/SD
EC/SD
EC/SD
Plant1"
Barley
Bean
Sunflower
Barley
Effects on
Foliar
Yield
Alfalfa
Pea
0.38
1.15
1/90
0.40
0.40
0.50
0.60
0.40
0.40
0.45
0.46
0.50- ,
6.00
14 days
3 hr
4 hr
6 hr
6 hr/2 wk
(1-2 exposures)
3 hr for 7
exposures for
growing season
7 hr
30 min
0
EC/SD
EC/SD
EC/SD
F/CC
EC
F/CC
Radish
Oats
Tomato
Apples
Alfalfa
Cotton
Pecan
Pepper
Wheat
Buckwheat
Soybean
X
X
X
X
Species Effect0 Reference
11% foliar injury; 38% decrease in 292
dry weight shoot
1% foliar injury; 38% decrease in
dry weight shoot
5% foliar injury; 41% decrease in
dry weight shoot
21% foliar injury; 26% decrease in
dry weight shoot
2% foliar injury; 15% decrease in dry
weight shoot
16% foliar injury; 29% decrease in
dry weight shoot
80% decrease in apparent photosynthesis 472
Increase in glutamine content, decrease 205
in glutamic acid and protein content
inorganic S accumulated
Necrosis and growth inhibition at 0.35
x 14 days
Decrease injury at 0.38 and above, inhibited 111
seed germination, formation of green leaf-
lets of sprouts, and root growth
Threshold for leaf injury 174
Increase accumulation total and soluble 29
S content
No effect 227
No difference found in total N, protein/total 424
N ration, chlorophyll, all plants, all
treatments
No accumulative effect on yield, no effect on
average head length or number of grains/head 394
Injury threshold
Very significant negative linear relation-
ship between percent leaf area destroyed
and percent crop loss; 0.66% yield loss
for every 1% increase 1n foliar injury;
asymtomatnc plots Increased yield by
6.02% over controls
498
102
-------
TABLE 1-5. (continued)
en
CO
ConCi
ppn
0.50
0.50
0.50
Exposure
time
1.5 hr
1.5 hr
2 hr »-
Exposure.
condition0
EC/SO
EC
P1antc
Soybean
Oats
Begonia
Petunia
Coleus
Effects on
Foliar Yield
X X
X
X
Snapdragon
0.50
0.50
0.50
0.50
0.50
2 hr EC
4 hr EC/SO
4 hr EC/SD
4 hrfday EC/SO
for 14 days
5 hr/day
6 days/wk
for 12 days
26 days
Grape
Alfalfa
Broccol 1
Radish
Tobacco
Tomato
Oats
Radish
Soybean
Tobacco
Oats
Barley
Bean
Sunflower
Barley
Bean
Sunflower
X
X
X
X
X
X
Species Effect0 Reference
71 decrease In short fresh weight; 172
trace foliar Injury
Inltlon of leaf Injury 174
No effect
30t decrease In flower number; 151 2
decrease In shoot weight '
No effect; 12% decrease In shoot weight
lit decrease In number of flowers; no
effect
1901 Increase In stomatal resistance 371
191 leaf Injury
41 leaf Injury 431
11 leaf Injury
11 leaf Injury
11 leaf Injury
Foliar Injury occurred to all crops 441
321 decrease top dry weight; 131 decrease 175
1n root dry weight; number of heads
unchanged
241 foliar Injury; 421 decrease In dry 292
weight shoot
71 foliar Injury; 311 decrease 1n dry
weight shoot
181 foliar Injury; 441 decrease In dry
weight shoot
361 foliar Injury; 451 decrease In dry
weight shoot
121 foliar Injury; 341 decrease In dry
weight shoot
261 foliar Injury; 351 decrease In dry
weights shoot
-------
TABLE 1-5. (continued)
Cone,
ppm
0.50
1.00
1.50
0.50
0.50
0.56
0.77
0.92
0.60
0.60
0.70
0.70
Exposure
time
4 hr
5 hr/day
6 days/ week
for 14 days
6 days
4 hr
6 hr
6 hr/day
5 days/week
for 14 days
8 hr/day for
3 days
6/12, 24 hr/day
1-7 days
Exposure^
condition
GC/G
EC
0
EC/SD
EC/SD
EC/SD
EC/SD
EC
Plant1"
Red Clover
Broadbean
Sunflower
Tobacco
Pea
Cucumber
Apples
Soybean
White bean
Broadbean
Bean
Effects on
Foliar YTeTd
0.75
0.80
0.80-
2.00
0.80-
2.00
3 hr
2 hr
4 hr, 20 min
4 hr, 20 min
EC/SD
F/ZAP
F/ZAP
Al fal fa
Alfalfa
Soybean
Soybean
X
X
X
Species Effect0 Reference
Increase in vitamin A, fat, protein content 196
Significant change in plant nutritional
components
Under drought conditions exposure caused 379
wider opening of stomata, no effect on
diffusive resistance
Increased glutamate dehydrogenase; increased 465
peroxidase activity
Accumulation of significant total and soluble 29
sulfur
7.3% increase in foliage injury; 5% increase 227
leaf abscission
No effects
Bifacial necrotic lesion on mature leaves 183
20% decrease in photosynthesis after 2 hr
fumigation; after 3-day fumigation, 1 24
hr to full recovery in light condition;
no foliage injury
Increase in total amino acids and ammonium,
decrease in aspartic acid glutamic acid
and protein synthesis, all before visible
injury present
No injury developed 192
Threshold dose for foliar necrosis; 25-50% 30
decrease in net photosynthesis
4.5% decrease in yield at 1.4 ppm
11% decrease in yield at 1.7 ppm 310
15% decrease in yield at 2.0 ppm
Epidermal and mesophyll cell death, the number
of dead mesophyll cells highly correlated
with increase in S02
Highest S02 concentration, significant decrease
in seed yield
-------
TABLE 1-5. (continued)
Cone,
ppm
0.90
1.00
1.00
1.00
1.00
1.00
1.0
1.0
1.00
1.50
1.50
Exposure
time
2 hr
2 hr
3 hr
3 hr
2 hr
Exposure.
condition
EC/SO
GC
EC/SO
EC
4 hr
EC/SO
1 hr/2 days
for 4 days
6 hr/day
for 3 days
1.5 hr
3 hr
0.75-3 hr
3 hr
EC/SO
EC/SO
EC/SO
EC/SD
EC/SO
P1antc
Broadbean
Barley
Polnsettla
eight cultlvars
Alfalfa
Begonia
Petunia
Coleus
Snapdragon
Broccol 1
Bromegrass
Cabbage
Lima bean
Radish
Spinach
Tomato
Geranium
Strawberry
Soybean
Soybean
Alfalfa
Effects
Foliar
X
X
X
X
X
X
X
X
X
X
X
X
X
on
Yield
X
X
X
X
X
X
30
177
192
Species Effect6 Reference
261 decrease In net photosynthesis under 39
saturated light conditions; 52t decrease
1n net photosynthesis under nonsaturated
light conditions
Threshold dose for foliar necrosis; 30-60
decrease In net photosynthesis ( |
No effect
Leaf necrosis at 315 ppm CO. was 2.8x that
Induced under 645 ppm C0_
No effect
30t decrease 1n flower number; 191 decrease 1n
shoot weight
271 decrease In flower number; 191 decrease In
shoot weight
141 decrease In flower number; 161 decrease 1n
shoot weight
381 leaf Injury
651 leaf Injury
701 leaf Injury
251 leaf Injury
461 leaf Injury
491 leaf Injury
331 leaf Injury
Rapid closing of stomata In low-RH air after
exposure; slow closing 1n hlgh-RH conditions,
stomata remained open
No effect on growth and development
Necrotlc lesions, lower leaf surface
91 decrease 1n shoot fresh weights, 41 leaf
Injury 211-291 decrease 1n shoot fresh weight
24-941 decrease In shoot fresh weight; 63-931
foliar Injury
Leaf necrosis at 315 ppm CO. was 2.5x that
Induced under 645 ppm CO,
431
47
358
172
172
192
-------
TABLE 1-5. (continued)
Cone,
PPM
1.50
2.00
2.0
2.5
3.0
4.00
Exposure
time
To maturation
Exposure,.
condition0
EC/SO
2 hr
EC
3 hr
6 hr
1 hr
2 hr
3 hr
2 hr
GC
EC/SO
GC
GC
GC
EC
Mante
Kidney bean
Effects on
Foliar
Begonia
Petunia
Coleus
Snapdragon
Polnsettla X
eight cultlvars
Apples X
Polnsettla X
eight cultlvars X
X
Begonia
Petunia
Coleus
Snapdragon
Yield
X
X
Species Effect0
Reference
201 decrease In root dry weight; 141
decrease 1n legume dry weight;
171 decrease 1n seed dry weight; 33
10-301 decrease In apparent photo*
synthesis. Increase In chlorophyll
a and b content
141 decrease flower number; 221 decrease ' ' 2
1n shoot weight
321 decrease In flower number; 241 decrease
1n shoot weight
301 decrease in flower number; 201 decrease
In shoot weight
151 decrease In flower number; 151 decrease
In shoot weight
0-18.31 foliar Injury 177
171 Increase In foliar Injury; 621 Increase 227
in lead abscission; 191 decrease In shoot
growth
0.13.81 foliar Injury 177
1.8-26.81 foliar Injury
18.8-96.51 foliar Injury
271 decrease in flower number; 331 decrease
In shoot weight 2
421 decrease In flower number; 321 decrease
in shoot weight
301 decrease in flower number, 211 decrease in
shoot weight
201 decrease In flower number; 191 decrease In
shoot weight
second-order divisions. Doses within a single study that
concentration that Induced said effect.
*Table arranged by Increasing SO, concentration as first-order and exposure time as
did not Induce specifically different effects are listed along with the lowest SO.
F » field or forest surveys
F/CC • Field, closed chambers
0/OT • Field, open-top chambers
F/ZAP • Field, zonal air pollution system
6 • Greenhouse
GC • Growth chambers
EC • Exposure chambers
EC/SO • Exposure chamber, special design
0 • Other
c$ee Appendix A for most scientific latin binomials of plants.
*l Indicates study examinated foliar and/or yield effects. The X does not necessarily imply that an effect was found.
'most prominent or significant effect reported.
-------
the quality and quantity of the crop can be negatively affected by the concen-
tration of pollutant multiplied by the length of the exposure period. Unfortunate
variations in exposure regimens and response measurements make it difficult to
compare the results of different studies.
The concept of dose-response can be demonstrated by a synthesis of the
data presented in Tables 1-5, 1-6, 1-7. The following conclusions were formed
by summarizing the dose-response data without designating specific associated
exposure conditions:
o Yield of economically important agricultural spe cies can be signifi-
cantly suppressed by SO- concentrations in the range of 0.05 to 0.06
ppm if the exposure period is sufficiently long (2 weeks). Both
crop quality and quantity can be negatively affected.
o Fluctuating, long-term (seasonal, annual) SO- exposures averaging
0.05 ppm or less can cause economically and ecologically undesirable
effects on productivity and stability of range and forest ecosystems.
o As SO- concentrations increase to 0.25 ppm, a variety of agricultural
crops (such as alfalfa, timothy, range grasses, soybean, barley,
wheat, cabbage, lettuce, spinach, tobacco, cucumber, eggplant, pea,
and kidney bean) respond with necrotic foliar injury or suppressed
yield. Approximately 70 percent of the cultivated agronomic crop
species exposed to 0.25 ppm or less respond to the SO- exposure of 1
hour or longer with changes in stomatal aperture (leaf openings),
foliar injury or yield effects. Foliar injury on vegetables and
suppression of yield are directly related to economic values.
o Forest trees species (such as pine, spruce, fir, beech, alder and
poplar, representing coniferous and deciduous forest ecosystems)
respond to 0.25 ppm or less of S02. Approximately 90 percent of the
1-62
-------
TABLE 1-6. DOSE-RESPONSE INFORMATION SUMMARIZED FROM LITERATURE PERTAINING TO FOREST TREE SPECIES
AS RELATED TO FOLIAR, -YI£LB, AND SPECIFIC EFFECTS INDUCED BY INCREASING S02 DOSE
Cone,
ppm
0.001
0.003-
0.09
0.09-
0.12
0.004
0.35
0.006
i- 0.007-
cn 0.01
co
0.007-
0.01
t
0.008
0.011
0.017
0.015
0.019
0.023
0.025
Exposure
time
10 yr avg.
i"
Annual avg
Annual avg
Annual avg
Annual avg
(exposed 5 mo)
Growing season
Annual avg
Annual avg
10 yr avg
Growing season
10 yr avg
Annual avg
Annual avg
Annual avg
6 hr
Exposure.
condition1'
f
F
F
F
F
F
F
F
F
F
F
F
EC/SO
Hwite
Forest trees
Scotch pine
Eastern white pine
White birch
Fir forests
Fir forests
Forest trees
White birch
Forest trees
Conifers
Conifers
Conifers
Eastern white pine
Effects on
Foliar Y1e1d~
X X
X X
X
X X
X X
X
X X
X
X
X
X
Species Effect'
Reference
Mo Injury
i I
Decreased photosynthesis leading to the
death of tree
No significant difference In S content of
foliage
No effect on foliar S content
20 +_ 5X growth Increase
20 + 5* growth decrease
Premature defoliation
Very little chronic foliar Injury
Increased foliar S content; trace to light
foliar Injury
Mostly chronic foliar Injury; some acute Injury
301 decrease In growth
52t decrease in growth
54! decrease In growth
Threshold dose for needle damage; most
sensitive clones only
355
369
273
297
296
273
.54
454
454
193
-------
TABLE 1-6. (continued)
Cone,
ppm
0.025-
0.037
0.026
0.026-
0.037
0.035
0.038-
0.057
Exposure
time
Annual avg
Growing season
Annual avg
5 mo
Annual avg
Exposure.
condition"
F
F
F
F
F
P1antc
Fir
White birch
F1r
Eastern white pine
Scotch pine
Effects
Foliar
X
X
X
X
on
Yield
X
X
X
0.045
0.045
0.048
0.05
0.15
0.05
0.05
0.05
0.10
0.20
0.05
0.05-
0.10
0.05
0.10
0.20
Growing season
10 yr avg
Growing season
6 hr
49 days
10 wk
5 mo
9 mo
9 mo
9 mo
F
F
F
EC/SO
F/CC
F/CC
F/CC
F/CC
F/CC
Jack pine
Forest trees
White birch
Eastern white pine
Norway spruce
Spruce
Beech
Spruce
Scotch pine
Fir
X
X
X
X
X
X
I I
273
296
369
133
Species Effect0 Reference
Death of groups of trees 297
Moderate to severe foliar Injury
Rapid death of groups of trees
Foliar Injury
Species occurrence negatively correlated
with SO. ambient cone.; foliar S content
positively correlated with SO- ambient
cone.; foliar S content positively
correlated with SO. ambient cone.; by
distance from source
Reduced chlorophyll content, tissue death 273
Acute and chronic foliar Injury
Severe foliar Injury; foliar S concentration 273
3x normal
60! of tolerant clones foliar Injury 193
developed
Foliar Injury 491
No effects 224
Increase In S concentration proportional 223
to Increase 1n SO. exposure cone.;
terminal bud deatn
Decreased foliar buffering capacity; In- 222
peroxldase activity
Decreased photosynthesis 222
Decreased pollen viability 222
-------
TABLE 1-6. (continued)
I—1
1
cr>
tn
Cone,
ppm
0.069
0.07
0.10
0.20
0.10
0.15
0.30
0.18-
0.20
0.20-
1.00
Exposure
time
Annual evg
3 days
10 wk
76 days
9 wk
24 hr
1 hr
Exposure.
condition0
r
EC
'" F/CC
Mantc
Conifer
Eastern white pine
Spruce
Black alder
Poplar
Jack pine
Azalea
Flrethom
White ash
White birch
Effects on
Foliar
9.20
D.20
0.025
0.2S
0.27
0.35
0.40
0.50
12 hr/day
for 7 wk
110 days
2 hr
2 hr
3 mo
3 hr
74 hr
EC/SO
EC/SO
EC/SO
EC
EC/SO
EC
Hybrid poplar
English birch
Eastern white pine
Jack pine
Red pine
Loblolly pine
Short leaf pine
Slash pine
Virginia pine
Pin oak
White birch
Trembling aspen
Yellow pine
Eastern white pine
X
X
X
X
X
X
X
X
X
X
X
Species Effect0 Reference
701 decrease In growth 454
Chlorotlc spotting and death of needle tips 19
Decreased CO- uptake; positive correlation 223
between COf uptake and cambium growth; (.
Increase in cone. Induced annual ring
width
Increase phenoloxldase activity 491
Decreased leaf area Index and foliar 209
growth
Inhibited foliar llpld synthesis. Inhibition 286
reversible; Increase In dose • Increase 1n
recovery time
No appreciable effect on foliar sorptlon of SO. 367
Slightly decreased height; decreased relative 208
growth rate, relative leaf growth rate, and
relative leaf area expansion rate
No effect on phenyloxldase activity 491
6.51 foliar Injury 34
4.51 foliar Injury
0.5* foliar Injury
All equally sensitive; most sensitive period 35
8-10 wk of age or older
45X decrease In height growth 368
1071 Increase In height growth
21 foliar Injury 217
Chlorophyll content varied Inversely with 100
concentration
-------
TABLE 1-6. (continued)
Cone,
ppm
0.45
0.45
0.5
Exposure
time
6 hr
9 hr/day
for 8 wk
15 mln
30 m1n
60 m1n
120 m1n
Exposure.
condition
EC/SO
,, EC
Plant0
Eastern white pine
Ponderosa pine
Red pine
Effects on
Foliar Yield
X
X
0.50
2 hr
EC/SD
1—
1
en
en
0.50
0.50
3 hr
5 hr
EC/SD
GC
Eastern white pine
Jack pine
Red pine
Trembling aspen
Austrian, Ponderosa
Scotch pine. Balsam,
Fraser, White fir
Blue, White spruce
Douglas fir
0.50
0.65
1.00
1.00
1.07-
6.41
1.83
30 day
3 hr
4 hr
8 hr
30 m1n
to 6 hr
50 m1n
GC
EC/SD
GC
EC
0
0
Chinese elm
Gingko
Norway maple
Pin oak
Trembling aspen
Austrian, Ponderosa.
Scotch pine. Balsam,
Fraser, White fir.
White spruce, Douglas
fir
American elm
Scotch pine
P. pinea
V. nigra
Pine
Spruce
X
X
X
X
X
X
X
X
Species Effect0 Reference
All tolerant clones developed foliar 193
Injury
Severe needle tip chlorosis and necrosis 128
i I
•Decreased primary needle chlorophyll con- 78
tent.
Decrease dry weight of primary needles and
cotyledons
Further Increase of all of above effects
121 foliar Injury 34
lit foliar Injury
ZX foliar Injury
US foliar Injury 217
No Injury 399
Severe chlorosis and necrosis 416
Moderate marginal chlorosis
Moderate marginal chlorosis
Slight overall chlorosis
23X foliar Injury 217
Less than 4t foliar Injury all species 399
Inhibition of stomatal closing 335
Visible Injury was proportional to foliage 65
S content
SO. absorbed by exposed foliage In winter- 298
time; S stored In new shoots
-------
TABLE 1-6. (continued)
Cone,
ppm
2.00
2.00
2.00
2.00
2.00
2.00
3.00
Exposure
time
2 hr
6 hr
6 hr
6 hr
6.5 hr
12 hr
6 hr
Exposure,
condition0
CG
GC
GC
GC
0
GC
P1antc
Effects on
Foliar
Yield
Austrian, Ponderosa X
Scotch pine, Balsam,
Fraser, White fir. Blue,
white spruce, Douglas fir
American elm X
American elm
Chinese elm
Glnkgo
American elm
Glnkgo
Norway maple
Species Effect' Reference
No foliar injury on Douglas fir, firs, 399
spruce
Pine foliar injury threshold, necrotic
tips
Induce severe foliar injury; defoliation in 76
older leaves; significant reduced expansion
of new leaves; number of emerging leaves
and root dry weight reduced
No significant reduction in lipid content; 79
significant decrease 1n new leaf protein
content; significant decrease In leaf,
stem, root carbohydrate content
lOOt leaf necrosis 416
Water-stressed plant increased uptake 335
of S02
Induced stonatal closing; S content 416
Increased In plants fumigated In
light
SOX leaf necrosis 416
Table arranged by Increasing SO, concentration as first-order and exposure time as second-order divisions. Doses within a single study
that did not Induce specifically different effects are listed along with the lowest
S0_ concentration that Induced said effect.
F • field or forest surveys
F/CC « field, closed chambers
D/OT » field, open-top chambers
G/ZAP « field, zonal air pollution system
G • greenhouse
GC » growth chambers
EC • exposure chambers
EC/SO • exposure chamber, special design
0 « other
cS*e Appendix A for most scientific latin binomials of plants
-------
species tested in this range of SO. concentrations (with 2 hour
exposures) responded with physiological modifications, suppressed
photosynthesis, foliar injury, death of buds, or suppressed foliar
or woody growth.
o Non-woody components of native ecosystems such as lichens and grasses
also respond to S0? concentrations below 0.025 ppm. Responses
include suppressed growth, death, and reduced diversity in lichen
populations and suppressed photosynthesis and growth of leaves,
tillers, and stubble of grasses.
o At S0_ concentrations between 0.25 and 0.50 ppm (1-8 hours exposure),
less than 50 percent of the agronomic species tested responded
negatively to the sulfur dioxide treatments. At S0? concentrations
less than 0.25 ppm but for multiple days, 70 to 90 percent of the
species responded negatively.
A comparison of the species response to the latter treatments at the
concentrations ranging between 0.25 to 0.50 ppm with the response to concentrations
below 0.25 ppm might be interpreted as suggesting that plant response is not
positively correlated with dose. This is not the case. Exposure durations
used in the studies at the higher SO- concentrations generally ranged from 1-8
hours while multiple-day exposures were frequently used at the lower concentrations
of SO-- All agronomic species responded to a concentration of 0.50 ppm at
exposure durations ranging from 1.5 to 5 hours. A variety of responses occurred,
including physiological modifications, foliar injury, and suppressed growth.
These trends suggest that as the S0» concentration is increased (a) a shorter
exposure duration is sufficient to elicit a plant response equal to or greater
than that which occurred at a lower concentration; (b) responses become more
severe; (c) plants tolerant at lower concentrations become sensitive.
1-68
-------
TABLE 1-7. DOSE-RESPONSE INFORMATION SUMMARIZED FROM LITERATURE PERTAINING TO NATIVE PLANTS AS
RELATED TO FOLIAR, YIELD, AND SPECIFIC EFFECTS INDUCED BY INCREASING S02 DOSE
Cone,
ppn
0.006
0.018
0.01-
0.02
0.015
0.017
0.02
Exposure Exposure.
time condition0 Plant5
6 mo
Annual avg '"
Annual avg
6 mo
29 days
F
f
F
F
GC
Lichens
Lichens
Bryophytes
Lichens
Bryophytes
Ryegrass
Effects on
Foliar
Yield
0.14
22 days In
2 consecutive
growing seasons
29 days GC
22 days 1n 2
consecutive growling
seasons
Ryegrass
0.02
0.02
0.03
0.04
0.08
0.15
0.03
0.04
0.04
0.15
0.05
85 days
Growing season
10 wk
6 mo
51 days
Growing season
GC
GC
GC
F
GC
GC
Ryegrass
Ryegrass
Ryegrass
Lichens
Ryegrass
Ryegrass
Species Effect0 Reference
Loss of chlorophyll, decreased growth 255
Elimination of many lichen species 256«
i l
Decreased lichen diversity 397
Elimination of species 147
No effect on net photosynthesis, dark resplra- 87
tlon, transpiration coefficients, number of
tillers and yield
As above effects except visible foliar Injury
and reduction of specific leaf area
Increased organic S content 89
Increased organic and Inorganic S content 86
Alleviated S deficiency 277
Alleviated S deficiency
Alleviated S deficiency
Reduction In yield without symptoms
Tissue death ?' Sa
Decreased concentration glyclne and serlne; 234
Inhibited photoresplratlon pathway
Alleviated S deficiency symptoms 86
-------
TABLE 1-7. (continued)
Cone,
ppm
0.06
0.067
0.073
0.074
0.08
0.09
(Peak)
0.11
0.11
0.11
0.12
0.13
0.25
0.50
1.00
0.13
Exposure
time
Growing season
26 wk
26 wk
Exposure.
Condition" Plantc
Ryegrass
Effects on
Foliar"
4 wk
103 5 hr/wk
for 20 wk
9 wk
6 wk
6 wk
GC
EC
EC
Ryegrass
Ryegrass
18 hr
13 hr/day for
28 days
115 days
4 Mk
0
EC
F/CC
0
Splderwort
Foxtail grass
Ryegrass
Cocksfoot
X
X
EC
EC/SO
Ryegrass
Grass
Ryegrass
Ryegrass
EC/SD Ryegrass
Species Effect0 Reference
Increase in photosynthesis, respiration 135
and chlorophyll content; light Increase
In productivity
Increase In dry weight of leaves, number 24
of tillers, dry weight of stubble and |
leaf area; SIS decrease In yield
501 decrease 1n shoot dry weight 241 25
decrease In chlorophyll a content;
261 decrease In chlorophyll b content
Increase In chromosome aberration rate of 282
germinating pollen
Foliar Injury as caused by heavy metals was 236
Increased by SO- exposure
Decrease 1n weight; accelerated leaf 42
senesence
301 decrease 1n leaf area; 45! decrease 12
In dry weight; decrease In number tillers;
decrease In number of green leaves;
decrease root/shoot ratios
20% decrease 1n leaf area; 4OT decrease 1n 14
dry weight; decreased root/shoot ratio
Significant decrease In leaf area and all dry 13
weight fractions; decrease In number of leaves
and tillers
Decrease 1n dry weight of leaves, number 24
of tillers, dry weights of stubble and
leaf area; 461 decrease In yield
Foliar necrotlc lesions and decrease In 135
net primary productivity at 0.13 ppm
and above
Decreased productivity 135
-------
TABLE 1-7. (continued)
Cone,
ppm
0.15
0.15
0.15
0.20
0.20
(peak)
0.25
0.27
0.27
0.38
(peak)
0.50-
11.00
0.71
2.00-
0.00
Exposure
time
6 wk
51 days
Growing season
2 hr
55 days
5 wk
14 days
8 wk
6-43 wk
2 hr
1 hr
2 hr
5 hr
8 hr
Exposure.
condition"
GC
GC
F/CC
EC/SD
EC/SO
EC/SO
EC/SO
F/CC
EC/SO
EC/SO
Plant0
Duckweed
Duckweed
Ryegrass
Kentucky
bJuegrass
Ryegrass
Ryegrass
Ryegrass
Ryegrass
Ryegrass
87 Desert
species
Lily
Diplacus
meteromeles
Effects on
Yield
Foliar
3.50
1 hr
Acacia
Species Effect0 Reference
Decrease in starch content and size of 131
fronds
Decrease in starch content and growth; 131
decrease 1n surface area dry weight ' I
Alleviated S deficiency symptoms; Increase 85
in S content of foliage, free amino acid
content, and N/S ratio
Visible foliar Injury 3?0
Decrease In weight; accelerated leaf sensence 4?
No effect on number of tillers; 17t decrease 188
In yield
Increase in free amino acid content 11
381 decrease In green weight; 301 decrease 188
in total dry weight; no reduction in
number of tillers; 2* senesence
36X decrease In total dry weight 93
Most plants required more than 2.00 ppm 182
S0_ to produce foliar injury
Inhibited pollen tube elongation at all 294
exposure durations
Increase In SO. dose Induced a progressive 478
decrease In photosynthesis and transpiration
Foliar Injury 397
-------
TABLE 1-7. (continued)
1 Table arranged by Increasing SO. concentration as first-order and exposure time as second-order divisions. Doses within a single study
that did not Induce specifically different effects are listed along with the lowest SO. concentration that Induced said effect.
1 F • field or forest surveys
F/CC • field, closed chambers
0/OT • field, open-top chambers
G/ZAP « field, zonal air pollution system
G • greenhouse
GC • growth chambers
EC • exposure chambers
EC/SO « exposure chamber, special design
0 • other
Se« Appendix A for most scientific latin binomials of plants.
-------
The lack of short- or long-term monitoring data make^ it difficult to
assess the results of dose-response studies in the field. When data are not
available to determine whether short-term spike concentrations occurred, then
only long-term averages have been used to define the dose. Obvious differences
between forests in areas with high SO- concentrations have been observed.
There is usually no exact dose information for short-term influences; therefore,
in most field studies, only long-term averages are used to define the dose.
As the dose of S02 increases, plants develop more predictable and more
obvious visible symptoms. Foliar symptoms advance from chlorosis or other
types of pigmentation changes to actual necrotic areas, and the extent of
necrosis increases with exposure. Studies of the effects of SO- on growth and
yield have demonstrated a reduction in the dry weight of foliage, shoots,
roots, and seeds, as well as a reduction in the number of seeds. At still
higher doses, reductions in growth and yield increase. Extensive mortality
has been noted in forests continuously exposed to S0_ for many years.
The amount of sulfur accumulated from the atmosphere by leaf tissues is
influenced by the amount of sulfur in soil relative to the sulfur requirement
of the plant. After low-dose exposure to S0_, plants grown in sulfur-deficient
soils have exhibited increased productivity.
Sulfur dioxide and particulate sulfate are the main forms of sulfur in
the atmosphere, and a plant may be exposed to these pollutants in several
different ways. Dry deposition of particulate matter and wet deposition of
gases and particles bring sulfur compounds into contact with plant surfaces
and soil substrates. The effects of such exposures are more difficult to
assess than those associated with the entry of S0_ through plant stomata.
Plant response to dynamic physical factors such as light, leaf surface moisture,
relative humidity, and soil moisture may influence stomatal opening and closing,
1-73
-------
and hence, play a major role in determining sensitivities of species and
cultivars or the time of sensitivity of each on a seasonal basis. Dose-response
relationships are significantly conditioned by environmental conditions before,
during, and following exposure to S02>
Sulfur accumulation in plants has been suggested as a tool for determining
the levels of sulfur in the atmosphere of a given area over time. Such data,
however, cannot accurately define the dose received by a plant.
Very few studies have been conducted to determine the sensitivity of
microorganisms to SO- or to explore the interactions of SO- with plant pathogens
such as fungi, nematodes, or bacteria. Both inhibition and enhancement of
disease processes have been reported, but more data are needed to provide
reliable information on trends.
In ambient atmospheres, SO- and other pollutants usually exist as diverse
mixtures in which a multiplicity of chemical combinations can take place.
Therefore, with the possible exception of a point source in which vegetation
is exposed to a high dose of SO-, the direct and indirect influences of other
air pollutants in combination with SO- must be considered. Major phytotoxic
air pollutants that have been studied in combination with S0_ include ozone,
oxides of nitrogen, and hydrogen fluoride. The interactions of SO- and 0_
have been most extensively investigated because of the incursions of ozone and
other oxidant precursors into many rural areas, as well as their presence in
urban areas. Many studies have demonstrated more than additive effects in
symptom expressions, but relatively few studies have attempted to evaluate the
impacts on growth and yield. Additionally, pollutant combinations with SO-
have caused less than additive and/or additive effects depending upon doses
applied. The influence of various physical and biological factors of the
environment increase in complexity as the pollutants are combined together
during exposures.
1-74
-------
A few studies have examined effects of combinations rf particulate matter
and SO- or participate matter and other pollutants: (1) increase foliar
uptake of S02; (2) increase foliar injury of vegetation by heavy metals; and
(3) reduce growth and yield. Because of the complex nature of particulate
pollutants, conventional methods for assessing pollutant injury to vegetation,
such as dose-response relationships, are poorly developed. Studies have
generally reported vegetation responses relative to a given source and the
physical size or chemical composition of the particles. For the most part,
studies have not focused on effects associated with specific ambient concentrations.
Coarse particles such as dust directly deposited on the leaf surfaces result
in reduced gas exchange, increased leaf surface temperature, reduced photo-
synthesis, chlorosis, reduced growth, and leaf necrosis. Heavy metals deposited
either on leaf surfaces or in the soil (and subsequently taken up by the
plant) can result in the accumulation of toxic concentrations of the metals
within the tissue.
Particulate matter is heterogeneous in size and composition, ranging in
mean diameter from <0.005 urn (molecular clusters) to >100 urn (visible dust
particles). Particles occur in both solid and liquid phases and vary in
chemical composition from a single chemical species (e.g., H-SO^) to complex
combinations of chemical species. They are produced directly from stationary
and mobile sources and are also formed secondarily in the atmosphere through
chemical reactions.
While coarse particles (>2.0 pm in mean diameter) settle rapidly, fine
particles (<2.0 urn in mean diameter) have prolonged atmospheric residence
times and are not strongly influenced by gravity. Because of the complex
nature of particulate matter, dose-response studies are very difficult to
conduct and data, therefore, are not available for making generalized statements.
1-75
-------
Vegetation within terrestrial ecosystems is sensitive to direct SO-
toxicity, as evidenced by changes in physiology, growth, development, sur-
vival, fecundity, and community composition. Responses of individual organisms
reflect both direct or indirect effects. Habitat modification by S02 results
in indirect effects occurring. Nutrient cycling appears to be a sensitive
indicator of subtle, yet important, environmental modification. The removal
of certain lichen species can reduce nitrogen fixation in forest ecosystems.
At the community level, chronic exposure to S0_ may result in a shift in
the species composition due to the elimination of individuals or populations
sensitive to the pollutant. The tendency for SO- derivatives to accumulate in
the soil may have consequences for the microbiota inhabiting the upper soil
horizons. The gradual accumulation of pollutant derivatives may cause a
change in soil chemistry and influence nutrient cycling and ecosystem pro-
ductivity.
Particulate emissions have their greatest impact on terrestrial ecosystems
near large emission sources, and participate matter in itself constitutes a
problem only in those areas where deposition rates are high. Most of the
material deposited by wet and dry deposition on foliar surfaces in vegetated
areas is transferred to the soil. Foliage may serve as a transitional site of
accumulation if previously deposited dry material becomes highly concentrated
during precipitation. Ecological modifications may occur if the particles
contain toxic elements, even though deposition rates are moderate. Solubility
of particle constituents is critical, since water-insoluble elements are not
mobile within an ecosystem. Soils are long-term sites for the retention of
many constituents found in particulate matter. Accumulation in the soil-litter
layer influences ecological processes such as decomposition, mineralization,
nutrient cycling, and primary production.
1-76
-------
1.4.2 Acidic Precipitation
Acidic precipitation is a major environmental concern in many regions of
the United States, Canada, northern Europe, and Japan. It has caused measurable
damage to aquatic ecosystems in Scandinavia, eastern Canada, and the northeastern
United States. Acidic precipitation has, by acidifying lakes, induced the
extinction of fish, caused the breakdown of nutritional food webs, and reduced
life in lakes to a few acid-tolerant species. Acidic precipitation, in addition,
has damaged national monuments and buildings made of stone. It also has the
potential for impoverishing sensitive soils, degrading natural terrestrial
ecosystems and for damaging forest ecosystems over the long-term (several
decades).
In an atmosphere relatively free of natural or man-made emissions of
sulfur and nitrogen oxides, precipitation would be expected to have a pH of
5.6 due to the presence of carbonic acid formed when atmospheric carbon dioxide
dissolves in water vapor. The precipitation has become acidic chiefly due to
the large amounts of sulfur and nitrogen oxides being emitted from the combustion
of fossil fuels (particularly coal and oil). In addition, substances present
in other gases, aerosols, and particulate matter from natural and man-made
sources also contribute to the acidity of precipitation.
Precipitation acts as a scavenger, bringing to earth substances present
in the atmosphere. The chemical composition of rain, therefore, depends on
the substances present in the atmosphere. On a global scale natural emissions
far exceed man's contributions; however, man-made emissions are localized in
specific geographic areas, where they may be concentrated in the atmosphere or
transported meteorologically to other areas downwind.
Sulfur and nitrogen oxides are transformed in the atmosphere to sulfates
and nitrates. Sulfates and nitrates upon hydrolysis in the the atmosphere
1-77
-------
contribute hydrogen ions (H+). If hydrogen ions are present in significant
quantities, precipitation becomes acidic. The acidity of precipitation is a
reflection of the balance between the major cations and anions in precipitation,
however, when determining the pH of precipitation all cations and anions
should be measured. Currently the acidity of precipitation in the northeastern
United States is between 4.0 and 5.0.
The ratio of sulfuric to nitric acids in precipitation varies from time
to time and place to place. In much of the eastern United States, the average
annual ratio of sulfuric to nitric acids is currently 2:1; however, nitric
acid is apparently becoming progressively more important as a contributor of
hydrogen ions. Preliminary estimates suggest that two-thirds of the sulfur
emitted into the atmosphere of eastern North America is probably deposited
there with the remainder leaving the atmosphere of the region and moving
primarily to the east.
Tall stacks (some as high as 1200 ft.) from power plants have decreased
local pollution problems but may have increased the widespread wet and dry
deposition of sulfur and nitrogen oxides by permitting them to be carried long
distances by air streams. Analysis of air-mass movements and chemical trans-
formations in the atmosphere indicates that acidic precipitation in one state
or region of the United States or Canada results in large part from emissions
which enter the atmosphere in other states or regions, often many hundreds of
miles from the original source.
Acidic precipitation is only one special feature of the general phenome-
non of atmospheric deposition. In addition to precipitation (wet deposition),
dry deposition also occurs. The major chemical substances that are transferred
into ecosystems via acidic precipitation (wet deposition) are also transferred
into ecosystems by dry deposition when it is not raining or snowing. It is,
1-78
-------
therefore, Impossible to chemically distinguish the biolos'cal effects of
acidic precipitations from those of dry deposition.
The increased deposition of acidic substances into aquatic ecosystems
has, as of 1979, caused hundreds of lakes in the Adirondack Mountain region of
New York State, certain lakes in northern Minnesota, and many hundreds of
lakes in various parts of southern Ontario and Quebec to show acid stress.
Reduction or extinction of fish and other plant and animal populations has
occurred. Lakes and streams in other regions of the United States and Canada
are also potentially vulnerable to stress by acidic precipitation. Damage or
injury to aquatic or terrestrial organisms is most likely to occur when a
particularly sensitive life form or life stage (one with a narrow range of
tolerance), developing in poorly buffered waters or soils, coincides in time
and/or space with major episodic injections of acidic precipitation or other
injurious substance.
The disappearance of fish from lakes and streams usually follows two
general patterns. One pattern results when sudden short-term shifts in pH
occur, the other results from the long-term decrease in the pH of the water.
A major episodic injection of acids and other soluble substances occurs when
these substances present in polluted snow are released in the meltwater during
warm periods in winter or early spring. The release of pollutants can cause
major and rapid short-term changes in acidity and chemical properties of
stream and lake waters. Fish kills are a dramatic consequence of such episodic
injections into aquatic ecosystems.
Equally dramatic long-term changes in aquatic ecosystems also occur from
the wet and dry deposition of acidic substances because the chemical compo-
sition of precipitation and dry deposition determine in part the chemical
composition of lake, stream, and ground waters. The terrestrial watershed
1-79
-------
also plays a significant role as the chemical composition of precipitation is
modified by chemical and biological weathering and exchange processes which
take place as precipitation washes over vegetation, percolates through the
soil and interacts with the underlying bedrock of the drainage basin in which
the precipitation occurs. The situation is analogous to a gigantic, regional
scale titration with the lakes and streams acting as receiving vessels for
acidic additions from the atmosphere. The titration end point of each lake is
predetermined by its hydrology and the capacity of the soils in the drainage
basin to assimilate the incoming acid. If the soils and drainage basin can no
longer assimilate the incoming acids, the lake and stream waters are changed
from conditions that are favorable for fish and other aquatic organisms to
conditions that inhibit reproduction and/or recruitment of populations of fish
and other aquatic organisms, some of which are food for fish.
Prolonged acidity interferes with reproduction and spawning so that
changes in the structure of a population occur over time. These changes
include a decrease in population density and a shift in size and age of the
population to one consisting primarily of larger and older fish. The process
is insidious, and effects on yield are often delayed and not recognizable
until the population is close to extinction. This is particularly true for
late maturing species with long lives. Large increases in the mortality rate
are not necessary to bring about population extinction. Even relatively small
increases (5 to 50 percent) in mortality of fish eggs and fry can decrease
yield and bring about population extinction. Many populations of freshwater
fish become extinct at a pH below 5.0 while the reproduction of many species
of aquatic organisms is inhibited at pH's between 6.0 and 5.0. Increasing
acidity of freshwater habitats causes shifts in species, populations and
communities of most aquatic organisms to occur. Virtually all trophic (food)
levels are affected.
1-80
-------
Interference with normal reproductive processes in fish occurs, not only
because of acidity of the water, but also due to increased concentrations of
certain metallic cations, notably aluminum, which become mobilized in acidified
lakes and streams. The metallic cations may originate from the lake or stream
watershed or be introduced in wet and dry deposition. Among the inorganic
substances are elements such as manganese, zinc, copper, iron, boron, flourine,
bromine, aluminum, lead, iodine, nickel, cadmium, vanadium, mercury, and
arsenic. The underlined elements in the list above are essential micronutrients
that are required by plants in small amounts. However, at concentrations
above the amount required, these same elements, can be toxic to plants and
animals. The non-esssential elements can also be toxic to plants or animals
when present in large amounts or when their mobility and solubility is increased
due to soil acidity. Also the deposition of these metallic substances in
precipitation can affect the foliage and roots of plants and injure microorganisms
or animals that may ingest the plants. In addition, these substances can harm
animals (including man) that may drink water containing these elements as well
as aquatic animals (especially fish) that live in the water.
An indirect effect of acidification, which is potentially of concern to
human health is the possible contamination of edible fish and of human water
supplies. Studies in the United States, Canada and Sweden have revealed the
presence of high mercury concentrations in fish from acidified regions. Lead
and copper have been found in household plumbing systems with acidified water
supplies. Persons drinking the water could suffer from lead or copper poisoning.
The drainage basins, soils, and wet and dry deposition of acidic substances
link aquatic and terrestrial ecosystems. Effects of acidic precipitation on
soils may indirectly influence plant productivity by altering the supply and
availability of soil nutrients. Increased additions of hydrogen ions may
1-81
-------
result in a gradual acidification of the soil. Soil acidification increases
leaching of plant nutrients (such as calcium, magnesium, potassium, iron, and
manganese) and increases the rate of weathering of most minerals. It also
makes phosphorous less available to plants. Acidification also decreases the
rate of many soil microbiological processes such as nitrogen fixation and the
breakdown of organic matter. Various processes important in nutrient cycling
and critical in most ecosystems are known to be inhibited by increasing the
soil acidity. Included among these processes are: nitrogen fixation by
Rhizobium bacteria on legumes and by the free-living Azotobacter; mineralization
of nitrogen from forest litter; nitrification of ammonium compounds; and
overall decay rates of forest floor litter.
Acidic precipitation increases the solubility and mobility of many cations
in the soil, thus increasing the concentration of trace metal cations such as
aluminum, manganese, and zinc to toxic concentrations in soil solutions.
Solubility and mobility of other heavy metals is also enhanced. These toxic
or nutrient ions leached from the soil are transferred into surface and ground
waters from which they may enter lakes or streams and drinking water. Plant
nutrients leached in the same way are lost to vegetation.
Large quantities of hydrogen ions are added to soils as acidic precipitation
and, also, as a result of soil amendment and fertilization practices. Acidifica-
tion by these processes can be readily controlled through normal soil management
practices such as liming. Large areas of the United States, however, are not
cultivated and have soils that are poorly buffered. These soils are sus-
ceptible to further acidification. Many of these soils occur in forest and
wilderness areas. Some of these soils could benefit from significant quantities
of plant nutrients, including nitrogen and sulfur being added to soil in
precipitation and by dry deposition. In some ecosystems these additions may
1-82
-------
be important in the overall nutrient budgets; however, the additions are
subject to the vagaries of wind and weather. Some substances, such as ammonium
and sulfate ions, are acidifying in their effects when taken up by plants from
the soil.
Various specific biological effects of simulated acidic rain have been
demonstrated in controlled field and laboratory experiments. But reliable
evidence of economic damage to agricultural crops, forests, and other natural
vegetation and to biological processes in soil by naturally occurring acidic
precipitation have been very rarely reported.
Dry deposition of toxic gases, aerosols, and particulate matter causes
substantial damage to crops in certain regions of the United States. The
possible effects of acid deposition must be considered together with the
serious economic crop damage caused by sulfur dioxide, ozone, oxides of nitrogen,
fluoride, and hydrogen chloride.
Direct and indirect injury to crops and forests has been reported based
on laboratory, greenhouse, and field experiments in which simulated acidic
rains were used, the following biological effects were observed:
o Formation of necrotic lesions and spots on foliage.
o Accelerated erosion of waxes on leaf surfaces.
o Loss of nutrients due to leaching from exposed plant surfaces.
o Inhibition of nodulation of legumes leading to decreased nitrogen
fixation by symbiotic bacteria.
o Reduced yield of marketable crops.
o Reduced rates of leaf litter decomposition leading to decreased
mineralization of organically-bound nutrients.
1-83
-------
Soils differ by orders of magnitude in their susceptibility to acidifi-
cation. Acidic additions are unlikely to damage calcareous (calcium carbonate)
soils, but metal deposition may. Soils with very low cation exchange capacities
are very susceptible to increased acidification. In addition, the consequences
of acidic additions to soils vary greatly. Such variations depend upon the
rates and recent history of the acidic additions, the character of the vegetation,
natural rates of acid formation in the soil, and the physical-chemical properties
of the parent material of the soil.
Acidic precipitation plays an important role in the deterioration of
stone buildings, monuments, and a variety of materials. Stone has traditionally
been considered one of the most durable building materials used by man. What
is forgotten is that the structures built with stone which were not durable
have long since disappeared.
High acidity promotes corrosion of metals because hydrogen ions act as a
sink for the electrons liberated during the critical corrosion process.
Acidic precipitation forms a layer of moisture on the surface of material and
by adding hydrogen and sulfate ions increases corrosion. Atmospheric sulfur
compounds react with the carbonates in limestone and dolomites, calcareous
sandstone, and mortars to form calcium sulfate. Blistering, scaling, and loss
of surface cohesion occurs, which in turn induces similar effects in neighboring
materials not in themselves subject to direct attack. Acid rain may also
leach ions from stonework just as acidic runoff and ground water leaches ions
from soil bedrock.
1.4.3 Effects on Visibility and Climate
Pollutants released into the atmosphere alter the environment in several
ways, such as by reducing visibility and affecting climate. Visibility refers
to various characteristics of the optical environment such as clarity and
1-84
-------
trueness of color, as well as the distance over which one can see distant
objects. Climate is defined as the long-term manifestation of weather at a
given location over a specified period, usually several decades.
Meteorologically, visibility refers to the greatest distance at which a
bTack object can be distinguished against the horizon sky. Visibility and
visual range are reduced by atmospheric particles which scatter and absorb
light. Gases play a relatively minor role in visibility reduction. Data on
scattering and absorption are used to determine the extinction coefficient, a
measure of the visible range.
Light scattering by particles, especially fine particles, is the most
important cause of lowered visibility. Particles with diameters similar to
the wave-length of light (0.1 to 1.0 urn) are the most efficient light scatterers
per unit mass. Sulfates generally fall within this size range. An aerosol
composed of particles of 0.5 urn diameter scatters about a billion times more
light than does the same mass of air.
Sulfates reduce visibility on both local and regional scales. Considerable
evidence from chemical-mass balance methods indicates that sulfates, which
constitute approximately 50 percent of the fine aerosol mass in the atmosphere,
cause more visibility degradation than do other chemical species. Ambient
sulfate concentrations and sulfur oxides emission trends from the early and
middle 1960s to the early 1970s closely parallel visibility trends. A 30-year
record of spatial and temporal trends of coal use suggests that increases in
haziness have been associated with increases in sulfur oxides emissions since
the 1950s.
Visibility or visual air quality can be measured by determining total
extinction, which is the sum of light scattering, and absorption. In many
cases, absorption is assumed to be small and total extinction is estimated by
1-85
-------
light scattering. Visibility measurement approaches include: the observe
method, in which a black object is viewed against the horizon; the contrast
telephotometric method, in which the brightness of a black object is compared
with the brightness of the horizon; and the long-path extinction method, in
which the decrease in intensity of a beam of light is measured as a function
of range. Indirect methods to measure visibility include scattering and
absorption measurements. The nephelometric method measures the scattering
component of extinction. Absorption can be measured in several ways, such as
by determining the difference between extinction and scattering; however, no
single method has been proven totally effective.
Pollutants released into the atmosphere may lead to slow and subtle
changes in atmospheric composition and, possibly, climate. For example, a
fraction of the solar radiation may be absorbed by aerosols, further reducing
the amount of radiation reaching the earth's surface and, at the same time,
heating the aerosol layer itself. On a hazy day, direct solar radiation is
reduced to about one-half of that on a clear day. Most of this light is
diffused as skylight, while there is an overall loss of up to about 10 to 20
percent of the radiation reaching the surface. Changes in certain measures of
daylight in the Eastern U.S. are consistent with observed patterns of haziness
and can be attributed to man-made fine particles, including sulfates and
nitrates. A possible link between haziness, decreased solar radiation, and
decreased surface temperature in the East Central U.S. has been pointed out
recently.
Atmospheric aerosols, primarily those with strong water affinity, influence
cloud formation. Essentially all water vapor condensation occurs via nucleation,
i.e., by deposition on cloud condensation nuclei or ice nuclei. Both types of
nuclei are aerosol particles emitted from natural or manmade sources. Their
1-86
-------
quantity and nucleation properties have an effect on cloudiness, which in turn
influences the amount of solar energy reaching the earth's surface. In urban
areas, increases in cloudiness and the quantity of precipitation are well
established. The incorporation of particles into cloud and fog droplets can
change the quality and quantity of precipitation by altering its chemical
composition. Effects on visibility and climate of the type summarized above
may exert significant impacts on aesthetic and economic values.
1.4.4 Materials Damage and Soiling Effects
Atmospheric sulfur oxides and particulate matter damage materials through
deterioration and soiling. Physical and economic damage functions have been
developed to estimate materials damage from sulfur oxides and particulate
matter. Their accuracy is hampered by problems in identifying dose-response
relationships for specific damage from specific pollutants because of many
variables influencing exposure in the environment. Damage functions indicate
that reductions in SO- and particulate matter will decrease economic damage.
In most cases, the cost of replacing a product that has suffered premature
damage is far greater than the cost of using protective measures or alternative
materials resistant to damage.
A principal deleterious effect of sulfur oxides is to accelerate the
corrosion of metals to form metal sulfates. Soluble sulfates in rust can
stimulate further corrosion because of their hygroscopicity and electrical
conductivity, while insoluble sulfates in rust provide corrosion protective
properties.
Laboratory studies show that corrosion is most severe under conditions of
high SO- concentration and high humidity. Field studies show that corrosion
rates are related to the amount of sulfur compounds on exposed surfaces of
susceptible metals.
1-87
-------
Many nonmetallic materials are also damaged by sulfur oxides and/or
soiled by particles. These materials include paints and other protective
coatings, fabrics, building materials, electrical components, paper, leather,
plastics, and works of art.
The chemical action of SO- erodes paint layers; light and ozone cause
degradation of the polymer. Paint films permeable to water may be penetrated
by sulfur dioxide and aerosols containing sulfuric acid. SO- can sensitize
dried paint film, permitting water to be absorbed during the weathering cycle,
especially at high humidity.
Cotton, rayon, and nylon fabrics are damaged by acids derived from SO-
Polyester, acrylic, and polypropylene fibers are damaged—by—ammefmnn—s-tri-fa-te-
.-particles by acid hydrolysis-
Certain types of building stone adsorb S0_ and undergo chemical changes
that weaken the material and lead to erosion. Concrete reacts with SO- and
suffers erosion and spelling if not protected by paint. Concrete is also
subject to chemical damage by sodium sulfate. The action of SO- has been
implicated in the deterioration of ancient buildings. Sulfate damage has been
found in medieval stained glass windows, bronze sculptures, marble and stone
statues, and fresco paintings on lime plaster.
Sulfur dioxide and particles have deleterious effects on electrical
contacts. To reduce damage, contacts are electroplated with corrosion-resistant
metals.
Sulfur dioxide is readily absorbed by paper and oxidized to sulfuric acid
by metallic impurities; the paper then hydrolyzes and loses strength. Leather
also has a high capacity for absorption of SO • the material is weakened by
hydrolysis of the proteins that make up the collagens in leather. The weathering
of plastics has been attributed in part to the joint action of SO- and light.
1-88
-------
Airborne particles contribute to corrosion by producing acid electrolytes;
by functioning as nuclei to promote the condensation of water containing SO-
or sulfates; and by forming a solid structure to retain active pollutants such
as chlorides, organic matter, and sulfates. Deposition of dust and soot on
building materials reduces the esthetic appeal of structures and can also
result in erosion and direct chemical attack. Fabrics soiled by airborne
particles require more frequent cleaning, which leads to increased costs and
reduces their life.
Exterior paints are soiled by particles of soot, tarry acids, and various
other constituents such as sulfates of iron, copper, calcium, and zinc. Staining
and pitting of auto finishes have been traced to iron particles from nearby
industrial operations and to alkali mortar dust from buildings being demolished.
It has been suggested that particles promote the chemical deterioration of
paint by acting as wicks to transfer SO- to the underlying surface. Acid
smut, emitted mainly from large industrial operations such as oil-fired boilers,
has been shown to cause significant damage to auto finishes, paints, and
fabrics in the area near the source. The effect is localized because the
emitted particles are very large and tend to be deposited quickly. Damage may
be severe because the smut, which may contain as much as 30 percent sulfuric
acid, is highly corrosive.
A number of investigators have produced estimates of the economic costs
of materials damage. A significant economic cost has been attributed to this
damage, although the estimates produced vary greatly.
1-89
-------
1.5 HEALTH EFFECTS ASPECTS
In the first three chapters (Chapters 11, 12, and 13) of the fourth
volume, information is assessed regarding the uptake, deposition, and
absorption of sulfur oxides and particulate matter and various health effects
demonstrated to be associated with these pollutants by means of animal toxicology
and human clinical studies. Such studies offer the advantage of being able to
study biological processes specifically associated with particular pollutant
exposures under highly controlled laboratory conditions.
The animal toxicology studies are particularly valuable in providing both
qualitative characterization of the full ranges of health effects caused in
mammalian species by SOp and particulate matter exposures and information on
the mechanisms of action underlying such effects. However, considerable
caution must be applied in extrapolating quantitative dose-effect relationships
defined in animal studies to humans.
Of course, some such definition of quantitative dose-effect relationships
can be more directly ascertained by means of human clinical studies. Such
studies, however, are also somewhat limited, in terms of the kinds of health
effects potentially characterized by them. More specifically, only the effects
of short-term (a few hours) exposures or perhaps a few repeated short exposures
are typically investigated in such studies. Also, the nature of the effects
studied are generally limited to detection of onset of relatively transient
changes in pulmonary or cardiac functions and, at times, related physiological
or biochemical parameters. In addition, restrictions arising from human rights
considerations often result in limitations that preclude thorough investigation
of health effects experienced by the most sensitive members of the population.
1-90
-------
Community health (epidemiology) studies offer several Advantages that go
beyond what can be determined by animal toxicology or human clinical studies,
in that health effects of both short- and long-term pollutant exposures (including
the presence of other pollutants) can be studied and sensitive members of
populations at special risk for particular effects identified. In addition,
epidemiology evaluations are not limited to the study of more or less transient
physiological or biochemical effects but also include investigation of both
acute and chronic disease effects induced by S0x and particulate matter pollution
and associated human mortality as well. Information from epidemiology studies,
then, together with the results from animal and human clinical studies, help
to provide more complete understanding of the health effects of environmental
air pollutants such as sulfur oxides and particulate matter.
1.5.1 Respiratory Tract Deposition and Biological Fate
Airborne particles and sulfur dioxide are deposited in the various regions
of human and animal respiratory tracts. Particles are deposited by gravitational
settling, impaction, diffusion, interception, and electrostatic attraction.
Gases are deposited due to convective and diffusional processes. Chemical and
physical properties of particles and SO^, respiratory tract anatomy, and
airflow patterns during respiration also influence deposition. In addition,
deposition is influenced by individual respiration and anatomical features,
which can vary considerably. Nevertheless, the state of knowledge concerning
the deposition of inhaled aerosols and gases is sufficient to predict regional
deposition.
The respiratory tract is usually described in terms of three functional
regions: (1) nasopharyngeal, (2) tracheobronchial, and (3) pulmonary. The
deposition of inhaled material depends on its aerodynamic properties. Material
1-91
-------
soluble in body fluids will readily enter the blood stream. Hygroscopic and
deliquescent solid particles can grow in the humid respiratory tract, depositing
in greater proportion and in larger airways than can insoluble or hydrophobic
particles.
Relatively insoluble material deposits in various regions according to
particle size. If it lands on ciliated epithelium, either in the nasopharyngeal
or tracheobronchial airways, it will be moved by mucous flow to the throat and
be swallowed. Relatively insoluble material deposited on nonciliated surfaces
in the pulmonary regions may be phagocytized by alveolar macrophages, may
enter the interstitium and remain in the lung for an extended period, or may
be translocated via lymphatic drainage. Some material from the pulmonary
region may enter the tracheobronchial region and be cleared via the mucociliary
conveyor. Very insoluble particles deposited in the pulmonary region may
remain there for months or years, while similar particles deposited in the
tracheobronchial or nasopharyngeal regions are cleared in a few days.
Sulfur dioxide may be deposited directly in the airways or enter into a
variety of gas-particle chemical and physical reactions. S0? may dissolve in
liquid droplet aerosols or hygroscopic and deliquescent particles and thereby
lead to increased deposition deep in the respiratory tract. The aerodynamic
properties of particles and the route of breathing affect deposition.
A nose-breathing person taking 15 breaths per minute with a tidal volume
of 1450 ml would have deposited in the deep lung: 35% of the 0.2 pm particles
inhaled; 25% of the 1 pm particles inhaled; 10% of the 5 urn particles inhaled;
and almost none of the 10 urn particles inhaled. During mouth breathing these
deposition percentages would be expected to be about 35 percent, 30 percent,
1-92
-------
55 percent and 10 percent, respectively. This difference shows that pulmonary
deposition of large particles is greater during mouth breathing than during
nose breathing. Some particles that normally are collected in the nasopharyngeal
region during nose breathing may pass the glottis and be deposited in the
upper part of, the tracheobronchial tree during mouth breathing. One study- .<
'/TO lo *
that about 56- percent of 15 urn unit density spheres may enter the
tracheobronchial tree during mouth breathing, while almost none of these
particles enter the tracheobronchial tree during nose breathing.
Since many biological studies of inhaled aerosols and SOp involve experi-
mental animals, it is important to understand how humans and animals differ
with respect to pulmonary deposition and clearance. Small rodents tend to
have smaller lung deposition fractions of particles less than 3 urn D and
QI
somewhat higher pulmonary clearance rates than humans. In dogs, however,
particle deposition and clearance parallel that in humans. Mouth-breathing
people bypass most of the nasal filtration of S02 and have much higher lung
exposure than nose-breathing experimental animals. This may be a very important
factor for humans under heavy work or exercise conditions. The three functional
regions of the respiratory tract can each be characterized by mechanisms of
deposition, clearance, and potential biological responses. The respiratory
tract is itself a target of inhaled particles and gases, and it is also the
portal of entry to other organs that may be affected.
1.5.2 Animal Toxicology Studies
Although major gaps exist in the animal toxicological data base on sulfur
oxides (SO ) and particulate matter (PM), such data can be useful in:
y\
(1) delineating the full range of toxicological effects of SQX and
PM, including the effects of both very high-level exposures and
prolonged low-level exposures;
1-93
-------
(2) elucidating potential mechanisms of toxicity and defining
structure-function relationships between physico-chemical
properties of these agents and particular health effects;
(3) supporting the findings of human clinical and epidemiological
studies investigating analogous health endpoints; and
(4) investigating of health effects (e.g., potential mutagenesis or
carcinogenesis effects), which neither can be experimentally
induced in human clinical studies nor easily or precisely
assessed with current epidemiology methods.
The use of the toxicological data base from animal studies in order to
make comparisons between man and animals involves assumptions that a chemical
has ultimate mechanisms of toxicity involving chemical and biological structures
which are similar in man and animals. Although the precise mechanisms underlying
most effects are currently unknown, it is typically acknowledged that common
biochemical events are probably involved for both man and other mammalian
animal species. For example, if a pollutant is observed to cause toxicity by
destroying a chemical structure essential for normal activity of a cell membrane
in one mammalian species, then that cell structure is likely at risk in other
mammalian species and humans. The critical issue then becomes delivery of
the pollutant to that structure in different animal species and human beings.
Thus, qualitative extrapolation of the type of effect from animal to man has a
theoretical basis; and that basis becomes even stronger if the same effect is
observed in a number of animal species, since species differences in susceptibility
can exist. As will be discussed later, certain experimental evidence indicates
the occurrence of similar effects in man and animals do occur in cases where
the same biological endpoint has been examined.
1-94
-------
The current animal toxicological data base on the effects of participate
matter and sulfur oxides is relatively limited. In addition, sulfuric acid
^K^J-<^
studies which used whole body exposure chambers a&e- confounded by an ammonia-
induced conversion of I^SO^. Since such an event would not occur to an
equivalent degree in human exposure chambers, direct quantitative comparison
of animal and human HLSO. exposure effects would not be accurate, and failure
to find an effect in animal studies does not rule out its possible occurence
in humans in the absence of ammonia*or other agents more often fetmd—rn^anrnraHH
Hv-JB9.-ABd--.testing conditions-. On the other hand, many animal studies which
have been conducted with SO^, (NH.^SO., and sulfuric acid used very high
concentrations relative to ambient air and with exposures of short duration
(less than 1 day) and the direct relevance of such exposure conditions to
assessment of quantitative dose-effect/dose-response relationships for humans
can be questioned. Turning to a summary of animal study results, note that
certain studies have looked at the effects of relatively high levels of SOp,
particles, or aerosols on pulmonary system morphology or function in various
animal species. Some animal studies have also examined the effects of such
exposures on immune system functions and susceptibility to bacterial or viral
infections. Certain key findings are summarized below.
1.5.2.1 Effects of Acute and Chronic Exposure to Particles or S0,,--The results
of most of the animal studies discussed in Chapter 12 regarding the,effects of
(Jut~CJK£**J*fiUsX^> ^ I.TK^t/s
acute and chronic exposure to various particles or SO^ alone/\jan^summarized in
Tables 1-8, 1-9, and 1-10. Key findings and their possible implications for
human health assessment are discussed in what follows.
1-95
-------
TABLE 1-8. SUMMARY OF EFFECTS OF ACUTE EXPOSURE TO <1 mg/m3 PARTICLES3 IN ANIMALS
Concentration
0.05 mg/m3 CdCl2
0.1 mg/m3 NiCl2
0.1 mg/m3 CdCl2 or
0.2 mg/m3 CdS04
0.1 mg/m3 H2S04
0.19-1.4 mg/m3 H2S04
0.19 mg/m3 CdCl2
0.25 mg/m3 NiCl2
0.25 mg/m3
ZnS04, (NH4)2 S04
0.4-2.1 mg/m3
(NH4)2S04
0.43 mg/m3
CuS04
0.5 mg/m3
(NH4)2S04
0.5 or 1.0 mg/m3
Duration
2 hr
2 hr
2 hr (CdCl2)
3 hr (CdS04)
1 hr
1 hr
2 hr
2 hr
1 hr
1 hr
1 hr
1 hr
1 hr
Species
Hamster
Hamster
Mouse
Guinea pig
Donkey
Mouse
Mouse
Guinea pig
Donkey
Guinea pig
Guinea pig
Dog
Results
Decreased ciliary beat frequency in trachea.
Decreased ciliary beating frequency in trachea.
Increased susceptibility to streptococcal lung infection.
41% increase in flow resistance and 27% decrease in
compliance. (Higher concentrations did not cause these
effects in a similar study, Reference 53).
Bronchial mucociliary clearance was slowed.
Decreased number of antibody-producing spleen cells.
Decreased number of antibody-producing spleen cells
22% increase in flow resistance
No change in pulmonary resistance or dynamic compliance
No significant change in flow resistance. 11%
increase in compliance.
23% increase in flow resistance and 27% decrease in
compliance.
At the lower concentration, tracheal mucociliary transport
Reference
Adalis et al.157
Adalis et al.156
Gardner et al. ,7Q
Ehrlich et al.
179
Amdur «t,al. "~
Amdur1 /J
Schlesinger et al.
Graham et al.160
Graham et al.160
123
Amdur and Corn
Amdur J
999
Schlesinger et al.
Amdur et a I.130
Amdur et al.130
Wolff et al.224
0.5 mg/m3 NiCl2
2 hr
Mouse
was accelerated, but 1 wk later, it was depressed. At
1 mg/m3, the rate was depressed, even at 1 wk post-
exposure.
Increased susceptibility to streptococcal lung
infection.
Adkins et al.155
-------
TABLE 1-8 (continued).
t— *
1
UD
•-J
Concentration
0.6 mg/m3 CuS04
0.8-1.51 mg/m3
H2S04
0.9 mg/m3
H2S04
0.9 mg/m3
H2S04
0.91 mg/m3
ZnS04
0.93 mg/m3
NH4HS04
a
Duration
3 hr
1 hr
2 hr
2 hr
1 hr
1 hr
Species
Mouse
Donkey
Hamster
Mouse
Guinea pig
Guinea pig
Results
Increased susceptibility to streptococcal lung infection
No change in pulmonary resistance or dynamic compliance.
Decreased tracheal ciliary beat frequency.
No effect on susceptibility to infectious bacterial
pulmonary disease.
4158 increase in flow resistance.
15% increase in flow resistance and 15% decrease in
compliance
•
Reference
Ehrlich et al.
Schlesinger et
-ip-
Grose et al.
Schiff et al.
Gardner *t al.
Amdur and Corn'
AmdurI/0
Amdur et al.13(
179
222
182
145
123
this issue completely. Data are presented for lowest effective concentration tested. See the text and tables
of Chapter 12 for details. (See Table 12-8 for more complete details on effects of particles on airway resistance
in guinea pigs.)
-------
TABLE 1-9. SUMMARY OF EFFECTS OF CHRONIC EXPOSURE TO <1 mg/m3 PARTICLES3 IN ANNALS
Concentration
0.08, 0.1 mg/m3
H2S04
0.38, 0.48 mg/m3
H2S04
0.89 mg/m3
H2S04
0.1 mg/m3
H2S04
0.01, 0.15 mg/m3
Pb203; 0.01 mg/m3
PbCl2, 0.11 mg/m3
NiCl2, 0.12 mg/m3
NiO
Duration Species
52 wk, Guinea pig
continuous
78 wk, Monkey
continuous
21 hr/day, Dog
225 or 620
days
1 hr/day, Donkey
5 days/wk,
several mo
3 mo con- Rat
tinuous for
Pb203; 12
hr/day, 6
days/wk, 2 mo
Results
No effects on hematology, pulmonary function.
No hematological changes. At 0.38 mg/m3 there were
bronchiolar epithelial hyperplasia and thickening
of the respiratory bronchioles. Particle size
influenced the effects. Exposure to 0.48 mg/m3 altered
distribution of ventilation early in the exposure
period. Respiratory rate increased at 0.38 mg/m3.
Other functional parameters were not affected.
No morphological changes at 620 days. At 225 days,
CO diffusing capacity was decreased. At 620 days,
CO diffusing capacity was decreased and other pulmonary
function measurements were affected.
After 4 wk erratic bronchial mucociliary rates were
observed. During the second 3 mo of exposure, clearance
was slowed in animals never pre-exposed.
Decreased number of alveolar macrophages after Pb203.
NiO increased the number of alveolar macrophages.
The soluble metals did not change the number of
alveolar macrophages.
Reference
Q? 197
Alarie et al. >iy/
1Q7
Alarie et al .
Lewis et al.89'104
223
Schlesinger et al .
Bingham et al.152'153
aThe toxicity of particles is dependent on particle size. For simplicity, this table does not address this issue
compiletely. Data are presented for lowest effective concentration tested. See the text and tables of Chapter 12
for details.
-------
11
TABLE 1-10. SUMMARY OF EFFECTS OF EXPOSURE TO <13.1 mg/m3 (5 ppm) SULFUR DIOXIDE IN ANIMALS
i
10
V£>
Concentration
0.26 or 2.62 mg/m3
(0.1 or 1 ppm)
0.37, 1.7 or
3.35 mg/m3 (0.14,
0.64, or 1.28 ppm)
0.42 or 0.84 mg/m3
(0.16 or 0.32 ppm)
2.62, 5.24, 13.1 mg/m3
(1, 2, or 5 ppm)
2.62 mg/m3
(1 ppm)
2.62 or 13.1 mg/m3
(1 or 5 ppm)
9.43 mg/m3
(3.6 ppm)
13.1 mg/m3 (5 ppm)
Duration
7 hr/day,
5 days/wk,
25 days
78 wk,
continuous
1 hr
1 hr
1.5 hr/day
24 hr
7 days,
continuous
6 hr/day,
20 days
Species
Rat
Monkey
Guinea pig
Dog
Dog
Rat
Mouse
Guinea pig
Results
0.26 mg/m3 (0.1 ppm) accelerated tracheobronchial
clearance at day 10 and 23. The higher concentration
accelerated clearance at day 10, but at day 25,
clearance was decreased.
No effects on pulmonary morphology or morphology.
Increase in flow resistance.
-
Increased bronchial reactivity to aerosols of a
bronchoconstrictor agent (acetylcholine).
Decreased mucous flow.
Phagocytosis of alveolar macrophages increased after
3 or 4 days in culture.
Exposure to S02, whether alone or in combination with
a virus, produced weight loss.
No change in bacterial clearance.
Reference
Ferin and Leach
on QI
Alarie et al. u>*1
Amdur and Underbill
Amdur et al.
Islam et al.102
Hirsch et al.111
iqc
Katz and Laskfn133
Lebowitz and.
Fairchild
Ry lander ,«-
Rylander et al.
13.1 mg/m3 (5 ppm)
3 hr/day, Mouse
1-15 days and
24 hr/day,
1-3 mo
No change in mortality due to a laboratory-induced
streptococcal infection.
Ehrlich
178
-------
As indicated by the results summarized- in Tables
.
d-jrlfr virtually *
i£*£ JZtf&sv&si*-' J&*feaJ
all particles examined to date cause health effects^ TWy |nd^vei^ty ,of^
the effects are chemical-specific and concentration dependelit/t For woclasses /
of effects (histamine release and increased susceptibility to infectious
dfsease), the cation of a given particle species was found to have more influence
on the toxicity than the anion. For example, in regard to increasing susceptibility
to infection, CdSO. was more toxic than ZnSO., while NaSO. and (NH. )„$(). had
no significant effect even at high concentrations. Guinea pigs experienced
approximately a 40% increase in airway resistance with exposure either to 0.1
mg/m HpS04 or 0.5 mg/m -ZnSO. and- (Nlh)2SO^. The ranking of toxicological
potency varies with site of deposition or physiological process. For example,
the ranking of sul fates for airway resistance does not precisely agree with
that for effects on susceptibility to pulmonary infection. These findings
illustrate the complexity of toxicological responses and show the need for a
broad data base.
The size of the particle plays a role in the health effects observed.
Unfortunately, no known experimental data compare the effects of particles
which would predominately be deposited in the head vs. the respiratory tract.
Studies generally have been conducted with particles that tend to enter the
deep portions of the lung where gas-exchange occurs. In studies employing
^SO^ or ZnSO. and (NH.KSO. particles of various sizes that would predominantly
deposit in the gaseous exchange region, it was found that pulmonary flow
resistance increased as particle size decreased.
As for particulate matter dose-effect relationships, in contrast to
results discussed below for SOp alone, acute (1-2 hr) exposure of the several
different animal species listed to much lower levels of different particulate
1-100
-------
substances have been shown to lead to various -pafcfcephysiological changes,
e.g., slowed mucociliary clearance, increased flow resistance and compliance,
^anjLJncr4taj>jiU-Jll^F-reyrstflftfHt- and comp44tmce, and increased susceptibility to
an infectious bacterial agent. Such effects occurred with exposures to such
pa~rticulate substances as H2$04, ZnS04, CuS04, and Ni and CdCl2 at levels as
3
low as 100-200 (jg/m . Also chronic exposures of various animal species to
sulfate aerosols, as shown on Table 1-8, results in such effects as bronchiolar
epithelial hyperplasia, slowed mucociliary clearance rates, and decreased CO
diffusing capacity. Certain of these changes occur at chronic exposure levels
3
of 100 to 500 |jg/m of sulfate aerosol. Other effects, e.g., altered numbers
of alveolar macrophages, have also been observed with chronic exposures to
lead and nickel oxide particle concentrations of around 100 ug/m .
Some, but not all, animal and controlled human studies have shown that
3
acute exposure to high (>1 ppm; 2860 ug/m ) concentrations of S0? increases
airway resistance. At lower concentrations, only in animal studies on the
guinea pig, has increased airway resistance been observed, i.e., at 420 to 840
o
ug/m . It appears, however, that the mechanism of 50,,-induced flow resistance
is likely similar in man and animals, since atropine inhibits the response in
both types of subjects. This suggests that the effect of SOp is mediated by
parasympathetic motor pathways which alter airway smooth muscle tone.
In contrast to the above, human and animal studies of effects of H2S04 on
pulmonary function, have yielded variable results. Certain concentrations of
H^SO. increased resistance in some guinea pig studies, but not in human studies.
In guinea pigs, but not in humans, H2$04 is far more potent than S02 in increasing
resistance. In guinea pigs, salt (NaCl) potentiates the response to S02. In
some human studies, a similar potentiation occurred. If the hypothesis of
1-101
-------
H2S04 formation as a possible explanation were true, then H2$04 can likely
affect human airway resistance. However, in other human studies, SO2 plus
NaCl was not found to result in potentiation.
Some possible reasons for this difference in H2$04 action between man and
animals can be offered. Perhaps the mechanism for H^SO^ has a species
specificity. In animals histamine release is one of the hypothesized mechanism.
Man also has histamine stores and reacts to histamine release. Nonetheless,
the detailed metabolism and action of histamine in man and animals is not fully
known. Another possible difference between human and animal studies is the
extent neutralization of KUSO. by ammonia in exposure chambers, with much
higher ammonia levels present in animal chambers from the urine of the test
subjects. Both human and animal S02 studies, however, have shown approximately
a 10 percent incidence of more responsive individuals. It could be possible
that the positive animal studies had an unusually high incidence of these
"sensitives" and the humans an unusually low incidence of "sensitives".
Another possibility is that the negative guinea pig studies more closely
reflect the human condition, and neither the resistance of man nor animals is
highly affected by H2S04-
Mucociliary clearance has also been investigated. In dogs, rats, and
man, short-term exposure to lower concentrations of S0? and H?S04 generally
accelerated bronchial clearance of particles, while higher concentrations
slowed clearance. In donkeys, single or repeated exposures to low concen-
trations of H2S04 appeared to slow bronchial clearance. Such findings would
imply that these chemicals might increase the residence times of substances in
the lungs that are normally removed faster. This slowed clearance could
influence susceptibility to infectious disease. Guinea pigs exposed to S02
experienced an increase in laboratory-induced viral pneumonia. Some animal
1-102
-------
studies indicate that acute exposure to a variety of sulfates increases
susceptibility to infectious (bacterial) pulmonary disease in mice, probably
by affecting alveolar macrophages and additional host defense mechanisms.
Metal sulfates caused these effects, but much higher concentrations of HpSO.,
(NH4)2S04, and S02 did not.
1.5.2.2 Effects of Exposures to Combinations of SO and Particles—Some
investigations with animals have focused on the effects of S02 in combination
with particles. The results of such studies are summarized in Table 1-11.
Long-term exposure of mice to carbon plus S02 or carbon alone caused
alterations in the immune system which were more extensive than those caused
by SOp alone. Combinations of H^SO. and carbon resulted in greater morphological
effects in the lungs of mice than exposure to carbon alone. Similar combination
exposures caused more effects on tracheal tissue than did carbon or H2SO.
alone. In mice exposed acutely to 0- and then to H^SO., an additive effect
was observed on susceptibility to infectious respiratory disease. However,
this same regimen resulted in antagonistic effects on reduction of tracheal
ciliary beating frequency. For long-term exposure, the effects of 03 plus
H2S04 on lung morphology species were attributed to 0, alone. The mechanisms
for the toxicological interactions described above are not known; the effects
of ammonia neutralization of H2$0. during long exposure studies, however,
cannot be ruled out.
A variety of other chronic studies have been conducted with single chemicals
3
and pollutant combinations. At concentrations below 13.4 mg/m (5.12 ppm) S02
for up to 18 months, there were no morphological or pulmonary functional
changes in monkeys. A similar level of S02 caused increased pulmonary flow
resistance and decreased lung compliance in dogs after 225 days of exposure.
1-103
-------
TABLE 1-11. SUMMARY OF EFFECTS OF COMBINATIONS OF PARTICLES3 (<1 mg/m3) AND SULFUR DIOXIDE
413.1 mg/m3, 5 ppm) IN ANIMALS
Concentration
0.79-0.84 mg/m3
(0.3-0.32 ppm) S02
+0.9 mg/m3
(NH4)2S04, NH4HS04,
or Na2S04
0.94 mg/m3
(0.36 ppm) S02
+0.4 mg/m3 CuS04
(— •
^ 2.62 mg/m3 (1 ppm)
§ S02 + 1 mg/m3 Nad
2.62 mg/m3 (1 ppm)
S02 + aerosols of
metal salts
2. 6 mg/m3 (1 ppm)
S02 + 0.9 mg/m3
H2S04
0.08 mg/m3 H2S04 +
0.45 mg/m3 fly ash
0.28, 2.62, or 13.1
mg/m3 (0.11, 1, or
5 ppm) S02 + 0.56
mg/m3 fly ash
5.24 mg/m3 (2 ppm)
S02 +0.56 mg/m3
carbon
Duration
1 hr
1 hr
1 hr
1 hr
18 mo,
continuous
12 mo,
continuous
78 wk,
continuous
52 wk,
continuous
100 hr/wk,
192 days
Species
Guinea pig
Guinea pig
Guinea pig
Guinea pig
Monkey
Guinea pig
Monkey
Guinea pig
Mouse
Results Reference
130
Additive effect on increased flow resistance. Amdur et al.
Potentiation of increased flow resistance Amdur et al.
At 80% relative humidity, NaCl potentiated the response McJilton et al.
to S02 flow (increased flow resistance).
96
Manganous chloride, ferrous sulfate and sodium Amdur and Underhill
orthovanadate potentiated the response to S02
flow (increased flow resistance).
92
No effects on hematology or pulmonary function. Alarie et al.
Morphological changes in bronchial mucosa. Addition of
0.41 mg/m3 fly ash did not alter the results.
92
No changes on hematology, pulmonary function, or Alarie et al.
morphology.
199
No effects on pulmonary function, morphology, or Alarie et al.
hematology.
Alterations in the pulmonary and systemic humoral immune Zarkower
system.
aThe toxicity of particles is dependent on particle
text and tables of Chapter 12.
size which is not presented on this table. For details, see the
-------
TABLE 1-11. continued).
Concentration
Duration
Species
Results
Reference
Various combinations,
including 1.1 mg/m3
(0.42 ppm) S02 +
0.09 mg/m3 H2S04
16 hr/day
68 mo
Dog
No changes in pulmonary function after 18 or 36 mo of
exposure. Some alterations in pulmonary function after
61 mo. Thirty-two to 36 mo after exposure ceased,
morphological effects were observed (i.e., enlarged air
spaces and increased number and size of interalveolar
pores, loss of cilia, nonciliated bronchiolar cell
hyperplasia and loss of interalveolar septa in alveolar
ducts). These changes appear similar to centrilobular
emphysema.
Lewis et al.'
Hyde et al. 1B6
Vaughan et al.
,_, The toxicity of particles is dependent on particle size which is not presented on this table.
For details, see the text and tables of Chapter 12,
o
01
-------
3
Long-term exposure of monkeys to H^SO. (as low as 0.38 mg/m ) caused morpholo-
gical and some pulmonary function changes. A higher concentration (0.89
mg/m ) caused no structural changes in dogs, but pulmonary function decrements
were observed. Guinea pigs were not affected after a one year exposure to 0.1
~3
mg/m H2S04-
Effects were also noted after chronic exposure to pollutant combinations,
but in most cases the contributions of the individual chemical species to the
effects of the mixture are obscure. Although various mixtures of S0?, hLSO.,
and fly ash caused pulmonary morphologic changes, no effects on pulmonary
function were found in monkeys. It appeared that fly ash did not contribute
to the effects. Fly ash, when combined with S02 or H^SO. did not significantly
affect guinea pigs. When dogs received mixtures of SO^ and HpSO. for 620
days, the effects on pulmonary function were attributed to the HLSO^ (0.89
3
mg/m ). Several morphological changes were observed in the lungs of dogs 32
to 36 months after a 68 month exposure to 1.1 mg/m (0.42 ppm) S02 plus 0.09
3
mg/m H^SO. ceased. It was hypothesized that these changes were analagous to
an incipient stage of human proximal acinar (centrilobular) emphysema. It is
not known whether the disease state progressed, abated, or remained stable
over this post-exposure period. A shorter exposure (approximately 18 months)
to a higher concentration of KLSO. (0.9 mg/m ) plus 2.62 mg/m (1 ppm) S0?
caused less serious morphological effects in monkeys. Concentration-time
relationships of effects in the above studies are not clear. But from the
above-mentioned dog and monkey study, it appears that time of exposure is an
important factor in the development of disease. However, concentration does
play a role, as evidenced from the concentration-dependent effects after
either acute or chronic exposure.
1-106
-------
1.5.3 Studies on the Oncogenic Properties of SO and PM
/\
The oncogenic potential of SCL has been studied, but the findings are not
yet conclusive. In one series of experiments, hamsters and rats were exposed
for 98 weeks or a lifetime to various regimens of SOp and benzo(a)pyrene, a
known carcinogen. No lung tumors or other pathological effects were observed
for hamsters. Rats, however, had an increased incidence of tumors after
3 3
exposure to 26.2 mg/m (10 ppm) for 6 hr/day and 9.17 mg/m (3.5 ppm) S02 plus
3
10 mg/m benzo(a)pyrene for 1 hr/day. In a later study, rats exposed for 6
hr/day to 26.2 mg/m3 (10 ppm) S02 and for 1 hr/day to 10.5 mg/m3 (4 ppm) S02
3
plus 10 mg/m benzo(a)pyrene had a cancer (squamous cell carcinoma) incidence
of 19.6 percent. This incidence contrasts with a zero percent incidence in
o
the group receiving air or 26.2 mg/m (10 ppm) S02 for 6 hr/day and a 9 percent
incidence in the group exposed for 1 hr/day to 10.5 mg/m (4 ppm) S0« plus 10
mg/m benzo(a)pyrene. (See Table 12-3 in Chapter 12 for details.) The reports
do not specifically state whether the 26.2 mg/m (10 ppm) SOp was given before,
during, or after the exposure to 10.5 mg/m (4 ppm) SOp plus benzo(a)pyrene.
Thus, it is difficult to speculate on potential prompter/or initiator effects.
In another study, mice were exposed to T310 mg^m (500/ppm) 50^ for a
min/day, 5 days/wk for a lifetime. This exposure increased the carcinoma
incidence in females from 0 to 18 percent. No such change occurred in the
males. Primary pulmonary neoplasias increased in both males and females. The
investigators conclude that the increased incidence of primary lung tumors is
a result of an S0?-induced inflammatory reaction, "followed by a state of
apparent tolerance, which accelerates the inherent tendency of these mice to
develop lung tumor spontaneously but does not justify the classification of
S02 as a chemical carcinogen as generally understood." From the single study
1-107
-------
of S02 alone in mice and the two studies of S02 plus benzo(a)pyrene in rats,
no definitive conclusions can be drawn about the carcinogenic or co-carcinogenic
potential of S02< However, the data reported thus far do justify some concern,
and these issues need to be more extensively addressed in further experimental
research.
It is commonly believed that fundamental similarities exist between the
molecular mechanisms of both mutagenesis and carcinogenesis. This assumption
is based on the theory that a chemical interaction with DMA and/or other
critical cellular macromolecules initiates a genetic change which may lead to
carcinogenic transformation. Therefore, the demonstration of mutagenic activity
for a substance is generally taken as strong presumptive evidence for the
existence of carcinogenic activity. It follows that an investigation of the
mutagenicity of a substance may be predictive of carcinogenic potential, and
may serve as an early warning of a possible threat to human health.
1.5.4 Experimental Investigations of Human Subjects
Studies of human subjects exposed to sulfur oxides and particulate matter
explore physiological and sensory responses to these pollutants under controlled
conditions. A major limitation of such studies is that they do not reflect
the long exposure durations which typically occur in the ambient environment.
However, they may reflect short-term peak exposures, such as may occur in
certain urban situations. The results of human clinical studies discussed in
Chapter 13 are summarized in Tables 1-12 to 1-14 below.
The typical response to S02 exposures is an increase in pulmonary flow
resistance. Most subjects respond to short exposures of 5 ppm (13,000 ug/tn3),
while other subjects react to lower levels of S02- Significant decreases in
nasal mucous flow rate have been demonstrated at 1 ppm; 2620 ug/m . Pulmonary
1-108
-------
1-u. rutMDMMrr EFFECTS of AEMSOLS
Duration of
Concentration exposure (.Ins)
SO. (1.6 - 5 pp.) 5
MCI 0.22 u. MB
SO (9-60 pp.) S
MCI (OB - 0.95 UB)
SO (0.5, 1.0 and 5.0 pp.) IS
Saline panicles 7.0 u.
Ibid 30
SO (1.1 - 3.6 pp.) . 30
NK1 2.0-2.7 M«/m
MB • 0.25 M"
SO (1-2. 4-7. 14-17 pp.) 30
Nici 10-30 mg/m
MB 0.15 UB
SO (1 pp.) 60
NiCl 1 «g/«
MB 0.9 M og « 2.0 M"
Ibid 60
AsmnnluB sulfata 150
100 MO/"
6
Aamonlua blsulfate 150
85 MO/" aerosol s1u
distribution
g,4 jaa_tn»»4ll__^
0. 35-5.0 BB/B3 H-SO. 15
MB 1 M"
3-39 Bo/"3 H-SO. 10-60
MB 1-1.5 M"
SO- (1-60 pp.) plus Variable
H;O- to for. H-SO.
atrosol 2 4
OB 1.8 and 4.6 MB
H.SO. .1st 120
f 1000 MO/.
MB 0.5 M" (00 * 2-59)
H SO aerosol . 10
10, 100. 1000 MB/"
MB 0.1 MB
H-SO. (75 MO/"3) i20
HMD 0.48 - 0.81 M"
H-SO. (0. 100, -300, 60
Or 1,000 MO/"
MMO 0.5 MB
(oo • 1-9)
Number of '
subjects
13
10
9
9
(asthmatics)
10
12
9
(asthmatic!)
(normals)
5 (normal)
4 (ozone
sensitive)
(asthmatics)
16
15
Variable
24
10
6 normal
6 asthmatics
6 normal
6 asthmatics
10
Source
Mask
Mask
Orel
(Mask
(Exercise for
10 .1 mites)
Oral
Oral
Oral
Mask
Chamber
(exercise)
Chamber
(exercise)
Naik (rest)
Hask (rest)
Chamber (rest)
(Rest)
Chamber
(exercise)
Oral
Chamber
(exercise)
Nasal
Effect!
Synerolsttc Increases In
airway resistance with aerosol
Airway resistance greater after
exposure to aerosol than to
exposure to SO. alone
NEF-— significantly greater
dKreeses In aerosol (NaCl)
condition
y y
•flv?0?'ano*tT7oecrease
significantly In aerosol
condition
No effect on pulannary functions
Change! In pulannary function
sl.llar to changes due to SO.
alone not Influenced by aerolol
Significant decreases In V .-_
• Ml V mmmUt MB
max 75X
No pulmonary effects demon
strata*
No changes In pulmonary
functions
No changes In pulmonary
functions
Respiratory rates Increased,
•ax. Imp. and explretory
flow rates and tidal
decreased volu.es
Longer particles due to "wet
•1st" resulted In Increased
flow resistance cough, rales
bronchoconstrlctlon
Airway resistance
Increased especially
with larger particles
No pulmonary function
changet but Increased
trecheobronchlel clearance
No pulmonary function
changes, no alterations
In gas transport
No pulmonary effects
In either group
No pulennary function
effects
Rronclal mucoclllary
clearance t following
100 MB/" »^ut * following
1000 MO/" mucoclllary
clearance distal to trachea
Reference
1 1'
Toyama. 1962
Nakamura. 1964
Snail and Luchilnger.
1969
Koentg et al., 1979
Burton et al., 1969
Frank et al . , 1964
Koenlo. 1979
Koanlg, 1979
•ell and Hackney, 1977;
Klalnman and Hackney, 1978;
Aral et al., 1979
Amour etTl . . 1952
SI. and Pattle, 1957
Toyama and Nakamura,
1964
Newhouse et al., 1978
Sachner et al., 1978
Klalnman and Hackney,
1978; Avol et al., 1979
Ltppmann et el., 1979
-------
TMU 1-13. EFFECTS OF SO,
I
I—>
t—'
O
Pollutant
Concentration (PPM)
.17
.17
.40
.SO
1.0
1.0. S.O
1.0. 1.0, S.O
1. 5. 13
1. S. 15
1.1-1.6
1.0. 5.0, 25.0
S.O
1-23
1.1-80
2.5-50
4-6
1-8
5.0
Duration of
Exposure
120 Minutes
120 Minutes
120 Minutes
180 Minutes
120 Minutes
60 Minutes
120 Minutes
1 Minutes
10- 10 Minute*
30 Minutes
10 Minute*
HP to 6 hr/day
4.5 hrs.
60 Minutes
10 Minutes
10 Minute*
10 Minutes
10 Minutes
120 Minutes
Effects Reference
No pulMonary effects Bates and Huucha. 1973;
Hazucha and lates, 197S
No pulannary effects tall et al.. 1977
No pulMonary effects Bodl et al. , 1979
Horvath and Follnsgea. 1977
HNFR decreased 2.7X; delayed Jaeger, et al.. 1979
•sthM* *
ffj^af^tf^^^^^taU. .„, lllMCta UTJ-
Ho effects observed HcJIltan, 1976
Increase In nasal flow Andersen, et al.. 1974
resistance, decrees* In na**l
MUCUS flow
Light exercls* potentiates effect KralSMan, et al. , 1976
af S02; NFF 40V decreased
No change* In pulse rate. Frank, et al. , 1962
respiratory rate. pulMonary flow
resistance Increased at S and 13
POM but less during nasal breathing; I ^ff)
&T t fl&rt '/ jyti. 4ief^TJ^U4^J~ " fi J^&t**^*- ^*?
Increase In U at S0| «bov«7 Frank et al.. 1964 &3L0-&.
Deep breathing produced n* effect* Burton et al. , 1969
Significant decree*** In expire- Anderson, et •!., 1974
tor* flow and FEV, n. decreased
MUCUS flow
Nuttier ef colds slMllar In both Andersen, et al.. 1977
group* but severity let* In SO.
axposed group* *
BronchoconttrlctloM SIM and Pattle, 1169
BronchocomtrlctlM Sla and fettle. 1957
Increased Insptretory and Aba. 1967
respiratory resistance
Airway conductan* decreased Nadel, 1965
reflex effect
Pulse rate, respiratory rate Amur, et al.. 1951
Increased; tidal voluM* decreased
NNFR decreased 8.SX; Increased Nevnouse. et al. 1978
tricheobronchlal clearance
-------
1-14. PULMONARY EFFECTS OF SO. AND OTHER AIR POLLUTANTS
Concentration
S02 (0.37 ppM)
and
03 (0.37 ppM)
S02 (0.37 ppM)
and
03 (0.37 ppH)
S02 (0.40 ppm)
and
0, (0.40 ppM
S02 (5 ppM)
and
N02 (5 ppM)
S02 (5 ppn.)
M02 (5 ppm)
and
03 (0.1 ppM)
S02 (0.12 ppM)
N02 (0.06 ppM)
and
03 (0.025 pp.)
Duration of Nuaber of
exposure (mlns) subjects Source
120 8 Chamber
(exercise)
120 4 (normal) Chamber
4 (ozone (exercise)
sensitive)
4 (from Bates)
120 9 Chamber
(exercise)
120 11 Chamber
(exercise)
120 11 Chamber
(exercise)
120 11 Chamber
(exercise)
Effects
Decrease pulmonary functions
(In synerglstlc effect of
S02 on 03) FRC, FEV-j^ Q.
MMFR, MEFR_~~
Unable to confirm
synerglstlc effects
pulmonary decrement due
to 0, alone
Unable to confirm
synerglstlc effects
changes due to ozone
alone
No changes 1n P , P
pHa or TGr -Rt °2 C02
Increased
No changes In P , P ,
pHa or TGr -RtC02 °2
Increased .
No changes 1n pulmonary
functions
TT
Reference
Hazucha and
Bates, 1973, 1975
Bell et al., 1977
Horvath and Follnsbee
(1977);
Bedl et al. (1979)
von Nledlng et al., 1979
von Nledlng et al., 1979
von Nledlng et al., 1979
-------
responses to SOp exposure apparently persist for some 5 to 10 minutes, while
the change in mucous flow rates persists for several hours following exposure.
Approximately 10 to 20 percent of individuals studied under controlled conditions
appear to be sensitive to exposures as low as 1 ppm SO^. Effects on certain
sensory functions, (e.g., dark adaptation, odor perception) are evident at S02
concentrations below 1 ppm (2600 ug/m ).
A number of factors influence responses to SO,,. Sulfur dioxide is removed
more efficiently if individuals breathe by nose than by mouth. This removal
is related to the high solubility of SOp in water; most of the SO,, in inhaled
air is absorbed in the moist linings of the nose (and upper airways). When
subjects breathe by mouth, more SOp reaches deeper areas of the lungs. Subjects
who exercise at a level requiring mouth breathing show significant decreases
o
in pulmonary function at an SOp concentration of 0.75 ppm (1,950 ug/m ).
Asthmatics may be more sensitive to S02, but the data on such individuals
are not definitive. Sensory awareness of the presence of SOp may be decreased
in individuals chronically exposed to SOp.
The presence of particles with S02 may influence physiologic responses.
Particulate matter may function as a carrier, bringing more SOp into the
lungs, or may induce chemical reactions to convert SOp into sulfates. The
presence of aerosols such as NaCl does not appear to modify the pulmonary
response to S02 in most normal subjects. However, it has been found that
adolescent extrinsic asthmatic subjects exhibit pulmonary function changes in
the small airways when orally breathing 1 ppm S02 and NaCl aerosols. One
study shows that brief exercise during this exposure induces changes in both
large and small airways.
1-112
-------
Results of a chamber study indicated synergistic effects on pulmonary
functions in response to a combination of ozone (0.37 ppm) and sulfur dioxide
(0.37 ppm). This synergisic effect has not been demonstrated in more recent
studies suggesting that the effect reported in the earlier study may have been
caused by the presence of some other substances in the exposure chamber.
Few studies have examined the effects of exposure to sulfuric acid and
sulfates. Sensory responses to sulfuric acid have not been clearly defined.
Pulmonary functions do not appear to be influenced by exposure to hLSO. at
2
concentrations up to 1,000 pg/m . Mucociliary transport in the airways distal
to the trachea is affected more by H^SO. exposure than is transport in the
trachea. This effect depends upon the concentration levels of HpSO., i.e.,
3
exposure to 100 ug/m increases bronchial clearance, while exposure to 1,000
ug/m reduces clearance. No adverse pulmonary effects have been reported in
3
normal or asthmatic subjects after 2.5 hour exposures to 100 ug/m ammonium
3
bisulfate or 85 ug/m ammonium sulfate.
1.5.5 Community Health Observational Studies
Animal and clinical studies provide incomplete information about the
health effects of sulfur oxides and particulate matter. The principal
limitations of animal studies are problems with extrapolating quantitative
dose effect dose-response relationships from animal to man and the difficulties
of mounting studies large enough to detect small health effects, particularly
small mortality effects, that may be of public health significance. Health
effects of chronic exposure typically cannot be assessed in human clinical
studies, and ethical constraints limit the range and quality of acute exposures
that can be investigated in such studies. Both types of investigations are
limited in that the mix of pollutants found in the ambient air is difficult to
1-113
-------
characterize and duplicate in the laboratory. Thus, the evidence from
observational studies plays an important and crucial role in assessing the
health effects of sulfur oxides and particulate matter.
Observational or community health studies also have important limitations.
Since the level of exposure is not under the control of the investigator,
observational studies can demonstrate associations but not necessarily cause
and effect relationships. Also, since sulfur oxides and particulate matter
have common sources, they frequently covary in level in the observational
setting, making it difficult to distinguish health effects of these pollutants
individually. In addition, other factors such as temperature or frequency of
smoking may covary with the level of sulfur oxides and particulate matter, so
that effects of pollutants cannot readily be separated from effects of these
other factors. A final limitation of observational studies to be mentioned
here is the difficulty of relating pollution levels at one or a few community
air monitoring stations to health effects seen in persons living in the vicinity
of those stations. Further discussion of these and other methodological
issues of importance in evaluating observational studies can be found in
Chapter 14.
Despite the above limitations, an enormous number of epidemiology studies
have been conducted over the past 30 years in order to better define both
qualitative and quantitative relationships between various health effects and
sulfur oxides and particulate matter in the ambient air; and Chapter 14 of
this document provides a rather comprehensive and thorough critical assessment
of such studies. The information contained in Chapter 14 is summarized below.
1-114
-------
1.5.5.1 Overview Summary of Chapter 14 Contents—In Chapter 14, an extensive
array of information is discussed concerning: (1) methodological considerations
that must be taken into account in evaluating community health epidemiology
studies (Section 14.1); (2) critical assessment of practical applications of
air quality measurement techniques employed in the collection of sulfur oxides
and particulate matter data utilized in related community health studies
(Section 14.2); (3) critical review of such studies on mortality effects
associated with acute and chronic exposures to sulfur oxides and particulates
(Section 14.3); (4) critical review of studies of morbidity associated with
acute exposures to the same pollutants (Section 14.4); and (5) critical
assessment of morbidity effects associated with chronic exposures to sulfur
oxides and particulate matter. In addition, in the summary and conclusions
section of the chapter, there are discussed other published reviews and critical
evaluations of the subject material and the interpretations derived from such
reviews are compared.
Through the discussion in Section 14.1 of Chapter 14, it is seen that
numerous methodological factors, including covarying or confounding variables,
can potentially affect the results and interpretation of community health
studies. It is also seen, through material summarized in Section 14.2 of the
chapter, that a number of sources of errors have been identified as having
affected sulfur oxides and particulate matter air quality measurements obtained
in both the United Kingdom and the United States and used in British and
American epidemiology studies which provide the bulk of the information reviewed
in the chapter. It was further noted that while such errors in air measurements
can at times be fairly large, they also often act to introduce both positive
and negative biases into air quality data sets that tend to cancel each other
1-115
-------
out, especially when considering data grouped or averaged over long time
periods (monthly; annually) from the same sites or across several geographic
areas classed as "low" or "high" pollution areas. At other times, however, it
also became clear that certain measurement errors were such as to introduce
either consistently negative or positive bias into particular British or
American sulfur oxides or particulate matter data sets used in various community
epidemiology studies providing information on quantitative air pollution/health
effects relationships. It was further noted that such biases due to air
quality measurement errors must be taken into account in evaluating such
epidemiology studies -- not for the purpose of discrediting such studies but
rather to understand better the error limits likely associated with the reported
quantitative findings derived from them and to thereby allow for more accurate
interpretation of overall patterns of pertinent results.
Turning to the critical assessments of pertinent community health mortality
and morbidity studies contained in Sections 4.3, 4.4 and 4.5 of Chapter 14,
results of many of the better known and often cited quantitative and qualitative
studies discussed in the chapter are summarized in a series of tables presented
below along with summary statements regarding the results of community health
studies of mortality and morbidity affects associated with short- and long-term
exposures to sulfur oxides and particulate matter.
Health Effects of Acute Exposure to SO,, and Particulate Matter—Qualitative
studies demonstrating increased mortality effects to be associated with sulfur
oxides and particulate matter air pollution are summarized in Table 1-14.
Note that these studies demonstrate, for example, associations between such
pollution and significantly increased mortality from bronchitis, pneumonia and
heart disease. Also note that essentially all population groups are affected,
1-116
-------
TABLE l-14a QUALITATIVE ASSOCIATION OF GEOGRAPHIC DIFFERENCES IN MORTALITY
WITH RESIDENCE IN AREAS OF HEAVY AIR POLLUTION
Pemberton?and
Goldberg
Stocks
138,164-167
224-225
Gorham
Q
Gore and Shaddick
and Hewitt:
Haastrom et al.
Zeidberg et al
Sprague et al.
16
17
Lepper et al.
227
Jacobs and
Landoc1/b
1950-1952 bronchitis mortality
rates in men 45 years of age
and older in county boroughs
of England and Wales
Bronchitis mortality, 1950-1953,
in urban and rural areas of
Britian, with adjustments for
population density and social
i ndex
1950-1954 deaths, 53 counties
of England, Scotland, and
Wales
Mortality in London, 1954-1958
and in 1950-1952, respectively
1949-1960 deaths for each cause
in Nashville, Tenn., categor-
ized by census tract into 3
degrees of air pollution and
3 econimic classes (levels
not accurately determined)
1964/1965 mortality rates in
Chicago census tracts strati--
fied by socioeconomic class and
SO concentration
1968/1970 mortality rates
in Charleston, S.C. ,
industrial vs. non-indus-
trial areas
Sulfur oxide concentrations
(sulfation rates) were con-
sistently correlated with
bronchitis death rates in the
35 county boroughs analyzed
Significant correlation of mor-
tality from bronchitis and
pneumonia among men, and from
bronchitis among females, with
smoke density
Bronchitis mortality was strongly
correlated with acidity of
winter precipitation
Duration of residence in London
significantly correlated with
bronchitis mortality, after
adjusting for social class
Within the middle social class,
total respiratory disease
mortality, but not bronchitis
and emphysema mortality, were
significantly assoicated with
sulfation rates and social index.
White infant mortality rates
were significantly related to
sulfation rates
Increased respiratory disease
death rates in areas of inter-
mediate and high SO- concen-
tration, within a socioeconomic
status, without a consistent
mortality gradient between the
areas of intermediate and high
S0_ concentration
Higher total and heart disease
mortality rates in industrial
area
1-117
-------
TABLE l-l4a.(continued)
Morn's et al.
24
Collins et al.
287
Beaker et al.
323
Toyama
330
Lindeberg
321
1960-72 mortality rates
compared to 1959-60 air
pollution levels
Death rates in children 0-14
years of age, 1958-1964,
in relation to social and air
pollution indices in 83 county
boroughs of England and Wales
Thanksgiving 1966 Fog,
New York
Mortality in districts
of Tokyo
Deaths in Oslo winters
Mortality higher in smokers
with lower air pollution
exposures
Partial correlation analysis
suggested that indices of
domestic and industrial
pollution account for a
differences in mortality
from bronchopneumonia and
all respiratory diseases among
children 0-1 year of age
Complaints of cough, phlegm,
wheezing, breath!essness, eye
irritation increased with in-
creasing air pollution
Bronchitis mortatliy associated
with dustfall (but not cardio-
vascular, pneumonia or cancer
mortality)
Average deaths per week, 1958-65
winter, correlated with pollution
1-118
-------
both males and females and both the young and the old; the very young (infants)
and the very old, especially the infirm and those with preexisting respiratory
or cardiovascular diseases, however, appear to be at most risk. Essentially
the same patterns can be discerned for morbidity effects as demonstrated by
qualitative studies summarized in Table 1-15.
Studies providing evidence of quantitative associations between acute
health effects and air levels of sulfur oxides and particulate matter are
summarized in Table 1-16. Overall, various British, Dutch, Japanese and
American episodic mortality studies have yielded results that appear to suggest
that mortality effects might occur at or above 300-500 ug/m S02> The three
non-episodic mortality studies listed in the table suggest that mortality
3
effects can be seen when TSP levels reach 500 to 600 ug/m and S0? concentrations
reach 300 to 500 ug/m . These three studies summarize a relatively small body
of data from two winters in London and five winters in New York City. The
stated effect levels may be conservative (high), however, since examination of
the detailed evidence from these studies presented in Section 14.3 of Chapter
14 suggests the possibility of an exposure-response relationship at lower
levels of these pollutants. More complex time series studies of daily mortality
have also found associations between mortality and these pollutants at lower
levels. The size of the estimated effects has proved to be sensitive to model
specification and choice of other adjustment variables. Although the possibility
of mortality effects of TSP and S02 levels below those cited in Table 1-16
cannot be excluded, it is unlikely that this question can be resolved in the
near future by observational studies. Thus, the minimum air levels at which
3
acute mortality increases might be projected to be seen would be 300-500 ug/m
for both TSP and S02, based on the results summarized in Table 1-16.
1-119
-------
TABLE 1-15. QUALITATIVE STUDIES OF AIR POLLUTION AND ACUTE
RESPIRATORY DISEASE
Study
Characteristics
Findings
Angel et ai.69
Attack rates of minor respiratory
illness among 85 London workers,
examined every 3 weeks, October
1962-May 1963.
Attack rates were associated
with weekly average smoke
and S02 concentrations.
Levy et al.70
Schoettlin and
Landau288
Ze-fdberg et a I.289
Cowan et al.290
Greenberg et al.291
Weill et al.292
Carroll293
Hospital admissions for respira-
tory disease in Hamilton,
Ontario, correlated with
sulfur oxide/particulate air
pollution index.
137 asthmatics reporting attacks
on daily occurrence of asthma,
September-December, 1956, in
Los Angeles Basin.
Study during 1 year of 49 adults
and 34 children with asthma in
Nashville, Tenn.
History of asthma, and skin tests
of University of Minnesota
students, in relation to dust
from nearby grain elevator.
New York City hospital emergency
room visits for asthma in
month of September.
Retrospective study of emergency
room visits to New Orleans
Charity Hospital.
Increased hospital admissions on'
heavy pollution days, except at
one hospital far removed from
major pollution sources.
Significantly more asthma on days
of heavier oxidant pollution.
No adjustment was made for
variations in temperature or season
Doubling of asthma attack rates
in persons living in more
polluted neighborhoods. No
adjustment for demographic or
social factors.
Significant association between
grain-dust exposure and
asthma attacks.
Emergency room visits strongly
associated with onset of cold
weather but not with degrees of
air pollution during the one
month of study.
Periodic "epidemics" of asthma
in New Orleans could not be
traced to any common pollutant
exposure.
1-120
-------
TABLE 1-15 (continued)
Study
Characteristics
Findings
Phelps294
Meyer295_
"Tokyo-Yokohama asthma" in
American servicemen stationed
in Japan after World War II.
Glassej et al.
Chiaramonte
et al.297
296
Derrick57
Rao298
Goldstein and
Black58
Emergency room visits in seven
New York city hospitals during
the November 1966 air pollution
episode.
Emergency room visits at a
Brooklyn hospital during a
November 1966 air pollution
episode.
Nighttime emergency room visits
for asthma in Brisbane,
Australia.
Pediatric emergency room visits
for asthma at Kings County
Hospital, Brooklyn, October
1970-March 1971.
Emergency room visits for
asthma at a hospital in
Harlem and in Brooklyn,
September-December 1970 and
September-December 1971.
Disease primarily in smokers
attributed to allergic response
to atmospheric substances that
could not be characterized.
Patients improved after leaving
the area and were immediately
affected on return. Some had
long-term effects afterwards.
Increased emergency room visits
for asthma in three of seven
hospitals studied.
Statistically significant
increase in emergency room
visits for asthma and for
all respiratory diseases, con-
tinuing to 3 days after the
peak air pollution concen-
trations.
Negative correlation between
asthma visits with degrees
of smoke shade.
Negative correlation of asthma
visits with degrees of smoke
shade. Lack of temperature
adjustments. Considerable
distance of hospital district
from air monitoring stations.
Temperature adjusted asthma
rates positively correlated with
S02 values in Brooklyn but
not in Harlem. In 1971 period,
50-90% increase in asthma
visits on 12 days of heaviest
pollution.
1-121
-------
TABLE 1-15 (continued)
Study
Characteristics
Findings
Finklea et al.117
Finklea
et al.122 123
Incidence of acute respiratory
disease, determined at 2-week
intervals, in parents of
nursery schoolchildren residing
in Chicago, December 1969-
November 1970.
Daily diaries kept by 50
asthmatics in each of three
New York City area communi-
ties, October 1970-May 1971.
Acute lower respiratory
illness rates were signifi-
cantly lower among families
living in neighborhoods
where air pollution had been
substantially decreased.
Rates were adjusted for social
class, smoking, residential
mobility, and season of year.
Cannot quantitate pollutant
exposures.
Temperature-adjusted attack
rates significantly correlated
with total particulates in two
of the communities. Increase
in relative risk from days of
light to heavy pollution was
relatively small. High turnover
in reporting panels.
*Reference 251
1-122
-------
TABLE 1-16 SUMMARY TABLE - ACUTE EXPOSURE EFFECTS
Type of Study
Mortality (episodic)
British
Dutch
Japanese
USA
(Non-episodic)
Reference
Table 14-1
Table 14-2
Table 14-2
Table 14-2
Martin and Bradley11
Martin6
Glasser and
Greenburg222
24-hour average pollutant levels
at which effects appear
Effects observed
Excess deaths
Excess deaths
Excess deaths
Excess deaths
Increases in daily mortality
Increases in daily mortality
above the 15 moving average
Increases in daily mortality
TSP ((jg/mj)
546*
300-500
285
570 (5 CoH)
500*
500*
350-450**
S02 ((jg/mj)
994
500
1800
400-532
(1 hr max: 2288)
300
400
524
Morbidity
Martin16
Lawther et al . 53
Greenberg et al.196
Lawther et al . 52
Stebbings. and
Hayes190'
Increases in hospital admissions 500*
for cardiac or respiratory illness
Worsening of health status among
195 bronchi tics
Increased cardio-respiratory
ER visits
Increased clinical condition
in CB patients
Increased symptoms in chronic
bronchitis (CB) patients
344* (250 BS)
357** (260 BS)
529* (400 BS)
344* (250-350 BS)
200 (60 RSP)
(12SS) 8 SN)
400
300-500
715
450
300
100
-------
TABLE 1-16 (continued).
Type of Study
Reference
Cohen et al.55
Effects observed
Increased AS attacks
24-hour average pollutant levels
at which effects appear
TSP (Mg/m') S02 (pg/m3)
150 (20SS) 200
McCarroll et al.163 Increased ARI daily 160* (1.2 COH) 372
inc/prev
Cassell et al.208 209 Increased ARI average 205* (2 COH) 452
daily inc/prev
Stebbings and Decreased FEV0<75 (children) 700 300
Fogleman et al.216
"Converted from BS (British Smoke).
-------
Numerous studies reporting morbidity effects associated with acute
exposures are also listed in Table 1-16. Worsening of symptoms in bronchitis
patients and increased hospital admissions in Britain were reported to occur
at_TSP and S02 levels of 300 or 350 to 500 ug/m3 or more. A United States
study, however, found exacerbation of symptoms among bronchi tics at 200 ug/m3
3
TSP and 100 ug/m SO,, and asthmatics were reported to show increased attacks
3 3
at 150 ug/m TSP and 200 ug/m SO,,. Also, spirometry tests were reported to
show decreases in lung function at 700 ug/m3 TSP and 300 ug/m3 S02. However,
in another study not listed, van der Lende saw improvement in lung function
o
among adults when pollution levels were reduced from 245 ug/m (TSP) and 300
ug/m (SO,,). Acute upper and/or lower respiratory illness also has been
reported to occur at levels as low as at 160 ug/m TSP (24-hour averages).
Overall, then, the summarized results suggest that (1) very severe morbidity
effects, e.g., worsening of symptoms in bronchitic patients, clearly occur at
TSP and SOp levels of approximately 300 or 350 to 500 ug/m , and (2) less
severe but significant morbidity effects may occur with acute exposure at
levels of approximately 150-300 ug/m . These studies do not, however, provide
a basis for separately estimating the health effects of S02 and particulates.
Since these two forms of pollution have important common sources, their levels
tend to usually vary together over time.
Health Effects of Chronic Exposure to SO,, and Particulate Matter—The results
of qualitative studies on relationships between sulfur oxides and particulate
matter pollution are summarized in Table 117. Many of the same types of
observations as stated above for population groups at apparent special risk
also apply here; that is, the elderly, the infirm, and children appear to be
most severely affected health-wise by chronic exposures.
1-125
-------
Many well-known and often-cited mortality and morbidity studies that have
been reported as demonstrating quantitative associations between mortality,
illnesses, or decrements in pulmonary function with chronic (monthly or annual
average) levels of particulate matter of S02 are summarized in Table 1-18. As
seen in that table, the two mortality studies suggest that mortality effects
can occur at annual levels of 125 to 140 ug/m3 or less of TSP and S02- In the
morbidity studies, lower respiratory disease, chronic bronchitis, and reduced
pulmonary function results were reported that are indicative of morbidity
effects likely clearly occurring at annual average TSP or SO^ levels of 150 to
o
250 ug/m or more. Other study results summarized in the table suggest an
association of various morbidity effects with concentrations in excess of
3 3
about 70 to 80 ug/m TSP and SOp concentration in excess of 96 to 107 ug/m .
As with studies of acute effects, many of these studies could be further
interpreted not only as demonstrating that health effects are exposure-related
but also that they increase as these pollutants increase over the entire range
of exposures studied and no clear-cut "no effect" level can be determined on
the basis of presently available information. Also, in general, these studies
cannot be used to distinguish between the effects of sulfur oxides and particulate
In several studies, however, TSP effects were reported to occur in the presence
of low or non-significant levels of SOp. (Reference 188, 212, 213, 215, and
257 of Chapter 14, as shown in Table 1-18.)
Health Effects of Atmospheric Sulfates—Conversion to sulfate compounds,
including sulfuric acid, has been proposed as a major pathway by which sulfur
dioxide and possible other sulfur compounds may exert toxic effects. However,
only a few community health studies have attempted to measure and assess
health effects associated with suspended sulfates (SS). Stebbings and Hays,190
1-126
-------
TABLE 1-17 QUALITATIVE STUDIES OF AIR POLLUTION AND PREVALENCE OF CHRONIC
' RESPIRATORY SYMPTOMS AND PULMONARY FUNCTION DECLINES
Study
Characteristics
Findings
Fairbatrn and
Reid265
Mork
266
Deane et al.267
Cederlof,39
Hrubec et al.40
Comparison of respiratory illness
among British postmen living
in areas of heavy and light
pollution
Questionnaire and ventilatory
function tests of male trans-
port workers 40-59 years of age
in Bergen, Norway and London,
England
Questionnaire and ventilatory
function survey of outdoor
telephone workers 40-59 years
of age on the west coast of U.S.
Chronic respiratory symptom
prevalence in large panels of
twins in Sweden and in the
U.S. Index of air pollution
based on estimated residential
and occupational exposures to
S02, particulates, and CO
Sick leave, premature
retirement, and death
due to bronchitis or
pneumonia were closely
related to pollution
index based on visibility
Greater frequency of
symptoms and lower
average peak flow rates
in London. Differences
were not explained by
smoking habits or socio-
economic factors
Increased prevalence of
respiratory symptoms,
adjusted for smoking and
age, a larger volume of
morning sputum and a lower
average ventilatory function
in London workers, and in
the English compared with
American workers. No
differences in symptom
prevalence between
San Francisco and
Los Angeles workers,
although particulate
concentrations were
approximately twice as
high in Los Angeles
Increased prevalence of
respiratory symptoms in
twins related to smoking,
alcohol consumption,
socioeconomic character-
istics, and urban residence,
but not to indices of air
pollution
1-127
-------
TABLE 1-17. QUALITATIVE STUDIES OF AIR POLLUTION AND PREVALENCE OF CHRONIC
RESPIRATORY SYMPTOMS AND PULMONARY FUNCTION DECLINES
Study
Characteristics
Findings
Bates ef al.268-270
Bates271
Yashizo272
Winkelstein and
Kantor273
Ishikawa et al.275
Fujita et al.276
Comparison of symptom prevalence,
work absences, and ventilatory
function in Canadian veterans
residing in 4 Canadian cities
10-year follow-up study of
Canadian veterans initially
evaluated in 1960, and
followed at yearly intervals
with pulmonary function tests
and clinical evaluations
Bronchitis survey of 7 areas of
Osaka, Japan, 1966, among
adults 40 years of age and over
Survey of respiratory symptoms
in a random sample of white
women in Buffalo, New York
Comparison of lungs obtained at
autopsy from residents of
St. Louis and Winnipeg
Prevalence survey (Medical
Research Council questionnaire)
of post office employees in
Tokyo and adjacent areas, 1962
and re-surveyed in 1967
Lower prevalence of symptoms
and work absences and better
ventilatory function in
veterans living in the lest
polluted city
Least decline in pulmonary
function with age in veterans
from least polluted city
Bronchitis rates, standardized
for sex, age, and smoking
were greater among men and
women in the more polluted
areas. Bronchitis rates
followed the air pollution
gradient.
In nonsmokers 45 years of age
and over, and among smokers
who did not change residence,
respiratory symptoms were
correlated with particulate
concentrations obtained in
the neighborhood of residence.
No association of symptom pre-
valence with S02 concentrations
Autopsy sets, matched for age,
sex and race, showed more
emphysema in the more polluted
city. Autopsied groups may
not reflect prevalence of
disease in general population
Two-fold increase over time in
prevalence of cough and sputum
production in same persons,
irrespective of smoking habits.
Change was attributed to
increasing degrees of air
pollution
1-128
-------
TABLE 1-17., QUALITATIVE STUDIES OF AIR POLLUTION AND PREVALENCE OF CHRONIC
RESPIRATORY SYMPTOMS AND PULMONARY FUNCTION DECLINES
Study
Characteristics
Findings
Reichel,277
Ulmer et al.278
Nobuhiro et al.279
Comstock et al.280
Speizer and
Ferris281-282
Linn et al.283
Prindle et al.284
Respiratory morbidity prevalence
surveys of random samples of
population in 3 areas of West
Germany with different degrees
of air pollution
Chronic respiratory symptom
survey of high and low exposure
areas of Osaka and Ako City,
Japan
Repeat survey in 1968/1969 of east
coast telephone workers and of
telephone workers in Tokyo
Comparison of respiratory
symptoms and ventilatory
function in central city and
suburban Boston traffic poll ice-
men
Respiratory symptoms and function
in office working population
in Los Angeles and San Francisco,
1973
Comparison of respiratory
disease and lung function
in residents of Seward and
New Florence, PA
No differences in respiratory
morbidity, standardized for
age, sex, smoking habits,
and social conditions,
between populations living
in the different areas
Higher prevalence of chronic
respiratory symptoms in more
polluted areas
After adjustment for age and
smoking, no significant
association of respiratory
symptom prevalence with
place of residence
Slight but insignificant
increase in symptoms pre-
valence among non-smokers
and smokers, but not
exsmokers, from the central
city group. No group
differences in ventilatory
function
No significant difference in
chronic respiratory symptom
prevalence between cities;
women in the more polluted
community more often reported
nonpersistent (<2 years)
production of cough and sputum
Increased airway resistance in
inhabitants of more polluted
community. Differences in
occupation, smoking, and
socioeconomic level could
account for these differences
1-129
-------
TABL€-1-17 QUALITATIVE STUDIES OF AIR POLLUTION AND PREVALENCE OF CHRONIC
RESPIRATORY SYMPTOMS AND PULMONARY FUNCTION DECLINES
Study
Characteristics
Findings
Watanabe285
Anderson and
Larsen286
Collins et al.287
Peak flow rates in Japanese
school children residing in
Osaka
Peak flow rates and school
absence rates in children 6-7
years of age from 3 towns in
British Columbia
Death rates in children 0-14
years of age, 1958-1964,
in relation to social and air
pollution indices in 83 county
boroughs of England and Wales
Lower peak flow rates in
children from more polluted
communities. Improved peak
flow rates when air pollution
levels decreased
Significant decrease in peak
flow rates in 2 towns
affected by Kraft pulp
mill emissions. No effect
on school absences.
Ethnic differences were
not studied
Partial correlation analysis
suggested that indices of
domestic and industrial
pollution account for a
greater part of the area
differences in mortality
from bronchopneumonia and
all respiratory diseases
among children 0-1 year
of age
1-130
-------
TABLE 1-18 SUMMARY TABLE - CHRONIC EXPOSURE EFFECTS
i !•
Type of Study
Mortality (geog. )
Reference
Winkelstein188
Zeidberg and
colleagues16-18
Annual average pollutant levels
at which effect occurred
Effects observed
Increased mortality
Increased mortality
TSP (ug/m3)
125-140
55-60
S02 (ug/m3)
not significant
30
Morbidity
Longitudinal and
cross-sectional
Ferris
et al.41 42 46 47
Higher rate of respiratory
symptoms; and decreased lung
function
180
55
Cross-sectional
(2 areas)
Sawicki (1972)31
More chronic bronchitis,
asthmatic disease in smokers;
reduced FEV%
250"
125
Cross-sectional
study of school-
children in 4 areas
Lunn et al.96 97
Increased frequency of res-
piratory symptoms; decreased
lung function in 5-year olds
260*
190
Follow-up of school-
children in 4 areas
Douglas and Waller90
Increased lower respiratory
tract infection
197* (130 BS)
130
Cross-sectional study
of children in 4 areas
Hammer et al.214
Increased incidence of lower
respiratory diseases
85-110
175-250
Cross-sectional study
of high school
children in 2 areas
Mostardi and
colleagues177 258
Lower FVC, FEV0>75 and maximal
oxygen consumption
77-109
96-100
Cross-sectional
(multiple areas)
Lambert and Reid28
Increased respiratory symptoms
160* (100 BS)
100-150
Cross-sectional
(3 areas)
Goldberg et al.109 Increased CRD
78-82
69-160
-------
TABLE 1-18 (continued)
I
t-«
ro
Type of Study
Cross-sectional
(4 areas)
Cross-sectional
and Long (2 areas)
Cross-sectional
(3 areas)
Cross-sectional and
retro- long in 4 areas
(children)
Cross-sectional
2 areas
Cross-sectional
3 areas (children)
Cross-sectional
2 areas (children)
Reference
House et al.108
Sawicki and
Lawrence (1977)181
Rudnick182
Nelson et al.114
Hammer113 257
Shy et al.215
Shy et al.215
Chapman et al.213
i r •
Annual average pollutant levels
at which effect occurred
Effects observed
Increased CRD
Increased Prev CB and AS
Increased persistence, Males
31-50; Increased incidence,
Females, some ages
Increased respiratory symptoms
in boys. Increased Rh in girls
Increased LRD
Increased LRD
Decreased adjusted FEV>75 in
children > 8 years
Decreased adjusted FEV>75
TSP (|jg/mj)
70 (15SS)
169+
221-316*
(150-227 BS)
70
133
(SS=14)
78-82
96-114 (45 RSP)
S02 (pg/m3)
100-150
114-130
108-148
107
<25
69-160
(= and low)
*Converted from BS (British Smoke).
**Converted from CoH.
-------
for example, reported increased symptoms in patients with 24-hour averages of
12 ug/m3 SS (200 TSP, 60 RSP, 8 SN, 100 S02). Also, Chapman et al.212 reported
increased chronic respiratory disease prevalence rates in a high pollution
community with an annual average of 15 ug/m3 SS (70 TSP and 107 SOO. Hammer257
further reported increased lower respiratory disease prevalence rates in a
high pollution community with an annual average of 14 ug/m SS (133 TSP and
S02 >17). Thus, suspended sulfate levels of 12 ug/m3 (daily) or more and 15
ug/m (annually) might be interpreted as being important based on those results;
however, certain methodological considerations discussed in Chapter 14 must be
taken into account in qualifying these results.
Respirable Particulates Effects—As discussed in Chapter 11, particles below
3
15 ug/m MMAD are important. Respirable suspended particulates (RSP) £10 urn,
have been measured in only a few American epidemiology studies, e.g., those by
~i~n 9^7 iqn ?~R ?TI
Hammer, '" Stebbings and Hayes, and Shy and Chapman et al. J' The
later study was reported as demonstrating decreased adjusted FEV 75 in children
3 3
in an area with higher pollution with RSP of 45 ug/m (96 to 114 ug/m TSP and
3
S02 very low). Thus RSP of 45+ ug/m may be important but, again, acceptance
of these results must be qualified based on certain methodological considerations
discussed in Chapter 14.
1.5.5.2 Methodological Factors Impacting Interpretation of Results—If it
were assumed that all of the results summarized in Tables 1-16 and 1-18 were
derived from methodologically sound studies and were universally accepted as
valid, then the above summary of their results could be accepted as a reasonable
representation of the likely atmospheric particulate and sulfur oxides levels
found to be associated with mortality and morbidity effects. However, the
matter of the methodological soundness and validity of various studies has
1-133
-------
been a matter of considerable controversy and discussion during the past
decade.
Such controversy has derived, in large part, from the fact that certain
additional risk factors can often be as important as the air pollution variables
studied in affecting human health. For example, as was alluded to earlier and
discussed more thoroughly in Chapter 14 (Sections 14.1, 14.3, 14.4), it has
been strongly emphasized that smoking is one such factor, as are occupational
exposures. Furthermore, age and sex co-variables can also be critical in the
evaluation of health effects. Race or ethnic group characteristics likely
fall into this catagory as well. In addition, numerous social variables may
be highly critical in terms of their existing direct effects on human health,
as well as how they may modify the health effects of environmental pollutants.
Such social factors include social economic status (income, education, and
occupational levels and associated social class status), migration, and household
characteristics. Also, meterological variables such as sudden temperature
changes or shifts in humidity levels may also be critical co-variables which,
along with air pollutants, might affect health in a deleterious manner.
Parental smoking and other sources of indoor pollutants may also be critical.
Other less-well defined social/ environmental variables, such as a greater
degree of crowding in housing conditions, too, may represent a set of "urban
factors" differentially acting to affect health in comparison to "rural"
conditions.
Many of the studies have attempted to control for at least some of the
above and other factors. Most of the studies analyzed, however, have not
controlled for all possible confounding or covarying factors, an extremely
difficult task in practice. Nor have all been able to exquisitely, control for
1-134
-------
what factors were taken into account. In fact, there is not a single study
that has controlled for everything that might have affected its results.
Thus, the likely validity of each study has to be appraised after evaluating
the_ importance of possible confounding variables and co-variables controlled
for or not taken into account by that study. In many cases, known highly
correlated variables have been controlled for or taken into account and, in
other cases, comparison study groups have been chosen so as to be similar in
terms of crucial characteristics, making it relatively easy to ascribe likely
validity to their observed results.
In regard to evaluating other (less well-designed) studies, it should be
noted that some studies exist which indicate that possible confounding variables
are not always as important as they were originally thought to be. For example,
follow-up studies on an adult cohort previously studied as children by one
group of investigators did not confirm original social class differences
between the groups to be of much significance in accounting for health findings
for the groups later in life, and other studies have shown that household/familial
factors are not necessarily important in all cases in accounting for observed
results. Therefore, care must be taken not to over-emphasize the relative
importance of potential confounding or covarying factors not ruled out as
possible alternative explanations for the results of a given study. In other
words, being overly critical where information is lacking to support the
likelihood of a specific confounding factor or co-variable affecting the
pattern of results obtained in a study at a particular time represents as much
of a disservice in trying to achieve an objective, balanced appraisal of study
results under discussion as would any countervailing lack of reasonable regard
for the potential importance of such factors.
1-135
-------
It must also be recognized that no single study alone, no matter how
well-designed or conducted in and of itself completely establishes what comes
to be accepted as a "scientific fact" defining either a relationship between
two or more variables studied or a lack thereof. Rather, excellence in the
design and conduction of a given study, internal consistency and biological
plausibility of its results, and their consistency with other known results or
information all help to heighten confidence in the likely existence of relation-
ships indicated by that study's results. Even greater certainty is attributed
to the probable existence of such relationships if further independent studies,
regardless of particular individual flaws, yield results consistent with such
relationships. Thus, consistency in the overall pattern of results indicative
of particular relationships or the overall "weight of the evidence" from more
than one study are crucial in establishing given relationships as "scientific
facts" or in determining the degree of certainty ascribed to them.
1.5.5.3 Quantitative Dose-Response Relationships Defined by Community Health
Studies—In order to elucidate dose-response relationships established
by community health epidemiology studies of the type reviewed above, numerous
attempts besides the present one have been made to examine both negative and
positive information concerning such studies. This has usually been done to
determine which are sufficiently sound methodologically to allow for reasonable
conclusions to be drawn from them in evaluating the overall meaning of their
results individually and collectively. Such attempts include critical reviews
OAt: 7Aft
and commentaries written by Rail (1974), Higgins et al. (1974), Goldsmith
and Friberg (1977),247 Ferris (1978),314a and Waller (1978).314b They also
include the following evaluative documents appearing in 1978: an American
Thoraic Society (ATS) review of Health Effects of Air Pollution (Shy et al.,
1-136
-------
251
1978); a National Research Council /National Academy of Science (NRC/NAS)
document on Airborne Particles by Higgins and Ferris (1978) and an NRC/NAS
ono
document on Sulfur Oxides by Speizer and Ferris (1978) More recent such
reviews and commentary appearing in 1979 include: the 1979 World Health
Organization (WHO) document, Environmental Health (8): Sulfur Oxides and
31? ^m
Suspended Particulate Matter; a report by Holland et al. (1979) written
for the American Iron and Steel Institute and appearing in the American Journal
of Epidemiology; and a reply to that report in the same journal by Shy (1979).
Some of the more salient points of these reviews and commentaries are concisely
highlighted below.
As will quickly become apparent through the course of the discussion
below, there are certain studies that many reviews consistently rate as being
methodologically sound and their results valid. Also, when those study results
are viewed together, collectively, fairly consistent patterns of quantitative
relationships emarge regarding exposures to sulfur oxides and parti cul ate
matter associated with the occurence of various types of health effects,
including (1) mortality and morbidity effects associated with acute exposures
to fairly high ranges of air concentrations of those substances and (2) mor-
bidity effects associated with chronic exposures to lower atmospheric levels
of the same agents. Given the general concensus that appears to exist regarding
the validity of these studies, then, there seems to exist very good support
for placing considerable confidence in the overall patterns of quantitative
relationships defined by their collective evaluations.
In regard to other reasonably well-designed studies, but for which less
of a concensus exists regarding their likely validity, several interesting
points emerge from the subsequent discussion. First, it becomes apparent
1-137
-------
that, beyond some small modicum of agreement among the reviews concerning
problems associated with certain studies, the various reviews often differ
considerably in regard to their assessments of the methodological soundness or
validity of any given individual study. This derives mainly from different
reviewers emphasizing or citing different possible confounding or covarying
factors as potentially being important in affecting the results of a given
study. However, despite whatever flaws might be evaluated by different reviewers
to be associated with the particular individual studies, a surprisingly great
degree of consistency exists both between most of the "flawed" study results
and, also, in comparison with the findings of the other studies alluded to
above as being widely recognized as being valid. In some cases, however, the
results of some of the supposedly "flawed" studies point toward still lower
levels of sulfur oxides and particulate matter being associated with significant
mortality or morbidity effects. Thus, whereas not as much confidence can yet
be placed in such findings as those from the more universally accepted studies,
it is still not appropriate scientifically to completely disregard or ignore
them. This is especially true in view of the fact that, all too often, relationships
indicated to exist by "suggestive" evidence derived from numerous "flawed"
studies are later confirmed by more carefully designed and conducted "definitive"
studies.
Tables 1-19 to 1-21 summarize conclusions regarding particular study
results based on critical assessments contained in Chapter 14 of this document
and assessments from other reviews of observational studies of the health
effects of sulfur dioxide, expressed in ug/m of SOp, and particulate matter,
3
expressed in terms of ug/m of total suspended particulates (TSP). When
exposures were originally obtained in units of black smoke or coefficient of
1-138
-------
haze, they have been converted to TSP for these tables. The conversion
relationships are discussed in Chapters 3 and 14. Because information bearing
on the health effects of atmospheric sulfates and the fine fraction of
pa_rticulate matter is insufficient to separately assess the health effects of
these fractions of atmospheric particulates, these measures of pollution level
are not considered in these tables. These issues are discussed in more detail
in Chapter 14.
Each row in Tables 1-19 to 1-21 corresponds to an observational study
that has been cited in at least one review. These studies are grouped
according to whether acute effects were studied, that is, effects associated
with fluctuations in 24-hour average level, or chronic effects were studied,
that is, differences in effects associated with differences in annual average
level. Within each exposure category, studies showing mortality effects are
listed separately from those showing morbidity effects. The geographic
location(s) of the populations studied and approximate dates or time periods
covered by each study are given.
Each column corresponds to one of the particular reviews referenced and
discussed in Chapter 14; and the entries are the air concentrations of of SC^
and TSP interpreted to be associated with health effects demonstrated by the
different studies listed. The absence of stated quantitative values for
TSP/S02 for a given study does not necessarily mean that it was judged to be
an inadequate study by the particular reviewer(s); rather, in some cases,
either no levels were clearly stated in their review or the particular study
may not have been considered (some studies, for example, have appeared after
the publication of several of the reviews).
1-139
-------
TULE 1-19. SUMMV Of VARIOUS REVIEWERS' EVALIMTIWK Of QUNHTITATIVE
DOSE-RESPOWSE RELATIONSHIPS DERIVED FRO* STUOIES OF
PDRTAUTY EFFECTS ASSOCIATED KITH ACUTE EXPOSURES TO
S0? AW PARTICULATE MATTER
' " — ,
Study
:
Scott/Btirj«M/
Gore
Lavther13
Glasser 4...
Greenburo."2
Martin L
Bradley"
Martin6
U. K. Ministry
of Pensions
McCarroll
et al.
Greenburg,
„ itil.1S1
1
g Beuchley
Vatanab* ...
i Kanefco""
Composite
Dutch
Compos 1U
British
Conpoilt*
USA
«ppl1nj it «1.
I Waller
Rlggan
et al.
Date
(1954-56)
(1958-59)
(1960-64)
(1958-59)
(1959-60)
(195Z)
(1962-64)
(1953-64)
(1962-66)
(1965-66)
(1960s)
(1955-62)
(1952-64)
(1975)
(1975)
. ,, Goldsulth &
Population (lSJi) M^?* Fr1lwr9 F«"'« W»S $0 IMS TSP WHO "'I"?
(1,74) ,,„,, ,„„, (W7|>) {M7§J WSJSP (WH09) rt^.,. ^^
London 2000/1144* 2000/1000 «»„«.
»00/1040 1000/750
London 750/700 Txinin
750'710 750/710
New York
291/520 750/710
lon*m 750/710
Iond0n 1000/500 *IT««/,»T
417BS/277 500/500 750/700
U. K.
>1000/>1000
H*» York 720/1500 720/1500
800A450
Nev York
720/500 > 570/850*
N~Y°rk -/*» -/300 -/300 -/500
Osaki
Kotterdn
Large
U.K. cities
Urge
U.S. cities
London
54C/994
Pittsburgh
Ware EM
et al. Ch 14
) (1980) (1980)
2000/1000
580/780 33D/524
530/300 500/300
500/400 500/400
570/1500
570/1000
-/500
300/266
300/500
250/250
570/1000
500/7*0
700/300
(•») Judg«d to be •etKodologlc^lly sound/valid ttudy.
(-) Judged to be Methodologically flawed/Invalid ttudy.
(7) Judged to be Methodologically sound, but reservations enpressed.
* Each entry represents TSP/SO. levels 1n pg/m3 as reported by given
reviewer. Exceptions: BS=Br1t1sh smoke shade, CoH= coefficient of haze.
-------
TABLE 1-20. SUMMARY OF VARIOUS REVIEWERS' EVALUATIONS OF QUANTITATIVE
DOSE-RESPONSE RELATIONSHIPS DERIVED FROM STUDIES OF
MORBIDITY EFFECTS ASSOCIATED WITH ACUTE EXPOSURES TO
SO AND PARTICULATE HATTER
Study
IQfi
Greenberg
Lawther52
Lawther53
Martin16
Waller7
Angel59/ .„
Fletcher*"
McCarrol205
Cassell208'209
.- Lawther as.oer
~ Goldsmith3
Lawther as.Der
Golds«1th
Carnow174
Cohen36
Vander Lende74
Hanwer214
Chapman
HaMter113.257
GervoU61
Stebblngs216
Date
1953
1954
1954-68
1959-60
1961-64
1962-63
1963-65
1964-67
1964-65
1966-68
1968
1968-69
1969
1969-71
1969-71
1970s
1975
Rail
Population (1974)
New York
London
London
London
London
London
New York
New York
London
London
Chicago
West Virginia
Netherlands
New York
Birmingham
(S.E. USA)
France
Pittsburgh
Goldsnith & Holland
H'fgglns Frlberg Ferris NAS SOW NAS TSP WHO et al. Shy
(1974) (1977) (1978) (1978? (1978) (1979) (1979) (1979)
3 CoH/700
250/250-500 350/250 350/250 500/500 250 BS/500 350/500
516/340
250/500 350/500
MO/250 250/250 230/250 200/400 (?)
129/264 (*)
68/204 (t)
-/700
150/200 150/200 150/200 (-)
230/300 245/300 140 85/300 (?) 160 BS/-
145/286 (-)
180/26
-------
TAILE 1-21 Sl»«A«Y OF VARIOUS REVIEWERS' EVALUATIONS OF QUANTITATIVE
DOSE-RESPONSE RELATIONSHIPS DERIVED FROM STUDIES OF
NOM10ITT EFFECTS ASSOCIATED WITH CHRONIC EXPOSURES TO
SO, AMD PMT1CUUTE MATTE*
Study
U.K. Ministry
of Pension*
DouglasJL
Waller90
w
Ferris43
Fletcher274
ItasUrdl25*
1
C shy**
Bennett
Tessler322
Col ley t Reid
Suzuki *,,.
Hltosugr"
Irving et al.98
Rudnlck182
French306
Data
1946-65
19*6-68
1963-65
1966-67
1950V65
1966-73
1968
1972
1968
1966-67
1971-74
1966
1970
1971-72
1972-73
1972-74
1969-71
1969-71
Gold»1th i
Rill Hlgglni Frlberg Ferris
Population (1974) (1974) (1977) (1978)
BrIUIn 200/200*
BrIUIn 70/90 140/130 230/120
BrIUIn 100/100 100-200/100-200
BrIUIn 180/120
Berlin, N.H. 180/55*
London 420 > 100/260 250/250
Cracow 240/130
Ohio 93/98
*•" York 85-195/50-450
Cincinnati, OH
Kent, Eng. 708V-
France
BrIUIn -/100
Tokyo
B1r>lngha», AL
(S.E. USA)
BHUln
Poland
Chicago
BlretnflhM. AL
(S.E. USA)
Holland
HAS SO MAS TSP WHO «t il. Shy
(1978) (1978) (1979) (1979) (1979)
>100/>100 (-)
230/120 70/90 140BS/140 (?)
100/100 (-) 1-200BS/-
180/120 100/120 200 BS/200 330/180
180/55 180/73 180/- (?) 180/-
250/250 (-)
250/125 170/125 (7) 170/-
110/110 (-)
100/200 (-)
(-)
(7) <100BS/-
(7)
-------
One very notable feature of all three tables is the variety of levels
cited for the same studies by different reviewers. When one considers the
continuous relationship between exposure level and response seen in many of
these studies (for example, the studies by Martin discussed in detail in
Chapter 14), this variation in cited levels can be attributed largely to the
lack of a clear threshold level for effects being defined by these studies.
The other important feature of the table is the variation among reviewers in
the choice of studies considered to have demonstrated health effects. The
next sections are devoted to a discussion of these differences in interpretation
among reviewers and to a discussion of the studies which bear critically on
differences in conclusions drawn in these various reviews.
Acute Mortality
Examination of Table 1-19 reveals that considerable agreement exists to
the effect that episodic mortality has occurred definitely above levels of 750
"3 13 9A~1 307 31 ? ^01
ug/ni TSP and SQ^ > > > *> jn London. Holland et al. concluded,
o
mostly from British studies, that the critical values were 500 ug/m TSP and
700 ug/m S0?. Interpretation of daily mortality by Martin et al. ' indicate
the effective levels in London could be as low as 500 ug/m TSP and 300 to 400
ug/m S02.304'312 Also, new analyses contained in Chapter 14 suggest that the
3
TSP effect levels may be as low as 200-400 ug/m in the absence of significant
temperature or other confounding effects. Similar studies by Glasser et
220
al. in New York City would indicate that the levels where mortality has
been seen there could be 2.5 to 5 CoHs (190 to 580 ug/m3 TSP) and 520 ug/m
307
S02. This is not all that different from results found for London, but
different reviewers may interpret these results somewhat differently. WHO
concluded that levels above 500 ug/m3 of each (SOp and particulate matter
1-143
-------
312
expressed as BS) could definitely produce increased mortality. On the
other hand, some studies in the Netherlands and in Japan might indicate that
3
such mortality increases could occur at levels of TSP around 300 pg/m with
SO^ at 500 (jg/m3 or above.100'232'302 All of these effects are tempered by
the particular meteorological conditions present during the study period and
whether only central pollution monitoring states or multiple numbers of
geographically representative sites were used in determining the stated levels
of S0? and particulate matter. All studies showed mortality effects to occur
predominantly in the infirm, the elderly, and infants.
There is more certainty of effects occurring when either TSP or S02 are at or
above 500 (jg/m . Mortality effects may be less certain at levels of TSP
between 300 and 500 pg/m3 at levels of S02 of 500 (jg/m3.
Acute Morbidity
There is a broader range of opinion, and estimations of effect levels,
associated with the studies of acute morbidity. The earlier studies by Lawther,
Waller et al. of pulmonary function and of exacerbation in bronchitics show
various effects when the S0? and BS levels were higher in London. Most reviewers
3
agree that these effects occurred when TSP levels were 250 to 350 pg/m and
3
SOp levels were 250 to 500 |jg/m . As the levels decreased, the acute effects
decreased. By the winter of 1964-65, the exacerbations were slightly reduced
3
and less consistent; daily average BS was 129 pg/m (196 TSP) and daily average
SOp was 264 pg/m in that winter. Even during the winter of 1967-68, although
the effects had decreased further, they were significantly correlated with
3
pollutant concentrations; daily average BS was 68 pg/m (121 TSP) and daily
average SO^ was 204 p.g/m during that winter. The Greenberg et al. studies in
New York City in the 1960s had also shown increased morbidity during episodes
1-144
-------
3 3
at levels above 300 ug/m TSP and 700 ug/m S02_ WHO concluded that daily
levels above 250 (jg/m of S02 and BS (-330 TSP) would produce acute effects.
However, several studies have shown effects at levels around or below those
values. Studies of daily acute respiratory disease symptoms in New York City
families, which included statistical analyses or controls for possible confounders,
showed increased daily symptoms at levels of 160 to 205 ug/m3 TSP and 370 to
3 205 208 209
450 ug/m SO,,. ' ' However, these were only average values over the
winters of the study. Studies of prevalence rates of acute respiratory symptoms
children in New York City, controlling for confounders indicated increased
3 1 9~\9
rates associated with levels of 145 ug/m TSP and 286 ug/m S02. ')< A
similar study in Birmingham showed increased rates at levels of 180 to 220
•3 •}
ug/rn TSP and 26 ug/ni S02.
Probably one of the most critical s4udy to be considered is that of Cohen
oc
et al. They studied 20 asthmatics around a coal -fired power plant. They
found temperature was most important, but within temperature ranges, pollution
3 3
also caused asthma attacks. The levels cited were 150 ug/m TSP and 200 ug/m
S0?, both daily averages. Various reviews have considered this study
weak, ' ' ' while others consider it a significant demonstration of
251
competition of environmental factors. After temperature and one pollutant
were removed, none of the other pollutants entered into the relationship;
thus, no single pollutant could be considered solely responsible for the
increased attack rates.
On the basis of the above studies, one could conclude that acute morbidity
could be seen at levels between 145 to 220 ug/m3 TSP and approximately 200 to
400 ug/m3 SO
1-145
-------
Chronic Morbidity
The studies summarized under the section of chronic morbidity include a
mixture of investigations carried out in population samples composed of random
samples of adults, and adults selected from working groups and populations of
children selected by areas of residence. In addition, all but a few of the
studies use a cross sectional method, that is, they examine the population in
question at one point in time and determine the prevalence of whatever morbid
condition is being assessed. Alternatively, a few of the investigations
follow a group of people over a period of time. Although in these studies the
opportunity to measure the onset of new morbid events exists, eventually all
of the studies actually measure the prevalence of conditions at different
points in time and cannot be used to determine the incidence of new conditions
related to exposure to air pollutants. The morbid events in each study generally
relate to the prevalence of chronic respiratory symptoms or disease states as
ascertained by standardized questionnaires and/or levels or changes in levels
of pulmonary function measured by generally reliable equipment.
The variation in reported levels of effective exposures from the reviews
which indicate the same studies as providing reliable information probably
reflect one or more of the various problems of with aerometry or health measurement
errors encountered in this observational assessments (see above). The inclusion
(or exclusion) of any given study in any particular review in part reflects the
interpretation by the reviewers of the validity (reliability?) of the measurement
of exposure or health outcomes in that particular study.
Except for the Holland report which found only one group of acceptable
studies which provided information on chronic morbidity, >97 all of the other
reviews found several studies with which to provide estimates of lowest levels
1-146
-------
of health effects. In general, there is a range of agreement between the
reviewers that health effects are measurable at levels of TSP above 180 ug/m3
generally in conjunction with levels of S02 above 120 ug/m3. There are notable
exceptions in which lower levels are reported to be associated with health
effects. For example, Ferris reports TSP levels of 180 ug/m3 in association
with S02 levels of 55 ug/m , however, the S02 measures were made from 30-day
sulfation rates and may not reflect considerably higher peak exposures which
may have occurred.
More striking in indicating a divergence of levels is the study of Mostardi
and certain other studies reviewed in Chapter 14. Mostardi indicates effects
at levels of both TSP and S02 approximately 100 ug/m , however, some of the
reviewers rejected his findings because he did not analyze his data on pulmonary
function in adolescents by race in spite of having mentioned that there were
some blacks who would be expected to have lower levels of pulmonary function
in the high exposure group.
The remaining studies were generally part of EPA sponsored CHESS Program
investigations which have received independent peer review and have been
published in the open literature over the last several years. In these studies,
increases in chronic morbidity have consistently been found where levels of
3
TSP and/or S02 have exceeded 100 ug/m . Discussions of published critiques of
shortcomings of a number of these studies are included in Chapters 3 and 14 of
this document and, also, in some of the other reviews listed in Tables 1-19 to
1-21. Due to considerable controversy over the results and interpretation of
many of the CHESS studies, many such quantitative findings regarding air
pollution/health effects relationships need to be qualified to take into
account, for example, certain air quality measurement errors that tended to
1-147
-------
bias detected effect levels in a negative direction (i.e., toward somewhat
lower levels than those suggested by proximally located local air sampling
monitors.) Still, at least some of the CHESS studies, e.g., that by
Hammer113'2 on children in the Southeastern United States (which was
approved as a Harvard doctoral dissertation), appear to be methodologically
sound, their results well-analyzed statistically and accurately interpreted,
and likely valid. However, in the absence of much, if any, comment on some
such studies in the published reviews listed in Tables 1-19 to 1-24, it is
difficult to assign the same level of certainty to their reported effect
levels as those indicated by results from others more widely critiqued to date
and agreed upon as being valid. Nevertheless, it should be noted that many of
the quantitative results derived from such studies, especially when known or
suspected errors in aerometry are taken into account, are not badly divergent
from the results obtained in some number of other studies.
The quantitative relationships defined by many of the studies listed in
Tables 1-19 to 1-21 are depicted in Figure 1-4. Also provided in that figure
are some indications of notable divergence of opinion between certain reviewers,
especially Holland et al. (1979) versus the WHO (1979) appraisal, the present
EPA (1980) evaluation and several other published appraisals. The acute and
chronic exposure levels for SOp and particulate matter (BS translated to TSp)
evaluated by the WHO (1979) to be associated with mortality and morbidity
effects are indicated by dashed lines in Figure 1-4. Also, presented below
are tables (Tables 1-22 to 1-24) summarizing the conclusions of the WHO (1979)
regarding levels at which mortality and morbidity effects can be expected to
occur and their recommended guidelines for exposure limits consistent with the
protection of human health.
1-148
-------
| OSAKA (1962)^ | 1
EPA (1980)^)'^
900 — ^
X
800 — ^',
EPA (1980)0' ^
700 —
600 —
r)
* -ACUTE MORTALITY
UJ
2 ROTTERDAM (1960'i)0 ^'
g-iOO _ _ _^-__ EP*J.!98°'O'W
5 *™ LONDON (1950-60f[J J"J
§ ACUTE MORBIDITY *
u.
-J
V)
400 ~ LONDON (1958-60)1 1
300— LONDON 0 NETHERLANDS (1969-72)
NEW YORK CITY> (1964-65) /
(1960-70) LJ ~
CHRONIC MORBIDITY ,ri D "•
WEST ^ SHEFFIt ^nSSiS&HOLLAND. ET AL
200 ~ VS'AE WA0~"VRANCE,1973>
(1979)
UK (1946-66) & WHO (1979)
fWlf Af*n MO79I
^7 f*R Af*mAl 1 Qfift 7TI
TOKYO (1970|V 1 j
BERLIN, NH (1967 73)^ WHO (1979)
IT SOUTHEAST U.S.A. (196971) | j
,CHOLLAND, ET AL (1979) | |
N
^©HOLLAND. ET AL (1979)
F —
s
^OHOLLAND, ET AL (1979) —
s
*$' -
<$&'
/
/
HO (1979) —
ACUTE MORTALITY
O MARTIN, ET AL. (1960-64) - LONDON
0 GLASSER & GREENBURG (1971) - NYC
C APLING. ET AL, WALLER (1977-78) LONDON
• OTHER STUDIES
ACUTE MORBIDITY
D LAWTHER (1970) - LONDON 1950-1975
(E VAN DER LENDE (1975) - NETHERLANDS
f\ COHEN. ET AL (1972) WEST VIRGINIA
(1979) | OTHER STUDIES
CHRONIC MORBIDITY
A LUNN. ET AL (1967, 1970) • SHEFFIELD. UK
A DOUGLAS & WALLER (1966) - UK
A FERRIS. ET AL (1973. 1976) - BERLIN. NH
V SAWICKI (1972) -CRACOW. POLAND —
Y OTHER STUDIES
I 1 1 1
100
200
TOTAL SUSPENDED PARTICULATES,
Figure
Comparison of interpretations,of studies evaluated by Holland
et al. (1979), WHO (lij^tnft or otner reviews such as those
in the NRC/NAS documents ' and the present chapter. Aside
from the British studies noted for London and Sheffield,and the
1960-64 New Youk City mortality study, Holland et al. either
ignored the other studies shown or evaluated them as being in-
valid based on methodological flaws or reinterpretation of their
findings. "OTHER STUDIES" not specifically identified in the
above key include those reported by: Gervois et al.,17n_Ecance
(1973); Martin10 D London Q958-60); Mostardi et al. /f"° V
Chicago (1972); Hammer11"5'"7 V Southeast USA (1969-71);
Suzuki and3|jljtosugi V Tokyo (1970). The dashed lines depict
WHO (1979) conclusions regarding S0? and particulate levels
associated with acute (24-hr) mortality, acute morbidity, and
chronic (annual) morbidity.
1-149
-------
TABLE 1-22. EXPECTED EFFECTS OF AIR POLLUTANTS ON HEALTH IN SELECTED
SEGMENTS OF THE POPULATION: EFFECTS OF SHORT-TERM EXPOSURES^*
n3
Expected effects Sulfur dioxideSmoke
24-h mean concentration (ug/m )
in
Excess mortality among the elderly 500 500
or the chronically sick
Worsening of the condition of patients 250 250
with existing respiratory disease
"Concentrations of sulfur dioxide and smoke as measured by OECD or British
daily smoke/sulfur dioxide method (Ministry of Technology, UK, 1966;
Organization for Economic Cooperation and Development, 1965). These
values may have to be adjusted in terms of measurements made by other
procedures.
*From WHO 1979 Criteria Document for Sulfur Oxides and Particulate Matter.
1-150
-------
TABLE 1-23. EXPECTED EFFECTS OF AIR POLLUTANTS ON HEALTH IN SELECTED
SEGMENTS OF THE POPULATION: EFFECTS OF LONG-TERM EXPOSURES3*
Annual mean concentration (ug/m )
Expected effects Sulfur dioxideSmoke
Increased respiratory symptoms 100 100
among samples of the general
population (adults and children)
and increased frequencies of
respiratory illnesses among
children
Concentrations of sulfur dioxide and smoke as measured by OECD or British
daily smoke/sulfur dioxide method (Ministry of Technology, UK, 1966;
Organization for Economic Cooperation and Development, 1965). These values
may have to be adjusted in terms of measurements made by other procedures-,2
*From WHO 1979 Criteria Document for Sulfur Oxides and Particulate Matter.
1-151
-------
TABLE 1-24. WORLD HEALTH ORGANIZATION GUIDELINES FOR EXPOSURE
LIMITS CONSISTENT WITH THE PROTECTION OF PUBLIC HEALTH3'*
Concentration (ug/m )
Expected effects Sulfur dioxideSmoke
24-h mean 100-150 100-150
Annual..arithmetic mean 40-60 40-60
_%
*Values for sulfur dioxide and smoke as measured by OECD or British daily
smoke/sulfur dioxide method (Ministry of Technology, UK, 1966; Organization
for Economic Cooperation and Development, 1965). Adjustments may be necessary
where measurements are made by other methods. For example» smoke conccntra-t4onr.
*f-*eO-15D uy/iii convert to approximately 200-300 po/nT TSP and smptt levels ~
of 40-607up*rr-ctmven*Uto approximate4y-S^4£0-ug/m3 TSPr
*From WHO ir
1-152
-------
7/9/80
Chapter 1 Introduction, Summary, and Conclusions SO/PM
A
Corrigenda
Before listing specific errata (deletions/insertions) for Chapter 1 (Volume
I) of the April, 1980, External Review Draft of the EPA criteria document for
sulfur oxides and particulate matter, certain general comments should be noted
reguarding anticipated revisions in Chapter 1.
First, major revisions planned to be made in later current chapters (2-14) of
the document, as indicated in ensuing corrigenda materials, will also be appropriately
reflected in revisions to be made in Chapter 1. For example, certain major revisions
in the text of Chapter 3 noted in corrigenda comments for that chapter will be
appropriately reflected in revision of text in Section 1.3.2 (pg. 1-19 to 1-43).
This especially includes introductory materials (4 main points) to be inserted on
pg. 3-84 at the start of the discussion of comparison of particulate matter measurement
techniques, as noted later in corrigenda comments for Chapter 3. Similarly, revisions
noted in those corregenda comments to be made in Chapter 3 regarding the discussions
of specific studies comparing COM versus TSP and BS versus TSP measurement results
will be appropriately reflected in Chapter 1 revisions.
Other major revisions in Chapter 14, noted in the later corrigenda comments
for that chapter, will also be reflected in revisions of Section 1.5.5 (Community
Health Observation Studies) of Chapter 1. Of particular importance are major
changes to be made in Tables 1-19 to 1-21 (on pg. 1-140 to 1-42) and accompanying
text regarding summarization of various expert reviewers' evaluations of key quanti-
tative community health studies. Specific changes in those tables will include the
following:
13
(1) In Table 1-19, deletion of all entries except those for studies by Lawther ,
pop 11 c
Glasser and Greenburg , Martin and Bradley , and Martin .
-1-
-------
(2) In Table 1-20, deletion of entries for all studies except those by Greenburg
Lawther52'53, Martin16, Waller7, and Van der Lende74.
(3) In Table 1-21, deletion of entries for all studies except those by Douglas and
Waller90, Lambert and Reid28, Lunn et al96'97, Ferris43, Sawicki181, Mostardi,117'258
Shy215, and Rudnick182.
Discussion of tables 1-19 to 1-21, in the text on pg. 1-143 to 1-152 is to be
revised such that comments on quantitive air quality levels associated with observed
health effects will generally be in terms of the original (COM, BS, TSP) particulate
matter measurement units employed in specific studies summarized in the table,
except for comments on interpretative evaluations by particular expert reviewers
that involved "translation" of COM or BS units into TSP units. Further evaluative
comments are to be added on whether reasonable interconversions between COH, BS,
and TSP measurement units can be made and, if so, in what manner and under what
circumstances. The impact of such interconversion or lack of sound bases to do so
on interpretation of the epidemiology data base for SO /PM will then be taken into
/\
account in text revisions more specifically delineating key conclusions based on
the epidemiology literature.
The implications of those conclusions, and others based on information discussed
in Chapters 11, 12, and 13, for development of health criteria for sulfur oxide and
particulate matter are to be delineated in an integrative health summary and conclusions
chapter still in the process of being prepared for addition to the document.
Relevant text summarizing the most salient features of that chapter, once completed.
is to be added as the final portion of Chapter 1 (Volume 1).
-2-
------- |