xvEPA

                United States
                Environmental Protection
                Agency
                  Office of Research and
                  Development
                  Washington DC 20460
EPA/600/P-99/002aB   :
March 2001
Second External Review Draft
US EPA Office e-' Research and Davsioamanf
Air Quality Criteria for
Particulate Matter
                Volume  I
                                   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.

-------

-------
                                          EPA 600/P-99/002aB
                                               March 2001
                                   Second External Review Draft
  Air Quality Criteria for
      Particulate Matter
                 Volume I
                    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.
     National Center for Environmental Assessment
        Office of Research and Development
        U.S. Environmental Protection Agency
         Research Triangle Park, NC 27711
                                     Printed on Recycled Paper

-------
                                    Disclaimer

     This document is an external review draft for review purposes only and does not constitute
U.S. Environmental Protection Agency policy.  Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
March 2001
I-ii
DRAFT-DO NOT QUOTE OR CITE

-------
                                         Preface

      National Ambient Air Quality Standards (NAAQS) are promulgated by the United States
 Environmental Protection Agency (EPA) to meet requirements set forth in Sections 108 and 109
 of the U.S. Clean Air Act (CAA). Sections 108 and 109 require the EPA Administrator (1) to list
 widespread air pollutants that reasonably may be expected to endanger public health or welfare;
 (2) to issue air quality criteria for them that assess the latest available scientific information on
 nature and effects of ambient exposure to them; (3) to set "primary" NAAQS to protect human
 health with adequate margin of safety and to set "secondary" NAAQS to protect against welfare
 effects (e.g., effects on vegetation, ecosystems,  visibility, climate, manmade materials, etc); and
 (5) to periodically (every 5 years) review and revise, as appropriate, the criteria and NAAQS for
 a given listed pollutant or class of pollutants.
      The original NAAQS for particulate matter (PM), issued in 1971 as "total  suspended
 particulate" (TSP) standards, were revised in 1987 to focus on protecting against human health
 effects associated with exposure to ambient PM less than 10 microns (<10 //m) that are capable
 of being deposited in thoracic (tracheobronchial and alveolar) portions of the lower respiratory
 tract. Later periodic reevaluation of newly available scientific information, as presented in the
 last previous version of this "Air Quality Criteria for Particulate Matter" document published in
 1996, provided key scientific bases for PM NAAQS decisions published in July  1997.  More
 specifically, the PM10 NAAQS set in 1987 (150  /^g/m3, 24-h; 50 f^g/m3, annual average) were
 retained in modified form and new standards (65 ju.g/m3, 24-h;  15 yUg/m3, annual  average) for
 particles <2.5 ^.m (PM25) were promulgated in July 1997.
      This Second External Review Draft of revised Air Quality Criteria  for Particulate Matter
 assesses new scientific information that has become available mainly between early 1996 through
 December 2000. The present draft is being released for public comment  and review by the Clean
 Air Scientific Advisory Committee (CASAC) to obtain comments on the organization and
 structure of the document, the issues addressed,  the approaches employed in assessing and
 interpreting the newly available information on PM exposures and effects, and the key findings
 and conclusions arrived at as a consequence of this assessment.  Extensive additional pertinent
 information is expected to be published during the next 6 to 9 mo (including results from a vastly
 expanded EPA PM Research program and from  other federal and state agencies,  as well as other
March 2001                               I-iii        DRAFT-DO NOT QUOTE OR CITE

-------
partners in the general scientific community) and, as such, the findings and conclusions presented
in this draft document must be considered only provisional at this time. Public comments and
CASAC review recommendations will be taken into account, along with any pertinent newly
available information published or accepted for peer-reviewed publication by May/June 2001, in
making any appropriate further revisions to this document for incorporation into a Third External
Review Draft. That draft is expected to be released in September/October, 2001 for further
public comment and CASAC review (December 2001) in time for a final version to be
completed by early 2002. Evaluations contained in the present document will be drawn on to
provide inputs to associated PM Staff Paper analyses prepared by EPA's Office of Air Quality
Planning and Standards (OAQPS) to pose options for consideration by the EPA Administrator
with regard to proposal and, ultimately, promulgation of decisions on potential retention or
revision of the current PM NAAQS.
      Preparation of this document was coordinated by staff of EPA's National Center for
Environmental Assessment in Research Triangle Park (NCEA-RTP). NCEA-RTP scientific
staff, together with experts from other EPA/ORD laboratories and academia, contributed to
writing of document chapters, and earlier drafts of this document were reviewed by experts from
federal and state government agencies, academia, industry, and NGO's for use by EPA in support
of decision making on potential public health and environmental risks of ambient PM. The
document describes the nature, sources, distribution, measurement, and concentrations of PM in
outdoor (ambient) and indoor environments. It also evaluates the latest data on human exposures
to ambient PM and consequent health effects in exposed human populations (to support decision
making regarding primary, health-related PM NAAQS). The document also evaluates ambient
PM environmental effects on vegetation and ecosystems, visibility, and man-made materials, as
well as atmospheric PM effects on climate change processes associated with alterations in
atmospheric transmission of solar radiation or its reflectance from the Earth's surface or
atmosphere (to support decision making on secondary PM NAAQS).
      The NCEA of EPA acknowledges the contributions provided by authors, contributors, and
reviewers and the diligence of its staff and contractors in the preparation of this document.
 March 2001
I-iv
DRAFT-DO NOT QUOTE OR CITE

-------
                Air Quality Criteria for Particulate Matter
                                VOLUME I
   1.   INTRODUCTION  	1-1

   2.   PHYSICS, CHEMISTRY, AND MEASUREMENT OF
       PARTICULATE MATTER :	2-1

   3.   CONCENTRATIONS, SOURCES, AND EMISSIONS OF
       ATMOSPHERIC PARTICULATE MATTER  	3-1
       Appendix 3 A: Organic Composition of Particulate Matter	 3A-1
       Appendix 3B: Composition of Particulate Matter Source Emissions	3B-1

   4.   ENVIRONMENTAL EFFECTS OF PARTICULATE MATTER  	4-1
       Appendix 4A: Excerpted Key Points from the Executive Summary of
                   the World Meteorological Organization 1998 Assessment
                   of Stratospheric Ozone Depletion	 4A-1
       Appendix 4B: Excerpted Key Points from the Executive Summary of
                   the United Nations Environment Programme 1998
                   Assessment of Environmental Effects of Ozone
                   Depletion	4B-1
       Appendix 4C: Excerpted Key Points from the Executive Summary of
                   the Special Report of the Intergovernmental Panel on
                   Climate Change Working Group n on the Regional
                   Impacts of Climate Change: An Assessment of
                   Vulnerability	4C-1
       Appendix 4D: Excerpted Materials from the U.S. Global Change
                   Research Program Assessment Overview Report on
                   Climate Change Impacts on the United States and
                   Subsidiary Regional Assessment Reports	 4D-1
       Appendix 4E: Recent Model Projections of Excess Mortality Expected
                   in U.S. Cities During Summer and Winter Seasons
                   Because of Future Climate Change, Based on Kalkstein
                   and Greene (1997)	4E-1

   5.   HUMAN EXPOSURE TO PARTICULATE MATTER AND ITS
       CONSTITUENTS  	5-1
March 2001
I-v
DRAFT-DO NOT QUOTE OR CITE

-------
                Air Quality Criteria for Particulate Matter
                                  (cont'd)
                               VOLUME II
   6.  EPIDEMIOLOGY OF HUMAN HEALTH EFFECTS FROM
       AMBIENT PARTICULATE MATTER	 6-1
       Appendix 6A: Demographic and Pollution Data for 90-City Analysis of
                   NMMAPS Project	  6A-1
       Appendix 6B: Heart Rate Variability as a Predictor of Serious
                   Cardiac Outcomes	6B-1

   7.  DOSIMETRY OF PARTICULATE MATTER	7-1

   8.  TOXICOLOGY OF PARTICULATE MATTER  	8-1

   9.  INTEGRATIVE SYNTHESIS OF KEY POINTS: PARTICULATE
       MATTER ATMOSPHERIC SCIENCE, AIR QUALITY, HUMAN
       EXPOSURE, DOSIMETRY, AND HEALTH RISKS 	.. 9-1
       Appendix 9A: Key Quantitative Estimates of Relative Risk for
                   Particulate Matter-Related Health Effects Based on
                   Epidemiologic Studies of North American Cities
                   Assessed in the 1996 Particulate Matter Air Quality
                   Criteria Document	  9A-1

   EXECUTIVE SUMMARY	E-l
March 2001
I-vi
DRAFT-DO NOT QUOTE OR CITE

-------
                               Table of Contents
                                                                             'age
 List of Tables	I-xiv
 List of Figures	I-xviii
 Authors, Contributors, and Reviewers	I-xxv
 U.S. Environmental Protection Agency Project Team for Development of
  Air Quality Criteria for Particulate Matter	 I-xxxv

 1.  INTRODUCTION	   1_!
    1.1  LEGISLATIVE REQUIREMENTS	'.'.'.'.'.'.'.'.'.'.'.'.'.'.'. 1-1
    1.2  HISTORY OF PREVIOUS PARTICULATE MATTER CRITERIA
        AND NATIONAL AMBIENT AIR QUALITY STANDARDS REVIEWS	1-2
        1.2.1  The 1997 Particulate Matter National Ambient Air Quality
              Standards Revision	1_3
        1.2.2  Presidential Memorandum: Next Particulate Matter Review
              and Research	1_5
    1.3  CURRENT PARTICULATE MATTER CRITERIA AND NATIONAL
        AMBIENT AIR QUALITY STANDARDS REVIEW	1-6
        1.3.1  Criteria Review	j_g
        1.3.2  Methods and Procedures for Document Preparation 	1-8
        1.3.3  Approach	;	J.JQ
        1.3.4  Key Issues of Concern	1_11
    1.4  DOCUMENT CONTENT AND ORGANIZATION	     	1-13
    REFERENCES	'.'.'.'.'.'.'.'. 1-16

2.  PHYSICS, CHEMISTRY, AND MEASUREMENT OF PARTICULATE MATTER    2-1
    2.1   PHYSICS AND CHEMISTRY OF PARTICULATE MATTER	2-1
        2.1.1  Definitions	2-1
        2.1.2  Physical Properties and Processes	2-3
              2.1.2.1 Definitions of Particle Diameter  	2-3
              2.1.2.2 Aerosol Size Distributions	2-3
              2.1.2.3 Nuclei-Mode Particles	2-15
        2.1.3  Chemistry of Atmospheric Particulate Matter  	2-19
              2.1.3.1 Chemical Composition and Its Dependence on Particle Size  ....2-19
              2.1.3.2 Primary and Secondary Particulate Matter	2-20
              2.1.3.3 Particle-Vapor Partitioning 	2-27
              2.1.3.4 Removal Processes	2-30
              2.1.3.5 Particulate Matter and Welfare Effects  	2-33
        2.1.4  Summary	2-34
   2.2   MEASUREMENT OF PARTICULATE MATTER  	'.'.'.'.'.'.'.'.'.'.'.'. 2-34
        2.2.1  Problems in Measuring Particulate Matter	2-34
              2.2.1.1  Treatment of Semivolatile Components of Particulate Matter  ... 2-36
              2.2.1.2 Upper Cut Point	2-38
March 2001
I-vii
                                                DRAFT-DO NOT QUOTE OR CITE

-------
                                Table of Contents
                                      (cont'd)
              2.2.1.3  Cut Point for Separation of Fine-Mode and Coarse-Mode
                      Particulate Matter	2-38
              2.2.1.4  Treatment of Pressure, Temperature, and Relative Humidity	2-41
              2.2.1.5  No Way to Determine Accuracy for Ambient Particulate
                      Matter Mass Measurement	2-42
        2.2.2  Why Measure Particles	2-44
              2.2.2.1  Attainment of a Standard	2-44
              2.2.2.2  Implementation of a Standard  	2-44
              2.2.2.3  Determination of Health Effects  	2-45
              2.2.2.4  Determination of Ecological Effects	2-45
              2.2.2.5  Determination of Radiative Effects	2-45
              2.2.2.6  Particulate Matter Components/Parameters That Need
                      To Be Measured	2-45
        2.2.3  Problems Associated with Semivolatile Particulate Matter	2-46
              2.2.3.1  Particulate Nitrates		2-48
              2.2.3.2  Semivolatile Organic Compounds	2-51
              2.2.3.3  Use of Denuder Systems To  Measure Semivolatile
                      Compounds  	2-53
              2.2.3.4  Particle-Bound Water	2-66
        2.2.4  U. S. Environmental Protection Agency Monitoring Programs	2-71
              2.2.4.1  The Federal Reference Methods for Equilibrated Mass	2-71
              2.2.4.2  Speciation Monitoring  	2-74
        2.2.5  Continuous Monitoring	•	2-86
              2.2.5.1  Tapered Element Oscillating Microbalance	 2-86
              2.2.5.2  Real-Time Total Ambient Mass Sampler 	2-89
              2.2.5.3  Continuous Ambient Mass Monitor  	2-89
              2.2.5.4  Light Scattering	2-91
              2.2.5.5  Beta-Gauge Techniques   	2-92
              2.2.5.6  Measurements of Individual Particles	2-92
              2.2.5.7  Automated Fine Particulate Nitrate	2-94
              2.2.5.8  Semi-continuous Carbon Analysis	2-95
              2.2.5.9  Determination of Aerosol Surface Area in Real Time	2-96
        2.2.6  Data Quality 	,	2-97
              2.2.6.1  Errors in Gravimetric Analyses	2-97
              2.2.6.2  Quality Assurance	 2-98
   2.3  SUMMARY	2-98
   REFERENCES	 2-106

3. CONCENTRATIONS, SOURCES, AND EMISSIONS OF ATMOSPHERIC
   PARTICULATE MATTER	3-1
March 2001
I-viii
DRAFT-DO NOT QUOTE OR CITE

-------
                                Table of Contents
                                     (cont'd)
   3.1
   3.2
INTRODUCTION	3-1
TRENDS AND PATTERNS IN AMBIENT PM25 CONCENTRATIONS	3-2
        3.2.1  Daily and Seasonal Variability	3-12
        3.2.2  Diurnal (Circadian) Variability	3-19
        3.2.3  Relations Among Particulate Matter in Different Size Fractions	3-24
        3.2.4  Relations Between Mass and Chemical Component Concentrations	3-26
        3.2.5  Spatial Variability	3-28
   3.3  SOURCES OF PRIMARY AND SECONDARY PARTICULATE MATTER .... 3-34
        3.3.1  Source Contributions to Ambient Particulate Matter	3-38
        3.3.2  Long-Range Transport of Particulate Matter from Sources Outside
              the United States	3-44
   3.4  EMISSIONS ESTIMATES AND THEIR UNCERTAINTIES 	3-49
        3.4.1  Emissions Estimates for Primary Particulate Matter and Sulfur Dioxide,
              Nitrogen Oxides, and Volatile Organic Compounds in the United States .. 3-49
        3.4.2  Uncertainties of Emissions Inventories	3-54
   3.5  SUMMARY AND CONCLUSIONS	3-57
   REFERENCES  	3-61

   APPENDK 3 A: Organic Composition of Particulate Matter 	  3A-1

   APPENDIX 3B: Composition of Particulate Matter Source Emissions	3B-1

4.  ENVIRONMENTAL EFFECTS OF PARTICULATE MATTER	4-1
   4.1  INTRODUCTION	4-1
   4.2  EFFECTS ON VEGETATION AND ECOSYSTEMS 	4-1
        4.2.1  Direct Effects of Particulate Matter on Individual Plant Species	4-5
              4.2.1.1   Effects of Coarse Particles	4-6
              4.2.1.2  Effects of Fine Particles	4-9
        4.2.2  Particulate Matter Effects on Natural Ecosystems 	4-20
              4.2.2.1   Introduction	4-20
              4.2.2.2  Ecosystem Responses to Stress	4-22
              4.2.2.3  Ecosystem Response to Direct Plant Effects	4-25
              4.2.2.4  Indirect Effects of Particulate Matter in Ecosystems	4-33
        4.2.3  Ecosystem Goods and Services and Their Economic Valuation  	4-81
   4.3  EFFECTS ON VISIBILITY	4-85
        4.3.1  Introduction	4-85
        4.3.2  Factors Affecting Atmospheric Visibility	4-85
              4.3.2.1   Anthropogenic Pollutants	4-85
              4.3.2.2  Human Vision 	;	4-86
              4.3.2.3   Characteristics of the Atmosphere	4-86
        4.3.3  Optical Properties of Particles 	4-88
March 2001
                               I-ix
DRAFT-DO NOT QUOTE OR CITE

-------
                                 Table of Contents
                                       (cont'd)
                                                                               Page
        4.3.4  Effect of Relative Humidity on Particle Size and Light-Scattering
               Properties  	4-91
        4.3.5  Measures of Visibility	4-94
               4.3.5.1  Human Observations	4-94
               4.3.5.2  Light-Extinction Coefficient and Parameters Related to the
                       Light-Extinction Coefficient	4-94
               4.3.5.3  Light-Scattering Coefficient	4-96
               4.3.5.4  Fine Particulate Matter Concentrations	4-96
               4.3.5.5  Discoloration	4-98
        4.3.6  Visibility Monitoring Methods and Networks	4-99
        4.3.7  Visibility Modeling	4-102
               4.3.7.1  Regional Haze  	4-102
               4.3.7.2  Plume Models  	4-106
               4.3.7.3  Photographs	4-107
        4.3.8  Trends in Visibility Impairment	 4-108
        4.3.9  Economics of Particulate Matter Visibility Effects  	4-111
   4.4  EFFECTS ONMATERIALS	4-114
        4.4.1  Effects of Particles  and Sulfur Dioxide on Man-Made Surfaces  	4-115
               4.4.1.1  Metals	4-115
               4.4.1.2  Painted Finishes	4-116
               4.4.1.3  Stone and Concrete  	4-119
        4.4.2  Soiling and Discoloration of Man-Made Surfaces	4-126
               4.4.2.1  Stones and Concrete	4-126
               4.4.2.2  Household and Industrial Paints	4-127
   4.5  EFFECTS OF ATMOSPHERIC PARTICULATE MATTER ON CLIMATE
        CHANGE PROCESSES AND THEIR POTENTIAL HUMAN HEALTH
        AND ENVIRONMENTAL IMPACTS  	4-128
        4.5.1  Solar Ultraviolet Radiation Transmission Impacts on Human Health
               and the Environment: Atmospheric Particulate Matter Effects	4-130
               4.5.1.1 Bases for  Concern Regarding Increased Ultraviolet
                      Radiation Transmission	4-130
               4.5.1.2 Airborne Particle Impacts on Atmospheric Ultraviolet
                      Radiation Transmission	4-133
        4.5.2  Global Warming Processes, Human Health and Environmental
               Impacts, and Atmospheric Particle Roles	4-137
               4.5.2.1 Bases for  Concern Regarding Global Warming and Climate
                      Change	4-137
               4.5.2.2 Airborne Particle Relationships to Global Warming and
                      Climate Change	4-150
   4.6  SUMMARY	,	4-156
        4.6.1  Particulate Matter Effects on Vegetation and Ecosystems	4-156
March 2001
I-x
DRAFT-DO NOT QUOTE OR CITE

-------
                                 Table of Contents
                                       (cont'd)
         4.6.2  Particulate Matter-Related Effects on Visibility	4-161
         4.6.3  Particulate Matter-Related Effects on Materials	4-162
         4.6.4  Effects of Particulate Matter on the Transmission of Solar Ultraviolet
               Radiation and Global Wanning Processes	                   4-163
    REFERENCES  	.	" " ' 4_165

    APPENDIX 4 A: Excerpted Key Points from the Executive Summary of the
                   World Meteorological Organization 1998 Assessment of
                   Stratospheric Ozone Depletion	  4A-1

    APPENDIX 4B: Excerpted Key Points from the Executive Summary of the
                   United Nations Environment Programme 1998 Assessment
                   of Environmental Effects of Ozone Depletion	 4B-1

    APPENDIX 4C: Excerpted Key Points from the Executive Summary of the
                   Special Report of the Intergovernmental Panel on Climate
                   Change Working Group II on the Regional Impacts of Climate
                   Change:  An Assessment of Vulnerability 	4C-1

    APPENDIX 4D: Excerpted Materials from the U.S. Global Change Research
                   Program Assessment Overview Report on Climate Change
                   Impacts on the United States and Subsidiary Regional
                   Assessment Reports 	       4D-1

    APPENDIX 4E:  Recent Model Projections of Excess Mortality Expected in
                   U.S. Cities During Summer and Winter Seasons Because of
                   Future Climate Change, Based on Kalkstein and Greene (1997)	4E-1

5. HUMAN EXPOSURE TO PARTICULATE MATTER AND ITS CONSTITUENTS     5-1
    5.1  INTRODUCTION	5.!
        5.1.1  Purpose	5_j
        5.1.2  Particulate Matter Mass and Constituents	5-2
        5.1.3  Relationship to Past Documents	                       5.3
    5.2  STRUCTURE FOR THE CHAPTER  	'.'.'.'.'.'.'.'.'.'.'.'.'.	5-4
    5.3  BASIC CONCEPTS OF EXPOSURE	.....5-4
        5.3.1  Components of Exposure	5.4
        5.3.2  Methods To Estimate Personal Exposure	5.5
              5.3.2.1  Direct Measurement Methods 	5.7
              5 3.2.2  Indirect Methods (Modeling Methods)  	5-8
              5.3.2.3  Methods of Estimating Personal Exposure to Ambient
                      Particulate Matter	5_14
March 2001
I-xi
                                                  DRAFT-DO NOT QUOTE OR CITE

-------
                                Table of Contents
                                      (cont'd)
   5.4   SUMMARY OF PARTICULATE MATTER MASS DATA 	5-19
        5.4.1  Types of Particulate Matter Measurement Studies	5-19
        5.4.2  Available Data 	5-19
              5.4.2.1  Personal Exposure Data	,	5-19
              5.4.2.2  Microenvironmental Data 	5-23
              5.4.2.3  Interpretation of Particulate Matter Exposure Data	5-24
        5.4.3  Factors Influencing and Key Findings on Particulate Matter Exposures ...  5-30
              5.4.3.1  Correlations of Personal/Microenvironmental Particulate
                      Matter with Ambient Particulate Matter 	5-30
              5.4.3.2  Factors That Affect Correlations	5-47
              5.4.3.3  Impact of Ambient Sources on Exposures to Particulate Matter ..  5-64
              5.4.3.4  Correlations of Particulate Matter with Other Pollutants  	5-67
   5.5.  SUMMARY OF PARTICULATE MATTER CONSTITUENT DATA	5-70
        5.5.1  Introduction	5-70
        5.5.2  Monitoring Studies That Address Particulate Matter Constituents  	5-70
        5.5.3  Key Findings	5-73
              5.5.3.1  Correlations of Personal and Indoor Concentrations with
                      Ambient Concentrations of Particulate Matter Constituents	5-73
        5.5.4  Factors Affecting Correlations Between Ambient Measurements and
              Personal or Microenvironmental Measurements of Particulate Matter
              Constituents	5-79
        5.5.5  Limitations of Available Data	5-80
   5.6.  IMPLICATIONS OF USING AMBIENT PARTICULATE MATTER
        CONCENTRATIONS IN EPIDEMIOLOGIC STUDIES OF PARTICULATE
        MATTERHEALTH EFFECTS	5-80
        5.6.1  Potential Sources of Error Resulting from Using Ambient Particulate
              Matter Concentrations in Epidemiologic Analyses 	5-81
        5.6.2  Associations Between Personal Exposures and Ambient Particulate
              Matter Concentrations	5-83
        5.6.3  Role of Compositional Differences hi Exposure Characterization for
              Epidemiology	•	5-85
        5.6.4  Role of Spatial Variability in Exposure Characterization for
              Epidemiology	5-87
        5.6.5  Analysis of Exposure Measurement Error Issues in
              Particulate Matter Epidemiology	5-88
              5.6.5.1  Analysis of Exposure Measurement Errors in Time-Series
                      Studies	5-89
              5.6.5.2  Analysis of Exposure Measurement Errors in
                      Long-Term Epidemiology Studies	5-92
              5.6.5.3  Conclusions from Analysis of Exposure Measurement Errors
                      on Particulate Matter Epidemiology  	5-93
March 2001
I-xii
DRAFT-DO NOT QUOTE OR CITE

-------
                             Table of Contents
                                  (cont'd)
   5.7  SUMMARY OF KEY FINDINGS AND LIMITATIONS	5-95
   REFERENCES 	5_102
March 2001
I-xiii
DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Tables

Number                                                                          Page

2-1     Comparison of Ambient Particles, Fine Mode and Coarse Mode  	2-35

2-2     Particulate Matter Components/Parameters of Interest for Health, Ecological,
        or Radiative Effects; for Source Category Apportionment Studies; or for Air
        Quality Validation Studies 	2-46
3-1
3-2
3-3


3-4


3-5



3-6


3-7

3-8

3-9

3-10
Gross Chemical Composition of PM25 Particles Obtained in Rural Areas
of the Eastern and Western United States by the IMPROVE Network and in
Mixed Rural, Suburban, and Urban Areas Obtained by Studies Summarized
in the 1996 Particulate Matter Air Quality Criteria Document	3-3

Distribution of Ratios of PM2.5 to PM10 and Correlations Between PM2.5
and PMIO, PM2.5, and PM(10.2.5), and PM(10.2.5) and PM10 Found at Collocated
Monitoring Sites in Seven Aerosol Characteristic (U.S. Environmental
Protection Agency/Health Effects Institute) Regions in 1999  	3-25

Concentrations of PM25 and Selected Elements in the PM2 5 Size Range
and Correlations Between Elements and PM2 5 Mass	3-27

Pearson Correlation Coefficients for the Spatial Variation of PM2 5
Concentrations in Selected Metropolitan Statistical Areas	3-29
Correlation Coefficients for Spatial Variation of PM2 5 Mass and Different
Sources for Pairs of Sampling Sites in the South Coast Air Basin
(1986)  	
                             3-33
Correlation Coefficients for Spatial Variation of PM2.5 Mass and Different
Components for Pairs of Sampling Sites in Philadelphia (1994)	3-34

Constituents of Atmospheric Particles and Their Major Sources	3-35

Receptor Model Source Contributions to PM2 5	3-39

Receptor Model Source Contributions to PM10 	3-40

Nationwide Changes in Ambient Concentrations and Emissions of PM10
and Gaseous Precursors to Secondary Particulate Matter
from 1989 to 1998	3-54
 3A-1    Particulate Organic and Elemental Carbon Concentrations Based on
         Studies Published After 1995	 3A-2
March 2001
                                 I-xiv
DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Tables
                                       (cont'd)
 Number
 3A-2    Particulate Organic Compound Concentrations Based on Studies Published
         after 1990 at Selected Sites	 3A-6

 3B-1    Average Abundances of Major Elements in Soil and Crustal Rock	3B-2

 3B-2    Summary of Particle-Phase Organic Compounds Present in Fine Particle
         Road Dust Sample	33.5

 3B-3    Composition of Fine Particles Released by Various Stationary Sources in
         the Philadelphia Area 	33.7

 3B-4a   Organic and Elemental Carbon Fractions of Diesel and Gasoline Engine
         Particulate Matter Exhaust	        3B-13

 3B-4b   Contribution of Organic Carbon to Particulate Matter Carbon Emissions
         in Motor Vehicle Exhaust Collected from Vehicles Operated on Chassis
         Dynamometers	3B-14

 3B-5    Emission Rates for Constituents of Particulate Matter from Gasoline and
         Diesel Vehicles	3B-15

 3B-6    Summary of Particle-Phase Organic Compounds Emitted from Motor
         Vehicles	                     3B-17

 3B-7    Mass Emissions, Organic Carbon, and Elemental Carbon Emissions from
         Residential Combustion of Wood	3B-20

 3B-8    Summary of Particle-Phase Organic Compounds Emitted from the
         Combustion of Wood in Fireplaces	       3B-21

 3B-9    Mean Aerosol Composition at Tropical Site Affected Heavily by Biomass
         Burning Emissions 	3B-23

 4-1      Relative Importance of Wet, Dry, Particulate, and Total Deposition to Three
        Forest Sites	         4.4

4-2     Ecosystem Services	4_2Q

4-3     Ecosystem Functions Impacted by Air Pollution Effects on Temperate Forest
        Ecosystems „	   4_2y
March 2001
I-xv       DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Tables
                                       (cont'd)

Number

4-4     Types of Plant Responses to Ultraviolet-B Radiation  	4-40

4-5     Nitrogen-Saturated Forests in North America, Including Estimated Nitrogen
        Inputs and Outputs 	4-44

4-6     Primary Goods and Services Provided by Ecosystems	4-82

4-7     Averaged Regional PM2 5 Mass and Extinction Summaries for the Years
        1988 to 1996	'.	....-'	4-101

4-8     Corrosive Effects of Particulate Matter and Sulfur Dioxide on Metals  	4-117

4-9     Effects of Particulate Matter and Sulfur Dioxide on Stone  	4-120

4D-1   Range of Projected Warming in the 21st Century  	  4D-5

4D-2   Types of Water Concerns Projected To Be Important for U.S. Regions
        Consequent to Future Climate Change	  4D-11

4D-3   Projected Future Climate-Change-Induced Impacts on Types of Ecosystems
        of Concern to Different U.S. Regions	  4D-12

4E-1   Modeled Projections of Direct Human Health Impacts of Climate Change:
        Estimated Total Excess Mortality in U.S. Urban Areas For an Average
        Summer Season, Assuming Full Acclimatization3	4E-2

4E-2   Modeled Projections of Direct Human Health Impacts of Global Climate
        Change: Estimated Total Excess Mortality in U.S. Urban Areas for an
        Average Winter Season, Assuming Full Acclimatization3	4E-4

5-1     Classes of Particulate Matter Exposure and Concentration Definitions	5-6

5-2     Activity Pattern Studies Included in the Consolidated Human Activity Database .. 5-11

5-3     Personal Exposure Models for Particulate Matter 	5-12

5-4      Summary of Recent Personal Exposure Studies	5-20

5-5      Summary of Recent Microenvironmental Measurement Studies	5-25

5-6     Papers Interpreting Particulate Matter Exposure Studies	5-29
 March 2001
I-xvi
DRAFT-DO NOT QUOTE OR CITE

-------
                                    List of Tables
                                        (cont'd)
 Number
                                                                                   age
 5-7     Personal Monitoring Studies for Particulate Matter:  Measured Concentrations
         and Correlation Coefficients	5.33

 5-8     Volume Mean Diameter and Maximum PM2 5 Concentrations of
         Indoor Particle Sources	5_62

 5-9     Correlations Between Personal PM2 5 and Ambient Pollutant Concentrations	5-68

 5-10     Correlations Between Hourly Personal PM2 5 and Gaseous Pollutants	5-69

 5-11     Studies That Have Measured Particulate Matter Constituents in
         Personal Exposure Samples	5.71

 5-12     Studies That Have Measured Particulate Matter Constituents in
         Microenvironmental Samples	5.72

 5-13     Summary Statistics for Personal, Indoor, and Outdoor Concentrations of
         Selected Aerosol Components in Two Pennsylvania Communities	5-75

 5-14     Statistical Correlation of Outdoor Versus Indoor Concentration for
         Measured Species	5.77
March 2001
I-xvii
                                                    DRAFT-DO NOT QUOTE OR CITE

-------
                                    List of Figures

Number                                                                           2^ge

2-1     Number of particles as a function of particle diameter	2-5

2-2     Particle volume distribution as a function of particle diameter 	2-6

2-3     Distribution of coarse, accumulation, and nuclei- or ultrafine-mode
        particles by three characteristics: number, surface area, and volume
        for the grand average continental size distribution	2-7

2-4     Volume size distribution, measured in traffic, showing fine-mode and
        coarse-mode particles and the nuclei and accumulation modes within
        the fine-particle mode	2-8

2-5     An idealized distribution of ambient particulate matter showing
        fine-mode particles and coarse-mode particles and the fractions
        collected by size-selective samplers	2-11

2-6     Specified particle penetration through an ideal inlet for five different
        size-selective sampling criteria	•	2-12

2-7     Comparison of penetration  curves for two PM10 beta gauge samplers using
        cyclone inlets	• • • •	2-15

2-8     Particle growth curves showing fully reversible hygroscopic growth of
        sulfuric acid particles, deliquescent growth of ammonium sulfate particles
        at about 80% relative humidity (RH), hygroscopic growth of ammonium
        sulfate solution droplets at RH greater than 80%, and hysteresis (the
        droplet remains supersaturated as the RH decreases below 80%) until the
         crystallization point is reached	2-31

2-9     Theoretical predictions and experimental measurements of growth of
        ammonium hydrogen sulfate particles at relative humidity between
        95 and 100%	2-32

2-10    Schematic showing major nonvolatile and semivolatile components of PM2.5  .... 2-37

2-11   Particulate matter concentrations in Spokane, WA, during August 30, 1996,
         dust storm	2-40

2-12   Amount of ammonium nitrate volatilized from Teflon filters, expressed as
         a percentage of the measured PM25 mass, for the Southern California Air
         Quality and CalTech studies, for spring and fall sampling periods 	2-49
 March 2001
I-xviii
DRAFT-DO NOT QUOTE OR CITE

-------
                                    List of Figures
                                        (cont'd)
 Number
                                                                                    age
 2-13
 2-14
 2-15
2-16


3-1

3-2

3-3


3-4a

3-4b

3-5


3-6

3-7a



3-7b
 Aerosol water content expressed as a mass percentage, as a function of
 relative humidity	2-69

 This thermogram, for a sample containing rock dust (a carbonate source) and
 diesel exhaust, shows three traces that correspond to temperature, filter
 transmittance, and flame ionization detector response	2-82

 Comparison of mass measurements with collocated real-time total ambient
 mass sampler (real-time data), the Particle Concentration-Brigham Young
 University Organic Sampling System (1-hour data), the Federal Reference
 Method PM2 5 sampler (24-hour data), and a conventional tapered-element
 oscillating microbalance monitor (real-time data) 	2-90

 Size distribution of particles divided by chemical classification into organic,
 marine, and crustal	2-93

 1999 annual mean PM,0 concentrations	3.4

 Nationwide trend in ambient PM10 concentration from 1989 through 1998	3-5

 Trend in PM,0 annual mean concentrations by U.S. Environmental Protection
 Agency region, 1989 through 1998	3-6

 1999 annual mean PM25 concentrations	3-7

 1999 98th percentile 24-hour average PM25 concentrations	3-8

 Annual distribution of 24-hour average PM25 concentrations observed in
 U.S. and Canadian health studies	3.9
1999 annual mean PM,
                    (10-2.5)
concentrations	3-12
Quarterly distributions of 24-hour average PM25 concentrations obtained in
eight eastern U.S. cities by the nationwide SLAMS/NAMS network of PM25
Federal Reference Method monitors during 1999  	3-13

Quarterly distributions of 24-hour average PM2 5 concentrations obtained in six
central and mountain U.S. cities by the nationwide SLAMS/NAMS network
of PM25 Federal Reference Method monitors during 1999  	3-14
March 2001
                                 I-xix
                    DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Figures
                                       (cont'd)
Number
                                       Page
3-7c    Quarterly distributions of 24-hour average PM2 5 concentrations obtained in
        seven western U.S. cities and San Juan, PR, by the nationwide
        SLAMS/NAMS network of PM2 5 Federal Reference Method monitors
        during 1999	3-15

3-8     Concentrations of PM2 5 and PM10 measured in the four Metropolitan Acid
        Aerosol Characterization Study cities	3-16

3-9     Frequency distribution of 24-hour average PM2.5 concentrations measured at
        the PBY site in southwestern Philadelphia	3-17

3-10    Concentrations of 24-hour average PM25 measured at the U.S. Environmental
        Protection Agency site in Phoenix, AZ	 3-18

3-11    Frequency distribution of 24-hour average PM2 5 concentrations measured at
        the U.S. Environmental Protection Agency site in Phoenix, AZ 	3-18

3-12    Frequency distribution of 24-hour average PM2 5 measurements obtained
        from all California Air Resources Board dichotomous sampler sites from
        1989 to 1998	3-20

3-13    Frequency distribution of 24-hour average PM(10.2 5) concentrations obtained
        from all California Air Resource Board Dichotomous sampler sites from
        1989 to 1998	3-20

3-14    Concentrations of 24-hour average PM2 5 measured at the Riverside-Rubidoux
        site	.'	3-21

3-15    Frequency distribution of 24-hour average PM2 5 concentrations measured at
        the Riverside-Rubidoux site	3-21

3-16    rntraday variability of hourly average PM2 5 concentrations across the
        United States	'.	3-23

3-17    PM25 chemical components in downtown Los Angeles and Burbank (1986)
        have similar characteristics	3-31

3-18    Concentrations of PM25 chemical components in Rubidoux and downtown
        Los Angeles (1986) ..'	3-32
March 2001
I-xx
DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Figures
                                        (cont'd)
Number
3-19    Monthly average Saharan dust components of the aerosol sampled in Miami,
        FL, during 1974 to 1996	3-46

3-20    PM2 5 and PM10 concentrations measured at Chilliwack Airport, located
        in northwestern Washington State, just before and during the Asian desert
        dust episode of April and May 1998	3-47

3-21    Time series of 24-hour average PMIO concentrations observed in the
        Rio Grande Valley during May 1998 	3-48

3-22    PMIO concentrations observed in St. Louis, MO, during May 1998	3-48

3-23    1998 directly emitted national particulate matter (PM25) emissions by
        principal source categories for nonfugitive dust sources  	3-50

3-24    Nationwide emissions of sulfur dioxide, nitrogen oxides, volatile organic
        compounds, and ammonia from various  source categories  	3-52

3B-1    Size distribution of particles generated in a laboratory resuspension chamber  .... 3B-3

3B-2    Size distribution of California source emissions, 1986  	3B-4

3B-3    Chemical abundances for PM2 5 emissions from paved road dust in
        Denver, CO	3B-5

3B-4    Chemical abundances for PM2 5 emissions from wood burning in Denver, CO ... 3B-19

4-1     Effects of environmental stress on forest trees are presented on a hierarchial
        scale for the leaf, branch, tree, and stand levels  of organization 	4-23

4-2     Nitrogen cycle	4-45

4-3     Diagrammatic overview of nitrogen excess in North America	4-48

4-4     Schematic of sources and sinks of hydrogen ions in a forest	4-56

4-5     Calcium deposition in >2-//m particles, <2-//m particles, and wet forms
        and leaching in the Integrated Forest Study sites	4-64

4-6     Magnesium deposition in >2-//m particles, <2-yum particles, and wet forms
        and leaching in the Integrated Forest Study sites 	4-65
March 2001
I-xxi
DRAFT-DO NOT QUOTE OR CITE

-------
                                    List of Figures
                                         (cont'd)

Number
4-7      Potassium deposition in >2-yum particles, <2-fj.ro. particles, and wet forms
         and leaching in the Integrated Forest Study sites  	4-66

4-8      Base cation deposition in >2-//m particles, <2-//m particles, and wet forms
         and leaching in the Integrated Forest Study sites	4-67

4-9      Total cation leaching balanced by sulfate and nitrate, estimated from
         particulate deposition and by other sources of sulfate and nitrate and by
         other anions in the Integrated Forest Study sites	 4-68

4-10     Soil exchangeable Ca2+ pools and net annual export of Ca2+ in the Integrated
         Forest Study sites  	4-69

4-11     Soil exchangeable Mg2"1" pools and net annual export of Mg2"1" in the Integrated
         Forest Study sites  	4-70

4-12     Soil exchangeable K2+ pools and net annual export of K2+ in the Integrated
         Forest Study sites  	4-71

4-13     Simulated soil solution  mineral acid anions and base cations in the red spruce
         site with no change, 50% N and S deposition, and 50% base cation deposition  ... 4-74

4-14     Simulated soil solution  Al and soil base saturation in the red spruce site
         with no change, 50% N and S deposition, and 50% base cation deposition	4-75

4-15     Simulated soil solution  mineral acid anions and base cations in the Coweeta
         site with no change, 50% N and S deposition, and 50% base cation deposition  ... 4-76

4-16     Relationship of plant nutrients and trace metals with vegetation	4-80

4-17     Light reflected from a target toward an observer	4-87

4-18     Humidity effect on scattering coefficients computed for internal and external
         mixtures of the mixed-salt aerosol:  Na2SO4(x2 = 0.5)-(NH4)2 SO4 (x3 = 0.5),
         for two dry-salt particle size distributions, where x is the mass fraction of
         the dry solutes	4-92

4-19     Scattering ratios for different chemical compositions as a function
         of relative humidity	4-93

4-20     Comparison of extinction and visual range  	4-95
March 2001
I-xxii
DRAFT-DO NOT QUOTE OR CITE

-------
                                    List of Figures
                                        (cont'd)
 Number
                                                                                    age
 4-21a,b  Plots of the 10th, 50th, and 90th percentile groups for PM25 and deciview
         at the Badlands National Park	4.97

 4-22     Reduction in visual range as a function of increasing fine and coarse
         particle concentrations	4-98

 4-23     Five-season average haze patterns for the periods 1980 to 1985
         and 1990 to 1995	4_110

 4-24     Secular haze trends for 1940 to 1990 for six regions, summer and winter
         in the 75th percentile	4-112

 4-25     Processes involved in stratospheric ozone depletion because of man's
         production of chlorofluorcarbons, halons, and other trace gases are shown	4-131

 4-26     Bases for concern about global warming and climate change effects on the
         environment and human health	4-142

 4-27     Estimated global mean radiative forcing exerted by gas and various particle
         phase species from 1850 to 1950 .	4-154

 4D-1     Records of Northern Hemisphere surface temperatures, carbon dioxide
         concentrations, and carbon emissions show a close correlation	  4D-2

 4D-2     Simulation of decadal average changes in temperature from leading
         climate models on historic and projected changes in carbon dioxide
         and sulfate atmospheric concentrations  	  4D-4

 4D-3     Historic and projected changes in sea level based on the Canadian
         and Hadley model simulations	  4D-6

 4D-4     This map is a preliminary classification of annual shoreline erosion
         throughout the United States, in coarse detail and resolution	  4D-6

 4D-5     Breakout of United States regions as evaluated by the U.S. Global Change
         Research Program, based on the overview assessment depiction	  4D-9

4D-6     Projected climate-change impacts in the Mid-Atlantic Region of the
         United States	 4D-13
March 2001
I-xxiii
DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Figures
                                        (cont'd)

Number                                                                           Page

5-1      Comparison of correlation coefficients for longitudinal analysis of
         personal exposure versus ambient concentrations for individual subjects
         for PM2.5 and sulfate	5-42

5-2      Personal exposure versus ambient concentrations for PM2 5 and sulfate 	5-44

5-3      Regression analyses of aspects of daytime personal exposure to PM10 estimated
         using data from the Particle Total Exposure Assessment Methodology study	5-46

5-4      Air-exchange rates measured in homes throughout the United States	5-53

5-5      Box plots of hourly air-exchange rates stratified by season in Boston, MA,
         during 1998	5-54

5-6      Geometric mean infiltration factor (indoor/outdoor ratio) for hourly nighttime,
         nonsource data for two seasons	5-55

5-7      Comparison of deposition rates from this study with literature values	5-57

5-8      Penetration efficiencies and deposition rates from models of nightly
         average data	5-58

5-9      Mean hourly indoor/outdoor particle concentration ratio from an unoccupied
         residence in Fresno, CA, during spring 1999	5-63

5-10     Personal versus outdoor SO4= in State College, PA	5-76

5-11     Plots of nonambient exposure to PM10, daytime individual values from the
         Particle Total Exposure Assessment Methodology study data and daily-average
         values from THEES data 	5-86
March 2001
I-xxiv
DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                           CHAPTER 1. INTRODUCTION
Principal Authors

Dr. Lester D. Grant—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

Dr. Dennis J. Kotchmar—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC  27711
            CHAPTER 2. PHYSICS, CHEMISTRY, AND MEASUREMENT
                            OF PARTICULATE MATTER
Principal Authors

Dr. William Wilson—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

Dr. Candis S. Claiborn—Washington State University, Laboratory for Atmospheric Research,
Department of Civil and Environmental Engineering, P.O. Box 642910, Pullman, WA  99164

Contributing Authors

Dr. Barbara J. Turpin, The State University of New Jersey, Rutgers, Department of
Environmental Sciences and Rutgers Cooperative Extension, New Brunswick, NJ  08901-8551

Dr. James J. Schauer, University of Wisconsin, College of Engineering, Department of Civil and
Environmental Engineering, Madison, WI 53706

Contributors and Reviewers

Dr. Timothy Buckley—Johns Hopkins University, Department of Environmental Health
Sciences, 615 North Wolfe Street,, Baltimore, MD  21205

Ms. Lee Byrd—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

Dr. Lyle Chinkin—Sonoma Technology, 1360 Redwood Way, Suite C,
Petaluma, CA 94549
March 2001
I-xxv
DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
Contributors and Reviewers
(cont'd)

Dr. Steven Colome—Integrated Environmental Services, 5319 University Drive, #430,
Irvine, CA 92612

Dr. Delbert Eatough—Brigham Young University, E 114 BNSN,
Department of Chemistry and Biochemistry, Provo, UT  84602

Dr. Edward O. Edney—National Exposure Research Laboratory (MD-84)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Mr. William Ewald—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Sheldon Friedlander—University of California at Los Angeles, Department of Chemical
Engineering, 5531 Boelter Hall, Los Angeles, CA  90095

Dr. Judith Graham—National Exposure Research Laboratory (MD-75),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Lynn Hildemann—Stanford University, Civil and Environmental Engineering Department,
Stanford, CA 94305

Mr. Jim Homolya—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Rudolf Husar—CAPITA, Washington University, Campus Box 1124,
One Brookings Drive, St. Louis, MO 63130

Dr. Charles W. Lewis—National Exposure Research Laboratory (MD-47),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Mr. Scott Mathias—Office of Air Quality Planning and Standards (MD-15)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Mr. Tom McCurdy—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Mr. Frank McElroy—National Exposure Research Laboratory (MD-46),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001
I-xxvi
DRAFT-DO NOT QUOTE OR CITE

-------
                      Authors, Contributors, and Reviewers
                                       (cont'd)
 Contributors and Reviewers
 (cont'd)

 Dr. Haluk Ozkaynak—National Exposure Research Laboratory (MD-56),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Mr. Tom Pace—Office of Air Quality Planning and Standards (MD-14),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Joseph Pinto—National Center for Environmental Assessment (MD-52),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Richard Poirot—VT Air Program, Building 3 South, 103 South Main Street,
 Waterbury, VT  05671

 Dr. Linda Sheldon—National Exposure Research Laboratory (MD-77),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Helen Suh—Harvard School of Public Health, 665 Huntington Avenue,
 Boston, MA 02461

 Mr. Robert Wayland—Office of Air Quality Planning and Standards (MD-15),
 U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

 Dr. Russell Weiner—National Exposure Research Laboratory (MD-46)
 U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

 Mr. Dane Westerdahl—California Air Resources Board, 2020 L Street,
 Sacramento, CA 95814
                 CHAPTER 3. CONCENTRATIONS, SOURCES, AND
              EMISSIONS OF A TMOSPHERIC PARTICULA TE MA TTER
Principal Author

Dr. Joseph P. Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001
I-xxvii
                                                  DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
Contributing Authors

Dr. Barbara J. Turpin, The State University of New Jersey, Rutgers, Department of
Environmental Sciences and Rutgers Cooperative Extension, New Brunswick, NJ 08901-8551

Dr. James J. Schauer, University of Wisconsin, College of Engineering, Department of Civil and
Environmental Engineering, Madison, WI 53706

Contributors and Reviewers

Dr. Timothy Buckley—Johns Hopkins University, Department of Environmental Health
Sciences, 615 North Wolfe Street, Baltimore, MD 21205

Ms. Lee Byrd—Office of Air Quality Planning and Standards (MD-14),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Lyle Chinkin—Sonoma Technology, 1360 Redwood Way, Suite C,
Petaluma, CA 94549

Dr. Steven Colome—Integrated Environmental Services, 5319 University Drive, #430,
Irvine, CA 92612

Dr. Delbert Eatough—Brigham Young University, E 114 BNSN,
Department of Chemistry and Biochemistry, Provo, UT 84602

Mr. William Ewald—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

Dr. Sheldon Friedlander—-University of California at Los Angeles, Department of Chemical
Engineering, 5531 Boelter Hall, Los Angeles, CA  90095

Dr. Judith Graham—National Exposure Research Laboratory (MD-75),
U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

Dr. Lynn Hildemann—Civil and Environmental Engineering Department,
Stanford University, Stanford, CA 94305

Mr. Jim Homolya—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Rudolf Husar—CAPITA, Washington University, Campus Box 1124,
One Brookings Drive, St. Louis, MO 63130

March 2001                            I-xxviii      DRAFT-DO NOT QUOTE OR CITE

-------
                      Authors, Contributors, and Reviewers
                                        (cont'd)
 Contributors and Reviewers
 (cont'd)

 Dr. Charles W. Lewis—National Exposure Research Laboratory (MD-47),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Mr. Scott Mathias—Office of Air Quality Planning and Standards (MD-15)
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Mr. Tom McCurdy—National Exposure Research Laboratory (MD-56),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Mr. Frank McElroy—National Exposure Research Laboratory (MD-46),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Haluk Ozkaynak—National Exposure Research Laboratory (MD-56),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Mr. Tom Pace—Office of Air Quality Planning and Standards (MD-14),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Richard Poirot—VT Air Program, Building 3 South, 103 South Main Street
 Waterbury,VT 05671

 Dr. Linda Sheldon—National Exposure Research Laboratory (MD-77),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Helen Suh—Harvard School of Public Health, 665 Huntington Avenue,
 Boston, MA 02461

 Mr. Robert Wayland—Office of Air Quality Planning and Standards (MD-15),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Mr. Dane Westerdahl—California Air Resources Board, 2020 L Street, Sacramento, CA 95814

 Dr. William Wilson—National Center for Environmental Assessment (MD-52),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001
I-xxix
                                                   DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
       CHAPTER 4. ENVIRONMENTAL EFFECTS OF PARTICULATE MATTER
Principal Authors

Ms. Beverly Comfort—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Mr. William Ewald—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. J.H.B. Gamer—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Lester D. Grant—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. David A. Grantz—University of California/Riverside, Kearney Agricultural Center,
9240 South Riverbend Avenue, Parlier, CA 93648

Dr. Paul J. Hanson—Environmental Sciences Division, Oak Ridge National Laboratory,
P.O. Box 2008, Bethel Valley Road, Building 1059, Oak Ridge, TN 37831-6422

Dr. Dale W. Johnson—Environmental and Resource Science, 1000 Valley Road, University of
Nevada, Reno, NV 89512

Dr. Joseph P. Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. William H. Smith—Yale University School of Forestry and Environmental Studies,
370 Prospect Street, New Haven, CT  06511

Ms. Debra Meyer Wefering—Duckterather Weg 61, Bergisch Gladbach, Germany 54169
(formerly with National Exposure Research Laboratory [MD-56], U.S. Environmental Protection
Agency, Research Triangle Park, NC  27711)

Contributors and Reviewers

Dr. Larry T. Cupitt—National Exposure Research Laboratory (MD-75),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001
I-xxx
DRAFT-DO NOT QUOTE OR CITE

-------
                       Authors, Contributors, and Reviewers
                                       (cont'd)
 Contributors and Reviewers
 (cont'd)

 Dr. Russell R. Dickerson—University of Maryland, Department of Meteorology,
 Stadium Drive, College Park, MD 20742

 Dr. Anne Grambsch—National Center for Environmental Assessment (860ID),
 U. S. Environmental Protection Agency, Washington, DC 20036

 Dr. Sagar V. Krupa—University of Minnesota, Department of Plant Pathology,
 495 Borlaug Hall, 1991 Upper Buford Circle, St. Paul, MN 55108

 Dr. Alan J. Krupnick—Quality of the Environment Division, Resources for the Future
 1616 P Street, NW, Washington, DC 20036

 Mr. Paul T. Roberts—Sonoma Technology, Inc., 1360 Redwood Way - Suite C
 Petaluma, CA 94954

 Mr. John Spence—1206 Sturdivant Drive, Gary, NC 27511

 Dr. Richard Zepp—National Exposure Research Laboratory (IOD),
 U. S. Environmental Protection Agency, Athens, GA
           CHAPTER 5. HUMAN EXPOSURE TO PARTICULA TE MA TTER
                             AND ITS CONSTITUENTS
Principal Authors

Dr. David T. Mage—Institute for Survey Research, Temple University,
Philadelphia, PA 19122-6099 (formerly with the National Exposure Research
Laboratory (MD-56), U.S. Environmental Protection Agency, Research Triangle
Park,NC 27711)

Mr. Thomas McCurdy—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Linda S. Sheldon—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 277\ 1
March 2001
                                       I-xxxi
DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
Principal Authors
(cont'd)

Dr. Haluk Ozkaynak—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. William E. Wilson—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Contributing Authors

Dr. Janet Burke—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Roy Fortmann—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Gary Norris—National Exposure Research Laboratory (MD-47),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Anne Rea—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Alan Vette—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Contributors and Reviewers

Dr. Timothy Buckley—Johns Hopkins University, Department of Environmental Health
Sciences, 615 North Wolfe Street,, Baltimore, MD  21205

Ms. Lee Byrd—National Exposure Research Laboratory (MD-75),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Lyle Chinkin—Sonoma Technology, 1360 Redwood Way, Suite C, Petaluma, CA 94549

Dr. Steven Colome—Integrated Environmental Services, 5319 University Drive, #430,
Irvine, CA 92612

Dr. Delbert Eatough—Brigham Young University, E 114 BNSN,
Department of Chemistry and Biochemistry, Provo, UT 84602
 March 2001
I-xxxii
DRAFT-DO NOT QUOTE OR CITE

-------
                      Authors, Contributors, and Reviewers
                                       (cont'd)
 Contributors and Reviewers
 (cont'd)

 Dr. Sheldon Friedlander—University of California at Los Angeles, Department of Chemical
 Engineering, 5531 Boelter Hall, Los Angeles, CA 90095

 Dr. Judith Graham—National Exposure Research Laboratory (MD-75),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Lester D. Grant—National Center for Environmental Assessment (MD-52),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Vic Hasselblad—29 Autumn Woods Drive, Durham, NC 27713

 Dr. Lynn Hildemann—Civil and Environmental Engineering Department
 Stanford University,  Stanford, CA 94305

 Mr. Jim Homolya—Office of Air Quality Planning and Standards (MD-14),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Rudolf Husar—CAPITA, Washington University, Campus Box 1124,
 One Brookings Drive, St. Louis, MO 63130

 Dr. Dennis J. Kotchmar—National Center for Environmental Assessment (MD-52),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Charles W. Lewis—National Exposure Research Laboratory (MD-47),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Allan Marcus—National Center for Environmental Assessment (MD-52),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Mr. Frank McElroy—National Exposure Research Laboratory (MD-46),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Mr. Tom Pace—Office of Air Quality Planning and Standards (MD-14),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Joseph Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
March 2001
I-xxxiii
DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
Contributors and Reviewers
(cont'd)

Dr. Richard Poirot—VT Air Program, Building 3 South, 103 South Main Street,
Waterbury.VT 05671

Mr. Harvey Richmond—Office of Air Quality Planning and Standards (MD-15),
U. S. Environmental Protection Agency, Research Triangle Park, NC  27711

Dr. Linda Sheldon—National Exposure Research Laboratory (MD-77),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Helen Suh—Harvard School of Public Health, 665 Huntington Avenue,
Boston, MA 02461

Mr. Robert Wayland—Office of Air Quality Planning and Standards (MD-15),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Mr. Dane Westerdahl—California Air Resources Board, 2020 L Street, Sacramento, CA 95814

Dr. Jim Xue—Harvard School of Public Health, 665 Huntington Avenue, Boston, MA  02115
March 2001
I-xxxiv
DRAFT-DO NOT QUOTE OR CITE

-------
               U.S. ENVIRONMENTAL PROTECTION AGENCY
   PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
                         FOR PARTICULATE MATTER


 Scientific Staff

 Dr. Lester D. Grant—Director, National Center for Environmental Assessment (MD-52),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. William E. Wilson—Air Quality Coordinator, Physical Scientist, National Center for
 Environmental Assessment (MD-52), U.S. Environmental Protection Agency, Research Trianele
 Park,NC 27711

 Dr. Lawrence J. Folinsbee—Health Coordinator, Chief, Environmental Media Assessment
 Group, National Center for Environmental Assessment (MD-52), U.S. Environmental Protection
 Agency, Research Triangle Park, NC  27711

 Dr. J.H.B. Garner—Ecological Scientist, National Center for Environmental Assessment
 (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Dennis J. Kotchmar—Project Manager, Medical Officer, National Center for Environmental
 Assessment (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC


 Dr. Robert Chapman—Medical Officer, National Center for Environmental Assessment
 (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Ms. Beverly Comfort—Health Scientist, National Center for Environmental Assessment
 (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Mr. William Ewald—Health Scientist, National Center for Environmental Assessment (MD-52),
 U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

 Dr. David Mage—Physical Scientist, Institute for Survey Research, Temple University
 Philadelphia, PA 19122-6099

 Dr. Allan Marcus—Statistician, National Center for Environmental Assessment (MD-52),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. James McGrath—Visiting Senior Health Scientist, National Center for Environmental
 Assessment (MD-52), U.S. Environmental Protection Agency, Research Trianele Park NC
 27711
March 2001
                                      I-xxxv
DRAFT-DO NOT QUOTE OR CITE

-------
             U.S. ENVIRONMENTAL PROTECTION AGENCY
 PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
                       FOR PARTICULATE MATTER
                                      (cont'd)
Scientific Staff
(cont'd)

Dr. Joseph P. Pinto—Physical Scientist, National Center for Environmental Assessment,
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Mr. James A. Raub—Health Scientist, National Center for Environmental Assessment (MD-52),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Technical Support Staff

Mr. Randy Brady—Deputy Directory, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Mr. Douglas B. Fennell—Technical Information Specialist, National Center for Environmental
Assessment (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711

Ms. Emily R. Lee—Management Analyst, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

Ms. Diane H. Ray—Program Specialist, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Ms. Eleanor Speh—Office Manager, Environmental Media Assessment Branch, National Center
for Environmental Assessment (MD-52), U.S. Environmental Protection Agency, Research
Triangle Park, NC 27711

Ms. Donna Wicker—Administrative Officer, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Mr. Richard Wilson—Clerk, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
 March 2001
I-xxxvi
DRAFT-DO NOT QUOTE OR CITE

-------
              U.S. ENVIRONMENTAL PROTECTION AGENCY
  PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
                        FOR PARTICULATE MATTER
                                      (cont'd)


Document Production Staff

Mr. John R. Barton—Document Processing Coordinator, OAO Corporation,
2222 Chapel Hill-Nelson Highway, Beta Building, Suite 100, Durham, NC 27713

Ms. Diane G. Caudill—Graphic Artist, OAO Corporation, 2222 Chapel Hill-Nelson Highway,
Beta Building, Suite 100, Durham, NC 27713

Ms. Yvonne A. Harrison—Word Processor, OAO Corporation, 2222 Chapel Hill-Nelson
Highway, Beta Building, Suite 100, Durham, NC  27713

Ms. Bettye B. Kirkland—Word Processor, OAO Corporation, 2222 Chapel Hill-Nelson
Highway, Beta Building, Suite 100, Durham, NC  27713

Mr. David E. Leonhard—Graphic Artist, OAO Corporation, 2222 Chapel Hill-Nelson Highway,
Beta Building, Suite 100, Durham, NC 27713

Ms. Phyllis H. Noell—Technical Editor, OAO Corporation, 2222 Chapel Hill-Nelson Highway,
Beta Building, Suite 100, Durham, NC 27713

Ms. Carolyn T. Perry—Word Processor, OAO Corporation, 2222 Chapel Hill-Nelson Highway,
Beta Building, Suite 100, Durham, NC 27713

Technical Reference Staff

Mr. R. David Belton—Reference Specialist, OAO Corporation, 2222 Chapel Hill-Nelson
Highway, Beta Building, Suite 100, Durham, NC  27713

Mr. John A. Bennett—Technical Information Specialist, OAO Corporation, 2222 Chapel
Hill-Nelson Highway, Beta Building, Suite 100, Durham, NC 27713

Ms. Sandra L. Hughey—Technical Information Specialist, OAO Corporation, 2222 Chapel
Hill-Nelson Highway, Beta Building, Suite 100, Durham, NC 27713

Mr. Jian Ping Yu—Reference Retrieval and Database Entry Clerk, OAO Corporation,
2222 Chapel Hill-Nelson Highway, Beta Building, Suite 100, Durham, NC  27713

Ms. Kun Zhang—Records Management Technician, OAO Corporation, 2222 Chapel Hill-Nelson
Highway, Beta Building, Suite 100, Durham, NC 27713
March 2001
I-xxxvii
DRAFT-DO NOT QUOTE OR CITE

-------

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
                             1.  INTRODUCTION
      This document is an update of "Air Quality Criteria for Particulate Matter" published by the
 U.S. Environmental Protection Agency (EPA) in 1996, and it will serve as the basis for
 Congressionally-mandated periodic review of the National Ambient Air Quality Standards
 (NAAQS) for particulate matter (PM). The present document critically assesses the latest
 scientific information relative to determining the health and welfare effects associated with
 exposure to various concentrations of PM in ambient air. The document is not intended as a
 complete and detailed literature review, but rather focuses on assessment and integration of
 information most relevant to PM NAAQS criteria development, based on pertinent literature
 mainly available through December 2000. This introductory chapter presents a brief summary of
 the history of the PM NAAQS, provides an overview of issues addressed and procedures utilized
 in the preparation of the present document, and provides orientation to the general organizational
 structure of this document.
 1.1  LEGISLATIVE REQUIREMENTS
      Sections 108 and 109 of the U.S. Clean Air Act (CAA) (U.S. Code, 1991) govern the
 establishment, review, and revision of National Ambient Air Quality Standards (NAAQS).
 Section 108 directs the EPA Administrator to list pollutants that may reasonably be anticipated to
 endanger public health or welfare and to issue air quality criteria for them. The air quality
 criteria are to reflect the latest scientific information useful in indicating the kind and extent of all
 exposure-related effects on public health and welfare that may be expected from the presence of
 the pollutant in ambient air.
      Section 109(a,b) directs the Administrator of EPA to propose and promulgate "primary"
 and "secondary" NAAQS for pollutants identified under Section 108.  Section 109(b)(l) defines
 a primary standard as a level of air quality, the attainment and maintenance of which, in the
judgment of the Administrator, based on the criteria and allowing for an adequate margin of
 safety, is requisite to protect the public health. Under Section 109(b) of the CAA, the EPA
       March 2001
                                          1-1
DRAFT-DO NOT QUOTE OR CITE

-------
 1      Administrator must consider available information to set secondary NAAQS that are based on
 2      the criteria and are requisite to protect public welfare from any known or anticipated adverse
 3      effects associated with the presence of such pollutants. Welfare effects include effects on
 4      vegetation, crops, soils, water, animals, manufactured materials, visibility, weather, and climate,
 5      as well as damage to and deterioration of property, hazards to transportation, and effects on
 6      economic value and personal comfort and well-being.  Section 109(d) also requires periodic
 7      review and, if appropriate, revision of existing criteria and standards, and it requires an
 8      independent committee of non-EPA experts, the Clean Air Scientific Advisory Committee
 9      (CASAC), to provide the EPA Administrator advice and recommendations regarding the
10      scientific soundness and appropriateness of criteria and NAAQS for PM and other "criteria air
11      pollutants" (e.g., ozone, nitrogen oxides, sulfate dioxide, carbon monoxide, lead) regulated under
12     CAA Sections 108-109.
13
14
15     1.2  HISTORY OF PREVIOUS PARTICIPATE MATTER CRITERIA AND
16          NATIONAL AMBIENT AIR QUALITY STANDARDS REVIEWS
17           On April 30,1971 (Federal Register, 1971), EPA promulgated the original primary and
18     secondary NAAQS for PM under Section 109 of the CAA. The reference method for measuring
19     attainment of these standards was the "high-volume" sampler (Code of Federal Regulations,
20     1977), which collects ambient PM up to a nominal size of 25 to 45 micrometers (jj.m) (i.e.,
21     so-called "total suspended particulate" or "TSP"). Thus, TSP was the original indicator for the
22     PM NAAQS. The primary standards for PM (measured as TSP) were 260 //g/m3 (24-h average),
23     not to be exceeded more than once per year, and 75 /zg/m3 (annual geometric mean). The
24     secondary standard (measured as TSP) was 150 Mg/m3 (24-h average), not to be exceeded more
25     than once per year. The next review of PM air quality criteria and standards was completed in
26     July 1987, when the original TSP NAAQS set in 1971 were revised to protect against adverse
27     health effects of inhalable airborne particles with an upper 50% cut-point of 10-yUm aerodynamic
28     diameter (PM10), which can be deposited in the lower (thoracic) regions of the human respiratory
29     tract (Federal Register, 1987). Identical primary and secondary PM10 standards were set for two
30     averaging times:  150 ^g/m3 (24-h average), with no more than one expected exceedance per
31     year; and 50 /^g/m3 (expected annual arithmetic mean), averaged over 3 years.
       March 2001
1-2
DRAFT-DO NOT QUOTE OR CITE

-------
 1      1.2.1  The 1997 Particulate Matter National Ambient Air Quality Standards
 2            Revision
 3          The EPA initiated the last previous review of the air quality criteria and standards for PM
 4      in April 1994 by announcing its intention to develop revised Air Quality Criteria for Particulate
 5      Matter (henceforth, the "PM Air Quality Criteria Document" or PM AQCD). Thereafter, the
 6      EPA presented its plans for review of the criteria and standards for PM under a highly
 7      accelerated, court-ordered schedule at a public meeting of the CAS AC in December 1994.
 8      A court order entered in American Lung Association v. Browner, CIV-93-643-TUC-ACM (U.S.
 9      District Court of Arizona, 1995), as subsequently modified, required publication of EPA's final
10      decision on the review of the PM NAAQS by July 19,1997.
11          Several workshops were held by EPA's National Center for Environmental Assessment
12      RTP Division (NCEA-RTP) in November 1994 and January 1995 to discuss important new
13      health effects information useful in preparing initial PM AQCD draft materials.  External review
14      drafts of the PM AQCD then were made available for public comment and were reviewed by
15      CASAC at public meetings held in August 1995, December 1995, and February 1996.  The
16      CASAC came to closure in its review of the PM AQCD, advising the EPA Administrator in a
17      March 15,1996, closure letter (Wolff, 1996) that "although our understanding of the health
18      effects of PM is far from complete, a revised Criteria Document which incorporates the Panel's
19      latest comments will provide an adequate review of the available scientific data and relevant
20      studies of PM." Revisions in response to public  and CASAC comments were incorporated as
21      appropriate in the final 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a).  A PM
22      Staff Paper (SP), prepared by EPA's Office of Air Quality Planning and Standards (OAQPS) and
23      drawing on the 1996 PM AQCD and other exposure and risk assessments to pose options for PM
24      NAAQS decisions, also underwent similar CASAC review and public comment, with consequent
25      revision to its July 1996 final form (U.S. Environmental Protection Agency, 1996b).
26          The SP analyses served as key inputs to subsequently published proposals for revision of
27      the primary PM NAAQS. Taking into account information and assessments presented in the PM
28      AQCD and the SP, advice and recommendations of CASAC, and public comments received on
29      the proposal, the EPA Administrator revised the PM NAAQS by adding new PM2 5 standards and
30      by revising the form of the 24-h PM10 standard. Specifically, in July 1997, the Administrator
31      made the following revisions to the PM NAAQS:
       March 2001
1-3
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
(1) The suite of PM standards was revised to include an annual primary PM2 5 standard and a
    24-h PM2 5 standard.
(2) The 24-h PM2S standard is met when the 3-year average of the 98th percentile of 24-h PM25
    concentrations at each population-oriented monitor within an area is less than or equal to
    65 //g/m3, with fractional parts of 0.5 or greater rounding up.
(3) The annual PM2 5 standard is met when the 3-year average of the annual arithmetic mean
    PM2 j concentrations, from single or multiple community-oriented monitors is less than or
    equal to 15 /u.g/m3, with fractional parts of 0.05 or greater rounding up.
(4) The form of the 24-h PMIO (150 /^g/m3) standard was revised to be based on the 3-year
    average of the 99* percentile of 24-h PM10 concentrations at each monitor within an area.
(5) In addition, the Administrator retained the annual PM10 standard at the level of 50 /ug/m3,
    which is met when the 3-year average of the annual arithmetic mean PM10 concentrations at
    each monitor within an area is less than or equal to 50 //g/m3, with fractional parts of 0.5 or
    greater rounding up.
     The principal focus of the last review of the air quality criteria and standards for PM was on
recent epidemiological evidence  reporting associations between ambient concentrations of PM
and a range of serious health effects.  Particular attention was given to several size-specific
classes of particles, including PM10 and the principal fractions of PM10, referred to as the fine
(PM2 5) and coarse (PM10.2.5) fractions. PM2 5 refers to particles with an upper 50% cutpoint of
2.5-jum aerodynamic diameter. PM10_2 5 refers to those particles with an upper 50% cutpoint of
10 fj.ro. and a lower 50% cut point of 2.5-/^m aerodynamic diameter. In other words, the coarse
fraction (PM10.25) refers to the inhalable particles that remain if fine (PM25) particles are removed
from a sample of PM10 particles.  As discussed in the 1996 PM AQCD, fine and coarse fraction
particles can be differentiated by their sources and formation processes and by their chemical and
physical properties, including behavior in the atmosphere.  Detailed discussions of atmospheric
formation, ambient concentrations, and health effects of ambient air PM, as well as quantitative
estimates of human health risks associated with exposure to ambient PM, can be found in the
1996 PM AQCD and in the 1996 OAQPS SP (U.S. Environmental Protection Agency, 1996b).
       March 2001
                                          1-4
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
' 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
31
 1.2.2  Presidential Memorandum: Next Particulate Matter Review
        and Research
      On July 18, 1997, the EPA published a final rule revising the NAAQS for PM (Federal
 Register, 1997a) and, on the same day, a final rule revising the Ozone NAAQS (Federal Register,
 1997b). A Presidential Memorandum (Federal Register, 1997c) also was published outlining the
 Administration's goals for implementing the revised PM and Ozone NAAQS. The
 Memorandum directed EPA to provide to CASAC within 90 days and to publish a notice
 outlining its schedule for the next periodic review of PM and to complete the next review,
 including review by CASAC, within 5 years after issuance of the revised standards (i.e., by July
 2002).  Such a schedule would ensure that EPA's review of newly emerging scientific
 information, which forms the criteria on which the standards are based, and of the standards
 themselves will have been completed prior to any areas being designated as "nonattainment"
 under the newly established standards for fine particles (i.e., PM2 5 standards) and prior to the
 imposition of any new controls related to the revised standards. The Presidential Memorandum
 also directed EPA and other relevant Federal agencies to develop and implement a greatly
 expanded, coordinated research program. To facilitate timely scientific research within this
 review period, EPA initiated certain activities immediately, as noted below in the discussion of
 the PM Research Program.

 Particulate Matter Research Program
     The EPA broadened its ongoing PM research activities by developing, in partnership with
 other Federal agencies, a coordinated interagency PM research program. This interagency
 program has and continues to focus mainly on expanding scientific knowledge of ambient PM
 exposure and health effects, as well as including development of improved monitoring methods
 and cost-effective mitigation strategies.  The interagency effort also promotes further
 coordination with other research organizations, including the Health Effects Institute and other
 state-, university-, and industry-sponsored research groups. Beginning in the fall of 1997, public
participation has been and continues to be encouraged through workshops and review of program
documentation.
     To aid identification of needed research efforts, EPA published a particulate matter health
risk research needs document (U.S. Environmental Protection Agency, 1998a). That document
        March 2001
                                                 1-5
                                                    DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
identifies research needed to improve scientific information supporting future health risk
assessment and review of the PM NAAQS. The document aimed to provide a foundation for PM
research coordination among Federal agencies and other research organizations and served as one
useful input to National Research Council PM research deliberations.  In January 1998, the
National Research Council (NRC) established its Committee on Research Priorities for Airborne
Particulate Matter in response to a request from Congress in the Fiscal 1998 appropriation to
EPA. This Committee is charged to identify the most important research priorities relevant to
setting particulate matter standards, to develop a conceptual plan for particulate matter research,
and to monitor research progress toward improved understanding of the relationship between
particulate matter and public health. The Committee issued its first report in early 1998
(National Research Council, 1998) and a second one in 1999 (National Research Council, 1999).
     The EPA's PM Research Program includes studies to improve understanding of the
formation and composition of fine PM, the characteristics or components of PM that are
responsible for its health effects, the mechanisms by which these effects are produced, and
improved measurements and estimation of population exposures to PM. Specific EPA research
efforts include controlled human exposure studies, in vivo and in vitro toxicology, epidemiology,
atmospheric sciences including monitoring and modeling studies, development of data on
emissions of fine particles from stationary and mobile sources, and identification and evaluation
of risk management options.  The results from these efforts, as well as related efforts by other
Federal agencies and the general scientific community, are expected to enhance substantially the
scientific and technical bases for future decisions on the PM NAAQS and for the implementation
of PM monitoring and control efforts.
 1.3 CURRENT PARTICULATE MATTER CRITERIA AND NATIONAL
     AMBIENT AIR QUALITY STANDARDS REVIEW
 1.3.1 Criteria Review
      As with all NAAQS reviews, the purpose is to update the criteria and to determine whether
 it is appropriate to revise existing standards in light of new scientific and technical information.
        March 2001
                                          1-6
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
27
28
29
30
 Although the EPA concluded in its most recent final rule on the PM standards (Federal Register,
 1997a) that the current scientific knowledge provides a strong basis for the revised PM standards,
 including the establishment of PM25 standards, there remain scientific uncertainties associated
 with the health effects of PM and with the means of reducing such effects. Recognizing the
 importance of developing a better understanding of the effects of fine particles on human health,
 including their causes and mechanisms, as well as the species and sources of PM2 5, EPA has and
 will continue to sponsor research to address these uncertainties even as this criteria review
 progresses.
     As with other NAAQS reviews, a rigorous assessment of relevant scientific information is
 to be presented in this updated, revised PM AQCD being prepared by EPA's NCEA-RTP.
 Development of the document has and will continue to involve substantial external peer review
 through (a) public workshops involving the general aerosol scientific community, (b) iterative
 reviews of successive drafts by CASAC, and (c) comments from the public. The final document
 will reflect input received through these reviews and will serve to evaluate and integrate the latest
 available scientific information to ensure that the review of the PM standards is based on sound
 science. An earlier (October 1999) First External Review Draft of this updated document was
 released in the fall of 1999 for public comment and CASAC review.  This Second External
 Review Draft takes into account the earlier public comments and CASAC review
 recommendations  and includes consideration of relevant new peer-reviewed scientific studies
 published or accepted for publication from January 1996 through December 2000.
     Following CASAC review of the First External Review Draft of this revised PM AQCD in
 December 1999, EPA's OAQPS started to prepare an SP for the EPA Administrator. Drawing
 on information in this newly revised PM AQCD, the SP will evaluate policy implications of the
 key studies and scientific information contained in the AQCD and identify critical elements that
 EPA staff believes should be considered in reviewing the PM standards.  The SP is intended to
bridge the gap between the scientific review in the AQCD and the public health and welfare
policy judgments required of the Administrator in reviewing the PM NAAQS. For that purpose,
the SP will present technical analyses, including air quality analyses and a quantitative health risk
assessment, and other factors relevant to the evaluation of the PM NAAQS, as well as staff
conclusions and recommendations of options for the EPA Administrator's consideration. The SP
       March 2001
                                         1-7
DRAFT-DO NOT QUOTE OR CITE

-------
 1     also will be reviewed by CASAC and the public, and the final SP will reflect the input received
 2     through these reviews.
 3          Following completion of the final SP, the Administrator will then announce in the Federal
 4     Register proposals for retaining or revising the current PM NAAQS, and opportunities will be
 5     provided for public comment and CASAC review of those proposals.  Taking into account public
 6     comments and CASAC recommendations, final decisions regarding the current PM NAAQS
 7     review are scheduled to be promulgated by July 2002.
 8
 9     1.3.2  Methods and Procedures for Document Preparation
10          The procedures followed for developing this revised PM AQCD build on the knowledge
11     and methods derived from the most recent previous PM, Ozone, arid CO AQCD preparation
12     efforts.  Briefly, the respective responsibilities for production of the present PM AQCD are as
13     follows. An NCEA-RTP PM team was formed to be responsible for developing and
14     implementing the project plan for preparation of the PM AQCD, taking into account inputs from
15     individuals in other EPA program and policy offices identified as part of the EPA PM Work
16     Group.  The resulting project plan (i.e., the PM Document Development Plan) was then
17     discussed with CASAC (May 1998) and appropriately revised. An ongoing literature search has
18     continued to be conducted to identify, to the extent possible, all PM literature published since
19     early 1996. Additionally,  EPA published (1) a request for information in the Federal Register
20     asking for recently available research information on PM that may not yet be published and
21     (2) a request for individuals with the appropriate type and level of expertise to contribute to the
22     writing  of PM AQCD materials to identify themselves (U.S. Environmental Protection Agency,
23     1998b). Specific authors of chapters or sections of the proposed document were selected on the
24     basis of their expertise on the subject areas and their familiarity with the relevant literature; these
25     include both EPA and non-EPA scientific experts. The project team defined critical issues and
26     topics to be addressed by the authors and provided direction in order to emphasize evaluation of
27     those studies most clearly identified as important for standard setting.
28          The main focus of this revised criteria document is the evaluation and interpretation of air
29     quality data, human exposure information, and health and welfare effects information newly
30     published since that assessed in the 1996 PM AQCD and likely to be useful in deriving criteria
31     for PM NAAQS. Initial draft versions of AQCD chapters were evaluated via expert workshops
       March 2001                              1-8         DRAFT-DO NOT QUOTE OR CITE

-------
   1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
27
28
29
30
 and/or expert written peer reviews, which focused on the selection of pertinent studies included
 in the chapters, the potential need for additional information to be added to the chapters, and the
 quality of the summarization and interpretation of the literature.  The authors of the draft chapters
 then revised them on the basis of the workshop and/or written expert review recommendations.
 These and other integrative summary materials were incorporated into the First External Review
 Draft of the PM AQCD (October, 1999), which was released for public comment and reviewed at
 a December 1999 CAS AC public meeting. Necessary revisions, based on public comments and
 the recommendations derived from the December 1999 CAS AC review, as well as evaluation of
 newly emerging research results, have been incorporated into this Second External Review Draft.
 The final version of the newly revised PM AQCD will incorporate changes made in response to
 public comments and CAS AC review of this Second External Review Draft.
     New research results are being incorporated into this document as they become available.
 In order to foster timely presentation and publication of newly emerging PM research findings,
 EPA co-sponsored an Air and Waste Management Association International Speciality
 Conference, entitled "PM 2000:  Particulate Matter and Health", which was held in Charleston,
 SC, in January 2000. The conference  was co-sponsored in cooperation with several other
 government agencies and/or private organizations that also fund PM research. Topics covered
 included new research results concerning the latest advances in PM atmospheric sciences (e.g.,
 PM formation, transport, transformation), PM exposure, PM dosimetry and extrapolation
 modeling, PM toxicology (e.g., mechanisms, laboratory animal models, human clinical
 responses), and PM epidemiology. The main purpose of the conference was to facilitate having
 the latest scientific information available in time for incorporation into this revised draft EPA
 PM AQCD so as to allow for its release for public comment and CAS AC review by December
 2000. Arrangements were made for scientists to submit written manuscripts on papers or posters
presented at the PM 2000 Conference for expedited peer-review by several major journals, so
that decisions on acceptance for publication could be made by mid-2000.  The evaluations and
findings set forth in this Second External Review Draft of the revised PM AQCD include
consideration of such PM 2000 papers and extensive additional information published elsewhere
since completion of the previous First External Review Draft.
       March 2001
                                                 1-9
                                                     DRAFT-DO NOT QUOTE OR CITE

-------
 1      1.3.3 Approach
 2           The approach to organization and content of this revised PM AQCD is somewhat different
 3      from those used for previous criteria documents.  Because the most recent prior document (U.S.
 4      Environmental Protection Agency, 1996a) provides an extensive discussion of most topic areas,
 5      this new document focuses more specifically on critical issues that have been identified as areas
 6      needed to improve the scientific basis (criteria) for PM NAAQS, particularly for those areas in
 7      which the information database has continued to evolve rapidly.
 8           An initial step was to review the available scientific literature and to focus on the selection
 9      of pertinent issues to include in the document as the basis for the development of PM NAAQS
10      criteria. Preliminary issues were identified by the NCEA PM Team and through input from other
11      EPA program and policy offices. Identification of issue topics was derived from the 1996 PM
12     AQCD and SP, their CASAC and public reviews, from the standard promulgation process, and
13      from EPA's PM Research Needs Document. Further identification and clarification of issues
14     resulted from the NRC review and reports on PM research priorities. The CASAC review of the
15     PM AQCD Development Plan and public comments on draft AQCD materials at various stages
16     of their development also has played an important role in issue identification.
17          In developing draft materials for inclusion in the revised PM AQCD, detailed review of key
18     new research was undertaken as a first step.  However, instead of presenting a comprehensive
19     review of all the literature, emphasis in this revised AQCD is placed on (1) the concise summary
20     of key findings derived from previous PM criteria reviews and (2) evaluation  of the most
21     pertinent new key information, with greater emphasis on more interpretive assessment. This
22     approach reflects recommendations made by CASAC.
23           Building on the previous PM AQCD, most of the scientific information  selected for review
24     and discussion hi the text is from literature published since completion of the  previous
25     PM AQCD (U.S. Environmental Protection Agency, 1996a). To aid in development of a concise
26     document, compilation of summary tables of the relevant published literature  and selective
27     discussion of the literature has been undertaken,  and increased emphasis has been placed hi text
28     discussions on interpretive evaluation and integration of key points derived from the newly
29     summarized research results.
30
        March 2001
1-10
DRAFT-DQ NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
 1.3.4 Key Issues of Concern
      Several broad topics related to the main issues of concern addressed by this revised
 PM AQCD are summarized below.  The document reviews and assesses available data bearing
 on each of the issues identified below.
 1. Causality. Evaluation of the evidence for or against a causal relationship between health
   outcomes and ambient PM and/or specific physical-chemical components.
   • Specific components of interest include size classes such as PM10, PM10.25, PM25, and
     ultrafine particles.  Chemical components include transition metals, acidity, sulfates,
     nitrates, and organics.
   • Expand review of foundations of causal inference for associated PM air pollution health
     effects.
   • Access new long-term PM exposure and health data to broaden interpretation of long-term
     exposure findings.
   • Review data exploring potential mechanisms of response to PM physical-chemical
     characteristics, response pathway, and exposure-dose-response relationships (laboratory and
     clinical research).
2. Uncertainties. In carrying out overall assessment, address the following types of uncertainty.
   •  Uncertainties between stationary PM monitoring instruments and personal exposure to PM
     of ambient origin, especially for susceptible groups and their related activity patterns.
     Specific topics include measurement error in outdoor monitors themselves, use of central
     monitors for estimates of community concentrations, and the use of community
     concentrations as a surrogate for personal exposure to particles of ambient origin.
   •  Uncertainties related to particulate matter size fraction, particle number, surface area, and
     content of semi-volatile components.
   • Uncertainties about the effects of long-term PM exposure, such as life shortening, and
    development and progression of disease.
   • Uncertainties because of coexposure to other pollutants such as O3, SO2, CO, andNO2, and
    because of meterological factors.
   • Uncertainties because of potential confounding in epidemiologic studies (e.g., economic
    factors, demographic and lifestyle attributes, genetic susceptibility factors, occupational
    exposure, medical care).
       March 2001
                                         1-11
                                                             DRAFT-DO NOT QUOTE OR CITE

-------
 1        • Uncertainty about shape of concentration-response (CR) relationships and associated
 2          community risks (linear and threshold models for CR).
 3        • Uncertainty about methods for synthesis of health outcome studies and evaluation of
 4          sensitivity and confounding aspects, including but not limited to meta-analyses.
 5     3. Biological Mechanisms of Action. Evaluate data examining mechanisms underlying health
 6        outcomes of PM. Mechanistic information aids judgment about causality.
 7        • New studies have examined mechanisms of action of PM constituents, including transition
 8          metals, airborne allergens, and the generation of reactive oxygen species.  Different cell
 9          types have differing responses to PM components.
10        • Newly published studies also have identified potential mechanisms for the production of
11          cardiac arrhythmias by PM constituents, especially in animal models of disease and suggest
12          that particular attention should be accorded to PM metal constituents.
13        • Although many new animal toxicology studies involve instillation of previously collected
14          particles and this technique is appropriate to study mechanisms of action, extrapolation to
15          human equivalent exposure/doses is uncertain.
16        • Ongoing work on the effects of lung inflammation and PM phagocytosis on subsequent
17          systemic effects, especially cardiac or vascular effects, is needed to provide further
1 g          information on the relationship between inhaled pollutants and cardiac events.
19        • Interpretation of concentrated ambient particles studies. Newly available information is
20          examined from toxicology studies using devices that concentrate (to variable extents)
21          ambient PM to determine PM concentration-response relationships. Again, difficulties are
22          encountered regarding extrapolation to comparable human exposures to ambient PM levels.
23      4. Susceptible Populations. Examine health outcome data to determine specific risk groups that
24        are more susceptible than normal healthy adults to adverse effects from PM exposure.
25        • Preexisting respiratory or cardiovascular disease in conjunction with advanced age appear to
26          be important factors in PM mortality susceptibility.
27        • For morbidity health endpoints, children and asthmatics potentially may display increased
28          sensitivity to PM exposure.  Data will be examined for coherence.
29        • Patterns of respiratory tract deposition, clearance, and retention in susceptible populations
30          have been studied recently and provide evidence of differences in respiratory tract PM
31          deposition for children and small-sized adults and for those with lung diseases.
        March 2001
1-12
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
30
31
     • Animal models of lung disease exposed to PM constituents suggest a role for PM in cardiac
      death.
  5.  Environmental Effects. Evaluate several types of PM welfare effects.
     • Vegetation and ecosystem effects.
     • Visibility effects.
     • Materials damage.
     • Role of PM in atmospheric radiative transfer and potential consequences for penetration of
      biologically harmful UVB to the earth's surface and for climate change.
 6. Background Information Topics Useful in Evaluating Health Risks. Topics include the
    following.
    • New monitoring methods, especially methods used in epidemiology studies.
    • Indicator topics such as PM2.5 versus PM, 0; ultrafme; and PM25 versus PM10.2S.
    • New data patterns of daily and annual concentrations for PM2 5, PM10.2 5, and PMI0.
 1.4 DOCUMENT CONTENT AND ORGANIZATION
      The present draft document attempts to critically review and assess relevant scientific
 literature on PM through December 2000. The material selected for review and comment in the
 text generally comes from the more recent literature published since early 1996, with emphasis
 on studies conducted at or near PM pollutant concentrations found in ambient air. Literature
 discussed in detail in the previous 1996 EPA PM AQCD (U.S. Environmental Protection
 Agency, 1996a) generally is not discussed in depth in this document.  However, some limited
 treatment is included of the earlier studies judged to be potentially useful in deriving PM
 NAAQS. Key literature is presented mainly in tables and overall interpretive points are
 discussed mainly in the text.
     The primary emphasis is on consideration of published material that has undergone
 scientific peer review. However, in the interest of admitting new and important information
 expected to become available shortly, some material not yet fully published in the open literature
but meeting other standards of scientific reporting (i.e., peer review, quality assurance) are now
provisionally included. As noted earlier, emphasis has been placed on studies in the range of
current ambient levels. However, studies examining effects of higher concentrations have been
       March 2001
                                                 1-13
                                                     DRAFT-DO NOT QUOTE OR CITE

-------
 1      included if they contain unique data or documentation of a previously unreported effect or
 2      mechanism. In reviewing and summarizing the literature, an attempt has been made to present
 3      alternative points of view where scientific controversy exists.
 4           The present document includes nine chapters presented in two volumes. Volume 1
 5      contains this general introduction (Chapter 1).  It also includes Chapters 2 and 3, which provide
 6      background information on physical and chemical properties of PM and related compounds;
 7      sources and emissions; atmospheric transport, transformation, and fate of PM; methods for the
 8      collection and measurement of PM; and ambient air concentrations. Next, Chapter 4 describes
 9      PM environmental effects on vegetation and ecosystems, visibility, man-made materials, and
10     climate, as well as economic impacts of such welfare effects. Chapter 5, which discusses factors
11      affecting exposure of the general population to ambient PM, is also included in Volume 1.  The
12     second volume contains Chapters 6 through 9 and the Executive Summary for the entire
13     document. Chapters 6 through 8 evaluate information concerning the health effects of PM
14     (Chapter 6 discusses epidemiological studies; Chapter 7, dosimetry of inhaled particles in the
15     respiratory tract, and Chapter 8, the toxicology of specific types of PM constituents, including
16     laboratory animal studies and controlled human exposure studies).  Chapter 9 integrates key
17     information on exposure, dosimetry, and critical health risk issues derived from studies reviewed
18     in the prior chapters.
19           Neither control techniques nor control strategies for abatement of PM are discussed in this
20     document, although some topics covered may be incidentally relevant to control strategies.
21     Issues germane to the scientific basis for control strategies, but not pertinent to the development
22     of NAAQS criteria, are addressed in numerous other documents issued by EPA's OAQPS.
23     Technologies for controlling PM emissions also are discussed in other documents issued by
24     OAQPS. Also, certain issues of direct relevance to standard setting are not addressed explicitly
25      in this document, but instead are analyzed in documentation prepared by OAQPS as part of its
26      regulatory analyses materials. Such analyses include (1) delineation of particular adverse effects
27      that the primary and secondary NAAQS are intended to protect against, (2) exposure analyses
28      and assessment of consequent risk, and (3) discussion of factors to be considered in determining
29      an adequate margin of safely. Key points and conclusions from such analyses will be presented
 30      in the PM SP prepared by OAQPS for review by CASAC.  Although scientific data contribute
 31      significantly to decisions regarding the above issues, their resolution cannot be achieved solely
        March 2001
1-14
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
on the basis of experimentally acquired information. Final decisions on items (1) and (3) are
made by the EPA Administrator, as mandated by the CAA.
     A fourth issue directly pertinent to standard setting is identification of populations at risk,
which is basically a selection by EPA of the subpopulation(s) to be protected by the promulgation
of a given standard. This issue is addressed only partially in this document. For example,
information is presented on factors, such as preexisting disease, that may biologically predispose
individuals and subpopulations to adverse effects from exposures to PM.  The characterization of
population risk, however, requires information above and beyond data on biological
predisposition (e.g., information on estimated exposure, activity patterns, and personal habits).
Such information is typically addressed in the SP developed by OAQPS.
       March 2001
                                         1-15
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1
REFERENCES
 2       Code of Federal Regulations. (1977) Appendix B—Reference method for the determination of suspended
 3             particulates in the atmosphere (high-volume method). C. F. R. 40: §50.
 4       Federal Register. (1971) National primary and secondary ambient air quality standards. F. R. (April 30)
 5             36:8186-8201.
 6       Federal Register. (1987) Revisions to the national ambient air quality standards for particulate matter. F. R. (July 1)
 7             52:24,634-24,669.
 8       Federal Register. (1997a) National ambient air quality standards for particulate matter; final rule. F. R. (July 18)
 9             62:38,652-38,752.
10       Federal Register. (1997b) National ambient air quality standards for ozone; final rule. F. R. (July 18)
11             62:38,856-38,896.
12       Federal Register. (1997c) Implementation of revised air quality standards for ozone and particulate matter. F. R.
13             (July 18) 62: 38,421-38,422.
14       National Research Council. (1998) Research priorities for airborne  particulate matter. I. Immediate priorities and a
15             long-range research portfolio. Washington, DC: National Academy Press.
16       National Research Council. (1999) Research priorities for airborne  particulate matter. II. Evaluating research
17             progress and updating the portfolio. Washington, DC: National Academy Press.
18       U.S. Code. (1991) Clean Air Act, §108, air quality criteria and control techniques, §109, national ambient air
19             quality standards. U. S. C. 42:  §§7408-7409.
20       U.S. District Court of Arizona. (1995) American Lung Association v. Browner. West's Federal Supplement 884
21             F.Supp. 345 (No. CIV 93-643  TUC ACM).
22       U.S. Environmental Protection Agency. (1996a) Air quality criteria for particulate matter. Research Triangle Park,
23             NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001 aF-cF. 3v.
24       U.S. Environmental Protection Agency. (1996b) Review of the national ambient air quality standards for particulate
25             matter: policy assessment of scientific and technical  information. O AQPS staff paper. Research Triangle
26             Park, NC: Office of Air Quality Planning and Standards; report no. EPA/452/R-96-013. Available from:
27             NTIS, Springfield, VA; PB97-115406REB.
28       U.S. Environmental Protection Agency. (1998a) Particulate matter research needs for human health risk assessment
29             to support future reviews of the national ambient air quality standards for particulate matter. Research
30             Triangle Park, NC: National Center for Environmental Assessment; report no. EPA/600/R-97/132F.
31       U.S. Environmental Protection Agency. (1998b) Review of national ambient air quality standards for particulate
32             matter. Commer. Bus. Daily: February 19. Available: http://cbdnet.access.gpo.gov/index.html [1999,
33             November 24].
34       Wolff, G. T. (1996) Closure by the Clean Air Scientific Advisory Committee (CASAC) on the staff paper for
35             particulate matter [letter to Carol M. Browner, Administrator, U.S. EPA]. Washington, DC: U.S.
36             Environmental Protection Agency, Clean Air Scientific Advisory Committee;
37             EPA-SAB-CASAC-LTR-96-008; June 13.
         March 2001
                                                 1-16
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
      2.  PHYSICS, CHEMISTRY, AND MEASUREMENT
                     OF PARTICIPATE MATTER
      An extensive review of the physics and chemistry of participate matter (PM) was included
 in Chapter 3 of the 1996 EPA document Air Quality Criteria for Particulate Matter (U.S.
 Environmental Protection Agency, 1996).  Chapter 2 of this new version of the PM Air Quality
 Criteria Document (PM AQCD) provides background information on the physics and chemistry
 of atmospheric particles that may be useful in reading subsequent sections and chapters.
 New information needed to understand risk assessment is discussed, with emphasis placed on
 differences between fine and coarse particles and differences between the nuclei mode and the
 accumulation mode within fine particles.
      Chapter 4 of the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996) contained
 a review of the state-of-the-art of PM measurement technology.  Since that time, considerable
 progress has been made in understanding problems in the measurement of PM mass, chemical
 composition, and physical parameters. There also has been some progress in developing new and
 improved measurement techniques.  Therefore, a more extensive survey on measurement
 problems and on newly developed measurement techniques is included below in Section 2.2.
 For more detail and older references, the reader is referred to Chapter 3 and 4 of the 1996 PM
 AQCD (U.S. Environmental Protection Agency, 1996).
2.1  PHYSICS AND CHEMISTRY OF PARTICULATE MATTER
2.1.1  Definitions
     Atmospheric particles originate from a variety of sources and possess a range of
morphological, chemical, physical, and thermodynamic properties. Examples include
combustion-generated particles, such as diesel soot or fly ash; photochemically produced
particles, such as those found in urban haze; salt particles formed from sea spray; and soil-like
particles from resuspended dust. Some particles are liquid; some are solid. Others may contain a
solid core surrounded by liquid. Atmospheric particles contain inorganic ions, metallic
       March 2001
                                               2-1
                                                  DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
 19
20
21
22
 23
 24
 25
 26
 27
 28
 29
 30
 31
compounds, elemental carbon, organic compounds, and crustal compounds. Some atmospheric
particles are hygroscopic and contain particle-bound water. The organic fraction is especially
complex, containing hundreds of organic compounds. Primary particles are emitted directly from
sources. Secondary particles are formed in the atmosphere from products of chemical reactions
of gases from natural and anthropogenic sources such as SO2, NOX, and certain organic
compounds. The particle formation process includes nucleation of particles from low-vapor
pressure gases emitted from sources or formed in the atmosphere by chemical reactions;
condensation of low-vapor pressure gases on existing particles; and coagulation of particles.
Thus, any given particle may contain PM from many sources.
     The composition and behavior of airborne particles are fundamentally linked with those of
the surrounding gas.  Aerosol may be defined as a suspension of solid or liquid particles in air.
The term aerosol includes both the particles and all vapor or gas phase components of air.
However, the term aerosol is often used to refer to the suspended particles only. "Particulate" is
an adjective and should only be used as a modifier, as in particulate matter.
     A complete description of the atmospheric aerosol would include an accounting of the
chemical composition, morphology, and size of each particle and the relative abundance of each
particle type as a function of particle size (Friedlander, 1970). However, most often the physical
and chemical characteristics of particles are measured separately.  Size distributions by particle
number, from which surface area and volume distributions are calculated, often are determined
by physical means, such as electrical mobility or light scattering of suspended particles.
 Chemical composition usually is determined by analysis of collected samples, although sulfate
 can be measured in situ. The mass  and average chemical composition of particles, segregated
 according to aerodynamic diameter by cyclones or impactors, can also be determined. However,
 recent developments in single particle analysis techniques, by electron microscopy with X-ray
 analysis of single particles (but not agglomerates) collected on a substrate or by mass
 spectroscopy of suspended particles passing through a sensing volume, provide elemental
 composition of individual particles by particle size and, thus, are bringing the description
 envisioned by Friedlander (1970) closer to reality.
        March 2001
                                                  2-2
                                                      DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 2.1.2  Physical Properties and Processes
 2.1.2.1  Definitions of Particle Diameter
      The diameter of a spherical particle may be determined by optical or electron microscopy,
 by light scattering and Mie theory, by its electrical mobility, or by its aerodynamic behavior.
 However, atmospheric particles often are not spherical. Therefore, their diameters are often
 described by an "equivalent" diameter (i.e., that of a unit density sphere that would have the same
 physical behavior). The aerodynamic diameter is important for particle transport, collection,
 and respiratory tract deposition.  The aerodynamic diameter (Da) depends on the density of the
 particle. It is defined as the diameter of a spherical particle with a settling velocity equal to that
 of the particle in question, but with a density of 1 g/cm3.  Particles with the same physical size
 and shape but different densities will have different aerodynamic diameters. Detailed definitions
 of the various sizes and their relationships are given in standard aerosol textbooks (e.g.,
 Friedlander [1977], Reist [1984, 1993], Seinfeld and Pandis [1998], Hinds [1999], Vincent
 [1989, 1995], Willeke and Baron [1993], and Fuchs [1964, 1989]).

 2.1.2.2   Aerosol Size Distributions
      Particle size, as indexed by one of the "equivalent" diameters, is an important parameter in
 determining the properties, effects and fate of atmospheric particles. The atmospheric deposition
 rates of particles,  and therefore their residence times in the atmosphere, are a strong function of
 their aerodynamic diameters.  The aerodynamic diameter also influences deposition patterns of
 particles within the lung. Light scattering is strongly dependent on the optical particle size.
 Particle size distributions, therefore, have a strong influence on atmospheric visibility and,
 through their effect on radiative balance, on climate.  Studies using impactors or cyclones
 measure the particle-size distribution directly in aerodynamic diameter.  The diameters of
 atmospheric particles range from 1  nm to 100 yum, thus spanning 5 orders of magnitude.
 A variety of different instruments, measuring a variety of equivalent diameters, are required to
 cover this range.
     Older particle counting studies used optical particle counters to cover the range of 0.3 to
 30 ^m diameter. Diameters of particles below 0.5 ju.m were measured as mobility diameters.
 The particle diameters used in size distribution graphs from these studies usually are given as
physical diameters rather than aerodynamic diameters.  In recent years, aerodynamic particle
March 2001                                2-3        DRAFT-DO NOT QUOTE OR CITE

-------
 1     sizers, which give a direct measurement of the aerodynamic diameter in the range of
 2     approximately 0.7 to 10 //m diameter, have been used with electrical mobility analyzers, which
 3     measure the mobility diameter from approximately 0.5 /zm to very small particles of the order of
 4     0.005 jum, to cover the range of regulatory interest. Unfortunately, there is no agreed-upon
 5     technique for combining the various equivalent diameters.  Some workers use various
 6     assumptions to combine the various measurements into one presentation; others report each
 7     instrument separately. Therefore, the user of size distribution data must be careful to determine
 8     exactly which equivalent diameter is reported. Aerodynamic diameter is the most widely used
 9     equivalent diameter.  Therefore, particle diameters, unless otherwise indicated, refer to the
10     aerodynamic diameter in the discussions which follow below.
11
12     Particle Size Distribution Functions
13           The distribution of particles with respect to size is an important physical parameter
14     governing their behavior. Because atmospheric particles cover several orders of magnitude in
15     particle size, size distributions often are expressed in terms of the logarithm of the particle
16     diameter, on the X-axis, and the measured differential concentration on the Y-axis:
17     AN/A(logDp) = the number of particles per cm3 of air having diameters in the size range from
18     log Dp to log(Dp + ADp). Because logarithms do not have dimensions, it is necessary to think of
19     the distribution as a function of log(Dp/Dp0), where the reference diameter Dpo = 1 /^m is not
20     explicitly stated. If AN/A(logDp) is plotted on a linear scale, the number of particles between
21     Dp and Dp + ADp is proportional to the area under the curve of AN/A(logDp) versus logDp.
22     Similar considerations apply to distributions of surface, volume, and mass. It has been found that
23     atmospheric aerosol size distributions frequently may be approximated by a sum of log-normal
24     distributions corresponding to the various modes or fractions. When approximated by a function,
25     the distributions are usually given as dN/d(log Dp) rather than AN/A(log Dp).
26
27     Atmospheric Aerosol Size Distributions
28           Averaged atmospheric size distributions are shown in Figures 2-1 through 2-3 (Whitby,
29      1978; Whitby and Sverdrup, 1980).  Figure 2-1 describes the number of particles as a function of
30     particle diameter for rural, urban-influenced rural, urban, and freeway-influenced urban aerosols.
31     For some of the same data, the particle volume distribution is shown in Figure 2-2.  Figure 2-3
        March 2001
2-4
DRAFT-DO NOT QUOTE OR CITE

-------
            1,000,000

              10,000

                100 -
          :§     0.01 -
              0.0001 -

            0.000001 -
                   • Clean Rural
                    Urban Influenced Rural
                     	Average Urban
                     - • - • — Urban + Freeway
                   0.01    0.1     1     10    100
                    Particle Diameter, Dp (u.m)
                                                       200,000
                                                150,000 -
                                                    o
                                             Q.
                                             Q  100,000 -
                                             I
                                             TJ
                                                 50,000 -
                                                      0.01       0.1        1        10
                                                        Particle Diameter, Dp (|jm)
       Figure 2-1. Number of particles as a function of particle diameter: (a) number
                  concentrations are shown on a logarithmic scale to display the wide range by
                  site and size; (b) number concentrations for the average urban distribution are
                  shown on a linear scale for which the area under any part of the curve is
                  proportional to particle number in that size range.
       Source:  Whitby and Sverdrup (1980).
1
2
3
4
5
6
7
8
9
shows the number, surface, and volume distribution for the grand average continental size
distribution. Note that the particle diameter is always shown on a logarithmic scale. The particle
number is frequently shown on a logarithmic scale in order to display the wide range in number
concentration for different particle sizes and different sites. Volume and surface area, and
sometimes number, are shown on an arithmetic scale with the distributions plotted such that the
volume, surface area, or number of particles in any specified size range is proportional to the
corresponding area under the curve.  These distributions show that most of the particles are quite
small, below 0.1 y.m, whereas most of the .particle volume (and therefore most of the mass) is
found in particles >0.1  jj.m.
      March 2001
                                          2-5
DRAFT-DO NOT QUOTE OR CITE

-------
           70
           65
           60
           55
           50
           45
        S.30 -
        =§25 -
        1 20
           15
           10 -
            5 -
            0
Clean Rural
Urban Influenced
Rural
South-Central
New Mexico
   70
   65
   60  -
   55  -
   50  -

1 40
 I 35  H
 •a
 Q.30  -
 i25  '
 ^ 20  -
   15  -
   10  -
    5  -
    0
                    Average Urban
                    Urban + Freeway
             0.01      0.1       1        10
                       Particle Diameter, Dp (pm)
           100     0.01     0.1       1      10      100
                            Particle Diameter, Dp (|jm)
      Figure 2-2. Particle volume distribution as a function of particle diameter:  (a) for the
                  averaged rural and urban-influenced rural number distributions shown in
                  Figure 2-1 and a distribution from south central New Mexico, and (b) for the
                  averaged urban and freeway-influenced urban number distributions shown in
                  Figure 2-1.
      Source: Whitby and Sverdrup (1980) and Kim et al. (1993).
1           An important feature of the mass or volume size distributions of atmospheric aerosols is
2     their multimodal nature. Volume distributions, measured in ambient air in the United States, are
3     almost always found to be bimodal, with a minimum between 1 and 3 /um. The distribution of
4     particles that are mostly larger than the minimum is termed "coarse." The distribution of
5     particles that are mostly smaller than the minimum is termed "fine." Whitby and Sverdrup
6     (1980), Whitby (1978), and Willeke and Whitby (1975) identified three modes: (1) nuclei,
7     (2) accumulation, and (3) coarse. The three modes are most apparent in the freeway-influenced
8     size distribution of Figure 2-2b, in the surface area distribution of Figure 2-3b, and in the
9     in-traffic volume distribution of Figure 2-4. However, the nuclei mode, corresponding to
      March 2001
           2-6
DRAFT-DO NOT QUOTE OR CITE

-------
     (D
     .Q
     E
         I
           .
         O

         o
     CO  cvT
         O
         O)
         O
               15-
               10 -
5-
             600-
     8   -  400-
              200-
        a>
              30-
            Q
            o>
            ^
            <]  10-
                                    Nn = 7.7x 10
                                 DGNn = 0.013
                                                         (a)
   Na = 1.3x10
DGNa = 0.069
      = 2.03
   Nc = 4.2
DGNC = 0.97
  a   = 2.15
                                                                        (b)
                                   Sa = 535
                                 DGSa = 0,19
        Sn = 74
     DGSn = 0.023
                      Sc=41
                          Vn = 0.33
                        DGVn = 0.031
                  0.001
             0.01
                           10
                  100
Figure 2-3. Distribution of coarse (c), accumulation (a), and nuclei- or ultrafine (n) -mode
           particles by three characteristics, a) number (N), b) surface area (S) and
           c) volume (V) for the grand average continental size distribution. DGV =
           geometric mean diameter by volume; DGS = geometric mean diameter by
           surface area; DGN = geometric mean diameter by number; Dp = geometric
           diameter.
Source: Whitby(1978).

March 2001
                        2-7
             DRAFT-DO NOT QUOTE OR CITE

-------
                                                           Mechanically
                                                            Generated
        0.002
                Nuclei Mode
 0.1              1
Particle Diameter, Dp, um
Accumulation Mode
                                                                                100
                       Fine-Mode Particles
        Coarse Mode
    Coarse-Mode Particles
Figure 2-4.  Volume size distribution, measured in traffic, showing fine-mode and
            coarse-mode particles and the nuclei and accumulation modes within the
            fine-particle mode. DGV (geometric mean diameter by volume, equivalent to
            volume median diameter) and og (geometric standard deviation) are shown for
            each mode. Also shown are transformation and growth mechanisms (e.g.,
            nucleation, condensation, and coagulation).
Source: Adapted from Wilson and Suh( 1997).
particles below about 0.1 ftm, may not be noticeable in volume or mass distributions.  The
middle mode, from 0.1 to 1 or 2 /urn, is the accumulation mode. Fine particles include both the
accumulation and the nuclei modes. The third mode, containing particles larger than 1 or 2 fj-m,
is known as the coarse particle mode. The number concentrations of coarse particles are usually
too small to be seen in arithmetic plots (Figures 2-lb and 2-3a) but can be seen in a logarithmic
plot (Figure 2-lb). Whitby and Sverdrup (1980) observed that rural aerosols, not influenced by
sources, have a small accumulation mode and no observable nuclei mode.  For urban aerosols,
March 2001
        2-8
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
the accumulation and coarse particle modes are comparable in volume. The nuclei mode is small
in volume but it dominates the number distributions of urban aerosols.  Whitby's conclusions
were based on extensive studies of size distributions in a number of western and midwestern
locations during the 1970s (Whitby, 1978; Whitby and Sverdrup, 1980). No size-distribution
studies of similar scope have been published since then.  Newer results from particle counting
and impactor techniques, including data from Europe (U.S. Environmental Protection Agency,
1996) and Australia (Keywood et al., 1999, 2000), show similar results.

Definitions of Particle Size Fractions
     In the preceding discussion several subdivisions of the aerosol size distribution were
identified. Aerosol scientists use four different approaches or conventions in the classification of
particles by size:  (1) modes, based on the observed size distributions and formation mechanisms;
(2) cut point, usually based on the 50% cut point of the specific sampling device; (3) dosimetry
or occupational health sizes, based on the entrance into various compartments of the respiratory
system; and (4) legally specified, regulatory sizes for air quality standards.

     Modal.  The modal classification, first proposed by Whitby (1978), is shown in Figure 2-3.
The nuclei mode  can be seen clearly in the volume distribution only in traffic or near traffic or
other sources of nuclei mode particles (Figure 2-4).  The observed modal structure is frequently
approximated by  several log-normal distributions. Definitions of terms used to describe size
distributions in modal terms are given below.

     Coarse Mode:  The distribution of particles with diameters mostly greater than the
     minimum in the particle mass or volume distributions, which generally occurs between
     1 and 3 /urn. These particles are usually mechanically generated (e.g., from wind erosion of
     crustal material).

     Fine Mode: The distribution of particles with diameters mostly smaller than the minimum
     in the particle mass or volume distributions, which generally occurs between 1  and 3 ura..
     These particles are generated in combustion or formed from gases.  The fine mode includes
     the accumulation mode and the nuclei mode.
       March 2001
                                         2-9
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
     Nuclei Mode:  That portion of the fine particle mode with diameters below about 0.1 //m.
     The nuclei mode can be observed as a separate mode in mass or volume distributions only
     in clean or remote areas or near sources of new particle formation by nucleation.
     Toxicologists and epidemiologists use ultrafine to refer to particles in the nuclei-mode size
     range. Aerosol physicists and material scientists tend to use nanoparticles to refer to
     particles in this size range generated in the laboratory.

     Accumulation Mode: That portion of the fine particle mode with diameters above about
     0.1 ,um. Accumulation-mode particles normally do not grow into the coarse mode.
     Nuclei-mode particles grow by coagulation (two particles combining to form one) or by
     condensation (low-equilibrium vapor pressure gas molecules condensing on a particle) and
     "accumulate" in this size range.

     Over the years, the terms fine and coarse, as applied to particle sizes, have lost the precise
meaning given in Whitby's (1978) definition. In any given article, therefore, the meaning of fine
and coarse, unless defined, must be inferred from the author's usage.  In particular, PM2.5 and
fine-mode particles are not equivalent.  In this document, the term mode is used with fine and
coarse when it is desired to specify the distribution of fine-mode particles or coarse-mode
particles as shown in Figures 2-4 and 2-5.

     Sampler Cut Point. Another set of definitions of particle size fractions arises from
considerations of size-selective sampling. Size-selective sampling refers to the collection of
particles below or within a specified aerodynamic size range, usually defined by the upper 50%
cut point size, and has arisen in an effort to measure particle size fractions with some special
significance (e.g., health, visibility, source apportionment, etc.). Dichotomous samplers split the
particles into smaller and larger fractions, which may be collected on separate filters. However,
some fine particles (=10%) are collected with the coarse particle fraction. Cascade impactors use
multiple size cuts to obtain a distribution of size cuts for mass or chemical composition
measurements.  One-filter samplers with a variety of upper size cuts also have been used.
       March 2001
                                          2-10
DRAFT-DO NOT QUOTE OR CITE

-------
            70
           60   ~
           50   -
        J8 40
        D)
        O
 CO
 CO
           30

           20
           10   -
               0.1
                      Fine-Mode Particles
                                                     Coarse-Mode Particles
                0.2
0.5     1.0      2         5     10
  Aerodynamic Particle Diameter Da, urn
                                 Total Suspended Particles (TSP)
                                       PM
                                          10
                                PM,
                                   '2.5
                                                PM
                                                   (10-2.5f
                                                                          TSP
                                                                          HiVol
                                                                          WRAC
      Figure 2-5.  An idealized distribution of ambient particulate matter showing fine-mode
                  particles and coarse-mode particles and the fractions collected by size-selective
                  samplers.  (WRAC is the Wide Range Aerosol Classifier which collects the
                  entire coarse mode [Lundgren and Burton, 1995].)
      Source: Adapted from Wilson and Suh (1997).
1
2
3
4
5
     Occupational Health or Dosimetric Size Cuts. The occupational health community has
defined size fractions for use in the protection of human health. This convention classifies
particles into inhalable, thoracic, and respirable particles according to their upper size cuts.
However, these size fractions may also be characterized in terms of their entrance into various
compartments of the  respiratory system.  Thus, inhalable particles enter the respiratory tract,
      March 2001
                                        2-11
                          DRAFT-DO NOT QUOTE OR CITE

-------
1     including the head airways. Thoracic particles travel past the larynx and reach the lung airways
2     and the gas-exchange regions of the lung. Respirable particles are a subset of thoracic particles

3     which are more likely to reach the gas-exchange region of the lung.  In the past exact definitions
4     of these terms have varied among organizations. As of 1993, a unified set of definitions was

5     adopted by the American Conference of Governmental Industrial Hygienists (ACGIH) (1994),

6     the International Standards Organization (ISO), and the European Standardization Committee

7     (CEN). The curves which define inhalable (IPM), thoracic (TPM), and respirable (RPM)

8     particulate matter are shown in Figure 2-6.

9
                                                                     •  IPM
                                                                     •  TPM
                                                                     ORPM
                                                                     V  PM
                                       4            10    20          50
                                     Aerodynamic Diameter (|jm)
                                 100
      Figure 2-6.  Specified particle penetration (size-cut curves) through an ideal (no-particle-
                  loss) inlet for five different size-selective sampling criteria. PM10 is defined in
                  the Code of Federal Regulations (1991a). PM25 is also defined in the Federal
                  Register (1997).  Size-cut curves for inhalable particulate matter (IPM),
                  thoracic particulate matter (TPM) and respirable particulate matter (RPM)
                  size cuts are computed from definitions given by American Conference of
                  Governmental and Industrial Hygienists (1994).
      March 2001
2-12
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
      Regulatory Size Cuts.  In 1987, the NAAQS for PM were revised to use PM10, rather than
 total suspended particulate matter (TSP), as the indicator for the NAAQS for PM (Federal
 Register, 1987). The use of PM10 as an indicator is an example of size-selective sampling based
 on a regulatory size cut (Federal Register, 1987). The selection of PM10 as an indicator was
 based on health considerations and was intended to focus regulatory concern on those particles
 small enough to enter the thoracic region of the human respiratory tract. The PM2 5 standard, set
 in 1997, is also an example of size-selective sampling based on a regulatory  size cut (Federal
 Register, 1997). The PM25 standard was based primarily on epidemiological studies using
 concentrations measured with PM2 5 samplers as an exposure index. However, the PM2 5 sampler
 was not designed to collect respirable particles. It was designed to collect fine-mode particles
 because of their different sources (Whitby et al., 1974). Thus, the need to attain a PM2 5 standard
 will tend to focus regulatory concern on control of sources of fine-mode particles.
     Prior to 1997, the indicator for the NAAQS for PM was TSP.  TSP is defined by the design
 of the High Volume Sampler (hivol), which collects all of the fine particles but only part of the
 coarse particles. The upper cut-off size of the hivol depends on the wind speed and direction and
 may vary from 25 to 40 ^m.  The Wide Range Aerosol Classifier (WRAC) was designed
 specifically to collect-the entire coarse mode (Lundgren and Burton, 1995).
     An idealized distribution, showing the normally observed division of ambient aerosols into
 fine-mode particles and coarse-mode particles and the size fractions collected by the WRAC,
 TSP, PM10, PM2 5 and PM(10.2 5) samplers, is shown in Figure 2-5.  PM10 samplers, as defined in
 Appendix J to 40 Code of Federal Regulations (CFR) Part 50 (Code of Federal Regulations,
 1991a; Federal Register,  1987), collect all of the fine particles and part of the coarse particles.
 The upper cut point is defined as having a 50% collection efficiency at 10 ± 0.5 yum aerodynamic
 diameter. The slope of the collection efficiency curve is defined in amendments to 40 CFR,
 Part 53, (Code of Federal Regulations, 1991b).  An example of a PM10 size-cut curve is shown in
 Figure 2-6.
     An example of a PM2 5 size-cut curve is also shown in Figure 2-6. The  PM2 5 size-cut
 curve, however, is defined by the design of the Federal Reference Method Sampler.  The basic
 design of the FRM is  given in the Federal Register (1997, 1998) and as 40 CFR Part 50,
Appendix L in the Code of Federal Regulations (Code of Federal Regulations, 1999a).
Additional performance specifications are given in 40 CFR Parts 53 and 58 (Code of Federal
       March 2001
                                         2-13
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Regulations, 1999b). Each actual PM2 s reference method, as represented by a specific sampler
design and associated manual operational procedures, must be designated as a reference method
under Part 53 (see Section 1.2 of Appendix L). Thus there may be many somewhat different
PM2.s FRMs (currently, 6 have been designated).
     Papers discussing PM10 or PM2 5 frequently insert an explanation such as PMX (particles less
than x Aim diameter) or PMX (nominally, particles with aerodynamic diameter x are collected and not all particles 
-------
       c
       O
      W->
       E
       0
     100
      90-
      80-
      70-
      60-
      50-
      40-
      30-
      20-
      10-
0-f
  1
                                                              Wedding Cyclonic Inlet
                                                                 O  u=2 km/h
                                                                 D  U=8km/h
                                                                 A  U=24 km/h
                                                                Wedding and Weigand
                                                                       (1993)
                 Kimoto cyclonic inlet
                 ——Manufacturer
                    •  Tsai and Cheng (1996)
                             2             4       6     8   10
                                   Aerodynamic diameter, |jm
^—S-
 20
                                                                              30
      Figure 2-7.  Comparison of penetration curves for two PM,0 beta gauge samplers using
                 cyclone inlets. The Wedding PMJO sampler uses the U.S. EPA definition of
                 PMX as x = 50% cut point. The Kimoto PM,0 defines PMX as x = the 100% cut
                 point (or zero penetration).
      Source: Tsai and Cheng (1996).
1
2
3
4
5
6
2.1.2.3  Nuclei-Mode Particles
     As discussed in Section 8.5.5 of Chapter 8, Toxicology of Particulate Matter, and in
Chapter 6, Epidemiology of Human Health Effects from Ambient Particulate Matter, some
scientists argue that ultrafme (nuclei-mode) particles pose potential health problems and that
some health effects may be more closely associated with particle number or particle surface area
than particle mass. Because nuclei-mode particles contribute the major portion of particle
      March 2001
                                      2-15
                                                       DRAFT-DO NOT QUOTE OR CITE

-------
 1      number and a significant portion of particle surface area, some further attention to nuclei-mode
 2      particles is justified.
 3
 4      Formation and Growth of Fine Particles
 5            Several processes influence the formation and growth of particles.  New particles may be
 6      formed by nucleation from gas phase material. Particles may grow by condensation as gas phase
 7      material condenses on existing particles. Particles also may grow by coagulation as two particles
 8      combine to form one. Gas phase material condenses preferentially on smaller particles, and the
 9     rate constant for coagulation of two particles decreases as the particle size increases. Therefore,
10     nuclei mode particles grow into the accumulation mode, but accumulation mode particles do not
11      grow into the coarse mode (see Figure 2-4). More information and references on formation and
12     growth of fine particles may be found in the AQC PM 1996 (U.S. Environmental Protection
13     Agency, 1996).
14
15     Equilibrium Vapor Pressures
16           An important parameter in particle nucleation and in particle growth by condensation is the
17     saturation ratio S, defined as the ratio of the partial pressure of a species, p, to its equilibrium
18     vapor pressure above a flat surface, p0:  S = p/p0. For either condensation or nucleation to occur,
19     the species vapor pressure must exceed its equilibrium vapor pressure.  For particles, the
20     equilibrium vapor pressure is not the same as p0. Two effects are important: (1) the Kelvin
21     effect, which is an increase in the equilibrium vapor pressure above the surface due to its
22     curvature; thus very small particles have higher vapor pressures and will not be stable to
23     evaporation until they attain a critical size; and (2) the solute effect, which is a decrease in the
24      equilibrium vapor pressure of the liquid due to the presence of other compounds in solution.
25      Organic compounds may also be adsorbed on ultrafine carbonaceous particles.
26           For an aqueous solution of a nonvolatile salt, the presence of the salt decreases the
27      equilibrium vapor pressure of the water over the droplet. This effect is in the opposite direction
28      of the Kelvin effect, which increases the equilibrium vapor pressure above a droplet because of
29      its curvature. The existence of an aqueous solution will also influence the vapor pressure of
 30      water-soluble species. The  vapor pressure behavior of mixtures of several liquids or of liquids
 31      containing several solutes is complex.
        March 2001
2-16
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
 New Particle Formation
      When the vapor concentration of a species exceeds its equilibrium concentration (expressed
 as its equilibrium vapor pressure), it is considered condensable. Condensable species can either
 condense on the surface of existing particles or can form new particles. The relative importance
 of nucleation versus condensation depends on the rate of formation of the condensable species
 and on the surface or cross-sectional area of existing particles (McMurry and Friedlander, 1979).
 In ambient urban environments, the available particle surface area is sufficient to rapidly
 scavenge the newly formed condensable species.  Formation of new particles (nuclei mode) is
 usually not important except near sources of condensable species. Wilson et al. (1977) report
 observations of the nuclei mode in traffic.  New particle formation also can be observed in
 cleaner, remote regions. Bursts of new particle formation in the atmosphere under clean
 conditions usually occur when aerosol surface area concentrations are low (Covert et al., 1992).
 High concentrations of nuclei mode particles have been observed in regions with low particle
 mass concentrations, indicating that new particle formation is inversely related to the available
 aerosol surface area (Clarke, 1992).

Sources of Nuclei-Mode Particles
     Nuclei mode particles are the result of nucleation of gas phase species to form condensed
phase species with very low equilibrium vapor pressure, hi the atmosphere there are four major
classes of sources that yield particulate matter with equilibrium vapor pressures low enough to
form nuclei mode particles:
     (1)  Particles containing heavy metals. Nuclei mode particles of metal oxides or other
     metal compounds are generated when metallic impurities in coal or oil are vaporized during
     combustion and the vapor undergoes nucleation. Metallic ultrafine particles also may be
     formed from metals in lubricating oil or fuel additives that are vaporized during
     combustion of gasoline or diesel fuels. Nuclei-mode metallic particles were discussed in
     Section 6.9 of the 1996 PM AQCD (U.S. Environmental Protection Agency,  1996).
     (2)  Elemental carbon or soot (EC). EC particles are formed primarily by condensation of
     C2 molecules generated during the combustion process.  Because EC has a very low
     equilibrium vapor pressure, ultrafine EC particles can nucleate even at high temperatures
     (Kittelson, 1998; Morawska et al., 1998a).
       March 2001
                                         2-17
                                                             DRAFT-DO NOT QUOTE OR CITE

-------
 1           (3)  Sulfates and nitrates. Sulfuric acid (H2SO4), or its neutralization products with
 2           ammonia (NH3), ammonium sulfate ((NH4)2SO4) or ammonium acid sulfate (NH4HSO4),
 3           are generated in the atmosphere by conversion of sulfur dioxide (SO2) to H2SO4.  As H2SO4
 4           is formed, it can either nucleate to form new ultrafine particles, or it can condense on
 5           existing nuclei mode or accumulation mode particles (Clark and Whitby, 1975; Whitby,
 6           1978). The possible formation of ultrafine NH4NO3 by reaction of NH3 and HNO3
 7           apparently has not been investigated.
 8           (4)  Organic carbon.  Recent smog chamber studies and indoor experiments show that
 9           atmospheric oxidation of certain organic compounds found in the atmosphere can produce
10           highly oxidized organic compounds with an equilibrium vapor pressure sufficiently low to
11           result in nucleation (Kamens et al., 1999; Weschler and Shields,  1999).
12
13     Concentration of Nuclei-Mode Particles: A Balance Between Formation and Removal
14          Nuclei-mode particles may be removed by dry deposition or by growth into the
15     accumulation mode.  This growth takes place as other low vapor pressure material condenses on
16     the particles or as nuclei-mode particles coagulate with themselves or with accumulation mode
17     particles. Because the rate of coagulation would vary with the concentration of accumulation-
18     mode particles, it might be expected that the concentration of nuclei-mode particles would
19     increase with a decrease in accumulation-mode mass. On the other hand, the concentration of
20     particles would be expected to decrease with a decrease in the rate of generation of particles by
21     reduction in emissions of metal and carbon particles or a decrease in the rate of generation of
22     H2SO4 or condensable organic vapor. The rate of generation of H2SO4 depends on the
23     concentration of SO2 and OH, which is generated primarily by the photolysis of O3. Thus, the
24     reductions in SO2 and O3 that are expected to form a major basis for attaining PM2 5 and O3
25      standards and the implementation of Title H and Title IV Clean Air Act programs should lead to
26      a decrease in the rate of generation of H2SO4 and condensable organic vapor and to a decrease in
 27      the concentration of nuclei-mode particles.
 28           The balance between formation and removal is uncertain.  However, these processes can be
 29      modeled using a general dynamic equation for particle size distribution (Friedlander, 1977) or by
 30      aerosol dynamics modules in newer air quality models (Binkowski and Shanker, 1995;
 31      Binkowski and Ching, 1995).
        March 2001
2-18
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
      Further research is important due to the possibility of enhanced toxicity of ultrafine
 particles. It is possible that freshly generated ultrafine particles relatively near significant sources
 could present an additional risk to health, above those associated with particle mass per se.
 It will, therefore, be important to monitor particle number and surface as well as mass to further
 delineate the relative effectiveness of strategies for reducing particle mass, surface, and number.

 2.1.3 Chemistry of Atmospheric Particulate Matter
      The major constituents of atmospheric PM are sulfate,  nitrate, ammonium, and hydrogen
 ions; particle-bound water; elemental carbon; a great variety of organic compounds; and crustal
 material. Atmospheric PM also contains a large number of elements in various compounds and
 concentrations. More information and references on the composition of PM, measured in a large
 number of studies in the United States, may be found in 1996 PM AQCD (U.S. Environmental
 Protection Agency, 1996).  The composition and concentrations of PM are discussed in
 Chapter 3 of this document.

 2.1.3.1  Chemical Composition and Its Dependence on Particle Size
      Studies conducted in most parts of the United States indicate that sulfate, ammonium, and
 hydrogen ions; elemental carbon, secondary organic compounds and some primary organic
 compounds; and certain transition metals are found predominantly in the fine particle mode.
 Crustal materials such as calcium, aluminum, silicon, magnesium, and iron are found
 predominately in the coarse particles. Some organic materials such as pollen, spores, and plant
 and animal debris are also found predominantly in the coarse mode. Some components such as
 potassium and nitrate may be found in both the fine and coarse particle modes but from different
 sources or mechanisms. Potassium in coarse particles comes from soil.  Potassium also is found
 in fine particles in emissions from burning wood or cooking meat.  Nitrate in fine particles comes
primarily from the reaction of gas-phase nitric acid with gas-phase ammonia to form particulate
ammonium nitrate.  Nitrate  in coarse particles comes primarily from the reaction of gas-phase
nitric acid with preexisting coarse particles.
       March 2001
                                        2-19
                                                           DRAFT-DO NOT QUOTE OR CITE

-------
 1      2.1.3.2  Primary and Secondary Particulate Matter
 2           Particulate material can be primary or secondary. PM is called primary if it is in the same
 3      chemical form in which it was emitted into the atmosphere. PM is called secondary if it is
 4      formed by chemical reactions in the atmosphere. Primary coarse particles are usually formed by
 5      mechanical processes. This includes material emitted in particulate form such as wind-blown
 6      dust, sea salt, road dust, and combustion-generated particles such as fly ash and soot. Primary
 7      fine particles are emitted from sources, either directly as particles or as vapors that rapidly
 8      condense to form ultrafine or nuclei-mode particles.  This includes soot from diesel engines,
 9     a great variety of organic compounds condensed from incomplete combustion or cooking, and
10     compounds of As, Se, Zn, etc., which condense from vapor formed during combustion or
11      smelting. The concentration of primary particles depends on their emission rate, transport and
12     dispersion, and removal rate from the atmosphere.
13           Secondary PM is formed by chemical reactions of free, adsorbed, or dissolved gases.  Most
14     secondary fine PM is formed from condensable vapors generated by chemical reactions of
15     gas-phase precursors.  Secondary formation processes can result in either the formation of new
16     particles or the addition of particulate material to preexisting particles. Most of the sulfate and
17     nitrate and a portion of the organic compounds in atmospheric particles are formed by chemical
18     reactions in the atmosphere. Secondary aerosol formation depends on numerous factors
19     including the concentrations of precursors; the concentrations of other gaseous reactive species
20     such as ozone, hydroxyl radical, peroxy radicals, or hydrogen peroxide; atmospheric conditions
21     including solar radiation and relative humidity; and the interactions of precursors and preexisting
22     particles within cloud or fog droplets or on or in the liquid film on solid particles. As a result, it
23     is considerably more difficult to relate ambient concentrations of secondary species to sources of
24     precursor emissions than it is to identify the sources of primary particles. A significant effort is
25      currently being directed toward the identification and modeling of organic products of
26     photochemical smog including the conversion of gases to particulate matter.
27
28      Formation of Sulfates and Nitrates
29           A substantial fraction of the fine particle mass, especially during the warmer months of the
30      year, is secondary sulfate and nitrate, formed as a result of atmospheric reactions. Such reactions
 31      involve the gas phase conversion of SO2 to H2SO4 by OH radicals and aqueous-phase reactions of
        March 2001
2-20
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
 SO2 with H2O2, O3, or O2 (catalyzed by Fe and Mn).  These heterogeneous reactions may occur in
 cloud and fog droplets or in films on atmospheric particles. The NO2 portion of NOX can be
 converted to HNO3 by reaction with OH radicals during the day. At night, NOX also is oxidized
 to nitric acid by a sequence of reactions initiated by O3, that include nitrate radicals (NO3) and
 dinitrogenpentoxide (N2O5). Both H2SO4 and HNO3  react with atmospheric ammonia (NH3).
 Gaseous NH3 reacts with gaseous HNO3 to form particulate NH4NO3. Gaseous NH3 reacts with
 H2SO4 to form acidic HSO; (in NH4 HSO4) as well as in SO=4  in (NH4)2SO4.  In addition, acid
 gases such as SO2 and HNO3 may react with coarse particles to form coarse secondary PM
 containing sulfate and nitrate. Examples include reactions with basic compounds resulting in
 neutralization (e.g., CaCo3 + 2NHO3 - Ca (NO3)2 + H2CO31) or with salts of volatile acids
 resulting in release of the volatile acid (e.g., SO2 + 2NaCl + H2O - Na^O;, + 2HC11).
     If particulate NH4NO3 coagulates with an acidic sulfate particle (H2SO4 or HSO;), gaseous
 HNO3 will be released and the NH3 will increase the neutralization of the acidic sulfate. Thus, in
 the eastern United States where PM tends to be acidic, sulfate is usually a larger fraction of PM
 mass than nitrate. However, in the western United States, where higher NH3 and lower SO2
 emissions permit  complete neutralization of H2SO4, the concentration of nitrate may be higher
 than that of sulfate.  As SO2 concentrations in the atmosphere in the eastern United States are
 reduced, the NH3  left in the atmosphere after neutralization of H2SO4 will be able to react with
 HNO3 to form NH4NO3. Therefore, a reduction in SO2 emissions, especially without a reduction
 in NOX emissions, could lead to an increase in NH4NO3 concentrations (West et al., 1999; Ansari
 and Pandis, 1998). Thus, possible environmental effects of NH4NO3 are of interest for both the
 western and eastern United States.
     Chemical reactions of SO2 and NOX within plumes are an important source of H+,  SO4,  and
NO3. These conversions can occur by gas-phase and aqueous-phase mechanisms.  In power-
plant or smelter plumes containing SO2 and NOX, the gas-phase chemistry depends on plume
dilution,  sunlight,  and volatile organic compounds, either in the plume or in the ambient air
mixing into and diluting the plume.  For the conversion of SO2 to H2SO4, the gas-phase rate in
such plumes during summer midday conditions in the eastern United States typically varies
between  1 arid 3% h'1 but in the cleaner western United States rarely exceeds 1% h'1. For the
conversion of NOX to HNO3, the gas-phase rates appear to be approximately three times faster
       March 2001
                                        • 2-21
                                                             DRAFT-DO NOT QUOTE OR CITE

-------
 1      than the SO2 conversion rates. Winter rates for SO2 conversion are approximately an order of
 2      magnitude lower than summer rates.
 3           The contribution of aqueous-phase chemistry to particle formation in point-source plumes
 4      is highly variable, depending on the availability of the aqueous phase (wetted aerosols, clouds,
 5      fog, and light rain) and the photochemically generated gas-phase oxidizing agents, especially
 6      H2O2 for SO2 chemistry. The in-cloud conversion rates of SO2 to SO^ can be several tunes
 7      larger than the gas-phase rates given above. Overall, it appears that SO2 oxidation rates to SO;
 8      by gas-phase and aqueous-phase mechanisms may be comparable in summer, but aqueous phase
 9      chemistry may dominate in winter.
10          In the western United States, markedly higher SO2 conversion rates have been reported in
11      smelter plumes than in power plant plumes. The conversion is predominantly by a gas-phase
12     mechanism. This result is attributed to the lower NOX in smelter plumes.  In power plant plumes,
13     NO2 depletes OH and competes with SO2 for OH.
14          In urban plumes, the upper limit for the gas-phase SO2 conversion rate appears to be about
15     5% fr1 under the more polluted conditions.  For NO2, the rates appear to be approximately three
16     times faster than the SO2 conversion rates.  Conversion rates of SO2 and NOX in background air
17     are comparable to the peak rates in diluted plumes. Neutralization of H2SO4 formed by SO2
18     conversion increases with plume age and background NH3 concentration. If the NH3
19     concentrations are more than sufficient to neutralize H2SO4 to (NH4)2SO4, the HNO3 formed from
20     NOX conversions may be converted to NH4NO3.
21
22     Organic Aerosol
23           Organic compounds contribute from 20 to 60% of the dry fine particle mass in the
24     atmosphere (Gray et al., 1984; Shah et al., 1986; U.S. Environmental Protection Agency, 1996).
25     However, organic PM concentrations, composition, and formation mechanisms are poorly
26     understood. Particulate organic matter is an aggregate of hundreds of individual compounds
27     spanning a wide range of chemical and thermodynamic properties (Saxena and Hildemann,
28      1996). Some of the organic compounds are "semivolatile" (i.e., they have atmospheric
 29     concentrations and saturation vapor pressures), such that both gaseous and condensed phases
 30      exist in equilibrium hi the atmosphere. The presence of semivolatile or multiphase organic
 31      compounds complicates the sampling process. Organic compounds, originally in the gas phase,
        March 2001
2-22
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
 may be absorbed on glass or quartz filter fibers (positive artifact). Semivolatile compounds,
 originally present in the condensed phase, may evaporate from particles collected on glass,
 quartz, or Teflon filters (negative artifact). In addition, no single analytical technique is currently
 capable of analyzing the entire range of organic compounds present in the atmosphere in PM.
 Even rigorous analytical methods are able to identify only 10 to 20% of the organic PM mass on
 the molecular level (Rogge et al., 1993a). Even in smog chamber studies of specific compounds,
 only about 50% of the condensed phase compounds could be identified (Forstner et al., 1997a,b).
 Measurement techniques are discussed in Section 2.2.3.2. Information on the identification and
 concentration of the many different organic compounds identified in atmospheric samples is
 given in Chapter 3.

 Formationof Secondary Organic Paniculate Matter
     Atmospheric reactions involving volatile organic compounds such as alkanes, alkenes,
 aromatics, cyclic olefins, and terperies (or any reactive organic gas that contains at least seven
 carbon atoms) yield organic compounds with low ambient temperature, saturation vapor
 pressures. Such reactions may occur in the gas phase, in  fog or cloud droplets (Graedel and
 Goldberg, 1983; Faust, 1994) or possibly in aqueous aerosols (Aumont et al., 2000). Reaction
 products from the oxidation of reactive organic gases also may nucleate to form new particles or
 condense on existing particles to form secondary organic PM. Organic compounds with two
 double bounds may react to form dicarboxylic acids, which, with four or more carbon atoms, also
 may condense.  Both biogenic and anthropogenic sources contribute to primary and secondary
 organic particulate matter (Grosjean, 1992; Hildemann et al., 1996; Mazurek et al., 1997;
 Schauer et al., 1996). Oxalic acid was the most abundant organic acid found in PM2 5 in
 California (Poore, 2000).
     Although the mechanisms and pathways for forming inorganic secondary particulate matter
 are fairly well known, those for forming secondary organic PM are not as well understood.
 Ozone and the hydroxyl radical are thought to be the major initiating reactants. However, HO2
 and NO3 radicals also may initiate reactions and organic radicals may be nitrated by HNO2,
 HNO3, or NO2.  Pun et al. (2000) discuss formation mechanisms for highly oxidized,
 multifunctional organic compounds. The production of such species has been included in a
photochemical model by Aumont et al. (2000).  Understanding the mechanisms of formation of
       March 2001
                                         2-23
DRAFT-DO NOT QUOTE OR CITE

-------
 1     secondary organic PM is important because secondary organic PM can contribute in a significant
 2     way to ambient PM levels, especially during photochemical smog episodes. Experimental
 3     studies of the production of secondary organic PM in ambient air have focused on the Los
 4     Angeles Basin. Turpin and Huntzicker (1991, 1995) and Turpin et al. (1991) provided strong
 5     evidence that secondary PM formation occurs during periods of photochemical ozone formation
 6     in Los Angeles and that as much as 70% of the organic carbon hi ambient PM was secondary in
 7     origin during a smog episode in 1987. Schauer et al. (1996) estimated that 20 to 30% of the total
 8     organic carbon PM in the <2.1 //m size range in the Los Angeles airshed is secondary in origin
 9     on an annually averaged basis.
10           Pandis et al. (1992) identified three mechanisms for formation of secondary organic PM:
11     (1) condensation of oxidized end-products of photochemical reactions (e.g., ketones, aldehydes,
12     organic acids, hydroperoxides), (2) adsorption of organic gases onto existing solid particles (e.g.,
13     polycyclic aromatic hydrocarbons), and (3) dissolution of soluble gases that can undergo
14     reactions  in particles (e.g., aldehydes).  The first and third mechanisms are expected to be of
15     major importance during the summertime when photochemistry is at its peak. The second
16     pathway can be driven by diurnal and seasonal temperature and humidity variations at any time
17     of the year.  With regard to the first mechanism, Odum et al. (1996) suggested that the products
18     of the photochemical oxidation of reactive organic gases are semivolatile and can partition
19     themselves onto existing organic carbon at concentrations below their saturation concentrations.
20     Thus, the yield of secondary organic PM depends not only  on the identity of the precursor
21     organic gas but also on the ambient levels of organic carbon capable of absorbing the oxidation
22     product.
23           Haagen-Smit (1952) first demonstrated that hydrocarbons irradiated in the presence of NOX
24     produce light scattering aerosols.  The aerosol forming potentials of a wide variety of individual
25     anthropogenic and biogenic hydrocarbons were compiled by Pandis et al. (1992) based mainly on
26     estimates made by Grosjean and Seinfeld (1989) and data from Pandis et al. (1991) for p-pinene
27     and Izumi and Fukuyama (1990) for aromatic hydrocarbons. Zhang et al. (1992) examined the
28     oxidation of a-pinene.  Pandis et al. (1991) found no aerosol products formed in the
29     photochemical oxidation of isoprene, although they and Zhang et al.  (1992) found that the
30     addition of isoprene to reaction mixtures increased the reactivity of the systems studied.  Further
31     details about the oxidation mechanisms and secondary organic PM yields from various reactive
        March 2001
2-24
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
31
 organic gases given in the above studies. Estimates of the production rate of secondary organic
 PM in the Los Angeles airshed are provided in the previous PM AQCD (U.S. Environmental
 Protection Agency, 1996).
      More recently, Odum et al. (1997a,b) have found that the aerosol formation potential of
 whole gasoline vapor can be accounted for solely by summing the contributions of the individual
 aromatic compounds in the fuel.  In general, data for yields for secondary organic PM formation
 can be broken into two distinct categories.  The oxidation of toluene and aromatic compounds
 containing ethyl or propyl groups (i.e., ethylbenzene, ethyltoluene, n-propylbenzene) produced
 higher yields of secondary organic PM than did the oxidation of aromatic compounds containing
 two or more methyl groups (i.e., xylenes, di-, tri-, tetra-methylbenzenes),  Yields in the first
 group ranged from about 7 to 10% and in the second group were generally between 3 and 4%
 within a range of existing organic carbon levels between 13 and 100 yUg/m:3. This grouping is
 consistent with those found by Izumi and Fukuyama (1990). Reasons for the differences in
 secondary organic PM yields found between the two classes of compounds are not clear.
      Kao and Friedlander (1995) examined the statistical properties of a number of PM
 components in the South Coast Air Basin (Los Angeles area). They found that, regardless of
 source type and location within their study area, the concentrations of nonreactive, primary
 components of PMIO had approximately log normal frequency distributions with constant values
 of the geometric standard deviations (GSDs). However, aerosol constituents of secondary origin
 (e.g.,  SO4=, NH4+, and NO3") were found to have much higher GSDs. Surprisingly, the GSDs of
 organic (1.87) and elemental (1.74) carbon were both found to be within la (0.14) of the mean
 GSD  (1.85) for nonreactive primary species, compared to GSD's of 2.1 for sulfate, 3.5 for
 nitrate, and 2.6 for ammonium.  These results suggest that most of the organic carbon seen in
 ambient samples in the South Coast Air Basin was of primary origin.  Pinto et al. (1995) found
 similar results  for data obtained during the summer of 1994.  Further studies are needed to
 determine if these relations are valid at other locations and to what extent the results might be
 influenced by the evaporation of volatile constituents during or after sampling.  It must be
 emphasized that the inferences drawn from field studies in the Los Angeles Basin are unique to
that area and cannot be extrapolated to other areas of the country.
       March 2001
                                         2-25
DRAFT-DO NOT QUOTE OR CITE

-------
 1      Secondary Organic Aerosol Formation from Oxidation Products ofBiogenic Hydrocarbons
 2           The formation of atmospheric aerosols from biogenic emissions has been of interest for
 3      many years.  Recently, more quantitative results have been reported. Hoffmann et al. (1997)
 4      found secondary organic PM yields of ~ 5% for open-chain biogenic hydrocarbons such as
 5      ocimene and linalool, 5 to 25% for monounsaturated cyclic monoterpenes such as a-pinene,
 6      d-3 carene and terpinene-4-ol, and =40% for a cyclic monoterpene with two, double bonds such
 7      as d-limonene.  Secondary organic PM yields of close to 100% were observed during the
 8      photochemical oxidation of one sesquiterpene, trans-caryophyllene. These results were all
 9      obtained for initial hydrocarbon mixing ratios of 100 ppb.
10          Kamens et al. (1999) observed secondary organic PM yields of 20 to 40% for a-pinene.
11      Using information on the composition of secondary PM formed from a-pinene (Jang and.
12     Kamens, 1999), they were able to calculate formation rates with a kinetic model including
13     formation mechanisms for O3 + a-pinene reaction products.  Griffin et al. (1999) introduced the
14     concept of incremental aerosol reactivity, the change in the secondary organic aerosol mass
15     produced (in Mg/m3) per unit change of parent organic reacted (in ppb), as a measure of the
16     aerosol-forming capability of a given parent organic compound in a prescribed mixture of other
17     organic compounds.  They measured the incremental aerosol reactivity for a number of aromatic
18     and biogenic compounds for four initial mixtures. Incremental aerosol reactivity ranged from
19     0.133 to 10.352 //gm'3 ppb'1 and varied by almost a factor of two depending on the initial
20     mixture.
21           Recent laboratory and field studies support the concept that nonvolatile and semivolatile
22     oxidation products from the photooxidation of biogenic hydrocarbons contribute significantly to
23     ambient PM concentrations in both urban and rural environments.  A number of multifunctional
24     oxidation products have been identified in laboratory studies (Yu et al., 1998; Glasius et al.,
25     2000; Christoffersen et al., 1998; Koch et al., 2000; Leach et al., 1999). Many of these
26      compounds have subsequently been identified in field investigations (Yu et al., 1999; Kavouras
27      et al., 1998,1999a,b; Casimiro et al., 2001; Castro et al., 1999). However, further investigations
28      are needed  to accurately access their overall contributions to PM2 5 concentrations.
 29           Sampling and characterization of PM in the ambient atmosphere and in important
 30      microenvironments is required to address important issues in exposure, toxicology, and
 31      compliance. Currently, it is not possible to fully quantify the concentration, composition, or
        March 2001
2-26
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 sources of the organic components. Many of the secondary organic aerosol components are
 highly oxidized, difficult to measure, multifunctional compounds. Additional laboratory studies
 are needed to identify such compounds, strategies need to be developed to sample and measure
 such compounds in the atmosphere, and models of secondary organic aerosol formation need to
 be improved and added to air quality models in order to address compliance issues related to
 reducing PM mass concentrations that affect human exposure.
      A high degree of uncertainty is associated with all aspects of the calculation of secondary
 organic PM concentrations. This is compounded by the volatilization of organic carbon from
 filter substrates during and after sampling as well as potential positive artifact formation from the
 absorption of gaseous hydrocarbon on quartz filters. Significant uncertainties always arise in the
 interpretation of smog chamber data because of wall reactions.  Limitations also exist in
 extrapolating the results of smog chamber studies to ambient conditions found in urban airsheds
 and forest canopies. Concentrations of terpenes and NOX are much lower in forest canopies
 (Altshuller,  1983) than the levels commonly used in smog chamber studies. The identification of
 aerosol products of terpene oxidation has seldom been a specific aim of field studies, making it
 difficult to judge the results of model calculations of secondary organic PM formation.
 Uncertainties also arise because of the methods used to measure biogenic hydrocarbon emissions.
 Khalil and Rasmussen (1992) found much lower ratios of terpenes to other hydrocarbons (e.g.,
 isoprene) in forest air than were expected, based on their relative emissions strengths and rate
 coefficients  for reaction with OH radicals and O3.  They offered two explanations: (1) either the
 terpenes were being removed rapidly by some heterogeneous process, or (2) emissions were
 enhanced artificially by feedbacks caused by the bag enclosures they used. If the former
 consideration is correct, then the production of aerosol carbon from terpene emissions could be
 substantial; if the latter is  correct, then terpene emissions could have been overestimated by the
 techniques used.

 2.1.3.3  Particle-Vapor Partitioning
     Several atmospheric aerosol species, such as ammonium nitrate and  certain organic
 compounds, are semivolatile and are found in both gas and particle phases. A variety of
thermodynamic models have been developed to predict the temperature and relative humidity
dependence of the ammonium nitrate equilibria with gaseous nitric acid and ammonia. However,
       March 2001
                                         2-27
DRAFT-DO NOT QUOTE OR CITE

-------
 1     under some atmospheric conditions, such as cool, cold, or very clean air, the relative
 2     concentrations of the gas and solid phases are not accurately predicted by equilibrium
 3     considerations alone, and transport kinetics can be important.  The gas-particle distribution, of
 4     semivolatile organic compounds depends on the equilibrium vapor pressure of the compound,
 5     total particle surface area, particle composition, atmospheric temperature, and relative humidity.
 6     Although it generally is assumed that the gas-particle partitioning of semivolatile organics is in
 7     equilibrium in the atmosphere, neither the equilibria nor the kinetics of redistribution are well
 8     understood. Diurnal temperature fluctuations, which cause gas-particle partitioning to be
 9     dynamic on a time scale of a few hours, can cause semivolatile compounds to evaporate during
10     the sampling process. The pressure drop across the filter can also contribute to loss of
11     semivolatile compounds. The dynamic changes in gas-particle partitioning, caused by changes in
12     temperature, pressure, and gas-phase concentration, both in the atmosphere and after collection,
13     cause serious sampling problems that are discussed in Section 3.2.3.
14
15     Equilibria with Water Vapor
16           As a result of the equilibrium of water vapor with liquid water in hygroscopic particles,
17     many ambient particles contain liquid water (particle-bound water). Unless removed, this
18     particle-bound water will be measured as a component of the particle mass.  Particle-bound water
19     is important in that it influences the size of the particles and in turn their aerodynamic properties
20     (important for deposition to surfaces, to airways following inhalation, and in sampling
21     instrumentation) and their light scattering properties. The aqueous solution provides a medium
22     for reactions of dissolved gases, including reactions that do not take place in the gas phase. The
23     aqueous solutions also may act as a carrier to convey soluble toxic species to the gas-exchange
24     regions of the respiratory system,  including species that would be removed by deposition in the
25     upper airways if in the gas phase (Friedlander and Yeh, 1998; Kao and Friedlander, 1995;  ,
26     Wilson, 1995).  An extensive review of this equilibrium as it pertains to ambient aerosols was
27     given in Chapter 3 of the 1996 PM AQCD (U.S. Environmental Protection, Agency, 1996).
28           The interaction of particles with water vapor may be described briefly as follows.
29     As relative humidity increases, particles of crystalline soluble salts, such as (NH4) 2SO4,
30     NH4HSO4, or NH4NO3, undergo a phase transition to become aqueous solution particles.
31     According to the phase rule, for particles consisting of a single component, this phase transition
        March 2001
2-28
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
30
31
  is abrupt, taking place at a relative humidity that corresponds to the vapor pressure of water
  above the saturated solution (the deliquescence point). With further increase in relative
  humidity, the solution particle adds water (and the concentration of the solute decreases) so that
  the vapor pressure of the solution is maintained equal to that of the surrounding relative
  humidity; thus, the solution particle tends to follow the equilibrium growth curve. As relative
  humidity decreases, the solution particle follows the equilibrium curve to the deliquescence
  point. However, rather than crystallizing at the deliquescence relative humidity, the solute
  remains dissolved in a supersaturated solution to considerably lower relative humidities.
  Ultimately the solution particle abruptly loses its water vapor (efflorescence), returning typically
 to the initial crystalline form.
      For particles consisting of more than one component, the solid to liquid transition will take
 place over a range of relative humidities, with an abrupt onset at the lowest deliquescence point
 of the several components, and with subsequent growth as crystalline material in the particle
 dissolves according to the phase diagram for the particular multicomponent system.  Under such
 circumstances, a single particle may undergo several more or less abrupt phase transitions until
 the soluble material is folly dissolved. At decreasing relative humidity, such particles tend to
 remain in solution to relative humidities well below the several deliquescence points. In the case
 of the sulfuric acid-ammonium sulfate-water system, the phase diagram is fairly completely
 worked out.  Mixed anion systems containing nitrate are more difficult because of the
 equilibrium between paniculate NH4NO3 and gaseous NH3 arid HNO3.  For particles of
 composition intermediate between NH4HSO4 and (NH4)2SO4, this transition occurs in the range'
 from 40% to below 10%, indicating that for certain compositions the solution cannot be dried in
 the atmosphere. At low relative humidities, particles of this composition would likely be present
 in the atmosphere as supersaturated solution droplets (liquid particles) rather than as solid
 particles.  Thus, they would exhibit hygroscopic rather than deliquescent behavior during relative
 humidity cycles.
      Other pure compounds,  such as sulfuric acid (H2SO4), are hygroscopic (i.e., they form water
 solutions at any relative humidity and maintain a solution vapor pressure over the entire range of
relative humidity). Soluble organic compounds may also contribute to the hygroscopicity of the
atmospheric aerosol (Saxena et al., 1995; Saxena and Hildeman, 1996), but the equilibria
involving  organic compounds and water vapor, and especially for mixtures of salts, organic
        March 2001
                                                  2-29
                                                      DRAFT-DO NOT QUOTE OR CITE

-------
 1      compounds, and water, are not so well understood. These equilibrium processes may cause an
 2      ambient particle to significantly increase its diameter at relative humidities above about 40%
 3      (Figure 2-8). A particle can grow to five times its dry diameter as the RH approaches 100%
 4      (Figure 2-9). The Federal Reference Methods, for filter measurements of PM2.5 and PM10 mass,
 5      require, after collection, equilibration at a specified, low relative humidity (=40% RH). This
 6      equilibration removes much of the particle-bound water and provides a stable PM mass (see
 7      Section 2.2 for details and references). Otherwise, particle mass would be a function of relative
 8      humidity and, at higher relative humidities, the particle mass would be largely particle-bound
 9      water.
10            Continuous monitoring techniques must remove particle-bound water before measurement,
11      either by heating or dehumidification. Semivolatile material may be lost during sampling or
12     equilibration. It is certainly lost when the collected sample is heated above ambient. In addition
13     to problems due to the loss of semivolatile species, recent studies have shown that significant
14     amounts of particle-bound water are retained in particles collected on impaction surfaces even
15     after equilibration and that the amount of retained particle-bound water increases with relative
16     humidity during collection (Hitzenberger et al., 1997).  Large increases in mass with increasing
17     relative humidity were observed for the accumulation mode. The change in particle size with
18     relative humidity also means that particle measurements such as surface area or volume, or
19     composition as a function of size, all must be made at the same RH if the results are to be
20     comparable.  These problems are addressed below in more detail, in Section 2.2 on Measurement
21     of Particulate Matter.
22
23      2.1.3.4  Removal Processes
24           The lifetimes of particles vary with size. Coarse  particles can settle rapidly from the
 25      atmosphere within hours, and normally travel only short distances. However, when mixed high
 26      into the atmosphere, as in dust storms, the smaller-sized coarse-mode particles may have longer
 27      lives and travel distances. Nuclei mode particles rapidly grow into the accumulation mode.
 28      However, the accumulation mode does not grow into the coarse mode.  Accumulation-mode fine
 29      particles are kept suspended by normal air motions and have very low deposition rates to
 30      surfaces. They can be transported thousands of km and remain in the atmosphere for a number of
 31      days. Coarse-mode particles of less than ~ 10 //m diameter, as well as accumulation-mode and
        March 2001
2-30
DRAFT-DO NOT QUOTE OR CITE

-------
                                                                        90
       Figure 2-8. Particle growth curves showing fully reversible hygroscopic growth of sulfuric
                  acid (H2SO4) particles, deliquescent growth of ammonium sulfate [(NH4)2 SO4]
                  particles at about 80% relative humidity (RH), hygroscopic growth of
                  ammonium sulfate solution droplets at RH greater than 80%, and hysteresis
                  (the droplet remains supersaturated as the RH decreases below 80%) until the
                  crystallization point is reached.
       Source: National Research Council (1993) adapted from Tang (1980),
1
2
3
4
5
nuclei-mode (or ultrafine) particles, all have the ability to penetrate deep into the lungs and to be
removed by deposition in the lungs. Dry deposition rates are expressed in terms of a deposition
velocity that varies with particle size, reaching a minimum between 0.1 and 1.0 jura aerodynamic
diameter. Accumulation-mode particles are removed from the atmosphere primarily by cloud
processes. Fine particles, especially particles with a hygroscopic component, grow as the relative
      March 2001
                                         2-31
                                                           DRAFT-DO NOT QUOTE OR CITE

-------
          5-
      o
     Q
2  4H
'•s
a:
I3
CD
          2 ~
                            T
                                      T
             0
                  	 Theoretical Prediction at 22°c
                  ooooo Experimental Measurements
                                                         RH=99.8%
                                                         RH=99.7%
                                                         !————""""""
                                                         RH=99.5%
                                                                            216
                    -125
                                                                           -64
                                                                           -27
                                                                           -8
                 T—i—|—T—i—i—T—T—r—FT—r—i—i—r—i—r
                      50            100            150           200
                   NH4  HSO4  Dry Particle Diameter (nm)
                                                                                  f
      Figure 2-9. Theoretical predictions and experimental measurements of growth of
                NH4HSO4 particles at relative humidity between 95 and 100%.
      Source: Lietal. (1992).
1
2
3
4
5
6
1
8
humidity increases, serve as cloud condensation nuclei, and grow into cloud droplets. If the
cloud droplets grow large enough to form rain, the particles are removed in the rain. Falling rain
drops impact coarse particles and remove them. Ultrafine or nuclei mode particles are small
enough to diffuse to the falling drop, be captured, and be removed in rain. Falling rain drops,
however, are not effective in removing accumulation-mode particles.
      March 2001
                                      2-32
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
26
27
28
29
30
 2.1.3-5  Particulate Matter and Welfare Effects
      The EPA is required by law to set primary standards to protect human health and secondary
 standards  to mitigate welfare effects. The role of particles in reducing visibility and affecting
 radiative balance through scattering and absorption of light is evident, as are the effects of
 particles in soiling and damaging materials. Visibility effects are addressed through regional
 haze regulations and are considered in establishing secondary NAAQS.  The direct effects of
 particles in scattering and absorbing light and the indirect effects of particles on clouds are being
 addressed in climate change programs in several government agencies with the lead role assigned
 to the Department of Energy. These welfare effects are discussed briefly in Chapter 4. The
 effects on  vegetation resulting from the direct and indirect effects of particles on light flux also
 are discussed in Chapter 4.
      Concerns over the possible ecological effects of acid deposition in the United States led to
 the creation of a major research program in 1980 under the new National Acid Precipitation
 Assessment Program (NAPAP).  However, the role of PM in acid deposition has not always been
 recognized. Acid deposition and PM are intimately related, however, first because particles
 contribute  significantly to the acidification of rain and secondly because the gas phase species
 that lead to dry deposition of acidity are also precursors of particles. Therefore, reductions in
 SO2 and NOX emissions will decrease both acidic deposition and PM concentrations.
      Sulfate, nitrate, and some partially oxidized organic compounds are hygroscopic and act as
 nuclei for the formation of cloud droplets. These droplets serve as chemical reactors in which
 (even slightly) soluble gases can dissolve and react.  Thus, SO2 can dissolve in cloud droplets and
 be oxidized to sulfuric acid by dissolved ozone or hydrogen peroxide. These  reactions do not
 take place in the gas phase but only in solution in water.  Sulfur dioxide also may be oxidized by
 dissolved oxygen. This process will be faster if metal catalysts such as iron or manganese are
 present in solution.  If the droplets evaporate, larger particles are left behind.  If the droplets grow
 large  enough, they will fall as rain, and the particles will be removed from the atmosphere with
 potential effects on the materials, plants, or soil on which the rain falls. (Similar considerations
 apply to dew.) Atmospheric particles that nucleate cloud droplets also may contain other soluble
 or nonsoluble materials such as metal salts and PNA organic compounds that may add to the
toxicity of the rain.  Thus, the adverse effects of acid deposition on soils, plants, and trees as well
        March 2001
                                          2-33
                                                              DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
 30
as lakes, streams and fish must be taken into account in setting secondary PM standards. These
effects are discussed in Chapter 4.
     Sulfuric acid, ammonium nitrate, ammonium sulfates, and organic particles also are
deposited on surfaces by dry deposition. The utilization of ammonium by plants leads to the
production of acidity. Therefore, dry deposition of particles can also contribute to the ecological
damages caused by acid deposition.                       j3

2.1.4 Summary
     The physical and chemical properties of ultrafine mode, accumulation mode, and coarse
mode particles are summarized in Table 2-1.
2.2  MEASUREMENT OF PARTICIPATE MATTER
     The 1996 PM AQCD (U.S. Environmental Protection Agency, 1996) summarized sampling
and analytical techniques for PM and acid deposition that had appeared in the literature since the
earlier 1982 PM AQCD (U.S. Environmental Protection Agency, 1982). Excellent reviews have
been published by Chow (1995) and McMurry (2000).  This section discusses problems in
measuring PM; new techniques that attempt to alleviate these problems or measure problem
species; the current EPA monitoring program (including measurements with Federal Reference
Methods, speciation monitors, and continuous monitors); and the importance of intercomparison
studies in the absence of any reference standard for suspended atmospheric particles.

2.2.1  Problems in Measuring Particulate Matter
     The EPA decision to revise the PM standards by adding daily and yearly standards for
PM2 5 has led to a renewed interest in the measurement of atmospheric particles and also to a
better understanding of the problems in developing precise and accurate measurements of
particles. Unfortunately, it is very difficult to measure and characterize particles suspended in
the atmosphere.
     The U.S. Federal Reference Methods (FRM) for PM2 5 and PM10 provide relatively precise
 (±10 %) methods for determining the mass of material remaining on a Teflon filter after
        March 2001
                                         2-34
DRAFT-DO NOT QUOTE OR CITE

-------
                  TABLE 2-1. COMPARISON OF AMBIENT PARTICLES,
        FINE MODE (Nuclei Mode Plus Accumulation Mode) AND COARSE MODE
                                       Fine
                                                                               Coarse
                        Nuclei
                           Accumulation
  Formed from:


  Formed by:
          Combustion, high-temperature
       processes, and atmospheric reactions
  Composed of:
  Solubility:
  Sources:
 Atmospheric
 half-life:

 Removal
 Processes:
 Nucleation
 Condensation
 Coagulation
Sulfates
Elemental carbon
Metal compounds
Organic compounds
with very low,
saturation vapor
pressure at ambient
temperature
Probably less
soluble than
accumulation mode

Combustion
Atmospheric
transformation of
SO2 and some
organic compounds
High temperature
processes
Minutes to hours
Grows into
accumulation mode
 Travel distance:  <1 to 10s of km
 Condensation
 Coagulation
 Evaporation of fog and
 cloud droplets in which
 gases have dissolved and
 reacted

 Sulfate, SO;;
 Nitrate, NOj
 Ammonium, NHj
 Hydrogen ion, H*
 Elemental carbon,
 Large variety of organic
 compounds
 Metals: compounds of Pb,
 Cd, V, Ni, Cu, Zn, Mn, Fe,
 etc.
 Particle-bound water

 Largely soluble,
 hygroscopic, and
 deliquescent

 Combustion of coal, oil,
 gasoline, diesel fuel, wood
 Atmospheric transformation
 products of NOX, SO2, and
 organic compounds,
 including biogenic organic
 species (e.g., terpenes)
 High-temperature
 processes, smelters, steel
 mills, etc.

 Days to weeks
Forms cloud droplets and
rains out
Dry deposition

100s to 1000s of km
 Break-up of large solids/droplets


 Mechanical disruption (crushing,
 grinding, abrasion of surfaces)
 Evaporation of sprays
 Suspension of dusts
 Reactions of gases in or on particles


 Suspended soil or street dust
 Fly ash from uncontrolled combustion
 of coal, oil, and wood
 Nitrates/chlorides from HNO3/HC1
 Oxides of crustal elements
 (Si,Al,Ti,Fe)
 CaCO3, NaCl, sea salt
 Pollen, mold, fungal spores
 Plant and animal fragments
 Tire, brake pad, and road wear debris


 Largely insoluble and nonhygroscopic
                                                                 Resuspension of industrial dust and
                                                                 soil tracked onto roads and streets
                                                                 Suspension from disturbed soil (e.g.,
                                                                 farming, mining, unpaved roads)
                                                                 Construction and demolition
                                                                 Uncontrolled coal and oil combustion
                                                                 Ocean spray
                                                                 Biological sources
                                               Minutes to hours
Dry deposition by fallout
Scavenging by falling rain drops
                                                                <1 to 10s of km
                                                                (100s to 1000s in dust storms')
 Source: Adapted from Wilson and Suh (1997).
March 2001
                              2-35
                       DRAFT-DO NOT QUOTE OR CITE

-------
 1     equilibration. However, numerous uncertainties remain as to the relationship between the mass
 2     and composition of material remaining on the filter, as measured by the FRMs, and the mass and
 3     composition of material that existed in the atmosphere as suspended PM.  The goal of a PM
 4     indicator might be to measure accurately what exists as a particle in the atmosphere. However,
 5     this is not currently possible, in part because of the difficulty of creating a reference standard for
 6     particles suspended in the atmosphere. As a result, EPA defines accuracy for PM measurements
 7     in terms of agreement of a candidate sampler with a reference sampler. Therefore,
 8     intercomparisons of samplers become very important in determining how well various samplers
 9     agree and how various design choices influence what is actually measured.
10          There are five  general areas where choices must be made in designing an aerosol indicator,
11     These include (1) treatment of semivolatile components; (2) selection of an upper cut point;
12     (3) separation of fine-mode and coarse-mode PM; (4) treatment of pressure, temperature, and
13     relative humidity; and (5) assessment of the reliability of the measurement technique.  In many
14     cases, choices have been made without adequate knowledge or understanding of the
15     consequences. As a result, measurement methods developed by different organizations may give
16     different results when sampling the same atmosphere, even though the techniques appear to be
                                                  \
17     identical.
18
19     2.2.1.1  Treatment of Semivolatile Components of Particulate Matter
20          Current filtration-based mass measurements lead to significant evaporative losses, during
21     and possibly after collection,  of a variety of semivolatile components (i.e., species that exist in
22     the atmosphere in dynamic equilibrium between the condensed phase and gas phase).  Important
23     examples include ammonium nitrate, semivolatile organic compounds, and particle-bound water.
24     This problem is illustrated in Figure 2-10.
25          Possible approaches that have been used to address the problem of potentially lost
26     semivolatile components include those that follow, which will be discussed in more detail in
27     subsequent sections.
28     1. Collect/measure  all components present in the atmosphere in the condensed phase except
29        particle-bound water.  (Examples: Brigham Young absorptive sampler, Harvard pressure drop
30        monitor. Both require preconcentration of the accumulation mode and reduction of ambient
31        humidity.)
        March 2001
2-36
DRAFT-DO NOT QUOTE OR CITE

-------
                Should be
                 retained
                                                           Particle-bound water
                                                           should be removed
                                   (NH4)XS04
                                   X= 0 to 2
                                Mineral/Metal
                0.1                                       1.0         2.5
                                  Aerodynamic Diameter, |jm
                   Semivolatile components subject to evaporation during or after sampling
      Figure 2-10. Schematic showing major nonvolatile and semivolatile components of PM2 5.
                  Semivolatile components are subject to partial to complete loss during
                  equilibration or heating. The optimal technique would be to remove all
                  particle-bound water but no ammonium nitrate or semivolatile organic PM.
1
2
3
4
5
6
7
2.  Stabilize PM at a specified temperature high enough to remove all particle-bound water. This
   results in loss of most of the semivolatile PM. (Examples: TEOM operated at 50 °C beta
   gauge with heated inlet.)
3.  Equilibrate collected material at fixed, near-room temperature and moderate relative humidity
   (RH) to remove most particle-bound water. Accept the loss of an unknown but possibly
   significant fraction of semivolatile PM.  (Example: U.S. Federal Reference Method and most
   filter-weighing techniques.) (Note: Equilibration originally was designed to remove adsorbed
      March 2001
                                       2-37
DRAFT-DO NOT QUOTE OR CITE

-------
 1        water vapor from glass fiber filters in order to maintain a stable filter weight. The designated
 2        RH (40%) was a compromise. If the RH is too low, electrostatic charging becomes a
 3        problem. The equilibration process does help provide a stable and reproducible mass. It also
 4        reduces the particle-bound water. However, it may not remove all particle-bound water.
 5          The amount of semivolatile material lost is  dependent on the concentration and
 6     composition of the semivolatile components and  is, therefore, also dependent on season and
 7     location.  The amount of semivolatile material lost has been found to be significant in air sheds
 8     with high nitrate, wood smoke, or secondary organic aerosols.
 9
10     2.2.1.2   Upper Cut Point
11          A technique must be used that gives an upper cut-point, and its standard deviation, that is
12     independent of wind speed and direction (the  classical high volume sampler head was
13     unsatisfactory because of radial asymmetry).  A separation that simulates the removal of particles
14     by the upper part of the human respiratory system would appear to be a good choice (i.e.,
15     measure what gets into the lungs). The ACGIH-ISO-CEN penetration curve for thoracic
16     particles, with a 50% cut-point at 10 ^m aerodynamic diameter (AD), would be an appropriate
17     choice. (Thoracic particles are able to pass the larynx and penetrate into the bronchial and
18     alveolar regions of the lung.)  Some countries, however, use PM10 to refer not to samplers with a
19     50% cut at 10 jum AD but samplers with 100% rejection of all particles greater than 10 //m AD.
20     Such samplers miss too much of the thoracic  PM. The U.S. PM10 separation curve, while sharper
21     than the thoracic curve, is probably satisfactory both for regulatory and health risk monitoring.
22     It has the advantage of reducing the problem of maintaining the finite collection efficiency
23     specified by the thoracic penetration curve for particles larger than 10 yum AD.  (See Figure 2-6
24     and Section 2.1.2.2.)
25
26      2.2.1.3  Cut Point for Separation of Fine-Mode and Coarse-Mode Particulate Matter
27           Fine-mode and coarse-mode particles differ not only in size and morphology (e.g., smooth
28      droplets versus rough solid particles) but also in  formation mechanisms; sources; and chemical,
29      physical, and biological properties. They also differ in terms of dosimetry (deposition in the
30      respiratory system), toxicity, and health effects as observed by epidemiologic studies. The many
31      reasons for wanting to collect fine and coarse particles separately and considerations as to the
        March 2001
2-38
DRAFT-DO NOT QUOTE OR CITE

-------
 1      appropriate cutpoint for separating fine and coarse particles were discussed in Chapter 3 of the
 2      1996 PM AQCD (U.S. Environmental Protection Agency, 1996). A review of atmospheric
 3      particle-size-distribution data did not provide a clear or obvious rationale for selection of an
 4      appropriate cutpoint. Depending on conditions, a significant amount of either fine- or
 5      coarse-mode material may be found in the intermodal region between 1 and 3 /zm. However, the
 6      analysis of the existing data did demonstrate the important role of relative humidity in
 7      influencing the size of particles in the accumulation mode.
 8           At high relative humidity, such as that found in fog and clouds, hygroscopic fine-mode
 9      particles will increase in size due to accumulation of particle-bound water.  Under such
10      conditions, some, originally submicrometer, fine-mode PM may be found with an AD above
11      1 /^m. At very low relative humidity, coarse-mode particles may be fragmented into smaller
12      sizes, and small amounts of coarse-mode PM may be found with an AD below 1 //m (Lundgren
13      et al., 1984).  Thus, a PM25 sample will contain most of the fine-mode material, except during
14      periods of RH near 100 %. However, especially under conditions of low RH, it may contain 5 to
15      20% of the coarse-mode material below 10 //m hi diameter. A cut point of 1.0 /j.m would reduce
16      the misclassification of coarse-mode material as fine, but under high RH conditions could result
17      in some fine-mode material being misclassified as coarse. A reduction in RH, either
18      intentionally or inadvertently, will reduce the size of the fine mode. A sufficient reduction hi RH
19      will yield a dry fine-particle mode with very little material above 1.0 /j.m. Studies of the changes
20      in particle size with changes hi relative humidity suggest that only a small fraction of
21      accumulation mode particles will be above 1 //m in diameter at RH below 60% but a substantial
22      fraction will grow above 1 /urn  for RH above 80% (Hitzenberger et al., 1997; McMurry and
23      Stolzenburg, 1989; U.S. Environmental Protection Agency, 1996).
24           It is desirable to separate fine-mode PM and coarse-mode PM as cleanly as possible in
25      order to properly allocate health effects to either fine-mode PM or coarse-mode PM and to
26      correctly determine sources by  factor analysis and/or chemical mass balance. For example,
27      sulfate in the fine-mode is associated with hydrogen or ammonium ions; sulfate  in the coarse
28      mode is associated with basic metal ions. The sources are different and the health effects may be
29      different. Transition metals hi the coarse mode are likely to be associated with soil and tend to
30      be less soluble than transition metals in the fine mode, which may be found hi fresh combustion
31      particles.
       March 2001
2-39
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
     In areas where winds cause high concentrations of wind blown soil, the current practice of
separating fine-mode and coarse-mode particles at 2.5 fj.ro. AD may not provide the best
separation for exposure or epidemiologic studies. An example, taken from data collected during
the August 1996 dust storm in Spokane, WA, is shown in Figure 2-11.  Note that the PM10 scale
is 10 times that of the other size fractions. PM15 although high in the morning, goes down as the
wind increases and PM10, PM25, and PM(2 5.,) go up.  During the peak of the dust storm, PM(2 -5_,)
was 88% of PM2.5. For the 24-h period, PM(2 5.^ was 54% of PM2 5. However, PM, was not
biased by the intrusion of coarse-mode particles.
                                                                                    600
                                                                                         Q.
                i   r  i   'i  '   i   '   i   '   r  •   i   '   i  '   r  '   i   •  i   '   r
              12am 2am  4am  6am  Sam 10am 12pm 2pm  4pm  6pm  8pm  10pm
                                   Local Time, August 30,1996
       Figure 2-11.  Participate matter concentrations in Spokane, WA, during the August 30,
                    1996 dust storm.
       Source: Claiborn et al. (2000).
 1          Under conditions of high relative humidity, a cut point near 1 /^m AD may reject some
 2     fine-mode material. Under these circumstances, a monitor using a 1.0 yum AD cut point can
       March 2001
                                        2-40
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
27
28
29
30
 achieve better modal separation if the air stream is dehumidified to some fixed humidity that
 would remove all or most particle-bound water without evaporating semivolatile components.
 New techniques have been developed for both integrated and continuous measurement of fine
 particulate matter minus particle-bound water, but including semivolatile nitrate and organic
 compounds.  These techniques (see Section 2.2.5) require reduction of RH prior to collection.
 With such techniques, PM, would be a good indicator of fine-mode particles.

 2.2.1.4  Treatment of Pressure, Temperature, and Relative Humidity
      There are a variety of techniques for defining (or ignoring) the pressure, temperature, and
 relative humidity during or after sampling.
 Temperature and Pressure
      (a) Sample volume based on mass or volumetric flow corrected to standard temperature
         and pressure (273 K and 1 atm.).
      (b) Sample volume based on volumetric flow at ambient conditions of temperature and
         pressure.
 Temperature During Collection
      (a) Heat enough to remove all particle-bound water (i.e., TEOM at 50 °C).
      (b) Heat several degrees to prevent condensation of water in sampling system.
      (c) Try to maintain sampler near ambient temperature.
      (d) Maintain sampler at constant temperature inside heated/air conditioned shelter.
 Temperature After Collection
     (a) No control
     (b) Constant Temperature (room temperature)
     (c) Store at cool temperature (4 °C)
Relative Humidity
     Changes in relative humidity cause changes in particle size of hygroscopic or deliquescent
     particles. Changing relative humidity by adding or removing water vapor affects
     measurements of the following items.
     (a) Particle number, particle surface area and particle size distribution
     (b) Amount of overlap of fine-mode and coarse-mode particles
       March 2001
                                                2-41
                                                    DRAFT-DO NOT QUOTE OR CITE

-------
 1           Changing relative humidity by intentional or inadvertent changes in temperature affects
 2           above measurements plus the following.
 3           (c) Amount of loss of ammonium nitrate and semivolatile organic compounds.
 4           Studies of relationships between personal/indoor/outdoor measurements present special
 5      problems.  Indoor environments are typically dryer than outdoors and may be warmer or, if
 6      air-conditioned, cooler. These differences may change particle size and the amount of
 7      volatilization of semivolatile components. Such changes between indoors and outdoors will
 8      complicate the comparison of indoor to outdoor concentrations, the modeling of personal
 9      exposure to all particles, and exposure apportionment by the disaggregation of personal exposure
10     into exposure to particles of ambient origin and exposure to particles of indoor origin.
11
12     2.2.1.5  No Way To Determine Accuracy for Ambient Participate Matter Mass
13               Measurement
14          Precision is typically determined by comparison of collocated samplers or through replicate
15     analyses, while accuracy is determined through the use of traceable calibration standards.
16     Unfortunately, no standard reference calibration material or procedure has been developed for
17     suspended, atmospheric PM.  It is possible to determine the accuracy of certain components of
18     the PM measurement system (e.g., flow control, inlet aspiration, PM2 5 cut, weighing, etc.). The
19     absolute accuracy for collecting a test aerosol can also be determined by isokinetic sampling in a
20     wind tunnel. However, it is not currently feasible to provide a simulated atmospheric aerosol
21     with naturally occurring semivolatile components. It is particularly challenging to develop an
22     atmospheric aerosol calibration standard suitable for testing samplers in the field. Therefore, it is
23     not possible at the present time to establish the absolute accuracy of a PM monitoring technique.
24     Intercomparison studies, to establish the precision of identical monitors and the extent of
25     agreement between different types of monitors, are essential for establishing the reliability of PM
26     measurements. Intercomparison studies have contributed greatly to our understanding of the
27     problems in PM measurement.  Such studies will be discussed as they apply to specific
28     measurement problems, monitoring instruments, or analytical techniques.
29           Some measurement errors of concern  in PM10 sampling, including those that arise due to
30     uncertainty tolerances in cutpoint, particle bounce and reentrainment, impactor surface
31      overloading, and losses to sampler internal surfaces, were discussed in detail in the 1996 PM
        March 2001
2-42
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
27
28
29
30
31
 AQCD (U.S. Environmental Protection Agency, 1996). Other measurement errors of concern in
 PM2 5 sampling arise because of our inability to assess accuracy in an absolute sense due to a lack
 of an atmospheric aerosol calibration standard, because of the inclusion in PM25 of a small
 amount of coarse particles, and because of problems associated with the definition of PM2 5 as
 what remains after collection on a filter and equilibration rather than the mass of particles as they
 exist in the air. However, it is possible to measure PM indicators with high precision.
      Because of the difficulties associated with determining the accuracy of PM measurements,
 EPA has sought to make FRM measurements equivalent by specifying operating conditions and,
 in the case of PM2 5 samplers, by specifying details of the sampler design. Thus, both the PM,0 as
 well as the PM2 5 standards are defined with consistency of measurement technique, rather than
 accuracy of the true mass concentration measurement, in mind (McMurry, 2000). It is
 acknowledged in the Federal Register (1997) that, "because the size and volatility of the particles
 making up ambient particulate matter vary over a wide range and the mass concentration of
 particles varies with particle size, it is difficult to define the accuracy of PM2 5 measurements in
 an absolute sense...."  Thus, accuracy is defined as the degree of agreement between a field PM2 5
 sampler and a collocated PM2 5 reference method audit sampler (McMurry, 2000). The Federal
 Reference Method (FRM) for PM25 is discussed below in Section 2.2.3.3. As mentioned earlier,
 volatilization losses, during sampling or post-sampling handling, of some organics as well as
 ammonium nitrate can lead to significant underestimation of the true fine particulate mass
 concentration in some locations. Sources of error in the measurement of true PM2 5 mass also
 arise because of adsorption or desorption of semivolatile vapors onto or from collected PM, filter
 media, or other sampler surfaces; neutralization of acid or basic vapors on either filter media or
 collected PM; and artifacts associated with particle-bound water.
     During the past 25 years, there have been advancements in the generation and classification
 of monodisperse aerosols, as well as in  the development of electron microscopy and imaging
 analysis, that have contributed to the advancement in aerosol calibration (Chen, 1993). Still, one
 of the limitations in PM sampling and analysis remains the lack of primary calibration standards
 for evaluating analytical methods and for intercomparing laboratories. Klouda et al. (1996)
examined the possibility of resuspending the NIST Standard Reference Material 1649 (Urban
Dust) in air for collection on up to 320 filters simultaneously, using SRI, International's dust
generation and collection system.  However, the fine component is not resuspended and the
       March 2001
                                         2-43
DRAFT-DO NOT QUOTE OR CITE

-------
 1      semivolatile component has evaporated so this material is not a suitable standard for suspended
 2     PM. Little additional workin this area has been reported.
 3          Methods validation was discussed in the 1996 PM AQCD (U.S. Environmental Protection
 4     Agency, 1996), and the usefulness of intercomparisons and "internal redundancy" was
 5     emphasized. For example, a number of internal consistency checks are applied to the IMPROVE
 6     network (Malm et al., 1994).  These include mass balances, sulfur measurements by both proton •
 7     induced X-ray emission (PIXE) and ion chromatography (1C), and comparison of organic matter
 8     by combustion and by proton elastic scattering analysis (PES A) analysis of hydrogen.  Mass
 9     balances compare the gravimetrically determined mass with the mass calculated from the sum of
10     the major chemical components (i.e., crustal elements plus associated oxygen, organic carbon,
11     elemental carbon, sulfate, nitrate, ammonium, and hydrogen ions).  Mass balances are useful
12     validation techniques; however, they do not check for, or account for, artifacts associated with
13     the absorption of gases during sampling, or the loss of semi-volatile material during sampling.
14     The mass balance check may appear reasonable even if such artifacts are present because only the
15     material collected on the filter is included in the balance.
16
17     2.2.2 Why Measure Particles
18     2.2.2.1 Attainment of a Standard
19           A critical need for particle measurements is to determine if a location is in compliance with
20     an existing standard and to determine if trends show improvements in air quality. For this
21     purpose, precision of the measurement by the variety of indicators  in use is the most important
22     consideration. Therefore, intercomparisons of various potential indicators, under a variety of
23     atmospheric and air quality conditions, are essential.
24
25     2.2.2.2  Implementation of a Standard
26           In order to reduce pollution to attain a standard, local agencies and national research
27     organizations need measurements to identify source categories and to develop and validate air
28     quality models. For these purposes, PM parameters other than mass, such as chemical
29     composition and size distribution, must also be measured.  Moreover, measurements are needed
30     with shorter tune resolution in order to match changes in pollution with diurnal changes in the
31     boundary layer.
       March 2001                              2-44        DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 2.2.2.3  Determination of Health Effects
      PM measurements are needed to determine exposure for use in epidemiological studies, to
 assess exposure for risk assessment, and to determine components of PM to guide planning and
 interpretation of toxicologic experiments.  For these purposes, size and chemical composition
 may be needed. For exposure assessment, different measurement time intervals may be needed.
 For epidemiologic studies of acute (short-term PM exposures), 1-h of continuous measurements
 may be needed as well as 24-h measurements. However, for epidemiologic studies of chronic
 PM exposures, measurements that integrate over longer intervals (e.g., a week to a month) may
 be more cost effective.  For dosimetric studies and modeling, information will be needed on the
 particle size distribution and on the behavior of particles as the relative humidity and temperature
 are increased to those in the respiratory system.

 2.2.2.4 Determination of Ecological Effects
      Measurements of particles, and of the chemical components of particulate matter in rain,
 fog and dew, are needed to understand the contributions of PM to spiling of surfaces and damage
 to materials and to understand the wet and dry deposition of acidity and toxic substances to
 surface water, soil, and plants. Some differentiation into particle size is needed to determine dry
 deposition. Information on chemical composition is also needed to understand materials damage
 and ecological damage.

 2.2.2.5  Determination of Radiative Effects
     Particles reduce visibility by scattering and absorbing light. They also have a direct effect
 on the climate by scattering visible and ultraviolet light back into space and, indirectly, as cloud
 condensation nuclei, by changing the albedo and stability of clouds.  For understanding these
 effects, information is needed on refractive index (including ratio of scattering to absorption),
 size distribution, and change in particle size with change in relative humidity.

2.2.2.6  Particulate Matter Components/Parameters That Need To Be Measured
     PM parameters and components of PM that need to be measured for various purposes are
summarized in Table 2-2.
       March 2001
                                         2-45
DRAFT-DO NOT QUOTE OR CITE

-------
             TABLE 2-2. PARTICULATE MATTER COMPONENTS/PARAMETERS
           OF INTEREST FOR HEALTH, ECOLOGICAL, OR RADIATIVE EFFECTS;
                   FOR SOURCE CATEGORY APPORTIONMENT STUDIES;
      	OR FOR AIR QUALITY VALIDATION STUDIES
       Particle number
       Particle surface area
       Particle size distribution
       PM mass (fine-mode [PM, 0] and coarse-mode [PM10.,] mass as well as PM2 5 and PM10);
       nonvolatile mass, Federal Reference mass, and mass including semivolatile components such
       as ammonium nitrate and semivolatile organic compounds, but not particle-bound water
       Ions (sulfate, nitrate, and ammonium)
       Strong acidity (H+)
       Elemental carbon
       Organic carbon (total, nonvolatile, and semivolatile; functional groups and individual species)
       Transition metals (water soluble, bioavailable, oxidant generation)
       Specific toxic elements and organic compounds
       Crustal elements
       Bioaerosols
       Particle refractive index (real and imaginary)
       Particle density
       Particle size change with changes in relative humidity	,_=__
1     2.2.3 Problems Associated with Semivolatile Particulate Matter
2          It is becoming increasingly apparent that the semivolatile component of PM may impact
3     significantly the quality of the measurement and can lead to both positive and negative sampling
4     artifacts.  Losses of semivolatile species, like ammonium nitrate and many organic species, may
5     occur during sampling because of changes in temperature, relative humidity, or composition of
6     the aerosol or because of the pressure drop across the filter (McMurry, 1999). Gas phase organic
7     species, both volatile and semivolatile, may adsorb onto or react with filter media or collected
8     PM, leading to a positive sampling artifact. Quartz fiber filters have a large specific surface area
      March 2001                              2-46        DRAFT-DO NOT QUOTE OR CITE

-------
  1      on which adsorption of gases can occur. A number of other types of filters (e.g., stretched Teflon
  2      membrane filters) have much smaller exposed surface areas (Turpin et al., 1994) and appear to be
  3      subject to less adsorption (Kirchstetter et al., 2000; Turpin et al., 1994).  Tsai and Huang (1995)
  4      observed positive sulfate and nitrate artifacts on high-volume PM10 quartz filters and attributed
  5      the artifacts to interactions between acidic gases SO2, HONO, and HNO3 and both the filter
  6      media (either glass fiber or quartz) and the coarse particles collected on the filter. Volatilization
  7      losses also have been reported to occur during sample transport and storage (Chow, 1995).
  8      Evaporative losses of particulate nitrates have been investigated in laboratory and field
  9      experiments (e.g., Wang and John, 1988),  and in theoretical studies (Zhang and McMurry, 1992).
 10      It has been known for some time that volatilization losses of SVOC can be significant (e.g.,
 11      Eatough et al., 1993).
 12           The theory describing phase equilibria of SVOC continues to be developed. Liang et al.
 13      (1997), Jang et al. (1997), and Strommen and Kamens (1997) have modeled the gas/particle
 14      partitioning of SVOC on inorganic, organic, and ambient smog aerosols.
 15           Adsorption of organic vapors onto quartz filters is recognized as a source of positive
 16      sampling error.  This artifact has been examined in experiments in which two quartz fiber filters
 17      were deployed in series.  The second quartz filter may indicate gaseous VOC adsorbed on both
 18      filters (positive artifact) or SVOC evaporated from particles on the first filter and subsequently
 19      adsorbed on the second filter (negative artifact), or a combination of both effects.  Unless the
20      individual compounds are identified, the investigator does not know what to do with the loading
21      value on the second filter (i.e., to add or subtract from the first filter loading value).
22           The developing state of the art in which diffusion denuder technology is being applied to
23      SVOC sampling (e.g., Eatough et al., 1993; Gundel et al., 1995), as well as for sampling of gas
24      and particulate phase organic acids (Lawrence and Koutrakis, 1996a,b), holds promise for
25      improving the understanding of SVOC sampling artifacts. In a denuder-based system, gas-phase
26      organics are removed by diffusion to an adsorbent surface (e.g., activated carbon, special
27      polymer resins, etc.).  Particles then are collected on a filter downstream of the denuder and the
28      remaining organic vapors (i.e., from denuder breakthrough and volatile losses from the collected
29      particles) are collected in an adsorbent downstream of the filter (e.g., charcoal or carbon-
30      impregnated filters, polyurethane foam, or polystyrene-divinylbenzene resin [XAD]).
        March 2001
2-47
DRAFT-DO NOT QUOTE OR CITE

-------
 1          Finally, Eatough et al. (1999a) have reported on a batch sampler (the Particle Concentrator-
 2     Brigham Young University Organic Sampling System [PC-BOSS]) and a continuous sampler
 3     (Real-Time Air Monitoring System or RAMS), which attempt to correct simultaneously for
 4     volatilization losses of both nitrate and SVOC. These samplers are discussed in more detail in
 5     Section 2.2.3.2.
 6
 7     2.2.3.1  Particulate Nitrates
 8          It is well known that volatilization losses of particulate nitrates (e.g., Zhang and McMurry
 9     [1992]; see also Hering and Cass [1999] and references therein) occur during sampling on Teflon
10     filters. The impact on the accuracy of atmospheric particulate measurements from these
11     volatilization losses is more significant for PM2 5 than for PMI0. The FRM for PM2 5 suffers loss
12     of nitrates, similar to the losses experienced with other simple filter collection systems.
13     Sampling artifacts resulting from the loss of particulate nitrates represents a significant problem
14     in areas such as southern California that experience high amounts of nitrates. Hering and Cass
15     (1999) examined the errors in PM2 5 mass measurements because of volatilization of particulate
16     nitrate by looking at data from two field measurement campaigns conducted in southern
17     California: (1) the Southern California Air Quality Study (SCAQS, Lawson, 1990) and (2) the
18     1986 CalTech study (Solomon et al., 1992). In both these studies, side-by-side sampling of PM2 5
19     was conducted. One sampler collected particles directly onto a Teflon filter. The second
20     sampler consisted of a denuder to remove gaseous nitric acid followed by a nylon filter that
21     absorbs the HNO3 which evaporates from ammonium nitrate.  In both studies, the denuder
22     consisted of MgO-coated glass tubes (Appel et al., 1981). Fine particulate nitrate collected on
23     the Teflon filter was compared to fine particulate nitrate collected on the denuded nylon filter.
24     In both studies, the PM2 5 mass lost because of volatilization of ammonium nitrate represented a
25     significant fraction of the total PM25 mass.  The fraction of mass lost was higher during summer
26     than during fall (17% versus  9% during the SCAQS study and 21% versus 13% during the
27     CalTech study) (Figure 2-12). In regard to percentage loss of nitrate, as opposed  to percentage
28     loss of mass discussed above, Hering and Cass (1999) found that nitrate remaining on the Teflon
29     filter samples was on average 28% lower than that on the denuded nylon filters.
30          Hering and Cass (1999) also analyzed these data by extending the evaporative model
31     developed by Zhang and McMurry (1987).  The extended model utilized by Hering and Cass
        March 2001
2-48
DRAFT-DO NOT QUOTE OR CITE

-------
         0)
         I
            80%
            60%
            40%-
                                 SCAQS Data Set
                                 o Summer Measurements
                                 • Fall Measurements
            20% -
                     50     100    150    200
                     PM2.5 Gravimetric Mass (M9/m3)
                                        250
                                                                         Caltech Data Set
                                                                         o April - September
                                                                         • October-March
                                                            50     100    150    200     250
                                                             PM2S Gravimetric Mass (pg/m3)
        Figure 2-12. Amount of ammonium nitrate volatilized from Teflon filters, expressed as a
                     percentage of the measured PM25 mass, for the SCAQS and CalTech studies,
                     for spring and fall sampling periods.
        Source: Herring and Cass (1999).
  1
  2
  3
  4
  5
  6
  1
  8
  9
10
11
12
13
14
15
 (1999) takes into account dissociation of collected particulate ammonium nitrate on Teflon filters
 into nitric acid and ammonia via three mechanisms: (1) scrubbing of nitric acid and ammonia in
 the sampler inlet (John et al. [1988] showed that clean PMIO inlet surfaces serve as an effective
 denuder for nitric acid), (2) heating of the filter substrate above ambient temperature by
 sampling, and (3) pressure drop across the Teflon filter.  For the sampling systems modeled, the
 flow-induced pressure drop was measured to be less than 0.02 atm, and the corresponding change
 in vapor pressure was 2%, so losses driven by pressure drop were not considered to be significant
 in this work. Losses from Teflon filters were found to be higher during the summer compared to
 the winter, higher during the day compared to night, and reasonably consistent with modeled
predictions.
     Finally, during the SCAQS study, particulate samples also were collected using a Berner
impactor and greased Tedlar substrates, in size ranges from 0.05 to 10 /u.m in aerodynamic
diameter.  The Berner impactor PM2 5 nitrate values were much closer to those from the denuded
nylon filter than those from the Teflon filter, with the impactor nitrate being approximately
2% lower than the nylon filter nitrate for the  fall measurements, and approximately 7% lower
       March 2001
                                          2-49
                                                             DRAFT-DO NOT QUOTE OR CITE

-------
 1      during the summer measurements.  When the impactor collection was compared to the Teflon
 2      filter collection for a nonvolatile species (sulfate), the results were in agreement.
 3           It should be noted that during these intercomparison studies, filters or collection surfaces
 4      were removed immediately after sampling and placed into vials containing a basic extraction
 5      solution. Therefore, losses that might occur during handling, storage, and equilibration of filters
 6      or impaction surfaces were avoided.  The loss of nitrate observed from Teflon filters and
 7      impaction surfaces in this study, therefore, is a lower limit compared to losses that might occur
 8      during the normal processes involved in equilibration and weighing of filters and impaction
 9      surfaces. Brook and Dann (1999) measured particulate nitrate in Windsor and Hamilton,
10     Ontario, Canada, by three techniques: (1) a single Teflon filter in a dichotomous sampler, (2) the
11      Teflon filter in an annular denuder system (ADS), and (3) total nitrate including both the Teflon
12     filter and the nylon back-up filter from the ADS. The dicot Teflon filter averaged only 13% of
13     the total nitrate.  The Teflon filter from the ADS averaged 46% of the total nitrate. The authors
14     conclude that considerable nitrate was lost from the dicot filters during handling, which included
15     weighing and XRF measurement in a vacuum.
16           Kim et al. (1999) also examined nitrate sampling artifacts by comparing denuded and
17     undenuded quartz and nylon filters, during the PM10 Technical Enhancement Program (PTEP) in
18     the South Coast Air Basin of California.  They observed negative nitrate artifacts (losses) for
19     most measurements; .however, for a significant number of measurements they observed positive
20     nitrate artifacts.  Kim et al. (1999) pointed out that random measurement errors make it difficult
21     to measure true amounts of nitrate loss.
22           Several diffusion denuder samplers have been developed to account for the nitrate lost
23      because of volatilization from filters, many of which were discussed in the 1996 PM AQCD
24      (U.S. Environmental Protection Agency, 1996). Eatough et al. (1999a) developed a high-volume
25      diffusion denuder system in which diffusion denuder and particle concentrator techniques were
 26      combined (see Section 2.2.3.2). The particle concentrator reduces the flow through the denuder
 27      so that the denuder can be operated for weeks without a loss of collection efficiency, thus making
 28      the sampler suitable for routine field sampling.  The system was evaluated for the collection of
 29      fine particulate sulfate and nitrate in Riverside,  CA (Eatough et al., 1999b). Concentrations of
 30      PM2 5 nitrate obtained from the PC-BOSS agreed with those obtained using the Harvard-EPA
 31      Annular Denuder Sampler, HEADS (Koutrakis et al., 1988a).
        March 2001
2-50
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
26
27
28
29
30
31
      In atmospheres with high sulfate and low ammonia, the PM tends to be acidic (NH4HSO4
 or H2SO4), and nitric acid remains in the vapor phase.  In atmospheres with lower sulfate and
 higher ammonia, there may be sufficient ammonia to folly neutralize the H2SO4 and also react
 with HNO3 to form NH4NO3 particles. In the United States, therefore, loss of nitrate will be a
 bigger problem in the western United States than in the eastern United States. However, as SO2
 emissions are reduced in the eastern United States, nitrate may become a larger fraction of the
 suspended PM.

 2.2.3.2  Semivolatile Organic Compounds
      Semivolatile organic compounds (SVOC) can similarly be lost from Teflon filters because
 of volatilization, causing the PM2 5 mass to be significantly underestimated  (negative artifact).
 Like particulate nitrates, the FRM for PM25 will suffer loss of SVOC, similar to the losses
 experienced with other simple filter collection systems. When PM is collected on a quartz filter
 in a system without a denuder, the quartz filter may adsorb some gas phase  organic compounds
 (positive artifact) as well as SVOC that evaporate from collected particles.  A second quartz
 filter, placed directly after either a quartz or Teflon first filter, could also collect some gas phase
 organic compounds passing through the first filter as well as  SVOC that evaporated from
 particles collected on the first  filter. Some workers (Turpin et al., 2000) suggest subtracting the
 organic carbon mass on the quartz second filter from that on the quartz first filter to correct for
 the positive artifact. However, if some SVOC, lost by evaporation from particles collected on
 the first filter, are adsorbed on the quartz second filter, the negative artifact would be doubled
 (Eatough et al., 1994; Cui et al., 1998). Using their multichannel diffusion denuder sampling
 system (BOSS), Eatough et al. (1995) reported that, for samples collected at the South Coast Air
 Quality Management District sampling site at Azusa, CA, changes in the phase distribution of
 SVOC could result in a loss on average of 35% of the particulate organic material. Cui et al.
 (1998) found that losses of SVOC from particles in the  Los Angeles Basin during the summer
were greater during the night (average, 62%) than during the day (average, 42%).
     The percent SVOC lost from the front filter in a filter-denuder system may be greater than
that lost in a filter-only system such as the FRM. In a filter-denuder system, the gas-phase
component of the SVOC is removed.  Because of the  absence of the gas phase, SVOC collected
on the filter might evaporate more rapidly in a filter-denuder system than in a filter-only
       March 2001
                                         2-51
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1      collection system.  To determine the fraction of SVOC lost from a Teflon filter in a filter-only
 2      system, it is necessary to compare the amount measured by a nondenuder system with that
 3      measured by a denuder system.  At present, little information is available on the volatilization
 4      losses of SVOC. However, in one study (Pang et al., 2000), the total mass on denuded and
 5      undenuded filters were compared and found to be identical within error limits (R2 = 0.816, slope
 6      = 0.961 ± 0.027 for total mass compared to R2 = 0.940, slope = 0.986 ± 0.020 for sulfate). This
 7      suggests that the major cause of loss of SVOC is the pressure drop across the filter.
 8            In addition to their contribution to suspended PM mass, SVOC are also of interest because
 9      of their possible health effects.  SVOC include products of incomplete combustion such as
10     polycyclic aromatic hydrocarbons (PAHs) and polycyclic organic matter, which has been
11      identified as a hazardous air pollutant. PAHs also have been suggested as alternative particulate
12     tracers for automobile emissions, because the phase-out of organo-lead additives to gasoline
13     means that lead is no longer a good tracer for automobiles (Venkataraman et al., 1994). PAHs
14     also are emitted during biomass burning, including burning of cereal crop residues and wood
15     fuels (Jenkins et al., 1996; Roberts and Corkill, 1998).
16           Several investigators have observed that collection of particles on a filter can result in
17     underestimation of particulate organic compounds because of losses of semivolatile organic
18     material during sample collection (negative sampling artifact) (Eatough et al., 1993; Tang et al.,
19      1994; Eatough et al., 1995; Gundel et al., 1995; Finn et al., 2000). Positive sampling artifacts
20     also can occur because of the adsorption of gases onto the filter materials (e.g., Gundel et al.,
21      1995).  There appears to be a larger positive artifact because of adsorption of organic vapor onto
22     quartz fiber filters than to Teflon filters (Turpin et al., 1994; Chow et al., 1994, 1996; Eatough et
23     al., 1996; Finn et al., 2000). When samples for organic analysis are collected on quartz fiber
24      filters, the amount of adsorbed  organic vapor on the quartz filter is sometimes estimated by the
25      amount collected on a second quartz  fiber filter behind the first, or by the amount collected on a
26      quartz fiber filter placed behind a Teflon filter in a parallel sampling port (Novakov et al., 1997).
27      Many, but by no means all, investigators subtract this adsorption estimate from the front filter
28      quantity to obtain the mass of collected particulate organic (Turpin et al., 2000).
29           Kirchstetter et al. (2000) report that adsorptive properties of quartz fiber filters vary with lot
30      number, and therefore front and back-up filters should be taken from the same lot. Recent
31      literature suggests that a Teflon-quartz back-up filter appears to provide a better estimate of the
        March 2001
2-52
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 adsorption of gases on a quartz fiber front filter than does a quartz-quartz backup, and that the
 difference between these two adsorption estimates can be substantial for short durations
 (Kirchstetter et al., 2000; Turpin et al., 2000). The typically lower organic carbon loadings on
 quartz-quartz back-up filters, relative to Teflon-quartz back-up samples collected concurrently, is
 believed to occur because adsorption on the quartz front filter acts to reduce the gas-phase
 concentration downstream until gas phase (i.e., adsorbed phase equilibrium has been achieved in
 the vicinity of the front quartz filter surface). Because Teflon filters have little affinity for
 organic vapors, this equilibrium occurs almost instantaneously for Teflon filters, and the Teflon-
 quartz back-up filter is exposed to the ambient concentration of organic vapors from the
 beginning of the sampling period.  It might be expected that the quantity of organic vapor
 adsorbed on a quartz filters would depend on the organic composition and would vary by season
 and location.

 2.2.3.3  Use of Denuder Systems To Measure Semivolatile Compounds
     Phase distribution of semivolatile organic species has been the subject  of several studies
 that have employed denuder technology (see Gundel et al., 1995; Gundel and Lane, 1999) to
 directly determine the phase distributions while avoiding some of the positive and negative
 sampling artifacts associated with using back-up quartz filters, hi an ideal system with a denuder
 that is 100% efficient, the gas phase would be collected in the denuder and the particle phase
 would be the sum of the material collected on the filter and the adsorbent downstream. Denuder
 collection efficiency depends on the denuder surface area (+), the diffusivity (+) and vapor
 pressure (-) of the compound, the temperature (-) and flow rate (-) of the air stream, and the
 presence of competing species (-), including water vapor (Cui et  al., 1998; Kamens and Coe,
 1997; Lane et al., 1988). (The + and - symbols in parentheses indicate qualitatively the effect
 increasing each parameter would have on efficiency). In a system with a denuder collection
 efficiency less than 100%, the collection efficiency must be known to accurately attribute
 adsorbed organics from denuder breakthrough to the gas phase and adsorbed organics volatilized
 from collected particles to the particle phase. In calculating the overall phase distributions of
 SVOC PAH from a denuder system, the collection efficiency for each compound is needed.
     The efficiency of silicone-grease-coated denuders for the collection of polynuclear aromatic
hydrocarbons was examined by Coutant et al. (1992), who examined the effects of uncertainties
        March 2001
                                         2-53
DRAFT-DO NOT QUOTE OR CITE

-------
 1     in the diffusion coefficients, and in the collisional reaction efficiencies, on the overall phase
 2     distributions of SVOC PAH calculated using denuder technology. In their study, they used a
 3     single stage, silicone-grease-coated aluminum annular denuder, with a filter holder mounted
 4     ahead of the denuder, and an XAD trap deployed downstream of the denuder. In a series of
 5     laboratory experiments, they spiked the filter with a mixture of perdeuterated PAH, then swept
 6     the system with ultra-high purity air for several hours, and then analyzed the filter and the XAD.
 7     They found that the effects of these uncertainties, introduced by using a single compound as a
 8     surrogate  PAH (hi their case, naphthalene) for validation of the denuder collection efficiency, are
 9     less significant than normal variations because of sampling and analytical effects. Results on
10     field studies using their sampling system have not been published.
11           For  measuring participate phase organic compounds, the denuder-based sampling system
12     represents an improvement over the filter/adsorbent collection method (Turpin et al., 1993).
13     Some researchers, however, have reported that denuder coatings themselves can introduce
14     contamination (Mukerjee et al., 1997), or the adsorbed species may be difficult to remove from
15     the coating (Eatough et al., 1993).
16           In a study conducted in southern California (Eatough et al., 1995), the Brigham Young
17     University Organic Sampling System (BOSS) (Eatough et al.,  1993) was used for determining
18     POM composition, and a high-volume version (BIG BOSS) (flowrate 200 L/min) was utilized
19     for determining the particulate size distribution and the chemical composition of SVOC in fine
20     particles.  The BOSS, a multi-channel diffusion denuder sampling system, consists of two
21     separate samplers (each operating at 35 L/min).  The first sampler consists of a multi-parallel
22     plate diffusion denuder with charcoal-impregnated filter papers as the collection surfaces,
23     followed  by a two-stage quartz filter pack, followed by a two-stage charcoal-impregnated filter
24     pack. The second sampler operating in parallel with the first consists of a two-stage quartz filter
25     pack, followed by the parallel plate denuder, followed by the two-stage charcoal-impregnated
26     filter pack. The filter samples collected by the BOSS sampler were analyzed by temperature-
27     programmed volatilization analysis. Eatough et al. (1995) also operated a two-stage quartz filter
28     pack alongside the BOSS sampler. The BIG BOSS system (Tang et al.,  1994) consists of four
29     systems (each with a flowrate of 200 L/min).  Particle size cuts of 2.5, 0.8, and 0.4 /u.m are
30     achieved  by virtual impaction, and the sample subsequently flows through a denuder, then is
31     split, with the major flow (150 L/min) flowing through a quartz filter followed by an XAD-II
        March 2001
2-54
DRAFT-DO NOT QUOTE OR CITE

-------
  1      bed. The minor flow is sampled through a quartz filter backed by a charcoal-impregnated filter
  2      paper.  The samples derived from the major flow (quartz filters and XAD-II traps) were extracted
  3      with organic solvents and analyzed by gas chromatography and GC-mass spectroscopy. The
  4      organic material lost from the particles was found to represent all classes of organic compounds.
  5           Eatough et al. (1996) operated the BOSS sampler for a year at the IMPROVE site at
  6      Canyonlands National Park, UT, alongside the IMPROVE monitor and alongside a separate
  7      sampler consisting of a two-stage quartz filter pack. They found that concentrations of
  8      participate carbon determined from the quartz filter pack sampling system were low on average
  9      by 39%, and this was attributed to volatilization losses of SVOC from the quartz filters.
 10      In another study conducted with the BOSS in southern California, losses of 35% of the POM, on
 11      average, were found and attributed to losses of the SVOC during sampling (Eatough et al.,
 12      1995).                                             '
 13           Ding et al. (1998a) developed a method for the determination of total n-nitroso compounds
 14      in air samples, and used the method to examine organic compounds formed from NOX chemistry
 15      in Provo, UT (Ding et al., 1998b).  In their method, n-nitroso compounds are selectively
 16      decomposed to yield nitric oxide, which is then detected using chemiluminescence. From the
 17      samples from Provo, they found that the majority of the n-nitroso and nitrite organic compounds
 18      that were present in fine particulate matter were semivolatile organic compounds that could be
 19      evaporated from the particles during sampling. They found particulate  n-nitroso compound
20      concentrations ranging between <1 and 3  nmoles/m3, and gas-phase n-nitroso compound
21      concentrations in the same range. Particulate organic nitrite concentrations were found in the
22      range of <1 to ~5 nmoles/m3, and gas-phase concentrations as high as 10 nmoles/m3 were found.
23           The PC-BOSS system of Eatough et al. (1999a) includes a virtual impactor upstream of the
24      denuder to improve the denuder collection efficiency by removing a majority of the gases from
25      the aerosol flow (i.e., gases and particles smaller than 0.1 jj.ro. are removed with the major flow of
26      the virtual impactor and the remaining aerosol enters the denuder). Particulate OC estimates are
27      corrected for particle losses of 46 to 48% in the inlet. The denuder consists of charcoal-
28      impregnated cellulose fiber filter material, and denuder collection efficiencies of greater than
29      98% are reported for organic gases that adsorb on quartz and charcoal-impregnated filters.
30           Turpin et al. (1993) developed a sampling system based upon a diffusion separator, which
31      corrects for the loss of semivolatile organic compounds during sampling by removal of most of
       March 2001
2-55
DRAFT-DO NOT QUOTE OR CITE

-------
 1      the gas phase material from the particles in a diffusion separator sampling system.  Unlike the
 2      previously mentioned systems, wherein the particulate phase is measured directly, hi the system
 3      of Turpin et al., the gas-phase is measured directly.  In the laminar flow system, ambient,
 4      particle-laden air enters the sampler as an annular flow. Clean, particle-free air is pushed through
 5      the core inlet of the separator.  The clean air and ambient aerosol join downstream of the core
 6      inlet section, and flow parallel to each other through the diffusion zone.  Because of the much
 7      higher diffusivities for gases compared to particles, the SVOC in the ambient air diffuse to the
 8      clean, core flow. The aerosol exits the separator in the annular flow, and the core flow exiting
 9      the separator now contains a known fraction of the ambient SVOC.  Downstream of the diffusion
10      separator, the core exit flow goes into a PUF plug, where the SVOC is collected.  The adsorbed
11      gas phase on the PUF plug is extracted with supercritical fluid CO2, and analyzed by gas
12     chromatography/mass-selective detection (GC/MSD). The gas-phase SVOC is thus determined.
13     Ultimately, to determine particulate phase SVOC concentrations, the total compound
14     concentration will also be measured, and the particulate phase obtained by difference.  The
15     system was tested for the collection of PAH. The diffusional transport of gas-phase PAHs and
16     particle concentrations agreed well with theory.  Breakthrough was problematic for low
17     molecular weight PAHs (MW < 160). Detection limits ranged from 20 to 50 pg of injected mass
18     for all PAHs.
19           Gundel et al. (1995) recently developed a technique for the direct determination of phase
20     distributions of semivolatile polcyclic aromatic hydrocarbons using annular denuder technology
21     instead of the different method.  The method, called the integrated organic vapor/particle sampler
22     (IOVPS), uses a cyclone inlet with a D50 cutpoint of 2.5 /zm at a sampling rate of 10 L/min. The
23     airstream then goes through two or three sandblasted glass annular denuders that are coated with
24     ground adsorbent resin material (XAD-4 was initially examined) that traps vapor-phase organics.
25     The airstream subsequently passes through a filter, followed by a backup denuder.  The denuder
26     collection efficiency is high and compares well with predictions based on the diffusivity of the
27     compounds.  The denuder can also be extracted to obtain gas-phase concentrations directly
28     (Gundel and Lane, 1999).  Particle-phase PAHs are taken to be the sum of material on the filter
29     and XAD adsorbent downstream after correction for denuder collection efficiency. The IOVPS
30     was tested for sampling semivolatile PAH in laboratory indoor air, and environmental tobacco
31     smoke (ETS).  After exposure, the denuders, filters, and sorbent traps were extracted with
        March 2001
2-56
DRAFT-DO NOT QUOTE OR CITE

-------
  1      cyclohexane (Gundel et al., 1995) and analyzed for PAHs from naphthalene to chrysene using
  2      dual-fluorescence detection (Mahanama et al., 1994).  Recoveries from both denuders and filters
  3      were approximately 70% for 30 samples.  Detection limits (lower limits of detection, defined as
  4      3 times the standard deviation of the blanks) for gas-phase SVOC PAHs ranged from 0.06 ng for
  5      anthracene to 19 ng for 2-methylnaphthalene.  The 95% confidence interval for reproduction of
  6      an internal standard concentration was 6.5% of the mean value.  Relative precision as determined
  7      either from a propagation of errors analysis, or from the 95% confidence interval from replicate
  8      analyses of standard reference material SRM 1649 (urban dust/organics) was 12% on average,
  9      and ranged from 8% for naphthalene to 22% for fluorene.  Sources of error included sampling
10      flow rate, internal standard concentration, and co-eluting peaks.  Gundel and Lane (1999)
11      reported that roughly two-thirds of particulate PAH fluoranthene, pyrene, benz[a]anthracene, and
12      chrysene were found on the postfilter denuders, so that it is likely that considerable desorption
13      from the collected particles took place.
14           Solid adsorbent-based denuder systems have been investigated by other researchers, as
15      well. Bertoni et al. (1984) described the development of a charcoal-based denuder system, for
16      the collection of organic vapors. Risse et al. (1996) developed a diffusion denuder system to
17      sample aromatic hydrocarbons.  In their system, denuder tubes with charcoal coating and
18      charcoal paper precede a filter pack for particulate collection and an adsorption tube to capture
19      particle blow-off from the filter sample. Breakthrough curves for benzene, toluene, ortho-xylene,
20      and meta-xylene were developed for 60-, 90-, and 120-cm denuder tubes. The effects of relative
21      humidity on the adsorption capacities of the denuder system were examined, and it was found
22      that the capacity of the charcoal was not impacted significantly by increases in relative humidity.
23      The feasibility of outdoor air sampling with the system was demonstrated.  Risse et al. (1996)
24      developed a diffusion denuder system for sampling aromatic hydrocarbons in which denuder
25      tubes were coated with charcoal.
26           Krieger and Kites (1992) designed a diffusion denuder system that uses capillary gas
27      chromatographic columns as the tubes for SVOC collection. The denuder was followed by a
28      filter to collect particles, which in turn was followed by a polyurethane foam (PUF) plug to
29      collect organic material volatilizing off the filter. Denuder samples were analyzed by liquid
30      solvent extraction (CH2C12) followed by GC-MS analysis.  The PUF plugs and filters were
31      extracted with supercritical fluid extraction using supercritical N2O. Using this system, an indoor
       March 2001
2-57
DRAFT-DO NOT QUOTE OR CITE

-------
 1     air sample was found to contain primarily chlorinated biphenyls, ranging from trichlorobiphenyls
 2     (vapor pressures 10'3 - lO"4 Torr at 25 °C) to octachlorobiphenyls (10'6 - 10'7 Torr), which
 3     demonstrated that the sampler collects compounds with a wide range of volatility. They also
 4     found that on-line desorption is successful in maintaining good chromatographic peak shape and
 5     resolution.  The entire method, from sample collection to the end of the chromatographic
 6     separation, took 2 h.
 7          Organic acids in both the vapor and particulate phases may be important contributors to
 8     ambient acidity, as well as representing an important fraction of organic particulate matter.
 9     Lawrence and Koutrakis (1996a,b) used a modified Harvard/EPA annular denuder system
10     (HEADS) to sample both gas and particulate phase organic acids in Philadelphia, PA, in the
11     summer of 1992.  The HEADS sampler inlet had a 2.1 -/-on cutpoint impactor (at 10 1pm),
12     followed by two denuder tubes, and finally a filter pack with a Teflon filter. The first denuder
13     tube was coated with KOH to trap gas phase organic acids.  The second denuder tube was coated
14     with citric acid to remove ammonia and thus to avoid neutralizing particle phase acids collected
15     on the filter.  The KOH-coated denuder tube was reported to collect gas phase formic and acetic
16     acids at better than 98.5% efficiency, and with precisions of 5% or better (Lawrence and
17     Koutrakis,  1994). It was noted that for future field measurements of particulate organic acids,
18     a Na2CO3-coated filter should be deployed downstream of the Teflon filter to trap organic acids
19     that may evaporate from the Teflon filter during sampling.
20
21     Role of the Collection Media
22          The role of the collection media was recently examined in a study conducted in Seattle
23     (Lewtas et al., 2001). In that study, the influence of denuder sampling methods and filter
24     collection media on the measurement of SVOC associated with PM2.5 was evaluated.  Activated
25     carbon and XAD collection media were used in diffusion denuders and impregnated back-up
26     filters in two different samplers, the VAPS and the PC-BOSS.  XAD-coated glass annular
27     denuders and charcoal-impregnated cellulose fiber (GIF) filter denuders also were used. GIF
28     filters also were compared to XAD-coated quartz filters as backup filter collection media.
29     Lewtas et al. (2001) found that the two denuder types resulted in equivalent measurement of
30     particulate organic carbon and particle mass. The carbon coated denuders in the BOSS sampler
31     were more efficient than the XAD coated denuders for the collection of more volatile carbon.
        March 2001
2-58
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Lewtas et al. (2001) concluded that the more volatile carbon that is collected in the carbon coated
BOSS denuder does not contribute substantially to the particle mass or to the SVOC measured as
OC on quartz filters. However, the more volatile carbon otherwise would be captured in carbon
impregnated filters placed behind quartz filters, so that, in the XAD denuder configuration, the
volatile carbon would result in a higher OC concentration and overestimation of the SVOC.
      Some of the recent research in denuder technology also has focused on reduction in the size
of the denuder, optimization of the residence time in the denuder, understanding the effect of
diffusion denuders on the positive quartz filter artifact, identifying changes in chemical
composition that occur during sampling, determining the effects of changes in temperature and
relative humidity, and identifying possible losses by absorption in coatings.

Reducing the Size of Denuders
      The typical denuder configuration is an annular diffusion denuder tube of significant length
(e.g., 26.5 cm for 10 L/min, Koutrakis et al., 1988a).  A more compact design based on a
honeycomb configuration was shown to significantly increase the capacity (Koutrakis et al.,
1993). However, in intercomparisons with an annular denuder/filter pack system (Koutrakis
et al., 1988a), significant losses of ammonia and nitric acid were observed for the honeycomb
configuration, and attributed to the large inlet surface area and long sample residence time of the
honeycomb design, relative to the annular denuder system. Sioutas et al. (1996a) subsequently
designed a modified glass honeycomb denuder/filter pack sampler (HDS) with an inlet that
minimizes vapor losses on the inlet surfaces.  The modified HDS has reduced inlet surfaces and
decreased residence time for sampled gases (NH3 and HNO3) compared to its predecessor
(Sioutas et al.,  1994). Sioutas et al. (1996b) tested various inlet materials (glass, PFA, and
PTFE) in laboratory tests and found that a PTFE Teflon coated inlet minimized loss of sampled
gases (1 to 8% losses of HNO3 observed, and -4 to 2% losses of NH3 observed). The highest
inlet losses were observed for HNO3 lost to PFA surfaces (14 to 25%). The modified HDS was
tested in laboratory and field tests and found to agree within 10% with the annular denuder
system.
       March 2001
                                         2-59
DRAFT-DO NOT QUOTE OR CITE

-------
 1     Residence Time in the Denuder
 2          The efficiency of a diffusion denuder sampler for the removal of gas phase material can be
 3     improved by increasing the residence time of the sampled aerosol in the denuder. However, the
 4     residence time can only be increased within certain limits. Because the diffusion denuder
 5     reduces the concentration of gas-phase semivolatile organic material, semivolatile organic matter
 6     present in the particles passing through the denuder will be in a thermodynamically unstable
 7     environment and will tend to outgas SVOC during passage through the denuder. The residence
 8     tune of the aerosol in the denuder, therefore, should be short enough to prevent significant loss of
 9     particulate phase SVOC to the denuder.  Various studies have suggested that the residence time
10     in the denuder should be less.than about 2 s (Gundel and Lane, 1999; Kamens and Coe, 1997;
11     Kamens et al., 1995). The residence tunes in the various denuder designs described by Gundel
12     and Lane (1999) are from 1.5 to 0.2 s. The equilibria and evaporation rates are not as well      ,
13     understood for organic components as they are for NH4NO3 (Zhang and McMurry, 1987,1992;
14     Hering and Cass,  1999).
15
16     Effect of Diffusion Denuders on the Positive Quartz Filter Artifact
17          To account for the volatilization losses of semivolatile organic compounds, Turpin et al.
18     (1994) recommended that a quartz filter be placed behind a Teflon filter in a parallel sampler.
19     Addition of a vapor trap (e.g., polyurethane foam plug) downstream of the filter also was
20     suggested as a method to collect semivolatile organic compounds. However, it was noted that
21     addition of some type of trap behind the Teflon filter collected both vapor phase organics as well
22     as "blow-off' from the Teflon filter (i.e., material vaporized from particles collected on Teflon
23     filter [Van Vaeck et al., 1984]). Kim et al. (2000) used a quartz filter behind a Teflon filter
24     recently to account for positive organic artifacts hi the South Coast Air Basin. They found that,
25     on an annual average basis, 30% of the PM2.5 organic carbon concentration resulted from positive
26     artifacts.
27          The adsorption of organic compounds by a second quartz filter has been shown to be
28     reduced, but not eliminated, in samples collected in the Los Angeles Basin, if a multi-channel
29     diffusion denuder with quartz filter material as the denuder collection surface preceded the quartz
30     filters (Fitz, 1990). This artifact can be further reduced by the use of activated charcoal as the
31     denuder surface and use of a particle concentrator to reduce the amount of gas phase organic
        March 2001
2-60
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
  compounds relative to condensed phase organic compounds (Cui et al., 1998, 1997; Eatough, <
  1999). Recent experiments (Gotham and Bidleman, 1992; Cui et al., 1998; Eatough et al., 1995,
  1996) have shown that the quartz filter artifact can result both from the collection of gas phase
  organic compounds and from the collection of semivolatile organic compounds lost from
  particles  during sampling.  Thus, results available to date suggest that both a "pbsitive" and a
  "negative" artifact can be present in the determination of particulate phase organic compounds
  using two tandem quartz filters.
      The importance of the adsorption of organic vapors on filters  or PM, relative to the
  volatilization of organic compounds from PM collected on a filter, continues to be a topic of
  active debate.  The relative importance of positive and negative artifacts will be different for
 denuded and undenuded filters; will depend on face velocity, sample loading, and the vapor
 pressures of the compounds of interest; and may vary with season and location because of
 variations in the composition of volatile and semivolatile organic material. Evidence exists for
 substantial positive and negative artifacts in the collection of organic PM.
      Undenuded quartz-quartz and Teflon-quartz back-up filters have been reported to collect
 10 to 50% of the organic mass found on quartz front filters that remove particulate matter with
 essentially 100% efficiency (Turpin; et al., 2000). Larger percentages were found for samples
 with shorter collection times and for cleaner locations. Kirchstetter  et al. (2000) and Turpin et al.
 (2000) argue that the quantity of organic material on a quartz back-up filter provides an estimate
 of the positive artifact (i.e., adsorbed organic vapors), but provides no information about the
 negative artifact (i.e., volatilized particulate organics). This argument is based on profiles of
 thermal carbon analyses (i.e., plots of evolved carbon with temperature created during Evolved
 Gas Analysis [EGA]) and the following argument. Material volatilized from the collected
 particles will not add significantly to the loading on the quartz backup filter unless the ratio of the
 mass of semivolatile vapor to the  mass of semivolatile condensed phase material is low and the
 rate of volatilization of the condensed phase semivolatile material  is  great enough to significantly
 increase the concentration of the semivolatile vapor passing through  the back-up filter (Zhang
 and McMurry, 1987).
     A net positive artifact for total particulate organic carbon was reported by Novakov et al.
(1997), whose filter-based aircraft measurements had carbon loadings that exceeded the total
aerosol mass. Novakov compared estimates of adsorption based on examination of EGA
       March 2001
                                                 2-61
                                                     DRAFT-DO NOT QUOTE OR CITE

-------
 1      thermograms and estimates of adsorption obtained from quartz-quartz back-up filters and
 2      concluded that, if anything, the quartz-quartz back-up filter underestimates the positive sampling
 3      artifact. Also, both McDow and Huntzicker (1990) and Turpin et al. (1994) observed that
 4      subtraction of the Teflon-quartz backup filter (an estimate of adsorbed organic gases)  from the
 5      quartz front filter loading removed the face velocity dependence of the particulate organic carbon
 6      concentrations obtained at face velocities of 20, 40 and 80 cm/s. Kirchstetter et al. (2000)
 7      reported that the organic carbon content of a denuded quartz filter collected in Berkeley, CA was
 8      comparable to the carbon content of a concurrently-collected undenuded quartz filter after
 9      subtraction of the matching Teflon-quartz backup (i.e., after correction for the positive artifact).
10      As a result, they concluded that volatile losses must not be important for this sample.  (Some
11      denuder breakthrough was noted in this study.)
12           Evidence of a net negative artifact is provided by Lewtas et al. (2001), who emphasized
13     that if particulate OC had been measured on a denuded quartz filter without an adsorbent
14     downstream, the negative bias would be large.  Their data showed that the sum of a denuded
15     quartz filter and absorbent downstream (average = 9.1 Mg/m3) was greater than a collocated
 16     undenuded quartz filter (average = 7.7 Mg/m3) in a PC-BOSS sampler after correction for losses
 17     (46 to 48%) in the virtual impactor inlet. A net negative artifact for total particulate OC has been
 18     reported by Eatough and colleagues in a number of studies (e.g., Cui et al., 1998; Eatough, 1999).
 19
 20      Changes in Chemical Composition During Sampling
 21           The use of sampling systems  designed to correctly identify the atmospheric gas and
 22      particulate phase distributions of collected organic material has been outlined above.
 23      An additional sampling artifact that has received little consideration in the collection of
 24      atmospheric sampling is the potential alteration of organic compounds as a result of the sampling
 25      process.  These alterations appear to result from the movement of ambient air containing
 26      oxidants and other reactive compounds past the collected particles. The addition of NO2
 27      (
-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
 mid-volume sampler (Ding et al., 1998a,b).  Similar results have been obtained for the exposure
 of a deuterated amine on a filter to NOX (Pellizzari and Krost, 1984).  When Tenax columns
 spiked with deuterated styrene and cyclohexene were exposed to ppm concentrations of ozone or
 halogens, oxygenated and halogenated compounds were shown to be formed (Pellizzari and
 Krost, 1984).  Similar oxidation of aldehydes and PAN during sampling has been observed
 (Grosjean and Parmar, 1990). Collected PAH compounds can be oxygenated or nitrated on a
 filter (Davis et al., 1987; Lindskog and Brorstrom-Lunden, 1987) but 1-nitropyrene has been
 shown to be resistant to additional nitration (Grosjean, 1983). These various chemical
 transformations of collected organic compounds can be eliminated by removal of the gas phase
 oxidants, NOX, HNO3, etc., prior to collection of the particles (Ding, 1998a,b; Grosjean and
 Parmar,  1990; Parmar and Grosjean, 1990; Pellizzari and Krost, 1984; Williams and Grosjean,
 1990). The BOSS denuder should be effective in eliminating most of the chemical
 transformation artifacts, because reactive gases are removed by the charcoal denuder that
 precedes the particle collection filter.

 Temperature and Relative Humidity Effects
      The problems of sampling artifacts associated with SVOC adsorption and evaporation are
 compounded by temperature and relative humidity effects (Pankow and Bidleman, 1991; Pankow
 et al., 1993; Falconer et al., 1995; Goss and Eisenreich, 1997). Effects of temperature on the
 partitioning of PAH were examined by Yamasaki et al. (1982), who found that the partition
 coefficient (PAHvapor/PAHpart) was inversely related to temperature and could be described using
 the Langmuir adsorption concept. The dissociation of ammonium nitrate aerosol is also a
 function  of temperature. Bunz et al. (1996) examined the dissociation and subsequent
 redistribution of NH4NO3 within a bimodal distribution, using a nine-stage low-pressure Berner
 impactor followed by analysis by ion chromatography and found a strong temperature
 dependency on the redistribution. Bunz et al. (1996) found that at lower temperatures (below
 10 °C), there was little change in the aerosol size distribution. At temperatures between 25 and
45 °C, however, the lifetime of NH4NO3 particles decreases by more than a factor of 10, and size
redistribution, as measured by average ending particle diameter, increased more for higher
temperatures than for lower temperatures.
       March 2001
                                         2-63
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1           The effects of relative humidity on the sorption of SVOC on particles are not well
 2      understood.  In a series of laboratory experiments, Goss and Eisenreich (1997) examined the
 3      sorption of both nonpolar (hydrocarbons and chlorinated hydrocarbons) and polar (ethyl ether
 4      and acetone) volatile organic compounds (VOC) onto combustion soot particles as a function of
 5      temperature and relative humidity. The soot particles used in their experiments were collected
 6      from oil furnaces and contained 60% (w/w) iron sulfate (water-soluble fraction) and 9% (w/w)
 7      elemental and organic carbon. The carbon and sulfate contents of their particulate matter are
 8      comparable to the chemical composition of ambient fine particles.  They found that, for all
 9      compounds, the sorption of VOC onto soot particles decreased with increasing relative humidity
10      over the range of 10 to 95%.  They also observed hysteresis in the relative humidity dependency,
11      with sorption coefficients at a given relative humidity higher when the RH is being increased
12     than when the RH is being decreased. The sorption coefficients were fit with an exponential
13     function to the RH so that the slope of the regression line would provide a measure of the
14     influence of relative humidity. Based on the magnitude of the slope, they concluded that the
15     RH-dependency of sorption was stronger for water-soluble organic compounds.
16           In another study by Jang and Kamens (1998), humidity effects on gas-particle partitioning
17     of SVOC were examined using outdoor environmental chambers and the experimentally
18     determined partitioning coefficients were compared to theoretical values. They examined the
19     partitioning of SVOC onto wood soot, diesel soot, and secondary aerosols and concluded that
20     "the humidity effect on partitioning was most significant for hydrophobic compounds adsorbing
21     onto polar aerosols." Although these two studies seem to be contradictory, on closer
22     examination, it is difficult to compare the two studies for several reasons. The experiments
23     conducted by Jang and Kamens (1998) were conducted in outdoor chambers at ambient
24     temperatures and humidities.  Their model was for absorptive partitioning of SVOC on
25     liquid-like atmospheric particulate matter. In contrast, the results of Goss and Eisenreich (1997)
26     were obtained from a gas chromatographic system operated at 70 °C higher than ambient
27      conditions.  The model of Goss and Eisenreich (1997) was for adsorptive partitioning of VOC on
28     solid-like atmospheric particulate matter. In the study of Jang and Kamens (1998), calculated
29      theoretical values for water activity coefficients for diesel soot were based on an inorganic  salt
30      content of 1 to 2%, whereas the combustion particles studied by Goss and Eisenreich (1997)
31      contained 60% water-soluble, inorganic salt content.  Jang and Kamens (1998) obtained their
        March 2001
2-64
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 diesel soot from their outdoor chamber, extracted it with organic solvent (mixtures of hexane and
 methylene chloride), and measured the organic fraction. The resulting salt content of 2% of the
 particulate matter studied in Jang and Kamens (1998) is enough to affect water uptake but
 presumably not to affect the sorption partitioning of organics.

 Impactor Coatings
      Impactors are used as a means to achieve a size cutpoint and as particle collection surfaces.
 Particles collected on impactors are exposed to smaller pressure drops than filter-collected
 particles, making them less susceptible to volatile losses (Zhang and McMurry, 1987). However,
 size resolution can be affected by bounce when samples are collected at low humidities (Stein
 et al., 1994). There are other sources of error inherent in some of the currently acceptable
 practices that could potentially affect particulate mass concentration measurements and that will
 surely become even more important as more emphasis is placed on chemical speciation.  Allen
 et al. (1999a) reported that the practice of greasing impaction substrates may introduce an artifact
 from the absorption of semivolatile species from the gas phase by the grease, which could
 artificially increase the amount of PAHs and other organic compounds attributed to the aerosol.
 Allen et al. (1999a)  offer several criteria to ensure that this absorption artifact is negligible,
 including selecting impaction  oils in which analytes of interest are negligibly soluble and
 ensuring that species do not have time to  equilibrate between the vapor and oil phases (criterion
 is met for nonvolatile species). They recommend using oiled impaction substrates only if the
 absorption artifact is negligible as determined from these criteria. Application of greases and
 impaction oils for preventing or reducing bounce when sampling with impactors is not suitable
 for carbon analysis because the greases contain carbon (Vasilou et al., 1999).
     Kavouras and Koutrakis  (2000) investigated the use of polyurethane foam (PUF) as a
 substrate for conventional inertial impactors. The PUF impactor substrate is not rigid like the
traditional impactor  substrate so particle bounce and reentrainment artifacts are reduced
 significantly. Kavouras and Koutrakis found that the PUF impaction substrate resulted in a much
smaller d50 at the same flow rate and Reynolds number. Moreover, the lower d50 was obtained at
a lower pressure drop than with the conventional substrate, which could lead to a reduction of
artifact vaporization of semivolatile components.
        March 2001
                                          2-65
                                                             DRAFT-DO NOT QUOTE OR CITE

-------
 1     2.2.3.4  Particle-Bound Water
 2          It is generally desirable to collect and measure ammonium nitrate and semivolatile organic
 3     compounds. However, for many measurements of suspended particle mass, it is desirable to
 4     remove the particle-bound water before determining the mass. In other situations it may be
 5     important to know how much of the suspended particle's mass or volume results from particle-
 6     bound water. The water content of PM is significant and highly variable. Moreover, there is
 7     significant hysteresis in the water adsorption-desorption pathways (Seinfeld and Pandis, 1998),
 8     further complicating the mass measurement. Figures 2-8 and 2-9 show the change in diameter of
 9     sulfate particles as a function of relative humidity.  Figure 2-8 shows the difference between
10     deliquescence and crystallization pornts.
11          Pilinis et al. (1989) calculated the water content of atmospheric particulate matter above
12     and below the deliquescent point.  They predicted that aerosol water content is strongly
13     dependent on composition, and concluded from their calculations that liquid water could
14     represent a significant mass fraction of aerosol concentration at relative humidities above 60%.
15     Since then, a few researchers have attempted to measure the water content of atmospheric
16     aerosol.  Most techniques have focused on tracking the particle mass as the relative humidity is
17     changed, and are still in the development phase.  There have been only a few demonstrations
18     using actual ambient aerosol, to date. Of interest, in particular, is the development of the
19     Tandem Differential Mobility Analyzer (TDMA) and its applications in investigations of the
20     effects of relative humidity on particle growth.
21           Lee et al. (1997) examined the influence of relative humidity on the size of atmospheric
22     aerosol using a TDMA coupled with a scanning mobility particle sizer (SMPS). They reported
23     that the use of the TDMA/SMPS system allowed for the abrupt size changes of aerosols at the
24     deliquescence point to be observed precisely. They also reported that, at relative humidities
25     between 81 and 89%, the water content of ammonium sulfate aerosols (by mass) ranged from
26     47 to 66%.
27           Andrews and Larson (1993) investigated the interactions of single aerosol particles coated
28     with an organic film with a humid environment. Using an electrodynamic balance, they
29      conducted laboratory experiments in which sodium chloride and carbon black particles were
30      coated with individual organic surfactants, intended to simulate the surface-active, organic films
31      that many atmospheric aerosol particles may exhibit, and their water sorption curves examined.
        March 2001
2-66
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
 '3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
 Their results showed that when ordinarily hydrophobic carbon black particles were coated with
 an organic surfactant, they sorbed significant amounts of water (20 to 40% of the dry mass of the
 particle).
      Liang and Chan (1997) developed a fast technique using the electrodynamic balance to
 measure the water activity of atmospheric aerosols. In their technique, the mass of a levitated
 particle is determined as the particle either evaporates or grows in response to a step change in
 the relative humidity. Their technique was demonstrated using laboratory experiments with
 NaCl, (NH4)2SO4, NaNO3, and (NH4)2SO4/NH4NO3 solutions. They concluded that one of the
 advantages of then- fast method is the ability to measure the water activity of aerosols containing
 volatile solutes such as  ammonium chloride and some organics.
      Mclnnes et al. (1996) measured aerosol mass concentration, ionic composition, and
 associated water mass of marine aerosol over the remote Pacific Ocean. The mass of
 particle-bound water was determined by taking the difference between the mass obtained at 48%
 RH and at 19% RH, assuming the aerosol particles were dry at 19% RH.  Based on a comparison
 of the remote Pacific aerosol to aerosol collected at a site at the marine/continental interface of
 the Washington coast, the amount of water associated with the aerosol was observed to be a
 function of the ammonium to sulfate ratio. They found that the amount of water associated with
 the submicrometer aerosol comprised 29% of the total aerosol mass collected at 47% RH and
 9% of the total mass at 35% RH.
      Ohta et al. (1998)  characterized the chemical composition of atmospheric fine particles
 (D50 = 2 fj.m) in Sapporo, Japan, and as part  of their measurements, determined the water
 content using the Karl Fischer method (Meyer and Boyd, 1959).  After exposing a Teflon filter, a
portion of the filter was equilibrated at 30% RH for 24 h. Then the filter piece was placed in a
water evaporator heated at 150 °C, vaporizing the particle-bound  water. The vapor evolved was
analyzed for water in an aqua-counter where it was titrated coulometrically in Karl Fischer
reagent solution (containing iodine, sulfur, and methanoi). The accuracy of the aqua-counter is
±1 mg. Using this technique, they determined that the water content of the particles ranged from
0.4 to 3.2% of the total particulate mass (at RH < 30%). This represents a smaller portion of
water compared to their previous reported values (Ohta and Okita, 1990) that were determined by
calculation at RH of 50%.
       March 2001
                                         2-67
DRAFT-DO NOT QUOTE OR CITE

-------
 1          Speer et al. (1997) developed an aerosol liquid water content analyzer (LWCA), in which
 2     aerosol samples are collected on PTFE filters, and then placed in a closed chamber in which the
 3     relative humidity is closely controlled.  The aerosol mass is monitored using a beta-gauge, first as
 4     the relative humidity is increased from low RH to high RH, and then as the RH is decreased
 5     again.  They demonstrated the LWCA on laboratory-generated aerosol and on an ambient PM2 5
 6     sample collected in Research Triangle Park, NC. The ambient aerosol sample was also analyzed
 7     for chemical constituents.  It is interesting to note that, although their laboratory-generated
 8     (NH4)2SO4 aerosol demonstrated a sharp deliquescent point, their atmospheric aerosol, which
 9     was essentially (NH4)2SO4, did not show a sharp deliquescent point.
10          Hygroscopic properties of aerosols have been studied from the viewpoint of their ability to
11     act as condensation nuclei. The hygroscopic properties of fresh and aged carbon and diesel soot
12     particles were examined by Weingartner et al. (1997) who found that fresh, submicron-size
13     particles tended to shrink with increasing relative humidity, because of a crystalline structural
14     change.  Lammel and Novakov (1995) found, in laboratory studies, that the hygroscopicity of.
15     soot particles could be increased by chemical modification, and that the cloud condensation
16     nucleation characteristics of diesel soot were similar to those of wood smoke aerosol.
17          The results of several of the above studies, in which aerosol water content as a function of
18     relative humidity was determined, are summarized in Figure 2-13. In this figure, the results of
19     Lee et al. (1997), Mclnnes et al. (1996), and Ohta et al. (1998) are included. Relative humidity
20     ranged from 9%, at which the aerosol water content was assumed to be zero (Mclnnes et al.,
21      1996), to 89%, at which the aerosol water content was determined to be 66% by mass (Lee et al.,
22      1997). Koutrakis et al. (1989) and Koutrakis and Kelly (1993) also have reported field
23     measurements of the equilibrium size of atmospheric sulfate particles as a function of relative
24     humidity and acidity.
25          The effects of relative humidity on particle growth were also examined in several studies.
26     Fang et al. (1991) investigated the effects of flow-induced relative humidity (RH) changes on
27     particle cut sizes for aqueous sulfuric acid particles in a multi-nozzle micro-orifice uniform
28      deposit impactor (MOUDI). Laboratory experiments were conducted in which polydisperse
29      sulfuric acid aerosols were generated and the RH was adjusted. The aerosols were analyzed by a
30      differential mobility analyzer.  Fang et al. (1991) observed that for inlet RH less than 80%, the
31      cut sizes for the sulfuric acid aerosols were within 5% of that for nonhygroscopic particles except
        March 2001
2-68
DRAFT-DO NOT QUOTE OR CITE

-------
ai
s
•o
cz
ZI
o
CO
i
CD
T5

CO
Q.

CD
Q
CO
CO
CO
0>
o
1±
CO
Q.
^

90-

80-


70-


60-


50-

, 40-

30-
20-


10-
0-

• Mclnnes
• Lee et al
A Ohta et £





,
t





•



•
	 ft 	 . 	 ^ 	 , 	 , 	 , 	 r- 	 _ 	
etal., 1996
., 1997
ll., 1998


















                      10      20      30      40      50     60
                                           Relative Humidity, %
                                                             70
                80
90
100
        Figure 2-13. Aerosol water content expressed as a mass percentage, as a function of
                    relative humidity.
        Source: Mclnnes et al. (1996); Lee etal. (1997); Ohta etal. (1998).
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
at the stage for which the cut size was 0.047 yum, where the cut size was 10.7% larger than the
nonhygroscopic particle cut size. They concluded that flow-induced RH changes would have
only a modest effect on MOUDI cut sizes at RH < 80%.
     Hitzenberger et al. (1997) collected atmospheric aerosol in the size range of 0.06 to 15 fj,m
in Vienna, Austria, using a nine-stage cascade impactor and measured the humidity-dependent
water uptake when the individual impaction foils were exposed to high RH. They observed
particle growth with varying growth patterns. Calculated extinction coefficients and single
scattering albedo increased with humidity.
     Hygroscopic properties, along with mixing characteristics, of submicrometer particles
sampled in Los Angeles, CA, during the summer of 1987 SCAQS study and at the Grand
Canyon, AZ, during the 1990 Navajo Generating Station Visibility Study were reported by Zhang
       March 2001
                                         2-69
DRAFT-DO NOT QUOTE OR CITE

-------
 1     et al. (1993). They used a tandem differential mobility analyzer (TDMA) (McMurry and
 2     Stolzenburg, 1989) to measure the hygroscopic properties for particles in the 0.05- to 0.5-ywm
 3     range. In their experimental technique, monodisperse particles of a known size are selected from
 4     the atmospheric aerosol with the first DMA.  Then, the relative humidity of the monodisperse
 5     aerosol is adjusted and the new particle size distribution is measured with the second DMA.
 6     At both sites, they observed that monodisperse particles could be classified according to "more"
 7     hygroscopic and "less" hygroscopic.  Aerosol behavior observed at the two sites differed
 8     markedly. Within the experimental uncertainty (±2%) the "less" hygroscopic particles sampled
 9     in Los Angeles did not grow when the RH was  increased to 90%, whereas at the Grand Canyon,
10     the growth of the "less" hygroscopic particles varied from day to day, but ranged from near 0 to
11     40% when the RH was increased to 90%. The growth of the "more" hygroscopic particles in
12     Los Angeles was dependent on particles size (15% at 0.05 //m to 60% at 0.5 y. m), whereas at the
13     Grand Canyon, the "more" hygroscopic particles grew by about 50%, with the growth not
14     varying significantly with particle size. By comparison of the TDMA data to impactor data,
15     Zhang et al. (1993) surmised that the more hygroscopic particles contained more sulfates and
16     nitrates, while the less hygroscopic particles contained more carbon and crustal components.
17           Although most of the work to date on the hygroscopic properties of atmospheric aerosols
18     has focused on the inorganic fraction, the determination of the contribution of particle-bound
19     water to atmospheric particulate mass is greatly complicated by the presence of organics. The
20     effect of RH on adsorption of semivolatile organic compounds is discussed elsewhere in this
21     chapter.  Saxena et al. (1995) observed that particulate organic compounds also can affect the
22     hygroscopic behavior of atmospheric particles.  They idealized the organic component of aerosol
23     as containing a hydrophobic fraction (high-molecular weight alkanes, alkanoic acids, alkenoic
24     acids, aldehydes, and ketones) and a hydrophilic fraction (e.g., lower molecular weight
25     carboxylic acids, dicarboxylic acids, alcohols, aldehydes, etc.) that would be likely to absorb
26     water. They then analyzed data from a tandem differential mobility analyzer in conjunction with
27     particle composition observations from an urban site (Claremont, CA) and from a nonurban site
28     (Grand Canyon) to test the hypothesis that, by adding particulate organics to  an inorganic aerosol,
29     the amount of water absorbed would be affected, and the effect could be positive or negative,
30     depending on the nature of the organics added.  They further presumed that the particulate
31     organic matter in nonurban areas would be predominantly secondary and thus hydrophilic,
        March 2001
2-70
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
 compared to the urban aerosol that was presumed to be derived from primary emissions and thus
 hydrophobic in nature. Their observations were consistent with their hypothesis, in that at the
 Grand Canyon, the presence of organics tended to increase the water uptake by aerosols, whereas
 at the Los Angeles site, the presence of organics tended to decrease water uptake.
      Nonequilibrium issues may be important for the TDMA, as well as for other methods of
 measuring water content. Although approach to equilibrium when the RH is increased is
 expected to be rapid for pure salts, it may be much slower for aerosols containing a complex mix
 of components (Saxena et al., 1995). For example, if an aerosol contains an organic film or
 coating, that film may impede the transport of water across the particle surface, thus increasing
 the time required for equilibrium (Saxena et al., 1995). Insufficient time to achieve equilibrium
 in the TDMA could result in underestimation of the water content.

 2.2.4 U. S. Environmental Protection Agency Monitoring Programs
 2.2.4.1  The Federal Reference Methods for Equilibrated Mass
      Federal Reference Methods (FRM) have been specified for measuring PM10 (Code of
 Federal Regulations, 1991a,b) and for measuring PM25 (Code of Federal Regulations, 1999a).
 The FRM for PM10 has been discussed in previous PM AQCD's and will only be briefly
 reviewed. The PM10 FRM defines performance specifications for samplers in which particles are
 inertially separated with a penetration efficiency of 50% at an aerodynamic diameter of
 10 ± 0.5 /zm. The collection efficiency increases to -100% for smaller particles and drops to
 -0% for larger particles. Particles are collected on filters and mass concentrations are
 determined gravimetrically. Instrument manufactures are required to demonstrate through field
 tests a measurement precision for 24-h samples of ± 5 Aig/m3 for PM10 concentrations below
 80 yUg/m3 and 7% above this value.
     As opposed to the performance-based FRM standard for PM10, the new FRM for PM2 5
 specifies certain details of the sampler design, as well as of sample handling and analysis,
whereas other aspects have performance specifications. The PM25 FRM sampler consists of a
PM10 inlet/impactor, a PM2 5 impactor with an oil-soaked impaction substrate to remove particles
larger than 2.5 /j,m, and a 47-//m polytetrafluoroethylene (PTFE) filter with a particle collection
efficiency greater than 99.7%.  The sample duration is 24 h, during which the sample temperature
is not to exceed ambient temperatures by more than 5 °C. After collection, samples are
       March 2001
                                         2-71
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1      equilibrated for 24-h at temperatures in the range of 20 to 23 °C (± 2 °C) and at relative
 2      humidities in the range of 30 to 40% (± 5%).  The equilibration tends to reduce particle-bound
 3      water and stabilizes the filter plus sample weight. Filters are weighed before and after sampling
 4      under the same temperature and relative humidity conditions. For sampling conducted at
 5      ambient relative humidity less than 30%, mass measurements at relative humidities down to 20%
 6      are permissible (Code of Federal Regulations, 1999a).
 7           The FRM also allows for Class I, II, and HI equivalent methods for PM2 5 (Code of Federal
 8      Regulations, 1999b). Class I equivalent methods use samplers with relatively small deviations
 9      from the sampler described in the FRM. Class II equivalent methods include "all other PM2 5
10     methods that are based upon 24-h integrated filter samplers that are subjected to subsequent
11      moisture equilibration and gravimetric mass analysis." Class IE equivalent methods include
12     non-filter-based methods such as beta attenuation, harmonic oscillating elements, or
13     nephelometry (McMurry, 2000).
14           The strength of the PM2 5 FRM is that specification of all details of the sampler design
15     ensures that measurements at all locations, if done properly, should be comparable. For example,
16     the FRM requires maintenance because the oil-soaked impaction substrate could otherwise
17     become loaded with coarse particles. Failure to do so could lead to coarse particle bounce, thus
18     artificially increasing the measured fine particle concentrations. Moreover, the specification of a
19     PM10 inlet requires the oil-soaked impaetion substrate to collect particles between 2.5 and 10 //m.
20     The implication is that, during sampling periods of high coarse PM concentrations, the impaction
21     substrate could become overloaded, leading to particle bounce. If an inlet with a cutpoint
22     diameter smaller than 10 jum were specified, coarse particle bounce could potentially be reduced,
23     and perhaps the maintenance frequency could be reduced (McMurry, 2000).
24           Since the implementation of the PM10 standard in 1987 (Federal Register, 1987)
25      considerable information has accumulated on the factors that affect the quality of the data
26      gathered from the EPA reference method for PMI0. These include inlet losses of coarse fraction
27     particles (e.g., Anand et al., 1992); biases in concentrations due to differences between  samplers
28      in large particle outpoints that are within the EPA's specified acceptable tolerances (Ranade
29      et al., 1990); and particle bounce tolerances and reentrainment leading to as much as 30% errors
30      (Wang and John, 1988). The sampling issues associated with cutpoint tolerances are predictable,
31      and the particle bounce and reentrainment problems have since been dealt with voluntarily by the
        March 2001
2-72
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 manufacturers by recommending operational procedures including oiling of impact surfaces and
 regular cleaning. The 1996 PM AQCD (U.S. Environmental Protection Agency, 1996)
 concluded that the PM10 sampling systems can be designed such that concentration measurements
 are precise to ±10%. For PM2 5, cutpoint tolerances are not expected to affect the mass
 concentration as much as for PM10, since the 2.5-^m cutpoint generally occurs near a minimum
 in the mass distribution (e.g., Figure 2-5).
      The PM25 mass concentration will be affected, on the other hand, by other sampling issues
 mentioned but not discussed extensively in the previous 1996 PM AQCD (U.S. Environmental
 Protection Agency, 1996). These issues have been discussed earlier in this chapter and include
 gas/particle and particle/substrate interactions for sulfates and nitrates (e.g., Appel et al., 1984),
 volatilization losses of nitrates (Zhang and McMurry, 1992), semivolatile organic compound
 (SVOC) artifacts (e.g., Eatough et al., 1993), and relative humidity effects (e.g., Keeler et al.,
 1988).
      Several studies now have been reported, in which the FRM was collocated with other PM25
 samplers in intercomparison studies. During the Aerosol Research and Inhalation Epidemiology
 Study (ARIES) several PM2 5 samplers were collocated at a mixed industrial-residential site near
 Atlanta, GA (Van Loy et al., 2000). These samplers included a standard PM2 5 FRM, a TEOM
 with Nafion drier, a particulate composition monitor (PCM) (Atmospheric Research and
 Analysis, Gary, NC), a high-volume carbon sampler operated by the Desert Research Institute, a
 HEADS  sampler, and a dichotomous sampler for coarse PM.  The PCM sampler has three
 channels, all of which have PMIO cyclone inlets. The first two channels both have two denuders
 preceding a 2.5-jum WINS impact and filter packs.  The first denuder is coated with sodium
 carbonate to remove acid gases, and the second is coated with citric acid to remove ammonia.
 The third channel has a carbon coated parallel-plate denuder preceding the WINS impactor.
 Measurements of 24-h mass from the FRM, PCM, and TEOM samplers, as well as reconstructed
 PM25 mass (RPM) were compared for a  12-mo period.  The slopes for the TEOM-FRM,
 PCM-FRM, and RPM-FRM correlations were 1.01, 0.94, and 0.91, respectively, whereas the
y-intercepts for each were 0.68, 0.04, and 0.98. Particulate sulfate measurements on the FRM
Teflon filter, the PCM Teflon filter, and PCM Nylon filter were nearly identical. Nitrate results
from the three filters were much less consistent, with the FRM collecting substantially less nitrate
than that collected on either the denuded nylon filter or a denuder followed by a Teflon-nylon
       March 2001
                                         2-73
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1      filter sandwich. Particulate ammonia measurements were also compared, and showed more
 2      scatter than the sulfate measurements, but less than the nitrate measurements.
 3           An intercomparison of both PM10 and PM2 5 mass measurements was conducted during the
 4      1998 Baltimore PM Study (Williams et al., 2000). PM monitors were collocated at a residential
 5      indoor, residential outdoor, and ambient monitoring site within Baltimore County, MD. PM
 6      samplers included TEOMs, PM2 5 FRMs, cyclone-based inlets manufactured by University
 7      Research Glassware (URG), and Versatile Air Pollution Samplers (VAPS). Personal
 8      Environmental Monitors (PEMs; MSP, Inc.) also were included but will not be discussed in this
 9      section.  The VAPS sampler is a dichotomous sampler operating at 33 L/min (one coarse particle
10      channel at 3 L/min, and two fine particle channels at 15 L/min, each). In the configuration
11      employed during this study, one fine particle channel was operated with a Teflon filter, backed
12     by a nylon filter and preceded by a sodium carbonate coated annular denuder; the second fine
13      particle channel has a quartz filter preceded by a citric acid-coated annular denuder; and the
14     coarse particle channel had a polycarbonate filter followed by a Zefluor filter for flow
15     distribution. Differences hi PM25 mass concentrations between the samplers, although not large,
16     were attributed to potential particle nitrate losses, denuder losses, and losses of SVOC for some
17     samplers. Differences between coarse particulate mass concentrations, on the other hand, varied
18     widely between the instruments.
19          In another intercomparison study, Tolocka et al. (2000) examined the magnitude of
20     potential sampling artifacts associated with the use of the FRM by collocating FRMs alongside
21     other chemical speciation samplers at four U.S. cities. The locations included a high nitrate and
22     carbon, low sulfate site (Rubidoux, CA); high crustal, moderate carbon and nitrate site
23     (Phoenix); high sulfate, moderate carbon, and low nitrate (Philadelphia); and low PM2 5 mass
24     (Research Triangle Park, NC).  The use of Teflon and heat-treated quartz filters also was
25     examined in this study. The Teflon filters collected less nitrate than the heat-treated quartz
26     filters. Filters in samplers using denuders to remove organic gases collected less organic PM
27     than filters in samplers without denuders.
28
29     2.2.4.2   Speciation Monitoring
30          In addition to FRM sampling to determine compliance with PM standards, EPA requires
31     states to conduct chemical speciation sampling primarily to determine source categories and
        March 2001
2-74
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 trends (Code of Federal Regulations, 1999c). A PM25 chemical speciation network is being
 deployed that will consist of 54 core National Ambient Monitoring Stations (NAMS) and
 250 State and Local Air Monitoring Stations (SLAMS). The overall goal of the speciation
 program is "to provide ambient data that support the Nation's air quality program objectives."
 (U.S. Environmental Protection Agency, 1999). The NAMS speciation sites will provide routine
 chemical speciation data that will be used to develop annual and seasonal aerosol
 characterization, air quality trends analysis, and emission control strategies. The SLAMS
 speciation sites will further support the NAMS network and provide information for
 development of State Implementation Plans (SIPs). At both types of sites, aerosol samples will
 be collected for analysis of trace elements, ions (sulfate, nitrate, ammonium, sodium, and
 potassium), and total carbon. The NAMS speciation sites will operate on a 1 in 3 day schedule,
 with 10 of these sites augmented for everyday operation. The SLAMS speciation sites will
 generally operate on a 1 in 6 day basis;  however, many sites may be operated on a 1  in 3 day
 basis in locations where increased data  collection is needed. The current samplers include three
 filters: (1) Teflon for equilibrated mass and elemental analysis (EDXRF), (2) a nitric acid
 denuded Nylon filter for ion analysis (ion chromatography), and (3) a quartz fiber filter for
 elemental and organic carbon (but without any correction for positive or negative artifacts caused
 by adsorption of organic gases or the quartz filters or evaporation of semivolatile organic
 compounds from the collected particles) (thermal optical analysis via NIOSH 5040 method).
 There are several samplers that are suitable for use in the NAMS/SLAMS network. These
 samples include an inlet cutpoint comparable to the WINS,  FRM; proven denuder technology for
 ions; and sampler face velocity and sample volume similar to that of the FRM with 46.2-mm
 diameter filters. Information and reports on EPA's speciation monitoring program may be found
 on EPA's Technology Transfer Network at http://www.epa.gov/ttn/amtic/pmspec.html.

Measurements for Source Category Apportionment
     Chemical analyses from the speciation network will be used for source category
apportionment via receptor modeling of PM. There are two major approaches to receptor
modeling: the chemical mass balance (CMB) receptor modeling approach, and statistically based
approaches. The CMB approach requires chemical characterization of all relevant sources.
Similar analyses should be used for characterization of receptor samples.  One of the advantages
       March 2001
                                         2-75
DRAFT-DO NOT QUOTE OR CITE

-------
 1     of using the CMB approach for receptor modeling is that it can be applied to a single sample, or
 2     to a limited number of samples.  CMB also uses chemical analyses that are performed routinely
 3     on speciation samples, such as EDXRF and ionic species. A considerable amount of receptor
 4     modeling work has been conducted with CMB using elemental analyses coupled with OC/EC
 5     and some ionic species (e.g., Watson et al., 1994; Hidy and Venkataraman, 1996; McLaren et al,
 6     1996; Vega et al., 1997).  Recent developments in receptor modeling include using organic
 7     analyses for tracers of specific sources (Benner et al., 1995), very detailed organic analyses for
 8     source fingerprinting (Rogge et al., 1991, 1993b,c,d,e, 1994,1997a,b, 1998), and chemical mass
 9     balance receptor modeling (Schauer et al., 1996).  Further detail on the organic analyses for these
10     studies is beyond the scope of this chapter.
11          Statistical models based upon factor analysis or principal component analysis (PCA) do not
12     require detailed source characterization information but have the drawback of requiring a large
13     data set of receptor sample analyses. These statistically based models have an additional benefit
14     in that they also can use other parameters such as meteorology. For a detailed review of factor
15     analysis and PCA, see Henry et al. (1984). In PCA, many intercorrelated variables within a large
16     data set are sorted into a smaller number of independent components, or factors, that account for
17     the variability in the data set. Veltkamp et al. (1996) performed a PCA for a study conducted at
18     Niwot Ridge, CO. Organic constituents of atmospheric aerosols were measured, along with
19     physical and meteorological data.  Organic compounds were thermally desorbed from the aerosol
20     particles at 250 °C in a pure helium atmosphere, separated by gas chromatography, and identified
21     by mass spectrometry. A principle component analysis was conducted using 31 variables that
22     included 18 particulate organic compounds, 11 gas-phase species (e.g., NO, NO2, HNO3, HONO,
23     PAN, H2O2, etc.), wind direction, and time of day. Several factors were identified that
24     distinguished various sources.. These included gas-phase internal combustion products;
25     particulate phase, oxygenated biogenic hydrocarbons; high molecular weight n-alkanes;
26     particulate phase anthropogenic products; and particulate phase biogenic aldehydes.
27          Pinto et al. (1998) also used a combination of PM25 chemical speciation and ambient
28     monitoring data hi a receptor modeling calculation to determine the relative sources of particulate
29     pollution in an industrial area in the northern Bohemia region of the Czech Republic. During
30     that study, a severe air pollution episode occurred in 1993 during which smoke and SO2
31      concentrations were 1800 and 1600 //g/m3, respectively.
        March 2001
2-76
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
      In addition to chemical speciation for factor analysis and source apportionment, Norris
 et al. (1999) showed that meteorological indices could prove useful in identifying sources of
 particulate matter that are responsible for observed health effects (specifically asthma) associated
 with exposure to particulate matter. They examined meteorology associated with elevated
 pollution events in Spokane and Seattle, WA, and identified a "stagnation index" that was
 associated with low wind speeds and increases in concentrations of combustion-related
 pollutants. Their factor analysis also identified a meteorological index (low relative humidity
 and high temperatures) that was associated with increases in soil-derived particulate matter, as
 well as a third factor (low temperatures and high relative humidity) that was associated with
 increasing concentrations of particulate sulfate and nitrate species (Norris, 1998).
      Ondov (1996) examined the feasibility of using sensitive isotopic and elemental tracer
 materials to determine the contributions of petroleum-fueled sources of PM10 in the San Joaquin
 Valley, CA. Costs of these experiments are affected not only by the tracer materials cost, but
 also by the sensitivities of the analytical methods for each, as well as the background levels.
 Suarez et al. (1996) used iridium tracer to tag emissions from diesel-burning sanitation trucks in
 Baltimore and determined the size distribution of soot from the trucks.

 Elemental Analyses
     X-ray emission, stimulated either by X rays (X-ray fluorescence, XRF) or by proton beams
 (Proton Induced X-ray Emission, PIXE) are standard techniques for nondestructive analysis of
 certain elements. Both were discussed in the previous 1996 PM AQCD. Some newer techniques
 with some advantages have become available in recent years.

     Energy Dispersive X-ray Fluorescence (EDXRF).  EDXRF by Method IO-4.4 is the
 method of choice for analysis of trace elements for the NAMS speciation program. EDXRF can
 accommodate small sample sizes and requires little sample preparation or operator time after the
 samples are placed into the analyzer.  It also leaves the sample intact after analysis so further
 analysis is possible.  The previous 1996 PM AQCD included a detailed discussion of EDXRF, so
that will not be repeated  here.
        March 2001
                                          2-77
DRAFT-DO NOT QUOTE OR CITE

-------
 1          Instrumental Neutron Activation Analysis (INAA).  INAA was mentioned only briefly in
 2     the 1996 PM AQCD and is expanded on here. INAA has been used to examine the chemical
 3     composition of atmospheric aerosols in several studies, either as the only method of analysis, or
 4     in addition to XRF (e.g., Yatin et al.,  1994; Gallorini, 1995). INAA has higher sensitivity for
 5     many trace species, and it is particularly useful in analyzing for many trace metals. Landsberger
 6     and Wu (1993) analyzed air samples collected near Lake Ontario for Sb, As, Cd, In, I, Mo, Si,
 7     and V, using INAA. They demonstrated that using INAA in conjunction with epithermal
 8     neutrons and Compton suppression produces very precise values with relatively low detection
 9     limits.
10          Enriched rare-earth isotopes have been analyzed via INAA and used to trace sources of
11     particulate matter from a coal-fired power plant (Ondov et al., 1992), from various sources in the
12     San Joaquin Valley (Ondov, 1996), from intentially tagged (iridium) diesel emissions from
13     sanitation trucks (Suarez et al., 1996; Wu et al., 1998), and from iridium-tagged emissions from
14     school buses (Wu et al., 1998).
15          An intercomparison was conducted in which 18 pairs of filters were sent to participants in
16     the Coordinated Research Program (CRP) on Applied Research on Waste Using Nuclear Related
17     Analytical Techniques (Landsberger et al., 1997). As part of that study, participants used PEXE,
18     INAA, XRF, or AAS to analyze the samples.  Many of the results for XRF and PIXE in the
19     coarse fraction were observed to be biased low compared to INAA. The authors speculated that
20     there is a systematic error because of self-attenuation of the X rays resulting from the particle
21     size effect.
22          In source apportionment studies, it is possible to use a combination of XRF and INAA to
23     develop a relatively complete set of elemental measurements.  Between these two analytical
24     techniques, good sensitivity is possible for many elements, including most of the toxic metals of
25     interest. The previous 1996 PM AQCD compared several methods for measuring elements.
26     In general, XRF provides better sensitivity for some metals (e.g., Ni, Pb, Cu, and Fe), whereas
27     INAA provides better sensitivity for others (Sb, As, Cr, Co, Se, and Cd). Both methods provide
28     similar detection limits for still other elements such as V, Zn, and Mn.
29
30          Atomic Absorption Spectrophotometry (AAS).  AAS was used to characterize the
31     atmospheric deposition of trace elements Zn, Ni, Cr, Cd, Pb, and Hg, to the Rouge River
       March 2001
2-78
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
 watershed by particulate deposition (Pirrone and Keeler, 1996). The modeled deposition rates
 were compared to annual emissions of trace elements that were estimated from the emissions
 inventory for coal and oil combustion utilities, iron and steel manufacturing, metal production,
 cement manufacturing, and solid waste and sewage sludge incinerators. They found generally
 good agreement between the trend observed in atmospheric inputs to the river (dry + wet
 deposition) and annual emissions of trace elements, with R2s varying from -0.84 to 0.98.  Both
 atmospheric inputs and emissions were found to have followed downward trends for Pb. For the
 period of 1987 to 1992, steady increases were observed for Cd (major sources are municipal solid
 waste incineration, coal combustion, sludge incineration, and iron and steel manufacturing),
 Cr and Ni (major sources are iron and steel production and coal combustion), and Hg (major
 sources are coal, the contribution from which had decreased from 53 to 45%, and municipal,
 solid, and medical waste incineration,  the contribution from which has increased).

     Inductively Coupled Plasma-Mass Spectroscopy (ICP-MS). Keeler and Pirrone (1996)
 also used ICP-MS to determine trace elements Cd, Mn, V, As, Se, and Pb in atmospheric
 particulate fine (PM2.5) and total suspended particulate samples collected in two Detroit sites.
 The results were then similarly used in a deposition model to estimate the dry deposition flux of
 trace elements to Lake Erie.

     Scanning Electron Microscopy (SEM).  Mamane et al. (2000) investigated the use of
 computer-controlled scanning electron microscopy (CCSEM) as a way of supplementing XRF
 analysis and providing automated analysis of particle size, chemistry, and particle classification.
An ambient coarse particulate sample from Baltimore was collected on a polycarbonate filter for
this analysis. CCSEM analyses were conducted for 2819 particles in 78 randomly selected fields
of view during an unattended 8-h run.  Mamane et al. confirmed the stability of the CCSEM
instrument over several hours of operation. The physical properties of the sample such as
particle diameter, mass loading per field, and particle number per field, were well represented by
analyzing approximately 360 particles, with little additional information gained by analyzing
more particles. Teflon filters are not well suited for SEM analyses. Analysis of fine PM is
expected to pose analytical challenges not addressed in the present study (Mamane et al., 2000).
       March 2001
                                         2-79
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Nelson et al. (2000) applied Raman chemical imaging and SEM (Raman/SEM) to study the
 2     size, morphology, elemental and molecular composition, and molecular structure of fine
 3     particulate matter. In their study, filter compatibility was examined, and Raman/SEM chemical
 4     imaging was conducted for several standard materials as well as for ambient PM2 5 samples.
 5     Polycarbonate was determined to be a suitable substrate for both SEM and Raman chemical
 6     imaging analysis.
 7               ,
 8          Elemental and Organic Carbon in Particulate Matter. Total carbon in aerosol particles ,
 9     (TC) can be expressed as the sum of organic .carbon (OC), elemental carbon (EC), and carbonate
10     carbon (CC), with the contribution of CC to TC usually on the order of 5% or less, for particulate
11     samples collected in urban areas (Appel, 1993). The 1996 PM AQCD  (U.S. Environmental
12     Protection Agency, 1996) listed several filter-based, thermal methods for measuring OC and EC,
13     and described the thermal/optical reflectance (TOR) method, which was noted, along with
14     thermal manganese oxidation, to be one of the most commonly applied methods in the United
15     States at the .time. In thermal separation methods, thermally evolved OC and EC are oxidized to
16     CO2 and quantified either by nondispersive infrared detection or electrochemically, or the CO2
17     can be reduced to CH4 and  quantified via flame ionization detection (FID). The various methods
18     give similar results for TC, but not for EC or OC.
19          In a methods comparison study (Countess,  1990), it was shown that it is necessary to
20     minimize or correct for pyrolytically generated EC ("char"), and that CC found in wood smoke
21     and automobile exhaust samples may interfere with some of the thermal methods. Recently,
22     Lavanchy et al.  (1999) reported on a study in which the operation of a catalytic oxidation system
23     was modified in an attempt to minimize pyrolysis of OC and at the same time minimize the
24     contribution of CaCO3. The system uses two ovens, one at 340  °C and one at 650 °C. The filter
25     sample is placed in a moveable sample boat.  In order to minimize charring, the sample is first
26     flash heated in the 650 °C oven for 1  min. Then it is inserted into the 340 °C stage of the two-
27     stage oven.  In both steps OC is oxidized to CO2 in the presence of O2. The second step requires
28     42 min. The filter, then is moved into the second-stage oven.  During this third step, EC is
29     oxidized to CO2 at 650 °C  for 32 min. This temperature is reported to be sufficient to completely
30     oxidize EC,  but with only about 1 % of the CaCO3 being vaporized (Lavanchy et al., 1999;
31     Petzold et al., 1997). To test for charring, they challenged their system with atmospheric samples
        March 2001
2-80
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17  .
 18
 19
20
21
22
23
24
25
26
27
28
29
30
 for which duplicates were analyzed via the German reference method for measuring OC and EC
 in atmospheric samples (Petzold and Niessner, 1995), in which a solvent extraction is used to
 remove organics before combustion. Lavanchy et al. (1999) reported a high correlation
 (R2 = 0.97) between their thermal oxidation method and the German method VDI. The slope of
 the EC:EC VDI line was 0.92, and the intercept was -^0.37 /ag cm2. They also reported detection
 limits of 1.3 Aig for EC and 1.8//g for OC.
      Pyrolytic char is corrected for in thermal-optical analysis, hi thermal-optical carbon
 analysis (Birch and Gary, 1996; Chow et al., 1993), punches from a quartz sampling filter are
 inserted into the carbon analyzer and heated in a helium atmosphere to volatilize organic carbon.
 Then, the temperature is reduced, and oxygen is added to the carrier gas, so that desorbed
 compounds are then oxidized to CO2, reduced to methane, and measured in a flame ionization
 detector.  In order to account for the portion of the OC that is pyrolyzed, a He-Ne laser monitors
 the sample reflectance (or transmittance).  As the pyrolysis occurs, the sample gets darker, and
 the reflectance decreases. Then, as elemental carbon is removed, the filter lightens, and the
 reflectance increases until all carbon has been removed from the filter. The split between organic
 and elemental carbon is considered to be the point at which the reflectance regains its
 prepyrolysis value, with material removed prior to this point being considered organic, and that
 after, elemental.
      The thermal/optical transmission  method (TOT) is similar to the TOR with the primary  '
 difference being that light transmission rather than reflectance is monitored on the filter
 throughout the analysis.  The TOT method of Birch and  Gary (1996) consists of a two-stage
 process, with the first stage being conducted in a pure helium atmosphere, and the second stage
 conducted in a 10% oxygen-helium mix. The temperature is ramped to about 820 °C in the
 helium phase,  during which organic and carbonate carbon are volatilized from the filter. In the
 second stage, the oven temperature is reduced, and then raised to about 860 °C. During the
 second stage, pyrolysis correction and EC measurement are made. Figure 2-14, an example of a
thermogram, shows temperature, transmittance, and FID response traces. Peaks are evident that
 correspond to OC, CC, EC, and pyrolitic carbon (PC). As can be seen in this figure, the high
temperature in the first stage allows for decomposition of CC. The ability to quantify PC is '
particularly important in high OC/EC regions (like wood-smoke-impacted airsheds), allowing for
       March 2001
                                         2-81
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
                                                       OC - EC split
                                        Time/min
      Figure 2-14. This thermogram, for a sample containing rock dust (a carbonate source) and
                  diesel exhaust, shows three traces that correspond to temperature, filter
                  transmittance, and FID detector response. Peaks correspond to organic (OC),
                  carbonate (CC), pyrolytic (PC), and elemental (EC) carbon. The final peak is
                  a methane calibration peak.
      Source: Birch and Gary (1996).
1     the volatilization of any remaining complex organic compounds so they are not apportioned to
2     the EC phase.
3          The National Institute for Occupational Safety and Health (NIOSH) Method 5040 for
4     monitoring elemental carbon as a.marker for particulate diesel exhaust is based upon a TOT
5     method analyzer (Birch and Gary, 1996), while the OC/EC method specified for the IMPROVE
6     network is the TOR method (Chow et al., 2000). Chow et al. (2000) compared the OC, EC, and
      March 2001
2-82
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 TC measurements from NIOSH and IMPROVE methods. The two methods differ in that
 temperature and atmospheric controls that are used to accomplish carbon speciation, in addition
 to the use of light transmission in the case of the NIOSH method, as compared to light
 reflectance in the IMPROVE method, of the filter is measured during the analysis.  The
 IMPROVE thermal protocol specifies organic carbon fractions at 120, 250, 450, and 550 °C in a
 nonoxidizing atmosphere (He), and elemental organic fractions at 550, 700, and 800 °C in an
 oxidizing atmosphere. The NIOSH method differs in its thermal protocol, which has organic
 carbon fractions at 250, 500, 650, and 850 °C in a nonoxidizing atmosphere (also He), and
 elemental carbon fractions at 650, 750, and 850 °C in an oxidizing atmosphere.  The high
 temperature before addition of oxygen in the NIOSH method is to quantify particulate carbonate,
 which evolves between 650 and 830 °C (Birch and Gary, 1996). The two methods also differ in
 the specified residence times at each temperature setpoint. The residence times  at each setpoint
 are typically longer for the IMPROVE analysis compared to the NIOSH analysis.
      Chow et al. (2000) analyzed 60 quartz filter samples that represented a wide variety of
 aerosol compositions and concentrations. The TC measurements from each protocol were in
 good agreement, with no statistically significant differences. A statistically significant difference
 was observed in the fraction of TC that is attributed to EC, as determined by the IMPROVE and
 NIOSH thermal evolution protocols, with the IMPROVE EC measurements typically higher than
 the NIOSH EC measurements. This difference was attributed to the 850 °C temperature step in
 the oxidizing atmosphere in the NIOSH protocol. Chow et al. compared the OC for each method
 and found that the two methods showed good agreement when the 850 °C nonoxidizing
 temperature step in the NIOSH method was not included in determination of OC. There was also
 a difference between the reflectance and transmittance detection methods in the pyrolysis
 adjustment, although this difference was most noticeable for very black filters for which neither
 reflectance nor transmittance was able to  accurately detect further blackening by pyrolysis.
 Because OC and EC are operationally defined parameters, Chow et al. pointed out that it is
 importance to retain ancillary information when reporting EC and OC by these analytical
methods, so that comparisons can be made among measurements taken at different sites using
these two methods.
     Further refinement of thermal techniques has resulted in the evolved gas analysis (EGA)
method, described by Grosjean et al. (1994).  This technique  involves combustion of particulate
       March 2001
                                        2-83
                                                           DRAFT-DO NOT QUOTE OR CITE

-------
 1      matter samples in an oxidizing environment while the temperature is raised from 100 to 600 °C.
 2      The amount of evolved CO2 contains information about the volatility of the organic aerosol
 3      compounds. Grosjean et al. (1994) present thermograms both for specific organic compounds
 4      (e.g., adipic acid) and for specific sources (e.g., vehicular traffic). They suggest that EGA may
 5      be useful for source apportionment applications.  Kirchstetter et al. (2000) and Novakov et al
 6      (1997) have used EGA to provide insights regarding organic sampling artifacts.
 7           Black carbon (BC) also is used, hi addition to the thermal and thermal/optical methods, for
 8      determining EC as a measure of soot (Penner and Novakov, 1996). Both EC and BC define a
 9     similar fraction of aerosol, but EC is determined based on thermal properties, whereas BC is
10     based on light-absorption properties. Optical methods for determining BC tend to suffer from
11      calibration problems (Hitzenberger et al., 1996). Lavanchy et al. (1999) compared their EC
12     concentrations as determined from their catalytic thermal oxidation method to BC concentrations
13     determined using an aethalometer operated at the same site, and found that the instrumental
14     calibration factor provided by the manufacturer was on the order of two times the calibration
15     factor they determined (9.3 ±0.4 m2g'')-  It is possible to calculate a theoretical specific
16     absorption coefficient (BJ from Mie theory given a known size distribution and refractive index,
17     and typically BC aerosols have values of Ba between 3 and 17 m2g-' (Hitzenberger et al. [ 1996]
18     and references therein). The Ba is defined as absorption per mass concentration and can be
19     calculated given the sample filter area, the total deposited mass, and absorption signals for both
20     the loaded and unloaded filters. Often, when no direct measurements are available, values of Ba
21     on the order of 10 mV1 have been used (Hitzenberger et al. (1996), and references therein).
22     European countries are trying to set air pollution standards that target diesel vehicles, one of the
23     principal sources of BC in urban areas (Hitzenberger et al. (1996), and references therein) and so
24      it is essential that accurate values  for Ba are available. Hitzenberger et al. (1996) investigated the
25      feasibility of using an integrating sphere photometer as an adequate measurement system for the
26      BC content and the absorption coefficient. Based on samples collected during a 10-day period in
27      May 1994, they determined that the usually assumed value of 10 m2g'1 was also applicable to
28      aerosol BC occurring in Vienna.
29           Hitzenberger et al. (1999) recently reported on a study in which the integrating sphere
 30      method was compared to an aethalometer (Hansen et al., 1984), the thermal method of Cachier
 31      et al. (1989), and the thermal/optical method of Birch and Gary (1996). The absorption
        March 2001
2-84
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 coefficients that were obtained from both the integrating sphere and the aethalometer were
 comparable. The BC mass concentrations obtained from the aethalometer were 23% of those
 obtained from the integrating sphere.  Compared to the thermal method, the integrating sphere
 overestimated the BC mass concentrations by 21%. Compared to the thermal/optical method, the
 integrating sphere was within 5% of the 1:1 line. However, the data were not so well correlated.
      In 1986, the Carbonaceous Species Methods Comparison Study (CSMCS) was conducted
 in Los Angeles, during which a number of methods for the measurement of this species were
 intercompared. The CSMCS was mentioned in the 1996 PM AQCD (U.S. Environmental
 Protection Agency,  1996). Hansen and McMurry (1990) specifically compared two very
 dissimilar methods for aerosol elemental carbon—collection of impactor samples backed by a
 quartz fiber afterfilter, followed by EC analysis by oxidation in helium over a MnO2 catalyst, and
 real-time measurements using an aethalometer (an optical absorption technique)—and found
 good agreement between these two, very different methods. The CSMCS interlaboratory
 precision for total carbon was 4.2% (Turpin et al., 2000). However, because the split between
 OC and EC is operationally defined, there was substantial interlaboratory variability in OC and
 EC (e.g., 34% for EC [Turpin et al., 1990]). The implications for data analysis  are twofold:
 (1) the analysis method used must be reported with particulate carbon data, and (2) comparative
 analyses should not be conducted with data analyzed by more than one carbon analysis method
 unless the mutual compatibility of the methods has been demonstrated.  Carbon analysis methods
 currently are being compared as a part of the Atlanta Supersite.
      Turpin et al. (1990) reported on an in situ, time-resolved analyzer for particulate organic
 and elemental carbon that could operate on a time cycle as short as 90 min. The analyzer
 consists of a filter-based sampling section and a thermal-optical carbon detector. Adsorbed
 organic material is thermally desorbed from the filter at 650 °C  and oxidized at  1000 °C over a
 MnO2 catalyst bed.  The evolved CO2 is converted to methane over a nickel catalyst, and the
 methane is measured in a flame ionization detector. Then the elemental carbon  is oxidized in a
 98% He-2% O2 atmosphere, at 350 °C.  Correction is made for pyrolytic conversion of some of
the organic particulate matter. The instrument was  operated with a 2-h time resolution during the
Southern California Air Quality Study (SCAQS) in 1987 (Turpin and Huntzicker, 1991), as well
as during the Carbonaceous Species Methods Comparison Study (CSMCS) in 1986.
       March 2001
                                         2-85
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
 27
 28
 29
 30
 31
     EC/OC Summary. The state of the art for soot measurements continues to develop, and,
although advances are being made, the definitions of EC and BC continue to be operational and
determined by the method employed. Similarly, the distinction between OC and EC also is
defined operationally. Therefore, reports of EC/OC measurements should include mention of the
method with which the species were determined. Finally, if possible, all ancillary data should be
retained, to allow later comparison to other methods.

     Ions. Aerosol ions refers to the water-soluble portion of suspended PM. Ion
chromatography (1C) is widely used for analyzing ionic species.  1C is the method of choice for
the measurement of sulfate, nitrate, ammonium, sodium, and potassium ions for the NAMS
program.  Aerosol strong acidity, H+, is determined by titration of a water solution of PM
collected following a series of annular denuders to remove acid and basic gases with back-up
filters to collect NH3 and HNO3 that might volatilize from the PM during collection.  The 1996
PM AQCD (U. S. Environmental Protection Agency, 1996) discussed measurement of ions by 1C
(Section 4.3.3.1) and of strong acidity (Sections 3.3.1.1 and 4.3.3.1) so  no further details will be
discussed here.

2.2.5  Continuous Monitoring
     The EPA expects that more than 200 local agency monitoring sites throughout the States
will operate continuous PM monitors. All currently available continuous measurements of
suspended particle mass share the problem of dealing with semivolatile PM components. So as
not to  include particle-bound water as part of the mass, the particle-bound water must be
removed by heating or dehumidification.  However, heating also causes loss of ammonium
nitrate and semivolatile organic components. A variety of potential candidates for continuous
measurement of mass or chemical components will be discussed in this section.

 2.2.5.1  Tapered Element Oscillating Microbalance
      The advantages of continuous PM monitoring, and the designation of the Tapered Element
 Oscillating Microbalance (TEOM) as an equivalent method for PM10, have led to the deployment
 of the TEOM at a number of air monitoring sites.  The TEOM also is being used to measure
 PM2 5. The TEOM differs philosophically from the federal reference methods for particulate
        March 2001
                                          2-86
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
 mass in that it does not require equilibration of the samples at a specified temperature and
 relative humidity.  Moreover, the TEOM samples at a constant temperature, typically heated to
 some temperature higher than the ambient temperature (Meyer et al., 1995; Meyer and.
 Rupprecht, 1996), whereas the  FRM samples at the ambient temperature. Thus, the TEOM does
 not provide data equivalent to the FRM because of losses of volatile species. Volatilization
 losses in the TEOM sampler can be reduced by operating the instrument heated to 30 °C rather
 than the 50 °C specified, during the cooler times of the year, and by using Nafion dryers on the
 inlet.
     This philosophical difference in operation and implications for fine particle measurements
 were examined by researchers at CSIRO Atmospheric Research in Australia (Ayers et al., 1999).
 That group compared 24-h mean PM2 5 mass concentrations as determined by a TEOM and by
 two manual, gravimetric samplers (a low-volume filter sampler and a MOUDI sampler) in four
 Australian cities, on 15 days in  the winter half-year. The TEOM was operated at 50 °C at one
 location and at 35 °C at the other three locations. A systematically low TEOM response in
 comparison to the integrated gravimetric methods was observed. In a comprehensive study,
 Allen et al. (1997) reported results in which TEOM data collected at 10 urban sites in the United
 States and Mexico were compared with 24-h integrated mass concentrations, for both PM10 and
 PM2 5.  They collected a large data set that included both winter and summer seasons. Allen et al.
 (1997) concluded that, especially for urban areas, a significant fraction of PM10 could be
 semivolatile compounds that could be lost from the heated filter in the TEOM, thus leading to a
 systematic difference between the TEOM and the EPA FRM for PMI0. They suggested that this
 difference is likely to be larger for PM25 than it is for PM10 (Allen et al., 1997).
     In a similar study conducted in Vancouver, British Columbia, the effect of equilibration
 temperature on PMIO concentrations  from the TEOM was examined. Two collocated TEOM
 monitors, operated at 30 and 50  °C, respectively, were operated in the Lower Fraser Valley in
 British Columbia for a period of approximately 17 mo (Mignacca and Stubbs, 1999).  A third
 TEOM operating at 40 °C was operated for 2 mo during this period.  They found that, on
 average, the 1-h PM10 from the TEOM operating at 30 °C was consistently greater than that from
the TEOM operated at 50 °C.  For the period during which the third TEOM was operated (at
40 °C), the PM10 from that instrument was between those values for the other two instruments.
They also found that the differences in masses were proportional to the PM10 loading, and more
       March 2001
                                        2-87
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1      strongly correlated to the PM10 from the TEOM operated at the lower temperature. They
 2      recommended that the TEOM monitors be operated at 40 °C as opposed to operating at 50 °C in
 3      summer and 30 °C in winter, hi order to avoid introducing a methodological seasonal bias.
 4      In most parts of Canada, and including the Greater Vancouver Regional District, TEOMs for
 5      both PM,0 and PM2.5 are operated with this revised protocol.
 6           A new sample equilibration system (SES) was developed to allow conditioning of the
 7      sample stream to lower humidity and temperature, to reduce losses of semivolatile species from
 8      the TEOM (Meyer et al., 2000). The SES utilizes a Nafion dryer designed for low particle loss,
 9      and humidity sensors.  The dryer fits between the flow splitter that follows the size-selective inlet
10     and the sensor unit. A dry purge gas flow over the exterior of the Nafion tubing allows for self-
11      regeneration.  A TEOM with PM2 5 inlet and equipped with an SES was operated at 30 °C
12     alongside another TEOM operating at 50 °C without the SES in Albany, NY, over a 6-day period
13     during a summertime high-temperature, high-relative-humidity episode.  The SES maintained the
14     sample air relative humidity under 30% and the TEOM with the  SES generally measured more
15     mass than the other TEOM. The TEOM with SES also was operated alongside an.FRM-type
16     sampler for the period of June 6 through September 25,1999. The correlation between the FRM
17     and TEOM/SES showed a slope of 1.0293 and R2 of 0.9352, whereas the correlation between the
18     FRM and the TEOM without SES and operating at 50 °C showed a slope of 0.8612 and R2 of
19     0.8209. The  SES can be installed on existing TEOM monitors.
20           Patashnick et al.  (2000) developed a differential TEOM system that is based on a pair of
21     TEOM sensors, each of which is preceded by its own electrostatic precipitator (ESP), and
22     downstream from a common size selective inlet.  By alternately  switching the ESPs on and off,
23     and out of phase with each other, the two sensors measure  "effective mass" that includes both the
24     nonvolatile component and the volatile component sampled by the TEOM, less the volatile
25     component that vaporized during the sampling interval. On the  sensor side with the ESP turned
26     on, there is no particle collection on that filter, so that only volatilization of previously collected
27     particles continues.  This would allow for correcting the effective mass as measured from the
28     first sensor, by subtracting out the volatilization artifact, and leaving the nonvolatile and volatile
29     components of the particulate matter.  This system has yet to be  well characterized for other
30     biases or interferences such as reactions on the filters, particle collection efficiency of the ESPs,
31     and particle and semivolatile material losses.
        March 2001
2-88
DRAFT-DO NOT QUOTE OR CITE

-------
   1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
27
28
29
30
31
 2.2.5.2  Real-Time Total Ambient Mass Sampler
      A Real-Time Total Ambient Mass Sampler, RAMS, based on diffusion denuder and
 TEOM monitor technology has been developed, validated, and field tested (Eatough et al.,
 1999a; Obeidi and Eatough, 1999; Pang et al., 2001) for the real-time determination of total fine
 paniculate mass, including semivolatile species.  The RAMS measures total mass of collected
 particles, including semivolatile species with a TEOM monitor using a "sandwich filter". The
 sandwich contains a Teflon coated particle collection filter followed by a charcoal impregnated
 filter (GIF) to collect any semivolatile species lost from the particles during sampling. Because
 the instrument measures total mass collected by the sandwich filter, all gas phase compounds that
 can be adsorbed by a CIF must be removed from the sampling stream prior to the TEOM
 monitor. Laboratory and field validation data indicate that the precision of fine PM mass
 determination is better than  10%. The RAMS uses a Nafion dryer to remove particle bound
 water from the suspended particles and a particle concentrator to reduce the amount of gas phase
 organics that must be removed by the denuder. An example of data from the RAMS, the TEOM,
 and the PC-BOSS is shown in Figure 2-15. This figure also shows the PM2 5 mass from the
 TEOM as being negative for the hours of 16 to 19. This likely results from the loss of volatile
 materials from the heated filter.

 2.2.5.3  Continuous Ambient Mass Monitor
     Koutrakis and colleagues (Koutrakis et al., 1996; Wang, 1997) have developed the
 Continuous Ambient Mass Monitor (CAMM), a technique for the continuous measurement of
 ambient particulate matter mass concentration, based on the measurement of pressure drop
 increase with particle loading across a membrane filter. Recently, Sioutas et al. (1999) examined
 the increase in pressure drop with increasing particle loading on Nuclepore filters. They tested
 filters with two pore diameters (2 and 5 //m) and filter face velocities ranging from 4 to 52 cm/s,
 and examined the effects of relative humidity in the range of 10 to 50%. They found that, for
 hygroscopic ammonium sulfate particles, the change in pressure drop per unit time and
 concentration was a strong function of relative humidity, decreasing with increasing relative
 humidity. These results suggest that particulate concentration measurements like the method of
Koutrakis et al. (1996) that use the pressure drop method may be subject to additional
uncertainties if used in an environment where the ambient relative humidity cannot be controlled
       March 2001
                                         2-89
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
                  PC-BOSS (Nonvolatile Material)
                     TEOM
                     at35C
—B—  RAMS
       at35C
PC-BOSS (Lost From Particles)

       	FRMPM2.5
             24 h average
                                      Riverside, CA
                |  | |  |  | | |  |  I I  I  I I  I  I I I  I  I I  I  I I  i  I  I I  I  I I  I  I I  I  I I  I  I  I I
             13  14 15  16 17 18  19 20  21  22 23 0   1
                                          Time of Day
      Figure 2-15. Comparison of mass measurements with collocated RAMS (real-time data),
                 PC-BOSS (1-h data), FRM PM2.5 sampler (24-h data), and a conventional
                 TEOM monitor (real-time data). The semivolatile fine particulate matter is
                 sampled with the RAMS and PC-BOSS, but not with the TEOM monitor or
                 the FRM PM2-S single filter sampler.  The PC-BOSS provides information on
                 both the nonvolatile component (NV) and the semivolatile organic component
                 (SVOC).

      Source: Eatoughetal.(1999a).
1     accurately.  Tlie current version of the CAMM (Wang, 1997) uses a particle concentrator, a
2     Nafion dryer, and frequent changes of the position on the filter tape where the pressure drop

3     measurement is made to avoid artifacts due to semivolatile components.
4          The CAMMS was recently operated alongside a gravimetric PM method (the Harvard
5     Impactor, or HI) in seven U.S. cities selected for their distinctly different ambient particulate
      March 2001
       2-90
DRAFT-DO NOT QUOTE OR CITE

-------
   1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
 compositions and densities. The correlation between the two methods was high, with an overall
 r2 of 0.90, and average CAMM/HI ratio of 1.07 (Babich et al., 2000).

 2.2.5.4  Light Scattering                                   ,
      The evaporation of ammonium nitrate aerosol in a heated nephelpmeter was examined by
 Bergin et al (1997). This is of interest, because nephelometers are frequently operated with the
 sampled airstream heated to a low reference relative humidity of 40%, in order .to measure the
 light scattering because of the dry aerosol rather than that caused by particle-bound water.
 Bergin et al. conducted laboratory experiments at low relative humidity (~ 10%) and as a function
 of temperature (300 to 320 K), mean residence time in the nephelometer, and initial particle size
 distribution. The evaporation of ammonium nitrate aerosol was also modeled, for comparison,
 and was found to accurately describe the decrease in aerosol scattering coefficient as a function
 of aerosol physical properties, and nephelometer operating conditions. Bergin et al. (1997)
 determined an upper limit estimate of the decrease in the aerosol light scattering coefficient at
 450 run from evaporation for typical field conditions. The model estimates for their worst^case
 scenario suggest that the decrease in the aerosol scattering coefficient could be roughly 40%.
 Under most conditions, however, they estimate that the decrease in aerosol scattering coefficient
 generally is expected to be less .than 20%.
     Morawska et al. (1996) examined the correlations between PMIO, visibility, and submicron
 concentration data in Brisbane, and concluded that the different principles of operation for, each
 instrument and the different aerosol characteristics measured by each technique make it difficult
 to observe any relationships. Morawska et al. (1998b) reported on a long-term monitoring
 program that included the criteria pollutants as well as light scattering, number/size distributions,
 number concentrations, and elemental analysis via inductively coupled plasma mass
 spectrometry. Particle size classification was conducted using a TSI scanning mobility particle
 sizer (SMPS) for the size range of 0.016 to 0.7 ^m, and a TSI aerodynamic particle sizer (APS)
for the size range of 0.7 to 30 /zm. They reported correlation coefficients between the light-
scattering coefficient and PM10, SMPS concentration, and APS concentration of 0.58, 0.38, and
0.37, respectively. They also reported a correlation coefficient between PMIO and the SMPS
concentration of 0.25. A lower correlation between PMIO mass and the SMPS concentration is
       March 2001
                                         2-91
                                                             DRAFT-DO NOT QUOTE OR CITE

-------
 1      consistent with the notion that PM10 mass measurements would provide less information about
 2      smaller particles in the 0.016 to 0.7//m range.
 3
 4      2.2.5.5  Beta-Gauge Techniques
 5           The use of absorption of beta radiation as a indicator of particle mass has been used
 6      effectively to measure the mass of equilibrated particulate matter collected on Teflon filters
 7      (Jaklevic et al.,  1981; Courtney et al., 1982).  The technique also has been used to provide near
 8      real-time measurements with time intervals on the order of an hour (Wedding and Weigand,
 9      1993). However, real-time beta gauge monitors experience the same problems as other
10      continuous or near real-time particulate matter mass monitoring techniques. Particle-bound
11      water must be removed to reduce the sensitivity of the indicated mass to relative humidity.
12     However, the simplest technique, mild heating, will remove a portion of the ammonium nitrate
13     and the semivolatile organic compounds as well as the particle-bound water.
14          An intercomparison study of two beta gauges at three sites indicated that the Wedding beta
15     gauge and the Sierra Anderson SA 1200 PM10 samplers were highly correlated, r > 0.97 (Tsai and
16     Cheng, 1996).  The Wedding beta gauge was not sensitive to relative humidity but was
17     approximately 7% lower. This suggests that the mild heating in the beta gauge causes losses
18     comparable to those caused by equilibration, although the differences could result from slight
19     differences in the upper cut points. The Kimoto beta gauge, however, which was operated at
20     ambient temperature, was sensitive to relative humidity, yielding significantly higher mass
21     concentrations  relative to the Sierra Anderson SA 1200 for RH > 80% than for RH < 80%, even
22     though the correlation with the SA 1200 was reasonable, r = 0.94 for RH > 80% and 0.83 for
23     RH<80%.
24
25      2.2.5.6  Measurements of Individual Particles
26           Recently, several researchers have developed instruments for real-time in situ analysis of
27      single particles (e.g., Noble and Prather, 1996; Gard et al., 1997; Johnson and Wexler, 1995;
28      Silva and Prather, 1997; Thomson and Murphy, 1994). Although the technique varies from one
 29      laboratory to another, the underlying principle is to fragment each particle into ions using either a
 30      high-power laser or a heated surface and to then use a time-of-flight mass spectrometer (TOFMS)
 31      to measure the ion fragments in a vacuum. Each particle is analyzed in a suspended state in the
        March 2001
2-92
DRAFT-DO NOT QUOTE OR CITE

-------
  1
 2
 3
 4
 5
 6
 7
 8
 air stream (i.e., without collection), avoiding sampling artifacts associated with impactors and
 filters. By measuring both positive and negative ions from the same particle, information can be
 obtained about the chemical composition, not just the elemental composition, of individual
 particles of known aerodynamic diameter. This information is especially useful in determining
 sources of particles. An example of the type of information that can be determined is shown in
 Figure 2-16.
                                                                                  Organic
                                                                                 Marine
                                                                               Soil
                   0.2
                   0.3 0.40.5  0.60.70.80.91.0         2.0
                              Aerodynamic Diameter
                                                                     3.0
                                                                           4.0
      Figure 2-16. Size distribution of particles divided by chemical classification into organic,
                   marine, and crustaL
1
2
3
4
5
     Because particles are analyzed individually, biases in particle sampling (the efficiency of
particle transmission into the sensor chamber as a function of size; particle size measurement,
and detection of particles prior to fragmentation) represent a major challenge for these
instruments. Moreover, the mass spectrometer has a relatively large variability in ion yields (i.e.,
identical samples would yield relatively large differences in MS signals [Thomson and Murphy,
      March 2001
                                          2-93
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1      1994]); therefore, quantitation is inherently difficult (Murphy and Thomson, 1997). Quantisation
 2      will be even more challenging for complex organic mixtures because of the following two
 3      reasons:  (1) a large number of fragments are generated from each molecule, and (2), ion peaks
 4      for organics can be influenced or obscured by inorganic ions (Middlebrook et al., 1998).
 5      Nonetheless, scientists have been successful in using these techniques to identify the presence of
 6      organics in atmospheric particles and laboratory-generated particles (i.e., as contaminants in    ,
 7      laboratory-generated sulfuric acid droplets) as well as the identification of specific compound
 8      classes such as PAHs in combustion emissions (Castaldi and Senkan, 1998; Hinz et al., 1994;
 9      Middlebrook et al., 1998; Murphy and Thomson, 1997; Neubauer et al., 1998; Noble and Prather,
10     1998; Reilly et al., 1998; Silva and Prather, 1997).
11           Until recently, single particle ATOFMS systems have only been able to characterize
12     particles that are larger than approximately 0.2 to 0.3 //m in diameter.  Wexler and colleagues
13     (Carson et al., 1997; Ge et al., 1998) have developed a single particle, TOFMS instrument that is
14     able to size, count, and provide chemical composition on individual particles ranging in size
15     fromlOnmto2//m.
16          Noble and Prather (1996) used ATOFMS to provide compositionally resolved particle-size
17     distributions. Their instrument is capable of analyzing, at typical ambient concentrations, size
 18     and chemical composition of 50 to  100 particles/min, and up to 600/min at high particle
 19     concentrations.  Data storage requirements are met using a Pentium 90 mHz personal computer.
20
21      2.2.5.7  Automated Fine Particulate Nitrate
 22           An integrated collection and vaporization cell was developed by Stolzenburg and Hering
 23      (2000) that provides automated, 10-min resolution monitoring of fine particulate nitrate. In this
 24      system, particles are collected by humidified impaction process and analyzed in place by flash
 25      vaporization and chemiluminescent detection of the evolved nitrogen oxides. In field tests in
 26      which the system was collocated with two FRM samplers, the automated nitrate sampler results
 27      followed the results from the FRM, but were offset lower. The system also was collocated with a
 28      Harvard EPA annular denuder sampler (HEADS), as well as a SASS speciation sampler
 29      (MetOne Instruments). In all these tests, the automated sampler was well correlated to other
 30      samplers, with slopes near 1 (ranging from 0.95 for the FRM to 1.06 for the HEADS) and
 31      correlation coefficients ranging from 0.94 to 0.996.
        March 2001
                                                  2-94
DRAFT-DO NOT QUOTE OR CITE

-------
   1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
 30
31
 2.2.5.8  Semi-continuous Carbon Analysis
      Testing and refinement of models that simulate aerosol concentrations from gas and
 particle emissions require air quality measurements of approximately 1-h time resolution to
 reflect the dynamics of atmospheric transport, dispersion, transformation and removal. Below
 we describe instruments that have been used to collect and analyze atmospheric organic PM with
 better than 2-h time resolution. These instruments were all present at the Atlanta Supersite
 experiment during the summer of 1999, and an intercomparison of results is underway.
      An "in situ carbon analyzer" measured total particulate organic and elemental carbon (i.e.,
 Mg of carbon/m3) with 1 to 2 h resolution in Glendora and Claremont, CA, during 1986 and 1987
 (Turpin and Huntzicker, 1991; Turpin and Huntzicker, 1995), and in Atlanta, GA, during 1999
 (Supersite experiment, unpublished). By using elemental carbon as a tracer for primary,
 combustion-generated organic carbon, these authors estimated the contributions of primary
 sources (i.e., material emitted in particulate  form) and secondary sources (i.e., particulate
 material formed in the atmosphere) to the total atmospheric particulate organic carbon
 concentrations in these locations.  This in situ carbon analyzer collects particulate matter on a
 quartz fiber filter mounted in a thermal-optical transmittance carbon analyzer (Turpin et al.,
 1990). The material on a quartz fiber filter behind a Teflon filter in the second sampling port
 provides an estimate of the positive sampling artifact (i.e., gas adsorption on the quartz sampling
 filter).
     An automated carbon analyzer with 15-min to 1-h time resolution is now commercially
 available (Rupprecht et al., 1995) and has been operated in several locations, including the
 Atlanta Supersite. It collects samples on a 0.1 -/am impactor downstream of an inlet with a
 2.5-fj.m cutpoint. Use of an impactor eliminates gas adsorption that must be addressed when
 filter collection is used. However, this collection system may experience substantial particle
 bounce, and a sizable fraction of EC is in particles < 0.2 /j.m.  In the analysis, step carbonaceous
 compounds are removed by heating in filtered ambient air.  Carbonaceous material removed
 below 340  °C is reported as organic carbon,  material removed between 340 and 750 °C is
 reported as elemental carbon. Turpin et al. (2000) comment that it would be more appropriate to
report carbon values obtained by this method as "low-" and "high-temperature" carbon, because
 some organics are known to evolve at temperatures greater than 340 °C (e.g., organics from
woodsmoke).
       March 2001
                                         2-95
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1           An aethalometer is an automated, time-resolved instrument (i.e., 5- to 15-min sample
 2      duration) that measures the light attenuation of aerosol particles collected on a filter tape (Hansen
 3      etal., 1984). It is also commercially available. The concentration of elemental carbon is derived
 4      from the light absorption measured on a filter using an estimate of the specific absorption (m2/g)
 5      of elemental carbon on the filter; the specific absorption value is derived from laboratory and
 6      atmospheric tests and is specified by the manufacturer. The specific absorption value could be
 7      expected to vary with location, season, and source mix. Comparisons in atmospheric
 8      experiments at some locations with EC values measured by thermal methods confirm that the
 9     aethalometer provides a statistically meaningful estimate of EC concentration (Allen et al.,
10     1999b; Liousse et al., 1993). For instance, Allen et al. (1999b) found the following statistical
11      relationship for Uniontown, PA, during summer 1990: Black Carbon (aethaometer) = 0.95*EC
12     (thermal) - 0.2 (r2 = 0.925, n not specified but appears to be >50, EC range from 0 to 9 Mg/m3).
13           The most recent semi-continuous carbon method is currently being tested by Dr. Susanne
14     Hering (unpublished). This is a flash volatilization method in which particles are impacted on a
15     surface and flash volatilized. Higher collection efficiencies are obtained for smaller particles by
16     growing the particles by humidification prior to impaction.  This device was first demonstrated at
17     the Atlanta Supersite.
18
19     2.2.5.9  Determination of Aerosol Surf ace Area in Real Time
20           Aerosol surface area is an important aerosol property for health effects research.  However,
21     methods for on-line measurement of surface area are not widely available. Woo et al. (2000)
22     used three  continuous aerosol sensors to determine aerosol surface area. They used a
23     condensation particle counter (CPC, TSI, Inc., Model 3020), an aerosol mass concentration
24     monitor (MCM, TSI, Inc., Model 8520), and an electrical aerosol detector for measuring particle
25     charge concentration (BAD, TSI, Inc., Model 3070).  The three sensor signals were inverted to
26     obtain the  aerosol size distribution using a lognormal size distribution model by minimizing the
27     difference  between the measured signals and the theoretical values based upon a size distribution
28     model, the instrument calibration, and its theoretical responses. The lognormal function was then
29      integrated  to calculate the total surface area concentration. Woo et al. demonstrated that this
30     method can give near real-time measurements of aerosol surface area.
31
        March 2001
2-96
DRAFT-DO NOT QUOTE OR CITE


-------
   1
   2
   3
  4
   5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
 2.2.6  Data Quality
      Although much of the recent work in instrument development for airborne particulate
 matter measurement has focused on addressing sampling artifacts associated with loss or gain of
 semivolatile species, or development of real-time measurements of aerosol concentrations, other
 recent efforts have examined issues associated with improving the quality of data being collected.

 2.2.6.1  Errors in Gravimetric Analyses
      Issues of precision and accuracy associated with gravimetric analyses have been examined
 in several studies. It is well known that weighing of particle sampler filters is subject to
 fluctuations of environmental conditions in the weighing. Gravimetric analysis issues reviewed
 by Allen et al. (1999c) include proper temperature and humidity controls, use of a high quality
 microbalance, 100% replicate weighings, control of static charge, aging of new filters, weighing
 of a sufficient number of laboratory blank filters, and accounting for buoyancy errors caused by
 variability in barometric pressure. Lawless and Rodes (1999) investigated the magnitude of
 uncertainties attributed to fluctuations in some of these parameters (humidity, temperature,
 drafts, vibration, and electrostatic charges) and recommended methods for improving their
 control. They noted that the role of humidity control in the laboratory did not seem to be as
 critical of a factor as the humidity under which the sample was collected.  Koistinen et al. (1999)
 give an excellant discussion of the procedures developed to overcome problems associated with
 gravimetric measurements of PM25 mass in the EXPOLIS Study.  They discuss factors such as
 corrections for buoyancy, elimination of static charge, and increases in the mass of blank filters
 with time.
     Mass concentration measurements of coarse particulate matter are inherently less precise
 than the corresponding PM2 5 or PM10 measurements (Allen et al.,  1999c). Coarse particulate
 mass concentrations are determined either by difference between collocated PM10 and PM25
 samplers or more directly by use of a dichotomous sampler.  The difference method suffers from
 errors because of the use of two independent measurements.  The dichotomous sampler also has
potential errors caused by postexposure loss of particles from unoiled filters and uncertainties in
the coarse mass channel enrichment factor. Allen et al. (1999c) summarized several sampling
issues to consider in measuring coarse particulate mass, including the use of identical
instrumentation (except cutpoints) such as filter media, filter face velocity, and ambient-filter
       March 2001
                                                 2-97
                                                     DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
 30
 31
temperature differences; common flow measurement devices; use of higher sampler flow rates
(10 L/min minimum for 24-h sample is recommended); avoiding excessive filter loading; and
full characterization of the cutpoint characteristics of the FRM without the PM2 5 WINS inlet.

2.2.6.2  Quality Assurance
     EPA has undertaken extensive studies to evaluate the quality of the 1999 PM25 FRM data
including precision and bias statistics, data reporting statistics, etc. Following a peer review,
EPA will issue a report that documents the quality assurance (QA) activities that were
undertaken for the PM25 environmental data operations for the calendar year January 1, 1999, to
December 31,1999, which was the first year of implementation the PM2 5 monitoring program.
The QA report will evaluate the adherence to the quality assurance requirements .described in
40 CFR Part 58 Appendix A (Code of Federal Regulations, 1999c) and evaluate the data quality
indicators of precision, accuracy/bias, completeness, comparability, and detectability.  The report
also will provide conclusions and recommendations for future improvements.
 2.3  SUMMARY
      Atmospheric particles originate from a variety of sources and possess a range of
 morphological, chemical, physical, and thermodynamic properties. The composition and
 behavior of airborne particles are linked with those of the surrounding gas. Aerosol may be
 defined as a suspension of solid or liquid particles in air and includes both the particles and all
 vapor or gas phase components of air. However, the term aerosol often is used to refer to the
 suspended particles only. Particulate is an adjective and should only be used as a modifier, as in
 particulate matter.
      A complete description of the atmospheric aerosol would include an accounting of the
 chemical composition, morphology, and size of each particle and the relative abundance of each
 particle type as a function of particle size.  Recent developments in single particle analysis
 techniques are bringing such a description closer to reality.
      The diameter of a spherical particle may be determined geometrically, from optical or
 electron microscopy; by light scattering and Mie theory; or by its behavior, such as its electrical
 mobility or its aerodynamic behavior. However, the various types of diameters may be different,
        March 2001
                                          2-98
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
26
27
28
29
30
 and atmospheric particles often are not spherical. Therefore, particle diameters are described by
 an "equivalent" diameter. Aerodynamic diameter (i.e., the diameter of a unit density sphere that
 would have the same terminal settling velocity as the real particle) is the most widely used
 equivalent diameter. Therefore, in this document, particle diameters, unless otherwise indicated,
 refer to the aerodynamic diameter.
      Atmospheric size distributions show that most atmospheric particles are quite small, below
 0.1 /^m, whereas most of the particle volume (and therefore most of the mass) is found in
 particles greater than 0.1 //m. An important feature of the mass or volume size distributions of
 atmospheric aerosols is their multimodal nature. Volume distributions, measured in ambient air
 in the United States, almost always are found to be bimodal, with a minimum between 1.0 and
 3.0 /^m. The distribution of particles that are mostly larger than the minimum is termed the
 coarse mode. The distribution of particles that are mostly smaller than the minimum is termed
 the fine mode.  Fine-mode particles include both the accumulation mode and the nuclei mode.
 Accumulation-mode particles are that portion of the fine particle fraction with diameters above
 about 0.1 Aim. The nuclei mode, that portion of the fine particle fraction with diameters below
 about 0.1 [j.m, can be observed as a separate mode in mass or volume distributions only in clean
 or remote areas or near sources of new particle formation by nucleation.  Toxicologists and
 epidemiologists use ultrafme to refer to particles in the nuclei-mode size range. Aerosol
 physicists and material scientists tend to use nanoparticles to refer to particles generated in the
 laboratory in this size range.
     The aerosol community uses four different approaches or conventions in the classification
 of particles by size:  (1) modes, based on the observed size distributions and formation
 mechanisms; (2) cut point, usually based on the 50% cut point of the specific sampling device
 (i.e., the particle size at which 50% of the particles enter and 50% of the particles  are rejected);
 (3) dosimetry or occupational sizes, based on the entrance into various compartments of the
respiratory system; and (4) legally specified, regulatory sizes for air quality standards. Over the
years, the terms fine and coarse,  as applied to particle sizes, have lost the original precise
meaning of fine mode and coarse mode, hi any given article, therefore, the meaning of fine and
coarse, unless defined, must be inferred from the author's usage. Li particular, PM25 and
fine-mode particles are not equivalent.  In this document, the term "mode" is used with fine and
       March 2001
                                         2-99
DRAFT-DO NOT QUOTE OR CITE

-------
 1      coarse when it is desired to specify the distribution of fine-mode particles or coarse-mode
 2      particles as shown in Figures 2-4 and 2-5.
 3           Several processes influence the formation and growth of particles. New particles may be
 4      formed by nucleation from gas phase material. Particles may grow by condensation as gas phase
 5      material condenses onto existing particles. Particles may also grow by coagulation as two
 6      particles combine to form one. Gas phase material condenses preferentially on smaller particles
 7      and the rate constant for coagulation of two particles decreases as the particle size increases.
 8      Therefore, nuclei mode particles grow into the accumulation mode but accumulation mode
 9      particles do not grow into the coarse mode.
10          The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen
11      ions; particle-bound water; elemental carbon; a great variety of organic compounds; and crustal
12     material. Atmospheric PM contains a large number of elements in various compounds and
13     concentrations and hundreds to thousands of specific organic compounds.  Particulate material
14     can be primary or secondary.  PM is called primary if it is in the same chemical form in which it
15     was emitted into the atmosphere. PM is called secondary if it is formed by chemical reactions in
16     the atmosphere.  Primary coarse particles are usually formed by mechanical processes. Primary
17     fine particles are emitted from sources, either directly as particles or as vapors that rapidly
18     condense to form particles.
19          Most of the sulfate and nitrate and a portion of the organic compounds in atmospheric
20     particles are secondary (i.e., they are formed by chemical reactions in the atmosphere).
21     Secondary aerosol formation depends on numerous factors including the concentrations  of
22     precursors; the concentrations of other gaseous reactive species such as ozone, hydroxyl radical,
23     peroxy radicals, or hydrogen peroxide; atmospheric conditions, including solar radiation and
24     relative humidity; and the interactions of precursors and preexisting particles within cloud or fog
25     droplets or on or in the liquid film on solid particles. As a result, it is considerably more difficult
26     to relate ambient concentrations of secondary species to sources of precursor emissions than it is
27     to identify the sources of primary particles.
28           The lifetimes of particles vary with particle size. Coarse particles can settle rapidly from
29     the atmosphere within minutes or hours, and normally travel only short distances.  However,
30     when mixed high into the atmosphere, as in dust storms, the smaller-sized, coarse-mode particles
31     may have longer lives and travel greater distances. Accumulation-mode fine particles are kept
        March 2001
2-100
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
  suspended by normal air motions and have very low deposition rates to surfaces. They can be
  transported thousands of kilometers and remain in the atmosphere for a number of days.
  Accumulation-mode particles are removed from the atmosphere primarily by cloud processes.
  Coarse mode particles of less than * 10-//m diameter as well as accumulation-mode and
  nuclei-mode (or ultrafine) particles all have the ability to penetrate deep into the lungs and be
  removed by deposition in the lungs. Dry deposition rates are expressed in terms of a deposition
  velocity that varies with the particle size, reaching a minimum between 0.1 and 1.0 /j.m
  aerodynamic diameter.
      The role of particles in reducing visibility and affecting radiative balance through scattering
 and absorption of light is evident as are the effects of particles in soiling and damaging materials.
 EPA addresses visibility effects through regional haze regulations. The direct effects of particles
 in scattering and absorbing light and the indirect effects of particles on clouds are being
 addressed in climate change programs in several government agencies.
      The role of PM in acid deposition has not always been recognized.  Acid deposition and
 PM are intimately related, however, first,  because particles contribute significantly to the
 acidification of rain and, second, because  the gas-phase species that lead to dry deposition of
 acidity are also precursors of particles. Therefore,  reductions in SO2 and NOX emissions will
 decrease both acid deposition and PM  concentrations. Sulruric acid, ammonium nitrate, and
 organic particles also are deposited on surfaces by  dry deposition.  The utilization of ammonium
 by plants leads to the production of acidity.  Therefore, dry deposition of particles also can
 contribute to the ecological damages caused by acid deposition.
      The decision by the EPA to revise the PM standards by adding daily and yearly standards
 for PM2 5 has led to a renewed interest in the measurement of atmospheric particles and also to a
 better understanding of the problems in developing precise and accurate measurements of
 particles.  Unfortunately, it is very difficult to measure and characterize particles suspended in
 the atmosphere.
     PM monitoring is needed to develop information to guide implementation of standard (i.e.,
by identifying sources of particles; to determine whether or not a standard has been attained; and
to determinate health, ecological, and radiative effects). Federal Reference Methods (FRM)
specify techniques for measuring PM10  and PM25. Particles are collected on filters and mass
concentrations are determined gravimetrically. The PM2 5 FRM sampler consists of a PM10
       March 2001
                                                 2-101
                                                      DRAFT-DO NOT QUOTE OR CITE

-------
 1      inlet/impactor, a PM2.S impactor with an oil-covered impaction substrate to remove particles
 2      larger than 2.5 ^m, and a 47-mm polytetrafluoroethylene (PTFE) filter with a particle collection
 3      efficiency greater than 99.7%. Both techniques provide relatively precise (±10 %) methods for
 4      determining the mass of material remaining on a Teflon filter after equilibration. The goal of a
 5      PM indicator might be to accurately measure the material that exists as a particle in the
 6      atmosphere. However, numerous uncertainties exist as to the relationship between the mass and
 7      composition of material remaining on the filter, as measured by the FRMs, and the mass and
 8      composition of material that exists in the atmosphere as suspended PM. It is currently not
 9      possible to accurately characterize the material that exists as a particle in the atmosphere. There
10     is no reference standard for particles suspended in the atmosphere; there is no accepted way to
11      remove particle-bound water without losing some of the semivolatile components of PM, such as
12     ammonium nitrate and semivolatile organic compounds and particle-bound water. It also is
13     difficult to cleanly separate fine-mode and coarse-mode PM. As a result, EPA defines accuracy
14     for PM measurements in terms of agreement of a candidate sampler with a reference sampler.
15     Therefore, intercomparisons of samplers become very important in determining how well various
16     samplers agree and how various design choices influence what is actually measured.
17           Fine-mode and coarse-mode particles differ not only in size and morphology (e.g., smooth
18     droplets versus rough solid particles), but also in formation mechanisms; sources; and chemical,
19     physical, and biological properties. It is desirable to separate fine-mode PM and coarse-mode
20     PM as cleanly as possible in order to properly allocate health effects to either fine-mode PM or
21      coarse-mode PM and to correctly determine sources by factor analysis or chemical mass balance.
22      In areas with high concentrations of wind-blown soil, the current practice of separating fine- and
23      coarse-mode particles at 2.5-//m AD may not provide the best separation of exposure,
24      epidemiologic, and source apportionment studies. A cut near 1  ^m would provide a good
 25      indicator of fine-mode PM if the air stream could be dehumidified to remove particle-bound
 26      water without evaporating semivolatile components.
 27           Current filtration-based mass measurements lead to significant evaporative losses, during
 28      and possibly after collection, of a variety of semivolatile components (i.e., species that exist in
 29      the atmosphere in dynamic equilibrium between the condensed phase and gas phase).  Important
 30      examples include ammonium nitrate, semivolatile organic compounds, and particle-bound water.
 31      Loss of these components may significantly impact the quality of the measurement, and can lead
        March 2001
2-102
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 to both positive and negative sampling artifacts. Negative artifacts, resulting from loss of
 ammonium nitrate and semivolatile organic compounds, may occur during sampling because of
 changes in temperature, relative humidity, or composition of the aerosol, or because of the
 pressure drop across the filter.  Negative artifacts also may occur during handling and storage
 because of evaporation. Positive artifacts occur when gas-phase compounds (H2O, HNO3, SO2,
 and organic compounds) absorb onto or react with filter media or collected PM, or when some
 particle-bound water is not removed.
      The loss of particulate nitrate may be determined by comparing nitrate collected on a
 Teflon filter to that collected on a nylon filter (which absorbs nitrate) preceded by a denuder to
 remove nitric acid.  In two studies in southern California, the PM2 5 mass lost because of
 volatilization of ammonium nitrate was found to represent 10 to 20% of the total PM2 5 mass and
 almost a third of the nitrate.  Semivolatile organic compounds (SVOCs) similarly can be lost
 from Teflon filters because of volatilization during or after collection. Such losses can cause the
 PM2 5 mass to be significantly underestimated. Positive sampling artifacts also can occur as the
 result of the adsorption of organic gases onto the filter materials. There is a larger positive
 artifact caused by adsorption of organic vapor onto quartz fiber filters than onto Teflon filters.
 Denuder-based sampling systems also have been developed for measuring particulate phase
 organic compounds.  This technique is an improvement over the filter/adsorbent collection
 method.  In most denuder systems, a denuder that removes gas-phase absorbable organic gases is
 followed by a filter pack. The first filter collects particles. It is followed by a charcoal-
 impregnated glass-fiber filter that absorbs semivolatile material that evaporates from particles on
 the front filter. The FRM for PM2 5 will suffer loss of particulate nitrates and SVOC, similar to
 the losses experienced with other single filter collection systems.
      It is generally desirable to collect and measure ammonium nitrate and semivolatile organic
 compounds as part of particulate matter mass. However, it is usually desirable to remove the
particle-bound water before determining the mass. In some situations it may be important to
know how much of the suspended particle's mass or volume results from particle-bound water.
Calculations and measurements indicate that aerosol water content is strongly dependent on
composition, but that liquid water can represent a significant mass fraction of the aerosol
concentration at relative humidities above 60%.  Relative humidity may affect particle size,
growth, and other properties. Accumulation-mode particles are usually hygroscopic. The more
       March 2001
                                         2-103
                                                             DRAFT-DO NOT QUOTE OR CITE

-------
 1      hygroscopic particles tend to contain more sulfates, nitrates, and secondary organic compounds,
 2      while the less hygroscopic particles tend to contain more elemental carbon, primary organic
 3      compounds, and crustal components.  Fresh, submicron-size soot particles may to shrink with
 4      increasing relative humidity because of a structural change. The effects of relative humidity on
 5      the sorption of SVOC on particles are not well understood. The amount of water sorbed to an
 6      atmospheric aerosol may be affected by the presence of an organic film on the particle, which
 7      may impede the transport of water across the surface.
 8           In addition to FRM sampling of equilibrated mass to determine compliance with PM
 9     standards, EPA requires states to conduct speciation sampling primarily to determine source
10     categories and trends.  The current speciation samplers include three filters: (1) Teflon for
11      equilibrated mass and elemental analysis, (2) a Nylon filter with a nitric acid denuder to collect
12     nitrate, and (3) a quartz fiber filter for elemental and organic carbon (but without any correction
13     for positive or negative artifacts because of adsorption of volatile organic compounds on the
14     quartz filters or evaporation of semivolatile organic compounds from the collected particles).
15     The IMPROVE network of samplers provides four 24-h integrated filter samples: .(1) a PM10
16     filter and (2) a PM2.5 Teflon filter for gravimetric determination of mass and for analysis of heavy
17     elements by X-ray fluorescence; (3) a Nylon filter, preceded by a nitric acid denuder, for artifact-
18     free determination of nitrate and measurement of other ionic species by ion chromatography; and
19     (4) a quartz filter for measurement of elemental carbon (EC) and organic carbon (OC) by thermal
20     optical analysis. The EC/OC measurement method utilized in the IMPROVE network is based
21     on optical correction of pyrolytic char using optical reflectance, whereas the EC/OC method
22     specified hi the NIOSH method 5040 (for diesel soot) is based on optical transmission for
23     correction for pyrolytic char. These methods also differ in their temperature profiles. The two
24     methods agree on total carbon but differ in the split of total carbon into EC and OC.
25           The EPA expects that more than 200 local agency monitoring sites throughout the states
26     will operate continuous PM monitors. However, EPA has not yet provided any guidance
27     regarding appropriate  continuous monitoring techniques. All currently available techniques for
28     continuous measurements of suspended particle mass, such as the integrating nephelometer, the
29     beta-absorption monitor, and the Tapered Element Oscillating Microbalance (TEOM) share the
30     problem of dealing with semivolatile PM components (i.e., so as not to include particle-bound
31     water as part of the mass, the particle-bound water must be removed by heating or
        March 2001
2-104
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
 dehumidification).  However, heating also causes ammonium nitrate and semivolatile organic
 compounds to evaporate. The TEOM monitor operates at a constant, but higher than ambient,
 temperature to remove particle-bound water. However, the FRM is required to operate at no
 more than 5 °C above the ambient temperature. This philosophical difference in operation leads
 to differences between the TEOM and integrated mass concentrations for both PM10 and PM2 5.
      Several candidates for continuous PM mass measurements are currently being field tested.
 The Real-Time Total Ambient Mass Sampler (RAMS) measures the total mass of collected
 particles, including semivolatile species with a TEOM monitor using a "sandwich filter". The
 sandwich contains a Teflon-coated particle-collection filter followed by a charcoal-impregnated
 filter to collect any semivolatile species lost from the particles during sampling. The RAMS uses
 a Nafion dryer to remove particle-bound water from the suspended particles and a particle
 concentrator to reduce the quantity of gas phase organic compounds that must be removed by the
 denuder. The Continuous Ambient Mass  Monitor (CAMM) estimates ambient particulate matter
 mass by measurement of the increase in the pressure drop across a membrane filter caused by
 particle loading.  It also uses a Nafion dryer to remove particle-bound water.  In addition to
 continuous mass measurement, a number of techniques for continuous measurement of sulfate,
 nitrate, or elements are being tested.
      Aerosol time-of-flight mass spectroscopy (ATOFMS) provides a new technique for
 real-time measurement of correlated size and composition profiles of individual atmospheric
 aerosol particles. Measurements are made in situ by combining a dual-laser aerodynamic particle
 sizing system to size and track individual particles through the instrument and laser
 desorption/ionization time-of-flight mass spectrometry to obtain correlated single particle
 composition data. ATOFMS technology is able to size, count, and provide chemical composition
on individual particles ranging in size from 10 nm to 2 ju.m.  However, there is still controversy
over the calibration of such techniques.
       March 2001
                                               2-105
                                                    DRAFT-DO NOT QUOTE OR CITE

-------
 1      REFERENCES

 2      Allen, G.; Sioutas, C.; Koutrakis, P.; Reiss, R.; Lurmann, F. W.; Roberts, P. T. (1997) Evaluation of the TEOM
 3            method for measurement of ambient paniculate mass in urban areas. J. Air Waste Manage. Assoc.
 4            47:682-689.
 5      Allen, J. O.; Paschkewitz, J. S.; Plummer, E. F.; Lafleur, A. L.; Sarofim, A. F.; Smith, K. A. (1999a) Absorption of
 6            semi-volatile compounds in oiled impaction substrates: measurement of pyrene absorption. Aerosol Sci.
 7            Technol. 30:16-29..
 8      Allen, G. A.; Lawrence, J.; Koutrakis, P. (1999b) Field validation of a semi-continuous method for aerosol black
 9            ' carbon (aethalometer) and temporal patterns of summertime hourly black carbon measurements in
10            southwestern PA. Atmos. Environ. 33: 817-823.
11      Allen, G. A.; Oh, J. A.; Koutrakis, P. (1999c) Techniques for high quality ambient coarse particulate mass
12            ' measurements. J. Air Waste Manage. Assoc. 49:  133-141.
13      Altshuller, A. P. (1983) Review: natural volatile organic substances and their effect on air quality in the United
14            States. Atmos. Environ. 17:2131-2165.                                                         .
15      American Conference of Governmental Industrial Hygienists. (1994) 1994-1995 threshold limit values for chemical
1 g            substances and physical agents and biological exposure indices. Cincinnati, OH: American Conference of
17            Governmental Industrial Hygienists.
18      Anand, N. K.; McFarland, A. R.; Kihm, K. D.; Wong, F. S. (1992) Optimization of aerosol penetration through
19            transport lines. Aerosol Sci. Technol. 16: 105-112.
20      Andrews, E.; Larson, S. M. (1993) Effect of surfactant layers on the size changes of aerosol particles as a function
21             of relative humidity. Environ. Sci. Technol. 27: 857-865.
22       Ansari, A. S.; Pandis, S. N. (1998) Response of inorganic PM to precursor concentrations. Environ. Sci. Technol.
23             32:2706-2714.
24       Appel,  B. R. (1993) Atmospheric sample analysis and sampling artifacts. In: Willeke, K.; Baron, P. A., eds. Aerosol
25             measurement: principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 233-259.
26       Appel,  B. R.; Toldwa, Y.; Haik, M.  (1981) Sampling of nitrates in ambient air. Atmos. Environ. 15: 283-289.
27       Appel', B. R.; Toldwa, Y.; Haik, M.; Kothny, E. L. (1984) Artifact particulate sulfate and nitrate formation on filter
28             media. Atmos. Environ. 18:409-416.
29       Aumont, B.; Madronich, S.; Bey, I.; Tyndall, G. S. (2000) Contribution of secondary VOC to the composition of
30            aqueous atmospheric particles: a modeling approach. J. Atmos. Chem. 35: 59-75.
31       Ayers,  G. P.; Keywood, M. D.; Gras, J. L. (1999) TEOM vs. manual gravimetric methods for determination of
32            PMjj aerosol mass concentrations. Atmos. Environ. 33:3717-3721.
33      Babich, P.; Davey, M.; Allen, G.; Koutrakis, P. (2000) Method comparisons for particulate nitrate, elemental
34            carbon, and PM2 5 mass in seven U.S. cities. J. Air Waste Manage. Assoc. 50: 1095-1105.
35      Benner, B. A., Jr.; Wise,  S. A.; Currie, L. A.; Klouda, G. A.; Klinedinst, D. B.; Zweidinger, R. B.; Stevens, R. K.;
35            Lewis, C. W.  (1995) Distinguishing the contributions of residential wood combustion and mobile source
37            emissions using relative concentrations of dimethylphenanthrene isomers. Environ. Sci. Technol.
38            29:2382-2389.
 39      Bergin, M. H.; Ogren, J. A.; Schwartz, S. E.; Mclnnes, L. M. (1997) Evaporation of ammonium nitrate aerosol in a
40            heated nephelometer: implications for field measurements. Environ. Sci. Technol. 31: 2878-2883.
41      Bertoni, G.; Febo, A.; Perrino, C.; Possanzini, M. (1984) Annular active diffuse sampler: a new device for the
 42             collection of organic vapours. Ann. Chim. (Rome) 74: 97-104.
 43      Binkowski, F. S.; Ching, J. K. S. (1995) Modeling fine particulate mass and visibility using the EPA regional
 44             particulate model. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research
 45             and Development, National Exposure Research Laboratory; report no. EPA/600/A-95/143.  Available from:
 46             NTIS, Springfield, VA; PB96-139100.                                                       .
 47      Binkowski, F.  S.; Shankar, U. (1995) The regional particulate matter model. I. Model description and preliminary
 48             results. J. Geophys. Res. [Atmos.] 100:26,191-26,209.
 49      Birch, M. E.; Gary, R. A. (1996) Elemental carbon-based method for occupational monitoring of particulate diesel
 50             exhaust: methodology and exposure issues. Analyst 121: 1183-1190.
 51      Brook, J. R.; Dann, T. F. (1999) Contribution of nitrate and carbonaceous species to PM2.5 observed in Canadian
, 52            'cities. J. Air Waste Manage. Assoc. 49: 193-199.
          March 2001
                                                         2-106
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
46
47
48
49
 50
 51
52
53
54
55
 Brorstrom, E.; Grennfelt, P.; Lindskog, A. (1983) The effect of nitrogen dioxide and ozone on the decomposition of
       particle-associated polycyclic aromatic hydrocarbons during sampling from the atmosphere. Atmos Environ
       17: 601-605.
 Bunz, H.; Koyro, M.; Mohler, O.; Saathoff, H. (1996) Temperature dependence of the redistribution mechanisms
       for volatile components of atmospheric aerosol particles. J. Aerosol Sci. 27(suppl. 1): S653-S654.
 Carson, P. G.; Johnston, M. V.; Wexler, A. S. (1997) Laser desorption/ionization of ultra fine aerosol particles.
       Rapid Commun. Mass Spectrom. 11: 993-996.
 Cachier, H.; Bremond, M.-P.; Buat-Menard, P. (1989) Determination of atmospheric soot carbon with a simple
       thermal method. Tellus 41B: 379-390.
 Castaldi, M. J.; Senkan, S. M. (1998) Real-time, ultrasensitive monitoring of air toxics by laser photoionization
       time-of-flight mass spectroscopy. J. Air Waste Manage. Assoc. 48: 77-81.
 Castro, L. M.; Pio, C. A.; Harrison, R. M.; Smith, D. J. T. (1999) Carbonaceous aerosol in urban and rural
       European atmospheres: estimation of secondary organic carbon concentrations. Atmos. Environ.
       33:2771-2781.
 Chen, B. T. (1993) Instrument calibration. In: Willeke, K.; Baron, P. A., eds. Aerosol measurement: principles,
       techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 493-520.
 Chow, J. C. (1995) Measurement methods to determine compliance with ambient air quality standards for
       suspended particles. J. Air Waste Manage. Assoc. 45:  320-382.
 Chow, J. C.; Watson, J. G.; Pritchett, L. C.; Pierson, W. R.; Frazier, C. A.; Purcell, R. G. (1993) The DRI
       thermal/optical reflectance carbon analysis system: description, evaluation and applications in U.S. air quality
       studies. In: Lodge, J. P., Jr. Fourth international conference on carbonaceous particles in the atmosphere;
       April 1991; Vienna, Austria. Atmos. Environ. Part A 27: 1185-1201.
 Chow, J. C.; Fujita, E. M.; Watson, J. G.; Lu, Z.; Lawson, D. R.; Ashbaugh, L. L. (1994) Evaluation of filter-based
       aerosol measurements during the 1987 Southern California Air Quality Study. Environ. Monit. Assess
       30: 49-80.
 Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Countess, R. J. (1996) Sources and chemistry of PM10 aerosol in
       Santa Barbara County, CA. Atmos. Environ. 30: 1489-1499.
 Chow, J. C.; Watson, J. G.; Crow, D.; Lowenthal, D. H.; Merrifield, T. (2000) Comparison of IMPROVE and
       NIOSH carbon measurements. Aerosol Sci. Technol.: accepted.
 Christoffersen, T. S.; Hjorth, J.; Hone, O.; Jensen, N. R.; Kotzias, D.; Molander, L. L.; Neeb, P.; Ruppert, L.;
       Winterhalter, R.; Virkkula, A.; Wirtz, K.; larsen, B. R. (1998) cis-Pinic acid, a possible precursor for organic
       aerosol formation from ozonolysis of "alpha"-pinerie. Atmos. Environ. 32: 1657-1661.
 Claiborn, C. S.; Finn, D.; Larson, T. V.; Koenig, J. Q. (2000) Windblown dust contributes to high PM25
       concentrations. J. Air Waste Manage. Assoc. 50: 1440-1445.
 Clark, W. E.; Whitby, K. T. (1975) Measurements of aerosols produced by the photochemical oxidation of SO, in
       air. J. Colloid Interface Sci. 51: 477-490.
 Clarke, A. D. (1992) Atmospheric nuclei in the remote free-troposphere. J. Atmos. Chem. 14:  479-488.
 Code of Federal Regulations. (1991a) Appendix J to Part 50-reference method for the determination of paniculate
       matter as PM10 in the atmosphere. C. F. R. 40: "section" 50.
 Code of Federal Regulations. (1991b) Ambient air monitoring reference and equivalent methods.  C. F. R.
       40: sect. 53.
 Code of Federal Regulations. (1999a) Appendix L to Part 50~reference method for the determination of fine
      particulate matter as PM25 in the atmosphere. C. F. R. 40: "section"50.
 Code of Federal Regulations. (1999b) Part 53-ambient air monitoring reference and  equivalent methods  C F R
      40: "sections"53, 58.
 Code of Federal Regulations. (1999c) Part 58-ambient air quality surveillance for particulate matter: final rule
      C. F. R. 40: "section"58.
 Gotham, W. E.; Bidleman, T. F. (1992) Laboratory investigations of the partitioning of organochlorine compounds
      between the gas phase and atmospheric aerosols on glass fiber filters. Environ. Sci. Technol. 26:469-478.
Countess, R. J. (1990) Intel-laboratory analyses of carbonaceous aerosol samples. Aerosol Sci. Technol
       12:114-121.
Courtney, W. J.; Shaw, R. W.; Dzubay, T. G. (1982) Precision and accuracy of a "beta" gauge for aerosol mass
      determinations. Environ. Sci. Technol. 16: 236-239.
Coutant, R. W.; Callahan, P. J.; Chuang, J. C.; Lewis, R. G. (1992) Efficiency of silicone-grease-coated denuders
      for collection of polynuclear aromatic hydrocarbons. Atmos. Environ. Part A 26: 2831-2834.
         March 2001
                                                2-107
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Covert, D. S.; Kapustin, V. N.; Quinn, P. K.; Bates, T. S. (1992) New particle formation in the marine boundary
 2             layer. J. Geophys. Res. [Atmos.] 97: 20,581 -20,589.
 3       Cui, W.; Machir, J.; Lewis, L.; Eatough, D. J.; Eatough, N. L. (1997) Fine particulate organic material at Meadview
 4             during the project MOHAVE summer intensive study. J. Air Waste Manage. Assoc. 47: 357-369.
 5       Cui, W.; Eatough, D. J.; Eatough, N. L. (1998) Fine particulate organic material in the Los Angeles Basin -1:
 6             assessment of the high-volume Brigham Young University organic sampling system, BIG BOSS. J. Air
 7             Waste Manage. Assoc. 48:1024-1037.
 8       Davis, C. S.; Fellin, P.; Otson, R. (1987) A review of sampling methods for polyaromatic hydrocarbons in air.
 9             JAPCA 37: 1397-1408.
10       Ding, Y.; Lee, M. L.; Eatough, D. J. (1998a) The determination of total nitrite and N-nitroso compounds in
11             atmospheric samples. Int. J. Environ. Anal. Chem. 69: 243-255.
12       Ding, Y.; Cui, W.; Lee, M. L.; Eatough, D. J. (1998b) Fine particulate N-nitroso and nitrite organic compounds in
13             the atmosphere. Appl. Occup. Environ. Hyg. 13: 432-438.
14       Eatough, D. J. (1999) BOSS, the Brigham Young University Organic Sampling System: determination of particulate
15             carbonaceous material using diffision denuder sampling technology. In:  Lane, D., ed. Gas/phase partition
16             measurements of atmospheric organic compounds. New York, NY: Gordon and Breach Science Publishers;
17             in press.
18       Eatough, D. J.; Wadsworth, A.; Eatough, D. A.; Crawford, J. W.; Hansen, L. D.; Lewis, E. A. (1993)
19             A multiple-system, multi-channel diffusion denuder sampler for the determination of fine-particulate organic
20             material in the atmosphere. Atmos. Environ. Part A 27:1213-1219.
21       Eatough, D. J.; Tang, H.; Cui, W.; Machir, J. (1995) Determination of the size distribution and chemical
22             composition of fine particulate semi-volatile organic material in urban environments using diffusion denuder
23             technology. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on particulate air pollution
24             and human mortality and morbidity, part II; Janaury 1994; Irvine, CA. Inhalation Toxicol. 7: 691-710.
25       Eatough, D. J.; Eatough, D. A.; Lewis, L.; Lewis, E. A. (1996) Fine particulate chemical composition and light
26             extinction at Canyonlands National Park using organic particulate material concentrations obtained with a
27             multisystem, multichannel diffusion denuder sampler. J. Geophys. Res. [Atmos.] 101: 19,515-19,531.
28       Eatough, D. J.; Obeidi, F.; Pang, Y.; Ding, Y.; Eatough, N. L.; Wilson, W. E. (1999a) Integrated and real-time
29             diffusion denuder sampler for PM2 5. Atmos. Environ. 33: 2835-2844.
30       Eatough, D. J.; Pang, Y.; Eatough, N. L. (1999b) Determination of PM25 sulfate and nitrate with a PC-BOSS
31             designed for routine sampling for semi-volatile particulate matter. J. Air Waste Manage. Assoc. 49: 69-75.
32       Falconer, R. L.; Bidleman, T. F.; Gotham, W. E. (1995) Preferential sorption of non- and
33             mono-ortho-polychlorinated biphenyls to urban aerosols. Environ. Sci. Technol. 29:1666-1673.
34       Fang, C. P.; McMurry, P. H.; Marple, V. A.; Rubow, K. L. (1991) Effect of flow-induced relative humidity changes
35             on size cuts for sulfuric acid droplets in the microorifice uniform deposit impactor (MOUDI). Aerosol Sci.
36             Technol. 14:266-277.
37       Faust, B. C. (1994) Photochemistry of clouds, fogs, and aerosols. Environ. Sci. Technol. 28: 217A-222A.
38       Federal Register. (1987) Revisions to the national ambient air quality standards for particulate matter. F. R. (July 1)
39             52:24,634-24,669.
40       Federal Register. (1997) National ambient air quality standards for particulate matter; final rule. F. R. (July 18)
41             62:38,652-38,752.
42      Federal Register. (1998) National ambient air quality standards for particulate matter and revised requirements for
43             designation of reference and equivalent methods for PM2 5 and ambient air quality surveillance for particulate
44            matter; correction. F. R. (February 17) 63: 7710-7718.
45       Finn, D.; Rumburg, B.; Claiborn, C.; Bamesberger, W.; Norris, G.; Koenig, J.; Larson, T. (2000) Sampling artifacts
46            from the use of Teflon filters and denuder tubes during the measurement of airborne carbonaceous particulate
47            matter. J. Air Waste Manage. Assoc.: submitted.
48       Fitz, D. R. (1990) Reduction of the positive organic artifact on quartz filters. Aerosol. Sci. Technol. 12: 142-148.
49      Forstner, J. H. L.; Flagan, R. C.; Seinfeld, J. H. (1997) Molecular speciation of secondary organic aerosol from
50            photooxidation of the higher alkenes: 1-octene and 1-decene. Atmos. Environ. 31: 1953-1964.
51       Forstner, J. H. L.; Flagan, R. C.; Seinfeld, J. H. (1997) Secondary organic aerosol from the photooxidation of
52            aromatic hydrocarbons: molecular composition. Environ. Sci. Technol. 31:1345-1358.
53       Friedlander, S. K. (1970) The characterization of aerosols distributed with respect to size and chemical composition.
54            J. Aerosol Sci. 1: 295-307.
         March 2001
2-108
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
49
 50
 51
 52
53
54
55
  Friedlander, S. K. (1977) Smoke, dust and haze: fundamentals of aerosol behavior. New York, NY: John Wiley &
        Sons.
  Friedlander, S. K.; Yeh, E. K. (1998) The submicron atmospheric aerosol as a carrier of reactive chemical species:
        case of peroxides. Appl. Occup. Environ. Hyg. 13: 416-420.
  Fuchs, N. A. (1964) The mechanics of aerosols. New York, NY: Pergamon Press.
  Fuchs, N. A. (1989) The mechanics of aerosols. New York, NY: Dover Publications.
  Gallorini, M. (1995) Trace element monitoring in atmospheric pollution processes by neutron activation analysis
        Microchem. J. 51: 127-137.
  Card, E.; Mayer, J. E.; Morrical, B. D.; Dienes, T.; Fergenson, D. P.; Prather, K. A. (1997) Real-time analysis of
        individual atmospheric aerosol particles: design and performance of a portable ATOFMS Anal Chem
        69:4083-4091.
  Ge, Z.; Wexler, A. S.; Johnston, M. V. (1998) Laser desorption/ionization of single ultrafine multicomponent
       aerosols. Environ. Sci. Technol. 32: 3218-3223.
  Glasius, M.; Lahaniati, M.; Calogirou, A.; Di Bella, D.; Jensen, N. R.; Hjorth, J.; Kotzias, D.; Larsen, B. R. (2000)
       .Carboxylic acids in secondary aerosols from oxidation of cyclic monoterpenes by ozone Environ Sci
       Technol. 34: 1001-1010.
 Goss, K.-U.; Eisenreich, S. J. (1997) Sorption of volatile organic compounds to particles from a combustion source
       at different temperatures and relative humidities. Atmos. Environ. 31: 2827-2834.
 Graedel, T. E.; Goldberg, K. I. (1983) Kinetic studies of raindrop chemistry: 1. inorganic and organic processes.
       J. Geophys. Res. C: Oceans Atmos. 88: 10,865-10,882.
 Gray, H. A.; Cass, G. R.; Huntzicker, J. J.; Heyerdahl, E. K.; Rau, J. A. (1984) Elemental and organic carbon
       particle concentrations: a long-term perspective. Sci. Total Environ. 36: 17-25.
 Griffin, R. J.; Cocker, D. R., Ill; Seinfeld, J. H. (1999) Incremental aerosol reactivity: application to aromatic and
       biogenic hydrocarbons. Environ. Sci. Technol. 33: 2403-2408.
 Grosjean, D. (1983) Polycyclic aromatic hydrocarbons in Los Angeles air from samples collected on Teflon glass
       and quartz filters. Atmos. Environ. 17: 2565-2573.
 Grosjean, D. (1992) In situ organic aerosol formation during a smog episode: estimated production and chemical
       functionality. Atmos. Environ. Part A 26: 953-963.
 Grosjean, D.; Parmar, S. S. (1990) Interferences from aldehydes and peroxyacetyl nitrate when sampling urban air
       organic acids on alkaline traps. Environ. Sci. Technol. 24: 1021-1026.
 Grosjean, D.; Seinfeld, J. H. (1989) Parameterization of the formation potential of secondary organic aerosols
       Atmos. Environ. 23:  1733-1747.
 Grosjean, D.; Williams, E. L., II; Grosjean, E.; Novakov,  T. (1994) Evolved gas analysis of secondary organic
    L  aerosols. Aerosol Sci. Technol. 21: 306-324.
 Gundel, L. A.; Lane, D. A. (1999) Sorbent-coated diffusion denuders for direct measurement of gas/particle
       partitioning by semi-volatile organic compounds. In: Lane, D. A., ed. Gas and particle phase measurements
       of atmospheric organic compounds. Amsterdam, The Netherlands: Gordon and Breach Science Publishers.
       (Advances  in environmental, industrial and process control technologies: v. 2).
 Gundel, L. A.; Lee, V. C.; Mahanama, K. R. R.; Stevens, R. K.; Daisey, J. M. (1995) Direct determination of the
       phase distributions of semi-volatile polycyclic aromatic hydrocarbons using annular denuders Atmos
       Environ. 29: 1719-1733.
 Haagen-Smit, A. J. (1952) Chemistry and physiology of Los Angeles smog. Ind. Eng. Chem. 44:  1342-1346
 Hansen, A. D. A.; McMurry, P. H. (1990) An intercomparison of measurements of aerosol elemental carbon during
      the 1986 carbonaceous species method comparison study. J. Air Waste Manage. Assoc. 40: 894-895
 Hansen, A. D. A.; Rosen, H.; Novakov, T. (1984) The aethalometer - an instrument for the real-time measurement
      of optical absorption by aerosol particles. Sci. Total Environ. 36: 191-196.
 Henry, R. C.; Lewis, C. W.; Hopke, P. K.; Williamson, H. J. (1984) Review of receptor model fundamentals
      Atmos. Environ. 18: 1507-1515.
Hering, S.; Cass, G. (1999) The magnitude of bias in the measurement of PM25 arising from volatilization of
      particulate nitrate from Teflon filters. J. Air Waste Manage. Assoc. 49: 725-733.
Hidy, G. M.; Venkataraman, C. (1996) The chemical mass balance method for estimating atmospheric particle
      sources in southern California. Chem. Eng. Commun. 151: 187-209.
Hildemann, L. M.; Rogge, W. F.; Cass, G. R.; Mazurek, M. A.; Simoneit, B. R. T. (1996) Contribution of primary
      aerosol emissions from vegetation-derived sources to fine particle concentrations in Los Angeles
      J. Geophys. Res. [Atmos.] 101: 19,541-19,549.
         March 2001
                                                         2-109
                                                              DRAFT-DO NOT QUOTE OR CITE

-------
 1      Hinds, W. C. (1999) Aerosol technology: properties, behavior, and measurement of airborne particles. 2nd ed.
 2            New York, NY: John Wiley & Sons.
 3      Hinz, K.-P.; Kaufmann, R.; Spengler, B. (1994) Laser-induced mass analysis of single particles in airborne state.
 4            Anal. Chem. 66:2071-2076.
 5      Hitzenberger, R.; Dusek, U.; Berner, A. (1996) Black carbon measurements using an integrating sphere. J. Geophys.
 6            Res. [Atmos.] 101:19,601-19,606.
 7      Hitzenberger, R.; Bemer, A.; Dusek, U.; Alabashi, R. (1997) Humidity-dependent growth of size-segregated aerosol
 8            samples. Aerosol Sci. Technol. 27: 116-130.
 9      Hitzenberger, R.; Jennings, S. G.; Larson, S. M.; Dillner, A.; Cachier, H.; Galambos, Z.; Roue, A.; Spain, T. G.
JO            (1999) Intercomparison of measurement methods for black carbon aerosols. Atmos. Environ. 33: 2823-2833.
11      Hoffmann, T.; Odum, J. R.; Bowman, F.; Collins, D.; Klockow, D.; Flagan, R. C.; Seinfeld, J. H. (1997) Formation
12            of organic aerosols from the oxidation of biogenic hydrocarbons. J. Atmos. Chem. 26:189-222.
13      Izumi, K.;  Fukuyama, T. (1990) Photochemical aerosol formation from aromatic hydrocarbons in the presence of
14            NOx. Atmos. Environ. Part A 24:1433-1441.
15      Jaklevic, J. M.; Gatti, R. C.; Goulding, F. S.;  Loo, B. W. (1981) A p-gauge method applied to aerosol samples.
16             Environ. Sci. Technol. 15:680-686.
17      Jang, M.; Kamens, R. M. (1998) A thermodynamic approach for modeling partitioning of semivolatile organic
1 g             compounds on atmospheric particulate matter: humidity effects. Environ. Sci. Technol. 32: 1237-1243.
19      Jang, M.; Kamens, R. M. (1999) Newly characterized products and composition of secondary aerosols from the
20             reaction of "alpha"-pinene with ozone. Atmos. Environ. 33: 459-474.
21      Jang, M.; Kamens, R. M.; Leach, K. B.; Strommen, M. R. (1997) A thermodynamic approach using group
22             contribution methods to model the partitioning of semivolatile organic compounds on atmospheric particulate
23             matter. Environ. Sci. Technol. 31: 2805-2811.
24       Jenkins, B. M.; Jones, A. D.; Turn, S. Q.; Williams, R. B. (1996) Particle concentrations, gas-particle partitioning,
25             and species intercorrelations for polycyclic aromatic hydrocarbons (PAH) emitted during biomass burning.
26             Atmos. Environ. 30: 3825-3835.
27       John, W.;  Wall, S. M.; Ondo, J. L. (1988) A new method for nitric acid and nitrate aerosol measurement using the
28             dichotomous sampler. Atmos. Environ. 22:1627-1635.
29       Johnson, M V.; Wexler, A. S. (1995) MS of individual aerosol particles. Anal. Chem. 67: 721A-726A.
30       Kamens, R.; Odum, J.; Fan, Z.-H. (1995) Some observations on times to equilibrium for semivolatile polycyclic
31             aromatic hydrocarbons. Environ. Sci. Technol. 29:43-50.
32       Kamens, R. M.; Coe, D. L. (1997) A large gas-phase stripping device to investigate rates of PAH evaporation from
33             airborne diesel soot particles. Environ. Sci. Technol. 31: 1830-1833.
34      Kamens, R.; Jang, M.; Chien, C.-J.; Leach, K. (1999) Aerosol formation from the reaction of "alpha"-pinene and
35             ozone using a gas-phase kinetics-aerosol partitioning model. Environ.  Sci. Technol. 33: 1430-1438.
36      Kao, A. S.; Friedlander, S. K. (1995) Frequency distributions of PM10 chemical components and their sources.
37            Environ.  Sci. Technol.  29: 19-28.
38      Kavouras, I. G.; Koutrakis, P. (2000) Use of polyurethane foam as the impaction substrate/collection medium in
39            conventional inertial impactors. Aerosol Sci. Technol.: accepted.
40      Kavouras, I. G.; Mihalopoulos, N.; Stephanou, E. G. (1998) Formation of atmospheric particles from organic acids
41             produced by forests. Nature (London) 395: 683-686.
42      Kavouras, I. G.; Mihalopoulus, N.; Stephanou, E. G. (1999a) Formation and gas/particle partitioning of
43            monoterpene photooxidation products over forests. Geophys. Res. Lett. 26: 55-58.
44      Kavouras, I. G.; Mihalopoulus, N.; Stephanou, E. G. (1999b) Secondary organic aerosol formation vs primary
45            organic aerosol emission: in situ evidence for the chemical coupling between monoterpene acidic
46            photooxidation products and new particle formation over forests. Environ. Sci. Technol. 33: 1028-1037.
47      Keeler, G. J.; Pirrone, N. (1996) Atmospheric transport and deposition of trace elements to Lake Erie from urban
48            areas. Water Sci. Technol. 33: 259-265.
49      Keeler, G. J.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Pierson, W. R. (1988) Effect of ambient humidity
 50            on  dichotomous sampler coarse/fine ratios. Atmos. Environ. 22: 1715-1720.
 51      Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
 52            segregated mass concentration data and ultrafme particle number concentrations in urban areas. Atmos.
 53            Environ. 33:2907-2913.
 54      Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (2000) Size distribution and sources of
 55            aerosol in Launceston, Australia, during winter 1997. J. Air Waste Manage. Assoc. 50: 418-427.
          March 2001
2-110
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
 Khalil, M. A. K.; Rasmussen, R. A. (1992) Forest hydrocarbon emissions: relationships between fluxes and ambient
       concentrations. J. Air Waste Manage. Assoc. 42: 810-813.
 Kim, Y. J.; Boatman, J. F.; Gunter, R. L.; Wellman, D. L.; Wilkison, S. W. (1993) Vertical distribution of
       atmospheric aerosol size distribution over south-central New Mexico. Atmos. Environ Part A
       27: 1351-1362.
 Kim, B. M.; Lester, J.;  Tisopulos, L.; Zeldin, M. D. (1999) Nitrate artifacts during PM25 sampling in the South
       Coast Air Basin of California. J. Air Waste Manage. Assoc. 49: 142-153.
 Kim, B.M.; Cassmassi, J.; Hogo, H.; Zeldin, M. D. (2000) Positive organic carbon artifacts on filter medium during
       PM25 sampling in the South Coast Air Basin. Aerosol. Sci. Technol.: submitted.
 Kirchstetter, T. W.; Corrigan, C. E.; Novakov, T. (2000) Laboratory and field investigation of the adsorption of
       gaseous organic compounds onto quartz filters. Atmos. Environ.: Accepted.
 Kittelson, D. B. (1998) Engines and nanoparticles: a review. J. Aerosol Sci. 29: 575-588.
 Klouda, G. A.; Klinedinst, D. B.; Steel, E. B.; Benner, B. A., Jr.; Parish, H. J. (1996) Exploring a method to
       produce an urban dust particle filter standard. J. Aerosol Sci. 27(suppl. 1): S351-S352.
 Koch, S.; Winterhalter, R.; Uherek, E.; Kolloff, A.; Neeb, P.; Moortgat, G. K. (2000) Formation of new particles in
       the gas-phase ozonolysis of monoterpenes. Atmos. Environ. 34:4031-4042.
 Koistinen, K. J.; Kousa, A.; Tenhola, V.; Hanninen, O.;  Jantunen, M. J.; Oglesby, L.; Kuenzli, N.; Georgoulis, L.
       (1999) Fine particle (PM2 5) measurement methodology, quality assurance procedures and pilot results of the
       EXPOLIS study. J. Air Waste Manage. Assoc. 49: 1212-1220.
 Koutrakis, P.; Kelly, B. P. (1993) Equilibrium size of atmospheric aerosol sulfates as a function of particle acidity
       and ambient relative humidity. J. Geophys. Res. [Atmos.] 98: 7141-7147.
 Koutrakis, P.; Wolfson, J. M.; Slater, J. L.; Brauer, M.; Spengler, J. D.; Stevens, R. K.; Stone, C. L. (1988a)
       Evaluation of an annular denuder/filter pack system to collect acidic aerosols and gases Environ Sci
       Technol. 22: 1463-1468.
 Koutrakis, P.; Wolfson, J. M.; Spengler, J. D. (1988b) An improved method for measuring aerosol strong acidity:
       results from a nine-month study in St. Louis, Missouri and Kingston, Tennessee. Atmos. Environ
       22: 157-162.
 Koutrakis, P.; Wolfson, J. M.; Spengler, J. D.; Stern, B.; Franklin, C. A. (1989) Equilibrium size of atmospheric
       aerosol sulfates as a function of the relative humidity.  J. Geophys. Res. [Atmos.] 94: 6442-6448.
 Koutrakis, P.; Wang, P.-Y.; Wolfson, J. M.; Soutas, C., inventors. (1996) Method and apparatus to measure
       paniculate matter in gas. U.S. patent 5,571,945. November 5. Available: www.patents.ibm.com [1999,
       November 24].
 Krieger, M. S.; Kites, R. A. (1992) Diffusion denuder for the collection of semivolatile organic compounds
       Environ. Sci. Technol. 26: 1551-1555.
 Lammel, G.; Novakov, T. (1995) Water nucleation properties of carbon black and diesel  soot particles. Atmos
       Environ. 29: 813-823.
 Landsberger, S.; Wu, D. (1993) Improvement of analytical sensitivities for the determination of antimony, arsenic,
       cadmium, indium, iodine, molybdenum, silicon and uranium in airborne particulate matter by epithermal
       neutron activation analysis. J. Radioanal. Nucl. Chem. 167: 219-225.
 Landsberger, S.;  Wu, D.; Vermette, S. J.; Cizek, W. (1997) Intercomparison of IAEA airborne particulate matter
       reference material. J. Radioanal. Nucl. Chem. 215: 117-127.
 Lane, D. A.; Johnson, N. D.; Barton, S. C.; Thomas, G. H. S.; Schroeder, W. H. (1988) Development and
       evaluation of a novel gas and particle sampler for semivolatile chlorinated organic compounds in ambient air.
       Environ. Sci. Technol. 22: 941-947.
 Lavanchy, V.  M. H.; Gaggeler, H. W.; Nyeki, S.; Baltensperger, U. (1999) Elemental carbon (EC) and black carbon
       (BC) measurements with a thermal method and an aethalometer at the high-alpine research station
      Jungfraujoch. Atmos. Environ. 33: 2759-2769.
 Lawless, P. A.; Rodes, C. E.; Evans, G.; Sheldon, L.; Creason, J. (2000) Aerosol concentrations during the 1999
      Fresno Exposure Studies as functions of size, season and meteorology. Aerosol Sci. Technol.: accepted.
 Lawrence, J. E.; Koutrakis, P. (1994) Measurement of atmospheric formic and acetic acids: methods evaluation and
      results from field studies. Environ. Sci. Technol. 28: 957-964.
Lawrence, J. E.; Koutrakis, P. (1996a) Measurement and speciation of gas and particulate phase organic acidity in
      an urban environment: 1. analytical. J. Geophys. Res. [Atmos.] 101: 9159-9169.
Lawrence, J.; Koutrakis, P. (1996b) Measurement and speciation of gas and particulate phase organic acidity in an
      urban environment: 2. speciation. J. Geophys. Res. [Atmos.]  101: 9171-9184.
         March 2001
                                                2-111
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Lawson, D. R. (1990) The Southern California Air Quality Study. J. Air Waste Manage. Assoc. 40: 156-165.
 2       Leach, K. B.; Kamens, R. M.; Strommen, M. R.; Jang, M. (1999) Partitioning of semivolatile organic compounds in
 3            the presence ofa secondary organic aerosol in a controlled atmosphere. J. Atmos. Chem. 33: 241-264.
 4       Lee, W.-M. G.; Shih, P.-M.; Wang, C.-S. (1997) The influence of relative humidity on the size of atmospheric
 5            aerosol. J. Environ. Sci. Health Part A: Environ. Sci. Eng. Toxic Hazard. Subst. Control 32: 1085-1097.
 6       Lewtas, J.; Pang, Y.; Booth, D.; Reimer, S.; Eatough, D. J.; Gundel, L. A. (2001) Comparison of sampling methods
 7            for semi-volatile organic carbon (SVOC) associated with PM2 5. Aerosol. Sci. Technol.: submitted.
 8       Li, W.; Montassier, N.; Hopke, P, K. (1992) A system to measure the hygroscopicity of aerosol particles. Aerosol
 9            Sci. Technol. 17:25-35.
10       Liang, Z,; Chan, C. K. (1997) A fast technique for measuring water activity of atmospheric aerosols. Aerosol Sci,
11            Technol. 26:255-268.
12       Liang, C.; Pankow, J. F.; Odum, J. R.; Seinfeld, J. H. (1997) Gas/particle partitioning of semivolatile organic
13            compounds to model inorganic, organic, and ambient smog aerosols. Environ. Sci. Technol. 31: 3086-3092.
14       Lindskog, A.; Brorstrom-Lunden, E.; Alfheim, I; Hagen, I. (1987) Chemical transformation of PAH on airborne
15            particles by exposure to NO2 during sampling: a comparison between two filter media. Sci. Total Environ.
16            61:51-57.
17       Liousse, C.; Cachier, H.; Jennings, S. G. (1993) Optical and thermal measurements of black carbon aerosol content
18             in different environments: variation of the specific attenuation cross-section, sigma. Atmos. Environ.
19             27A: 1203-1211.
20       Lundgren, D. A.;  Burton, R. M. (1995) Effect of particle size distribution on the cut point between fine and coarse
21             ambient mass fractions. In: Phalen, R: F.; Bates, D. V., eds. Proceedings of the colloquium on particulate air
22             pollution and human mortality and morbidity; January 1994; Irvine,  CA. Inhalation Toxicol. 7:131-148.
23       Lundgren, D. A.;  Hausknecht, B. J.; Burton, R. M. (1984) Large particle size distribution in five U.S. cities and the
24             effect on a new ambient particulate matter standard 0- Aerosol Sci. Technol. 7: 467-473.
25       Mahanama, K. R. R.; Gundel, L. A.; Daisey, J. M. (1994) Selective fluorescence detection of polycyclic aromatic
26             hydrocarbons in environmental tobacco smoke  and other airborne particles. Int. J. Environ. Anal. Chem.
27             56:289-309.
28       Malm, W. C.; Sisler, J. F.; Huffman, D.; Eldred, R. A-! Cahill, T. A. (1994) Spatial and seasonal trends in particle
29             concentration and optical extinction in the United States. J. Geophys. Res.  [Atmos.] 99:  1347-1370.
30       Mamane, Y.; Willis, R.; Conner, T. (2000) Evaluation of computer-controlled scanning electron microscopy applied
31             to an ambient urban aerosol sample. Aerosol Sci. Technol.: accepted.
32       Mazurek, M.; Masonjones, M. C.; Masonjones, H. D.; Salmon, L. G.; Cass, G. R.; Hallock, K. A.; Leach, M.
33             (1997) Visibility-reducing organic aerosols in the vicinity of Grand  Canyon National Park: properties
34             observed by high resolution gas chromatography. J. Geophys. Res. [Atmos.] 102: 3779-3793.
35       McDow, S. R.; Huntzicker, J. J. (1990) Vapor adsorption artifact in the sampling of organic aerosol: face velocity
36             effects. Atmos. Environ. Part A 24: 2563-2571.
37       Mclnnes, L. M.; Quinn, P. K.; Covert, D. S.; Anderson, T. L. (1996) Gravimetric analysis, ionic composition, and
38             associated water mass of the marine aerosol. Atmos. Environ. 30: 869-884.
39       McLaren, R.; Gertler, A.  W.; Wittorff, D. N.; Belzer,  W.; Dann, T.; Singleton, D. L. (1996) Real-world
40             measurements of exhaust and evaporative emissions in the Cassair tunnel predicted by chemical mass balance
41             modeling.  Environ. Sci. Technol. 30: 3001-3009.
42       McMurry, P. H. (2000) A review of atmospheric aerosol measurements. Atmos. Environ. 34:1959-1999.
43       McMurry, P. H.;  Friedlander, S. K. (1979) New particle formation in the presence of an aerosol. Atmos. Environ.
44            13:1635-1651.
45       McMurry, P. H.;  Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
46            aerosols. Atmos. Environ. 23: 497-507.
47      Meyer, A. S., Jr.; Boyd, C. M. (1959) Determination of water by titration with coulometrically generated Karl
48            Fischer reagent. Anal. Chem. 31:215-219.
49      Meyer, M. B.; Rupprecht, E. G. (1996) Particulate matter sampling methods: the importance of standardization.
50            J. Aerosol. Sci. 27(suppl. 1): S349-S350.
51       Meyer, M. B.; Rupprecht, E.; Patashnick, H. (1995) Considerations for the sampling and measurement of ambient
52            particulate matter. In: Particulate matter: health and regulatory issues: proceedings of an international
53            specialty conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association;
54            pp. 204-210. (A&WMA publication VIP-49).
          March 2001
2-112
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
45
46
47
48
49
50
51
52
53
54
 Meyer, M. B.; Patashnick, H.; Ambs, J. L.; Rupprecht, E. (2000) Development of a sample equilibration system for
       the TEOM continuous PM monitor. J. Air Waste Manage; Assoc. 50: 1345-1349.
 Middlebrook, A. M.; Murphy, D. M.; Thomson, D. S. (1998) Observations of organic material in individual marine
       particles at Cape Grim during the first aerosol characterization experiment (ACE 1). J. Geophys Res
     .. 103: 16475-16483.-
 Mignacca, D.; Stubbs, K. (1999) Effects of equilibration temperature on PM,0 concentrations from the TEOM
       method in the Lower Fraser Valley. J. Air Waste Manage. Assoc. 49: 1250-1254.
 Miller, F, J.; Gardner, D. E.; Graham, J. A.; Lee, R. E., Jr.; Wilson, W. E.; Bachmann, J. D. (1979) Size
       considerations for establishing a standard for inhalable particles. J. Air Pollut. Control Assoc: 29: 610-615.
 Morawska, L.; Thomas, S. B.; Bofinger, N. D.; Neale, D.; Wainwright, D. (1996) Characterisation of airborne
       particulate pollution. In: Smith, A., ed. 2000 Then What?: proceedings of the 13th international clean air and
       environment conference; September; Adelaide, Australia. Eastwood, New South Wales, Australia: Clean Air
       Society of Australia and New Zealand Inc.; pp. 285-290.
 Morawska, L.; Bofinger, N. D.; Kocis, L.; Nwankwoala, A. (1998a) Submicrometer and supermicrometer particles
       from diesel vehicle emissions. Environ. Sci. Technol. 32:  2033-2042.                   •
 Morawska, L.; Thomas, S.; Bofinger, N.; Wainwright, D.; Neale, D. (1998b) Comprehensive characterization of
       aerosols in a subtropical urban atmosphere: particle size distribution and correlation with gaseous pollutants
       Atmos. Environ. 32: 2467-2478.                        ,                                        •  '
 Mukerjee, S.; Ellenson, W. D.; Lewis, R. G.; Stevens, R. K.; Somerville, M. C.; Shadwick, D. S.; Willis, R. D.
       (1997) An .environmental scoping study in the lower Rio Grande Valley of Texas-Ill, residential'
       microenvironmental monitoring for air, house dust, and soil. Environ. Int. 23: 657-673.
 Murphy, D. M.; Thomson, D. S. (1997) Chemical composition of single aerosol particles at Idaho Hill: negative ion
       measurements. J. Geophys. Res.  102: 6353-6368.
 National Research Council. (1993) Protecting visibility in national parks and wilderness areas. Washington, DC:
       National Academy Press. 3v.
 Nelson, M. P.; Zugates, C. T.; Treado, P. J.; Casucio, G. S.; Exline, D. L.; Schlaegle,  S. F. (2000) Combining
       Raman chemical imaging scanning electron microscopy (SEM) to characterize  ambient fine particulate.
       Aerosol Sci. Technol.: accepted.,                                                        ,
 Neubauer, K. R.; Johnson, M. V,; Wexler, A. S. (1998) Humidity effects of the mass spectra of single aerosol
       particles. Atmos. Environ. 32:2521-2529.                                                  ,    .
 Noble, C. A.; Prather, K. A. (1996) Real-time measurement of correlated size and composition profiles,of individual
       atmospheric aerosol particles. Environ. Sci. Technol. 30:2667-2680.
 Noble* C. A.; Prather, K. A, (1998) Single particle characterization of albuterol metered dose inhaler aerosol in near
       real-time. Aerosol Sci. Technol. 29: 294-306.
 Norris, G. (1998) Air pollution and the exacerbation of asthma in an arid, western, US city [dissertation]. Seattle,
       WA: University of Washington. Available from: University Microfilms International,  Ann Arbor, MI;
       publication no. 9836228.                                       .
 Norris, G.; Young-Pong, S. N.; Koenig, J. Q.; Larson, T. V.; Sheppard, L.; Stout, J. W. (1999) An association
       between fine particles and asthma emergency department visits for children in Seattle  Environ Health
       Perspect. 107: 489-493.
 Obeidi, F. D.; Eatough, D. J. (1999) Continuous measurement of semi-volatile fine particulate mass in Provo, UT.
       J. Aerosol Sci.: submitted.
 Odum, J. R.; Hoffmann, T.; Bowman, F.; Collins, D.; Flagan, R. C.; Seinfeld, J. H. (1996) Gas/particle partitioning
       and secondary organic aerosol yields. Environ. Sci. Technol. 30: 2580-2585.     .   •
 Odum, J. R.; Jungkamp, T. P. W.; Griffin, R. J.; Forstner, H. J. L.; Flagan, R. C.; Seinfeld, J. H. (1997a) Aromatics,
      reformulated gasoline, and atmospheric organic aerosol formation. Environ. Sci. Technol. 31:1890-1897
 Odum, J. R.; Jungkamp, T. P. W.; Griffin, R. J.; Flagan, R. C.; Seinfeld, J. H. (1997b) The atmospheric
      aerosol-forming potential of whole gasoline vapor. Science (Washington, DC) 276: 96-99.
 Ohta, S.;Okita,T. (1990) A chemical characterization of atmospheric aerosol in Sapporo Atmos Environ Part A
      24A: 815-822.
Ohta, S.; Hori, M.; Yamagata, S.; Murao, N. (1998) Chemical characterization of atmospheric fine particles in
      Sapporo with determination of water content. Atmos. Environ. 32:1021-1025.
Ondov, J. M. (1996) Particulate tracers for source attribution: potential for application to California's San Joaquin
      Valley. J. Aerosol Sci. 27(suppl. 1): S687-S688.
         March 2001
                                                2-113
                                                                       DRAFT-DO NOT QUOTE OR CITE

-------
 1       Ondov, J. M.; Kelly, W. R.; Holland, J. Z.; Lin, Z. C.; Wight, S. A. (1992) Tracing fly ash emitted from a coal-fired
 2             power plant with enriched rare-earth isotopes: an urban scale test. Atmos. Environ. Part B 26B: 453-462.
 3       Pandis, S. N.; Paulson, S. E.; Seinfeld, J. H.; Flagan, R. C. (1991) Aerosol formation in the photooxidation of
 4             isoprene and "beta"-pinene. Atmos. Environ. Part A 25: 997-1008.
 5       Pandis, S. N.; Harley, R. A.; Cass, G. R.; Seinfeld, J. H. (1992) Secondary organic aerosol formation and transport.
 6             Atmos. Environ. Part A 26: 2269-2282.
 7       Pang, Y.; Ren, Y.; Obeidi, F.; Hastings, R.; Eatough, D. J.; Wilson, W. E. (2001) Semi-volatile species in PM25:
 8             comparison of integrated and continuous samplers for PM2 5 research or monitoring. J. Air Waste Manage.
 9             Assoc.51:174-185.
10       Pankow, J. F.; Bidleman, T. F. (1991) Effects of temperature, TSP and per cent non-exchangeable material in
11             determining the gas-particle partitioning of organic compounds. Atmos. Environ. Part A 25: 2241-2249.
12       Pankow, J. F.; Storey, J. M. E.; Yamasaki, H. (1993) Effects of relative humidity on gas/particle partitioning of
13             semivolatile organic compounds to urban particulate matter. Environ. Sci. Technol. 27:2220-2226.
14       Parmar, S. S.; Grosjean, D. (1990) Laboratory tests of KI and alkaline annular denuders. Atmos. Environ. Part A
15             24:2695-2698.
16       Patashnick, H.; Rupprecht, G.; Ambs, J. L.; Meyer, M. B. (2000) Development of a reference standard for
17             particulate matter mass in ambient air. Aerosol Sci. Technol.: accepted.
18       Pellizzari, E. D.; Krost, K. J. (1984) Chemical transformations during ambient air sampling for organic vapors.
19             Anal. Chem. 56:1813-1819.
20       Penner, J. E.; Novakov, T. (1996) Carbonaceous particles in the atmosphere: a historical perspective to the fifth
21             international conference on carbonaceous particles in the atmosphere. J. Geophys. Res. [Atmos.]
22             101: 19,373-19,378.
23       Petzold, A.; Kopp, C.; Niessner, R. (1997) The dependence of the specific attenuation cross-section on black carbon
24             mass fraction and particle size. Atmos. Environ. 31: 661-672.
25       Pilinis, C.; Seinfeld, J. H.; Grosjean, D. (1989) Water content of atmospheric aerosols. Atmos. Environ.
26             23:1601-1606.
27       Pinto, J. P.; Pleasant, M.; Kellogg, R.; Torok, S. Z.; Stiles, D.; Willis, R. D.; Koutrakis, P.; Allen, G. A. (1995)
28             Chemical characterization of the ambient aerosol in Philadelphia. Research Triangle Park, NC: U.S.
29             Environmental Protection Agency, National  Center for Environmental Assessment.
30       Pinto, J. P.; Stevens, R. K.; Willis, R. D.; Kellogg, R.; Mamane, Y.; Novak, J.; Santroch, J.; Benes, I.; Lenicek, J.;
31             Bures, V. (1998) Czech air quality monitoring and receptor modeling study. Environ. Sci. Technol.
32             32:843-854.
33       Pio, C.; Alves, C.; Duarte, A. (2001) Organic components of aerosols in a forested area of central Greece. Atmos.
34          '  Environ. 35:389-401.
35       Pirrone, N.; Keeler, G. J. (1996) The Rouge River watershed pollution by trace elements: atmospheric depositions
36            and emission sources. Water Sci. Technol. 33:267-275.
37      Poore, M. W. (2000) Oxalic acid in PM2J particulate matter in California. J. Air Waste Manage. Assoc.
38            50:1874-1875.
39      Pun, B. K.; Seigneur, C.; Grosjean, D.; Saxena, P. (2000) Gas-phase formation of water-soluble organic compounds
40            in the atmosphere: a retrosynthetic analysis.  J. Atmos. Chem. 35:199-223.
41       Reilly, P. T. A.; Gieray, R. A.; Whitten, W. B.; Ramsey, J. M. (1998) Real-time characterization of the organic
42            composition and size of individual diesel engine smoke particles. Environ. Sci. Technol. 32: 2672-2679.
43      Reist, P. C. (1984) Introduction to aerosol science. New York, NY: Macmillan Publishing Company.
44      Reist, P. C. (1993) Introduction to aerosol science. New York, NY: McGraw-Hill.
45      Risse, U.; Flammenkamp, E.; Kettrup, A. (1996) Determination of aromatic hydrocarbons in air using diffusion
46            denuders. Fresenius J. Anal. Chem. 356:390-395.
47      Roberts, R. A.; Corkill, J. (1998) Grass seed field smoke and its impact on respiratory health. J. Environ. Health 60
48            (June):  10-16.
49      Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991) Sources of fine organic
50             aerosol. 1. Charbroilers and meat cooking operations. Environ. Sci. Technol. 25: 1112-1125.
51      Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. (1993a) Quantification of urban
52             organic aerosols at a molecular level: identification, abundance and seasonal variation. Atmos. Environ. Part
 53             A 27:1309-1330.
          March 2001
2-114
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
49
 50
 51
 52
53
54
55
 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993b) Sources of fine organic
        aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ Sci
        Technol. 27:636-651.
 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993c) Sources of fine organic
        aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks Environ
        Sci. Technol. 27: 1892-1904.
 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993d) Sources of fine organic
        aerosol. 4. Particulate abrasion products from leaf surfaces of urban plants. Environ. Sci. Technol
        27:2700-2711.
 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993e) Sources of fine organic
        aerosol. 5. Natural gas home appliances. Environ. Sci. Technol. 27: 2736-2744.
 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1994) Sources of fine organic
       aerosol. 6. Cigarette smoke in the urban atmosphere. Environ. Sci. Technol. 28: 1375-1388.
 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997a) Sources of fine organic
       aerosol. 7. Hot asphalt roofing tar pot fumes. Environ. Sci. Technol. 31: 2726-2730.
 Rogge, W. F,; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997b) Sources of fine organic
       aerosol. 8. Boilers burning no. 2 distillate fuel oil. Environ. Sci. Technol. 31: 2731-2737.
 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1998) Sources of fine organic
       aerosol. 9. Pine, oak, and synthetic log combustion in residential fireplaces. Environ. Sci. Technol. 32: 13-22
 Rupprecht, G.; Patashnick, H.; Beeson, d. E.; Green, R. N.; Meyer, M. B. (1995) A new automated monitor for the
       measurement of particulate carbon in the atmosphere. In: Particulate matter: health and regulatory issues:
       proceedings of an international specialty conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste
       Management Association; pp. 309-319. (A&WMA publication VIP-49).
 Saxena, P.; Hildemann, L. M. (1996) Water-soluble organics in atmospheric particles: a critical review of the
       literature and applications of thermodynamics to identify candidate compounds. J. Atmos. Chem.  24: 57-109
 Saxena, P.; Hildemann, L. M.; McMurry, P. H.; Seinfeld, J. H.  (1995) Organics alter hygroscopic behavior of
       atmospheric particles. J. Geophys. Res. [Atmos.] 100: 18,755-18,770.
 Schauer, J. J.; Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R. (1996) Source apportionment of
       airborne particulate matter using organic compounds as tracers. Atmos. Environ. 30: 3837-3855.
 Seinfeld, J. H.; Pandis, S. N. (1998) Atmospheric chemistry and physics: from air pollution to climate change New
       York, NY: John Wiley &  Sons, Inc.
 Shah, J. J.; Johnson, R. L.; Heyerdahl, E. K.; Huntzicker, J. J. (1986) Carbonaceous aerosol at urban and rural sites
       in the United States. J. Air Pollut. Control Assoc. 36: 254-257.
 Silva, P. J.; Prather, K. A. (1997) On-line characterization of individual particles from automobile emissions
       Environ. Sci. Technol. 31: 3074-3080.
 Sioutas, C.; Koutrakis, P.; Wolfson, J. M. (1994) Particle losses in glass honeycomb denuder samplers. Aerosol Sci
       Technol. 21: 137-148.
 Sioutas, C.; Koutrakis, P.; Wolfson, J. M.; Azaroff, L. S.; Mulik, J. D. (1996a) Field evaluation of a glass
       honeycomb denuder/filter pack system to collect atmospheric gases and particles. In: Gay, B. W.; Fuerst,
       R. G.; Jayanty, R. K. M., eds. Measurement of toxic and  related air pollutants: proceedings of a conference-
       May, 1994; Durham, NC.  Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of '
       Research and Development; report no. EPA/600/R-94/136; pp. 426-435.
 Sioutas, C.; Wang, P. Y.; Ferguson, S. T.; Koutrakis, P.; Mulik,  J. D. (1996b) Laboratory and field evaluation of an
      improved glass honeycomb denuder/filter pack sampler. Atmos. Environ. 30: 885-895.
 Sioutas, C.; Koutrakis, P.; Wang, P.-Y.; Babich, P.; Wolfson, J.  M. (1999) Experimental investigation of pressure
      drop with particle loading  in Nuclepore filters. Aerosol Sci. Technol. 30: 71-83.
 Solomon, P. A.; Salmon, L. G.; Fall, T.; Cass, G. R. (1992) Spatial and temporal distribution of atmospheric nitric
      acid and particulate nitrate concentrations in the Los Angeles area. Environ. Sci. Technol. 26: 1596-1601.
Speer, R. E.; Barnes, H. M.; Brown, R. (1997) An instrument for measuring the liquid water content of aerosols
      Aerosol Sci. Technol. 27: 50-61.
Stein, S. W.; Turpin, B. J.; Cai, X.; Huang, P.-F.; McMurry, P. H. (1994) Measurements of relative
      humidity-dependent bounce and density for atmospheric particles using the DMA-impactor technique
      Atmos. Environ. 28: 1739-1746.
Stolzenburg, M. R,; Hering, S. V. (2000) Method for the automated measurement of fine particle nitrate in the
      atmosphere. Environ. Sci. Technol. 34: 907-914.
         March 2001
                                                2-115
                                                                       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Strommen, M. R.; Kamens, R. M. (1997) Development and application of a dual-impedance radial diffusion model
 2            to simulate the partitioning of semivolatile organic compounds in combustion aerosols. Environ. Sci.
 3            Technol. 31:2983-2990.
 4      Suarez, A. E.; Caffrey, P. F.; Ondov, J. M.; Thaung, K. C. (1996) Size distribution of Ir-tagged soot from diesel
 5            sanitation trucks. J. Aerosol Sci. 27(suppl. 1): S697-S698.
 6      Tang, I. N. (1980) Deliquescence properties and particle size change of hygroscopic aerosols. In: Willeke, K., ed.
 7            Generation of aerosols and facilities for exposure experiments. Ann Arbor, MI: Ann Arbor Science
 8            Publishers, Inc.; pp. 153-167.
 9      Tang, H.; Lewis, E. A.; Eatough, D. J.; Burton, R. M.; Farber, R. J. (1994) Determination of the particle size
10            distribution and chemical composition of semi-volatile organic compounds in atmospheric fine particles with
11            a diffusion denuder sampling system. Atmos. Environ. 28: 939-947.
12      Thomson, D. S.; Murphy, D. M. (1994) Analyzing single aerosol particles in real time. Chemtech 24: 30-35.
13      Tolocka, M. P.; Solomon, P. A.; Mitchell, W.; Morris, G. A.; Gemmill, D. B.; Wiener, R. W.; Vanderpool, R. W.;
14            Homolya, J. B.; Rice, J. (2000) East vs. west in the US: chemical characteristics of PM2 5 during the winter of
15             1999. Aerosol Sci. Technol.: accepted.
16      Tsai, C.-J.; Cheng, Y.-H. (1996) Comparison of two ambient beta gauge PM10 samplers. J. Air Waste Manage.
17            Assoc. 46:142-147.
18      Tsai, C.-J.; Huang, H.-Y. (1995) Atmospheric aerosol sampling by an annular denuder system and a high-volume
19            PM,0 sampler. Environ. Int. 21: 283-291.
20      Turpin, B. J.; Huntzicker, J. J. (1991) Secondary formation of organic aerosol in the Los Angeles basin:
21            a descriptive analysis of organic and elemental carbon concentrations. Atmos. Environ. Part A 25: 207-215.
22      Turpin, B. J.; Huntzicker, J. J. (1995) Identification of secondary organic aerosol episodes and quantitation of
23            primary and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29: 3527-3544.
24      Turpin, B. J.; Gary, R. A.; Huntzicker, J. J. (1990) An in situ, time-resolved analyzer for aerosol organic and
25            elemental carbon. Aerosol Sci. Technol. 12:161-171.
26       Turpin, B. J.; Liu, S.-P.; Podolske, K. S.; Gomes, M. S. P.; Eisenreich, S. J.; McMurry, P. H. (1993) Design and
27             evaluation of a novel diffusion separator for measuring gas/particle distributions of semivolatile organic
28             compounds. Environ. Sci. Technol. 27: 2441-2449.
29       Turpin, B. J.; Huntzicker, J. J.; Hering, S. V. (1994) Investigation of organic aerosol sampling artifacts in the Los
30             Angeles basin. Atmos. Environ. 28: 3061-3071.
31       U.S. Environmental Protection Agency. (1982) Air quality criteria for particulate matter and sulfur oxides. Research
32             Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
33             Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
34      U.S. Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park,
35             NC: National  Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
36       U.S. Environmental Protection Agency. (1999) Particulate matter (PM25) speciation guidance. Final draft. Edition 1.
37             Research Triangle Park, NC: Office of Air Quality Planning and Standards. Available:
38             www.epa.gov/ttn/amtic/files/ambient/pm25/spec/specfinl.pdf [2000, October 11].
39      Van Loy, M.; Bahadori, T.; Wyzga, R.; Hartsell, B.; Edgerton, E. (2000) The Aerosol Research and Inhalation
40             Epidemiology Study (ARIES): PM2 5 mass and aerosol component concentrations and sampler
41              intercomparisons. J. Air Waste Manage. Assoc. 50:1446-1458.
42      Van Vaeck, L.; Van  Cauwenberghe,  K.; Janssens, J. (1984) The gas-particle distribution of organic aerosol
43              constituents: measurements of the volatilisation artefact in hi-vol cascade impactor sampling. Atmos.
44             Environ. 18:417-430.
45      Vasilou, J. G.; Sorenson, D.; McMurry, P. H.  (1999) Sampling at controlled relative humidity with a cascade
46             impactor. Atmos. Environ. 30: 1049-1056.
47      Vega,  E.; Garcia, I.;  Apam, D.; Ruiz, M. E.; Barbiaux, M. (1997) Application of a chemical mass balance receptor
48             model to respirable particulate matter in Mexico City. J. Air Waste Manage. Assoc. 47: 524-529.
49      Veltkamp, P. R.; Hansen, K. J.; Barkley, R. M.; Sievers, R. E. (1996) Principal component analysis of summertime
 50             organic aerosols at Niwot Ridge, Colorado. J. Geophys. Res. [Atmos.] 101: 19,495-19,504.
 51      Venkataraman, C.; Lyons, J. M.; Friedlander, S. K. (1994) Size distributions of polycyclic aromatic hydrocarbons
 52             and elemental carbon. 1. Sampling, measurement methods, and source characterization. Environ. Sci.
 53             Technol. 28: 555-562.
 54      Vincent, J. H. (1995) Aerosol science for industrial hygienists. Oxford, United Kingdom: Pergamon.
          March 2001
2-116
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
48
49
 50
 51
52
53
54
55
  Vincent, J. H., ed. (1999) Particle size-selective sampling of participate air contaminants'. Cincinnati, OH: American
        Conference of Governmental Industrial Hygienists (ACGIH).         •            , ;,.;,!• i  ,.,,.-
  Vincent, J. H.; Jones, A. D.; Johnston, A. M.; McMillan, C.; Bolton, R. E.; Cowie, H. (1987) Accumulation of
        inhaled mineral dust in the lung and associated lymph nodes: implications for exposure arid dose in;  .
        occupational lung disease. Ann. Occup.Hyg. 31:375-393.                  .,   •  ...  .-.,!.•,.        ,
  Wang, P. Y. (1997) Continuous aerosol mass measurement by flow obstruction. Boston, MA;'Harvard University
        School of Public Health.                                           ,        .   ,,..,..   <•,  .     ,
  Wang, H.-C.; John, W. (1,988) Characteristics of the Berner impactor for sampling inorganic ions:-Aerosol Sci
       Technol. 8: 157-172.
  Watson, J. G.; Chow, J. C.; Lu, Z.; Fujita, E. M.; Lowenthal, D. H.; Lawson, D. R.; Ashbaugh, L. L. (1994)
       Chemical mass balance source apportionment of PM10 during the Southern California Air Quality Study
       Aerosol Sci. Technol. 21:1-36.
  Wedding, J. B.; Weigand, M. A.  (1993) An automatic particle sampler with beta gauging. Air Waste 43- 475-479
  Weingartner, E.; Burtscher, H.; Baltensperger, U. (1997) Hygroscopic properties of carbon and diesel soot particles
       Atmos. Environ. 31: 2311 -2327.
  Weschler, C. J.; Shields, H. C. (1999) Indoor ozone/terpene reactions as a source of indoor particles Atmos
       Environ. 33:2301-2312.
  West, J. J.; Ansari, A. S.; Pandis, S. N. (1999) Marginal PM2.5: nonlinear aerosol mass response to sulfate
       reductions in the eastern United States. J. Air Waste Manage. Assoc. 49: 1415-1424.
 Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
 Whitby, K. T.; Charlson, R. E.; Wilson, W. E.; Stevens, R. K. (1974) The size of suspended particle matter in air
       Science (Washington, DC) 183: 1098-1099.
 Whitby, K. T.; Sverdrup, G. M. (1980) California aerosols: their physical and chemical characteristics. In: Hidy
       G. M.; Mueller, P. K.; Grosjean, D.; Appel, B. R.; Wesolowski, J. J., eds. The character and origins of smog
       aerosols: a digest of results from the California Aerosol Characterization Experiment (ACHEX). New York
       NY: John Wiley & Sons, Inc.; pp. 477-517. (Advances in environmental science and technology: v. 9).
 Willeke, K.; Baron, P. A., eds. (1993) Aerosol measurement: principles, techniques and applications New York
       NY: Van Nosrrand Reinhold Publishers.
 Willeke, K.; Whitby, K. T. (1975) Atmosperic aerosols: size distribution interpretation. J. Air Pollut. Control Assoc
       25: 529-534.
 Williams, E. L., II; Grosjean, D. (1990) Removal of atmospheric oxidants with annular denuders Environ Sci
       Technol. 24:811-814.                                                           '        '    '
 Williams, R.; Suggs, J.; Zweidinger, R.; Evans, G.; Creason, J.; Kwok, R.; Rodes, C.; Lawless, P.; Sheldon, L.
       (2000) The 1998 Baltimore paniculate matter epidemiology-exposure study: part 1  - comparison of ambient,
       residential outdoor, indoor and apartment paniculate matter monitoring. J. Exposure Anal  Environ
       Epidemiol.: 10: 518ff.
 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
       epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
 Wilson, W. E.; Spiller, L. L.; Ellestad, T. G.; Lamothe, P. J.; Dzubay, T. G.; Stevens, R. K.; Macias, E. S.; Fletcher
       R. A.; Husar, J. D.; Husar, R. B.; Whitby, K. T.; Kittelson, D. B.; Cantrell, B. K. (1977) General Motors
       sulfate dispersion experiment: summary of EPA measurements. J. Air Pollut. Control Assoc 27- 46-51
 Woo, K.-S.; Chen, D.-R.; Pui, D. Y. H.; Wilson, W. E. (2000) Use of continuous measurements of integral aerosol
       parameters to estimate particle surface area. Aerosol Sci. Technol.: accepted.
 Wu, C. C.; Suarez, A. E.; Lin, Z.; Kidwell,  C. B.; Borgoul, P. V.; Caffrey, P. F.; Ondov, J. M.; Sattler, B. (1998)
       Application of an Ir tracer to determine soot exposure to students commuting to school on Baltimore public
       buses. Atmos. Environ. 32: 1911-1919.
 Yamasaki, H.; Kuwata, K.; Miyamoto, H. (1982) Effects of ambient temperature on aspects of airborne polycyclic
      aromatic hydrocarbons. Environ. Sci. Technol. 16: 189-194.
 Yatin, M.; Tuncel, S. G.; Tuncel, G.; Aras, N. K. (1994) Trace element composition of atmospheric aerosols in
      Ankara, Turkey, determined by instrumental neutron activation analysis. J. Radioanal Nucl  Chem
       181:401-411.
Yu, J.; Cocker, D. R., Ill; Griffin, R. J.; Flagan, R. C.; Seinfeld, J. H. (1998) Gas-phase ozone oxidation of
      monoterpenes: gaseous and particulate products. J. Atmos. Chem. 34: 207-258.
Yu, J.; Griffin, R. J.; Cocker, D. R., Ill; Flagan, R. C.; Seinfeld, J. H. (1999) Observation of gaseous and particulate
      products of monoterpene oxidation in forest atmospheres. Geophys. Res. Lett. 26: 1145-1148.
         March 2001
                                                        2-117
                                                              DRAFT-DO NOT QUOTE OR CITE

-------
1      Zhang, X. Q.; McMuny, P. H. (1987) Theoretical analysis of evaporative losses from impactor and filter deposits.
2            Atmos. Environ. 21:1779-1789.
3      Zhang, X.; McMurry, P. H. (1992) Evaporative losses of fine particulate nitrates during sampling. Atmos. Environ.
4            Part A 26: 3305-3312.
5      Zhang, S.-H.; Shaw, M.; Seinfeld, J. H.; Flagan, R. C. (1992) Photochemical aerosol formation from "alpha -pmene
6            and "beta"-pinene. J.  Geophys. Res. [Atmos.] 97: 20,717-20,729.
7      Zhang, X. Q.; McMurry, P. H.; Hering, S. V.; Casuccio, G. S. (1993) Mixing characteristics and water content of
8            submicron aerosols measured in Los Angeles and at the Grand Canyon. Atmos. Environ. Part A
9            27:1593-1607.
         March 2001
2-118
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
   3.  CONCENTRATIONS, SOURCES, AND EMISSIONS
        OF ATMOSPHERIC PARTICULATE MATTER
 3.1  INTRODUCTION
      This chapter incorporates material from Chapters 5 (Sources and Emissions of Atmospheric
 Particles) and Chapter 6 (Environmental Concentrations) of the previous document, Air Quality
 Criteria for Particulate Matter or "1996 PM AQCD" (U.S. Environmental Protection Agency,
 1996) and presents updates to these materials where available.
      Information about concentrations, the composition, and the spatial and temporal variability
 of ambient particles across the United States is presented in Section 3.2. Ambient concentration
 data obtained during the first year of operation of the recently deployed nationwide network of
 Federal Reference Method PM25 monitors are presented. Results of field studies that have
 characterized the composition of organic compounds in the ambient aerosol are summarized in
 Appendix 3 A as a  complement to the data for the inorganic composition of ambient particles that
 was presented in Appendix 6A in the 1996 PM AQCD.  Data for characterizing the daily and
 seasonal variability of PM25 concentrations are discussed in Section 3.2.1, the intraday variability
 of PM25 concentrations in Section 3.2.2, the relations among different size  fractions in
 Section 3.2.3, the interrelations and correlations among PM components in Section 3.2.4, and the
 spatial variability of various PM components in Section 3.2.5.
     Unlike gaseous criteria pollutants (SO2, NO2, CO, O3), which are well-defined chemical
 entities, atmospheric particulate matter (PM) is composed of a variety of particles  differing in
 size and chemical composition. Therefore, sources of each component of the atmospheric
 aerosol must be considered in turn. Differences in the composition of particles emitted by
 different sources also will lead to spatial and temporal heterogeneity in the composition of the
 atmospheric aerosol. The nature of the sources and the composition of the emissions from these
 sources are discussed in Section 3.3. Estimates of contributions of various  sources to ambient
 PM levels given by source apportionment studies also are presented in Section 3.3. More
 detailed information about the composition of emissions from various sources is given in
Appendix 3B. Because PM is composed of both primary and secondary constituents, emissions
       March 2001
                                        3-1
DRAFT-DO NOT QUOTE OR CITE

-------
 1      of both the primary components and the gaseous precursors of secondary PM must be considered.
 2      Nationwide emissions estimates of primary PM and precursors to secondary PM are discussed in
 3      Section 3.4.1 and uncertainties in emissions estimates in Section 3.4.2.
 4           The organization of topics in this chapter (ambient measurements, source characterization
 5      and apportionment, and emissions inventories) reflects, in a broad sense, the order in which these
 6      topics are addressed in scientific studies and arguably increasing uncertainties in these areas.
 7
 8
 9     3.2  TRENDS AND  PATTERNS IN AMBIENT PM2.S CONCENTRATIONS
10          A significant amount of data for characterizing PM10 mass concentrations and trends exists
11      and that available up to about 1994 was presented in the 1996 PM AQCD. However, data sets
12     for characterizing PM2 5 and PM(10.2 5) mass or trends were not as extensive.  Sources of data on
13     PM2 5 (fine) and PM(I0.2.5) (coarse), which were discussed in the 1996 PM AQCD, include EPA's
14     Aerometric Information Retrieval System (AIRS) (U. S. Environmental Protection Agency,
15     2000a), IMPROVE (Eldred and Cahill, 1994; Cahill, 1996), the California Air Resources Board
16     (CARB) Data Base (California Air Resources Board, 1995), the Harvard Six-Cities Data Base
17     (Spengler et al., 1986; Neas, 1996), and the Harvard Philadelphia Data Base (Koutrakis, 1995).
18     The Inhalable Particulate Network (EPN) (Inhalable Particulate Network, 1985; Rodes and Evans,
19     1985) provided TSP, PM15, and PM25 data but only a small amount of PM10 data.
20          New sources of PM data include the recently deployed nationwide PM2 5 compliance
21     monitoring network, which provides mass measurements using a Federal Reference Method
22     (FRM). This section summarizes calendar year 1999 data from this network, and provides an
23     approximate characterization of nationwide PM(10_2 5) by comparing PM10 to PM25 measurements
24     at sites where both types of compliance monitors are located.  In addition, a small number of
25     recent studies in which daily mass and composition measurements are available for extended
26     periods will be discussed in this section.
27           Summary tables giving the results of field studies that obtained data for the composition of
28     particles in the PM2.5, PM(10.2.S), or PM10 size ranges were presented in Appendix A to Chapter 6
29     of the 1996 PM AQCD.  The summary tables included data for mass, organic carbon, elemental
30     carbon, nitrate, sulfate, and trace elements. The results of 66 studies were separated and
31     presented for the eastern, western, and central United States.  It should be noted that these studies
        March 2001
3-2
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
 took place at various times and lasted for various durations over a 20-year period. Summary
 tables showing data for organic carbon, elemental carbon and organic compounds in ambient
 particles are given in Appendix 3 A. Data for the gross chemical composition of PM2 5 particles
 sampled in rural air by the IMPROVE network and for the aerosol composition studies cited in
 the 1996 PM AQCD (which were carried out mainly in urban areas) are summarized in
 Table 3-1. Data are shown separately for the eastern and western United States. The IMPROVE
 data are annual average concentrations for 1998. Quality assured aerosol composition data for
 urban areas from the PM2 5 speciation network are not yet available for comparison to the
 IMPROVE data set. Many features are reflected broadly in both data sets (i.e., increasing
 organic carbon, nitrate, and minerals and decreasing sulfate in going from east to west).  The
 annual average PM2 5 concentration of 11.0 /^g/m3 reported by eastern IMPROVE sites is almost
 three times higher than 3.9 ^g/m3 reported by IMPROVE sites in the western United States. The
 data shown in Table 3-1 refer only to components that have been identified and quantified.
There may be other unidentified components that can represent a significant contribution to the
total measured mass, as indicated in the 1996 PM AQCD.
              TABLE 3-1. GROSS CHEMICAL COMPOSITION OF PM2 5 PARTICLES
                OBTAINED IN RURAL AREAS OF THE EASTERN AND WESTERN
            UNITED STATES BY THE IMPROVE NETWORK AND IN MIXED RURAL,
          SUBURBAN, AND URBAN AREAS OBTAINED BY STUDIES SUMMARIZED IN



SO4=
EC
OC
NO3
Crustal

Eastern

56
5
27
5
7
IMPROVE1
U.S. Western U.S.
% Contribution
33
6
36
8
17
Reconstructed PM2 5 Concentration Og/m3)
PM,,
11.0
3.9

Eastern

44
, 5
27
1
6
1996PM AQCD2
U.S. Western U.S.
% Contribution3
11
14
38
15
14
PM2 5 Concentration 0/g/m3)


        "Note that contributions do not add to 100% because a portion of the measured total mass was not characterized chemically.
        Sources: (1) IMPROVE network (1998); (2) U.S. Environmental Protection Agency (1996).
      March 2001
                                               3-3
                                                  DRAFT-DO NOT QUOTE OR CITE

-------
1
2
3
4
5
6
7
PM,0 Concentrations and Trends
     Nationwide PM10 annual mean concentrations from AIRS for calendar year 1999 are shown
in Figure 3-1. Concentrations in most areas of the country were below the level of the annual
PM10 standard (50 //g/m3) in 1999.  Exceptions include central South Carolina, Puerto Rico, and
several places in the southwestern United States and central California.
                                      c i c    "~rr~S »„_
                       Alaska
                          o  < 4 quarters
                          C  one or more quarters with < 11 samples
                          O  All quarters with at least 11 samples
                          O  All quarters 75% or more complete
                                                           >50
                                                          30-50
                                                          20-30
                                                           0-20
        Figure 3-1.  1999 annual mean PM^ concentrations C"g/m3).

        Source: Fitz-Simons et al. (2000).
  1          Nationwide trends in annual mean PMIO concentrations from 1989 through 1998 based on
  2     data obtained at 138 rural sites, 355 suburban sites, and 413 urban sites reporting to AIRS are
        March 2001
                                                   3-4
                                                        DRAFT-DO NOT QUOTE OR CITE

-------
1
2
3
4
5
6
shown in Figure 3-2 (U.S. Environmental Protection Agency, 2000b). Though average
concentration levels differ among sites, with higher levels at urban and suburban sites, the overall
nationwide trend shows a decline. Figure 3-3 shows the annual mean PM10 trend summarized by
EPA region. Decreases were greater in the western United States than in the eastern United
States, ranging from about 20% in the East to about 38% in the Northwest.
                35

                30

          1=   25
           O)
           g  29
           CD
           O
          O
         15

         10

          5

          0
                \.
                                                	 Rural (138 sites)
                                                	Suburban (355 sites)
                                                	Urban (413 sites)
                    89    90    91    92
93    94
 Year
                                                     95   96    97    98
     Figure 3-2, Nationwide trend in ambient PM10 concentration from 1989 through 1998.
     Source: U.S. Environmental Protection Agency (2000b).
     March 2001  '
                                      3-5
                                                      DRAFT-DO NOT QUOTE OR CITE

-------
                 34.3
                 •-f 38%
                                  27.2
              413
                f 36%
                                               f-21%
                      f 22%
                                                                30.0
                                          28.7
                                           f 19%
                                                                  f 18%
              The National Trend
               31.7
             f 25%
       Alaska Is in EPA Region 10; Hawaii, EPA Region 9; and Puerto Rico, EPA Region 2.
       Concentrations are ug/m3.
                             Note: These trends ate
                             influenced by the
                             distribution of monitoring
                             locations in a given region
                             and, therefore, can be
                             driven largely by urban
                             concentrations. For this
                             reason, they are not
                             indicative of background
                             regional concentrations.
       Figure 3-3. Trend in PM10 annual mean concentrations by EPA region, 1989 through
                   1998
       Source: U. S. Environmental Protection Agency (2000a).
 1     PMZ5 Concentrations and Trends
 2           Nationwide PM2 s concentrations from the 1999 compliance network are shown in
 3     Figures 3-4a and 3-4b. By the end of 1999 the network consisted of over 1025 monitors. Annual
 4     mean PM2 s concentrations were above 15 jug/m3 in many areas of the country, especially
 5     throughout the eastern United States, and above 20 //g/m3 in several major urban locations.  The
 6     98th percentile 24-h average concentrations were generally below 65 yug/m3.  Most of the sites
 7     with levels above 65 /^g/rn3 are located in California. As shown by the size of the dots on the
 8     maps, the picture for 1999 is not complete because some monitoring locations did not record
 9     valid data for all four quarters, or recorded fewer than 11 samples in one or more quarters.
10     Further, at the time these maps were created some states such as Massachusetts and New
11     Hampshke had not reported valid data to AIRS from all monitoring sites. It is premature to
        March 2001
3-6
DRAFT-DO NOT QUOTE OR CITE

-------
                                    •v-"       \
                                      o      \
                                   Hawaii          >
                       Alaska
                                                            Puerto Rico
                          °  < 4 quarters                        0   >20   .
                          C  one or more quarters with < 11 samples    H  15-20
                          O  All quarters with at least 11 samples      © . 10 - 15
                          O  All quarters 75% or more complete       O  °-10
               Source: US EPA AIRS Data Base as of 7/12/00 without data flagged as 1 ,. 2, 3, 4, T, W, Y, or X.
       Figure 3-4a.  1999 annual mean PM25 concentrations (//g/m3).
       Source: Fitz-Simons et al. (2000).
 1
 2
 3
 4
.5
 6
 7
 8
 9
make judgments on whether an area likely will attain the 1997 PM25 standards based on this
single year of data, not only because the 1999 data is not complete, but also because the 1997
standard is defined in terms of 3-year average concentrations.
     Annual average PM2 5 obtained as part of health studies conducted in various locations in
the United States and Canada from the late 1980s to the early 1990s are shown in Figure 3-5
(Bahadori et al., 2000a). These studies include the Harvard six-cities study (Steubenville, OH;
Watertown, MA; Portage, WI; Topeka, KS; St. Louis, MO; and Kingston-Harriman, TN),
PTEAMS (Riverside, CA),  MAACS (Philadelphia,  PA; Washington, DC; and Nashville, TN),
South Boston Air Quality and Source Apportionment Study (Boston, MA); NPMRMN
       March 2001
                                           3-7
                                                             DRAFT-DO NOT QUOTE OR CITE

-------
                      Alaska
                                 Hawaii
                        o < 4 quarters of data
                        O one or more quarters with < 75% of scheduled samples
                        O All quarters with at least 75% of scheduled samples
    £  >65
    • 40-65
    ©  30-40
    O  0-30
            Source: US EPAAIRS Data Base as of 7/12/00 without data flagged as'1, 2, 3, 4, T, W, Y, orX.

      Figure 3-4b. 1999 98th percentile 24-h average PM2.5 concentrations C"g/m3).
      Source:  Fitz-Simons et al. (2000).
1      (Phoenix, AZ). Remaining sites were part of the 24-cities study.  Sufficient data are not yet
2      available to permit the calculation of nationwide trends of PM2.5 and PM(I0.2.5); however some
3      general conclusions can be reached. Darlington et al. (1997) proposed that because the consistent
4      reductions in PM10 levels were found in a wide variety of environments ranging from urban to
5      rural over large areas, that common factors or controls might be responsible for these reductions,
6      and that these factors affected fine particles more strongly than coarse particles because fine
7      particles can be transported over longer distances. The longest time series of PM2 5 concentration
8      and composition data have been obtained by the California Air Resources Board. Their data
9      show that annual average PM2 5 concentrations decreased about 50% in the South Coast Air
       March 2001
                                                   3-8
DRAFT-DO NOT QUOTE OR CITE

-------
          - uojsog

          - B>j9dqj_

          - 9691103

          - UOJM9N
                                                                         )OB|g

                                                 9|||AS9UBZ
                                   •o


                                   1
                                   09


                                   1
                                   a
                                   o



                                   I
                                   o>
                                                SU91JJV

                                                SUOSJBd
                                               - 9IIJALJSBN
                                   I
                                  JS
                                  •4
                                  CM
                                   O
                                   a
                                   o
                                                PB9L|9JO|AJ


                                                UMOJUOjUn
                                                9PJSJ8AJJJ
                                                                        •X   o
                                                                        •a   o
                                                                            I
   ?ui/6rl '

                                                                        OG
                                                                             03
                                                                            n
                                                                            I
March 2001
3-9
                                               DRAFT-DO NOT QUOTE OR CITE

-------
 1      Basin, 35% in the San Joaquin Valley, 30% in the San Francisco Bay Area, and 35% in the
 2      Sacramento Valley from 1990 to 1995 (Dolislager and Motallebi, 1999). PM2.5 data have been
 3      collected continuously since 1994 as part of the children's health study in 12 communities in
 4      southern California (Taylor et al., 1998). Data obtained at all sites show decreases ranging from
 5      2% at Santa Maria  to 37% at San Dimas/Glendora in PM25 from 1994 through 1998. These
 6      decreases were accompanied by decreases in major components such as nitrate, sulfate,
 7      ammonium, and acids.  However, undefined components showed a mixed pattern of increases
 8      and decreases at the same sites.
 9           In common usage, the term "background concentrations" refers to concentrations observed
10      in remote areas relatively unaffected by local pollution sources. However, as noted in Chapter 6
11      of the 1996 PM AQCD, several definitions of background concentrations are possible.
12          The two definitions chosen in that document as being most relevant for regulatory purposes
13     are based on  estimates of contributions from uncontrollable sources that can affect concentrations
14     in the United States. The first definition refers to the concentration resulting from anthropogenic
15     and natural emissions outside North America and natural sources within North America. The
16     second definition refers to the concentration resulting from natural sources only within and
17     outside of North America. Because of long-range transport from anthropogenic source regions in
 18     North America, it  is impossible to obtain background concentrations defined above solely on the
 19     basis of direct measurement in remote areas in North America. Annual average natural
20      background levels of PM10 (according to definition 1) have been estimated to range from 4 to
21      8 /zg/m3 in the western United States and 5 to 11 Mg/m3 in the eastern United States.
22      Corresponding PM2.5 levels have been estimated to range from 1 to 4 //g/m3 in the western
 23      United States and  from 2 to 5 Mg/m3 in the eastern United States (U.S. Environmental Protection
 24      Agency, 1996). Although the values shown in Table 3-1 are broadly consistent with those given
 25      above, the data shown in Table 3-1 represent only upper limits to background concentrations,
 26      because of contributions from long-range transport from anthropogenic sources within North
 27      America. Peak 24-h average natural background concentrations may be substantially higher than
 28      the annual or seasonal average natural background concentrations. Estimates of levels for
 29      background  2 are  not yet available. However, recent information about contributions to
 30      background  concentrations that fall under definitions 1 and 2 because of long-range transport
 31      from sources outside the United States is given in Section 3.3.2.
        March 2001
                                                  3-10
DRAFT-DO NOT QUOTE OR CITE.

-------
 1      PM(ig.25) Concentrations
 2           Using AIRS data from the 1999 PM10 and PM2 5 compliance networks it is possible to
 3      construct a preliminary picture of coarse PM across the country. This is accomplished by pairing
 4      data from nearly 400 compliance monitoring sites where PM10 and PM2 5 monitors are col-
 5      located, and subtracting the mass concentrations of PM2 5 from PM10. The results of this simple
 6      difference  method are shown in Figure 3-6. Because of potential problems with this approach,
 7      the results  should be viewed with caution. Using this approximate method, annual mean
 8      PM(10_2 5) concentrations are as high as 54 j^g/m3, with a nationwide median concentration of
 9      10 /^g/m3.  The higher values occur mainly in the western United States, particularly in
10      California.
                         O  < 4 quarters
                         C  one or more quarters with < 11 samples
                         O  All quarters with at least 11 samples
                         O  All quarters 75% or more complete
             Source: US EPA AIRS Data Base as of 7/12/00 without data flagged as 1, 2, 3, 4, T, W, Y, orX.

       Figure 3-6. 1999 annual mean PMao.2 5) concentrations
       Source: Fitz-Simons et al. (2000).
       March 2001
3-11
DRAFT-DO NOT QUOTE OR CITE

-------
 1     3.2.1  Daily and Seasonal Variability
 2          Data for PM2.5 concentrations obtained as part of the nationwide NAMS/ SLAMS network
 3     during 1999 are summarized in Figures 3-7a through c for individual sites in selected urban areas
 4     across the United States. As far as possible, the cities were chosen because air pollution-health
 5     outcome studies had been performed there, and others were added for the sake of geographic
 6     coverage. At least two sites within each of the seven aerosol characteristic regions of the United
 7     States identified in Chapter 6 of the 1996 PM AQCD and later adopted by the Health Effects
 8     Institute for grouping the results of air pollution-epidemiology studies were chosen. The figures
 9     show the range of 24-h average values within each calendar quarter as box and whisker plots and
10     the annual average concentrations for 1999 are shown above each figure. Because FRM
11     measurements of PM2 s began only in January, data tend to be limited in many areas, especially
12     for the first quarter. As can be seen from the figures, the pattern of seasonal variability for 1999
13     varied across the United States.  At all of the sites shown for the eastern United States, except for
14     the site in Miami, FL, highest quarterly mean values and maximum values occurred during the
15     third quarter (summer) of 1999. This pattern was found, in general, at other sites within the same
16     MSAs,  although there were exceptions.  At sites west of the Mississippi River, highest mean
17     values occurred during the first or fourth quarter (winter or autumn) of 1999, except for the site
18     in Kansas City, where the highest quarterly mean and maximum values occurred in the third
19     quarter. Generally, similar patterns of seasonal variability were found at all other sites within
20     MSAs sampled, although there were exceptions, which may have been related to contributions
21     from local sources as opposed to contributions from regional background sources. At the sites in
22     Miami and Puerto Rico, maximum concentrations occurred during the second quarter, and may
23     have been related to the transport of dust from the Sahara Desert.  Because of the limited nature
24     of these data, definitive conclusions regarding long-term patterns of seasonal variability cannot
25     be drawn from these data alone.
26           Longer time series for making more definitive statements about seasonal variations in PM2 5
27     concentrations are available from a few studies which have had as their goal the characterization
28     of PM2 5 and PM10 concentrations in major urban areas.  The Metropolitan Acid Aerosol
29     Characterization Study (MAACS) (Bahadori et al, 2000b) characterized the levels and the
30     spatial and temporal variability of PM2 5, PM10, and acidic sulfate concentrations in four cities in
31     the eastern United States (Philadelphia, PA; Washington, D.C.; Nashville, TN; and Boston, MA).
        March 2001
3-12
DRAFT-DO NOT QUOTE OR CITE

-------
50-

CO
.E
CO
^.
1
o
O
0-
Bangor, ME
8.9 ug/m3




29
*
25
*
17
*
* .
1
T 29
01
•51
1234
60-



30-

0-
Philadelphia, PA
14.3 |jg/m3
61
*

35
52
*
*
T
OFT
* -*-
* *
1 2 3


69
*
T
0
*
4
• 2301900021 4210100041
                                           150
                                            75-
Atlanta, GA
21.6|jg/m3

74
*
77
*
*
1 2
80
*
81
* *
3 4
1308920011
                         70
                         35-
                                                                      Miami, FL
                                                                     12.1 pg/m3
                                                                       77
                                                                       *
                                                                            81
                                                                            *
                                                                     1234
                                                                      1202510161
70-
CO
j 35-
1
o
O

0-
Detroit, Ml
13.2 ug/m3
24 25 28
* * * 27
*
TT-r
inUH
* * *
1234
2614700051
                        70
                       35-
Steubenville, OH Lexington, KY Baton Rouge, LA
18.3 ug/m3 15.4 ug/m3 15.1 ug/m3

83
*

71
*
62
*
'ml
***-*-
% 	
70-






35-


n .






30
* 31
* 28
-r *
21
*
* *
60-






30-


n



91
*

88
87
* 89
Trlj
* * i *

1234 1234 1234
3908110011 2106700121 2203300091
Figure 3-7a. Quarterly distributions of 24-h average PM25 concentrations obtained in
            eight eastern U.S. cities by the nationwide SLAMS/ NAMS network of PM2 s
            FRM monitors during 1999.  The data show the lowest, lowest tenth
            percentile, lower quartile, median, highest quartile, highest tenth percentile,
            and highest PM2 5 values. Values given above the highest extreme value in the
            graphs refer to the number of observations. Annual average concentrations
            are shown immediately above each graph.
March 2001
3-13
DRAFT-DO NOT QUOTE OR CITE

-------
         40-
    CO
    I
    •3   20
    s
     o
    o
               Fargo, ND
               9.4 (jg/m3
                1234
                3801710041
  Kansas City, MO-KS
      11.6 ug/m3
50 H
25-
                                              89
                                           *  *
       1234
       2904700411
                                                             60-
          30-
                Wichita, KS
                12.5|jg/m3
                                 23
                                 *
                 1234
                 2017300101
    co
    .e
     D)

    O
30-


0-
Cheyenne, WV
5.6 ug/rn3
27
*
T * 24 *
* * *
1234
5602100011
                                      Albuquerque, NM
                                         5.2 ug/m3
                                   40-I
                                   20-
                                            29 33 32
                                          1234
                                          3504310031
                           50-
                           25-
                                Phoenix, AZ
                                 13.1 ug/m3
                                                                           64

                                                                           *
                                                                      69
                                                                         87
                                  1234
                                  0401399911
Figure 3-7b.  Quarterly distributions of 24-h average PM2 5 concentrations obtained in
             six central and mountain U.S. cities by the nationwide SLAMS/ NAMS
             network of PM25 FRM monitors during 1999.  The data show the lowest,
             lowest tenth percentile, lower quartile, median, highest quartile, highest
             tenth percentile, and highest PM2 5 values. Values given above the highest
             extreme value in the graphs refer to the number of observations. Annual
             average concentrations are shown immediately above each graph.
March 2001
      3-14
DRAFT-DO NOT QUOTE OR CITE

-------
          Bellevue, WA
           10.3|jg/m3
70-


1
c
I 35-
IE
8
o
O


n-

88
*
86
*


89 83 T
T * * n
JL Fi 1*1
0 5 H u
* * * *
1234
5303300211
                  Sacramento, CA
                    16.2ug/m3
                        100-
                         50-
                                      61
                                      *
                               29
                               1234
                               0606700101
                                             140-
                                  70-
Bakersfield, CA E. Los Angeles, C>
27.9 ug/m3 25.9 ug/m3
70
*


29
*

-


^
- 88 *
SL^



K

r


•*••**#
100-







50-






n


29
*


29
*
17
*

T ~~
Ml T *
RPRM
x jr y
* * *
*

1234 1234
0602900141 0603716011
San Diego, CA
16.6 ug/m3
70-



co
^
D)
cr
i 35-
j=
s
c
o
O

o-
80
*


74
*


82
T 80 T
*
_L_ pL—
II II
l+l TP T" [+|
[I m R U
* * * *

80-







40-







Anchorage, AK Honolulu, HI San Juan, PR
6.5 ug/m3 5.0 ug/m3 10.3 ug/m3

46
*



28
*

54
*



-i-
* ™ T? *
24-







12-






n



78 :
*
' * 	 (105)
82


T * 81 71
I *
I
II ^r
i | i-*-i
MM*
* * * *
1234 1234 " 1234
40-







20-







0-





66

* 60
59 68
T *
* M *
I I r
j + l 1 —T"
~f" r— i p^| |
P U H n
* * * X
*
1234
0607300031 0202000181 1500310011 7206100051
Figure 3-7c.  Quarterly distributions of 24-h average PM2 5 concentrations obtained in
             seven western U.S. cities and San Juan, PR, by the nationwide SLAMS/
             NAMS network of PM25 FRM monitors during 1999. The data show the
             lowest, lowest tenth percentile, lower quartile, median, highest quartile,
             highest tenth percentile, and highest PM2 s values. Values given above the
             highest extreme value in the graphs refer to the number of observations. For
             Honolulu, HI the highest value observed is shown in parentheses. Annual
             average concentrations are shown immediately above each graph.
March 2001
                            3-15
                                                  DRAFT-DO NOT QUOTE OR CITE

-------
     Seasonal variations in PM2.5 and PM10 concentrations obtained during the course of this study are
     shown in Figure 3-8.
                 a
                 cf
                 o
                 03
                 O
                 O
                 O
                 o
                 05
                 a.
                                       Washington     Nashville      Boston
             -i	r	1	r
SPSU F W  SPSU F W  SPSU F W  SPSU  F W
                     Season
      Figure 3-8. Concentrations of PM2^ and PM10 measured in the four MAACS cities. The
                 data show the lowest, lowest tenth percentile, lowest quartile, median highest
                 quartile, highest tenth percentile, and highest PM2.S 24-h average values. The
                 dashed line shows the level of the annual PM2.5 standard.
      Source: Bahadori et al. (2000b).
1          The data for the four cities included in MAACS are presented as box plots showing the
2     lowest, lowest tenth percentile, lowest quartile, median, highest quartile, highest tenth percentile,
3     and highest PM2.5 and PM10 values in Figure 3-8. Highest PM2 5 and PM10 values are found
4     during the summer in all four cities, and mean values are highest during the summer in all cities,
5     although the seasonal pattern in Boston appears to be more nearly bi-modal with an additional
6     winter peak. This seasonal pattern, based on 2- to 3-year sampling periods for each city during
7      1992 through 1996, is in accord with that obtained from the FRM monitors in the NAMS and
8     SLAMS network.       .
       March 2001
                     3-16
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
     Frequency distributions for PM2 5 concentrations obtained in Philadelphia are shown in
Figure 3-9 (Bahadori et al., 2000b). Concentrations predicted from the log-normal distribution,
using geometric mean values and standard deviations derived from the data, also are shown.
In Philadelphia, the highest PM2 5 values were observed when winds were from the southwest
during sunny but hazy high pressure conditions. In contrast, the lowest values were found after
significant rainstorms during all seasons of the year. Day-to-day concentration differences in the
data set are 6.8 ± 6.5 /^g/m3 for PM25 and 8.6 ± 7.5 vg/m3 for PM10. Maximum day-to-day
                                                         3 for PM10.
 concentration differences are 54.7 fj.g/m3 for PM2 5 and 50.4
                          350
                              0
                                                               PM2.5
                                                 geometric mean = 15.2 yg/m3
                                                              Og=1.69
                                                   "Tnp  I  I  I  i
                                  20   30   40    50    60   70   80
                                     Concentration (ng/m3)
      Figure 3-9. Frequency distribution of 24-h average PM2S concentrations measured at the
                  PBY site in southwestern Philadelphia. Log-normal distribution fit to the data
                  shown as solid line.
1
2
3
4
     Different conclusions could be drawn about data collected elsewhere in the United States.
PM25 concentrations obtained in Phoenix, AZ, from 1995 through 1997 are summarized in
Figure 3-10 and frequency distributions of PM2 5 concentrations obtained in Phoenix are shown
in Figure 3-11. Day-to-day concentration differences in this data set are 2.9 ± 3.0 yug/m3 with a
      March 2001
                                         3-17
                                                          DRAFT-DO NOT QUOTE OR CITE

-------
40 i


£ 30-
OL
o 20-
2
§
£ 10-
o
O
n .
Phoenix, AZ y


PM2.5


^ (n = 876)











N
•
f











n
•
^













•

T








n

•


y




                  Mar-May   June-Aug     Sep-Nov    Dec-Feb
Figure 3-10.  Concentrations of 24-h average PM2.S measured at the EPA site in Phoenix,
            AZ. The data show the lowest, lowest tenth percentile, lowest quartile,
            median (black circles), highest quartile, highest tenth percentile, and highest
            PM2-5 values.
            200
                                                  PM2.5
                                         geometric mean = 10.5 |jg/m
                                                 o=1.70
                 0
10    15    20   25   30    35    40
  Concentration (|jg/m3)
Figure 3-11. Frequency distribution of 24-h average PM2-5 concentrations measured at the
            EPA site in Phoenix, AZ.
March 2001
         3-18
DRAFT-DO NOT QUOTE OR CITE

-------
   1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
 maximum day-to-day concentration difference of 23 /u.g/m3. PM2 5 and PM(10_2 5) data were
 obtained at a number of sites in California on a sampling schedule of every 6 days with
 dichotomous samplers from 1989 through 1998.  Histograms showing the frequency distribution
 of the entire set of PM2 5 and PM(10.2 5) concentrations obtained by the GARB network of
 dichotomous samplers from 1989 to 1998 are shown in Figures 3-12 and 3-13.  Also shown are
 log-normal distributions generated using geometric means and standard deviations derived from
 the data as input. Although the data for both size fractions appear to be reasonably well
 simulated by the function, data obtained at individual locations may not be. Data showing the
 seasonal variability of PM25 obtained at Riverside-Rubidoux are summarized in box plot form in
 Figure 3-14. The frequency distribution of PM25 concentrations obtained at Riverside-Rubidoux
 from 1989 to 1994 is shown in Figure 3-15. It can be seen that the data  are not as well fit by a
 log-normal distribution as can the data shown in Figure 3-9, for example, mainly as the result of
 a significant number of days with PM2 5  >100 ^g/m3.
      An examination of the data  from the four MAACS cities, Phoenix, AZ, and Riverside, CA,
 indicates that substantial differences exist in aerosol properties between  widely separated
 geographic regions.  Fine-mode particles make up most of the PM10 mass observed in the
 MAACS cities and appear to drive the daily and seasonal variability in PM10 concentrations
 there.  Coarse-mode particles represent a larger fraction of PMj0 mass in Phoenix and Riverside
 and drive the seasonal variability in  PM10 seen there. The ratio of PM2 5 to PM10 concentrations is
 much larger in the MAACS cities of Philadelphia (0.72); Washington, DC (0.74); and Nashville
 (0.63) than in either Phoenix (0.34) or Riverside (0.49). Differences between median and
 maximum concentrations in any size fraction are much larger at the Riverside site than at either
 the MAACS or Phoenix sites.  Many of these differences could reflect the more sporadic nature
 of dust suspension at Riverside. In addition, the seasonal variability of PM2 5 concentrations
 observed in Phoenix, AZ, and Riverside, CA, appears to be different from that observed in the
 MAACS cities.  These considerations demonstrate the hazards in extrapolating conclusions about
 the nature of variability in aerosol characteristics inferred at one location to another.

3.2.2  Diurnal (Circadian) Variability
     The variability of PM concentrations on time scales shorter than a day can,  in principle, be
characterized by measurements made by TEOMs and p-gauge monitors that are currently used
       March 2001
                                                3-19
                                                    DRAFT-DO NOT QUOTE OR CITE

-------
        500
          0
                                               PM2.5
                                     geometric mean = 12.8 M9/m3
                                             a. = 2.29
           0   10  20  30  40  50  60  70  80  90 100  110 120 130 140 150
Figure 3-12. Frequency distribution of 24-h average PM2.S measurements obtained from all
            California Air Resources Board dichotomous sampler sites from 1989 to 1998.
                                               PM(10-2.5)
                                    geometric mean = 15.7 pg/m3
                                             
-------


140-
120-
100-
80-
60-
40-
20-
o-l
Fine
(n = 382)














1













i fl 1 i
Tl fj
W

±
J i . i V



« i i i 	 1 	 	 •
Jan - Mar Apr - Jun Jul - Sept Oct - Dec
1st Qtr 2nd Qtr 3rd Qtr 4th Qtr
 Figure 3-14.  Concentrations of 24-h average PM2 5 measured at the Riverside-Rubidoux
              site. The data show the lowest, lowest tenth percentile, lowest quartile,
              median (black squares), highest quartile, highest tenth percentile, and highest
              PM2 5 values.
      100
      80-
    geometric mean = 26.6 M9/rn3
               20    40     60     80    100   120    140   160    180   200
Figure 3-15. Frequency distribution of 24-h average PM25 concentrations measured at the
            Riverside-Rubidoux site.
 March 2001
3-21
                                                  DRAFT-DO NOT QUOTE OR CITE

-------
 1      to provide Air Quality Index (API) information to the public. A description of these methods
 2      was given in Section 2.2.5.  However, as shown in Chapter 2, continuous methods are subject to
 3      artifacts because, in large part, of heating of their inlets, which results in the loss of components
 4      such as water, ammonium nitrate, and semi-volatile organic compounds (cf. Sections 2.2.1.1 and
 5      2.2.3 for further details concerning the chemistry of volatilizable components), so caution should
 6      be used in interpreting results obtained by these techniques.
 7           The composite diurnal variation of PM2 5 concentrations obtained throughout the
 8      continental United States by 31 TEOM and P-gauge monitors reporting to AIRS in 1999 is
 9      shown hi Figure 3-16. As can be seen^ there is a distinct pattern with maxima occurring during
10     the morning and evening. Notable exceptions to this pattern occur in California where broad
11      nighttime maxima and daytime minima occur, which may be related to the use of P-gauge
12     monitors with unheated inlets there.  It should be noted in examining the diurnal variations
13     shown in Figure 3-16, thatthere is substantial day-to-day variability in the diurnal profile of
14     PM2 5 measured at the same location which is then smoothed out after a suitably long averaging
15     period is chosen. The large ratio of the interquartile range to the median values supports the
16     view that there is substantial variability hi the diurnal profiles.
17          The diurnal variability of PM components is determined by interactions between variations
18     in emissions, the. rates of photochemical transformations, and the vertical extent and intensity of
19     turbulent mixing near the surface.  Wilson and Stockburger( 1990) characterized the diurnal
20     variability of sulfate and lead in Philadelphia. At that tune, Pb was emitted mainly by motor
21     vehicles. Pollutants emitted mainly by motor vehicles, such as carbon monoxide, show two
22     distinct peaks occurring during the morning and evening rush hours (see Chapter 3, U.S.
23     Environmental Protection Agency, 2000c). Pollutants, such as sulfate, which are transported
24     long distances hi the free troposphere (i.e.,  above the planetary boundary layer), tend to be mixed
25     downward and have their highest concentrations during the afternoon when the intensity and
26     vertical extent of turbulent mixing are greatest. Secondary aerosol components that are produced
27     by photochemical reactions such as secondary organic compounds may have a daily maximum in
28     the afternoon, by analogy with ozone. PM produced by residential heating (e.g., from wood
29     burning), on the other hand, reach maximum levels during the night.
30           Although the interquartile ranges for hour-to-hour changes in PM2 5 concentrations shown
31      in Figure 3-16  encompass several //g/m3, extreme values for the hour-to-hbur variations can be
        March 2001
3-22
DRAFT-DO NOT QUOTE OR CITE

-------
              30-
           O)
           o
           •S3
           2
           «4—•
           I
           O
           O
       15-
               0-
                                       6646
                                       i—1 6575
6576 6571 6605 6640 6649

    +
                                                                               6704
                                                                               f— i 6583
                   0  1
                                    ~1   I	1	1	1	1	1	1	1	1	1	1	1	1	1	r—
                                     8   9  10 11 12 13  14  15  16  17 18 19 20 21 22 23
                                             Hour
        Figure 3-16. Intraday variability of hourly average PM25 concentrations across the United
                     States. Interquartile ranges, median and mean (+) values are shown. Values
                     above the box plots refer to the number of observations during 1999.
                     Median, mean (+) and interquartile ranges are shown.
        Source: Fitz-Simons et al. (2000).
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
much larger (Fitz-Simons et al., 2000). The 98th percentile values for positive and negative
excursions in concentration are all less than 20 yUg/m3.  Maximum positive excursions were much
larger, ranging from 27 yug/m3 in the Northeast up to 220 //g/m3 in the Southwest, and with
maximum excursions in other regions all less than 125 yug/m3.  It should be borne in mind that
the hour-to-hour changes that are reported reflect the effects of a number of processes occurring
during passage through the sampler inlets and on the TEOM measurement elements. These
considerations add uncertainty to the interpretation of the hour-to-hour changes that are observed,
as discussed earlier in Chapter 2. However, because of the tendency of these monitoring
instruments to lose material by evaporation, the concentrations reported during excursions
probably represent lower limits to the true values that were present.
       March 2001
                                         3-23
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1      3.2.3  Relations Among Particulate Matter in Different Size Fractions
 2      Relations Among PMZ5, PM(I0.2_5), and PM10
 3           Data obtained in 1999 by collocated PM2 5 and PM10 FRM monitors have been used to
 4      calculate the ratio of PM2.5 to PMi0 concentrations and correlations among PM25, PM(10.2.5) and
 5      PMIO concentrations. Results are shown in Table 3-2 for each of the seven aerosol characteristic
 6      regions identified in Chapter 6 of CD 96. As can be seen from the table, the ratio of PM2 5 to
 7      PM10 concentrations tends to be higher in the eastern United States than in the western United
 8      States. This general pattern and the values are consistent with that found for the studies included
 9      in Appendix A to Chapter 6 of CD 96. In that compilation, the mean ratio of PM2 5 to PM10 was
10      0.75 in the East, 0.52 in the central United States, and 0.53 in the western United States.
11      Although a large number of paired entries have been included in Table 3-2, seasonal variations
12     and annual averages in a number of regions could not be determined from the data set because of
13     data sparseness, mainly during the early part of 1999.  It also can be seen in Table 3-2 that the
14     ratio of PM2 5 to PM10 was greater than one for a few hundred measurements.  There are a number
15     of reasons for these results, many of which arise because the ratios are based on two independent
16     measurements. Measurement imprecisions play a role when the ratios are large and especially
17     when concentrations are small. Differences in the behavior of semi-volatile components in the
18     two samplers also could occur. The results also may be the result of errors in sampler placement,
19     field, laboratory, or data processing procedures.
20
21      Ultrafine Particle Concentrations
22           Data for characterizing the levels of ultrafine particles (<0.15-yum AD) and the relations
23     between ultrafine particles and larger particles are sparse. Perhaps the most extensive data set for
24     ultrafine particle properties is that described by Woo et al. (2000) for a site located 10 km to the
25     northwest of downtown Atlanta, GA. Size distributions from 3 to 2000 nm were being measured
26      every 12 min for 24 mo beginning in August  1998. Approximately 89% of the total number of
27     particles were found to be smaller than 100 nm, whereas 26% were found to be smaller than
28      10 nm. Concentrations tend to be lower during the summer than during the winter.
29     No correlation was found between number concentration and either volume or surface area for
30      particle sizes up to 2 jum. Because the total number of particles is concentrated in the smallest
31      size ranges, these results also indicate that fine particle mass does not correlate with the number
        March 2001
3-24
DRAFT-DO NOT QUOTE OR CITE

-------
    o
            ~
                  I
                  3

                 §"
                 BH
     VO
     o
 CO

 CJ
o\

0
                        :
S
                          O
                           I
 E   g
 O   O
      I
     OS   —
     vo   c-
                          •*
                          o
                                   co
                                   d
                                       en
                                       VO
                          VD
                          •


    r-

    o
                                       "~i
                                       d
                       vq
                       c>
                       vo
                       to
                       o
     to
     o
                       VO
                       ir>

                       C5
                       10
                       vq
                       o
                       vq
                       o
                   
     VO
     •*
     rr
	,    O   CO

o    §   SS
     m   ~^
                  3
                     0
                     r-
                         VO
                         r-
    o
    r--
         S!
o
f-

          CO
          CS
        o
        >n
        d
dwi
                 e
                 s
                'So
Northeas
Southeas
         •c   ^
         -b    ^
         w    d>

         *Q    p\

         £   D
wes
         r-
         •*
         0
Ca
         £   tS   -S
          a   c   3
No
Sou
                                     o
                                     CJ
                             •a    E
                             2    s
                                     
-------
 1     of ultrafine particles. The high time resolution of the measurements allows some inferences to be
 2     made about the possible sources of the ultrafine particles. The number of particles larger than
 3     10 nm tends to peak during the morning rush hour (around 8 a.m.) and then to decrease through
 4     the day and to increase again after 6 p.m., consistent with a traffic source.  Particles smaller than
 5     10 nm tend to peak during the mid-afternoon, consistent with nucleation involving products of
 6     active photochemistry (McMurry et al., 2000). More direct relations between particle mass
 7     observed in different size ranges can be obtained using multi-stage impactors. Keywood et al.
 8     (1999) found a correlation between PM2 5 and PM015 of about 0.7, whereas they found
 9     correlations of about 0.96 between PM, and PM25 and between PM25 and PM10 based on samples
10     collected by MOUDIs (Multiple Orifice Uniform Deposit Impactors) in six Australian cities.
11
12     3.2.4  Relations Between Mass and Chemical Component Concentrations
13          Time series of elemental composition data for PM2 5 particles based on X-ray fluorescence
14     (XRF) analyses have been obtained at a few locations across the United States. Time series of
15     components of the organic carbon fraction of the aerosol have not yet been obtained. The filter
16     samples that were collected at the PB Y site in southwestern Philadelphia and were used hi the
17     construction of Figures 3-8 and 3-9 also were analyzed by XRF. Concentrations of the trace
18     elements and correlations between trace elements and the total mass of particles in the PM2 5 size
19     range are shown hi Table 3-3. Also shown hi Table 3-3 are similar results obtained for filter
20     samples collected in Phoenix, AZ.  Filters from both monitoring studies were analyzed by the
21     same X-ray spectrometer at the EPA facility hi Research Triangle Park, NC. As can be seen
22     from inspection of Table 3-3, the analytical uncertainty (given hi parentheses next to
23     concentrations) as a fraction of the  absolute concentration is highly variable, and it exceeds the
24     concentration for a number of trace metals whose absolute concentrations are low, whereas it is
25     very small for abundant elements such as sulfur.
26          There are a number of distinct differences between the two data sets. For instance, sulfate
27     and associated cations and water appear to constitute a major fraction of the composition of the
28     PM hi the Philadelphia data set, whereas they appear to constitute a much smaller fraction of the
29     Phoenix data set. The highest PM2.5 values were observed in Philadelphia during episodes driven
30     by high sulfate abundances and are caused, at least partly, by higher sulfate concentrations.
31     Correlation coefficients between SOJ andPM25 were 0.97 during the summer of 1993. Similar
        March 2001
3-26
DRAFT-DO NOT QUOTE OR CITE

-------
         TABLE 3-3. CONCENTRATIONS OF PM2 s Cug/m3) AND SELECTED ELEMENTS
        (ng/m3) IN THE PM2 5 SIZE RANGE AND CORRELATIONS BETWEEN ELEMENTS
                                      AND PM2 s MASS
              (Values in parentheses refer to analytical uncertainty in X-ray fluorescence
                                      determinations.)

PM2.5
Al
Si
P
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
As
Se
Br
Pb
Philadelphia, PAa
Concentration
17.0±0.8Mg/m3
4.0 (3 1.1) ng/m3
116(21.1)
8.6 (10.3)
2100(143)
5.1 (3.4)
60.4 (4.7)
46.6 (4.2)
4.9(4.1)
8.8(1.8)
0.7 (0.7)
- 3.1 (0.8)
109(10.5)
0.1 (1.4)
7.3 (1.4)
4.8(1.1)
36.9 (3.7)
0.6(1.2)
1.5(0.6)
5.0 (0.9)
17.6(2.5)

r
1
0.10
0.51
0.31
0.92
-0.01
0.50
0.39
0.44
0.37
0.15
0.39
0.50
0.04
0.22
0.25
0.21
0.18
0.63
0.11
0.19
Phoenix,
Concentration
9.4 ± 0.5 ^g/m3
68.9 (27.2) ng/m3
209 (48.4)
7.6(4.5)
408 (30.9)
11.4(2.4)
78.6 (8.2)
76.5 (9.7)
7.2 (3.3)
0.7(1.0)
0.4 (0.4)
4.3 (0.6)
112(15.1)
-0.2 (0.8)
0.4 (0.4)
3.3 (0.7)
12.7(1.7)
1.3(0.6)
0.3 (0.3)
3.1(0.6)

AZb
r
1
0.23
0.35
0.52
0.16
0.13
0.67
0.51
0.44
-0.28
0.41
0.64
0.80
-0.01
0.38
0.69
0.64
0.50
0.40
0.57

"11=1105.
bn = 643.
1
2
correlations between SO4 and PM2 5 were found at a site in northeastern Philadelphia (24 km
distant from the site under discussion) during the summer of 1993.
      March 2001
                                     3-27
                                                    DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
 30
 31
     Concentrations of "crustal elements" (e.g., Al, Si, K, Ca, Ti, Mn, and Fe) constitute a larger
fraction of PM2-5 mass in the Phoenix data set than they do in the Philadelphia data set. Sulfur is
very highly correlated with PM2.5 r = 0.92) in the Philadelphia data set, whereas it is only weakly
correlated r = 0.16) with PM2 5 in the Phoenix data set. Trace metals (e.g., Cr, Co, Ni, Cu, Zn,
As, and Pb) are not well correlated ( 0.04 < r < 0.25) with PM2 5 in the Philadelphia data set,
whereas they are more variably correlated (0.01 < r < 0.69) with PM25 in the Phoenix data set.
The uncertainty in the concentration measurement most probably plays a role in determining a
species' correlation with PM2 5, especially when the analytical uncertainty is high relative to
concentration (e.g., for trace metals such as Co).

3.2.5 Spatial Variability
      Intersite correlation coefficients for PM2 5 can be calculated based on the results of FRM
monitors placed at multiple sites within Metropolitan Statistical Areas (MSAs) across the United
States.  Pearson correlation coefficients for PM2 5  monitors located in the Atlanta, GA; Detroit,
MI; Phoenix, AZ; and Seattle, WA, MSAs are shown in Tables 3-4a through d.  Only sites with
at least 100 measurement days were chosen, and,  furthermore, only days with concurrent
measurements were selected from this subset of monitoring sites.  As can be seen from
Table 3-4a through d, PM2 5 concentrations tend to be highly correlated among sites within all of
the MSAs shown, although there can be exceptions, as shown in the results for Atlanta, GA.
There are a number of factors that could lower intersite correlations. These include field
measurement and laboratory analysis errors, placement of monitors close to active sources, and
transient local events.
      In the Philadelphia area, PM2 5 was found to be strongly correlated (r > 0.9) between seven
urban sites and one background site (Valley Forge, PA) during the summer of 1993 (Suh et al.,
 1995).  Similar relationships  also were found during the summer of 1994 at four monitoring sites
 as part of a separate study (Pinto et al., 1995). The data collected  in these studies also indicate
 that PM2.5 and SOJ concentrations are spatially uniform throughout the Philadelphia
 metropolitan area, and that variability in PMj0 levels is caused largely by variability in PM2 5
 (Wilson and Suh, 1997).
      Three methods for comparing the chemical composition of aerosol databases obtained at
 different locations and times were discussed by Wongphatarakul et al. (1998). Log-log plots of
        March 2001
                                           3-28
DRAFT-DO NOT QUOTE OR CITE

-------
    TABLE 3-4. PEARSON CORRELATION COEFFICIENTS FOR THE SPATIAL
    VARIATION OF PM2 5 CONCENTRATIONS IN SELECTED METROPOLITAN
                            STATISTICAL AREAS
         (AIRS site ID numbers without the state code are used to identify sites.)
(a) Atlanta, GA (n = 46)
0892001 1
0892001 1 1
12100321
12110011
06300911
12100391
08900021
(b) Detroit, MI (n = 52)
09900091
09900091 1
14700051
16300011
16300161
(c) Phoenix-Mesa, AZ (n = 45)
01399901
01399901 1
01399971
01399911
01399921
(d) Seattle-Bellevue-Everett,
WA (n = 78)
03300211
03300211 1
03300571
03300801
03320041 .
06110071

12100321 12110011 06300911 12100391
0.87 0.89 0.86 0.90
1 0.88 0.78 0.91
1 0.93 0.90
1 0.86
1


14700051 16300011 16300161
0.98 0.96 0.97
1 0.94 0.93
1 0.95
1

01399971 01399911 01399921
0.92 0.89 0.84
1 0.97 0.90
1 0.89
1

03300571 03300801 03320041 06110071
0.96 0.95 0.95 0.81
1 0.93 0.92 0.80
1 0.94 0.83
1 0.90
1

08900021
0.48
0.33
0.42
0.39
0.36
1



















 Source: Fitz-Simons et al. (2000); data from EPA Aerometric Information Retrieval System (AIRS).
March 2001
3-29
DRAFT-DO NOT QUOTE OR CITE

-------
 1      chemical concentrations, obtained at pairs of sampling sites accompanied by the coefficient of
 2      divergence (COD) were examined as a way tb provide an easily visualized means of comparing
 3      two data sets'. Examples comparing downtown Los Angeles with Burbank and with Riverside-
 4      Rubidoux are shown in Figures 3-17 and 3-18, respectively. As the composition of two sampling
 5      sites become more similar, the COD approaches zero; as then: compositions diverge, the COD
 6      approaches one. Cluster analyses based on the COD between individual data sets can be used to
 7      determine the degree of similarity among a number of data sets. Correlation coefficients
 8      calculated between components can be used to show the degree of similarity between pairs of
 9      sampling sites. In addition to calculating correlation coefficients for total mass or for individual
10     components, correlation coefficients for characterizing the spatial variation of the contributions
11      from given source types can also be calculated by averaging the correlation coefficients of the set
12     of chemical components that represent the source type. The first two methods could be applied
13     either to aerosol data sets collected at multiple sites within a given geographic region or to
14     aerosol data sets collected at widely different locations or times while the third method is best
15     used to characterize sites within a particular geographic region.
16           Correlation coefficients showing the spatial relations among PM2 5 (total) and contributions
17     from different source categories obtained at various sites in the South Coast Air Basin (SoCAB)
18     Study are shown in Table 3-5. In Wongphatarakul et al. (1998), crustal material (crustal), motor
19     vehicle exhaust (mv), residual oil emissions (residual oil), and secondary PM (sec) were
20     considered as source categories. Al, Si, Fe, and Ca were used as markers for crustal material
21     (crustal). V and Ni were used as markers for fuel oil combustion (residual oil). Pb, Br, and Mn
22     were used as markers for motor vehicle exhaust (mv). NO3", NH4+, and SO4= represent secondary
23     PM components (sec). The average of the correlation coefficients of marker elements within
24      each source category are shown in Table 3-3. Values of rsec and rmv are much higher than those
25      for rcrusta, and rresidual oil throughout the SoCAB suggesting a more uniform distribution of the
               'The COD for two sampling sites is defined as follows:
                                              PM
        where x^ represents the average concentration for a chemical component i at site j, j and k represent two sampling
        sites, and p is the number of chemical components.
        March 2001
3-30
DRAFT-DO NOT QUOTE OR CITE

-------
                                                  Source: V\fongphataraku| (1997)
         10
                            Downtown Los Angeles (pg/m3)

       "Jan. 2 - Dec. 28,1986 (63 data points), 24 hours sampling, sampling every 6 days, dp < 2.5 urn
       **Jan. 2 - Dec. 28,1986 (61 data points), 24 hours sampling, sampling every 6 days, d, < 2.5 urn


Figure 3-17. PM2S chemical components in downtown Los Angeles and Burbank (1986)
             have similar characteristics.

Source: Wongphatarakul, etal. (1998).
March 2001
                                        3-31
DRAFT-DO NOT QUOTE OR CITE

-------
      102
   E  10° -I
   x
   o
   I  1C'1
                 COD=0.230
            10
                                                     Unknown
                                               Source: Wongphatarakul(1997)
                       I	
                      10°
Downtown Los Angeles (|jg/m3)

                                                                       102
      *Jan. 2 - Dec. 28,1986 (63 data points), 24 hours sampling, sampling every 6 days, dp < 2.5 pm
      "Jan. 2 - Dec. 28,1986 (60 data points), 24 hours sampling, sampling every 6 days, dp < 2.5 urn


Figure 3-18. Concentrations of PM2.5 chemical components in Rubidoux and downtown
            Los Angeles (1986). The diagram shows a significant spread in the
            concentrations for the two sites compared with downtown Los Angeles and
            Burbank (Figure 3-13).

Source: Wongphatarakul (1998).
March 2001
                                       3-32
                        DRAFT-DO NOT QUOTE OR CITE

-------
           TABLE 3-5. CORRELATION COEFFICIENTS FOR SPATIAL VARIATION OF
               PM2 s MASS AND DIFFERENT SOURCES FOR PAIRS OF SAMPLING
                         SITES IN THE SOUTH COAST AIR BASIN (1986)

Hawthorne and Rubidoux
Long Beach and Rubidoux
Anaheim and Rubidoux
Downtown Los Angeles and Rubidoux
Burbank and Rubidoux
Hawthorne and Anaheim
Long Beach and Anaheim
Burbank and Anaheim
Downtown Los Angeles and Anaheim
Downtown Los Angeles and Hawthorne
Burbank and Hawthorne
Long Beach and Burbank
Long Beach and Hawthorne
Downtown Long Angeles and Long Beach
Downtown Los Angeles and Burbank
''total
-0.027
0.051
0.066
0.095
0.120
0.760
0.852
0.770
0.827
0.808
0.704
0.731
0.880
0.842
0.928






0.034
0.075
0.105
0.143
0.568
0.599
0.633
0.649
0.653
0.825
rKc





0.768
0.888
0.749
0.804
0.854
0.790
0.737
0.909
0.817
0.960
rmv





0.492
0.504
0.579
0.556
0.669
0.688
0.714
0.861
0.719

rresidualoil





0.170
0.150
0.161
0.233
0.533
0.491
0.295
0.482
0.378

        Source: Wongphatarakul et al. (1998).
1
2
3
4
5
6
7
8
9
contributions from secondary PM formation and automobiles than from crustal material and
localized stationary sources.
     Correlation coefficients in Philadelphia air for PM2 5 (total), crustal components (Al, Si, Ca,
and Fe), the major secondary component (sulfate), organic carbon (OC) and elemental carbon
(EC) are shown in Table 3-6. Because these data were obtained after Pb had been phased out of
gasoline, a motor vehicle contribution could not be estimated from the data. Pb also is emitted
by discrete point sources, such as the Franklin smelter. Sulfate in aerosol samples collected in
Philadelphia arises mainly from long-range transport from regionally dispersed sources (Dzubay
et al., 1988). This conclusion is strengthened by the high correlations in sulfate between different
      March 2001
                                       3-33
                                                        DRAFT-DO NOT QUOTE OR CITE

-------
          TABLE 3-6. CORRELATION COEFFICIENTS FOR SPATIAL VARIATION OF
            PM2 s MASS AND DIFFERENT COMPONENTS FOR PAIRS OF SAMPLING
                                SITES IN PHILADELPHIA (1994)

Castor Ave. and Roxboro
Castor Ave. and NE Airport
Castor Ave. and Broad St.
Roxboro and NE Airport
Roxboro and Broad St.
NE Airport and Broad St
r«
0.92
0.93
0.93
0.98
0.95
0.95
*crustal
0.52
0.47
0.57
0.67
0.90
0.69
rsec
0.98
0.99
0.99
0.98
0.98
0.99
TOC
0.88
0.88
0.85
0.83
0.86
0.84
TEC
0.84
0.77
0.89
0.82
0.79
0.63
rPb
0.43
-0.07
0.11
0.20
0.47
0.11
        Source: Pinto et al. (1995).
 1      monitoring sites and the uniformity in sulfate concentrations observed among the sites.
 2      Widespread area sources (e.g., motor vehicle traffic) also may emit pollutants that are relatively
 3      spatially uniform and are highly correlated between sites with uniform traffic density and
 4      emissions patterns. Very few studies have compared aerosol composition in urban areas to that
 5      in nearby rural areas. Tanner and Parkhurst (2000), for example, found that sulfate constituted a
 6      larger fraction of fine particle mass at rural sites in the Tennessee Valley PM2 5 monitoring
 7      network than did organic carbon. For urban sites, they found the situation was largely reversed,
 8      with organic carbon constituting a larger fraction of aerosol mass than sulfate.
 9
10
11      3.3  SOURCES OF PRIMARY AND SECONDARY PARTICULATE
12          MATTER
13           Information about the nature and relative importance of sources of ambient PM is presented
14     hi this section.  Table 3-7 summarize anthropogenic and natural sources for the major primary
15     and secondary aerosol constituents of fine and coarse particles. Major sources of each
16     constituent are  shown in boldface type.  Anthropogenic sources can be further divided into
17     stationary and mobile sources. Stationary sources include fuel combustion for electrical utilities,
18     residential space heating and industrial processes; construction and demolition; metals, minerals,
19     and petrochemicals; wood products processing; mills and elevators used in agriculture; erosion
       March 2001
3-34
DRAFT-DO NOT QUOTE OR CITE

-------







02
1
O
CO
S
s
w
1
 _
« » S.
o o s^,
en u Z ~ e.g I
O **• g *O **- e OJ
ill Jill
oil 5IJ1-
c3 2- ff> M
^ « M ^ S • s
oi u ~ a T3 —
Illl |i
^ii! ^«
o "S oa i-Q S H
•c 5 "2 •« -a £
=2 « § a 43 .
^o o ^ ^is w rs ^
o S s g £ o S
TJ
•a c
IlJ
^2 3
1 1 1||
S °- M
1 § «,
ft, « •§

C
^c
§* 1
Q, 1 
s • i 2
 3 jj
s =c "5 - .5 a
^ ^ > g >^


G" <«
O "S "3
fj ^^^ -i-i CO
ll lie 1 i
OS 53 g S S 23


























G
c


8
C*H
O
O
no known s
dface type.
S-
g c
SI
|2
§ g
either very
ich compon
w iv
a£
S °
:S U
II
ll
p^
March 2001
                                    3-35
DRAFT-DO NOT QUOTE OR CITE

-------
 1      from tilled lands; waste disposal and recycling; and fugitive dust from paved and unpaved roads.
 2      Mobile, or transportation-related, sources include direct emissions of primary PM and secondary
 3      PM precursors from highway and off-highway vehicles and nonroad sources. In addition to
 4      fossil fuel combustion, biomass in the form of wood can be burned for fuel. Vegetation can be
 5      burned to clear new land for agriculture and for building construction, to dispose of agricultural
 6      and domestic waste, to control the growth of animal or plant pets, and to mange forest resources
 7      (prescribed burning). Also shown are sources for precursor gases whose oxidation forms
 8      secondary particulate matter. A description of the atmospheric chemical processes producing
 9      secondary PM is given in Section 3.4.
10           In general, the nature of sources of fine particulate matter is very different from that for
11      coarse particulate matter. A large fraction of the mass in the fine size fraction is derived from
12     material that has been formed during combustion (primary), has been volatilized in combustion
13     chambers and then recondensed to form primary PM, or has been formed in the atmosphere from
14     precursor gases as secondary PM.  Because precursor gases and fine particulate matter are
15     capable of traveling great distances, it is difficult to identify individual sources of constituents.
16     The coarse PM constituents have shorter lifetimes in the atmosphere, so their impacts tend to be
17     more localized. Only major sources for each constituent within each broad category shown at the
18     top of Table 3-7 are listed. Chemical characterizations of primary particulate emissions from a
19     wide variety of natural and anthropogenic sources as shown in Table 3-7 were given in Chapter 5
20      of 1996 PM AQCD. Summary tables of the composition of source emissions presented in 1996
21      PM AQCD and updates are given in Appendix B. These profiles were based in large measure on
22      the results of various studies collecting source signatures for use in source apportionment studies.
23           Natural sources of primary PM include windblown dust from undisturbed land, sea spray,
24      and plant and insect debris. The oxidation of a fraction of terpenes emitted by vegetation and
25      reduced sulfur species from anaerobic environments leads to secondary PM formation.
26      Ammonium (NH4+) ions, that regulate the pH of particles, are derived from emissions of
 27      ammonia (NHs) gas. Source categories for NH3 have been divided into emissions from
 28      undisturbed soils (natural) and emissions that are related to human activities (e.g., fertilized
 29      lands, domestic and farm animal waste). There is considerable debate about characterizing
 30      emissions from wild fires (i.e., unwanted fires) as either natural or anthropogenic.
 31      Approximately 70 to 90% of wildfires may be ignited directly as the result of human activities,
        March 2001
3-36
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
30
31
 either deliberately through prescribed burning and arson, or inadvertently through the improper
 disposal of flammable material or fugitive sparks (Andreae, 1991), with the remainder ignited by
 lightning strikes.  On the other hand, human intervention also suppresses wildland fires that are
 ignited by natural causes (e.g., lightning strikes). Fire suppression allows the buildup of fire
 fuels and increase the susceptibility of forests to more severe and infrequent fires from whatever
 cause. Prescribed burning may limit the growth of these fuels and the chances for more
 catastrophic fires.
      Receptor models are perhaps the primary means used to estimate source category
 contributions to particulate matter at individual monitoring sites.  Dispersion models (i.e., three-
 dimensional chemistry and transport models) are formulated in a prognostic manner (i.e., they
 attempt to predict species  concentrations using a tendency equation that includes terms based on
 emissions inventories, atmospheric transport, chemical transformations, and deposition).
 Receptor models are diagnostic in their approach (i.e., they attempt to derive source contributions
 based either on ambient data alone or in combination with data from the chemical composition of
 sources). These methods have the advantage that they do not invoke all of the uncertainties
 inherent in emissions inventories or in parameterizing atmospheric transport processes in grid
 point models.  There are two main approaches to receptor modeling. Receptor models such as
 the chemical mass balance (CMB) model (Watson et al., 1990a) relate source category
 contributions to ambient concentrations based on analyses of the composition of ambient
 particulate matter and source emissions samples.  This technique has been developed for
 apportioning source categories of primary particulate matter and was not formulated to include
 the processes of secondary particulate matter formation. In the second approach, various forms
 of factor analysis are used. They rely on the analysis of time series of compositional data from
 ambient samples to derive  both the composition of sources and the source contributions.
 Standard approaches such as factor analysis or Principal Component Analysis (PCA) can
 apportion only the variance and not the mass in an aerosol composition data  set.  Positive matrix
 factorization (PMF) is a recently developed multivariate technique (Paatero and Tapper, 1993,
 1994) that overcomes many of the limitations of standard techniques, such as principal
 components analysis (PCA), by allowing for the treatment of missing data and data near or below
detection limits.  This is accomplished by weighting elements inversely according to their
uncertainties. Standard methods such as PCA weight elements equally regardless of their
       March 2001
                                         3-37
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1      uncertainty.  Solutions also are constrained to yield nonnegative factors. Both the CMB and the
 2      PMF approaches find a solution based on least squares fitting and minimize an object function
 3      and both methods provide error estimates.  For a complete apportionment of mass, all of the
 4      major sources affecting a monitoring site must be sampled for analysis by CMB, whereas there is
 5      no such restriction in the use of PMF.
 6           A number of specialty conference proceedings, review articles, and books have been
 7      published to provide greater detail about source category apportionment receptor models as
 8      described in CD 96. A review of the various methods used to apportion PM in ambient samples
 9      among its source categories was given in Section 5.5.2 of 1996 PM AQCD. The collection of the
10      source category characterization profiles shown in Appendix 3B has been motivated in many
11      cases by the need to use them in receptor modeling applications.
12
13     3.3.1  Source Contributions to Ambient Particulate Matter
14          The results of several source apportionment studies will be discussed in this section to
15     provide an indication of different sources of particulate matter across the United States. First,
16     results obtained mainly by using the chemical mass balance (CMB) approach for estimating
17     contributions to PM2 5 from different source categories at monitoring sites in the United States
18     will be discussed and presented in Table 3-8. More recent results using the PMF approach are
19     included for Phoenix, AZ. Results obtained at a number of monitoring sites hi the central and
20     western United States by using the CMB model for PM10 are shown in Table 3-9.  The sampling
21     sites represent a variety of different source characteristics within different regions of Arizona,
22     California, Colorado, Idaho, Illinois, Nevada, and Ohio.  Several of these are background sites,
23     specifically Estrella Park, Gunnery Range, Pinnacle Peak, and Corona de Tucson, AZ, and
24      San Nicolas Island, CA. Definitions of source categories also vary from study to study. The
25     results of the PM10 source apportionment studies were given in 1996 PM AQCD and are
26     presented here to allow easy comparison with results of PM2 5 source apportionment studies.
27           There are several differences between the source categories shown at the tops of Tables 3-8
28      and 3-9. These differences reflect the nature of sources that are important for producing fine and
29      coarse particulate matter shown in Table 3-7. They also are related to improvements in the
 30      ability to distinguish between sources of similar nature (e.g., diesel and gasoline vehicles, meat
 31      cooking and vegetation burning).  It has been only recently that motor vehicle emissions can be
        March 2001
3-38
DRAFT-DO NOT QUOTE OR CITE

-------








^
8
tr^
^
o

g
H
O
U
8
O
i/l
H-j

O
o
§
o
8.
•
bd
j-J
w

H

















|
1
 O
*=• {/I
o o
. 4)
o o
.£2 3 —
^ OT

« jj
1.J
B-a
o ~
•S05
i-s
^3 c
15
I"
D -"5

o
S 8
11
o >
I
3
•o
o
1
0.
O
p






Sampling Site

 "J — . W> „ 0 C1 ' 0 ~, 00
OON^CMOV !n 2 1— OO" vo' °°.^ •
csfocs^a-fo cscscs r^oo^

,
-- 1 -• 2 1 1 1 1 1 1 1 | 1 1 1 1 ^ 5
•n  co •* ^- i ' cj es' o ov' ~ cj co "* ^ o ° ON

^ fS oo C"~ ^
•^ ^ ^ s' | s s i » i i i i gsp."

* • • . V} ON co \O cN O *O ^O o o
SSS"^ <^'"-~-fm^^'S°^~
cscscs — ^ ooooooS?!oooo~~
5 5 S 2 i S 2 | | | - « >. S .3


3 . I; S -^5 M.. coro-*CSt-~.co — — vo — "^oo —



£2£2 III . 1 • . 1 _ s ^ 3 . S
001— OOOiX „„!,,,„ , _
odvoodcs'J vo m «n CN oq 1 ?c oo8o°> "O
— co — — 3 — — oo ov in 1 cs' — — go 2;
_ _ fc|.| | i 1 1 | | 'g i | n 1 fe~ ^
§§§2|2fifi § 1 | s -a a s b b | |3 1^
§|| 1 1 llillll |l§^2
"^"^Moricn — — — — ^^tn


Co" C" — ' v5 5v
OV '"' ^ O\ °v WV Q. ^.
S oo,-^ S°v~2?°
— ^x "> °° 2— •-_£?§;
If]!! islillliiSff! !
iijii l!!!llf|!lll 1
Iftf! !
Ill ! 11
ii
Is eb
fi .g
I -^1
I j 3 °
cs " o 'S
»1 g 0 g


*•• 3
^ • 1 I"--
w ^ S3 S §
•o * &T3 1
Jj^P^




1
T
i &
s -s

.1 i
g •S'g
g jj . g -a
s .g ta '-3 "2
"" 8 = !§ •"

ca
1
*o
i
>
« J

C3 • S3
^? a
"Secondary and other organic compoi
'Secondary ammonium.
cMeat cooking.
'Vegetative detritus.
'Value represents sum of diesel and g
March 2001
3-39
                                           DRAFT-DO NOT QUOTE OR CITE

-------
O
H
O
U

U
o
en

W

§
O

S
£

S

H
   .2

•a I

io. 1
s g
u
o t>
Ci ^.
C/3
. o
5§M
s^
. o
•-I"
II-
II 1
C O j-
III
11"
u
III
(3 o "H -
||i|
£2>iS
i1'!
_E £
£ |
|1
£ S
o


1
a
P

i
3
e
i
i
!
c

prfppppp»npp
^ co in ON in r** ts ON in ON
vocson
cnooo— ^ocdcirn^f
OtnOOCS'^rcS <5 ii> w o ^>
.s .s .s .s .s .s .s
i£ ^ ^ ^ ^ ? ^
^^ OS ON ON
oo •— —• ^^
oo «
^" ON ^^ r~ --J
S ^ o? f ^ 2 CT
O\ * 00 O OJ  > ^^^ON-
-Jes'S r 2 2 2 o\ S\ ^.
fsi?s&a=25
siell-s'llli
1 iisi n j j a
""••e 2'3£ S § 3citS
xy,S.dl<§f,Jg-«
I a. a a a a 1 1 ! i
i mi nil i
SsJKKSfiSSl^

so so
sS £


O O
O O

O 00
o o

O OS
o —
'P. CO
o —
0 O
O vo
en ON
oo vo
^ OO
—. r-
OO ON

«* o


*n ON



ON
OO
\O ON
oo -7
ON oo
— oo
ON


Bakersfield, CA (Magliano, 1988)
Bakerfield, CA (Chow et al., 1992b)

i/l\O — «nOOOON — — O^rooppTl;
cN^od — odp^ — v6voi^-P^Tj:cstj-p-
«n tn ^" f""" in ''S' in "^sf ON VO OO — ~ «-* OO — ^


ooooooooooooooo
OOOOOOOOOOOOOOO

sS§d§2S§§tS§§5§

2S§2^^33S3--3--S
~ f- O in *— • — t— t— •— • ^ "^1; ON CS ON ON
£J S2, ° — ******£ """(Srnc^cSfScS^
voO*nroroON — ro
inONroOO -'^SS'O'O'O'OJZ '^v
I 1 | 1 =1 i | f |l i 1 1 i ll
Otgfififi^j'jjsi.oiaSeieJcoS
March 2001
3-40
DRAFT-DO NOT QUOTE OR CITE

-------
O
H


I
H

tt
O
U
w
I
I
 u
 ^
         Ijl
tri
           So
           OT
                   p
                   o
                         p
                         o
                         p
                         o
                         0V  O
                                 O  O

                                 m  Tf
      p  p
      O  O
                                O  «

                                O  O
                                p  p
                                o  o
      °,  =?
      o  o
                                           p
                                           O
                                         p
                                         o
              p  p
              o  o
                                          p rn
                                          O oo*
                   p
                   •o
                   p
                   o
                        p
                        o
p
o
                                         o
                                         o
                           p

                           C5
                                   p
                                   C5
                         O

                         Tf
              p  p
              o  o
                                         —'   z

                                g" ^
                                ^  g.
March 2001
                                    3-41
                    DRAFT-DO NOT QUOTE OR CITE

-------
 1      broken down into contributions from diesel and gasoline vehicles through the use of organic
 2      tracers.  Meat cooking is also distinguished from vegetation burning in more recent studies,
 3      although both are considered to be part of biomass burning. Vegetation burning consists of
 4      contributions from residential fuel wood burning, wild fires, prescribed burning and burning of
 5      agricultural waste.  Miscellaneous sources of fine particles include contributions from
 6      combustion sources, whereas miscellaneous sources of coarse particles consist of contributions
 7      from soil and sea spray and industrial processing of geological material (e.g., cement
 8      manufacturing).  Although a large number of elements and chemical compounds are used to
 9     differentiate among source categories, it can be seen from Tables 3-8 and 3-9 that only a
10     relatively small number of sources are needed to account for the mass of PM25 and PM10.
11            Secondary sulfate is the dominant component of PM2 s samples collected in the studies of
12     Dzubay et al. (1988) and Glover et al. (1991).  Both studies found that sulfate at their monitoring
13     site arose from regionally dispersed sources.  Sulfate,  associated cations and water also represent
14     the major components of PM2 5 found in monitoring studies in the eastern United States
15     (Table 3-1). Motor vehicle emissions, arising mainly from diesels, are other major sources of
16     PM2 5. Contributions from road dust and soils are relatively minor, typically constituting less
17     than 10% of PM2 5 hi the studies shown in Table 3-8.  Studies hi the western United States shown
18     in Table 3-8 have found larger contributions from motor vehicles, fugitive dust and ammonium
19     nitrate. The most notable difference in the relative importance of major source categories of
20     PM2.S shown in Table 3-8 and PM10 shown in Table 3-9 involves crustal material, (e.g., soil, road
21     dust), which represents about 40% on average  of the total mass of PM10 in the studies shown in
22     Table 3-9.  The fraction is higher in locations located away from specific sources such as sea
23     spray or smelters. Emissions of fugitive dust are concentrated mainly in the PM(10.2 5) size range.
24     The average fugitive dust source contribution is highly variable among sampling sites within the
25     same urban areas, as seen by differences between the  Central Phoenix (33 yug/m3) and Scottsdale
26     (25 ^g/m3) sites in Arizona, and it is also seasonally variable, as evidenced by the summer and
27     fall contributions at Rubidoux, CA.  The variability in fugitive dust loadings reflects the sporadic
28     nature of its emissions and its short lifetime in the atmosphere.
29           In Table 3-9, primary motor vehicle exhaust contributions account for up to 40% of average
30     PMj0 at many of the sampling sites.  Vehicle exhaust  contributions are also variable at different
31      sites within the same study area.  The mean value and the variability of motor vehicle exhaust
        March 2001
3-42
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 contributions reflects the proximity of sampling sites to roadways and traffic conditions during
 the time of sampling. Many studies were conducted during the late 1980s, when a portion of the
 vehicle fleet still used leaded gasoline,  Pb and Br in motor vehicle emissions facilitated the
 distinction of motor vehicle contributions from other sources. Vehicles using leaded fuels have
 higher emission rates than vehicles using unleaded fuels.  Pb also poisons automobile exhaust
 catalysts and produces adverse human health effects. As a result, Pb virtually has been
 eliminated from vehicle fuels.  However, organic species have replaced Pb as a source marker for
 motor vehicle emissions (e.g., Rogge et al., 1993).
      Marine aerosol is found, as expected, at coastal sites such as Long Beach (average 3.8% of
 total mass) and San Nicolas Island (25%). These contributions are relatively variable and are
 larger at the more remote sites. Individual values reflect proximity to local sources.  Of great
 importance are the contributions from secondary ammonium sulfate in the eastern United States
 and ammonium nitrate in the western United States. These are especially noticeable at sites in
 California's San Joaquin Valley (Bakersfield, Crows Landing, Fellows, Fresno, Kern Wildlife,
 and Stockton) and in the Los Angeles area.
      Samples selected for chemical analysis are often biased toward the highest PM10 mass
 concentrations in the studies shown in Table 3-9, so average source contribution estimates are
 probably not representative of annual  averages. For example, the study by Motallebi (1999)
 considered only days when the PMIO concentration was greater than 40 vg/m3. Quoted
 uncertainties in the estimated contributions of the individual sources shown in Table 3-6 range
 from 10 to 50%. Uncertainties of source contribution estimates are not usually reported with the
 average values summarized in Table 3-9. Estimates of standard errors are calculated in source
 apportionment studies, and typically range from 15 to 30% of the source contribution estimate.
 They are much higher when the chemical source profiles for different sources are highly
 uncertain or too similar to distinguish  one source from another.
     Very few source apportionment studies using the CMB modeling technique have examined
the spatial variability of source contributions at different sites within an urban area. As can be
seen from Table 3-8, Dzubay et al. (1988) found a uniform distribution of sulfate among the NE
Airport in Philadelphia, PA; downtown Camden, NJ; and Clarksboro, NJ, during the summer of
 1982. The longest distance between two monitoring sites (NE Airport and Clarksboro) was
approximately 40 km. Magliano et al. (1998) examined the spatial variability of PMIO source
       March 2001
                                         3-43
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1      contributions at a number of sites in Fresno and Bakersfield, CA, during the winter of 1995-1996
 2      and reported values for 1 day, December 27,1995.  During that day, mobile sources contributed
 3      from 13.0 to!5.8 ^g/m3, vegetation burning contributed from 5.1 to 11.1 Aig/m3, ammonium
 4      sulfate contributed 2.4 to 3.4 (J-g/m?, and ammonium nitrate contributed 19.3 to 24.6 //g/m3 to
 5      PM10 at the sites in Bakersfield. Mobile sources contributed 13.9 to 22.5 Atg/m3, vegetation
 6      burning contributed 8.2 to 15.7 Aig/m3, ammonium sulfate contributed 1.8 to 2.3 Atg/m3, and
 7      ammonium nitrate contributed 14.5 to 18.9 Atg/m3 at the sites in Fresno. All of these components
 8      are expected to be found mainly in the PM2 5 size fraction. As can be seen, source contributions
 9      at different sites varied by factors of 1.2 to 2.2 in Bakersfield and by factors of 1.3 to 1.9 in
10      Fresno on that day.
11
12     3.3.2  Long-Range Transport  of Participate Matter from Sources Outside the
13            United States
14          Apart from sources within the continental United States, particulate matter can be brought
15     in by long-range transport from sources outside the United States. For example, the transport of
16     PM from uncontrolled biomass burning in Central America and southern Mexico resulted in
17     anomalously high PM levels observed in southern Texas and generally elevated PM
18     concentrations throughout the entire central and southeastern United States during the spring and
19     early summer of 1998.  Windblown dust from individual dust storms in the Sahara desert has
20     been observed in  satellite images as plumes crossing the Atlantic Ocean and reaching the
21     southeast coast of the United States (e.g., Ott et al., 1991). Dust transport from the deserts of
22     Asia across the Pacific Ocean also occurs (Prospero, 1996). Most dust storms in the deserts of
23     China occur in the spring following the passage of strong cold fronts after the snow has melted
24     and before a surface vegetation cover has been established.  Strong winds and unstable
25     conditions result in the rapid transport of dust into the middle and upper troposphere (4 to 5 km
26      altitude), where it is transported by strong westerly winds out over the Pacific Ocean (Duce,
27      1995). Satellite images have been used to track the progress of a dust cloud from the Gobi desert
28      to the northwestern United States during the spring of 1998 (Husar et al., 2000).
29           Satellite images obtained at visible wavelengths cannot track mineral dust across the
30      continents because of a lack of contrast between the plume and the underlying surface. Other
 31      means must be used to track the spread of North African dust through the eastern United States.
        March 2001
3-44
DRAFT-DO NOT QUOTE OR CITE

-------
  1      Perry et al. (1997) used two criteria (PM25 soil concentration > 3 //g m"3 and Al/Ca > 3.8) to
  2      distinguish between soil of local origin from soil originating in North Africa in characterizing the
  3      sources of PM in aerosol samples collected in the IMPROVE (Interagency Monitoring of
  4      Protected Visual Environments) network.  Their analysis indicates that incursions of Saharan
  5      dust into the continental United States have occurred, on average, about three times per year from
  6      1992 to 1995. These events have persisted for about 10 days, principally during the  summer.
  7      As can be expected, the frequency of dust events is highest in the southeastern United States;
  8      about half are observed only within the state  of Florida, with these being associated with dense
  9      hazes in Miami (Figure 3-19) during the summer (Prospero et al., 1987), such that African dust is
10      the dominant aerosol constituent in southern Florida during the summer (Prospero, 1999). The
11      mass median diameter of mineral dust over the oceans is typically between 2 and 3 //m (Duce,
12      1995).  North African dust has been tracked as far as Illinois (Gatz and Prospero, 1996) and to
13      Maine (Perry et al., 1997). Larger scale events typically covered from 15 to 30% of the area of
14      the continental United States and resulted in  increases of PM2 5 levels of 8.7 ± 2.3 //g m"3
15      throughout the affected areas, with mean maximum dust contributions of 19.7 ± 8.4 /^g m"3
16      during these events and a peak contribution of 32 /wg m"3 to 24-h average PM2 5 levels.
17           Husar et al. (2000) documented transport of dust from the Gobi and Taklimakan deserts to
18      North America during April 1998. The PM10 concentration averaged over 150 stations in
19      Washington, Oregon, California, Nevada, and Idaho reporting data to AIRS was 65 yug/m3
20      between April 26 and May 1, compared to about 20 yug/m3 during the rest of April and May.
21      Data from the IMPROVE network indicated  that PM10 concentrations were over 100 yug/m3 in
22      central  British Columbia, Washington State,  and Oregon. The highest PM concentrations
23      observed were 120 /^g/m3 for PM10 and 50 yug/m3 for PM2 5 at Chilliwack Airport in northwestern
24      Washington State (Figure 3-20). Aircraft measurements made over the northwestern United
25      States were consistent with a mass median diameter of the dust being between 2 and 3 yum.
26      Three-dimensional model simulations of the transport of inert tracers from Asia indicate that
27      substantial additions to PM concentrations also occurred throughout the north and mid-western
28      United  States and southwestern Canada (Hanna et al., 1999).
29           Biomass burning for agricultural purposes occurs normally during the spring of each year in
30      Central America and southern Mexico. During the spring of 1988, fires burned uncontrollably
31      because of abnormally hot and dry conditions associated with the intense El Nino of 1997 to
       March 2001
3-45
DRAFT-DO NOT QUOTE OR CITE

-------
                                                   Year
       Figure 3-19. Monthly average Saharan dust components of the aerosol sampled in Miami,
                    FL, during 1974 to 1996.
       Source: Prospero (1999).
 1     1998. PM10 concentrations observed in the southern Rio Grande Valley were elevated
 2     substantially during the passage northward of the biomass burning plume produced by these fires
 3     as shown in Figure 3-21. Elevated PM2 5 and PM10 concentrations also were found as far north as
 4     St. Louis, MO (Figure 3-22).  As can be seen from Figure 3-21 and Figure 3-22, the elevations in
 5     PM concentrations were limited in duration. .However, uncontrolled wildfires occur in the
 6     United States every year, but their effects on air quality throughout the United States still need to
 7     be evaluated systematically. These fires can be widespread.  For example, approximately
 ,8     26,000 km2 were consumed during 2000  in the western United States.
 9          Wildfires also occur in the boreal forests  of northwestern Canada. Wotowa and Trainer
10     (2000) suggested that the plume from fires occurring in the Northwest Territories of Canada in
       March 2001
3-46
DRAFT-DO NOT QUOTE OR CITE

-------
               140
                 4/25/98   4/27/98   4/29/98    5/1/98    5/3/98    5/5/98
       Figure 3-20.  PM2^ and PM10 concentrations measured at Chilliwack Airport, located in
                   northwestern Washington State, just before and during the Asian desert dust
                   episode of April and May 1998.
       Source: U.S. EPA Aerometric Information Retrieval System (AIRS).
1
2
3
4
5
6
early July 1995 may have extended throughout most of the eastern United States, resulting in
elevated levels of carbon monoxide (CO) and ozone. Simple scaling of their calculated excess
CO concentrations because of the fires by the ratio of emission factors of PM25 to CO indicates
that the excess PM25 concentrations in the plume may have ranged from about 5 /wg/m3 in the
Southeast and increasing to close to 100 ^g/m3 in the northern Plains States.
      March 2001
                                        3-47
                                                         DRAFT-DO NOT QUOTE OR CITE

-------
        600
    O
          5/5 5/8  5/7  5/8  5/9  5/10 5/11 5/12 5/13 5/14 5/15 5/16 5/17 5/18 5/19 5/20 5/21 5/22 5/23 5/24 5/25 5/26 5/27 5/28
                                          Time, days

Figure 3-21. Time series of 24-h average PM18 concentrations observed in the Rio Grande
             Valley during May 1998.

Source: U.S. EPA Aerometric Information Retrieval System (AIRS).
                            200
                        cb
    180-
    160-
    140-
    120-
    100-
.x
?    80-
>
     60-
     40-
     20
      0
                                                        Smoke
                                                        Event
                                 Pm10 24 hr Standard
                                5  6  7 8  9 10 11121314151617
                                             May 1998

Figure 3-22. PM,0 concentrations observed in St. Louis, MO, during May 1998.
Source: U.S. EPA Aerometric Information Retrieval System (AIRS).
March 2001                               3-48        DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
27
28
29
30
31
 3.4  EMISSIONS ESTIMATES AND THEIR UNCERTAINTIES
      In principle, source contributions to ambient PM also could be estimated on the basis of
 predictions made by chemistry-transport models (CTM) or even on the basis of emissions
 inventories alone. Uncertainties in emissions inventories have arguably been regarded as
 representing the largest source of uncertainty in CTMs (Calvert et al., 1993).  Apart from
 uncertainties in emission inventories, a number of other factors limit the ability of an emissions
 inventory driven CTM to determine the effects of various sources on particle samples obtained at
 a particular location.  Air pollution model predictions represent averages over the area of a grid
 cell, which in the case of the Urban Airshed Model typically has been 25 km2 (5 km x 5 km).
 The contributions of sources to pollutant concentrations at a monitoring site are controlled
 strongly by local conditions that cannot be resolved by an Eulerian grid-cell model. Examples
 would be the downward mixing of tall stack emissions and deviations from the mean flow caused
 by buildings.  The impact of local sources at a particular point in the model domain may not be
 predicted accurately, because their emissions would be smeared over the area of a grid cell or if
 the local wind fields at the sampling point deviated significantly from the mean wind fields
 calculated by the model. CTMs also have problems in predicting pollutant concentrations
 because of uncertainties in vertical mixing and in predicting concentrations of pollutants from
 stationary combustion sources resulting from uncertainties in estimates of plume rise.
     Estimates of nationwide emissions of primary PM2 s and gaseous precursors to secondary
 PM formation are given in Section 3.4.1. Uncertainties in emissions estimates are discussed in
 Section 3.4.2.

 3.4.1 Emissions Estimates for Primary Particulate Matter and Sulfur
      Dioxide, Nitrogen Oxides, and Volatile Organic Compounds in the
      United States
     Estimated emissions of primary PM2 5 from different sources in the United States are
 summarized in Figure 3-23. The estimates are based on information presented in the EPA
National Air Pollutant Emission Trends Report, 1900-1998 (U.S. Environmental Protection
Agency, 2000b), to which the reader is referred for detailed tables showing trends in PM2 5
emissions from a number of source categories froml 990 to 1998, descriptions of the
methodology used in the construction of these tables, and descriptions of the uncertainties
       March 2001
                                        3-49
DRAFT-DO NOT QUOTE OR CITE

-------
               PMo.Total Emissions (1998) = 8.4 Tg  yr
                                -1
                         On-Road
                           2.4%
        Non-Road
          4.9%
                  Incineration
                      0.6%
                  Industry
                    4.7%
                  Wind Erosion
                      9.5%
          Fossil Fuel
          Combustion
             5.2%
             Biomass
              Burning
              17.1%
                          Agriculture
                            11.5%
                                                                    Fugitive Dust
                                                                        44.2%
      Figure 3-23.  1998 directly emitted national particulate matter (PM2.5) emissions by
                  principal source categories for nonfugitive dust sources (see Section 3.4.2 for
                  discussion of uncertainties associated with emissions estimates).

      Source: U.S. Environmental Protection Agency (2000b).
1     involved in the estimation process. This document also provides information about emissions of

2     PM10, sulfur dioxide (SO2), nitrogen oxides (NOJ, volatile organic compounds (VOC), and

3     ammonia (NH3). Although uncertainties associated with these estimates are not given, a

4     discussion of uncertainties in emissions estimates is given in Section 3.4.2.

5          Estimated total nationwide emissions of primary PM2 5 were 8.4 Tg year' in 1998. The

6     category of fossil fuel combustion referred to in Figure 3-23 includes fossil fuel burning by

7     electric utilities, industry, and residences.  The industry category includes contributions from
      March 2001
3-50
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
31
 metals processing, petroleum refining, agricultural products processing, mining, and the storage
 and transport of industrial goods. Incineration refers to the burning of nonbiomass waste by
 residences and municipalities. The on-road vehicles category includes contributions from
 gasoline- and diesel-powered vehicles. The nonroad engines and vehicles category includes
 contributions from transportation (marine vessels, aircraft, trains, etc.); construction; and other
 commercial, industrial, and recreational activities. Wind erosion refers to the raising of crustal
 material by the wind. The biomass burning category includes contributions from residential
 wood burning, open burning of vegetation, and forest fires.  The agriculture category includes
 contributions from emissions of crustal material related to the production of agricultural crops
 and livestock. Fugitive dust refers mainly to crustal material raised by on-road and nonroad
 vehicles during their operation. As can be seen from inspection of Figure 3-23, the raising of
 crustal material by wind erosion, agriculture, and as fugitive dust emissions constitutes the
 largest source (65.2%) of primary PM2 5 on a nationwide basis. Note that wind erosion emissions
 are difficult to interpret, owing to the relatively short duration of wind gusts. Biomass burning
 constitutes the second largest source (17.1%) of primary PM25. The gross  composition of
 emissions from most of these categories is summarized in Table 3-5 in the EPA report, National
 Air Pollution Emission Trends, 1900-1998 (U. S. Environmental Protection Agency, 2000b).
 Total emissions of primary PM25, as well as contributions from individual  source categories,
 were relatively constant over the period from 1990 to 1998 (U.S. Environmental Protection
 Agency, 2000b).
     Estimated contributions from individual sources to emissions of gaseous precursors to
 secondary PM formation are summarized in Figure 3-24 for  SO2 , NOX , VOC, and NH3.
 Information about the yield of PM formed during the oxidation of VOC is given in Section 3.4.
     Although total emissions  of gaseous precursors (SO2, NOX, VOC, and NH3) are shown in
 Figure 3-24, it should be remembered that these values cannot be translated directly into
 production rates of PM. Dry deposition and precipitation scavenging of some of these gases can
 occur before they are converted to PM in the atmosphere.  In addition, some fraction of these
 gases are transported outside of the domain of the continental United States before being
 oxidized. Likewise, emissions of these gases from areas outside the United States can result in
the transport of their oxidation products into the United States.  Although the chemical oxidation
of SO2 will lead quantitatively to the formation of SO4=, the yield ofparticulate matter from the
       March 2001
                                                 3-51
                                                     DRAFT-DO NOT QUOTE OR CITE

-------
                 SO2 Total Emissions (1998)

                        = 18Tgyr1
       FbelCcmtantton-
       Efccfricil UBify ~
                                       Fuel Combustion-
                                         Industrial
Fuel Combustion-
   Other
                                          — Metals Processing

                                           On-Road and
                                          Non-Road Engines
                                           and Vehicles
                                                         On-Road and
                                                        Non-Road Engines
                                                         and Vehicles
                         NOX Total Emissions (1998)
                               = 21 Tg yr1
                                                                                         Combustion-
                                                                                       Electrical Utility
                                                                                           Fuel CombusBon-
                                                                                             Othor
                                                Fuel Combustion-
                                                  Inductrial
                 VOCs Total Emissions (1998)
                        = T7Tgyr1
                         NH3 Total Emissions (1998)

                               = 2.9Tgyr1
              Another
                                         Solvent Utilization
          On-Road and
         Non-Road Engines
          •ndVehidos
               On-Road and
              Non-Road Engines —v
               and Vehicles   V


                  MOlher
                                                         Chemical & Allied _^\
                                                          Product Mfg.
                                                    Miscellaneous
                                                   (includes livestock
                                                    and fertilizer)
       Figure 3-24. Nationwide emissions of SO2, NOX, VOC, and NH3 from various source ^
                     categories (see Section 3.4.2 for discussion of uncertainties associated with
                     emissions estimates).

       Source: U.S. Environmental Protection Agency (2000b).
 1       oxidation of VOC will be much less because only a small fraction of VOC react to form

 2       particles, and those that do have efficiencies less than 10% (see Section 3.4).

 3            The values shown in this section are based on annual totals. However, annual averages do

 4       not reflect the seasonality of a number of emissions categories.  Residential wood burning in

 5       fireplaces and stoves, for example, is a seasonal practice that reaches its peak during cold

 6      weather. Cold weather also affects motor vehicle exhaust particulate matter emissions, both in

 7      terms of chemical composition and emission rates (e.g., Watson et al., 1990b; Huang et al.,

 8      1994). Planting, fertilizing, and harvesting are also seasonal activities.  Forest fires occur mainly

 9      during the local dry season and during periods of drought. Maximum dust production by wind

10      erosion in the United States occurs during the spring, whereas the minimum occurs during the
        March 2001
           3-52
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
 summer (Gillette and Hanson, 1989). Efforts are being made to account for the seasonal
 variations of emissions in the nationwide emissions inventories.
      Trends in nationwide, annual average concentrations of PM10, and precursor gases (SO2,
 NO2, and VOC) over the 10 years from 1989 to 1998 are shown in Table 3-10. As can be seen
 from Table 3-10, there have been substantial decreases in the ambient concentrations of
 PM10,SO2 and NO2. Not enough data are available to define trends in concentrations of VOC;
 there also have been substantial decreases in the emissions of all the species shown in
 Table 3-10, except for NO2, although its average concentration has decreased by 14%. These
 entries suggest that decreases in the average ambient concentration of PM10 could have been
 produced by both decreases in emissions of primary PM10 and the formation of secondary PM,0.
 The large reductions in ambient SO2 concentrations have resulted in reductions in sulfate
 formation that would have been manifest in PM2_5 concentrations on the regional scale in the
 eastern and central United States, where sulfate has constituted a larger fraction of PM25 than in
 the West. Likewise, reductions in NO2 concentrations would have had a more noticeable impact
 on PM2 5 concentrations in the western United States  than in the eastern United States because
 nitrate is a larger component of the aerosol in the western United States. Trends hi aerosol
 components (i.e., nitrate, sulfate, carbon, etc.) are needed for a more quantitative assessment of
 the effects of changes in emissions of precursors.  Measurements of aerosol nitrate and sulfate
 concentrations have been obtained at North Long Beach and Riverside, CA, since 1978
 (Dolislager and Motallebi, 1999). Downward trends  in aerosol nitrate have tracked downward
 trends in NOX concentrations, and SO2 and sulfate concentrations have both decreased. However,
 the rate of decline of sulfate has been smaller than that of SO2 indicating the long range transport
 of sulfate from outside the air shed may be an important source in addition to the oxidation of
 locally generated SO2. There are a number of reasons why pollutant concentrations do not track
 estimated reductions in emissions. Some of these reasons are related to atmospheric effects such
 as meteorological variability and secular changes in the rates of photochemical transformations
 and deposition (U.S. Environmental Protection Agency, 2000b). Other reasons are related to
uncertainties in ambient measurements and in emissions inventories.
       March 2001
                                         3-53
DRAFT-DO NOT QUOTE OR CITE

-------
         TABLE 3-10.  NATIONWIDE CHANGES IN AMBIENT CONCENTRATIONS AND
              EMISSIONS OF PM10 AND GASEOUS PRECURSORS TO SECONDARY
                          PARTICULATE MATTER FROM 1989 TO 1998

PM10
PM2.5
SO47SO2
N03-/NOX
VOC
% Change
Ambient Concentration^
-25%
Urban east -5%(2)
Rural east +3%(2)
Rural west -11%(2)
-39% (sulfate)
-14% (nitrate)
—
1989-1999
Emissions(1)
-35%
+5% (1990 to 1998)
-16%(SO2)
+2% (NOX)
-20%
        Source: (1) U. S. Environmental Protection Agency (2000a); (2) U. S. Environmental Protection Agency (2000d).
 1     3.4.2  Uncertainties of Emissions Inventories
 2          As described in the 1996 PM AQCD, it is difficult to assign uncertainties quantitatively to
 3     entries in emissions inventories. Methods that can be used to verify or place constraints on
 4     emissions inventories are sparse. In general, the overall uncertainty in the emissions of a given
 5     pollutant includes contributions from all of the terms used to calculate emissions (i.e., activity
 6     rates, emissions factors, control device efficiencies). Additional uncertainties arise during the
 7     compilation of an emissions inventory because of missing sources and computational errors.
 8     The variability of emissions can cause errors when annual average emissions are applied to
 9     applications involving shorter time scales.
10          Activity rates for well-defined point sources (e.g., power plants) should have the smallest
11     uncertainty associated with their use, because accurate production records need to be kept.
12     On the other hand, activity rates for a number of really dispersed fugitive sources are difficult to
13     quantify. Emissions factors for easily measured fuel components that are released quantitatively
14     during combustion (e.g., CO2, SO2) should be the most reliable. Emissions of components
15     formed during combustion are more difficult to characterize as the emissions rates are dependent
       March 2001
3-54
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
31
 on factors specific to individual combustion units and on combustion stage (i.e., smoldering or
 active). Although the AP-42 emissions factors (U.S. Environmental Protection Agency, 1995)
 contain extensive information for a large number of source types, these data are very limited in
 the number of sources sampled. The efficiency of control devices is determined by their design,
 their age, their maintenance history, and operating conditions. It is virtually impossible to assign
 uncertainties in control device performance because of these factors. It should be noted that the
 largest uncertainties occur for those devices that have the highest efficiencies (>90%).  This
 occurs because the efficiencies are subtracted from one, and small errors in assigning efficiencies
 can lead to large errors in emissions,   -y
      Ideally, an emissions inventory should include all major sources of a given pollutant.  This
 may be an easy task for major point sources. However, area sources of both primary PM and
 precursors to secondary PM formation are more difficult to characterize than point sources and
 thus, they require special emphasis when preparing emission inventories.  Further research is
 needed to better characterize the sources of pollutants to reduce this source of uncertainty.  Errors
 can arise from the misreporting of data, and arithmetic errors can occur in the course of
 compiling entries from thousands of individual sources.  A quality assurance program is required
 to check for outliers and arithmetic errors.
     Because of the variability in emissions rates, there can be errors in the application of
 inventories developed on an annually averaged basis (as are the inventories shown in
 Figures 3-23 to 3-24) to episodes occurring on much shorter time scales.  As an example, most
 modeling studies of air pollution episodes are carried out for periods of a few days.
     Uncertainties in annual emissions were estimated to range from 4 to 9% for SO2 and from
 6 to 11% for NOX in the 1985 NAPAP inventories for the United States (Placet et al., 1991).
 Uncertainties in these estimates increase as the emissions are disaggregated both spatially and
 temporally.  The uncertainties quoted above are minimum estimates and refer only to random
 variability about the mean, assuming that the variability in emissions factors was adequately
 characterized, and that extrapolation of emissions factors to sources other than those for which
they were measured is valid.  The estimates do not consider the effects of weather or variations in
operating and maintenance procedures.
     Fugitive dust sources, as mentioned above, are extremely difficult to  quantify, and stated
emission rates may represent only order-of-magnitude estimates. Although crustal dust
        March 2001
                                          3-55
                                                             DRAFT-DO NOT QUOTE OR CITE

-------
 1      emissions constitute over 60% of the total PM2 5 inventory, they constitute less than about 15% of
 2      the source strengths inferred from the PM composition data shown in Table 3-1  and the receptor
 3      modeling studies shown in Table 3-6. However, it should be remembered that secondary
 4      components (sulfate, nitrate, and some fractions of organic carbon) often represent a significant
 5      fraction of ambient samples.  Therefore, this discrepancy is smaller than the factor of four that is
 6      obtained by comparing primary dust emissions to the sum of primary and secondary components
 7      in the ambient aerosol. The reasons for this apparent discrepancy are not clear.  In addition to
 8      errors in inventories or source apportionments, weather-related factors (wind speed and ground
 9     wetness) and the dominance of local sources on spatial scales too small to be captured in
10     inventories may be involved. It  should be remembered that dust  emissions are dispersed widely
11      and are highly sporadic. Dust particles also have short atmospheric residence times and  as a
12     result their dominance in emissions inventories may not be reflected in samples collected near
13     specific sources.
14          Although mineral dust sources account for most  of the emissions, their contributions are
15     distributed much more widely than are those from combustion sources. Watson and Chow
16     (1999) examined a number of possible causes for this discrepancy. In large part, it is related to
17     the method used to obtain emissions factors for fugitive dust. The standard methods use data
18     obtained by particle monitors stacked at several elevations from  1 to 2 m up to 7 to  10 m above
19     the surface. However, small-scale turbulent motions, not stable winds, characterize atmospheric
20     flow patterns immediately adjacent to the surface (Garrart, 1994). The depth of this turbulent
21     layer is determined by surface roughness elements, and, if particle monitors are sampling within
22     this layer, there is a high probability of particles being entrained  in turbulent eddies and
23     redepositing on the ground.  In addition to the source sampling problem referred to above, it
24      should be remembered that dust often is raised in remote areas far removed from population
25      centers. Gravitational settling can be an important loss mechanism for particles larger than a few
26      microns in aerodynamic diameter and precipitation or scavenging by cloud droplets also removes
27      smaller particles during transport from the source area.
 28           As rough estimates, uncertainties in emissions estimates could be as low as 10% for the
 29      best characterized source categories, whereas emissions figures  for windblown dust should be
 30      regarded as order-of-magnitude estimates. Given (a) uncertainties in the deposition of SO2 and
 31      its oxidation rate; (b) the variability seen in OC and EC emissions from motor vehicles, along
        March 2001
3-56
DRAFT-DO NOT QUOTE OR CITE

-------
  1      with the findings from past verification studies for NMHC and CO to NOX ratios; (c) ranges of
  2      values found among independent estimates for emissions of individual species (NH3, OC); and
  3      (d) the predominance of fugitive emissions, PM emissions rates should be regarded as
  4      order-of-magnitude estimates.
  5           There have been few field studies designed to test emissions inventories observationally.
  6      The most direct approach would be to use aircraft to obtain cross-sections of pollutants upwind
  7      and downwind of major urban areas. The computed mass flux through a cross section of the
  8      urban plume can then be equated to emissions from the city chosen. This approach has been
  9      attempted on a few occasions. Results have been ambiguous because of contributions from
10      fugitive sources, nonsteady wind flows, and general logistic difficulties.
11
12      :..-..-....•                  .  •   .         •            .          .   ••
13      3.5  SUMMARY AND CONCLUSIONS
14           The recently deployed PM2 5 FRM network has returned data for a large number of sites
15      across the United States. Annual mean PM2 5 concentrations range from about 5 //g/m3 to over
16      20 /^g/m3. In the eastern United States, the 1999 data indicate that highest quarterly mean
17      concentrations and maximum concentrations were reached during the summer. In the western
18      United States, highest quarterly mean values and maximum values occurred during the winter at
19      a number of sites, although there were exceptions to these general patterns.  These findings are
20      generally consistent with those based on longer term data sets such as MAAQS in the eastern
21      United States and the CARB network of dichotomous samplers in California.  The 1999 FRM
22      PM2 5 data indicate that, in general, PM2 5 concentrations are highly correlated among sites in
23      several MSAs (Atlanta, GA; Detroit, MI; Phoenix-Mesa, AZ; and Seattle-Bellevue-Everett,
24      WA), although there are exceptions to this rule. These findings are consistent  with those of
25      earlier studies in Philadelphia, PA, and Los Angeles, CA, examining the  spatial variability of
26      PM2 5 and its components. PM2 5 to PM10 ratios were generally higher in the East than in the
27      West, and values are consistent with those found in numerous earlier studies presented in the
28      1996PMAQCD.
29          Ambient particulate matter contains both primary and secondary components. The results
30      of ambient monitoring studies and receptor modeling studies in the eastern United States indicate
31      that PM2 5 is dominated by secondary components.  Depending on the origin of OC in ambient
       March 2001
3-57
DRAFT-DO NOT QUOTE OR CITE

-------
 1     samples, PM2-S, on average, also may be dominated by secondary components throughout the rest
 2     of the United States.  Primary constituents represent smaller but still important components of
 3     PM2-5, on average. Crustal materials constitute the largest fraction of PM(10_2 5) throughout the
 4     United States. Data collected in several airsheds including the Los Angeles Basin, Bakersfield
 5     and Fresno, CA; and Philadelphia, PA, suggest that secondary PM components are more
 6     uniformly distributed than are primary components. Compositional data obtained at multiple
 7     sites in other urban areas are sparse.
 8           Because of the  complexity of the composition of ambient PM2 5 and PM(IO_25), sources are
 9     best discussed in terms of individual constituents of both primary and secondary PM2 5 and
10     PM(10.2-5). Each of these constituents can have anthropogenic and natural sources, as shown in
11     Table 3-7. The distinction between natural and anthropogenic sources is not always obvious.
12     Although windblown dust might seem to be the result of natural processes, highest emission rates
13     are associated with agricultural activities in areas that are susceptible to periodic drought,
14     Examples include the dust bowl region of the midwestern United States and the Sahel of Africa.
15     Also, most forest fires in the United States could be classified as human in origin, either through
16     prescribed burning; by accident; or through forest management practices that allow the buildup
17     of combustible material, thereby increasing the likelihood of fire from whatever cause.
18           Emissions inventories are generally not the most appropriate way to apportion material in
19     ambient samples. Receptor modeling has proven to be an especially valuable tool in this regard.
20     Receptor modeling can help bound emission inventories and establish uncertainty estimates.
21     Compositional profiles developed for receptor modeling applications are perhaps the most
22     accessible and reliable means of characterizing the composition of emissions.  Techniques are
23     under development to use emission inventories and receptor modeling to reduce the uncertainty
24     in the overall source  apportionment (U. S. Environmental Protection Agency, 2000e).
25           The results of receptor modeling studies throughout the United States indicate that the
26     combustion of fossil and biomass fuels is a major source of PM2 5. Fugitive dust, found mainly in
27     the PM(10_25) range size, represents the largest source of PM10 in many locations in the western
28     United States. Quoted uncertainties in source apportionments of constituents in ambient aerosol
29     samples typically range from 10 to 50%. It is apparent that a relatively small number of source
30     categories, compared to the total number of chemical species that typically are measured in
        March 2001
3-58
DRAFT-DO NOT QUOTE OR CITE

-------
  1      ambient monitoring-squrce receptor modelstudies, are needed to account for most of the
 2      observed mass of PM in these studies.  :     •   ,    •  ,    ••   <. ,   ,.,   v     •             •
 3           Improvements in.the ability of receptor models to allocate sources, of .ambient PM continue
 4      to be made. 'Recently developed techniques ;such as positive matrix factorization allow
 5      quantitative determinations of contributions from different categories of PM sources to be made
 6      on the basis of ambient data alone.  Improvements in the accuracy of PM emissions inventories
 7      also continue to be made.  Recent studies have identified causes of the overrepresentation of
 8      crustal material in emissions inventories. The causes are related to the neglect of near-source PM
 9      deposition in the development of emissions factors.
10           As seen in Table 3-7, emissions of mineral dust, organic debris, and sea spray are
11      concentrated mainly in the coarse fraction of PM10 (>2.5 /urn aerodynamic diameter). A small
12      fraction of this material is in the PM2 5 size range (< 2.5 //m aerodynamic diameter).
13      Nevertheless, concentrations of crustal material can be appreciable, especially during dust events.
14   ,   It also should be remembered that much of the Saharan dust reaching the United States is in the
15      PM25 size range.  Emissions  from combustion sources (mobile and stationary sources and
16      biomass burning) are also predominantly in the PM25 size range.
17           Uncertainties in emissions inventories are difficult to quantify. They may be as low as 10%
18      for well-defined sources (e.g., for SO2) and may range up to a factor of 10 or so for windblown
19      dust. As a rule, total PM emissions rates should be regarded as order-of-magnitude estimates.
20      Because of the large uncertainty associated with emissions of suspended dust, trends of total
21      PM25 PM10 emissions should be viewed with caution, and emissions of specific components are
22      best discussed on an individual basis.  Receptor modeling, especially when coupled to accurate
23      measurements of the composition of emissions, can  be useful in providing bounds for emission
24      inventories.
25           Although most emphasis in this  chapter has been on sources within the United States,
26      it also should be remembered that sources outside the United States contribute to ambient PM
27      levels that can, at times, exceed the ambient NAAQS level for PM. Perry et al. (1997) have
28      found that the highest concentrations of mineral dust in the PM2 5 fraction are found in the eastern
29      United States during the summer and not in arid areas of the western United States. This  dust
30      has been emitted in the Sahara Desert and then transported across the Atlantic Ocean.
31      Large-scale dust storms in the deserts  of central Asia recently have been found to contribute to
        March 2001
3-59
DRAFT-DO NOT QUOTE OR CITE

-------
1
2
3
4
5
PM levels in the Northwest on an episodic basis. Uncontrolled biomass burning in central
America and Mexico may have contributed to elevated PM levels that exceeded the daily
NAAQS level for PM in Texas. Wildfires throughout the United States, Canada, Mexico, and
Central America all contribute to background concentrations of PM in the United States.
      March 2001
                                        3-60
DRAFT-DO NOT QUOTE OR CITE

-------
  1
 REFERENCES
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
45
46
47
48
49
 50
 51
 52
53
 Andreae, M. O. (1991) Biomass burning: its history, use, and distribution and its impact on environmental quality
       and global climate. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric
       implications. Cambridge, MA: MIT Press; pp. 1-21.
 Bahadori, T.; Suh, H.; Koutrakis, P. (2000a) A cross-study analysis of ambient concentrations of fine mass. J. Air
       Waste Manage. Assoc.: in press.
 Bahadori, T.; Suh, H.; Koutrakis, P. (2000b) PM2 5 and PM10 concentrations in four metropolitan cities in eastern
       United States. J. Air Waste Manage. Assoc.: in press.
 Cahill, T. A. (1996) [Data from IMPROVE data base.] Fort Collins, CO: Air Resource Specialists, Inc.
 California Air Resources Board. (1995) [Data base provided by John R. Holmes, Chief, Research Division.]
       Sacramento, CA: California Air Resources Board.
 Calvert, J. G.; Heywood, J. B.; Sawyer, R. F.; Seinfeld, J. H. (1993) Achieving acceptable air quality: some
       reflections on controlling vehicle emissions. Science (Washington, DC) 261: 37-45.
 Chow, J. C.; Watson, J. G.; Frazier, C. A.; Egami, R. T.; Goodrich, A.; Ralph, C. (1988) Spatial and temporal
       source contributions to PMIO and PM2 5 in Reno, NV. In: Mathai, C. V.; Stonefield, D. H., eds. PM.10:
       implementation of standards, an APCA/EPA international specialty conference; February; San Francisco,
       CA. Pittsburgh, PA: Air Pollution Control Association; pp. 438-457. (APCA transactions series TR-13).
 Chow, J. C.; Watson, J. G.; Richards, L. W.; Haase, D. L.; McDade, C.; Dietrich, D. L.; Moon, D.; Sloane, C.
       (1991) The 1989-90 Phoenix PM10 study. Volume  II: source apportionment. Final report. Phoenix, AZ:
       Arizona Department of Environmental Quality; Desert Research Institute document no. 8931.6F1.
 Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Frazier, C. A.; Hinsvark, B. A.; Prichett, L. C.; Neuroth, G. R.
       (1992a) Wintertime PM,0 and PM25 chemical compositions  and source contributions in Tucson, Arizona.
       In: Chow, J. C.; Ono, D. M., eds. PMIO standards and nontraditional particulate source controls, an
       A&WMA/EPA international specialty conference, v. I; January; Scottsdale,  AZ. Pittsburgh, PA: Air &
       Waste Management Association; pp. 231-243. (A&WMA transactions series no. 22).
 Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Solomon, P. A.; Magliano, K. L.; Ziman, S. D.; Richards, L. W.
       (1992b) PMIO source apportionment in California's San Joaquin Valley. Atmos. Environ. Part A
       26: 3335-3354.
 Chow, J. C.; Liu, C. S.; Cassmassi, J.; Watson, J. G.; Lu, Z.; Pritchett, L. C. (1992c) A neighborhood-scale study of
      PM10 source contributions in Rubidoux, California. Atmos. Environ. Part A 26: 693-706.
 Darlington, T. L.; Kahlbaum, D. F.; Heuss, J. M.; Wolff, G. T. (1997) Analysis of PM10 trends in the United States
      from 1988 through 1995. J. Air Waste Manage. Assoc. 47: 1070-1078.
 Dolislager, L. J.; Motallebi, N. (1999) Characterization of particulate matter in California. J. Air Waste Manage
      Assoc. 49: PM-45-56.
 Duce, R. A. (1995) Sources, distributions, and fluxes of mineral aerosols and their relationship to climate.
      In: Charlson, R. J.; Heintzenberg, J., eds. Aerosol forcing of climate: report of the Dahlem workshop on
      aerosol forcing of climate; April 1994; Berlin, Federal Republic of Germany. Chichester, United Kingdom:
      John Wiley & Sons, Ltd.; pp. 43-72.
 Dzubay, T. G.; Stevens, R. K.; Gordon, G. E.; Olmez, I.; Sheffield, A. E.; Courtney, W. J. (1988) A composite
      receptor method applied to Philadelphia aerosol. Environ. Sci. Technol. 22: 46-52.
 Eldred, R. A.; Cahill, T. A. (1994) Trends in elemental concentrations of fine particles at remote sites in the United
      States of America. Atmos. Environ. 28: 1009-1019.
 Fitz-Simons, T. S.; Mathias, M.; Rizzo, M. (2000) Analyses of 1999 PM data for the PM NAAQS review. U.S.
      Environmental Protection Agency, Office of Air Quality Planning and Standards; September 29.
Garratt, J. R. (1994) The atmospheric boundary layer. Cambridge, United Kingdom: Cambridge University Press.
      (Cambridge atmospheric and space science series).
Gatz, D. F.; Prospero, J. M. (1996) A large silicon-aluminum aerosol plume in central Illinois: North African Desert
      dust? Atmos. Environ. 30: 3789-3799.
Gillette, D. A.; Hanson, K. J. (1989) Spatial and temporal variability of dust production caused by wind erosion in
      the United States. J. Geophys. Res. [Atmos.] 94: 2197-2206.
Gillies, J. A.; Gertler, A. W.; Lowenthal, D.; Abu-Allaban, M.; Jennison, B.; Goodrich, A. (2000) Enhanced PM25
      source apportionment using chemical mass balance receptor  modeling and scanning electron microscopy.
      Aerosol Sci. Technol.: submitted.
         March2001
                                                3-61
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Glover, D. M.; Hopke, P. K.; Vermette, S. J.; Landsberger, S.; D'Auben, D. R. (1991) Source apportionment with
 2            site specific source profiles. J. Air Waste Manage. Assoc. 41: 294-305.
 3       Gray, H. A.; Landry, B.; Liu, C. S.; Henry, R. C.; Cooper, J. A.; Sherman, J. R. (1988) Receptor modeling for RM10
 4            source apportionment in the South Coast Air Basin of California. In: Mathai, C. V.; Stonefield, D. H., eds.
 5            PM.,0: implementation of standards, an APCA/EPA international specialty conference; February; San
 6            Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp. 399-418. (APCA transactions series
 7            TR-13).
 8       Hanna, A.; Mathur, R.; Alapaty, K.; Pinto, J. (1999) Modeling the episodic transport of air pollutants from Asia to
 9            North America. http://envpro.ncsc.org/projects/MITP/.
10       Hopke, P. K.; Wlaschin, W.; Landsberger, S.; Sweet, C.; Vermette, S. J. (1988) The source apportionment of PM10
11            in south Chicago. In: Mathai, C. V.; Stonefield, D. H., eds. PM.10: implementation of standards, an
12            APCA/EPA international specialty conference; February; San Francisco, CA. Pittsburgh, PA: Air Pollution
13            Control Association; pp. 484-494. (APCA transactions series TR-13).
14       Houck, J. E.; Rau, J'. A.; Body, S.; Chow, J. C. (1992) Source apportionment: Pocatello, Idaho PM,0 nonattainment
15            area. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional particulate source controls, an
16            A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air &
17            Waste Management Association; pp. 219-230. (A&WMA transactions series no. 22).
18       Huang, X.; Olmez, I.; Aras, N. K.; Gordon, G. E. (1994) Emissions of trace elements from motor vehicles: potential
19            marker elements and source composition profile. Atmos.  Environ. 28: 1385-1391.
20       Husar, R. B.; Schichtel, B. A.; Falke, S. R.; Li, F.; Wilson, W. E.; Pinto, J.; Malm, W. C.; Fox,  D. G.; Feldman,
21            G. C.; McClain, C.;  Kuring, N.; Holben, B. N.; Vermote, E. F.; Herman, J. R.; Elvidge, C. D. (2000) The
22            impact of the 1998 Central American smoke on the atmospheric environment of eastern North America.
23            J. Geophys. Res.: submitted.
24       IMPROVE: interagency monitoring of protected visual environments [database]. (1998) [Data on gross chemical
25            composition of PM2J particles in rural areas of the eastern and western United States as of 1998]. Fort
26            Collins, CO: National Park Service (NFS); Cooperative Institute for Research in the Atmosphere (CIRA).
27            Available at: http://vista.cira.colostate.edu/improve/ [2001, January 26].
28       Inhalable Particulate Network. (1985) [Inhalable Particulate Network Database.] Research Triangle Park, NC:
29            U.S. Environmental Protection Agency, National Exposure Research Laboratory; Jose M. Sune, contact.
30       Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
31             segregated mass concentration data and ultrafme particle number concentrations in urban areas. Atmos.
32             Environ. 33:2907-2913.
33       Kim, B. M.; Zeldin, M. D.; Liu, C. S. (1992) Source apportionment study for state implementation plan
34             development in the Coachella Valley. In: Chow, J. C.; Ono, D. M., eds. PM,0 standards and nontraditional
35             particulate source controls: an A&WMA/EPA international specialty conference, v. II; January; Scottsdale,
36             AZ. Pittsburgh, PA: Air & Waste Management Association; pp. 979-991. (A&WMA transactions series
37             no. 22).
38       Koutrakis, P. (1995) [PM2S and PMIO data for Philadelphia, 1992-1995]. Boston, MA: Harvard School of Public
39             Health.
40       Lawson, D. R.; Smith, R. E. (1998) The northern front range air quality study: a report to the Governor and General
41             Assembly. Fort Collins, CO: Colorado State University; December.
42       Magliano, K. L. (1988) Level  1 PM10 assessment in a California air basin. In: Mathai, C. V.; Stonefield, D. H., eds.
43             PM.,0: implementation of standards, an APCA/EPA international specialty conference; February;
44             San Francisco, CA.  Pittsburgh, PA: Air Pollution Control Association; pp. 508-517. (APCA transactions
45             series TR-13).
46       Magliano, K. L. (1998) Chemical mass balance modeling of data from the 1995 integrated monitoring study.
47             Sacramento, CA: California Air Resources Board.
48       McMurry, P. H.; Woo, K. S.; Weber, R.; Chen, D.-R.; Pui, D. Y. H. (2000) Size distributions of 3  to 10 nm
49             atmospheric particles: implications for nucleation mechanisms. Philos. Trans. R. Soc. London Ser. A 358:
50             2625-2642.
51       Motallebi, N. (1999) Wintertime PM25 and PM10 source apportionment at Sacramento, California.  J. Air Waste
52             Manage. Assoc. 49: PM-25-34.
53       Neas, L. M. (1996) [Data provided for 1986 from Harvard six cities data base.] Boston,  MA: Harvard School of
54             Public Health.
         March 2001
3-62
DRAFT-DO NOT QUOTE OR CITE

-------
   1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
47
48
49
 50
51
52
53
54
 Ott, S.-T.; Ott, A.; Martin, D. W.; Young, J. A. (1991) Analysis of a trans-Atlantic Saharan dust outbreak based on
       satellite and GATE data. Mon. Weather Rev. 119: 1832-1850.
 Paatero, P.; Tapper, U. (1993) Analysis of different modes of factor analysis as least squares fit problems. Chemom
       Intelligent Lab. Syst. 18: 183-194.,
 Paatero, P.; Tapper, U. (1994) Positive matrix factorization: a non-negative factor model with optimal utilization of
       error estimates of data values. Environmetrics 5: 111-126.
 Perry, K. D.; Cahill, T. A.; Eldred, R. A.; Dutcher, D. D.; Gill, T. E. (1997) Long-range transport of North African
       dust to the eastern United States. J. Geophys. Res. [Atmos] 102: 11,225-11,238.
 Pinto, J. P.; Pleasant, M.; Kellogg, R.; Torok, Sz.; Stiles, D.; Willis, R. D.; Koutrakis, P.- Allen, G. A. (1995)
       Chemical characterization of the ambient aerosol in Philadelphia. Presented at: Particulate matter: health and
       regulatory issues: an A&WMA international specialty conference; April; Pittsburgh, PA. Pittsburgh, PA:
       Air & Waste Management Association; submitted abstract.
 Placet, M.; Battye, R. E.; Fehsenfeld, F. C.; Bassett, G. W. (1991) Emissions involved in acidic deposition
       processes. In: Irving, P. M., ed. Acidic deposition: state of science and technology, volume I: emissions,
       atmospheric processes, and deposition. Washington, DC: The U.S. National Acid Precipitation Assessment
       Program. (State of science and technology report no. 1).
 Prospero, J. M. (1996) The atmospheric transport of particles to the ocean. In: Ittekkot, V.; Honjo, S.; Depetris, P.
       J., eds. Particle flux in the ocean. New York, NY: John Wiley & Sons, Ltd.; pp. 19-52. (SCOPE report no.

 Prospero, J. M.,(1999) Long-range transport of mineral dust in the global atmosphere: impact of African dust on the
       environment of the southeastern United States. Proc. Natl. Acad. Sci. U.S.A. 96: 3396-3403.
 Prospero, J. M.; Nees, R. T.; Uematsu, M. (1987) Deposition rate of paniculate and dissolved aluminum derived
       from Saharan dust in precipitation at Miami, Florida. J. Geophys. Res. [Atmos.] 92: 14,723-14,731.
 Ramadan, Z.; Song, X.-H.; Hopke, P. K. (2000) Identification of sources of Phoenix aerosol by positive matrix
       factorization. J. Air Waste Manage. Assoc. 50:  1308-1320.
 Rodes, C. E.; Evans, E. G. (1985) Preliminary assessment of 10 urn particulate sampling at eight locations in the
       United States. Atmos. Environ. 19: 293-303.
 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993) Sources of fine organic
       aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ Sci
       Technol. 27:636-651.
 Ryan, W. M.; Badgett-West, C. R.; Holtz, D. R.; Peters, T. A.; Cooper, J. A.; Ono, D. (1988) Reconciliation of
       receptor and dispersion modeling impacts of PM10 in Hayden, Arizona. In:  Mathai, C. V.; Stonefield, D. H.,
       eds. PM.10: implementation of standards, an APCA/EPA international specialty conference; February;
       San Francisco, CA, Pittsburgh, PA: Air Pollution Control Association; pp.  419-429. (APCA transactions
       series TR-13).
 Schauer, J. J.; Rogge, W. F.; Hildemann, L. M.; Mazurik, M. A.; Cass, G. R. (1996) Source apportionment of
       airborne particulate matter using organic compounds as tracers. Atmos. Environ. 30: 3837-3855.
 Skidmore, L. W.; Chow, J. C.; Tucker, T. T. (1992) PMIO air quality assessment for the Jefferson County, Ohio air
       quality control region. In: Chow, J.  C.; Ono, D. M., eds. PM,0 standards and nontraditional particulate source
       controls, an A&WMA/EPA international specialty conference, v. II; January; Scottsdale, AZ. Pittsburgh, PA:
      Air and Waste Management Association; pp.  1016-1031. (A&WMA transactions series no. 22).
 Spengler, J. D.; Briggs, S. L. K.; Ozkaynak, H. (1986) Relationship between TSP measurements and
      size-fractionated particle mass  measurements  in six cities participating in the Harvard Air Pollution Health
       Study. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning
      and Standards; EPA cooperative agreement no. CR-812667-01-0; December 5.
 Suh, H. H.; Allen, G. A.; Koutrakis, P.; Burton, R. M. (1995) Spatial variation in acidic sulfate and ammonia
      concentrations within metropolitan Philadelphia. J. Air Waste Manage. Assoc. 45: 442-452.
Tanner, R. L.; Parkhurst, W. J. (2000) Chemical composition of fine particles in the Tennessee Valley region. J. Air
      Waste Manage. Assoc. 50: 1299-1307.
Taylor, C. A., Jr.; Stover, C. A.; Westerdahl, F. D. (1998) Speciated fine particle (<2.5 urn aerodynamic diameter)
      and vapor-phase acid concentrations in southern California. Presented at: 91st annual meeting & exhibition
      of the Air & Waste Management Association; June; San Diego, CA. Pittsburgh, PA:  Air & Waste
      Management Association; paper no. WA7401.
         March 2001
                                                3-63
                                                                       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Thanukos, L. C.; Miller, T.; Mathai, C. V.; Reinholt, D.; Bennett, J. (1992) Intercomparison of PM10 samplers and
 2            source apportionment of ambient PMIO concentrations at Rillito, Arizona. In: Chow, J. C.; Ono, D. M., eds.
 3            PM10 standards and nontraditional particulate source controls, an A&WMA/EPA international specialty
 4            conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management Association;
 5            pp. 244-261. (A&WMA transactions series no. 22).
 6      U.S. Environmental Protection Agency. (1995) Compilation of air pollutant emission factors. Volume I: stationary
 7            point and area sources. 5th ed. Research Triangle Park, NC: Office of Air Quality Planning and Standards;
 8            report no. AP-42-ED-5. Available from: NTIS, Springfield, VA; PB95-196028.
 9      U.S. Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park,
10            NC: National Center for Environmental Assessment-RTP Office; report nos. EPA/600/P-95/001 aF-cF. 3v.
11      U.S. Environmental Protection Agency. (2000a) Aerometric Information Retrieval System (AIRS). Research
12            Triangle Park, NC: Office of Air Quality Planning and Standards. Available: www.epa.gov/ttn/airs/ [2000,
13            October 6].
14      U.S. Environmental Protection Agency. (2000b) National air pollutant emission trends, 1900-1998.  Research
15            Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/R-00-002. Available:
16            http://www.epa.gov/ttn/chief/trends/trends98/[2001, January 30].
17      U.S. Environmental Protection Agency. (2000c) Air quality criteria for carbon monoxide. Research Triangle Park,
18            NC: National Center for Environmental Assessment; report no. EPA/600/P-99/001F. Available:
19            www.epa.gov/ncea/co/  [2000, October 6].
20      U.S. Environmental Protection Agency. (2000d) National air quality and emissions trends report, 1998. Research
21            Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA/454/R-00-003. Available
22            from: NTIS, Springfield, VA; PB2000-104960.
23      U.S. Environmental Protection Agency. (2000e) Guidance for demonstrating attainment of air quality goals for
24            PM2J and regional haze. March.
25      Vermette, S. J.; Williams, A. L.; Landsberger, S. (1992) PM,0 source apportionment using local surface dust
26            profiles: examples from Chicago. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional
27            particulate source controls, an A&WMA/EPA international specialty conference, v. I; January; Scottsdale,
28            AZ. Pittsburgh, PA: Air & Waste Management Association; pp. 262-278. (A&WMA transactions series
29            no. 22).
30       Watson, J. G.; Chow, J. C. (1999) Reconciling urban fugitive dust emissions inventory and ambient source
31             contribution estimates:  summary of current knowledge and needed research [draft]. Reno, NV: Desert
32             Research Institute; DRI document no. 6110.4D2.
33       Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
34             (1990a) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
35       Watson, J. G.; Chow, J. C.; Pritchett, L. C.; Houck, J. A.; Ragazzi, R. A.; Burns, S. (1990b) Chemical source
36             profiles for particulate motor vehicle exhaust under cold and high altitude operating conditions. Sci. Total
37             Environ. 93:183-190.
38       Watson, J. G.; Chow, J. C.; Lu, Z.; Fujita, E. M.; Lowenthal, D. H.; Lawson,  D. R.; Ashbaugh, L. L. (1994)
39             Chemical mass balance source apportionment of PMIO during the Southern California Air Quality Study.
40             Aerosol Sci. Technol. 21:1-36.
41       Wilson, W. E.; Stockburger, L. (1990) Diurnal variations in aerosol composition and concentration. In: Masuda, S.;
42            Takahashi, K., eds. Aerosols: science, industry, health and environment:  proceedings of the third
43             international aerosol conference, v. 2; September; Kyoto, Japan. Oxford, United Kingdom: Pergamon Press
44            pic; pp. 962-965.
45      Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
46            epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
47      Wongphatarakul, V.; Friedlander, S. K.; Pinto, J. P. (1998) A comparative study of PM25 ambient aerosol chemical
48            databases. Environ. Sci. Technol. 32: 3926-3934.
49      Woo, K. S.; Chen, D. R.; Pui, D. Y. H.; McMurry, P. H. (2000) Measurement of Atlanta aerosol size distributions:
 50            observations of ultrafine particle events. Aerosol Sci. Technol.: accepted.
         March 2001
3-64
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
                               APPENDIX 3A

                   Organic Composition of Particulate Matter

     Although organic compounds typically constitute approximately 10 to 70% of the total dry
fine particle mass in the atmosphere, organic PM concentrations, composition, and formation
mechanisms are poorly understood. This is because particulate organic matter is an aggregate of
hundreds of individual compounds spanning a wide range of chemical and thermodynamic
properties (Saxena and Hildemann, 1996). The presence of multiphase or "semi-volatile"
compounds complicates collection of organic particulate matter.  Furthermore, no single
analytical technique currently is capable of analyzing the entire range of compounds present.
Rigorous analytical methods frequently identify only 10 to 20% of the organic mass on the
molecular level (Rogge et al., 1993). The data shown in Appendix 3A are meant to complement
the data given for the inorganic components of particles in Appendix 6 A of the 1996 PM AQCD
(U. S. Environmental Protection Agency, 1996). Table 3A-1 lists a number of recent urban and
some rural measurements of particulate organic and elemental carbon in ,ug of carbon/m3.
Emphasis was placed on measurements published after 1995.  The analysis method and artifact
correction procedure, if any, are indicated. Table 3A-2 presents information on recent (post
1990) studies concerning concentrations of particulate organic compounds found at selected U.S.
sites.
       March 2001
                                        3A-1
DRAFT-DO NOT QUOTE OR CITE

-------
 S
 u
i
H

1
Z in
OS
sa

   co
 p
 I
 p
 H
   pa
1
      £8
,0 g



H ^
      U
      PJ
      O
      O
        Q
ation
Refe
AN PM2J
                    oi
              H  H  O

              g  H  t
              ..  •-  CX
              &  &  GX
              6  £  Q
              s   s
              CS
              U
              1
              CQ
Offenberg
(2000)
               oo
               o
               o;
               ON


               <
               ^
                    i
                    D


                              oi
                              e
                              CX
                              CX
                                    s
                                    CU
                                    J=
                                    ^
                           •
                         Q
                          u
                                         0
Q  T.  CX


S  S  S"
Bu  CL,  CL.
                                   "**  JS  ""
                                   tS  VO  CN
                                    — o  —
                          OOOOOc'jr-;  ^°°  O_
                          *O Tj" C*) CO O^  CO '"^  OO
                   s. ds B" S 1  t^  ^  1
                   Jlpt  ii  I  1
                   ssioSjgS'S-s
                   CO BJ CXo! OQ  ^c«  oo  oo
                  .a

                
-------


">—
s
u
^
.s,

yy
JjjT
O
H
1
fc
o

ix\
^r JJ^
OS
2 P^
Up^
fe
Q
S !»
^
gr_T
" HH
Is
^^t ^^
U p
5 So
^ *z.
S Q
t'd). PARTICULATE Ol
BASED
c
o
1— 1
W
J
5
H




W
O








s
S
CU


we
VO






j^
06
*^s
en
en
^
en
^,
vq
10
c
S
3
ON
o
/%
"

Cu


•*







CN CN u-> CN t~- •*
O O O O O O


ON O O tN OO Tf
O — CN — • en tN



00
— .
o\
ON
Three Sisters Wilderness, OR
Rocky Mountains, CO
Brigantine, NJ
Acadia, MA
Jefferson; James River Face
Wilderness, VA
Glacier, MT


o
o .
o
fN^
1
&
HH


U
S
g
'o

2
Cu



£
^— ^
ON
CN

' :










ON
ON
>%
3
o
'c
a
I
's
o
§J


^
1
•3
ts
oo
00

en










CN
ON

-------
D
SE
c

00
55
0
63
H
W
U
*£'
O
u ^
oS

<3 S
^§
H °
y^ F™3
r*i CM
H 00
al
Mg
§ |5
u 5
H"< 2"v
5 oo
81
Op
w w
M
5
C_i
I
PH
•
^
5-
a
o
3

v*^
3
pq
H

|





|1

§
3?
^
t^, ^

g
S^
CJ §•
u

ra x— \
^» ^
*2 >2
O S
O

S3
"cd
Q


B
to
'a
u
3






u
i
I
C?
£
'3

o-


£
•*






o S 2 ol M S^5 5 S ^
TfVOOOOTfOvO\C OO ^-" Sw>/ ^-" t*^ ^"^
2 3 S 5 ^ 2 2 ? ^ 2

CTl
ON

CO
3
^.
"3
•-^

< ^ <

?5 H U eo >
«Usi§§lli
III Illllll



G^
o\
rl ^^
s ^-^
0- -3
^ s
1 •§
(Jj CJ
o
o
"*
S
cu


j=
CN





~
O,
o
o

x— N
VO
I—*
— ;

oo
^—
60
i
3
>— »

SJ
51
Grand Canyon, /


1

>^
Q
•a
S g
1 E
£>••
CX
*0
S"
04 .

JJ
^
cs







(S
0




vq

ON
O\
Jan-Dec 1



Bondville, IL



i1
ON
^
*c3
1
o .
erf
e
a
*o
S
eu


A
cs






S"
oT
*^

o"
vd

o
oo
ts
1
oo
2


N
Bullhead City, A


jv
ON
ON
^^
'"3
*j
^
O
oi ''
o . •
a
*0
S
o. ...


*
04

oo vo ir> rf — —
oo Tf ff) ro Ol ro





-






oo
ON
u
9
^->
*O S ^ ^"v (W
^ CQ .
^^ S -&^ § 
-------
S





o • k< S3 k. <3 i-
*~^ ^0 I^ ^O 1^ ^O !^

H^
S - !-•
"^ ^ -I
| .N |
'z, S u
S.' •: . ^
°3
Q- ' ;


' «J {C1
'•'.>» 00, .'
                       N __, "O
                       t: ^ £
                     .a - S
                      2 fa"
                              .«E   fi I S
                              111 .l|r
                              Hi -Mill
                              ilJVlil
                              O *j> »rn *"* S* ro **
                                     '
                         5 o
                                   8-1
March 2001
                            3A-5
                                    DRAFT-DO NOT QUOTE OR CITE

-------


WD
td
^^
W5
S
o
Fjrj
WO
^
"a
u
a1
JS
*&

*7

Q
R
t_,
z
rvi
8
|
1
O
S
o
u
n
Q
O
w


^,ON
— ""I***"
??t
fc

ll
•^"5 Is
Jl"


till

Ajs-i s.
68 o S o


jjjj O
S-ON
S ON
3 »1
"** •? °-
llc
g.
E-^?
"f^.a
U cs" 1"^?
es cs '5
111
1 a
CO

M^
t*> 'Ss
2 | |
t3 O "§ cu
u c c
Sn H? ^
O *«— '



-S*
"5
cs
CO
a<
^jO
3

|<
1 §
ii
g n


^
" ^»
Sz
 r-  o ™ o
?• •? a •? '? ••§ ?
c c £> e i v2. c

























TT % - P o
O vo co ON OO
r— * ^* f-*


CN OO ^«
« o o CN m
^ CN VO CO O
— CN

— co ON in in
VO CS ON CS vo



•— — r- o oo
m CN oo CN —

•eicosanoic acid
•heneicosanoic acid
•docosanoic acid
•tricosanoic acid
•tetracosanoic acid
c c c c c
March2001
3A-6
DRAFT-DO NOT QUOTE OR CITE

-------
 cc

 I

 H


 i
 Q
 VI
  D
  M
  a
  a
U

I
 §
S
         S3 °*
         w oC
         •Soif
         « co
         fa
       *
          •
        S 3
          '
       ii
        22

      as 1
               I


               1 §
              •ff
              2o
                                                           c> — ,o\ oo
                                                           " — 1-00
                     ^- «o o* (N o ^r<
                                       00    00
3 ^
I
u
•5
                     '§ 1

                     ll
                        '§ 1

                          l
S-o 'I
^'i_S
                         iS
                         ll
  ll
   I
                                     i '3

                                    i
March 2001
                                   3 A-7     DRAFT-DO NOT QUOTE OR CITE

-------
                         ON O tS
                         
                     oo —I ro rt1 t~ Wl
                         —   : cs r-
                                        o r-
                                        vo o
                                        en ~-
                                                                                       cs  e
                                                                                       co  in
                                                                                       2   fi
                                                       cs   o   rn   vq   o   —;
                                                       ^J  • vi   CO   vo   VO   ^
                                                       if)   ~   CS   ^-   •—   ^
                                                          o
                                                       VO   CO
                                                       vo   —•
                                                               en   *-*
March 2001
3A-8
DRAFT-DO NOT QUOTE OR CITE

-------
      •=" § J?
     Sil«
       Ss
ka
(2
95
de
     1SJ|Z
     h U ~
            So
                            •*
                            S?

                          000
                                   o
                                   o
                                   oo
                                   T}-
                                   0
        t-
        o'
f~ t— 00
ON 
-------
CO
H
p*j
H
Z
0
Q
W
CO
M
f*M

is
M
O
B
rRATIONS (in
ED SITES
g H
w H
rj W
il
i-
P
8|
y
S ^
o S
°i

g
u
p
1
PM
^
r*
a
o
•Q
^J
«
S
g
So1
OS m
•J ""1^
•sSf
IF
£

^
N Os
^•2 -? <*
3 o _o *^I

a g >2'0-1
I5

Khwaja (1995;
i96) October 1991
89 (semiurban)
no precut
2 2
Veltkamp et al. 1
July24-Aug4,
no precut
= '3'
f\ ""-'§• •»
_> 00 O r!
* 7J 5

U


5? ,-x
•t 8 ^
15 Q "3 a.
>3
r1






j?
ra
CO
s^.

0


cT
§ <
S-S
CO ~
2C
0
I e
r

Schenectady,
NY
oT
NiwotRidg
CO
1^
a
M

o1
w C_i
b^ -

a
O ^t
ld

s"
Jd
J










S 0 M (
o o o t
,

o1 r? os CM P- t- CM
f) ro ^? ^3 Os os so
3 O — —• O O —
OOOOO^-^^Ol^^fO
o o.o o o o o o o o













oo r- _
•* r~ o
Tf OO Jj


SO , —
•n •*' £J
-^ . f)

Tf O
in oo oo CM oo I*;
— oo o t~ o g
' . 1-H


cs ^^
oo oo in so r1— 9\
CM t^ ^^ ^^ ' ^^ "^
»^


o
"^ "P. 'a
S 'g .2 S •§
?i 1 • 1 2-§ ifl illlji ll













so in o so
in in m in
O O p O


 S« <
S* c
•a g a- a- S g. § tc •?.
1 „ 1 5^«s:l^g
|| B SS g'g'.gSJ
ll ll lilfil
•58 -Sg 22^g -§^2 g
6-E g gJS gSS^g-fS 1
^"5. So&gcilri1e-§c^&1
£^%S § lo^l§>lu-g
March 2001
3A-10
DRAFT-DO NOT QUOTE OR CITE

-------
t/3
S
g

H
.
1 r~t
O
'Q
§5
' *^N
gfl
^ s
"a
U
W>
S3
8
£.^1
wa
^s r-*-i
^j (*i

§ S
S H
Pyj^ ^3
ia
U ^
Q^
P o\
Q O\
£LH ^j

»*5 rvi
O 5
U b
U ^
£r Q
^ ly
O '@

o *-'
H S
i
!
?
o
s
F*H
^

Is
j'" '5'
2 S^-o
W 00 ^
S U
n 3 ;,
V g
1 1
r


^
(2
i>
s
J=
U
co
aT
M
lo
1
z

2
H

*H r^
-S
«
00


o"
1 u
bb

I
15
A
ft,
8"
"3
,3





•
1 — ' "^ O O ^-»C>OOiO OO ' fO ^J* C1^ ' 'O *O  ' O* O* O* O O



•\o r-- t — fi r'l
. O'i 1 — ^~ ^^ ^ " ' t
«o cs* CN o -— i— o r4'o cn'o V • ; •
ff) "'. ,L,


( •
•, o '
os CN o co o ^t* *o v^ co r^ f . • "
,VO ^Ol — '•OfNCNlO OO f> r— i-^ VI
r^ odorot^cd-^ \d^osoo\ .' :- _
«H
,. ; ......
0*0 ro ^r ON) Os m o L
(N..CO O> ^ ^- O Tf »-<
•-- o" o o o ~ ^ TH ."'"''
^~* C^ Os ^t ro ^3 ^f \O
-V1-; o o. o.'^ ^ ^ ,

(J
^s 0 ' "C «> • J,
T3 ' ro C O1 "O i
0,, HI i !.'• '• 1 .!*•!. *I 1? 1
E s--'i5 5 > - ^ "S *S-' -K idnc!^ u ^ 13'Sc2c c.S i 2 ' S
3 M ^C ^C CJ S S C fli W 3 ^1 xi "^L .^" ^J 'So So *^ S m ^3 "^ .7j
>-coHOH«g2gcc .^ yy. £> ,.•§'« A,£.eg.c o tf g — g . S
. ! 3 33:j |1 41 t
^> O OO OO r^P ""C? ^— 1 1—^ i^Tl • • • fr C MM: jr* Q LQ| g '. ^i +J S • |S ^, - S ,^L' ^ CU ^ H O« "c .Si O .52 1)
S.-G S ol Wj' | | S | | '^•s'fe §;.§ | 8 '.g § a J ' •§!.§.§• ? 8 § 9 §•"§ J ^ S 3
<£ £cJu ^ 1'^ ^ ^ s £ • S. s.2 J':l tg'J ;| . .§ s ^ o •!.£ 2- ^ S S 1.1. i 1 1 = 1
March 2001
3A-11.    DRAFT-DO NOT QUOTE OR CITE

-------
co
s


r.
*\
^
O
0

02
/-«.
n-*
g

D
on
B
,0
CO
Z CO
2g
555
c °
£N
» CONCEl
TSELEC
n
S ^
11
w C_i
u £
u ^

i§
JK CO
S3
&3 p
<

H
I
PM
?
TABLE 3A-2(cont
s-
ON p^
• ^" 'S*
73 oC w
** QQ *C
*j sEx
S &
I"

||
J~~" "~* •>,
^ s"
QJ S 3 OH
C § ^^ '

|co

,^-s

I?! 1
|jig
(-.
al
tkamp et al. (
Iy24-Aug4, 1
no precut
^^
§
Isf
o«-'& 3
13.11
" 8-
|«S
u


S«"M
2 o° ta
O<»> S
t5 Q 1 a.


,:•
j>
w
i! •<
D£ ^
B
•<
J

«<
CO .
||

S e
B BQ



•5"
iz
i

-r
Niwot Ridge
CO

s"
73
*&<
o (J
CO

f
|d
b

a
IN
«
eu
3
Id
3












fl OO »••* CO
O O i— " O













OO
• 00 00
VI ^"
0 ~



SO O
o^ oo



s
o


OO
o

M ii
illLtf ^iti

s- ~ ^ P
O f) i— • ^*
^ , o o cs
s^- >^/ ^_x ^^
c? o^ v* ^^ ^®
fsj ^5 ^5 v> • VJ
0' 0 0 0 • M



vo
3) 0 P
*— • ^5 O\













^



JS
PN


M 
-------
GO
H
s
H
fc
0
Q

52
2
PQ
1T
u
IDD
a -
JS
CO

o w
H ^
= 5
Z H
H W
y W
o w
U H
Q 55


^s PM
ff* rvi
O H
U fa
(j <
g Q
si
° PQ
H OH
J
p
^
5!-
OH
*
"3
o
jU^
r<
^
w
S
3.
ss-
li
"3 oC J
|~e
SM
£

r~
2 §^
*~? ~ 'c »
"" 8 -8 S"
w E 3,6^
e §
s "»


i^-
?^ ^ (3 -4-»
2 2 * a
tS U 3 §

* € 1 §
go-
Is
2 2
— Tf1 5
•2 s1 «
III
| V,
s-
o
o <^
CJ. » •§ .
**••' Os J=
i " -5,
O oo" u 3
"c 2 .2 a!
" «€
U
cn
x
^ CN "§
C-S 2
"«3.-8 «s"
•g Q "« o-
as 1
(S ""




e1
1
ea
J •<
|U
V)
J

a"
S 2
!»*
sf
is
G BQ
W

^
"1
H ^
c Z
£
u
en

of
TJ
So
lu
2
•rf
"a!

J -^J
(j

ea

•f<
£
1
^<
iu
n
rf
1)
< u
J































CJ l>- IO iO


o o *— < vo o •*!• r^j cJr-Hcs «~«'*j:i/So6|*ot«o r*
TH

OO'O^fOOC^OOOQQ *—< t** CS OO f^l fri {-». o ^^ f*^ O\ f«»
^3 c? I™* vo f^ cs t*~ r** ^— « ^^ c^ o ""^j* *™* ^" <|;^" cs "/^ v^ i*?
M -* g ^--s







1 -8 '
3 3 " '
1 ii iilji 1
li - !>.f i^2"ll --la If -lf| g'ltfJl's
11 8^:1. s i^-g-fil- aai^|.-l:llii.||^&8
Hf lillflll it|itll||ltl|il


















3










O — • 1-1
o t— t-
—<—•«>«


000
•/"> O NO









lit!
March 2001
3A-13
DRAFT-DO NOT QUOTE OR GITE

-------
CO
CsQ
H


H
CO
a
0
Q
w
S
n

^
u
taxi
a
o

>»^
Z CO
O g
^H P^
E^ ES
S Q
Fj M
) CONCEN'
tf SELECT
5 ^
K °
p ON
§? ff!
is
u E
[?j Cl
IB
55 co
§ tJ
w §
5 *
J
J3
U
P
3'
P-<
CO
S
3
So
ON '^
** ON
>^'^* "e1
"* cf M
t» oo 5
si o.'~/
c3 c/5
£

p1
ON ON
"3 "b Jl ^

e S ^^
£j oo

^^
§» °^ ^ **

»||^ i— « rN o
•I--M ^<
| 0 g 0
»5 ^^ ^"^
X

w£ ON
Veltkamp et a
(1996)
July 24-Aug 4,1
no preeut

S
CJ-ON-g
UoV!- 2
•0 r? g S
t« o «
Qj Q •— *
i° Jr
u
C/3


S ^
2 § 2?
^ •— i O _
* o ^ US
•s 91 a.
« ^

JS
"K
tt
en
£ •<
SjO
^
0
•J

f

»3.
4> C
£ O
o *••
i °
J


X

^
1 Z
^
u
I/I



Niwot Ridge.
CO

i
i d
ca
«

o-
1^


C8*
1^
£


J
Id
J


t^
«n
0
O
NO NO























oo
OO NO V)
fo c« o f) ;


ON M-
-^- cn C4 C5
ON (S M Wl , '



— ' « O c^ NO




NO t~* fl
OO ^ CN) OO C3
O *- O •-• T#
i
g 1 :| I
! -Jl IJ 1:1 If
f 'H *' 11. §11 fit
!II*S' lilllilil
Hill ^JlHlflt






oo





o







^p ;
* 7





§ • ; ;.




3 ?
S £ 'Is
t~ I/I NO
S $


» 3 •••
•-/ -^
ON ^e 0s
OO 5~ O
t~ in NO
*T iA
00 Tf
.a "g «> « j«
e E CA «5 TS
1 1 i 11
0 1 -1 -MS
I i 1 1!
ill III!
•M M* MM «, Oi fi • tU R
II 11 IS 11



































J*

I
ra
O<
.S
c
.2
@
8
naximum con
Mean values are provided with n



§
&
•o
g
•§
1

§
3
|

1
"^
a

*§*
o
rH
1

rt
«
1
J?
\, Downtown LA, I
j
S
I:-
i
CO
.S-.
"H.
.S
f;
DO
3,
3
i
*c3
,§
^
s
&
s
1

-------
 REFERENCES
 Allen, J. O.; Dookeran, N. M; Taghizadeh, K.; Lafleur, A. L.; Smith, K. A.; Sarofim, A. F. (1997) Measurement of
       oxygenated polycyclic aromatic hydrocarbons associated with a size-segregated urban aerosol. Environ  Sci
      . Technol. 31:2064-2070.                                              '         . ,
 Allen, G. A.; Lawrence, J.; Koutrakis, P. (1999) Field validation of a semi-continuous method for aerosol black
       carbon (aethalometer) and temporal patterns of summertime hourly black carbon measurements in
       southwestern PA. Atmos. Environ. 33: 817-823.
 Andrews, E.; Saxena, P.; Musarra, S.; Hildemann, L .M.; Koutrakis, P.;  McMurry, P. H.; Olmez, I.; White, W. H.
       (2000) Concentration and composition of atmospheric aerosols from the 1995 SEAVS  experiment and a
       review of the closure between chemical and gravimetric measurements. J. Air Waste Manage  Assoc 50-
       648-664.
 Chow, J. C.; Watson, J. G.; Lu, Z.; Lowenthal, D. H.; Frazier, C. A.; Solomon, P. A.; Thuillier, R. H.; Magliano, K.
       (1996) Descriptive analysis of PM25 and PM10 at regionally representative locations during
       SJVAQS/AUSPEX. In:  Parrish, D.; Trainer, M.; Rao, S. T.; Solomon, P. A., eds. A&WMA international
       specialty conference on regional photochemical measurements and modeling, part 2; November 1993; San
       Diego, CA. Atmos. Environ. 30: 2079-2112.
 Christoforou, C. S.; Salmon, L. G.; Hannigan, M. P.; Solomon, P. A.; Cass, G. R. (2000) Trends in fine particle
       concentration and chemical composition in southern California. J. Air Waste Manage. Assoc. 50: 43-53.
 Cui, W.; Machir, J.; Lewis, L.; Eatough, D. J.; Eatough, N. L. (1997) Fine particulate organic material at Meadview
       during the project MOHAVE summer intensive study. J. Air Waste Manage. Assoc. 47: 357-369.
 Fraser, M. P.; Cass, G. R.;  Simoneit,  B. R. T.; Rasmussen, R. A. (1998) Air quality model evaluation data for
       organics. 5. C6-C22 nonpolar and semipolar aromatic compounds. Environ. Sci. Technol. 32: 1760-1770.
 Gertler, A. W.; Lowenthal, D. A.; Coulombe, W. G. (1995) PM10 source apportionment study in Bullhead City,
       Arizona. J. Air Waste Manage. Assoc. 45: 75-82.
 Hegg, D. A.; Livingston, J.; Hobbs, P. V.; Novakov, T.; Russell, P. (1997) Chemical apportionment of aerosol
       column optical depth off the mid-Atlantic coast of the United States. J. Geophys. Res. 102: 25,293-25,303.
 IMPROVE: interagency monitoring of protected visual environments [database]. (2000) [Data on particulate'organic
       and elemental carbon concentrations after 1995]. Fort Collins, CO: National Park Service (NFS); Cooperative
       Institute for Research in the Atmosphere (CIRA). Available at: http://vista.cira.colostate.edu/improve/ [2001,
       January 26].
 Khwaja, H. (1995) Atmospheric concentrations of carboxylic acids and related compounds at a semiurban site.
       Atmos. Environ. 29: 127-139.
 Klinedinst, D. B.; Currie, L. A.  (1999) Direct quantification of PM25 fossil and biomass carbon within the northern
       front range air quality study's domain.  Environ. Sci. Technol. 33: 4146-4154.
 Lewtas, J.; Pang, Y.; Booth, D.; Reimer,  S.; Eatough, D. J.; Gundel, L. A. (2001) Comparison of sampling methods
       for semi-volatile organic carbon associated with PM2 5. Aerosol. Sci. Technol. 34: 9-22.
 Lioy, P. J.; Daisey, J. M. (1987) Toxic air pollution: a comprehensive study of non-criteria air pollutants. Chelsea,
       MI: Lewis Publishers.
 Malm, W. C.; Day, D. E. (2000) Optical properties of aerosols at Grand Canyon National Park. Atmos Environ 34-
       3373-3391.                                                                                   '  '
 Malm, W. C.; Gebhart, K. A. (1996) Source apportionment of organic and light-absorbing carbon using receptor
       modeling techniques. Atmos. Environ. 30: 843-855.
 Offenberg, J. H.; Baker, J. E. (2000) Aerosol  size distributions of elemental and organic carbon in urban and
       over-water atmospheres. Atmos. Environ. 34: 1509-1517.
 Omar, A. H.; Biegalski, S.;  Larson, S. M.; Landsberger, S. (1999) Particulate contributions to light extinction and
       local forcing at a rural Illinois site. Atmos. Environ. 33: 2637-2646.
 Pedersen, D. U.; Durant, J. L.; Penman, B. W.; Crespi, C. L.; Hemond, H. F.; Lafleur, A. L.; Cass, G. R. (1999)
      Seasonal and spatial variations in human cell mutagenicity of respirable airborne particles in the northeastern
      United States. Environ. Sci. Technol. 33: 4407-4415.
 Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. (1993) Quantification of urban
      organic aerosols at a molecular level: identification, abundance and seasonal variation. Atmos Environ Part A
      27: 1309-1330.
 Saxena, P.; Hildemann, L. M. (1996) Water-soluble organics in atmospheric particles: a critical review of the
      literature and applications of thermodynamics to identify candidate compounds. J. Atmos. Chem. 24:  57-109.
March 2001
3 A-15        DRAFT-DO NOT QUOTE OR CITE

-------
Schauer, J. J.; Cass, G. R. (2000) Source apportionment of wintertime gas-phase and particle-phase air pollutants
      using organic compounds as tracers. Environ. Sci. Technol. 34:1821-1832.
Turpin, B. J.; Huntzicker, J. J. (1995) Identification of secondary organic aerosol episodes and quantitation of
      primary and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29: 3527-3544.
U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
      NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
Veltkamp, P. R.; Hansen, K. J.; Barkley, R. M.; Sievers, R. E. (1996) Principal component analysis of summertime
      organic aerosols at Niwot Ridge, Colorado. J. Geophys. Res. [Atmos.] 101: 19,495-19,504.
 March 2001
3 A-16
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
32
                                APPENDIX 3B
              .Composition of Particulate Matter Source Emissions

     This appendix includes discussions of the elemental composition of emissions from various
source categories discussed in Table 3-7. Discussions in this appendix incorporate material
dealing with the inorganic components of source emissions from Chapter 5 of the 1996 PM
AQCD (U. S. Environmental Protection Agency, 1996), updates to that material, and material
describing the composition of organic components in source emissions. The primary emphasis in
the discussions is on the composition of PM25 particle sources.

Soil and Fugitive Dust
     The compositions of soils and average crustal material are shown in Table 3B-1 (adapted
from Warneck, 1988).  Two entries are shown as representations of average crustal material.
Differences from the mean soil composition shown can result from local geology and climate
conditions. Major elements in both soil and crustal profiles are Si, Al, and Fe, which are found in
the form of various minerals. In addition, organic matter constitutes a few percent, on average, of
soils.  In general, the soil profile is similar to the crustal profiles, except for the depletion of
soluble elements such as Ca, Mg, Na, and K. It should be noted that the composition of soils from
specific locations can vary considerably from these global averages, especially for elements like
Ca, Mg, Na, and K.
     Fugitive dust emissions arise from paved and unpaved roads, building construction and
demolition, parking lots, mining operations, storage piles, and agricultural tilling in addition to
wind erosion. Figure 3B-1 shows examples of size distributions in dust from paved and unpaved
roads, agricultural soil, sand and gravel, and alkaline lake bed sediments, which were measured in
a laboratory resuspension chamber as part of a study in California  (Chow et al., 1994).  This figure
shows substantial variation in particle size among some of these fugitive dust sources.  The PM, „
abundance (6.9%) in the total suspended PM (TSP) from alkaline lake bed dust is twice its
abundance in paved and unpaved road dust.  Approximately 10% of the TSP is in the PM2 5
fraction and approximately 50% of TSP is in the PM10 fraction. The sand/gravel dust sample
shows that  65% of the mass is in particles larger than the PM10 fraction. The PM2 5 fraction of
       March 2001
                                        3B-1
DRAFT-DO NOT QUOTE OR CITE

-------
                TABLE 3B-1.  AVERAGE ABUNDANCES OF MAJOR ELEMENTS IN
                                   SOIL AND CRUSTAL ROCK

Element
Si
Al
Fe
Ca
Mg
Na
K
Ti
Mn
Cr
V
Co
Elemental
Soil
0)
330,000
71,300
38,000
13,700
6,300
6,300
13,600
4,600
850
200
100
8
Abundances (ppmw)

(2)
277,200
81,300
50,000
36,300
20,900
28,300
25,900
4,400
950
100
135
25

Crustal Rock
(3)
311,000
77,400
34,300
25,700
33,000
31,900
29,500
4,400
670
48
98
12
        Source: (1) Vinogradov (1959); (2) Mason (1966); (3) Turekian (1971), Model A; as quoted in Warneck (1988).
 1     TSP is approximately 30 to 40% higher in alkaline lake beds and sand/gravel than in the other soil
 2     types. The tests were performed after sieving and with a short (<1 min) waiting period prior to
 3     sampling. It is expected that the fraction of PMj 0 and PM25 would increase with distance from a,
 4     fugitive dust emitter as the larger particles deposit to the surface faster than do the smaller
 5     particles.
 6          The size distribution of samples of paved road dust obtained from a source characterization
 7     study in California is shown in Figure 3B-2.  As might be expected, most of the emissions are in
 8     the coarse size mode. The chemical composition of paved road dust obtained in Denver, CO,
 9     during the winter of 1987-1988 is shown in Figure 3B-3.  The chemical composition of paved
10     road dust consists of a complex mixture of particulate matter from a wide variety of sources.
11     Hopke et al. (1980) found that the inorganic composition of urban roadway dust in samples from
       March 2001
3B-2
DRAFT-DO NOT QUOTE OR CITE

-------
             100
              80
            at
              60
              40
              20
                     Paved       Unpaved     Agricultural     Soil/Gravel
                   Road Dust    Road Dust        Soil
                           [Zl<1.0|jnn  IHl<2.5|jm Bl<10|jm  KDlSP
                             Alkaline
                            Lake Bed
       Figure 3B-1. Size distribution of particles generated in a laboratory resuspension
                    chamber.
       Source: Chow etal. (1994).
 1     Urbana, IL, could be described in terms of contributions from natural soil, automobile exhaust,
 2     rust, tire wear, and salt.  Automobile contributions arose from exhaust emissions enriched in Pb;
 3     from rust as Fe; tire wear particles enriched in Zn; brake linings enriched in Cr, Ba, and Mn; and
 4     cement particles derived from roadways by abrasion. In addition to organic compounds from
 5     combustion and secondary sources, road dust also contains biological material such as pollen and
 6     fungal spores.
 7          Very limited data exist for characterizing the composition in organic compounds
 8     resuspended paved road dust and soil dust. The only reported measurements are from Rogge et al.
 9     (1993a) and Schauer and Cass (2000), which consist of data for the fine particle fraction. The
10     resuspended road dust sample analyzed Rogge et al. (1993 a) was collected in Pasadena, California
       during May of 1988. The sample analyzed by Schauer and Cass (2000) is a composite sample
       March 2001
3B-3
DRAFT-DO NOT QUOTE OR CITE

-------
     100
      80
      60
    1
    DL
      40
      20
                                         95.8%
                                         93.1%
                                         (<2.5R
                                         92.4%
                          99.2%


                          97.4%
                          (<2.5|J)


                          87.4%
                                                                            34.9%
         Road and   Agricultural  Residential    Diesel      Crude ON  Construction
         Soil Dust     Burning       Wood      Truck     Combustion     Dust
                                Combustion   Exhaust
           Code:
2.5p-10M
1M-2.5|J
Figure 3B-2.  Size distribution of California source emissions, 1986.

Source: Houck et al. (1989,1990).

March 2001                               3B-4       DRAFT-DO NOT QUOTE OR CITE

-------
                                          Chemical Species

      Figure 3B-3. Chemical abundances for PM2 5 emissions from paved road dust in Denver,
                   CO. Solid bars represent fractional abundances, and the error bars
                   represent variability in species abundances. Error bars represent detection
                   limits when there are no solid bars.
      Source: Watson and Chow (1994).
1
2
3
4
5
6
7.
collected at several sites in the Central Valley of California in 1995. In both cases, road dust
samples were resuspended in the laboratory.  Samples were drawn through a PM2 0 cyclone
upstream of the collection substrate to remove particles with aerodynamic diameters greater than
2.0 /um. It is unclear if these samples are representative of road dust in other locations of the
United States. Table 3B-2 summarizes the organic compounds measured in these road dust
samples.
      March 2001
                                         3B-5
DRAFT-DO NOT QUOTE OR CITE

-------
             TABLE 3B-2.  SUMMARY OF PARTICLE-PHASE ORGANIC COMPOUNDS
                       PRESENT IN FINE PARTICLE ROAD DUST SAMPLE
Source
Pasadena Road Dust
(Roggeetal., 1993a)





San Joaquin Valley
Road Dust (Schauer
and Cass, 2000)

Contribution to Dominant Contributors to
Compound Class Particulate Mass (%) Emissions of Compound Class
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
Petroleum biomarkers
PAH
n-Alkanals
n-Alkanols
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
0.13
0.37
0.028
0.017
0.0059
0.046
0.021
0.023
0.23
0.095
C]7> Ci9, C2,
Palmitic acid and stearic acid
Oleic acid and linoleic Acid
Hopanes and steranes
No dominant compounds
Octacosanol and triacontanal
Hexacosanol and octacosanol
No dominant compounds
Palmitic acid and stearic acid
Oleic acid, linoleic acid, and
hexadecenoic acid
 1     Stationary Sources
 2          The elemental composition of primary particulate matter emitted in the fine fraction from a
 3     variety of power plants and industries in the Philadelphia area is shown in Table 3B-3 as a
 4     representative example of emissions from stationary fossil combustion sources (Olmez et al.,
 5     1988). Entries for the coal fired power plant show that Si and Al followed by sulfate are the
 6     major primary constituents produced by coal combustion, whereas fractional abundances of
 7     elemental carbon were much lower and organic carbon species were not detected. Sulfate is the
 8     major particulate constituent released by the oil fired power plants examined in this study, and,
 9     again, elemental and organic carbon are not among the major species emitted.  Olmez et al. (1988)
10     also compared their results to a number of similar studies and concluded that their data could have
11     much wider applicability to receptor model studies in other areas with some of the same source
12     types. The high temperature of combustion in power plants results in the almost complete
13     oxidation of the carbon in the fuel to CO2 and very small amounts of CO. Combustion conditions
14     in smaller boilers and furnaces allow the emission of unbumed carbon and sulfur in
       March 2001
3B-6
DRAFT-DO NOT QUOTE OR CITE

-------





^H
51
fc
g
H
CO
0
>
pa
Q
CO
h-i
CO
u
o
c
f^
M"j
s
^4
g
g
§
0
H
^••4
WD
o
CU
M

o
u
s

a
3
H

















3
1
ELPHIA
Q
j
a
PH
r 1
s
ffi
P""4
g
co
u
0
CO




















"c
_j
ji
i
*



c
^
5


c
a
C
0

41
NO
en
O

m
CN,
CN
0
41
O
CN

ON
41
CN

^


C--
i — i
41
f-.
m"

co


vO
. i— i
41
in
rn



m

ON
C3
41
ON
oo
"-•'

C?
5
2
en

in
en
41
NO


en

0.38 ±0.0
„

oo
O
41
en
NO

m


00
o
41
rn
rn

en


oo
o
41
p
en"



en

8
o
41
en
0

^N.
^a
^
en
O

0
41
in
CN
O

en

41-
00
NO
^_

S
C>
41

en

oo
O
o
41
Tf
ON
O

m

ON
0
O
41
in
"si*
o


m


41
2


^
^
<;
o

m
o
41
-

ON
O
o
CN
41
oo
ON

CN

CN
CN
41
en

~

^c
O
41
NO
CN"

ON
NO
0
41
ON
*™^

ON
NO
41
oq
g
K
0
CN

O
41
en
NO
C5


Q

CN

r-
CN
o
41
O

«

CN
C5
41
p

0\

O
41
""!
r~t

ON
0.62 ±0.10
^
a-
o

oo
«D
41
ON
CN

ON
NO
CN
41

CN


41
en

~


41
m

ON

CN
41
*"~*


ON
NO
O
41
en
g

-

en
CN
41
oo
NO

en
41
00
m

CN


CN
41
ON

*


41
^

cn

T}*
41
O


en
CN
41
ON
g>
cf
CO
en


in
41
ON
CN


Q

F— t


41
CN




Q


CN
ON
§
O
41
ON
O
O
m
0.022 ±0.11
g?
O
0

m
CN
41
NO

ON
O
O
O
41
m
o
o
CN

in
41
ON
O

~

8
o"
41
CN
O

ON
O
0
41
NO
O

ON
1.20 ±0.09
g
&
0
o

o
41
m
CN
0

ON
.030 ±0.004
o
CM

%
o
41
CN

m

p
41
en
CN"

en
p
_J
41
NO
en

en
0
41
rt
T— t
?
3
-
CN
0
O
41
O

en
O
41
CN

-
ON
en
o
.092 ±0.
o
en

8
o
41
S
0

en
s
o
41
f-
0

en
CN
41
a
u
GO
o
£
0
o
41
o
8
o"
m
0.38 ±0.1

CN
en
O
o
.024 ±0.

~

8
o
41
CN
O

ON
I
o
41

o
0
en
CN
O
41
g
P
CN

en
in
41
NO
00

en
o
>n
CN

-


41
NO
en

en
O
§
en
41
O
O
o
cT
CN
m
o
o
o
m
41
O
o
in
~
m
0
r—
41
O
>n
in
1

en


en
41
ON
' ON

en
oo
41
ON
m

—

O
CN
41
O

en

o
t-~
41
o
en
CN

m
O
*— *
41
m
m
CN

m
o
CN
41
m
I
5
en

0
^r
41
in
NO

en
en
41

-

m
41
CN

en

O
in
41
O
CN

en

^"
41
O
oo
m .

m
m
41
o
ON
CN
N§
J
m

o
o
41
CN
CN
0

ON
8
o
41
O
CN
o

-

CN
O
O
41
en
O

m

^
o
41


en
CN
O
41
NO
—•

m
C5
41
NO
g
U
tin
m

oo
c>
41
en

m
CN
41
in

~


CN
41
m

m

o
o
CN
41
O
O

m
o
»— i
41
O
?^

m
o
41
en
ON
I
o
u
O

0
TT
41
CN

ON
O
en
41
CN
CN

CN

O
41
§
en

~

i
CN
41
1
"
ON
g
O
in
41
o
o
o
in

ON
O
in
41
0
oo
en
|

March 2001
3B-7
DRAFT-DO NOT QUOTE OR CITE

-------



1ONARY
<
H
co
1
ps
•n t-~
Tt O
C5
«— CO CO
"-1

o o
^ 3 s
•H -H -H
g 00 0
2 <~: "">
•— • o


os co —

cX 2 *
41 -H •«
S 2 s
o\ *^
O\ CO CO
290 ±20
0.041 ±0.005
640 ±80

^ ~ t
a I a
a .g <
CO O

o
S §
•" «
9 1
CN
CO OS


oo
~H 1 1
Tl
t/-J


—• fN

0
co t--
-H -H
VO O
vo co
VO


CO -^
1-H

41 -H


CO OS

Os g
-H 41
VO o
^ §
CO CO
250 ±20
35 ±8

0. 0.
CO (S
CN

230 ±50



Q




_

00
CO
-H





Q




Q
Z
-
190 ±80

a.
a
£
o

00
Os ON


VO S
41 41
VO O
CO CO




Q Q
Z Z


.-. <— 1
*— 1 *— <

41 ^


OS Os

s §
il
OS OS
1290 ±60
490 ±190

It "s1
°- g;
* . fl
O CO

o .
CO V~l
41 41
O —
•^t r—
CN



0 ,Q
Z Z





ii


CO


1500 ±300
ND


CO

930 ±210
ND
fN
O
VO
-H Q
o Z
r-

1!
o on
S <
CO CO O
_ o
1 2 §
-H -H -H
| -3 |



Q Q Q
Z Z Z



fN

S
Q Q ^
Z Z 5
v>
in

^-
i— *

o
oo
§ i o
s


Os

O
CO
Q Q S
* 2 S
o co rN
41 41 41
O O 0
8 m vo
^ CO
fN


CO CO CO

00^
*~^ rn rN
o o •*-
S 2 S
CO fN
O O
^\ _LJ _|,l
1-1 -H -H
Zoo
CN OO

'e' 1- "i"
P. & OH
a a a
es BJ 1>
pa j c_>


Q Q
CO CO



S £


fN CO

t-;
i I
CO
CO CO
41 "H
g s

II
a a
Z on
March 2001
3B-8
DRAFT-DO NOT QUOTE OR CITE

-------



oa'
o

£
>"
03
Q |
|E
4> <
t>D



12



a 1
J J
&< -S
fe "»
o
CL,
E
O (o
I
•e


2
I
1 s,
Eddystone Cc
Fired Powe
Plant




§.!
oo 3


Q Q Q Q Q


en en en en en

T 0 ^ "* ~
O — — < O O
-H -H -H -H -H
^ f~ oo en ""!
O




Q Q Q Q Q
22222





en en


en ON
tN tN
.® Q ® Q Q
VO ON
G> 0
Q Q Q Q Q


en en «

>n m "1
S Q S ? Q
^, 2 m en 2
"> <^ 2



^-s ^v
s ft s 1. ^
3 T3 ^ JO 3
BO O H > ,3
en o . r~
tN tN
e Q Q ° ^ Q ^
^ ^ s
0

en en ON en ' r^-
fN
00 0 0
 2 vo
"~
"- fN CN ON tN VO
r^ »—>
p ^ p vq >r> *}
ON S ^|- oo ON ^~l
£ p — — &
O

en — H ON en VO
^
00 §
O °° G> & <& & & & & g















c
CQ
(H
U
e

u
^S
*o
c
1
en
•o
sntage of the total emitte
V
S §3
1 s-
c! OJQ
1 >
C ni
1 $
i §5 'Sb
«2 f- . .8
1 cH o §
§ O £ E
O t-4 O ^
.0 ^ ^ S
1 If?
p ill





























oo
00
•3
«
from Olmez i
1



u
o
w
March 2001
3B-9
                                         DRAFT-DO NOT QUOTE OR CITE

-------
 1      more reduced forms such as thiophenes and inorganic sulfides. A number of trace elements are
 2      greatly enriched over crustal abundances in different fuels, such as Se in coal and V, Zn, and Ni in
 3      oil. In fact, the higher V content of the fuel oil than in coal could help account for the higher
 4      sulfate seen in the profiles from the oil-fired power plant compared to the coal-fired power plant
 5      because V at combustion temperatures found in power plants is known to catalyze the oxidation
 6      of reduced sulfur species.  During combustion at lower temperatures, the emission of reduced
 7      sulfur species also occurs. For example, Huffman et al. (2000) identified sulfur species emitted
 8      by the combustion of several residual fuels oils (RFO) in a fire tube package boiler, which is
 9      meant to simulate conditions in small institutional and industrial boilers. They found that sulfur
10     was emitted not only as sulfate (26 to 84%), but as thiophenes (13 to 39%) with smaller amounts
11      of sulfides and elemental S. They also found that Ni, V, Fe, Cu, Zn, and Pb are present mainly as
12     sulfates in emissions. Linak et al. (2000) found, when burning RFO, that the fire tube package
13     boiler produced particles with a bimodal size distribution in which about 0.2% of the mass was
14     associated with particles smaller than 0.1-jum AD, with the rest of the mass lying between 0.5 and
15     100 //m. Miller et al. (1998) found that larger particles consisted mainly of cenospheric carbon,
16     whereas trace metals and sulfates were found concentrated in the smaller particles in a fire tube
17     package boiler. In contrast, when RFO was burning in a refractory-lined combustor, which is
18     meant to simulate combustion conditions in a large utility residual oil fired boiler, Linak et al.
19     (2000) found that particles were distributed essentially unimodally, with a mean diameter of about,
20     0.1 //m.
21           Apart from emissions in the combustion of fossil fuels, trace elements are emitted as the
22     result of various industrial processes such as steel and iron manufacturing and nonferrous metal
23     production (e.g., for Pb, Cu, Ni, Zn, and Cd). As may be expected, emissions factors for the
24     various trace elements are highly source-specific (Nriagu and Pacyna, 1988).  Inspection of
25     Table 3B-3 reveals that the emissions from the catalytic cracker and the oil-fired power plant are
26      greatly enriched in rare-earth  elements such as La compared to other sources.
 27           Emissions from municipal waste incinerators are heavily enriched in Cl arising mainly from
 28      the combustion of plastics and metals that form volatile chlorides. The metals can originate from
 29      cans or other metallic objects and some metals such as Zn and Cd are also additives in plastics or
 30      rubber.  Many elements such  as S, Cl, Zn, Br, Ag, Cd, Sn, In, and Sb are enormously enriched
 31      compared to then- crustal abundances. A comparison of the trace elemental composition of
        March 2001
3B-10       DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
31
 incinerator emissions in Philadelphia, PA (shown in Table 3B-3), with the composition of
 incinerator emissions in Washington DC, and Chicago, IL (Olmez et al., 1988), shows agreement
 for most constituents to better than a factor of two.
      Very limited data exist for characterizing the chemical composition of organic compounds
 present in particulate emissions from industrial-scale stationary fuel combustion.  Oros and
 Simoneit (2000) have presented the abundance and distribution of organic constituents in coal
 smokes that have been burned under laboratory conditions. This work provides the basis for
 further investigation addressing the emissions of coal fired boilers.
      Rogge et al. (1997a) measured the composition of the organic constituents in the particulate
 matter emissions from a 50 billion kj/h boiler that was operating at 60% capacity and was burning
 number 2 distillate fuel oil. The fine carbon particulate matter emissions from this boiler over
 five tests were composed of an average of 14% organic  carbon and 86% elemental carbon
 (Hildemann et al., 1991). Significant variability in the distribution of organic compounds present
 in the emissions from two separate tests was observed.  Most of the identified organic mass
 consisted of n-alkanonic acids, aromatic acids, n-alkanes, PAH, oxygeanted PAH, and chlorinated
 compounds. It is unclear if these emissions are representative of typical fuel oil combustion units
 in the United States.  Rogge et al. (1997b) measured the composition of hot asphalt roofing tar
 pots, and Rogge et al. (19.93b) measured the composition of emissions from home appliances that
 use natural gas.

 Motor Vehicles
      Exhaust emissions of particulate matter from gasoline powered motor vehicles and diesel
 powered vehicles have changed significantly over the past 25 years (Sawyer and Johnson, 1995;
 Cadle et al., 1999). These changes have resulted from reformulation of fuels, the wide application
 of exhaust gas treatment in gasoline-powered motor vehicles, and changes in engine design and
 operation. Because of these evolving tailpipe emissions, along with the wide variability of
 emissions between vehicles of the same class (Hildemann et al., 1991; Cadle et al., 1997; Sagebiel
 et al., 1997; Yanowitz et al., 2000), well-defined average emissions profiles for the major classes
of motor vehicles have not been established. Two sampling strategies have been employed to
obtain motor vehicle emissions profiles: (1) the measurement of exhaust emissions from vehicles
operating on dynamometers and (2) the measurement of integrated emissions of motor vehicles
       March 2001
                                                3B-11       DRAFT-DO NOT QUOTE OR CITE

-------
 1      driving through roadway tunnels.  Dynamometer testing can be used to measure vehicle emissions
 2      operating over an integrated driving cycle and allows the measurement of emissions from
 3      individual vehicles.  However, dynamometer testing requires considerable resources and usually
 4      precludes testing a very large number of vehicles. In contrast, a large number of vehicles can be
 5      readily sampled in tunnels, but vehicles driving through tunnels operate over limited driving
 6      conditions and the measurements represent contributions from a large number of vehicle types.
 7      As a result, except in a few cases, runnel tests have not been effective at developing chemically
 8      speciated particulate matter emissions profiles for individual motor vehicle classes. As a result,
 9      several studies have measured the contribution of both organic and elemental carbon to the
10     particulate matter emissions from different classes of motor vehicles operating on chassis
11      dynamometers.
12           The principal components emitted by diesel and gasoline fueled vehicles are organic carbon
13     (OC) and elemental carbon (EC) as shown in Tables 3B-4a and 4b. As can be seen, the variability
14     among entries for an individual fuel type is large and overlaps that found between different fuel
15     types. On average, the abundance of elemental carbon is larger than that of organic carbon in the
16     exhaust of diesel vehicles, whereas organic carbon is the dominant species in the exhaust of
17     gasoline fueled vehicles. Per vehicle, total carbon emissions from light and heavy duty diesel
 18     vehicles can range from 1 to 2 orders  of magnitude higher than those from gasoline vehicles.
 19     There appears to be a tendency for emissions of elemental carbon to increase relative to emissions
20      of organic carbon for gasoline fueled vehicles as simulated driving conditions are changed from a
21      steady 55 km/h to the various load conditions specified in the Federal Test Procedures (FTPs).
 22      Also shown are the results of sampling from mixed vehicle types along roadsides and in tunnels.
 23           As might be expected, most of the PM emitted by motor vehicles is in the PM2.5 size range.
 24      Particles in diesel exhaust are typically trimodal consisting of a nuclei mode, an accumulation
 25      mode and a coarse mode and are lognormal  in form (Kittelson, 1998). More than 90% of the total
 26     number of particles are in the nuclei mode, which contains only about 1  to 20% of the particle
 27     mass with a mass median diameter of about 0.02 /an, whereas the accumulation mode (with a
 28     mass median diameter of about 0.25 yum) contains most of the mass with a smaller fraction (5 to
 29     20%) contained in the coarse mode. Kerminin et al. (1997), Bagley et al. (1998), and Kleeman
 30     et al. (2000) also have shown that gasoline and diesel fueled vehicles produce particles that are
 31     mostly less than 2.0 too. in diameter.  Cadle et al. (1999) found that 91% of PM emitted by in-use
        March 2001
                                                  3B-12      DRAFT-DO NOT QUOTE OR CITE

-------
          TABLE 3B-4a. ORGANIC AND ELEMENTAL CARBON FRACTIONS OF DIESEL
                   AND GASOLINE ENGINE PARTICIPATE MATTER EXHAUST

Heavy-duty diesel engines3
Heavy-duty diesel engines (SPECIATE)b
Light-duty diesel engines0
Light-duty diesel engines (SPECIATE)"
Gasoline engines (hot stabilized)2
Gasoline engines ("smoker" and "high emitter")3'0
Gasoline engines (cold start)3
Organic Carbon
19 ±8%
21 - 36% .
30 ± 9%
22 - 43%
56 ±11%
76 ±10%
46 ± 14%
Elemental Carbon
75 ± 10%
52 - 54%
61 ± 16%
51 - 64%
25 ± 15%
7 ±6%
42 ± 14%
         "Fujita et al. (1998) and Watson et al. (1998).
         bU.S. EPA SPECIATE database.
         "Norbeck et al. (1998).
         Source: U.S. Environmental Protection Agency (1999).
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
gasoline vehicles in the Denver area was in the PM2.5 size range, which increased to 97% for
"smokers" (i.e., light-duty vehicles with visible smoke emitted from their tailpipes) and 98% for
diesels. Durbin et al. (1999) found that about 92% of the PM was smaller than 2.5 ^m for
smokers and diesels. The mass median diameter of the PM emitted by the gasoline vehicles
sampled by Cadle et al. (1999) was about 0.12 ,um, which increased to 0.18 /^m for smokers and
diesels. Corresponding average emissions rates of PM25 found by Cadle et al. (1999) for diesels
were 552 mg/mi; for smokers they were 222 mg/mi; and, for gasoline vehicles, they were
38 mg/mi. The values for smokers and for diesels appear to be somewhat lower than those given
in Table 3B-5, whereas the value for gasoline vehicles falls in the range given for low and
medium gasoline vehicle emissions.
     Examples of data for the trace elemental composition of the emissions from a number of
vehicle classes obtained as part of the North Frontal Range Air Quality Study (NFRAQS), which
took place in December 1997 in Colorado are shown in Table 3B-5. As can be seen from
Table 3B-5, emissions of total carbon (TC), which is equal to the sum of organic carbon (OC) and
elemental carbon (EC), from gasoline vehicles are highly variable.  Gillies and Gertler (2000)
point out that there is greater variability in the concentrations of trace elements and ionic
       March 2001
                                        3B-13
DRAFT-DO NOT QUOTE OR CITE

-------
           TABLE 3B-4b.  CONTRIBUTION OF ORGANIC CARBON TO PARTICIPATE
         MATTER CARBON EMISSIONS IN MOTOR VEHICLE EXHAUST COLLECTED
                  FROM VEHICLES OPERATED ON CHASSIS DYNAMOMETERS

GASOLINE POWERED VEHICLES
Light-duty vehicles
High-CO/VOC-emitting smokers
High-CO/VOC-emitting nonsmokers
Catalyst-equipped vehicles
Noncatalyst vehicles
DIESEL VEHICLES
Light-duty diesel vehicles
Medium-duty diesel vehicles
Heavy-duty diesel vehicles
Heavy-duty diesel vehicles
Year of Tests

1996-97
1994
1994
Mid-1980s
Mid-1980s

1996-1997
1996
1992
Mid-1980s
Test Cycle

FTP
IM-240
IM-240
FTP
FTP

FTP
FTP
c
c
Number of
Vehicles

195a
7
15
7
6

195a
2
6
2
OC%of
Total Carbon

70
91
76
69
89

40
50"
42
45
Notes

A
B
B
C
C

A
D
E
C
        Notes:
        A. From (Cadle et al., 1999). Average of summer and winter cold start emissions.
        B. From (Sagebiel et al., 1997). Hot start testing of vehicles identified as either high emitters of carbon
           monoxide or volatile organic compounds (VOCs).
        C. From (Hildemann et al., 1991). Cold start tests.
        D. From (Schauer et al., 1999). Hot start tests of medium duty vehicles operating on an FTP cycle.
        E. From (Lowenthal et al., 1994). Only includes measurement of vehicles powered by diesel fuel operated
           without an exhaust particulate trap.

        "A total of 195 light duty vehicles were tested that include both gasoline powered vehicles and diesel powered
        vehicles.
        bFraction of particulate matter consisting of organic carbon was measured with and without an organics denuder
        upstream of particulate filter. Results reported here represent measurement without an organics denuder for
        consistency with other measurements. Using an organics denuder, the organic carbon comprised 39% of the
        particulate matter carbon.
        'Driving cycle comprised of multiple idle, steady acceleration, constant speed, deceleration steps (see reference
        for more details).
1      species than for OC and EC among different source profiles  (e.g., SPECIATE, Lawson and Smith

2      (1998), Norbeck et al. (1998)). They suggest that this may arise because their emissions are not

3      related only to the combustion process, but also to their abundances in different fuels and
       March 2001
3B-14
DRAFT-DO NOT QUOTE OR CITE

-------
       TABLE 3B-5. EMISSION RATES (mg/mi) FOR CONSTITUENTS OF PARTICIPATE
                     MATTER FROM GASOLINE AND DIESEL VEHICLES
Gasoline Vehicles

TC
OC
EC
N03"
SO4=
Na
Mg
Al
Si
P
S
Cl
K
Ca
Fe
Ni
Cu
Zn
Br
Ba
Pb
Low
9.07 ± 0.75
6.35 ± 0.54
2.72 ± 0.52
0.039 ± 0.027
0.1 58 ±0.036
0.060 ± 0.063
0.036 ± 0.022
0.083 ±0.016
0.066 ± 0.008
0.035 ± 0.004
0.085 ± 0.006
0.024 ±0.01 2
0.010 ±0.009
0.060 ±0.010
0.1 43 ±0.004
0.001 ±0.004
0.002 ± 0.004
0.048 ± 0.003
0.001 ± 0.002
0.01 3 ±0.1 36
0.007 ±0.006
Medium
41. 30 ±1.68
26.02 ±1.31
15.28 ±0.99
0.057 ± 0.028
0.51 8 ±0.043
0.023 ±0.1 11
0.068 ± 0.027
0.078 ±0.01 6
0.279 ± 0.01 1
0.152 ±0.007
0.442 ±0.009
0.038 ±0.012
0.01 9 ±0.009
0.212 ±0.011
0.756 ± 0.005
0.005 ± 0.004
0.01 6 ±0.003
0.251 ±0.004
0.01 6 ±0.002
0.009 ±0.1 38
0.085 ± 0.005
High
207.44 ± 7.29
95.25 ± 4.28
112.19 ±5.82
0.141 ±0.031
0.651 ±0.052
0.052 ± 0.092
0.041 ± 0.033
0.057 ±0.014
0.714 ±0.012
0.1 13 ±0.007
0.822 ± 0.022
0.081 ±0.020
0.031 ±0.035
0.210 ±0.030
1.047 ±0.010
0.01 1 ± 0.005
0.021 ±0.005
0.265 ± 0.023
0.079 ± 0.003
0.011 ±0.299
0.255 ± 0.008
Smoker
456.38 ± 16.80
350.24 ±15.27
106.14 ±5.42
0.964 ±0.051
2.160 ±0.137
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
2.515 ±0.116
0.140 ±0.1 17
0.033 ± 0.386
0.362 ±0.250
2.438 ± 0.054
0.008 ±0.017
0.071 ±0.01 8
0.1 88 ±0.272
0.047 ±0.012
0.380 ±2.175
0.345 ± 0.032
Diesel
Light Duty
373.43 ±13.75
132.01 ± 5.82
241.42 ±12.11
1.474 ±0.071
2.902 ±0.1 65
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
2.458 ±0.124
0.228 ±0.1 14
0.000 ± 0.426
0.1 50 ±0.304
0.515 ±0.057
0.014 ±0.018
0.024 ±0.021
0.000 ± 0.299
0.003 ±0.014
0.428 ± 2.390
0.1 53 ±0.033
Vehicles
Heavy Duty
1570.69 ±58.24
253.94 ±16.12
1316.75 ±55.33
1.833 ±1.285
3.830 ±1.286
1.288 ±2. 160
1.061 ±0.729
0.321 ±0.543
8.01 8 ±0.221
0.407 ±0.1 36
3.717±0.111
0.881 ±0.221
0.064 ± 0.248
0.716 ±0.107
0.376 ± 0.055
0.002 ± 0.057
0.001 ±0.062
0.707 ± 0.032
0.01 2 ±0.050
0.493 ±3. 108
0.008 ±0.1 54
       Source: Lawson and Smith (1998).
1     lubricants and also to wear and tear during vehicle operation. Emissions from smokers are
2     comparable to those from diesel vehicles. Thus, older, poorly maintained gasoline vehicles could
3     be significant sources of PM2 5 (Sagebiel et al., 1997; Lawson and Smith, 1998), in addition to
      March 2001
3B-15
DRAFT-DO NOT QUOTE OR CITE

-------
 1      being significant sources of gaseous pollutants (e.g., Calvert et al., 1993). Durbin et al. (1999)
 2      point out that although smokers constitute only 1.1 to 1.7% of the light-duty fleet in the South
 3      Coast Air Quality Management District in California, they contribute roughly 20% of the total PM
 4      emissions from the light-duty fleet. In general, motor vehicles that are high emitters of
 5      hydrocarbons and carbon monoxide also will tend to be high emitters of PM (Sagebiel et al.,1997;
 6      Cadle et al., 1997). Particle emission rates also are correlated with vehicle acceleration and
 7      emissions occur predominantly during periods of heavy acceleration, even in newer vehicles
 8      (Maricq et al., 1999).
                                                                                      %.
 9           Although the data shown in Table 3B-5 indicate that S (mainly in the form of sulfate) is a
10      minor component of PM2 s emissions, S may be the major component of the ultrafine particles that
11      are emitted by either diesel or internal combustion engines (Gertler et al., 2000). It is not clear
12      what the source of the small amount of Pb seen in the auto exhaust profile is. It is extremely
13      difficult to find suitable tracers for automotive exhaust because Pb has been removed from
14      gasoline. However, it also should be remembered that restrictions in the use of leaded gasoline
15      have resulted in a dramatic lowering of ambient Pb levels.
16           Several tunnel studies have measured the distribution of organic and elemental carbon in the
17      integrated exhaust of motor vehicle fleets comprising several classes of motor vehicles (Pierson
18      and Brachaczek, 1983; Weingartner et al., 1997a; Fraser.et al., 1998a). The study by Fraser et al.
19      (1998a) found that organic carbon constituted 46% of the  carbonaceous particulate matter
20      emissions from the vehicles operating in the Van Nuys tunnel in Southern California in the
21      Summer of 1993. Although diesel vehicles constituted only 2.8% of the vehicles measured by
22      Fraser et al. (1998a), the contribution  of the organic carbon to the total particulate carbon
23      emissions obtained in the Van Nuys tunnels is in reasonable agreement with the dynamometer
24      measurements shown in Table 3B-4b.
25           Very few studies have reported comprehensive analyses of the organic composition of motor
26      vehicle exhaust. The measurements by Rogge et al. (1993c) are the most comprehensive, but are
27      not expected to be the best representation of current motor vehicle emissions because these
28      measurements were made in the mid-1980s.  Measurements reported by Fraser et al. (1999) were
29      made in a tunnel study conducted in 1993 and represent integrated diesel and  gasoline powered
30      vehicle emissions. In addition, exhaust emissions from two medium-duty diesel vehicles
31      operating over an FTP cycle were analyzed by Schauer et al. (1999). A unique feature of both the
       March 2001
3B-16
DRAFT-DO NOT QUOTE OR CITE

-------
1 measurements by Faser et al. (1999) and Schauer et al. (1999) is that they include the
2 quantification of unresolved complex mixture (UCM), which comprises aliphatic and cyclic
3 hydrocarbons that cannot be resolved by gas chromatography (Schauer et al., 1999). Schauer
4 etal. (1999) have shown that all of the organic compound mass in their diesel exhaust samples
5 could be extracted and eluted by CG/MS techniques, even though not all of the organic compound
6 mass can identified on a single compound basis. Table 3B-6 summarizes the composition of
7 motor vehicle exhaust
8
9
TABLE 3B-6.


Source
Gasoline and diesel-
powered vehicles
driving through the
Van Nuys Tunnel
(Fraser et al., 1999)a




Medium-duty diesel
vehicles operated over
an FTP Cycle
(Schauer et al., 1999)




measured by Fraser et al.


(1999) and Schauer


etal. (1999).


SUMMARY OF PARTICLE-PHASE ORGANIC COMPOUNDS
EMITTED FROM

Compound Class
n-Alkanes
Petroleum biomarkers
PAH
Aromatic acids
Aliphatic acids
Substituted aromatic
UCM"
n-Alkanes
Petroleum biomarkers
PAH
Aliphatic acids
Aromatic acids
Saturated cycloalkanes
UCMb
MOTOR VEHICLES
Contribution to
Particulate Mass (%)
0.009
0.078
0.38
0.29
0.21
0.042
23.0
0.22
0.027
0.54
0.24
0.014
0.037
22.2
Dominant Contributors to
Emissions of Compound Class
C2, through C29
Hopanes and steranes
No dominant compound
Benzenedicarboxylic acids
Palmitic and stearic acids
No dominant compound

C20 through C2g
Hopanes and steranes
No dominant compound
n-Octadecanoic acid
Methylbenzoic acid
C2i through C25

 "Includes emissions of brake wear, tire wear, and resuspension of road dust associated with motor vehicle traffic.
 bUnresolved complex mixture.
March 2001
3B-17
DRAFT-DO NOT QUOTE OR CITE

-------
  1           Several studies have measured the distribution of polycyclic aromatic hydrocarbons (PAHs)
  2      in motor vehicles exhaust from on-road vehicles (Westerholm et al., 1991; Lowenthal et al., 1994;
  3      Venkataraman et al., 1994; Westerholm and Egeback, 1994; Reilly et al., 1998; Cadle et al., 1999,
  4      Weingartner et al., 1997b; Marr et al., 1999).  Cadle et al. (1999) found high molecular weight
  5      PAHs (PAHs with molecular weights greater than or equal to 202 g/mole) to make up from 0.1 to
  6      7.0% of the particulate matter emissions from gasoline powered and diesel powered light duty
  7      vehicles.  It is important to note, however, that PAHs with molecular weights of
  8      202 (fluoranthene, acephenanthrylene, and pyrene), 226 (benzo[ghi]fluoranthene and
  9      cyclopenta[cd]pyrene), and 228 (benz[a]anthracene, chrysene, and triphenylene) exist in both the
10      gas-phase and particle-phase at atmospheric conditions (Fraser et al., 1998b). Excluding these
11      semi-volatile PAHs, the contribution of nonvolatile PAHs to the particulate matter emitted from
12      the light-duty vehicles sampled by Cadle etal. (1999) ranges from 0.013 to 0.18%.  These
13      measurements are in good agreement with the tunnel study conducted by Fraser et al. (1999) and
14      the heavy-duty diesel truck and bus exhaust measurements by Lowenthal et al. (1994), except that
15      the nonvolatile PAH emissions from the heavy duty diesel vehicles tested by Lowenthal et al.
16      (1994) were moderately higher, making up approximately 0.30% of the particulate matter mass
17      emissions.
18
19      Biomass Burning
20          In contrast to the mobile and stationary sources discussed earlier, emissions from biomass
21      burning in woodstoves and forest fires are strongly seasonal and can be highly episodic within
22      then: peak emissions seasons. The burning of fuelwood is confined mainly to the winter months
23      and is  acknowledged to be a major source of ambient air particulate matter in the northwestern
24      United States during the heating season. Forest fires occur primarily during the driest seasons of
25      the year in different areas of the country and are especially prevalent during prolonged droughts.
26      PM produced by biomass burning outside the United States (e.g., in Central America during the
27      spring of 1988) also can affect ambient air quality in the United States.
28          An example of the composition of fine particles (PM2 5) produced by woodstoves is shown
29      in Figure 3B-4. These data were obtained in Denver during the winter of 1987-1988 (Watson and
30      Chow, 1994).  As was the case for motor vehicle emissions, organic and elemental carbon are the
31      major  components of particulate emissions from wood burning.  It should be remembered
       March 2001
3B-18
DRAFT-DO NOT QUOTE OR CITE

-------
                                             Chemical Species
       Figure 3B-4. Chemical abundances for PM2 5 emissions from wood burning in Denver, CO.
                    Solid bars represent fractional abundances, and the error bars represent
                    variability in species abundances. Error bars represent detection limits when
                    there are no solid bars.
       Source: Watson and Chow (1994).
 1     that the relative amounts shown for organic carbon and elemental carbon vary with the type of
 2     stove, the stage of combustion and the type and condition of the fuelwood. Fine particles are
 3     dominant in studies of wood burning emissions. For instance, the mass median diameter of wood-
 4     smoke particles was found to be about 0.17 fj.m in a study of the emissions from burning
 5     hardwood, softwood, and synthetic logs (Dasch, 1982).
 6          Kleeman et al. (1999) showed that the particles emitted by the combustion of wood in
 7     fireplaces are predominately less than 1.0 /^m in diameter, such that the composition of fine
 8     particulate matter (PM2 5) emitted from fireplace combustion of wood is representative of the total
 9     particulate matter emissions from this source. Hildemann et al. (1991) and McDonald et al.
10     (2000) reported that smoke from fireplace and wood stove combustion consists of 48% to
11     71% OC and 2.9% to 15% EC. Average elemental and organic carbon contents for these
12     measurements are shown in Table 3B-7. It should be noted that the two methods used for the
       March 2001
3B-19
DRAFT-DO NOT QUOTE OR CITE

-------
            TABLE 3B-7. MASS EMISSIONS, ORGANIC CARBON, AND ELEMENTAL
              CARBON EMISSIONS FROM RESIDENTIAL COMBUSTION OF WOOD
Wood Type
Softwood
Softwood
Hardwood
Hardwood
Hardwood
Combustion
Type
Fireplace
Fireplace
Fireplace
Fireplace
Wood Stove
Average Mass
Emission Rate
(gkg"' of wood
burned)
13.0
5.14
5.28
5.66
3.96
Number
of Tests
2
5
3
5
8
Percent
Organic
Carbon3
48.4
58.5
. 48.4
63.2
71.2
Percent
Elemental
Carbon3
5.2
15.0
2,9
7.0
9.0
References
Hildemannetal. (1991)
McDonald et al. (2000)
Hildemann et al. (1991)
McDonald et al. (2000)
McDonald et al. (2000)
        •Hildemann et al. (1991) used the method described by Birch and Cary (1996) to measure EC and McDonald
        et al. (2000) used the method reported by Chow et al. (1993) to measure OC.
 1     measurements shown in Table 3B-7 have been reported to produce different relative amounts of
 2     OC and EC for wood smoke samples, but show good agreement for total carbon (OC + EC)
 3     measurements (Chow et al., 1993).
 4          Hawthorne et al. (1988) and Hawthorne et al. (1989) measured gas-phase and particle-phase
 5     derivatives of guaiacol (2-methoxyphenol), syringol (2,6-dimethoxyphenol), phenol, and catechol
 6     (1,2-benzenediol) in the downwind plume of 28 residential wood stoves and fireplaces. Rogge
 7     et al. (1998) reported a broad range of particle-phase organic compounds in the wood smoke
 8     samples collected by Hildemann et al. (1991), which include n-alkanes, n-alkanoic acids,
 9     n-alkenoic acids, dicarboxylic acids, resin acids, phytosterols, polycyclic aromatic hydrocarbons
10     (PAH), and the compounds reported by Hawthorne et al. (1989). Supplementing these
11     measurements, McDonald et al. (2000) reported the combined gas-phase and particle-phase
12     emissions of PAH and the compounds quantified by Hawthorne et al. (1989).  The measurements
13     by Rogge et al. (1998), which represent a comprehensive data set of the organic compounds
14     present in wood smoke aerosol, are summarized in Table 3B-8. It should be noted, however, that
15     these nearly 200 compounds account for only approximately 15 to 25% of the  organic carbon
16     particle mass emitted from the residential combustion of wood. Simoneit et al. (1999) have
17     shown that levoglucosan constitutes a noticeable portion of the organic compound mass not
18     identified by Rogge et al.  (1998).  In addition, Elias et al. (1999) used high-temperature gas
       March 2001
3B-20
DRAFT-DO NOT QUOTE OR CITE

-------
           TABLE 3B-8. SUMMARY OF PARTICLE-PHASE ORGANIC COMPOUNDS
               EMITTED FROM THE COMBUSTION OF WOOD IN FIREPLACES
                    (Measurements were made using a dilution sampler and no
                            semi-volatile organic compound sorbent.)
Biomass Type
Fireplace
combustion of
softwood







Fireplace
combustion of
hardwood







Contribution to Particulate Dominant Contributors to Emissions
Compound Class Mass (%) of Compound Class
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
Dicarboxylic acids
Resin acids
Substituted phenols
Phytosterols
PAH
Oxygenated PAH
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
Dicarboxylic acids
Resin acids
Substituted phenols
Phytosterols
PAH
Oxygenated PAH
0.039
0.45
0.12
0.36
1.28
3.30
0.37
0.092
0.019
0.044
1.33
0.049
0.42
0.11
8.23
0.21
0.13
0.020
C2) through C31
Cl6> Cl8> C20, C2], C22) C24
Oleic and linoleic acid
Malonic acid
Abietic, dehydroabietic, isopimaric,
pimaric, and sandaracopimaric acids
Benzenediols and guaiacols
P-Sitosterol
Fluoranthene and pyrene
IH-phenalen-l-one
C21 through C29
C|6> C22, C24, C26
Oleic and linoleic acid
Succinic acid
Dehydroabietic acid
Benzediols, guaiacols, and syringols
P-sitosterol
No dominant compounds
1 H-phenalen-1 -one
       Source: Roggeetal. (1998).
1     chromatography/mass spectrometry (HTGC-MS) to measure high-molecular-weight organic
2     compounds in smoke from South American leaf and steam litter biomass burning. These

3     compounds cannot be measured by the analytical techniques employed by Rogge et al. (1998) and,
4     therefore, are strong candidates to make up some of the unidentified organic mass in the wood
      March 2001
3B-21
DRAFT-DO NOT QUOTE OR CITE

-------
 1     smoke samples analyzed by Rogge et al. (1998). These compounds, which include triterpenyl
 2     fatty acid esters, wax esters, triglycerides, and high-molecular-weight n-alkan-2-ones, are
 3     expected to be present in North American biomass smoke originating from agricultural burning,
 4     forest fires, grassland fires, and wood smoke.
 5          Measurements of aerosol composition, size distributions, and aerosol emissions factors have
 6     been made'in biomass burning plumes either on towers (Susott et al., 1991) or aloft on fixed-wing
 7     aircraft (e.g., Radke et al., 1991) or on helicopters (e.g., Gofer et al., 1988). As was found for
 8     woodstove emissions, the composition of biomass burning emissions is strongly dependent on the
 9     stage of combustion (i.e., flaming, smoldering, or mixed), and the type of vegetation (e.g., forest,
10     grassland, scrub). Over 90% of the dry mass in particulate biomass burning emissions is
11     composed of organic carbon (Mazurek et al., 1991). Ratios of organic carbon to elemental carbon
12     are highly variable, ranging from 10:1 to 95:1, with the highest ratio found for smoldering
13     conditions and the lowest for flaming conditions. Emissions factors for total particulate emissions
14     increase by factors of two to four in going from flaming to smoldering stages in the individual
15     fires studied by Susott et al. (1991).
16          Particles in biomass burning plumes from a number of different fires were found to have
17     three distinguishable size modes, (1) a nucleation mode, (2) an accumulation mode, and
18     (3) a coarse mode (Radke et al., 1991). Based on an average of 81 samples, approximately 70%
19-    of the mass was found in particles <3.5 //m in aerodynamic diameter. The fine particle
20     composition was found to be dominated by tarlike, condensed hydrocarbons and the particles were
21     usually spherical in shape.  Additional information for the size distribution of particles produced
22     by vegetation burning was shown in Figure 3B-2.
23          An example of ambient data for the composition of PM25 collected at a tropical site that was
24     heavily affected by biomass burning is shown in Table 3B-9. The samples were collected during
25     November of 1997 on the campus of Sriwijaya University, which is located in a rural setting on
26     the island of Sumatra in Indonesia (Pinto et al., 1998). The site was subjected routinely to levels
27     of PM2 5 well in excess of the U.S. NAAQS as a result of the Indonesian biomass fires from the
28     summer of 1997 through the spring of 1988.  As can be seen from a comparison of the data shown
29     in Table 3B-9 with those shown in Figure 3B-4, there are a number of similarities and differences
30     (especially with regard to the heavy metal content) in the abundances of many species. The
       March 2001
3B-22
DRAFT-DO NOT QUOTE OR CITE

-------
               TABLE 3B-9. MEAN AEROSOL COMPOSITION AT TROPICAL SITE
                 (SRIWIJAYA UNIVERSITY, SUMATRA, INDONESIA) AFFECTED
                          HEAVILY BY BIOMASS BURNING EMISSIONS3
Component
OC
EC
S04=
Al
Si
Cl
K
Ca
Ti
v
Abundance (%)
76
1.2
11
BDb
9.3 x lO'2
4.4
0.7
4.5 x IQ-2
4.2 x IQ-3
BDb
Component
Cr
Mn
Fe
Ni
Cu
Zn
As
Se
Br
Pb
Abundance (%)
BDb
BDb
3.9 x lO'2
<3.8 x IQ-5
4.8 x 1Q-4
3.1 x IQ-3
6.4 x 1Q-4
2.8 x IQ-4
3.6 x lO'2
3.1 x 1Q-3
        aThe mean PM2 5 concentration during the sampling period (November 5 through 11, 1997) was 264 //g/m3.
        bBeneath detection limit.
        Source: Pinto et al. (1998).
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
abundances of some crustal elements (e.g., Si, Fe) are higher in Table 3B-9 than in Figure 3B-4,
perhaps reflecting additional contributions of entrained soil dust.
     Limited emissions data that includes organic compound speciation information have been
reported for agricultural burning (Jenkins et al., 1996), forest fires (Simoneit, 1985), and grassland
burning (Standley and Simoneit, 1987). Jenkins et al. (1996) present PAH emissions factors for
the combustion of cereals (barley, corn, rice, and wheat), along with PAH emissions factors for
wood burning.  Profiles of organic compounds in emissions from meat cooking (Rogge et al.,
1991) and cigarette smoke (Rogge et al., 1994) have been obtained.
       March 2001
                                        3B-23
DRAFT-DO NOT QUOTE OR CITE

-------
 1     Natural Sources
 2           Although sea-salt aerosol production is confined to salt water bodies, it is included here
 3     because many marine aerosols can exert a strong influence on the composition of the ambient
 4     aerosol in coastal areas. In some respects, the production of sea-salt aerosols is like that of
 5     windblown dust in that both are produced by wind agitation of the surface. The difference
 6     between the two categories arises because sea-salt particles are produced from the bursting of air
 7     bubbles rising to the sea surface.  Air bubbles are formed by the entrainment of air into the water
 8     by breaking waves. The surface energy of a collapsing bubble is converted to kinetic energy in
 9     the form of a jef of water that can eject drops above the sea surface.  The mean diameter of the jet
10     drops is about 15% of the bubble diameter (Wu, 1979). Bubbles in breaking waves range in size
11     from a few yum to several mm in diameter.  Field measurements by Johnson and Cooke (1979) of
12     bubble  size spectra show maxima in diameters at around 100 //m, with the bubble size distribution
13     varying as (d/do)'5 with d0 = 100 //m.
14           Because sea-salt particles receive water from the surface layer, which is enriched in organic
15     compounds, the aerosol drops are composed of this organic material in addition to sea salt (about
16     3.5% by weight in sea water). Na+ (30.7%), Cl" (55.0%), SO4= (7.7%), Mg2+ (3.6%), Ca2+ (1.2%),
17     K* (1.1%), HCO3" (0.4%), and Br" (0.2%) are the major ionic species by mass in sea water
18     (Wilson, 1975). The composition of the marine aerosol also reflects the occurrence of
19.     displacement reactions that enrich sea-salt particles in SO4" and NO3", while depleting them of Cl"
20     and Br.
21           Seasalt is concentrated in the coarse size mode with a mass median diameter of about 7 //m
22     for samples collected in Florida, the Canary Islands, and Barbados (Savoie and Prospero, 1982).
23     The size distribution of sulfate is distinctly bimodal. Sulfate in the coarse mode is derived from
24     sea water but sulfate in the submicron aerosol arises from the oxidation of dimethyl sulfide
25     (CH3SCH3) or DMS. DMS is produced during the decomposition of marine micro-organisms.
26     DMS is oxidized to methane sulfonic acid (MSA), a large fraction of which is oxidized to sulfate
27     (e.g., Herteletal., 1994).
28           Apart from sea spray, other natural sources of particles include the suspension of organic
29     debris and volcanism. Profiles of organic compounds in vegetative detritus have been obtained by
30     Rogge et al. (1993d). Particles are released from plants in the form of seeds, pollen, spores, leaf
31     waxes,  and resins, ranging in size from 1 to 250 yum (Warneck, 1988). Fungal spores and animal
        March 2001
3B-24
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
 debris, such as insect fragments, also are to be found in ambient aerosol samples in this size range.
 Although material from all the foregoing categories may exist as individual particles, bacteria
 usually are found attached to other dust particles (Warneck, 1988).  Smaller bioaerosol particles
 include viruses, individual bacteria, protozoa, and algae (Matthias-Maser and Jaenicke, 1994).
 hi addition to natural sources, other sources of bioaerosol include industry (e.g., textile mills),
 agriculture, and municipal waste disposal (Spendlove, 1974).  The size distribution of bioaerosols
 has not been characterized as well as it has for other categories.
     Trace metals are emitted to the atmosphere from a variety of sources such as sea spray,
 wind-blown dust, volcanoes, wildfires and biotic sources (Nriagu, 1989). Biologically mediated
 volatilization processes (e.g., biomethylation) are estimated to account for 30 to 50% of the
 worldwide total Hg, As, and Se emitted annually, whereas other metals are derived principally
 from pollens, spores, waxes, plant fragments, fungi, and algae.  It is not clear, however, how much
 of the biomethylated species are remobilized from anthropogenic inputs. Median ratios of the
natural contribution to globally averaged total sources for trace metals are estimated to be
 0.39 (As), 0.15 (Cd), 0.59 (Cr), 0.44 (Cu), 0.41 (Hg), 0.35 (Ni), 0.04 (Pb), 0.41 (Sb), 0.58 (Se),
0.25 (V), and 0.34 (Zn), suggesting a not insignificant natural source for many trace elements.
It should be noted though that these estimates are based on emissions estimates that have
uncertainty ranges of an order of magnitude.
       March 2001
                                         3B-25
DRAFT-DO NOT QUOTE OR CITE

-------
 1       REFERENCES

 2       Bagley, S. T.; Gratz, L. D.; Johnson, J. H.; McDonald J. F. (1998) Effects of an oxidation catalytic converter and a
 3             biodiesel fuel on the chemical, mutagenic, and particle size characteristics of emissions from a diesel engine.
 4             Environ. Sci. Technol. 32: 1183-1191.
 5       Birch, M. E.; Gary, R. A. (1996) Elemental carbon-based method for occupational monitoring of particulate diesel
 6             exhaust: methodology and exposure issues. Analyst 121: 1183-1190.
 7       Cadle, S. H.; Mulawa, P. H.; Ball, J.; Donase, C.; Weibel, A.; Sagebiel, J. C.; Knapp, K. T.; Snow, R. (1997)
 8             Particulate emission rates from in-use high-emitting vehicles recruited in Orange county, California. Environ.
 9       .     Sci. Technol. 31:3405-3412.
10       Cadle, S. H.; Mulawa, P.; Hunsanger, E. C.; Nelson, K.; Ragazzi, R. A.; Barrett, R.; Gallagher, G. L.; Lawson, D. R.;
11             Knapp, K. T.; Snow, R. (1999) Light-duty motor vehicle exhaust particulate matter measurement in the
12             Denver, Colorado, area. J. Air Waste Manage. Assoc. 49: PM-164-174.
13       Calvert, J. G.; Heywood, J. B.; Sawyer, R. F.; Seinfeld, J. H. (1993) Achieving acceptable air quality: some
14             reflections on controlling vehicle emissions. Science (Washington, DC) 261: 37-45.
15       Chow, J. C.; Watson, J. G.; Pritchett, L. C.;  Pierson, W. R.; Frazier, C. A.; Purcell, R. G. (1993) The DRI
16             thermal/optical reflectance carbon analysis system: description, evaluation and applications in U.S. air quality
17             studies. In: Lodge, J. P., Jr. Fourth international conference on carbonaceous particles in the atmosphere;
18             April 1991; Vienna, Austria. Atmos.  Environ. Part A 27: 1185-1201.
19       Chow, J. C.; Watson, J. G.; Houck, J. E.; Pritchett, L. C.;  Rogers, C. F.; Frazier, C. A.; Egami, R. T.; Ball, B. M.
20             (1994) A laboratory resuspension chamber to measure fugitive dust size distributions and chemical
21             compositions. Atmos. Environ. 28: 3463-3481.
22       Gofer, W. R., Ill; Levine, J. S.; Sebacher, D. I.; Winstead, E. L.; Riggin, P. J.; Brass, J. A.; Ambrosia, V. G. (1988)
23             Particulate emissions from a mid-latitude prescribed chaparral fire. J. Geophys. Res. [Atmos.] 93: 5207-5212.
24       Dasch, J. M. (1982) Particulate and gaseous emissions from wood-burning fireplaces. Environ. Sci. Technol.
25             16:639-645.
26       Durbin, T. D.; Smith, M. R.; Norbeck, J. M.; Truex, T. J.  (1999) Population density, particulate emission
27             characterization, and impact on the particulate inventory of smoking vehicles in the South Coast Air Quality
28             Management District. J. Air Waste Manage. Asssoc. 49: 28-38.
29       Elias, V. O.; Simoneit, B. R. T.; Pereira, A. S.; Cabral, J. A.; Cardoso, J. N. (1999) Detection of high molecular
30             weight organic tracers in vegetation smoke samples by high-temperature gas chromatography-mass
31             spectrometry. Environ. Sci.  Technol. 33: 2369-2376.
32.      Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T. (1998a) Gas-phase and particle-phase organic compounds emitted from
33             motor vehicle traffic in a Los Angeles roadway tunnel. Environ. Sci. Technol. 32: 2051 -2060.
34       Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T.; Rasmussen, R. A. (1998) Air quality model evaluation data for
35             organics. 5. C6-C22 nonpolar and semipolar aromatic compounds. Environ. Sci. Technol. 32: 1760-1770.
36       Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T. (1999) Particulate organic compounds emitted from motor vehicle
37             exhaust and in the urban atmosphere. Atmos. Environ. 33: 2715-2724.
38       Fujita, E.; Watson, J. G.; Chow, J. C.; Robinson, N. F.; Richards, L. W.; Kumar, N. (1998) Northern front range air
39             quality study, volume C: source apportionment and simulation methods and evaluation. Reno, NV: Desert
40             Research Institute; prepared for Colorado State University, Cooperative Institute for Research in the
41             Atmosphere.
42       Gertler, A. W.; Abu-Allaban, M.; Coulombe, W.; Gilles, J. A.; Pierson, W. R.; Rogers, C. F.; Sagebiel, J. C.; Tarnay,
43             L. (2000) Preliminary results of a tunnel study to characterize mobile source particulate emissions. Presented
44             at: PM2000: particulate matter and health—the scientific basis for regulatory decision-making, specialty
45             conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management Association.
46       Gillies, J. A.; Gertler, A. W. (2000) Comparison and evaluation of chemically speciated mobile source PM2 5
47             particulate matter profiles. J. Air Waste Manage. Assoc. 50: 1459-1480.
48       Hawthorne, S. B.; Miller, D. J.; Barkley, R. M.; Krieger, M. S. (1988) Identification of methoxylated phenols as
49             candidate tracers for atmospheric wood smoke pollution. Environ. Sci. Technol. 22:1191-1196.
50      Hawthorne, S. B.; Krieger, M. S.;  Miller, D. J.; Mathiason, M. B. (1989) Collection and quantitation of methoxylated
51             phenol tracers for atmospheric pollution from residential wood stoves. Environ.  Sci. Technol. 23:470-475.
52      Hertel, O.; Christensen, J.;  Hov, 0. (1994) Modelling of the end products of the chemical decomposition of DMS in
53             the marine boundary layer.  Atmos. Environ. 28:2431 -2449.
         March 2001
3B-26
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
48
49
 50
 51
52
53
54
55
  Hildemann, L; M.; Markowski, G. R.; Jones, M. C.; Cass, G. R. (1991) Submicrometer aerosol mass distributions of
       emissions from boilers, fireplaces, automobiles, diesel trucks, and meat-cooking operations Aerosol Sci
       Technol. 14: 138-152.
  Hopke, P. K.; Lamb, R. E.; Natusch, D. F. S. (1980) Multielemental characterization of urban roadway dust Environ
       Sci. Technol. 14: 164-172.
  Houck, J. E.; Chow, J. C.; Watson, J. G.; Simons, C. A.; Pritchett, L. C.; Goulet, J. M.; Frazier, C. A. (1989)
       Determination of particle size distribution and chemical composition of particulate matter from selected
       sources in California: volume I and executive summary (final report). Sacramento, CA: California Air
       Resources Board.
  Houck, J. E.; Goulet, J. M.; Chow, J. C.; Watson, J. G.; Pritchett, L. C. (1990) Chemical characterization of emission
       sources contributing to light extinction. In: Mathai, C. V., ed. Visibility and fine particles: an A&WMA/EPA
       international specialty conference; October 1989; Estes Park, CO. Pittsburgh, PA: Air & Waste Management
       Association; pp. 437-446. (A&WMA transactions series no. TR-17).
 Huffman, G. P.; Huggins, F. E.; Shah, N.; Huggins, R.; Linak, W. P.; Miller, C. A.; Pugmire, R. J.; Meuzelaar, H. L.
       C.; Seehra, M. S.; Manivannan, A. (2000) Characterization of fine particulate matter produced by combustion
       of residual fuel oil. J. Air Waste Manage. Assoc. 50: 1106-1114.                               ;
 Jenkins, B. M.; Jones, A. D.; Turn, S. Q.; Williams, R. B. (1996) Particle concentrations, gas-particle partitioning,
       and species intercorrelations for polycyclic aromatic hydrocarbons (PAH) emitted during biomass burning'
       Atmos. Environ. 30: 3825-3835.
 Johnson, B. D.; Cooke, R. C. (1979) Bubble populations and spectra in coastal waters: a photographic approach
       J. Geophys. Res. C: Oceans Atmos. 84: 3761-3766.
 Kerminen, V. M.; Makela, T. E.; Ojanen, C. H.; Hillamo, R. E.; Vilhunen, J. K.; Rantanen, L.; Havers, N.;
       Von Bohlen, A.; Klockow, D. (1997) Characterization of the particulate phase in the exhaust from a diesel
       car. Environ. Sci. Technol.'Sl:  1883-1889.
 Kittelson, D. B. (1998) Engines and nanoparticles: a review. J. Aerosol Sci. 29: 575-588.
 Kleeman, M. J.; Schauer, J. J.; Cass, G. R. (1999) Size and composition distribution of fine particulate matter emitted
       from wood burning, meat charbroiling, and cigarettes. Environ. Sci. Technol. 33: 3516-3523.
 Kleeman, M. J.; Schauer, J. J.; Cass, G. R. (2000) Size and composition distribution of fine particulate matter emitted
       from motor vehicles. Environ. Sci. Technol. 34: 1132-1142.
 Lawson, D. R.; Smith, R. E. (1998) The northern front range air quality study:  a report to the Governor and General
       Assembly. Fort Collins, CO: Colorado State University; December.
 Linak, W. P.; Miller, C. A.; Wendt, J. O. L.  (2000) Comparison of particle size distributions and elemental
       partitioning from the combustion of pulverized coal  and residual  fuel oil. J. Air Waste Manage Assoc
       50: 1532-1544.                                                                  .
 Lowenthal, D. H.; Zielinska, B.; Chow, J. C.; Watson,  J. G.; Gautam, M.; Ferguson, D. H.; Neuroth, G. R.;  Stevens,
       K. D. (1994) Characterization of heavy-duty diesel vehicle emissions. Atmos. Environ. 28: 731-743!
 Maricq, M. M.; Podsiadlik, D. H.; Chase, R. E. (1999) Gasoline vehicle particle size distributions: comparison of
       steady state, FTP,  and US06 measurements. Environ. Sci. Technol. 33: 2007-2015.
 Marr, L. C.; Kirchstetter, T. W.; Harley, R. A.; Miguel, A. H.; Hering, S. V.; Hammond, S. K. (1999)
       Characterization of polycyclic aromatic hydrocarbons in motor vehicle fuels and exhaust emissions Environ
       Sci. Technol. 33: 3091-3099.
 Mason, B. (1966) Principles of geochemistry. 3rd ed. New York, NY: John Wiley & Sons, Inc.
 Matthias-Maser, S.; Jaenicke, R. (1994) Examination of atmospheric bioaerosol particles with radii > 0 2 urn
      J. Aerosol Sci. 25: 1605-1613.
 Mazurek, M. A.; Cofer, W. R., Ill; Levine, J. S. (1991) Carbonaceous aerosols  from prescribed burning of a boreal
      forest ecosystem. In: Levine, J. S., ed. Global biomass burning: atmospheric,  climatic, and biospheric
      implications. Cambridge, MA: MIT Press; pp. 258-263.
 McDonald, J. D.; Zielinska, B.; Fujita, E. M.; Sagebiel, J. C.; Chow, J. C.; Watson, J. G. (2000) Fine particle and
      gaseous emission rates from residential wood combustion. Environ. Sci.  Technol. 34: 2080-2091
 Miller, C. A.; Linak, W. P.; King, C.; Wendt, J. O. L. (1998) Fine particle emissions from heavy fuel oil combustion
      in a firetube package boiler. Combust. Sci. Technol.  134: 477-502.
Norbeck, J. M.; Durbin, T. D.; Truex, T. J. (1998) Measurement of primary particulate matter emissions from
      light-duty motor vehicles. Riverside, CA: University  of California, College of Engineering, Center for
      Environmental Research and Technology; prepared for Coordinating Research Council, Inc. and South Coast
      Air Quality Management District.
         March 2001
                                                         3B-27       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Nriagu, J. O. (1989) A global assessment of natural sources of atmospheric trace metals. Nature (London)
 2            338:47-49.                                                                 .
 3      Nriagu, J. O.; Pacyna, J. M. (1988) Quantitative assessment of worldwide contamination of air, water and soils by
 4            trace metals. Nature (London) 333: 134-139.
 5      Olmez, I.; Sheffield, A. E.; Gordon, G. E.; Houck, J. E.; Pritchett, L. C.; Cooper, J. A.; Dzubay, T. G.; Bennett, R. L.
 6            (1988) Compositions of particles from selected sources in Philadelphia for receptor modeling applications.
 7            JAPCA 38:1392-1402.
 8      Oros, D. R.; Simoneit, B. R. T. (2000) Identification and emission rates of molecular tracers in coal smoke
 9            paniculate matter. Fuel 79: 515-536.
10      Pierson, W. R.; Brachaczek, W. W. (1983) Paniculate matter associated with vehicles on the road. II. Aerosol Sci.
11            Technol'. 2:1-40.
12      Pinto, J. P.; Grant, L. D.; Hartlage, T. A. (1998) Report on U.S. EPA air monitoring of haze from S.E. Asia biomass
13           ' fires'. Research Triangle Park, NC: U.S. Environmental Protection Agency, National Center for Environmental
14             Assessment; report no. EPA/600/R-98/071.
15      Radke, L. F.; Hegg, D. A.; Hobbs, P. V.; Nance, J. D.; Lyons, J. H.; Laursen, K. K.; Weiss, R. E.; Riggan, P. J.;
16             Ward, D. E. (1991) Particulate and trace gas emissions from large biomass fires in North America. In: Levine,
17             j. s., ed. Global biomass burning: atmospheric, climatic, and biospheric implications. Cambridge, MA: MIT
18             Press; pp. 209-224.
19       Reilly, P. T. A.; Gieray, R. A.; Whitten, W. B.; Ramsey, J. M. (1998) Real-time characterization of the organic
20            'composition and size of individual diesel engine smoke particles. Environ.  Sci. Technol. 32: 2672-2679.
21       Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991) Sources of fine organic
22             aerosol. 1. Charbroilers and meat cooking operations. Environ. Sci. Technol. 25: 1112-1125.
23       Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993a) Sources of fine organic
24             aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks. Environ.
25             Sci. Technol. 27:1892-1904.
26       Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993b) Sources of fine organic
27             aerosol. 5. Natural gas home appliances. Environ. Sci. Technol. 27: 2736-2744.
28       Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993c) Sources of fine organic
29             aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty  diesel trucks. Environ. Sci.
30             Technol. 27: 636-651.
31       Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993d) Sources of fine organic
32             aerosol. 4. Particulate abrasion products from leaf surfaces of urban plants. Environ. Sci. Technol.
33             27:2700-2711.
34'     Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1994) Sources of fine organic
35             aerosol. 6. Cigarette smoke in the urban atmosphere. Environ. Sci. Technol. 28: 1375-1388.
36      Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997a) Sources of fine organic
37             aerosol. 8. Boilers burning no. 2 distillate fuel oil. Environ. Sci. Technol. 31: 2731 -2737.
38      Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997b) Sources of fine organic
39             aerosol. 7. Hot asphalt roofing tar pot fumes. Environ. Sci. Technol. 31: 2726-2730.
40      Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1998) Sources of fine organic
41              aerosol. 9. Pine, oak, and synthetic log combustion in residential fireplaces. Environ. Sci. Technol. 32:13-22.
42      Sagebiel, J. C.;  Zielinska, B.; Walsh, P. A.; Chow, J. C.; Cadle, S. H.; Mulawa, P. A.; Knapp, K. T.; Zweidinger,
43             R. B.; Snow, R. (1997) PM-10 exhaust samples collected during IM-240 dynamometer tests of in-service
44             vehicles  in Nevada. Environ. Sci. Technol. 31: 75-83.
45      Savoie, D. L.; Prospero, J. M. (1982) Particle size distribution of nitrate and sulfate in the marine atmosphere.
46             Geophys. Res. Lett. 9:1207-1210.
47      Sawyer, R. F.; Johnson, J. H. (1995) Diesel emissions and control technology. Diesel exhaust: a critical analysis of
48             emissions, exposure, and health effects. Cambridge, MA: Health Effects Institute.
49      Schauer, J. J.; Cass, G. R. (2000) Source apportionment of wintertime gas-phase  and particle-phase air pollutants
 50             using organic compounds as tracers. Environ. Sci. Technol. 34: 1821-1832.
 51      Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T. (1999) Measurement of emissions from air pollution
 52             sources. 2. C, through C30 organic compounds from medium duty diesel trucks. Environ. Sci. Technol.
 53             33: 1578-1587.
 54      Simoneit, B. R. T. (1985) Application of molecular marker analysis to vehicular exhaust for source reconciliations.
 55             Int. J. Environ. Anal. Chem. 22:203-233.
          March 2001
3B-28
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
40
41
42
43
44
45
46
 Simoneit, B. R. T.; Schauer, J. J.; Nolte, C. G.; Oros, D. R.; Elias, V. O.; Fraser, M. P.; Rogge, W. F.; Cass, G. R.
       (1999) Levoglucosan, a tracer for cellulose in biomass burning and atmospheric particles. Atmos. Environ. 33-
       173-182.
 Spendlove, J. C. (1974) Industrial, agricultural, and municipal microbial aerosol problems. In: Current problems in
       aerobiology: proceedings of the thirtieth general meeting of the Society for Industrial Microbiology; August
       1973; Evanston, IL. Dev. Ind. Microbiol. 15: 20-27.
 Standley, L. J.; Simoneit, B. R. T. (1987) Characterization of extractable plant wax, resin, and thermally matured
       components in smoke particles from prescribed burns. Environ. Sci. Technol. 21: 163-169.
 Susott, R. A.; Ward, D. E.; Babbitt, R. E.; Latham, D. J. (1991) The measurement of trace emissions and combustion
       characteristics for a mass fire. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and
       biospheric implications. Cambridge, MA: MIT Press; pp, 245-257.
 Turekian, K. K. (1971) Elements, geochemical distribution of. In: Encyclopedia of science and technology, v. 4. New
       York, NY: McGraw-Hill Book Company; pp. 627-630.
 U.S. Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park,
       NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/001aF-cF. 3v.
 U.S. Environmental Protection Agency. (1999) Health assessment document for diesel emissions, SAB review draft.
       Washington, DC: Office of Research and Development; report no. EPA/600/8-90/057C.
 Venkataraman, C.; Lyons, J. M.; Friedlander, S. K. (1994) Size distributions of polycyclic aromatic hydrocarbons
       and elemental carbon. 1. Sampling, measurement methods, and source characterization. Environ. Sci. Technol
       28: 555-562.
 Vinogradov, A. P. (1959) The geochemistry of rare and dispersed chemical elements in soils. 2nd ed. New York,
       NY: Consultants Bureau, Inc.
 Warneck, P. (1988) Chemistry of the natural atmosphere. New York, NY: Academic Press, Inc.
 Watson, J. G.; Chow, J. C. (1994) Clear sky visibility as a challenge for society. Annu. Rev. Energy Environ.
       19: 241-266.
 Watson, J. G.; Fujita, E. M.; Chow, J. C.; Zielinska, B.; Richards, L. W.; Neff, W.; Dietrich, D. (1998) Northern fron
       range air quality study final report. Reno, NV: Desert Research Institute; prepared for Colorado State
       University, Cooperative Institute for Research in the Atmosphere.
 Weingartner, E.;  Burtscher, H.;  Baltensperger, U. (1997a) Hygroscopic  properties of carbon and diesel soot particles.
       Atmos. Environ. 31: 2311-2327.
 Weingartner, E.;  Keller, C.; Stahel, W. A.; Burtscher, H.; Baltensperger, U. (1997b) Aerosol emission in a road
       tunnel. Atmos. Environ. 31: 451-462.
 Westerholm, R.; Egeback, K.-E. (1994) Exhaust emissions from light- and heavy-duty vehicles: chemical
       composition, impact of exhaust after treatment, and fuel parameters. In: Symposium of risk assessment of
       urban air: emissions, risk identification, and risk quantitation; May-June 1992; Stockholm, Sweden. Environ.
       Health Perspect. 102(suppl. 4) 13-23.
 Westerholm, R. N.; Almen, J.; Li, H.; Rannug, J. U.; Egeback, K.-E.; Gragg, K. (1991) Chemical and biological
       characterization of particulate-, semi volatile-, and gas-phase-associated compounds in diluted heavy-duty
       diesel exhausts: a comparison of three different semivolatile-phase samplers. Environ. Sci. Technol
       25: 332-338.
Wilson, T. R. S. (1975) Salinity and the major elements of sea water. In: Riley, J. P.; Skirrow, G., eds. Chemical
      oceanography: v. 1. 2nd ed. London, United Kingdom: Academic Press Inc.; pp. 365-413.
Wu, J. (1979) Spray in the atmospheric surface layer: review and analysis of laboratory and oceanic results.
      J. Geophys. Res. C: Oceans Atmos. 84: 1693-1704.
Yanowitz, J.; McCormick, R. L.; Grabowski, M. S. (2000) In-use emissions from heavy-duty diesel vehicles.
      Environ. Sci. Technol. 34: 729-740.
         March 2001
                                                3B-29
DRAFT-DO NOT QUOTE OR CITE

-------

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
               4.  ENVIRONMENTAL EFFECTS OF
                       PARTICULATE MATTER
4.1  INTRODUCTION
     Several later chapters (Chapters 5 through 8) of this document assess the latest available
information on determinants of human exposures to particulate matter (PM); dosimetry of
particle deposition, clearance, and retention in human respiratory tract; epidemiologic analyses of
health effects associated with human exposures to ambient PM; and toxicologic evaluations of
pathophysiologic effects of PM and underlying mechanisms of action. The human exposure and
health-related findings assessed in those chapters provide key elements of the scientific bases to
support upcoming decision making regarding potential retention or revision of the primary PM
National Ambient Air Quality Standards (NAAQS). This chapter, in  contrast, assesses
information pertinent to decision making regarding secondary standards aimed at protecting
against welfare effects' of PM. More specifically, this chapter assesses environmental effects of
atmospheric PM, including discussion of PM effects on vegetation and ecosystems, PM effects
on visibility, PM effects on man-made materials, and relationships of ambient PM to global
climate change processes.
4.2  EFFECTS ON VEGETATION AND ECOSYSTEMS
     The Particulate Matter National Ambient Air Quality Standards (PM NAAQS) set in 1971
were specified in terms of total suspended particulates (TSP), which included both fine and
coarse mode particles (the latter ranging up to 25 to 40 /u.m in size). The 1987 revision of the
TSP NAAQS to PM10 standards focused attention on those particles (< 10 //m mean aerometic
diameter) capable of being deposited in lower (thoracic) portions of the human respiratory tract.
The subsequent  1997 PM NAAQS revisions retained the PM10 standards and added fine particle
(PM2 5) standards (both specified in terms of mass concentrations of particles undifferentiated in
terms of their specific chemical composition). The effects of PM on vegetation and ecosystems
as a basis for a secondary standard were not considered as part of the 1997 PM NAAQS
       March 2001
                                         4-1
DRAFT-DO NOT QUOTE OR CITE

-------
 1      revisions.  Vegetation and ecosystem effects of ambient PM evaluated in this chapter are
 2      dependent not so much just on PiVI size-related mass concentration, but rather on exposure of
 3      plants to PM components differentiated by chemical composition as well.
 4           Particulate matter deposition to vegetation is not well understood. Unlike gaseous dry
 5      deposition, neither the solubility of the particles nor the physiological activity of the surface are
 6      likely to be of first order importance in determining deposition velocity (Vd). Factors that
 7      contribute to surface wetness and stickiness may be critical determinants of sticking efficiency.
 8      Available tabulations of deposition velocities are highly variable and suspect. High-elevation
 9      forests receive larger particulate deposition loadings than equivalent lower elevation sites,
10      because of higher wind speeds and enhanced rates of aerosol impaction; orographic effects on
11      rainfall intensity and composition; increased duration of occult deposition; and, in many areas,
12      the dominance of coniferous species with needle-shaped leaves (Lovett, 1984).  Recent evidence
13      indicates that all three modes of deposition, (1) wet, (2) occult, and (3) dry, must be considered in
14      determining inputs to watersheds or ecosystems because each may dominate over specific
15      intervals of time or space.
16           Exposure to a given mass concentration of airborne PM may lead to widely differing
17      phytotoxic responses, depending on the particular mix of deposited particles. The most common
18      and useful subdivision of PM, derived from the typical bimodal distribution of atmospheric
19 •     particles, is into fine and coarse particles (Wilson and Suh, 1997). The smallest particle at or
20      near 1.0 to 2.5 ,um generally is taken as the division between fine and coarse, although this is not
21      an absolute and is subject to some shift (e.g., with changing ambient humidity). However, the
22      typical the rule of thumb, as previously used in the 1996 PM Air Quality Criteria Document or
23      "PM AQCD" (U.S. Environmental Protection Agency, 1996a), is that fine PM nominally falls in
24      the range of 0 to 2.5 yum and coarse-mode PM, 2.5 to 10.0 /u.m.
25           In general, fine-mode PM is secondary in nature, having condensed from the vapor phase or
26      been formed by chemical reaction from gaseous precursors in the atmosphere. These particles
27      exist in a nucleation mode (having a mass median aerodynamic diameter or MMAD of about
28      0.06 /um) and may grow by coagulation of existing particles or by condensation of additional
29      gases onto existing particles into an accumulation mode (about 0.5 fj.ro).  Sulfur and nitrogen
30      oxides (SOX and NOJ, as well as volatile  organic gases, are common precursors for fine PM,
31      often neutralized with ammonium cations as particulate salts. Condensation of volatilized metals
        March 2001
4-2
DRAFT-DO NOT QUOTE OR CITE

-------
  1      and products of incomplete combustion also are common precursors. Reactions of many of these
  2      materials with an oxidizing atmosphere lead to high secondary PM concentrations during
  3      summer months in many parts of the United States.
  4           In general, coarse-mode particles are primary in nature, having been produced and emitted
  5      from a point or area source as a fully formed particle. They range in size from ca. 2.5 to 100 //m.
  6      This material is created by abrasion and subsequent suspension by wind or mechanical means.
  7      Suspended geologic material contains the chemical and, potentially, the biological signature of
  8      the soil from which it derives (dominated by iron, silica, aluminum, and calcium). Additional
  9      anthropogenically derived coarse-mode PM derives from fly ash, automobile tires and brake
10      linings, and industrial effluent associated with crushing and grinding operations.  Coarse-mode
11      particles also include biogenically derived organic materials (e.g., fragments of plants and
12      insects, pollen, fungal spores, bacteria and viruses included in marine aerosols).
13           Atmospheric deposition of particles to ecosystems takes place via both wet and dry
14      processes through three major routes: (1) precipitation scavenging in which particles are
15      deposited in rain and snow; (2) fog, cloud-water, and mist interception; and (3) dry deposition,
16      a much slower, yet more continuous removal to surfaces (Hicks, 1986).
17           Precipitation scavenging includes rainout involving within-cloud nucleation phenomena
18      and washout involving below-cloud scavenging by impaction. Total inputs from wet deposition
19      to vegetative canopies can be significant (Table 4-1), although not all wet deposition involves
20      particle scavenging because gaseous pollutants also dissolve during precipitation.
21           Wet deposition is not affected by surface properties as much as is dry or occult deposition.
22      However, forested hillsides may receive much (four- to sixfold) greater precipitation than short
23      vegetation in nearby valleys because of a variety of orographic effects (Unsworth and Wilshaw,
24      1989). Additionally, closer aerodynamic coupling to the atmosphere of the tall forest canopy
25      than of the shorter canopies in the valleys leads to more rapid foliar drying, reduced residence
26      time of solubilized particulate materials available for foliar uptake, and, consequently, more rapid
27      and more extreme concentration of such materials on the cuticular surface. The results of direct
28      physical effects on leaves are not known.
29           Most of wet deposited particulate material passes through the plant canopy to the soil by
30      throughfall and stemflow, causing soil-mediated ecosystem-level responses. Rainfall also
31      removes much of the dry-deposited PM resident on foliar surfaces, reducing direct foliar effects
        March 2001
4-3
DRAFT-DO NOT QUOTE OR CITE

-------
               TABLE 4-1. RELATIVE IMPORTANCE OF WET, DRY, PARTICIPATE,
                       AND TOTAL DEPOSITION TO THREE FOREST SITES8
Deposition
Total Nitrogen1"
Site
Duke Forest
Gary Forest
Austin Forest
Wet
(%)
75
71
71
Dry
(%)
25
20
29
Particle
(%)
0.11
0.94
0.58
Total
(kg ha'1)
9.87
5.80
6.57
Wet
(%)
64
76
83
Total Sulfur*
Dry
(%)
33
20
13
Particle
(%)
2.7
4.2
4.3
Total
(kg ha1)
17.20
7.60
7.79
        "Data from Allen et al. (1994). Sampling was by triple filter pack, so that fine-mode particles could be sampled
        preferentially. An average particle deposition velocity of 0.9 cm s'1 was derived, as in Hicks et al. (1987).
        bWet nitrogen consists of NO3' and NH4+, dry nitrogen consists of vapor phase HNO3 and NO2, and particulate
        nitrogen consists of NO3".
        "Wet sulfur consists of SO4", dry sulfur consists of vapor phase SO2, and particulate sulfur consists of pSO4~.
 1      (Lovett and Lindberg, 1984). This washing effect, combined with differential foliar uptake and
 2      foliar leaching (both of which depend on the physiological status of the vegetation), alters the
 3      composition of rainwater that reaches the soil.  Dry deposition onto foliage and subsequent wet
 4      removal by runoff enhances soil-mediated effects of particulate deposition, both by enhancing
 5      total dry deposition relative to unvegetated surfaces nearby and by accelerating passage of
 6      deposited particles to the soil. The most significant effects of wet deposition occur through soil-
 7      mediated processes involving biogeochemcial cycling of major and minor nutrients and trace
 8      elements.
 9           Dry deposition is more effective for coarse particles  of natural origin and elements such as
10      iron and manganese, whereas wet deposition generally is more effective for fine PM of
11      atmospheric origin and elements such as cadmium, chromium, lead, nickel, and vanadium
12      (Smith, 1990a).  The actual importance of wet versus dry deposition, however, is highly variable,
13      depending on ecosystem type, location, and elevation.  For the Walker Branch Watershed, a
14      deciduous forest in rural eastern Tennessee, dry deposition constituted a major fraction of total
15      annual atmospheric input of cadmium and zinc (=20%), lead (=55%), and manganese (=90%),
16      but wet deposition rates for single precipitation events exceeded dry deposition rates by one to
17      four orders of magnitude (Lindberg and Harriss, 1981).  Miller et al. (1993) emphasized that
18      immersion of high-elevation forests in cloudwater for 10% or more of the year can enhance
19      significantly overall efficiency of transfer of atmospheric particles and gases to a forest canopy.
        March 2001                                4-4         DRAFT-DO NOT QUOTE OR CITE

-------
  1           Dry deposition of particles occurs to all vegetational surfaces exposed to the atmosphere
  2      (U.S. Environmental Protection Agency, 1982).  The range of particle sizes, the diversity of
  3      canopy surfaces, and the variety of chemical constituents in airborne PM have slowed progress in
  4      both prediction and measurement of dry particulate deposition. Wet deposition generally is
  5      confounded by fewer factors and has been easier to quantify (Chapter 2).
  6           Emphasis in this and the next section is placed on discussion of PM effects on individual
  7      plants in natural habitats and terrestrial ecosystems.  Except for the deposition of nitrogen and
  8      sulfur-containing compounds and their effects exerted via acidic precipitation, information
  9      concerning the effects of deposition of other specific  substances as PM on crops is not readily
10      available. The U.S. National Acid Precipitation Assessment Program (NAPAP) Biennial Report
11      to Congress: An Integrated Assessment presents an extensive overall discussion of the effects of
12      acidic deposition (National Science and Technology Council, 1998). The effects of gaseous
13      sulfur oxides and nitrogen oxides on crops are discussed in detail in EPA criteria documents for
14      those substances (U.S. Environmental Protection Agency, 1982, 1993).  A detailed discussion of
15      the ecological effects of acidic precipitation and nitrate deposition on aquatic ecosystems also
16      can be found in the EPA Nitrogen Oxides Air Quality Criteria Document (U.S. Environmental
17      Protection Agency, 1993). Neither nitrate or sulfate deposition on crops is discussed in this
18      chapter, as they are added frequently in fertilizers. Also, the effects of lead on crops, vegetation,
19      and ecosystems are discussed in the EPA document, Air Quality Criteria for Lead (U.S.
20      Environmental Protection Agency, 1986).
21           The effects of deposited PM may be direct or indirect.  Indirect effects are chiefly
22      nutritional responses mediated through the soil and result from the effects of PM components on
23      soil processes.  In the following sections, the direct effects on individual plants are discussed
24      first, followed by effects  on plant species and their interactions in ecosystems.

25      4.2.1  Direct Effects  of Particulate Matter on Individual Plant Species
26           Particulate matter in the atmosphere may affect vegetation directly following deposition on
27      foliar surfaces, indirectly by changing the soil chemistry, or through changes in the amount of
28      radiation reaching the Earth's surface through PM-induced climate change processes. Indirect
29      impacts, however, are usually the most significant because they can alter nutrient cycling and
        March 2001
4-5
DRAFT-DO NOT QUOTE OR CITE

-------
 1      inhibit plant nutrient uptake. The possible direct responses to PM deposition are considered in
 2      this section, and the indirect responses in the sections on ecosystems.
 3           Particles transferred from the atmosphere to foliar surfaces may reside on the leaf, twig or
 4      bark surface for extended periods; be taken up through the leaf surface; or be removed from the
 5      plant via resuspension to the atmosphere, washing by rainfall, or litter-fall with subsequent
 6      transfer to the soil. Any PM deposited on above-ground plant parts may exert physical or
 7      chemical impacts. The effects of "inert" PM are mainly physical, whereas those of toxic particles
 8      are both chemical and physical.  The chemical effects of dust deposited on plant surfaces or soil
 9      are more likely to be associated with their chemistry than simply with the mass of deposited
10      particles and may be more important than any physical effects (Farmer, 1993).
11           Studies  of the direct effects of chemical additions to foliage in particulate deposition have
12      found little or no effects of PM on foliar processes unless exposure levels were significantly
13      higher than typically would be experienced in the ambient environment. Interpretation of the
14      effects of atmospheric chemical deposition at the level of individual plants and ecosystems is
15      difficult because of the complex interactions that exist among biological, physicochemical, and
16      climatic factors. The majority of the easily identifiable direct and indirect effects, other than
17      climate, occur in severely polluted areas around heavily industrialized point sources, such as
18      limestone quarries, cement kilns, and smelting facilities for iron, lead, or various other metals.
19"     The diverse chemical nature and size characteristics of ambient airborne particles and the lack of
20      any clear distinction between effects attributed to phytotoxic particles and to other forms of air
21      pollutants confound the direct effects of PM on foliar surfaces. Most documented toxic effects of
22      particles on vegetation reflect their acidity, trace metal content, nutrient content, surfactant
23      properties, or salinity. These materials typically elicit similar biological effects, whether
24      deposited as coarse or fine particles, in wet, dry, or occult form, and, frequently, whether
25      deposited to foliage or to the soil.  Studies of direct effects of particles on vegetation have not yet
26      advanced to the stage of reproducible exposure experiments. Experimental difficulties in
27      application of ambient particles to vegetation have been discussed by Olszyk et al. (1989).
28      4.2.1.1 Effects of Coarse Particles
29           Coarse-mode particles, ranging in size from 2.5 to 100 fjm, are chemically diverse, are
30      dominated by local sources, and are typically deposited near the source because of their
        March 2001
4-6
DRAFT-DO NOT QUOTE OR CITE

-------
 1     sedimentation velocities. Airborne coarse particles are derived from road, cement kiln, and
 2     foundry dust; fly ash; tire particles and brake linings; soot and cooking oil droplets; biogenic
 3     materials (e.g., plant pollen, fragments of plants, fungal spores, bacteria and viruses) and sea salt.
 4     Most coarse particles in rural and some urban areas are composed of silicon, aluminum, calcium,
 5     and iron, suggesting that their main source is fugitive dust from disturbed land, roadways,
 6     agricultural tillage, or construction.  Rapid sedimentation of these particles tends to restrict their
 7     direct effects on vegetatioolargely to roadsides and forest edges.       L

 8           Physical Effects—Radiation.  Dust can have both a physical and chemical impact.
 9     Deposition of inert PM on above-ground plant organs may result in an increase in radiation
10     received, in leaf temperature and blockage of stomata. Increased leaf temperature, heat stress,
11     reduced net photosynthesis, and leaf chlorosis, necrosis, and abscission were reported by
12     Guderian (1986).  Road dust decreased the leaf temperature on Rhododendron catavshiense by
13     ca. 4 °C (Eller, 1977), whereas foundry dust caused an 8.7 °C increase in leaf temperature of
14     black poplar (Populus nigrd) (Guderian, 1986) under the conditions of the experiment.
15     Broad-leaved plants exhibited greater temperature increases because of particle loading than did
16     the needle-like leaves of conifers. Deciduous (broad) leaves exhibited larger temperature
17     increases because of particle loading than did conifer (needle) leaves,  a function of poorer
18     coupling to the atmosphere. Inert road dust caused a three- to fourfold increase in the absorption
19     coefficient of leaves ofHedera helix (Eller, 1977; Guderian, 1986) for near infrared radiation
20     (NIR; 750 to 1350 nm). Little change occurred in absorption for photosynthetically active
21     radiation (PAR; 400 to 700 nm).  The increase in NIR absorption was equally at the expense of
22     reflectance and transmission in these wavelengths. The net energy budget increased by ca. 30%
23     in the dust-affected leaves. Deposition of coarse particles increased leaf temperature and
24     contributed to heat stress, reduced net photosynthesis, and caused leaf chlorosis, necrosis, and
25     abscission (Dassler et al., 1972; Parish, 1910; Guderian, 1986; Spinka, 1971).
26           Starch storage in dust-affected leaves increased with dust loading under high (possibly
27     excessive) radiation, but decreased following dust deposition when radiation was limiting. These
28     modifications of the radiation environment had a large impact on single-leaf utilization of light.
29      The boundary layer properties, determined by leaf morphology and environmental conditions,
        March 2001
4-7
DRAFT-DO NOT QUOTE OR CITE

-------
  1     strongly influenced the direct effects of particle deposition on radiation heating (Eller, 1977;
  2     Guderian, 1986) and on gas exchange as well.
  3          Brandt and Rhoades (1973) attributed the reduction in growth of trees because of crust
  4     formation from limestone dust on the leaves. Crust formation reduced photosynthesis and
  5     formation of carbohydrate needed for normal growth, induced premature leaf-fall, destruction of
  6     leaf tissues, inhibited growth of new tissue,* and reduced storage. Dust may affect
  7     photosynthesis, respiration, and transpiration, and it may allow penetration of phytotoxic gaseous
  8     pollutants, thereby causing visible injury symptoms and decreased productivity.  Permeability of
  9     leaves to ammonia increased with increasing dust concentrations and decreasing particle size
 10     (Farmer, 1993).
 11          Dust also has been reported to physically block stomata (Krajickova and Mejstfik, 1984).
 12     Stomatal clogging by particulate matter from automobiles, stone quarries, and cement plants was
 13     also studied by Abdullah and Iqbal (1991).  The percentage of clogging was low in young leaves
 14     when compared with old and mature leaves and the amount of clogging varied with species and
 15     locality. The maximum clogging of stomata observed was about 25%.  The authors cited no
 16     evidence that stomatal clogging inhibited plant functioning. The heaviest deposit of dust is
 17     usually on the upper surface of broad-leaved plants, however, whereas the majority of the
 18     stomata are on the lower surface where stomatal clogging would be less likely.
 19          Chemical Effects.  The chemical composition of PM is usually the key phytotoxic factor
20     leading to plant injury. Cement-kiln dust on hydration liberates calcium hydroxide, which can
21      penetrate the epidermis and enter the mesophyll, and, in some cases, the leaf surface alkalinity
22     may reach to pH 12. Lipid hydrolysis coagulation of the protein compounds and ultimately
23      plasmolysis of the leaf tissue result in reduction in growth and quality of plants (Guderian, 1986).
24     In experimental studies, application of cement kiln dust of known composition for 2 to 3 days
25      yielded dose-response curves between net photosynthetic inhibition or foliar injury and dust
26      application rate (Darley, 1966). Lerman and Darley (1975) determined that leaves must be
27      misted regularly to produce large effects. ^Alkalinity was probably the essential phytotoxic
28      property of the applied dusts.
29           Particulate matter in the form of sea salt enters the atmosphere from oceans following
30      mixing of air into the water and subsequent bursting of bubbles at the surface. This process can
        March 2001
4-8
DRAFT-DO NOT QUOTE OR CITE

-------
 1      be a significant source of sulfate, sodium chloride, and trace elements in the atmosphere over
 2      coastal vegetation, resulting in the formation of the maritime forest, a specialized ecosystem.
 3      Sea-salt particles can serve as nuclei for the adsorption and subsequent reaction of other gaseous
 4      and particulate pollutants. Both nitrate and sulfate from the atmosphere have been found to be
 5      associated with coarse and fine sea-salt particles (Wu and Okada, 1994). Direct effects on
 6      vegetation reflect these inputs, as well as classical salt injury caused by the sodium and chloride
 7      that constitute the bulk of these particles. Salt pruning is a common phenomenon near the ocean
 8      (i.e., salt spray kills the buds on the windward side of trees and shrubs).

 9      4.2.1.2 Effects of Fine Particles
10           Fine PM is  generally secondary in nature, having condensed from the vapor phase or been
11      formed by chemical reaction from gaseous precursors in the atmosphere and is generally smaller
12      than 1 to 2.5 //m. Nitrogen and sulfur oxides, volatile organic gases, condensation of volatilized
13      metals, and products of incomplete combustion are common precursors for fine PM. Reactions
14      of many of these  materials with an oxidizing atmosphere contribute to high secondary PM
15      concentrations during summer months in many U.S. areas.  The conclusion reached in the 1982
16      PM AQCD  (U.S. Environmental Protection Agency, 1982) that sufficient data were not available
17      for adequate quantification of dose-response functions for direct effects of fine aerosols on
18      vegetation continues to be true today.  Only a few studies have been completed on the direct
19      effects of acid aerosols (U. S. Environmental Protection Agency, 1982). The major effects are
20      indirect and occur through the soil (Section 4.3).
21           Nitrogen.  Nitrate is observed in both fine and coarse particles. Nitrates from atmospheric
22      deposition represent a substantial fraction of total nitrogen inputs to southeastern forests (e.g.,
23      Lovett and Lindberg, 1986). However, much of this is contributed by gaseous nitric acid vapor,
24      and a considerable amount of the particulate nitrate is taken up indirectly, through the soil.
25      Garner et al. (1989) estimated deposition of nitrogen to forested landscapes in eastern North
26      America at 10 to 55 kg/ha/year for nitrate and 2 to 10  kg/ha/year for ammonium. About half of
27      these values were ascribed to dry deposition.
28           Atmospheric additions of particulate nitrogen in excess of vegetation needs are lost from
29      the system, mostly as leachate from the soil as nitrate.  Managed agricultural ecosystems may be
        March 2001
4-9
DRAFT-DO NOT QUOTE OR CITE

-------
         1      able to utilize deposited particulate nitrogen more efficiently than native ecosystems, although
         2      many cultivated systems also lose considerable nitrogen as nitrate in runoff, deep drainage, or tail
         3      water. It has proven difficult to quantify direct foliar fertilization by uptake of nitrogen from
         4      ambient particles.
         5           There is no doubt that foliar uptake of nitrate can occur, as clearly shown by the efficacy of
         6      foliar fertilization in horticultural systems.  Potassium nitrate was taken up by leave's of
         7      deciduous fruit trees (Weinbaum and Neumann, 1977) and resulted in increased foliar nitrogen
         8      concentrations. Not all forms of nitrogen are absorbed equally, nor are all equally benign.
         9      Following foliar application of 2600 ppm of nitrogen as Ca(NO3)2, (NH4)2SO4, or (NH2)2CO to
       10      apple canopies (Rodney, 1952; Norton and Childers, 1954), leaf nitrogen levels were observed to
       11      increase to similar levels, but calcium nitrate and ammonium sulfate caused visible foliar
       12      damage, whereas urea did not. Urea is generally the recommended horticultural foliar fertilizer.
       13           The mechanism of uptake of foliarly deposited nitrate is not well established. Nitrate
       14      reductase is generally a root-localized enzyme. It is generally not present in leaves, but is
       15      inducible there. This typically occurs when the soil is heavily enriched in NO3".  As the root
       16      complement of nitrate reductase becomes overloaded, unreduced nitrate reaches the leaves
       17      through the transpiration stream.  Nitrate metabolism has been demonstrated in  leaf tissue
       18      (Weinbaum and Neumann,  1977) following foliar fertilization. Residual nitrate reductase
       19      activity in leaves may be adequate to assimilate typical rates of particulate nitrate deposition.
       20      Uptake of nitrate may be facilitated by codeposited sulfur (Karmoker et al., 1991; Turner and
       21      Lambert, 1980).
       22           Nitrate reductase is feedback-inhibited by its reaction product, NH4+.  The common
       23      atmospheric aerosol, NH4NO3, therefore may be metabolized in two distinct biochemical steps,
       24      first the ammonium (probably leaving nitric acid)  and then the nitrate. Volatilization losses of
       25      nitric acid during this process, if they occur, have not been characterized.
       26           Direct foliar effects of particulate nitrogen have not been documented. Application of a
       27      variety of fine nitrogenous aerosol particles (0.25 ywm) ranging from 109 to 244  /ug/m3 nitrogen,
       28      with or without 637 //g/m3 sulfur, caused no consistent short-term (2- to 5-h) effect on gas
       29      exchange in oak, maize, or soybean leaves (Martin et al., 1992).
       30           Although no evidence exists for direct transfer of nutrient particulate aerosols into foliage,
       31      a few studies give insights into the potential for ammonium and nitrate transfer into leaves.
               March 2001
4-10
DRAFT-DO NOT QUOTE OR CITE
_

-------
 1     Fluxes of both NO3" and NH/, measured in wet deposition and in throughfall plus stemflow in
 2     forests, commonly indicate higher fluxes of nitrogen above the canopy (Parker, 1983; Lindberg
 3     et al., 1987; Sievering et al., 1996), indicating net foliar uptake. Lovett and Lindberg (1993)
 4     reported a linear relationship between inorganic nitrogen fluxes in deposition and throughfall,
 5     suggesting that uptake may be considered passive to some extent.
 6          Garten and Hanson (1990) studied the movement of I5N-labeled nitrate and ammonium
 7     across the cuticles of red maple (Acer rubrum) and white oak (Quercus alba) leaves when
 8     applied as an artificial rain mixture. Brumme et al. (1992), Bowden et al. (1989), and Vose and
 9     Swank (1990) have published similar data for conifers. These studies show the potential for
10     nitrate and ammonium to move into leaves, where it may contribute to normal physiological
11     processes (e.g., amino acid production; Wellburn,  1990). Garten (1988) showed that internally
12     translocated 35S was not leached readily from tree leaves of yellow poplar (Liriodendron
13     tulipifera) and red maple (Acer rubrum), suggesting that SO42" would not be as mobile as the
14     nitrogen-containing ions discussed by Garten and Hanson (1990). Further, when the foliar
15     extraction method is used it is not possible to distinguish sources of chemical deposited as gases
16     or particles (e.g., nitric acid [HNO3], nitrogen dioxides [NO2], nitrate [NO3"], or sources of
17     ammonium deposited as ammonia [NH3] or ammonium ion [NH4+]) (Garten and Hanson, 1990).
18          Particle deposition contributes only a portion of the total atmospheric nitrogen  deposition
19     reaching vegetation but, when combined with gaseous and precipitation-derived sources, total
20     nitrogen deposition to ecosystems has been identified as a possible causal factor leading to
21     changes in natural ecosystems (See Section 4.3).
22           Sulfur.  Anthropogenic sulfur emissions are >90% as SO2. Most of the remaining emission
23      of sulfur is directly as sulfate (U.S. Environmental Protection Agency, 1996a).  Sulfur dioxide is
24      hydrophilic and is rapidly hydrated and oxidized to sulfite and bisulfite and then to sulfate, which
25      is approximately 30-fold less phytotoxic. The ratio of sulfate/SO2 increases with aging of the air
26      mass and, therefore, with distance from the source. Sulfate is sufficiently hygroscopic that, in
27      humid air, it may exist significantly in the coarse particulate fraction. As dilution of both SO2
28      and particulate SO42" occurs with distance from the source, it is unusual for damaging levels of
29      particulate sulfate to be deposited.  Gas to particle conversion in this case is of benefit to
30      vegetation.
        March 2001
4-11
DRAFT-DO NOT QUOTE OR CITE

-------
  1           Sulfur is an essential plant nutrient.  Low dosages of sulfur serve as a fertilizer, particularly
  2     for plants growing in sulfur-deficient soil (Hogan et al., 1998). Current levels of sulfate
  3     deposition reportedly exceed the capacity of most vegetative canopies to immobilize the sulfur
  4     (Johnson, 1984). Nitrogen uptake in forests may be regulated loosely by sulfur availability, but
  5     sulfate additions in excess of needs do not typically lead to injury (Turner and Lambert, 1980).
  6           There are few field demonstrations of foliar sulfate uptake (Krupa and Legge, 1986, 1998).
  7     Sulfate in throughfall is often enriched above levels in precipitation.  The relative importance of
  8     foliar leachate and prior dry-deposited sulfate particles remains difficult to quantify (Cape et al.,
  9     1992). Leaching rates are not constant and may respond to levels of other pollutants, including
 10     acids.  Uptake and foliar retention of gaseous and particulate sulfur are confounded by variable
 11     rates of translocation and accessibility of deposited materials to removal and quantification by
 12     leaf washing. Following soil enrichment with 35SO42" in a Scots pine forest, the apparent
 13     contribution of leachate to throughfall was only a few percent, following an initial burst of over
 14     90% because of extreme disequilibrium in labeling of tissue sulfate pools (Cape et al., 1992).
 15          Olszyk et al. (1989) provide information on the impacts of multiple pollutant exposures
 16     including particles (NO3', 142 ^g/m3; NH/, 101 //g/m3; SO42', 107 //g/m3). They found that only
 17     gaseous pollutants produced direct (harmful) effects on vegetation for the concentrations
 18     documented, but the authors hypothesized that long-term accumulation of the nitrogen and sulfur
 19     compounds contributed from particle deposition might have effects on plant nutrition over long
20     periods of time. Martin et al. (1992) exposed oak (Quercus macrocarpd), soybean (Glycine
21     max), and maize (Zea mays) plants to acute exposures (2 to 5 h) of aerosols (0.25 //m) containing
22     only nitrate (109 /ug/m3), ammonium and nitrate (244 and 199 //g/m3, respectively), or
23     ammonium and sulfate (179 and 637 //g/m3, respectively). They found that these exposures,
24     which exceeded the range of naturally occurring aerosol concentrations, had little effect on foliar
25     photosynthesis and conductance. Martin et al. (1992) concluded that future investigations should
26     focus on the effects of particles on physiological characteristics of plants following chronic
27     exposures.
28           Acidic Deposition. The effects of acidic deposition have been accorded wide attention in
29      the media and elsewhere (Altshuller and Linthurst, 1984; Hogan et al., 1998). Probably the most
30      extensive assessment of acidic deposition processes and effects is the NAPAP Biennial Report to
        March 2001
4-12
DRAFT-DO NOT QUOTE OR CITE

-------
 1      Congress:  An Integrated Assessment (National Science and Technology Council, 1998).
 2      Concern regarding the effects of acidic deposition on crops and forest trees has resulted in
 3      extensive monitoring and research.  Exposures to acidic rain or clouds can be divided into
 4      "acute" exposures to higher ionic concentrations (several /^mol/L), and "chronic" long-term
 5      repeated exposures to lower concentrations (Cape, 1993). Pollutant concentrations in rainfall
 6      have been shown to have little capacity for producing direct effects on vegetation (Altshuller and
 7      Linthurst,  1984; Hogan et al., 1998); however, fog and clouds, which may contain solute
 8      concentrations up to 10 times those found in rain, have the potential for direct effects.  More than
 9      80% of the ionic composition of most cloud water is made up of four major pollutant ions: H+,
10      NH4+, NO3", and SO42". Ratios of hydrogen to ammonium and sulfate to nitrate vary from site to
11      site with all four ions usually present in approximately equal concentrations.  Available data from
12      plant effect studies suggest that hydrogen and sulfate ions are more likely to cause injury than
13      ions containing nitrogen (Cape, 1993).
14           The possible direct effects of acidic precipitation on forest trees have been evaluated by
15      experiments on seedlings and young trees. The size of mature trees makes experimental
16      exposure difficult, therefore necessitating extrapolations from experiments on seedlings and
17      saplings; however, such extrapolations must be used with caution (Cape, 1993).  Both conifers
18      and deciduous species have shown significant effects on leaf surface structures after exposure to
19      simulated  acid rain or acid mist at pH 3.5. Some species have shown subtle effects at pH 4 and
20      above.  Visible lesions have been observed on many species at pH 3 and on sensitive species at
21      pH 3.5 (Cape, 1993). The relative sensitivities of forest vegetation to acidic precipitation based
22      on macroscopic injury have been ranked as follows: herbaceous dicots > woody dicots >
23      monocots  > conifers (Percy 1991).
24           Huttunen (1994) described the direct effects of acid rain or acidic mist on epicuticular
25      waxes whose ultrastructure is affected by plant genotype and phenotype. The effects of air
26      pollutants on  epicuticular waxes of conifers have received greater study than the waxes of other
27      species. Leafage and shorter life span of broad-leaved trees make them less indicative of the
28      effects of acid precipitation. Many experimental studies indicate that epicuticular waxes that
29      function to prevent water loss from plant leaves can be destroyed by acid rain in a few weeks
30      (Huttunen, 1994). This function is crucial in conifers because of their longevity and evergreen
31      foliage.  Microscopic observations of epicuticular wax structures have, for a long time, suggested
        March 2001
4-13
DRAFT-DO NOT QUOTE OR CITE

-------
  1     links between acidic deposition and aging. In Norway spruce (Picea abies), acid rain causes not
  2     only the aging of needles (which in northern conditions normally last from 11 to 14 years) to be
  3     shortened but also accelerates the erosion rate of the waxes as the needles age.
  4          The effects of acidic precipitation and fog on red spruce (Picea rubens) have been studied
  5     extensively (Schier and Jensen, 1992). Visible foliar injury of the needles in the form of a
  6     reddish-brown discoloration has been observed on red spruce seedlings experimentally exposed
  7     to acidic mist, but this visible symptom has not been observed in the field. .Ultrastructural
  8     changes in the epicuticular wax were observed both experimentally and on spruce growing at
  9     high elevations.  Laboratory studies indicate that visible injury usually does not occur unless the
 10     pH is 3 or less (Schier and Jensen, 1992). Cape (1993) reported that, when compared with other
 11     species, red spruce seedlings appeared to be more sensitive to acid mist.  Huttunen (1994)
 12     concluded that his studies of conifers and review of the literature suggest that acidic precipitation
 13     causes direct injury to tree foliage and, also, indirect effects through the soil.  The indirect effects
 14     of acidic precipitation are discussed in Section 4.3.
 15          Based on his review of the many studies involving field and controlled laboratory
 16     experiments on crops in the literature, Cape (1993) drew a number of conclusions concerning the
 17     direct effects of acidic precipitation on crops:
 18          • foliar injury and growth reduction occurs below pH 3;
 19          • allocation of photosynthate is altered, with increased shoot to root ratios;
20          • expanded and recently expanded leaves are most susceptible, and injury occurs first to
21            epidermal cells;
22          •  leaf surface characteristics such as wettability, buffering capacity, and transport of
23            material across the leaf surface contribute to susceptibility and differ among species;
24          •  data obtained from experiments in greenhouses or controlled environmental chambers
25            cannot be used to predict effects on plants grown in the field;
26          •  quantitative data from experimental exposures cannot be extrapolated to field exposures
27            because of differences and fluctuations in concentrations, durations, and frequency of
28            exposure;
29          •  there are large differences in response within species;
30          •  timing of exposure in relation to phenology is of utmost importance;
31           •  plants may be able to recover from or adapt to injurious exposures; and
        March 2001
4-14
DRAFT-DO NOT QUOTE OR CITE

-------
  1           • sequential exposure to acidic precipitation and gaseous pollutants is unlikely to be more
 2            injurious than exposure to individual pollutants.
 3           Studies by Chevone et al. (1986), Krupa and Legge (1986), and Blaschke (1990) differ with
 4      the last conclusion of Cape listed above. Their studies indicate that interactions between acidic
 5      deposition and gaseous pollutants do occur. Acidity affects plant responses to both O3 and SO2.
 6      Chevone et al. (1986) observed increased visible injury on soybean and pinto bean when acid
 7      aerosol exposure preceded O3 exposure, whereas linear decreases in dry root weight of yellow
 8      poplar occurred as acidity increased with simultaneous exposures to O3 and simulated acid rain.
 9      Kfupa and Legge (1986) also noted increased visible injury to pinto bean when aerosol exposure
10      preceded O3 exposure. In none of the studies cited above did acid rain per se'produce significant
11      growth changes. Blaschke (1990) observed a decrease in ectomycorrhizal frequency and short
12      root distribution caused by acid rain exposure in combination with either SO2 or O3.

13           Trace Elements. All but 10 of the 90 elements that comprise the inorganic fraction of the
14      soil occur at concentrations of less than 0.1% (1000 |ig/g) and are termed "trace" elements.
15      Trace elements with a density greater than 6 g.cm"3, referred to as "heavy metals", are of
16      particular interest because of their potential toxicity for plant and animals. Although some trace
17      metals are essential for vegetative growth or animal health, they are all toxic in large quantities.
18      Combustion processes produce metal chlorides that tend to be volatile and metal oxides that tend
19      to be nonvolatile in the vapor phase (McGowan et al., 1993). Most trace elements exist in the
20      atmosphere in particulate form as metal oxides (Ormrod,  1984). Aerosols containing trace
21      elements derive predominantly from industrial activities (Ormrod, 1984).  Generally, only
22      cadmium, chromium, nickel, and mercury are released from stacks in the vapor phase (McGowan
23      et al., 1993). Concentrations of heavy metals in incinerator fly ash increase with decreasing
24      particle size.
25           Vegetational surfaces, especially the foliage, present a major reaction and filtration surface
26      to the atmosphere and act to accumulate particles deposited via wet and dry processes described
27      in Chapter 2 (Tong, 1991; Youngs et al., 1993).  Particles deposited on foliar surfaces may be
28      taken up through the leaf surface. The greatest particle loading is usually on the adaxial (upper)
29      leaf surface where particles accumulate in the mid-vein, center portion of the leaves.  The
        March 2001
4-15
DRAFT-DO NOT QUOTE OR CITE

-------
 1      mycelium of fungi becomes particularly abundant on leaf surfaces as the growing season
 2      progresses and is in intimate association with deposited particles (Smith, 1990b).
 3           Investigations of trace elements present along roadsides and in industrial and urban
 4      environments indicate that impressive burdens of particulate heavy metals can accumulate on
 5      vegetative surfaces. Foliar uptake of available metals could result in metabolic impact in above-
 6      ground tissues. Only a few metals, however, have been documented to cause direct phytotoxicity
 7      in field conditions.  Copper, zinc, and nickel toxicities have been observed most frequently. Low
 8      solubility, however, limits foliar uptake and direct heavy metal toxicity. A trace metal must be
 9      brought into solution before it can enter into leaves or bark of vascular plants, m those instances
10      when trace metals are absorbed, they are frequently bound in leaf tissue and are lost when the leaf
11      drops off (Hughes, 1981). Trace metals in mixtures may interact to cause a different plant
12      response when compared with a single element; however, there has been little research on this
13      aspect (Ormrod, 1984).  In experiments using chambers, Marchwinska and Kucharski (1987)
14      studied the effects of SO2 alone and hi combination with PM components (Pb, Cd, Zn, Fe, Cu,
15      and Mn) obtained from a zinc smelter bag filter.  The combined effects of SO2 and PM further
16      increased the reduction in yield of beans caused by SO2, whereas the combination, though
17      severely injuring the foliage, produced little effect on carrots and parsley roots, except after
18      long-term exposures (when there was a decrease in root weight).
19           Trace metal toxicity of lichens has been demonstrated in relatively few cases. Nash (1975)
20      documented zinc toxicity in the vicinity of a zinc smelter near Palmerton, PA. Lichen species
21      richness and abundance were reduced by approximately 90% in lichen communities at Lehigh
22      Water Gap near the zinc smelter when compared with those at Delaware Water Gap.  Zinc,
23      cadmium, and sulfur dioxide were present in concentrations toxic to some species near the
24      smelter; however, toxic zinc concentrations extended beyond the detectable limits of sulfur
25      dioxide (Nash, 1975). Experimental data suggest that lichen tolerance to Zn and Cd falls
26      between 200 and 600 ppm (Nash, 1975).
27           Though there has been no direct evidence of a physiological association between tree injury
28      and exposure to metals, heavy metals have been implicated because their deposition pattern is
29      correlated with forest decline. The role of heavy metals has been indicated by phytochelatin
30      measurements. Phytochelatins are intracellular metal-binding peptides that act as specific
31      indicators of metal stress. Because they are produced by plants as a response to sublethal
        March 2001
4-16
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
 concentrations of heavy metals, they can be used to indicate that heavy metals play a role in
 forest decline (Gawel et al., 1996). Concentrations of heavy metals increased with altitude, as
 did forest decline, and increased concentrations across the region showing increased levels of
 forest injury, as well.
      Phytochelatin concentrations were measured in red spruce and balsam fir (Abies balsamea)
 needles throughout the 1993 growing season at 1000 m on Whiteface Mountain in New York.
 Mean foliar concentrations in red spruce were consistently higher than in balsam fir from June
 until August, with the greatest and most significant difference occurring at the peak of the
 growing season in mid-July. In July, the phytochelatin concentrations were significantly higher
 than at any other time measured. Balsam fir did exhibit this peak, but maintained a consistently
 low level throughout the season.  Both the number of dead red spruce trees and phytochelatin
 concentrations increased sharply with elevation (Gawel et al., 1996).  The relationship between
 heavy metals and the decline of forests in northeastern United States was further tested by
 sampling red spruce stands showing varying degrees of decline at 1000 m on nine mountains
 spanning New Hampshire, Vermont, and New York. The collected samples indicated a
 systematic and significant increase in phytochelatin concentrations associated with the extent of
 tree injury. The highest phytochelatin concentrations were measured during 1994 from sites
 most severely affected by forest decline in the Green Mountains, VT, and the Adirondack
 Mountains, NY. These data strongly imply that metal stress is a cause of tree injury and,
 therefore, contributes to forest decline in the northeastern United States (Gawel et al., 1996).
     One potential direct impact of heavy metals is on the activity of microorganisms and
 arthropods resident on and in the  leaf surface ecosystem. The fungi and bacteria living on and  in
 the surfaces of leaves play an important role in the microbial succession that prepares leaves for
 decay and litter decomposition after their fall (U.S. Environmental Protection Agency, 1996b).
     Numerous fungi were consistently isolated from foliar surfaces, at various crown positions,
 from London plane trees growing in roadside environments in New Haven, CT.  Those existing
primarily as parasites included Aureobasidium pullulans, Chaetomium sp., Cladosporium sp.,
Epicoccum sp., and Philaphora verrucosa. Those existing primarily as parasites included
 Gnomoniaplatani, Pestalotiposis sp., andPleurophomella sp. The following cations were tested
in vitro for their ability to influence the growth of these fungi: cadmium, copper, manganese,
aluminum, chromium, nickel, iron, lead, sodium, and zinc.  Results indicated variable fungal
       March 2001
                                          4-17
                                                             DRAFT-DO NOT QUOTE OR CITE

-------
 1      response with no correlation between saprophytic or parasitic activity and sensitivity to heavy
 2      metals.  Both linear extension and dry weight data indicated that the saprophytic Chaetomum sp.
 3      was very sensitive to numerous metals. Aureobasidium pullulans, Epicoccum sp., and especially
 4      P.verrucosa, on the other hand, appeared to be much more tolerant. Of the parasites, G. platani
 5      appeared to be more tolerant than Pestalotiopsis sp. and Pleurophomella sp.  Metals exhibiting
 6      the broadest spectrum growth suppression were iron, aluminum, nickel, zinc, manganese, and
 7      lead (Smith and Staskawicz, 1977; Smith, 1990c). These in vitro studies employed soluble
 8      compounds containing heavy metals. Trace metals probably occur naturally on leaf surfaces as
 9      low-solubility oxides, halides, sulfates, sulfides, or phosphates (Clevenger et al., 1991; Koslow
10     et al., 1977). In the event of sufficient solubility and dose, however, changes in microbial
11      community structure on leaf surfaces because of heavy metal accumulation are possible.

12          Organic Compounds. Fine particles in the atmosphere reacting with volatilized chemical
13     compounds are partitioned between the gas and particle phases, depending on the liquid phase
14     vapor pressure at the ambient atmospheric temperature, the surface area of the particles per unit
15     volume of air, the nature of the particles and of the chemical being adsorbed and can be removed
16     by wet and dry deposition (McLachlan, 1996a).  Materials as diverse as DDT, polychlorinated
17     biphenyls (PCBs), and polynuclear aromatic hydrocarbons (PAHs) are being deposited from the
18     atmosphere on rural as well as urban landscapes (Kylin et al., 1994). Motor vehicles emit
19     particles to the atmosphere from several sources in addition to the tailpipe. Rogge et al. (1993)
20     inventoried the organic contaminants associated with fine particles (diameter <2.0 ^m) in road
21     dust, brake lining wear particles, and tire tread debris.  In excess of 100 organic compounds were
22     identified in these samples, including n-alkanols, benzoic acids, benaldehydes, polyalkylene
23     glycol ethers, PAHs, oxy-PAH, steranes, hopanes, natural resins, and other compound classes.
24     A large number of PAHs, ranging from naphthalene (C10H8) to 5- and 6-ring and higher PAHs,
25     their alkyl-substituted analogues, and their oxygen- and nitrogen-containing derivatives are
26     emitted from motor vehicle sources (Seinfeld, 1989).
27           Plants may be used as environmental  monitors to compare the deposition of PAH, POPs, or
28      SOCs between sites (e.g., urban versus rural) (Wagrowski and Kites, 1997; Ockenden et al.,
29      1998; McLachlan, 1999).  Vegetation can be used qualitatively to indicate organic pollutant
30     levels as long as the mechanism of accumulation is considered. The substance may enter the
        March 2001
4-18
DRAFT-DO NOT QUOTE OR CITE

-------
  1      plant via the roots or, as mentioned above, deposited as a particle onto the waxy cuticle of leaves
  2      or uptake thorough the stomata. The pathways are a function of the chemical and physical
  3      properties of the pollutant, such as its lipophilicity, water solubility, vapor pressure (which
  4      controls the vapro-particle partitioning) and Henry's law constant; environmental conditions,
  5      such as ambient temperature and the organic content of the soil; and the plant species, which
  6      controls the surface area and lipids available for accumulation (Simonich and Hites, 1995).
  7      Ockenden et al. (1998) have observed that, for lipophilic POPs, atmospheric transfer to plant has
  8      been the main avenue of accumulation.  Plants can differentially accumulate POPs.  Results have
  9      shown differences between species with higher concentrations in the lichen (Hypogymnia
 10      physiodes) than in pine needles (Pinus sylvestris). Even plants of the same species, because they
 11      have different growth rates and different lipid contents (depending on the habitat in which they
 12      are growing), have different rates of sequestering pollutants.  These facts confound data
 13      interpretations and must be taken into account when considering their use as passive samplers.
 14           Vegetation itself is an important source of hydrocarbon aerosols.  Terpenes, particularly
 15      a-pinene, p-pinene, and limonene released from tree foliage, may react in the atmosphere to form
 16      submicron particles. These naturally generated organic particles contribute significantly to the
 17      blue haze aerosols formed naturally over forested areas (Smith, 1990d).
 18           The low water solubility with high lipoaffinity of many of these organic xenobiotics
 19      strongly control their interaction with  the vegetative components of natural ecosystems. The
20      cuticles of foliar surfaces are covered  with a wax layer that helps protect plants from moisture
21      and short-wave radiation stress.  This  epicuticular wax, consisting mainly of long-chain esters,
22      polyesters, and paraffins, has been demonstrated to accumulate lipophilic compounds.  Organic
23      air contaminants, in the particulate or  vapor phase, are absorbed to and accumulate in the
24      epicuticular wax of vegetative surfaces (Gaggi et al., 1985; Kylin et al., 1994).  Direct uptake of
25      organic contaminants through the cuticle or the vapor-phase uptake through the stomates are
26      characterized poorly for most trace organics.
27           The phytotoxiciry and microbial toxicity of organic contaminants to soil microorganisms is
28      not well studied (Foster, 1991).
       March 2001
4-19
DRAFT-DO NOT QUOTE OR CITE

-------
 1     4.2.2 Particulate Matter Effects on Natural Ecosystems
 2     4.2.2.1 Introduction
 3          Human existence on this planet depends on nature and the life-support services ecosystems
 4     provide.  Ecosystem services (Table 4-2) are the conditions and processes through which natural
 5     ecosystems, and the species of which they are comprised, sustain and fulfill human life (Daily,
 6     1997). Both ecosystem structure and function play an essential role in providing societal
 7     benefits.  Society derives two types of benefits from the structural aspects of an ecosystem:
 8     (1) products with market value such as fish, minerals, forage, forest products, biomass fuels,
 9     natural fiber, and many Pharmaceuticals and the genetic resources of valuable species (e.g.,
10     plants for crops and timber, animals for domestication); and (2) ecosystem services (Table 4-2)
11     include the use and appreciation of ecosystems for recreation, aesthetic enjoyment, and study
12     (Westman, 1977; Daily, 1997).  Economic benefits and values associated with ecosystem
13     functions and services and the need to preserve them because of their value to human life are
14     discussed by Costanza et al. (1997) and (Pimentel et al., 1997). Services usually are not
15     considered to be items with market value.
                               TABLE 4-2. ECOSYSTEM SERVICES
         • Purification of air and water
         • Mitigation of floods and droughts
         • Detoxification and decomposition of wastes
         • Generation and renewal of soil and soil fertility
         • Pollination of crops and natural vegetation
         • Control of the vast majority of potential agricultural pests
         • Dispersal of seeds and translocation of nutrients
         • Maintenance of biodiversity, from which humanity has derived key elements of its
          agricultural, medicinal, and industrial enterprises
         • Protection from the sun's harmful rays
         • Partial stabilization of climate
         • Moderation of temperature extremes and the force of winds and waves
         • Support of diverse human cultures
         • Providing of aesthetic beauty and intellectual stimulation that lift the human spirit
         Source: Daily (1997).
        March 2001
4-20
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
      Ecosystems are structurally complex biotic communities consisting of populations of
plants, animals, insects, and microorganisms interacting with one another and with their abiotic
environment (Odum, 1993).  They are dynamic, self-adjusting, self-maintaining, complex
adaptive systems in which patterns at higher levels of organization emerge from localized
interactions and selection processes. Macroscopic ecosystem properties such as structure,
diversity-productivity relationships and patterns of nutrient flux emerge from the interactions
among components and may feed back to influence subsequent development of those
interactions. The relationship between structure and function is a fundamental one in ecosystem
science (Levin, 1998).  Structure refers to the species, their biodiversity, abundance, mass,  and
arrangement within an ecosystem. Ecosystem functions, energy flow, nutrient flux, and water
and material flow, are characterized by the way in which ecosystem components interact.
Elucidating these interactions across scales is fundamental to understanding the relationships
between biodiversity and ecosystem functioning (Levin, 1998).  To function properly and
maintain themselves, ecosystem components must have an adequate supply of energy, chemical
nutrients, and water. It is the flows of nutrients and energy, that provide the interconnectedness
between ecosystem parts and transforms the community from a random collection of species into
an integrated whole, an ecosystem in which the biotic and abiotic parts are interrelated (Levin,
1998).
     Growth of new trees and other vegetation requires energy in the form of carbon
compounds. Plants accumulate, store, and use carbon compounds to build their structures and
maintain physiological processes. Plants, using energy from sunlight, hi their leaves combine
carbon dioxide from the atmosphere and water  from the soil to produce the carbon compounds
(sugars) that provide the energy required by vegetation for growth and maintenance (Waring and
Schlesinger, 1985). Energy is transferred through an ecosystem from organism to organism in
food webs and, finally, is dissipated into the atmosphere as heat (Odum, 1993).  Chemical
nutrients, such as nitrogen, phosphorus, or sulfur, on the other hand, are taken up from the soil by
plants and are transferred to other species through the food webs. The process is cyclic with the
chemical nutrients eventually returning to the soil. This process is referred to as biogeochemical
cycling (Odum, 1993).  The biogeochemistry of an ecosystem is influenced by vegetation growth
characteristics  (Herbert et al., 1999).
       March 2001
                                         4-21
DRAFT-DO NOT QUOTE OR CITE

-------
 1          Ecosystem functions are characterized by the way components interact. These are the
 2     functions that maintain clean water, pure air, a green earth, and a balance of organisms, the
 3     functions that enable humans to survive. They are the dynamics of ecosystems.  The benefits
 4     they impart include absorption and breakdown of pollutants, cycling of nutrients, binding of soil,
 5     degradation of organic waste, maintenance of a balance of gases in the air, regulation of radiation
 6     balance, climate, and the fixation of solar energy (Table 4-2; Westman, 1977; Daily, 1997).
 7          Concern has risen in recent years regarding the consequences of changing biological
 8     diversity of ecosystems (Tilman, 2000; Ayensu et al, 1999; Wall, 1999; Hooper and Vitousek,
 9     1997; Chapin et al., 1998). The concerns arise because human activities are creating
10     disturbances that are causing the loss of biodiversity and altering the complexity and stability of
11     ecosystems and producing changes in nutrient cycling (structure and function) (Pimm, 1984;
12     Levin, 1998; Chapin et al., 1998; Peterson et al., 1998; Tilman, 1996; Tilman and Downing,
13     1994; Wall, 1999; Daily and Ehrlich,  1999). There are  few ecosystems on earth today that are
14     not influenced by humans (Freudenburg and Alario, 1999; Vitousek et al., 1997; Matson et al.
15     1997; Noble and Dirzo, 1997). The scientific literature is filled with references discussing the
16     importance of ecosystem structure and function. Ecorisk, complexity, stability, biodiversity,
17     resilience, sustainability, managing earth's ecosystems, and ecosystem health are frequently
18 .    discussed topics. There is a need, therefore, to understand how ecosystems respond to both
19     natural and anthropogenic stresses and, especially, the ways that anthropogenic stresses are
20     impacting ecosystem services and products.
21
22     4.2.2.2 Ecosystem Responses to Stress
23           Ecosystem responses to stresses begin at the population level. Population changes
24     however, begin with the response of individual plants or animals.  Plant responses, both
25     structural and functional, must be scaled hi both tune and space and propagated from the
26     individual to the more complex levels of community interaction to produce observable changes
27     in an ecosystem (Figure 4-1). At least three levels of biological interaction are involved:  (l)the
28     individual plant and its environment,  (2) the population and its environment, and
29     (3) the biological community composed of many species and its environment (Billings, 1978).
30     The response of individual organisms within a population based on their genetic constitution
31     (genotype), stage of growth at time of exposure, and the microhabitats in which they are growing
        March 2001
4-22
DRAFT-DO NOT QUOTE OR CITE

-------
"X. Reactic
Level of^
Organizatior
Leaf
(cm2)
Branch
(cm2)
Tree
(m2)
Stand
(ha)
nTime
Minute




Day

V
•^


Year

	 »•
k» 	 •
9r\
m •
^S 1
1-^3






Decade
2
3
»4
»-5

*•§
>9
|-— ^
Vte
M*
1



Century


10
11
12
13
L.14


16
Injury Symptom
Needle necrosis
and abscission
Branch length,
bifurcation ratio,
and ring-width
growth altered
Reduction in
diameter and death
of tree
Decreases in
stand productivity,
increases in mortality
and alterations in
regeneration patterns
Key Changes in Processes
Reduced carbon assimilation
because of reduced radiation
Reduced carbon available for foliage
replacement and branch growth/
export Synergistic interaction
between mistletoe and tephra
deposition
Reduced carbon available for
height, crown, and stem growth
Influence of crown class on initial '
impact and subsequent recovery
Interaction between stand
composition and recovery
For a given level, the dot associated with a line begins with a process (e.g., photosynthesis for #1 under leaf) an
ends with the associated structure (e.g., the needle).
Evaluating Impacts Within a Level of Organization
Leaf Level
Branch Level
Carbon exchange-1
Carbon pools-2
Needle number and size-3
Needle retention/abscission-4
Carbon allocation-5
Branch growth-6
Branch morphology-7
Branch vigor-8
Branch retention-9
Tree Level Height and diameter growth-10
Crown shape and size-1 1
Tree vigor-12
Mortality-13
Stand Level Productivity-14
Mortality-15
Species composition-16
  Evaluating Interactions Between Different Levels of Organization
               The diagonal arrow indicates the interaction between any two levels of organization.
               The types of interaction are due to the properties of variability and compensation.
               A - Refers to the interaction between the leaf and branch levels, where, for example,
                  variability at the branch level determines leaf quantity, and compensation at the leaf
                  level in photosynthesis may compensate for the reduction in foliage amount.
               B - Refers to the interaction between the branch and the tree, where variability in branches
                  determines initial interception, branch vigor, and branch location in the crown;
                  compensation may be related to increased radiation reaching lower branches.
               C - Refers to the interaction between the tree and the stand. Both genetic and
                  environmental variability, inter- and intraspecific compensations, and tree historical
                  and competitive synergisms are  involved.
Figure 4-1.  Effects of environmental stress on forest trees are presented on a hierarchial
              scale for the leaf, branch, tree, and stand levels of organization. The
              evaluation of impacts within a level of organization are indicated by horizontal
              arrows.  The evaluation of interactions between different levels of organization
              are indicated by diagonal arrows.

Source: Hinckleyetal. (1992).
March 2001
4-23
DRAFT-DO NOT QUOTE OR CITE

-------
 1     vary in their ability to withstand the stress of environmental changes (Levin, 1998). Individual
 2     organisms within a population vary in their ability to withstand the stress of environmental
 3     changes.  The range within which these organisms can exist and function determines the ability
 4     of the population to survive. Those able to cope with the stresses survive and reproduce.
 5     Competition among the different species results in succession (community change over time) and
 6     ultimately produces ecosystems composed of populations of plant species that have the capability
 7     to tolerate the stresses (Rapport and Whitford, 1999; Guderian, 1985).
 8           The number of species in a community usually increases during succession in unpolluted
 9     atmospheres. Productivity, biomass,  community height, and structural complexity increase.
10     Severe stresses, on the other hand, divert energy from growth and reproduction to maintenance,
11     and return succession to an earlier stage (Waring and Schlesinger, 1985). Ecosystems are subject
12     to natural periodic stresses, such as drought, flooding, fire, and attacks by biotic pathogens (e.g.,
13     fungi, insects). Ecosystem perturbation by natural stresses can be only a temporary setback.
14     Extremely severe natural perturbations return succession to an earlier stage, reduce ecosystem
15     structure (scarcity of life forms and no symbiotic interactions) and functions, disrupt the plant
16     processes of photosynthesis and nutrient uptake, carbon allocation and transformation that are
17     directly related to energy flow and nutrient cycling, shorten food chains, and reduce the total
18 •    nutrient inventory (Odum, 1993). This transformation, however, sets the stage for recovery,
19     which permits the perturbed ecosystem to adapt to changing environments (Rolling, 1986).
20     Therefore, these perturbations are seldom more than a temporary setback, and recovery can be
21     rapid (Odum, 1969).
22           In contrast, anthropogenic stresses usually are severe, debilitating stresses. Severely
23     stressed ecosystems do not recover readily, but may be further degraded (Odum, 1969; Rapport
24     and Whitford, 1999). Anthropogenic stresses can be classified into four main groups:
25     (1) physical restructuring (e.g., changes resulting from land use); (2) introduction of exotic
26     species; (3) over harvesting; and (4) discharge of toxic substances into the atmosphere, onto land,
27     and into water. Ecosystems lack the  capacity to adapt to the above stresses and maintain their
28     normal structure and functions unless the stress is removed (Rapport and Whitford, 1999). These
29     stresses result in a process of degradation marked by a decrease in biodiversity, reduced primary
30     and secondary production, and a lower capacity to recover and return to its original state.
31     In addition, there is an increased prevalence of disease, reduced nutrient cycling, increased
        March 2001
4-24
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 dominance of exotic species, and increased dominance by smaller, short-lived opportunistic
 species (Odum, 1985; Rapport and Whitford, 1999). Discharge of toxic substances into the
 atmosphere, onto land, and into water can cause acute and chronic stresses and, once the stress is
 removed, a process of succession begins which can ultimately return the ecosystem to a
 semblance of its former structure. Air pollution stresses, if acute, are usually short term and the
 effects soon visible. Chronic stresses, on the other hand, are long-term stresses whose effects
 occur at different levels of ecosystem organization and appear only after long-term exposures, as
 in the case of acidic deposition in the northeast or ozone in California (Shortle and Bondietti,
 1992; U.S. Environmental Protection Agency, 1996b).
      The possible effects of air pollutants on ecosystems have been categorized by Guderian
 (1977) as follows:
      (1)  accumulation of pollutants in the plant and other ecosystem components (such as soil
          and surface- and groundwater),
      (2)  damage to consumers as a result of pollutant accumulation,
      (3)  changes in species diversity because of shifts in competition,
      (4)  disruption of biogeochemical cycles,
      (5)  disruption of stability and reduction in the ability of self-regulation,
      (6)  breakdown of stands and associations, and
      (7)  expanses of denuded zones.
      How changes in these functions can result from PM deposition and influence ecosystems is
discussed in the following text. It should be remembered that, although the effects of PM are
being emphasized, the vegetational components of ecosystems also are responding to multiple
stresses from other sources.

4.2.2.3  Ecosystem Response to Direct Plant Effects
     The presence of PM in the atmosphere may affect vegetation directly, following physical
contact with the foliar surface (Section 4.2), but in most cases, the more significant impacts are
indirect.  These impacts may be mediated by suspended PM (i.e., through effects on radiation and
climate) and by particles that pass through the vegetative canopies to the soil.  Particulate matter,
as considered in this chapter is a heterogeneous mixture of particles differing in size, origin, and
chemical constituents and their impacts vary depending on the chemical nature of PM being
        March 2001
                                          4-25
DRAFT-DO NOT QUOTE OR CITE

-------
 1     deposited on vegetation or soil. Particulate inputs and ecosystem cycling of key elements are
 2     considered below.
 3           The majority of studies dealing with direct effects of particulate dust and trace metals on
 4     vegetation have focused on responses .of individual plant species and were conducted in the
 5     laboratory or in controlled environments (Saunders and Godzik, 1986). A few have considered
 6     the effects of particles on populations, communities, and ecosystems.  Most of these focused on
 7     ecosystems in industrialized areas heavily polluted by deposits of both chemically inert and
 8     active dusts. Effects can result from direct deposition or indirectly by deposition onto the soil.
 9     Reductions in growth, yield, flowering, and reproduction of plants from particulate deposition
10     have been reported (Saunders and Godzik, 1986). Sensitivities of individual species have been
11     associated with changes in composition and structure of natural ecosystems.
12           Evidence from studies of effects of PM deposition, specifically chemically inert and active
13     dusts indicates that, within a population, plants exhibit a wide range of sensitivity, which is the
14     basis for the natural selection of tolerant individuals (Saunders and Godzik, 1986). Rapid
15     evolution of certain populations of tolerant species at sites with heavy trace element and nitrate
16     deposition has been observed. Tolerant individuals present in low frequencies in populations
17     when growing in unpolluted areas have been selected for tolerance at both the seedling and adult
18 .   stages when exposed to trace metal or nitrate deposition (Ormrod, 1984; U.S. Environmental
19     Protection Agency, 1993). Chronic pollutant injury to a forest community may result in the loss
20     of sensitive species, loss of tree canopy, and maintenance of a residual cover of pollutant-tolerant
21     herbs or shrubs that are recognized as successional species (Table 4-3; Smith, 1974). Frequently,
22     trace metals that penetrate the above-ground plant parts are less injurious than when taken up
23     through the roots (Guderian, 1986).
24           Responses of ecosystems to stresses (unless severe or catastrophic) are difficult to
25      determine because the changes are subtle (Garner, 1991). This is particularly true of responses to
26      particles. Changes in the soil may not be observed until accumulation of the pollutant has
27      occurred for 10 or more years except in the severely polluted areas around heavily industrialized
28      point sources (Saunders and Godzik, 1986). hi addition, the presence of other co-occurring
29      pollutants makes it difficult to attribute the effects to PM alone. In other words, the potential for
30      alteration of ecosystem function and structure  exists, but it is difficult to quantify, especially
        March 2001
4-26
DRAFT-DO NOT QUOTE OR CITE

-------
              TABLE 4-3. ECOSYSTEM FUNCTIONS IMPACTED BY AIR POLLUTION
             	EFFECTS ON TEMPERATE FOREST ECOSYSTEMS
         Forest Soil and Vegetation: Activity and Response
           Ecosystem Consequence and Impact
         1.  Forest tree reproduction, alteration, or inhibition

         2.  Forest nutrient cycling, alteration
            a. Reduced litter decomposition
            b. Increased plant and soil leaching and soil
              weathering
            c. Disturbance of microbial symbioses

         3.  Forest metabolism
            a. Decreased photosynthesis
            b. Increased respiration
            c. Altered carbon allocation

         4.  Forest stress, alteration
            a. Phytophagous insects, increased or decreased
              activity
            b. Microbial pathogens, increased or decreased
              activity
            c. Foliar damage increased by direct air pollution
              influence
        1. Altered species composition

        2. Reduced growth, less biomass
       3. Reduced growth, less biomass
       4. Altered ecosystem stress:
          increased or decreased insect infestations;
          increased or decreased disease epidemics;
          and reduced growth, less biomass, and
          altered species composition
         Source:  Smith (1974).
 1 .     when there are other pollutants present in the ambient air, which may produce additive or

 2      synergistic responses, even though PM concentrations may not be elevated.
 3

 4      Physical Effects

 5           The direct effects of limestone dust on plants and ecosystems has been known for many

 6      years. Long-term changes in the structure and composition of the seedling-shrub and sapling

 7      strata of an experimental site near limestone quarries and processing plants in Giles County in

 8      southwestern Virginia were reported by Brandt and Rhoades (1972, 1973). Dominant trees in the

 9      control area, a part of the oak-chestnut association of the eastern deciduous forests of eastern

10      North America, were chestnut oak (Quercus prinus), red oak (Q. rubra), and red maple (Acer

11      rubrum). An abundance of uniformly distributed saplings and seedlings were visible under the

12      tree canopy, and herbs appeared in localized areas in canopy openings.  Q. prinus dominated the

13      area,  and the larger trees were  60 to 80 years old. The dusty site was dominated by white oak
       March 2001
4-27
DRAFT-DO NOT QUOTE OR CITE

-------
 1     (Q. alba), whereas Q. rubra and Tulip poplar (Liriodendron tulipifera) were subcodominants.
 2     The largest trees were 100 years old and had necrotic leaves, peeling bark, and appeared to be in
 3     generally poor condition except for L. tulipifera (which thrived in localized areas).  The site
" 4     contained a tangled growth of seedlings and shrubs, a few saplings, and a prevalence of green
 5     briar (Smilax spp.) and grape (Vitis spp).  The sapling strata in the area was represented by Acer
 6     rubrum, hickory (Carya spp.), dogwood (Cornusflorida), and hop-hornbeam (Ostrya
 7     virginiana). Saplings of none of the leading dominant trees were of importance in this stratum.
 8     The most obvious form of vegetation hi the seedling-shrub stratum, because of their tangled
 9     appearance, were C.florida, Ostrya virginiana, redbud (Cercis canadensis), and sugar maple
10     (Acersaccarum).
11          Crust formation reduced photosynthesis, induced premature leaf fall, destruction of leaf
12     tissues, inhibited growth of new tissue and reduced the formation of carbohydrate needed for
13     normal growth and storage (Brandt and Rhoades, 1973).  The authors (Brandt and Rhoades,
14     1972), citing Odum (1969), also stated that a result of the accumulation of toxic pollutants in the
15     biosphere as the result of human activities, is the simplification of both plant and animal
16     communities.  In plant communities, structure is determined by sampling various strata within
17     the community. Each stratum comprises a particular life form (e.g., herbs, seedlings, saplings,
18 '    trees). Dust accumulation favored growth of some species and limited others. For example,
19     Acer saccharum was more abundant in all strata of the dusty site when compared with the control
20     site where it was present only as a seedling. The growth of L. tulipifera, C.florida,
21     O. virginiana, black haw (Viburnum prunifolium), and C. canadensis appeared to be favored by
22     the dust. Growth of conifers and acidophiles such as rhododendron (Rhododendron maximum),
23     however, was limited.  Although dust accumulation began in 1945, the heaviest accumulation
24     occurred between 1967 and 1972 during the time of the study.
25           Changes in community composition were associated closely with changes in the growth of
26     the dominant trees. Decrease hi density of seedlings and saplings and in mean basal area, as well
27     as lateral growth of A. rubrum, Q. prinus, and Q. rubra, occurred in all strata. On the other hand,
28     all of these characteristics increased hi L. tulipifera, which was a subordinate species before dust
29     accumulation began but had assumed dominance at the tune of the study.  Reduction in growth of
30     the dominant trees had apparently given L. tulipifera competitive advantage because of its ability
 31     to tolerate  dust. Changes in soil alkalinity occurred because of the heavy deposition of limestone
        March 2001
4-28
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 dust; however, the facilities necessary for critical analysis of the soils were not available. From
 the foregoing, it is obvious that PM physical effects in the vicinity of limestone quarries and
 processing plants can impact ecosystems.

 Acidic Deposition
      The effects of acidic deposition have been discussed in several previous reports. The 1982
 EPA document, Air Quality Criteria for Paniculate Matter and Sulfur Oxides, devoted a chapter
 to the effects of acidic deposition (U.S.  Environmental Protection Agency, 1982). In 1984, EPA
 published The Acidic Deposition Phenomenon and Its Effects (Altshuller and Linthurst, 1984),
 and, in 1991, NAPAP published the result of its extensive study, Acidic Deposition: State of
 Science and Technology (Irving,  1991). The major effects of acidic deposition occur through the
 soil and are discussed under indirect effects. However, included among the direct responses of
 forest trees to acidic deposition are increased leaching of nutrients from foliage; accelerated
 weathering of leaf cuticular surfaces; increased permeability of leaf surfaces to toxic materials,
 water, and disease agents; and altered reproductive processes (Altshuller and Linthurst, 1984).

 Trace Elements                ,                                           ,
      Possible direct responses of trace elements on vegetation result from their deposition and  .
 residence on the phyllosphere (i.e., leaf surfaces).  Fungi and other microorganisms living on the
 leaves of trees and other vegetation play an important role in leaf decomposition after litterfall
 (Miller and McBride, 1999; Jensen, 1974; Millar, 1974). Early needle senescence and abscission
 in the San Bernardino Forest changed fungal microflora successional and decomposition patterns
 by altering the taxonomic diversity and population density of microflora that normally develop
 on needles while they are on the tree.  Changing the fungal community on the needles weakened
 the decomposer community, decreasing the rate of decomposition, and altered nutrient cycling
 (Bruhn, 1980). Nutrient availability was influenced by accumulation of carbohydrates and
 mineral nutrients in the heavy litter under those stands with the most severe needle injury and
 defoliation (U.S.  Environmental Protection Agency, 1996b). Possible impacts of heavy metals
on nutrient cycling and their effects on leaf microflora appear not to have been studied.,
     A trace metal must be brought into solution before it can enter into the leaves or bark of
vascular plants. Low solubility limits entry.  In those instances when trace metals are absorbed,
        March 2001
                                          4-29
DRAFT-DO NOT QUOTE OR CITE

-------
 1      they frequently are bound in the leaf tissue and then are lost when the leaf drops off (Hughes,
 2      1981) and can affect litter decomposition, an important source of soil nutrients. Changes in litter
 3      decomposition processes influence nutrient cycling in the soil and limit the supply of essential
 4      nutrients. Both Cotrufo et al. (1995) and Niklinska et al. (1998) point out that heavy metals have
 5      impacts on forest litter decomposition. Cotrufo et al. (1995) observed that decomposition of oak
 6      leaves containing Fe, Zn,  Cu, Cr, Ni, and Pb was influenced strongly during the early stages by
 7      metal contamination.  Fungal mycelium was significantly less abundant in litter and soil in
 8      contaminated sites, when compared with control sites. Niklinska et al. (1998) stated that toxic
 9      effects of heavy metals on soil respiration rate have been reported by many scientists, and that, in
10     polluted environments, this results in accumulation of undecomposed organic matter.  However,
11      they state that results of experiments should identify the most important "natural" factors
12     affecting soil/litter sensitivity because the effects of heavy metals on respiration rates depend on
13     the dose of heavy metals, the type of litter, types of metals deposited, and the storage time before
14     respiration tests are made.
15           Trace metals, particularly heavy metals  (e.g., cadmium, copper, lead, chromium, mercury,
16     nickel, zinc) have the greatest potential for influencing forest growth (Smith, 1991).
17     Experimental data indicate that the broadest spectrum of growth suppression of foliar microflora
18 •    resulted from iron, aluminum, and zinc. These three metals also inhibited spore formation, as did
19      cadmium, chromium, manganese, and nickel  (see Smith, 1990e). In the field, the greatest injury
20      occurs from pollution near rnining, smelting,  and other industrial sources (Ormrod, 1984). Direct
21      metal phytotoxicity can occur only if the metal can move from the  surface into the leaf or directly
22      from the soil into the root.
23
24      Organic Compounds
25           Secondary organic  compounds formed in the atmosphere, the effects of some of which are
 26      discussed below, have been referred to under the following terms:  toxic substances, pesticides,
 27      hazardous air pollutants (HAPS), air toxics, semivolatile organic compounds (SOCs), and
 28      persistent organic pollutants (POPS).  Again, it should be noted that the chemical substances
 29      denoted by such headings are not criteria air pollutants controlled by the NAAQS under
 30      Section 109 of the Clean Air Act (CAA) (U.S. Code,  1991), but rather are controlled under
 31      Sect. 112, Hazardous Air Pollutants. Their possible effects on humans and ecosystems are
        March 2001
4-30
DRAFT-DO NOT QUOTE OR CITE

-------
  1      discussed in a number of government documents and in many other publications. They are
  2      mentioned here because, in the atmosphere, many of the chemical compounds are partitioned
  3      between gas and particle phases.  As particles, they can become airborne, be distributed over a
  4      wide area, and impact remote ecosystems. Some of the chemical compounds are of concern
  5      because they may reach toxic levels in food chains of both animals and humans, whereas others
  6      tend to decrease or maintain the same toxicity as they move through the food chain.  Some
  7      examples of movement through food chains are provided below.
  8           Many chemical compounds from a variety of anthropogenic sources are released into the
  9      ambient air (See Section 4.2.1). In the atmosphere, the emitted compounds initially go through a
10      mixing process, and the airborne particles then  are distributed over a wide area and ultimately
11      deposited on ecosystem components. Atmospheric deposition of polychlorinated dibenzo-p-
12      dioxins and dibenzofurans (PCDD/Fs), as an example, can be divided into three different forms:
13      (1) dry gaseous, (2) dry particle-bound, and (3)  wet deposition. Dry particle-bound deposition
14      occurs when the PM containing the pollutant is deposited on the plant surface, whereas wet
15      deposition ranges from hail through rain to fog  and dew fall (McLachlan, 1996b).
16           Human exposure to PCDD/Fs has been demonstrated to be caused almost exclusively by
17      the ingestion of animal fat  from fish, meat, and dairy products. Almost half of human exposure
18      to PCDD/Fs is caused by consumption of beef and dairy products (McLachlan, 1996b). Cattle
19      obtain most of their PCCD/Fs though grass. Therefore, the grass-cattle-milk/beef pathway is
20      critical for human exposure.  It has been shown that root uptake/translocation is an insignificant
21      pathway of PCDD/Fs to aerial plant parts. Wet and dry particle deposition are the most
22      important for the accumulation of the higher chlorinated cogeners in vegetation. The persistence
23      of PCDD/Fs in plants has not been investigated extensively; however, biodegradation probably
24      does not occur in that these compounds are found primarily in the lipophilic cuticle and are very
25      resistant to microbial degradation (McLachlan,  1996b). Feed contaminated  with soil containing
26      the pollutant also can be another source of exposure of beef and dairy cattle  as well as chickens.
27      The PCDD/Fs are near a steady state in milk cows and laying hens; however, animals raised for
28      meat production (such as beef cattle and pigs) may accumulate them. The beef cattle and pigs
29      cannot excrete the contaminants in a lipid-rich matrix such as milk or eggs.  All of the PCDD/Fs,
30      ingested are stored in the body. In agricultural food chains, there is a biodilution of PCDD/Fs,
31      with the fugacity decreasing by up to three orders of magnitude between the air and cows milk
       March 2001
4-31
DRAFT-DO NOT QUOTE OR CITE

-------
 1     (McLachlan, 1996b). Fiirst et al. (1993), based on surveys to determine the factors that influence
 2     the presence of PCDD/PCDF in cows milk, earlier concluded that regardless of which pathway,
 3     soil - grass - cow or air - grass - cow, it was the congener of the chemical that was most
 4     important.
 5          Persistent polychlorinated pollutants (POPS), such as PCBs, PCDFs, and PCDDs, can be
 6     transported as particles through the atmosphere from industrial and agricultural sources; be
 7     brought down via wet and dry deposition in remote regions, such as the Arctic; and have been
 8     detected in all levels of the Arctic food chain (Oehme et al., 1995). High concentrations of PCB
 9     (1 to 10 ppm) were found in seals; but the concentrations increased to 10 to 100 ppm in polar
10     bears.  The polar bear is the top predator in the Arctic and feeds preferentially on ringed seals and
11     also, to a lesser extent, on other seal species.  Bioconcentration factors of organochlorines in the
12     Arctic food web, reaching 107 for fish and seals, are biomagnified in polar bears (Oehme et al.,
13     1995). Polychlorinated dibenzo-/?-dioxins (PCDDs) and polychlorinated dibenzofurans
14     (PDCF/s) also have also been found in seals (Oehme et al., 1995). Milk taken from
15     anaesthetized polar bears was also found to contain PCDD/PCDF.  Very little is known regarding
16     the intake of milk by polar bear cubs.  However, estimates of the intake of milk containing
17     detectable levels of PCDD/PCDF and PCB and the additional consumption of seal blubber
18 '    confirm that these pollutants are passed on to the next generation (Oehme et al., 1995).
19          Section 112 of the CAA, provides the legislative basis for U.S. hazardous air pollutant
20     (HAP) programs. In response to mounting evidence that air pollution contributes to water
21     pollution, Congress included Section 112m (Atmospheric Deposition to Great Lakes and Coastal
22     Waters) in the 1990 CAA Amendments, which directs the Environmental Protection Agency
23     (EPA) to establish a research program on atmospheric deposition of HAPS to the "Great
24     Waters".
25          Actions taken by EPA and others to evaluate and control sources of Great Waters pollutants
26     of concern appear to have positively affected trends in pollutant concentrations measured in air,
27     sediment, and biota. Details concerning these effects may be found in "Deposition of Air
28     Pollutants to the Great Waters", Third Report to Congress (U. S. Environmental Protection
29     Agency, 2000a). The Third Report (EPA-453/R-00-005, June 2000), like the First and Second
30     Reports to Congress, focuses on 15 pollutants of concern, including pesticides, metal
31     compounds, chlorinated organic compounds, and nitrogen compounds. The new scientific
        March 2001
4-32
DRAFT-DO NOT QUOTE OR CITE

-------
  1      information in the Third Report supports and builds on three broad conclusions presented in the
  2      previous two EPA Reports to Congress and discussed below.
  3      (1)  Atmospheric deposition from human activities can be a significant contributor of toxic
  4          chemicals and nitrogen compounds to the Great Waters. The relative importance of
  5          atmospheric loading for a particular chemical in a water body depends on many factors (e.g.,
  6          characteristics of the water body, properties of the chemical, and the kind and amount of
  7          atmospheric deposition versus or water discharges).
  8      (2)  A plausible link exists between emissions into the air of Great Waters toxic pollutants of
  9          concern; the atmospheric deposition of these  pollutants (and their transformation products);
10          and the concentrations of these pollutants found in the water, sediments, and biota, especially
11          fish and shellfish.  For mercury, fate and transport modeling and exposure assessments
12          predict that the anthropogenic contribution to the total amount of methylmercury in fish is, in
13          part, the result of anthropogenic mercury releases from industrial and combustion sources
14          increasing mercury body burdens (i.e., concentrations) in fish.  Also, the consumption of fish
15          is the dominant pathway of exposure to methylmercury for fish-consuming humans and
16          wildlife. However, what is known about each stage of this process varies with each pollutant
17          (for instance, the chemical species of the emissions and its transformation in the
18          atmosphere).
19      (3)  Airborne emissions from local as well as distant sources, from both within and outside the
20          United States, contribute pollutant loadings to waters through atmospheric deposition.
21          Determining the relative roles of particular sources—local, regional, national, and possibly
22          global, as well as anthropogenic, natural, and reemission of pollutants—contributing to
23          specific water bodies is complex, requiring careful monitoring, atmospheric modeling, and
24          other analytical techniques.
25
26      4.2.2.4 Indirect Effects of Particulate Matter In Ecosystems
27           The presence of PM in the atmosphere directly affects vegetation following physical
28      contact with foliar surfaces (as discussed above in Section 4.2.2.2),  but in many cases the more
29      significant impacts are indirect. These impacts may be mediated by suspended PM (i.e., through
30      effects on radiation and climate) and by particles that pass through vegetative  canopies to reach
31      the soil. Effects mediated in the atmosphere are considered briefly below and in greater detail
        March 2001
4-33
DRAFT-DO NOT QUOTE OR CITE

-------
 1     later, under Section 4.5. Indirect plant responses are chiefly soil mediated and depend primarily
 2     on the chemical composition of the individual elements deposited in PM.  The individual
 3     elements must be bioavailable to have an effect. The soil environment, composed of mineral and
 4     organic matter, water, air, and a vast array of bacteria, fungi, algae, actinomycetes, protozoa,
 5     nematodes, and arthropods, is one of the most dynamic sites of biological interactions in nature
 6     (Wall and Moore, 1999; Alexander, 1977). The quantity of organisms in soils varies by locality.
 7     Bacteria and fungi are usually most abundant in the rhizosphere, the soil around plant roots that
 8     all mineral "nutrients must pass through. Bacteria and fungi benefit from the nutrients in the root
 9     exudates (chiefly sugars) in the soil and, in turn, they play an essential role by making mineral
10     nutrients available for plant uptake (Wall and Moore, 1999; Rovira and Davey, 1974). Their
11     activities create chemical and biological changes in the rhizosphere by decomposing organic
12     matter and making inorganic minerals available for plant uptake. Bacteria are essential in the
13     nitrogen and sulfur cycles and make these elements available for plant uptake and growth (see
14     Section 4.3.3). Fungi are directly essential to plant growth. Attracted to the roots by the
15     exudates, they develop mycorrhizae, a mutualistic, symbiotic relationship, that is integral in the
16     uptake of the mineral nutrients (Allen, 1991). The impact in ecosystems of PM, particularly
17     nitrates, sulfates, and metals, is determined by their affect on the growth of the bacteria involved
18'   in nutrient cycling and the fungi involved in plant nutrient uptake.
19
20     Particulate Matter-Related Atmospheric Turbidity: Effects on Vegetative Processes
21           Photosynthetic processes underlie the contribution of vegetative surfaces to nutrient and
22     energy cycling. Photosynthesis and the heat-driven processes of water cycling depend on net
23     receipts and characteristics of the radiation environment. These characteristics may be altered
24     substantially when the atmosphere becomes turbid because of particulate loading.
25           Specific wavelengths of interest depend on the vegetation process under consideration.
26     Canopy temperature and water relations are particularly sensitive to long-wave, infrared
27     radiation, whereas primary photosynthetic charge separations depend on short-wave radiation in
28     the visible and photosynthetically active range (0.4 to 0.7/mi).
29           Effects of anthropogenic aerosols on the radiation  environment at the Earth's surface are
30     difficult to assess. The residence time of suspended particles varies with size and environmental
31     conditions (seconds to months or years), and concentrations are  spatially and temporally variable.
        March 2001
4-34
DRAFT-DO NOT QUOTE OR CITE

-------
  1      In particularly polluted urban and near-urban areas, unambiguous particulate impacts on radiation
  2      and local climate may be observed. Visibility was degraded by 50% in a large plume originating
  3      in the St. Louis urban area during the midweek, midday period (Pueschel, 1993). In contrast,
  4      visibility was reduced by only 20% on weekends, when traffic and industrial emissions were
  5      reduced. The area affected by this plume includes highly productive agricultural land.
  6           Empirical relationships between mass of specific components of the aerosol and radiation
  7      scattering have been developed (e.g., Pueschel, 1993), from which regional visibility (or
  8      radiation attenuation) isopleths can be constructed if appropriate mass data are available. These
  9      estimates support trends observed by direct measurement of turbidity (e.g., Flowers et al., 1969;
 10      U.S. Environmental Protection Agency, 1982).
 11           Sulfates, nitrates, and elemental carbon dominate effects on visibility, in part, because they
 12      frequently dominate the mass profiles and, in part, because they exhibit particularly large
 13      absorption coefficients (see Section 4.3). Absorption by particles containing carbon may range
 14      from 5 to 10% in rural areas to up to  50% in urban areas (U.S. Environmental Protection Agency,
 15      1982).  In west-coast cities with contrasting particulate sources and loadings, the common
 16      component that related PM to visibility degradation was sulfate between 0.65 and 3.6 yum
 17      (Barone et al., 1978).  For example, in Los Angeles, sulfate and nitrate had similar effects on
 18      visibility (White, 1976), despite the dominance of nitrate from transportation sources in the
 19      aerosol, although this is changing with controls on point sources of sulfate (Farber et al., 1994).
20           No long-term global trend of increasing atmospheric optical depth has been documented
21      (Bolle et al., 1986; Pueschel, 1993), although seasonal and  regional impacts are substantial. The
22      classic study by Flowers et al. (1969) demonstrated large regional distinctions in turbidity across
23      the United States.  Typically, the western deserts, plains, and Rocky Mountains exhibited low
24      mean annual turbidity, whereas the more humid and densely vegetated eastern half of the country
25      exhibited much greater turbidities.  In the mid-1970s, visible range hi the mountainous  southwest
26      exceeded 110 km and radiation attenuation was ca. 2.6%; whereas, in the east, visible range was
27      below 24 km and radiation attenuation was ca. 10%.  Visibility in the eastern United States has
28      decreased generally since the 1940s (Flowers et al., 1969; Trijonis and Shapland, 1979; U.S.
29      Environmental Protection Agency, 1982). Correlative trends in visibility degradation and
30      emissions of sulfur oxides suggest that particulate sulfate may account for much  of the turbidity.
        March 2001
4-35
DRAFT-DO NOT QUOTE OR CITE

-------
 1           These trends are typical of urban industrial areas around the world. Turbidity has increased
 2      above Mexico City (Binenko and Harshvardhan, 1993) since the 1911 to 1928 period. During
 3      this early period, a single annual peak of turbidity coincided with the end of the dry period, and
 4      natural sources dominated. By 1957 to 1962, the number of annual peaks had increased, as
 5      anthropogenic sources came to dominate. During this period, atmospheric transmission of direct-
 6      beam solar radiation decreased by about 10% (Binenko and Harshvardhan, 1993). Visibility in
 7      the Los Angeles basin has improved very slightly in the past decades (Farber et al., 1994), as
 8      sulfate emissions have been controlled by regulation.  The composition of the aerosol has
 9     changed, particularly in inland areas, as the former dominance of sulfate shifts to a
10     preponderance of secondary organics.
11           Particles interact with solar radiation through scattering and absorption. Absorption of
12     short-wavelength solar radiation reduces the amount of radiation reaching the Earth's surface and
13     leads to atmospheric heating.  If the absorbing particles reradiate in the infrared range, then some
14     of this energy is lost as long-wave reradiation to space. This loss mechanism is minimized
15     because most of the  anthropogenic aerosol in the troposphere resides in the planetary boundary
16     layer (Bolle et al., 1986), even within the lower 500 m (Binenko and Harshvardhan,  1993), where
17     the temperature is similar to that of the surface. Some of this energy is captured at the surface as
18     down-welling infrared radiation.
19           These wavelengths directly impact canopy temperatures and influence transpirational water
20     use by vegetation. The presence of absorbing aerosols reduces the ratio of photosynthetically
21     active radiation to total radiation received at the surface, potentially reducing photosynthetic
22     water use efficiency. The net effect of aerosol absorption on the surface depends on the relative
23     magnitudes of the particulate absorption coefficients in the visible  and infrared area  and on the
24      albedo of the Earth's surface.  In general, absorption is not a dominant particulate effect.
25           Scattering of radiation dominates the effects of particulate loading on visibility and
26      turbidity. Nonabsorbing, scattering aerosols raise the overall albedo of the atmosphere and
27      reduce the amount of radiation reaching the surface by the amount reflected or backscattered to
28      space.  As atmospheric turbidity increases, so does the scattering of light, including forward
29      scattering of photosynthetically active radiation that intercepts the  Earth's surface (Hoyt, 1978).
30           The largest effect is described by Mie-scattering theory. Forward scattering reduces the
31      intensity of direct radiation by disrupting the solar beam, thereby increasing the path length and
        March 2001
4-36
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 probability of absorption and also increases the intensity of diffuse (sky) radiation.  In a clear
 atmosphere, diffuse radiation may be on the order of 10% of total solar radiation (Choudhury,
 1987). However, in highly turbid, humid conditions, this fraction may increase, even up to 100%
 of solar radiation in extreme cases.  The direct-to-diffuse-radiation ratio is highest at solar noon
 and lowest near dawn or dusk, when the path length through the atmosphere is longest.
      Particle scattering is wavelength dependent, causing objects to appear blue- or red- tinged,
 depending on viewing and illumination angles and on the light quality, the alteration of which is
 a minor contributor to photosynthetic light-use efficiency. The wavelength dependence of
 scattering decreases rapidly from extreme sensitivity for very fine particles to little dependence at
 10 /j,m. Equations relating scattering at a reference wavelength to scattering at wavelengths of
 interest are rigorously applicable only to spherical particles but may be extended to nonspherical
 particles of equal volume (Janzen, 1980).
      World Meteorological Organization (WHO) data summarized in U.S. Environmental
 Protection Agency (1982) indicated that turbidity in the eastern United States commonly resulted
 in radiation losses of ca. 3.5% because of backscattered radiation and ca. 3.5% because of
 absorption, with a resulting total reduction of incident radiation to ca. 93% of total solar
 radiation. However, 28% of the radiation reaching the surface was converted from direct
 radiation to diffuse, or sky, radiation. Under more polluted condition's, losses were ca.' 9%
 backscattered and 9% absorbed, reducing total radiation to 82% of total solar radiation and
 converting 72% from direct beam to diffuse radiation. Photosynthetically active radiation (0.4 to
 0.7 //m) typically is enriched in diffuse radiation relative to total or direct beam radiation.

Altered Radiative Flux:  Effects on Vegetative Processes
      Canopy photosynthesis is typically a nearly linear function of incident radiation,
 overcoming saturation exhibited by individual leaves by distributing the light throughout the
 multilayer canopy. Light penetration into canopies limits photosynthetic productivity (Rosenberg
 et al., 1983).  The uppermost leaves  of many canopies are at or above light saturation for
photosynthetic processes. The simplest radiative transfer functions describing plant canopies
relate total down-welling radiation (direct plus diffuse radiation measured above the canopy) to
radiation interception at each leaf level through a Beer's Law analogy. The expected exponential
decline in radiation through the canopy depends only on total radiation and a bulk canopy
        March 2001
                                          4-37
DRAFT-DO NOT QUOTE OR CITE

-------
 1      extinction coefficient that depends on leaf size, orientation, and distribution, as well as on
 2      reflectance and absorption in wavelengths of interest.  These simplified models predict radiation
 3      distribution adequately for homogeneous canopies. Turbidity affects canopy processes only by
 4      attenuating the total radiation impinging on the canopy surface.
 5           In more complex, and more realistic, canopy-response models (e.g., Choudhury, 1987),
 6      radiation is considered in its direct and diffuse components.  Foliar interception by canopy
 7      elements is considered for both up- and down-welling radiation (a two-stream approximation).
 8      In this case, the effect of atmospheric PM on turbidity affects canopy processes both by radiation
 9      attenuation and by influencing the efficiency of radiation interception throughout the canopy
10      through conversion of direct to diffuse radiation (Hoyt, 1978).  Diffuse radiation is more
11      uniformly distributed throughout the canopy and increases canopy photosynthetic productivity by
12     distributing radiation to lower leaves. The treatment of down-welling direct-beam radiation in
13      the two-stream approach remains an elaboration of the simplified Beer's Law analogy, with solar
14     angle, leaf area distribution, and orientation individually parameterized (Choudhury,  1987).
15     Diffuse down-welling radiation is a function of diffuse and direct radiation at the top of the
16     canopy and penetration within the canopy, according to cumulative leaf area density and foliage
17     orientation. Up-welling (diffuse) radiation results from scattering and  reflectance within the
18     canopy, and by the soil, of both direct and diffuse down-welling radiation.
19           The effect of the altered distribution between diffuse and direct radiation impacts
20     photosynthesis in upper, exposed leaves as a function of leaf angle and in total canopy
21     photosynthesis as a function of penetration of radiation within the canopy.  This depends on
22     canopy structure, leaf optical properties, and leaf area density, as well as on solar angle and
23     atmospheric turbidity. Absorption of radiation by particles heats the upper atmosphere and
24     results in reduced vertical temperature gradients. This could reduce the intensity of atmospheric
25     turbulent mixing. The magnitude of such potential effects on turbulent transport within canopies
26      remains unknown, although damping of eddy transport could inhibit canopy gas exchange.
27      Suppressed tropospheric mixing also could intensify local temperature inversions and increase
28      the severity of pollution episodes (Pueschel, 1993), with direct inhibitory effects on
29      photosynthetic processes.
30           The most significant effect of aerosols on vegetation is probably through their role as cloud
31      condensation nuclei because clouds have substantial impact on radiation receipts at the surface.
        March 2001
4-38
DRAFT-DO NOT QUOTE OR CITE

-------
  1      An important characteristic of fine particles is their ability to affect the flux of solar radiation
  2      passing through the atmosphere directly, by scattering and absorbing solar radiation, and
  3      indirectly, by acting as cloud condensation nuclei which in turn influence the optical properties
  4      of clouds (Chameides et al., 1999).  Regional haze has been estimated to diminish surface solar
  5      visible radiation by approximately 8%. Crop yields have been reported as being sensitive to the
  6      amount of sunlight received. The potentially significant impact of regional haze on the yield of
  7      crops because of reduction in solar radiation has been examined by Chameides et al. (1999).
  8      Using a case study approach, Chameides et al. (1999), studied the affects of regional haze on
  9      crop production in China, where regional haze is especially severe. A rudimentary assessment of
10      the direct effect of atmospheric aerosols on agriculture suggests that optimal crop yields of
11      approximately 70% of the crops are being depressed by at least 5 to 3% by regional scale air
12      pollution and its associated haze (Chameides et al., 1999).
13
14      Effects of Solar Ultraviolet Radiation on Terrestrial Ecosystems
15           The transmission of solar UV-B radiation through the earth's atmosphere is controlled by
16      ozone, clouds and particles.  The depletion of stratospheric ozone caused by the release of
17      chlorofluorcarbons (CFCs) and other substances, such as halides, has resulted in heightened
18      concern about potentially deleterious increases in the amount of solar UV-B (SUVB) radiation
19      reaching the Earth's surface (see Section 4.5). One salient consideration is that, although CFC
20      production is at a peak level now, the problem likely will continue well into the future because of
21      the length of time it takes for molecules to reach the stratosphere (Greenberg et al, 1997).
22           The vulnerability of terrestrial plants to UV-B results from their requirement for sunlight
23      for photosynthesis. Each 1% decline in ozone has been predicted to decrease crops yield by 1%
24      (Greenberg et al., 1997). In addition to inhibiting photosynthesis, UV-B radiation triggers
25      numerous responses in plants (e.g., membrane, protein, and DNA damage; delayed maturation;
26      diminished growth; activation of chemical stress; flavonoid synthesis; leaf thickening)
27      (Table 4-4). It is not known which of the injury and damage effects are most detrimental to plant
28      growth (Table 4-4). Effects of increased UV-B on plant growth are likely to be incremental.
29      Because plants evolved under the selective pressure of ambient UV-B radiation in sunlight, they
30      have developed adaptive mechanisms (Greenberg et al., 1997). Although inhibition of
31      photosynthesis is a detrimental growth effect, flavonoid synthesis represents acclimation.
        March 2001
4-39
DRAFT-DO NOT QUOTE OR CITE

-------
         TABLE 4-4.  TYPES OF PLANT RESPONSES TO ULTRAVIOLET-B RADIATION8
        Acclimation and Morphological Responses  Damage and Injury Responses	      -
        Altered biomass distribution
        Altered leaf cell division
        Cotyledon curling
        Increased DNA repair
        Increased flavonoid biosynthesis
        Increased leaf thickness
        Increased leaf number
        Increased number of tillars
        Leaf wrinkling
        Reduced leaf area
        Reduced hypocotyl growth
        Reduced shoot height
        Reduced stomatal density
Altered gene expression
Degradation of auxin
Degradation of chlorophyll and carotenoids
Degradation of proteins
Diminished biomass
Epidermal collapse
Inhibition of growth
Inhibition of photosynthesis
Increased stomatal conductance
Lower seed yield
Oxidation of DNA
Peroxidation of lipids
Prymidine dimer formation	
        "Entries in alphabetical order.
 1     Plants growing under full light have been shown to be protected against UV-B effects but not
 2     when growing under weak visible light (Bjorn, 1996). A common adaptation is alteration in leaf
 3     transmission properties, which results in attenuation of UV-B in the epidermis before it can reach
 4     the leaf interior.
 5          Plant species vary enormously in their response to UV-B exposures, and large differences
 6     in response occur among different genotypes within a species. In general, dicotyledonous plants
 7     are more sensitive than monocotyledons from similar environments. In addition, plant responses
 8     may differ depending on stage of development.  Therefore, extrapolation of experimental
 9     responses from seedlings to mature plants must be taken with caution (Bjorn, 1996). The above
10     facts are especially important when considering the effects of UV-B on agricultural plants.
11     For example, among soybeans and rice, there are varieties for which growth and crop yield are
       March 2001
4-40
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 severely decreased by increased UV-B radiation and other varieties that are not affected or may
 even be stimulated. On the other hand, the growth of the same sensitive soybeans when grown
 under water stress was not inhibited.  Many crop plants grown in temperate regions originated in
 more tropical areas, hence, a gene pool for more resistant varieties is likely to exist (Bjorn, 1996).
 Crop plants, unlike forest trees and vegetation in natural ecosystems, are only exposed for one
 generation, and thus, it may be possible to readily change the genotype if a variety proves to be
 sensitive.
     Trees, forests, and perennial evergreen plants are long-lived when compared to agricultural
 systems, making it possible for UV-B exposure impacts to accumulate with time.  Saplings and
 young and small trees react differently when compared to mature trees; also, on evergreen trees,
 needles of different ages respond differently (Bjorn, 1996). Breeding and testing trees is a slow
 process, and, for this reason, much care needs to be taken when planting large areas with trees of
 a single species and one provenance (e.g., Stika Spruce [Picea sitchensis] in Britain). The
 response of only a few broad-leaved trees have been studied. The most investigated genus has
 been loblolly pine (Pinus taeda) (Bjorn, 1996).
     A few studies indicate that the photomorphogenesis (changes in leaf thickness under UV-B
 that results in a transition from shade  to sun  leaves, Table 4-4) and the variable responses of
 native plants in ecosystems to UV-B exposures results in changes in interactions between various
 plants species, changes between plants and other organisms, and between plants and their abiotic
 environment. These preliminary studies suggest that in natural ecosystems, composed of many
 different plant species, with complex interactions between plants and between plants and other
 organisms, there may develop effects  of UV-B that cannot be determined from experiments on
 single plant species. The effects of UV-B on natural plant systems, therefore, should be of
 greater concern than on agricultural crops (Bjorn, 1996).

Nitrogen Deposition Effects
     Nitrogen has long been recognized as the nutrient most important for plant growth. Plants
usually absorb nitrogen through their roots by absorbing NH4 + or NO3" or informed by symbiotic
organisms in the roots. Plants, however, vary in their ability to absorb ammonium and nitrate
(Chapin et al., 1987). Nitrogen is of overriding importance in plant metabolism and, to a large
extent, governs the utilization  of phosphorus, potassium, and other nutrients.  Most of the
       March 2001
                                          4-41
DRAFT-DO NOT QUOTE OR CITE

-------
 1      nitrogen in soils is associated with organic matter. Typically, the availability of nitrogen via the
 2      nitrogen cycle controls net primary productivity and possibly the decomposition rate of plant
 3      litter. Photosynthesis is influenced by nitrogen uptake in that ca. 75% of the nitrogen in a plant
 4      leaf is used during the process of photosynthesis. The nitrogen-photosynthesis relationship is,
 5      therefore, critical to the growth of trees and other plants (Chapin et al, 1987).
 6            Because nitrogen is not readily available and is usually in shortest supply, it is the chief
 7      element in agricultural fertilizers. Atmospherically deposited nitrogen also can act as a fertilizer
 8      in soil low in nitrogen. Not all plants, however, are capable of utilizing extra nitrogen. Inputs of
 9     nitrogen to natural ecosystems that alleviate deficiencies and increase growth of some plants can
10     impact competitive relationships and alter species composition and diversity (Ellenberg, 1987;
11      Kenk and Fischer, 1988; U.S. Environmental Protection Agency, 1993).
12           The impact of increasing nitrogen inputs (e.g., NOX, nitrates, nitric acid) on the nitrogen
13     cycle and forests, wetlands, and aquatic ecosystems is discussed in detail elsewhere (U.S.
14     Environmental Protection Agency, 1993,1997a; Garner, 1994; World Health Organization,
15      1997).  The most important effects of nitrogen deposition are accumulation of nitrogen
16     compounds resulting in the enhanced availability of nitrate or ammonium, soil-mediated effects
17     of acidification, and increased susceptibility to stress factors (Bobbink et al., 1998). A major
18     concern is "nitrogen saturation", the result of the deposition of large amounts of particulate
19     nitrates. Nitrogen saturation results when additions to soil background nitrogen (nitrogen
20     loading) exceed the capacity of plants and soil microorganisms to utilize and retain nitrogen
21      (Aber et al., 1989,1998; Garner, 1994; U.S. Environmental Protection Agency, 1993). Under
22     these circumstances, ecosystems become unable to utilize excessive nitrogen inputs and
23      disruptions of ecosystem functioning may result (Hornung and Langan,  1999).
24           Growth of most forests in North America is limited by the nitrogen supply.  Severe
25      symptoms of nitrogen saturation, however, have been observed in high-elevation, nonaggrading
26      spruce-fir ecosystems in the Appalachian Mountains, as well as in the eastern hardwood
27      watersheds at Fernow Experimental Forest near Parsons, WV. Mixed conifer forests and
28      chaparral watersheds with high smog exposure in the Los Angeles Air Basin also  are nitrogen
29      saturated and exhibit the highest stream water NO3" concentrations for wildlands in North
30      America (Bytnerowicz and Fenn, 1996; Fenn et al., 1998). Not all forest ecosystems react in the
31      same manner to nitrogen deposition.  High-elevation alpine watersheds in the Colorado Front
        March 2001
4-42
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
31
 Range and a deciduous forest in Ontario, Canada, also are naturally saturated even though
 nitrogen deposition has been moderate (=8 kg.ha ha'1 .year1). The Harvard Forest hardwood
 stand in Massachusetts, however, has absorbed >900 kg N/ha without significant NO3' leaching
 during a nitrogen amendment study of 8 years (Table 4-5; Fenn et al.,1998). Johnson et al.
 (199 la) reported that measurements showing the leaching of nitrates and aluminum (A13+) from
 high elevation forests in the Great Smoky Mountains indicates that these forests have reached
 saturation.
     Possible ecosystem responses to nitrate saturation, as postulated by Aber and coworkers
 (Aber et al., 1989), include a permanent increase in foliar nitrogen and reduced foliar phosphorus
 and lignin caused by the lower availability of carbon, phosphorus, and water; reduced
 productivity in conifer stands because of disruptions of physiological function; decreased root
 biomass and increased nitrification and nitrate leaching;  and (4) reduced soil fertility, resulting
 from increased cation leaching, increased nitrate and aluminum concentrations in streams, and
 decreased water quality. Saturation implies that some resource other than nitrogen is limiting
 biotic function.
     Water and phosphorus for plants and carbon for microorganisms are the resources most
 likely to be the secondary limiting factors. The appearance of nitrogen in soil solution is an  early
 symptom of excess nitrogen. In the final stage, disruption of forest structure becomes visible
 (Garner, 1994).
     Changes in nitrogen supply can have a considerable impact on an ecosystem's nutrient
 balance (Waring, 1987). Large chronic additions of nitrogen influence normal nutrient cycling
 and alter many plant and soil processes involved in nitrogen cycling (Aber et al., 1989).
Among the processes affected are (1) plant uptake and allocation, (2) litter production,
(3) immobilization (includes ammonification [the release of ammonia] and nitrificatrion
 [conversion of ammonia to nitrate during decay of litter and soil organic matter]), and (4) nitrate
leaching and trace gas emissions (Figure 4-2; Aber et al., 1989).
     Subsequent studies have shown that, although initially, there was an increase in nitrogen
mineralization (i.e., the conversion of soil organic matter to nitrogen in available form [see item
3 above]), nitrogen mineralization rates were reduced under nitrogen-enriched conditions. Also,
studies suggest that, during saturation, soil microbial communities change from predominantly
fungal (mycorrhizal) communities to those dominated by bacteria (Aber et al., 1998).
        March 2001
                                          4-43
DRAFT-DO NOT QUOTE OR CITE

-------
  TABLE 4-5. NITROGEN-SATURATED FORESTS IN NORTH AMERICA, INCLUDING
                                ESTIMATED N INPUTS AND OUTPUTS	
 Location
Forest Type
 Elevation   N Input         N Output
   (m)      (kg-ha"' -year')    (kg-ha"' -year'1)
                                                                                               Reference
Adirondack Mts.
northeastern New York
Catskill Mts.,
southeastern New York
Turkey Lakes Watershed,
Ontario, Canada

Whitetop Mt.,
southwestern Virginia

Northern hardwoods or '
hardwood/conifer mix
Mainly hardwood;
some eastern hemlock
Sugar maple and yellow
birch

Red spruce


396-661

,335-675

350-400


1650


9.3",

10,2s

7.0-7.7 (as
throughfall)

32C


Stage 1 N
loss"
Stage 1 and 2
N loss"
17.9-23.6


47C


Driscoll and
Van Dreason (1993)
Stoddard (1994)

Foster etal. (1989)
Johnson and
Lindberg(1992a)
Joslin and Wolfe
(1992)
Joslin etal. (1992)
 Fernow, West Virginia


 Great Smoky Mts.
  National Park, Tennessee
 Great Smoky Mts.
  National Park, Becking
  Site, North Carolina
 Great Smokey Mts.
  National Park, Tower
  Site, North Carolina
 Front Range, Colorado


 San'Dimas, San Gabriel
  Mts. southern California
 Camp Paivika,
  San Bemadino Mts.,
  southern California
 Klamath Mts,
  northern California
 Thompson Forest, Cascade
  Mts Washington
Mixed hardwood


American beech

Red spruce


Red spruce
Alpine tundra,
 subalpine conifer

Chapparral and
 grasslands
Mixed conifer
Western coniferous

Red alder
                                                   735-870      15-20
1600

1800


1740
            3.1",

            10.3d


            26.6
3000-4000    7.5-8.0


580-1680     23.3"

1600         30
NA
6.1


2.9

19.2


20.3


7.5


0.04-19.4

7-26f


NA8
             Mainly
             geologic8
220          4.7 plus >, 100   38.9
             as N, fixation
Gilliametal. (1996)
Peterjohn etal. (1996)

Johnson and Lindberg
(1992b)
Johnson etal. (199 la)
Johnson etal. (1991a)


Williams etal. (1996)


Riggan etal. (1985)

Fenn etal. (1996)

      '  !   • •
Dahlgren (1994)

Johnson and Lindberg
(1992b)	'
 "Estimated total N deposition from wet deposition data is from Driscoll et al. (1991) for the Adirondacks, and from Stoddard and
  Murdoch (1991) for the Catskills. Total deposition was estimated based on the wet deposition/total N deposition ratio (0.56) at
  Huntington Forest in the Adirondacks (Johnson and Lindberg, 1992b). Nitrogen deposition can be higher in some areas, especially
  at high-elevation sites such as Whiteface Mountain (15.9 kg-ha-'-year'1; Johnson and Lindberg, 1992b).
 'Stage 1 and 2 of N loss according to the watershed conceptual model of Stoddard (1994). Nitrogen discharge (kg-ha"'-year') data
  are not available; only stream water NO3" concentration trend data were collected.              ,
 ^Values appear high compared to other sites, especially N leaching losses. Joslin and Wolfe (1992) concede that "there is
  considerable uncertainty associated with the estimates of atmospheric deposition and leaching fluxes." However, elevated NO3"
  concentrations in soil solution, and lack of a growth response to N fertilization (Joslin and Wolfe, 1994) support the hypothesis that
  the forest at Whitetop Mountain is N saturated.
 •"Estimated total N deposition from throughfall data. Total deposition was estimated based on the throughfall/total N deposition
  ration (0.56) from the nearby Smokies Tower site (Johnson and Lindberg, 1992b).
 'Annual throughfall deposition to the chaparral ecosystem.
 'Nitrogen output is from unpublished streamwater data (M.E. Fenn and M.A. Poth). The low value represents a year of average
  precipitation, and the high value is for 1995, when precipitation was nearly double the long-term average.  Nitrogen output
  includes N export in streamwater and to groundwater.
 'Annual input and output data are not known, although N deposition in this forest'is probably typical for much of the rural western
  United States (2-3 kg N-ha^-year'1 (Young et al., 1988). Excess N is from weathering of ammonium in mica schist bedrock. The
  ammonium was rapidly nitrified, leading to  high NO3- concentrations in soil solution (Dahlgren,  1994).
March 2001
                          4-44
                 DRAFT-DO NOT QUOTE OR CITE

-------
                Nitrogen
                Oxides in
              Atmosphere
                          Gaseous
                         Nitrogen in
                         Atmosphere
                                                Photosynthesis
                                  Plant
                                Utilization
                                                          Animal
                                                          Proteins
                                                                                  Trace
                                                                                   Gas
                                                                                Emissions
                                            Litter
                                         Production
                                           (Death)
                        Nitrogen
                        Fixation
                                                         Microbial
                                                      Decomposition
                  Nitrates
                  in Soil
                              Microbial
                              Utilization
                                                                                  Urea
                                                                               (Ammonia)
                Nitrates in
                 Streams,
              Rivers, Lakes
               and Oceans
                                           Nitrites
                                             by
                                          Bacteria
                                                                —B	Process altered by
                                                                          nitrogen saturation

       Figure 4-2.  Nitrogen cycle (dotted lines indicate processes altered by nitrogen satuation).

       Source: Garner (1994).
1

2

3
     The availability of nutrients is an important factor in determining species composition, and

nitrogen is usually the growth-limiting nutrient (Bobbink, 1998).  Most of the plants growing iti

nutrient-poor habitats have become adapted to them over time and can only compete successfully
       March 2001
                                           4-45
DRAFT-DO NOT QUOTE OR CITE

-------
 1     on soils low in nitrogen (Bobbink, 1998; Chapin, 1991). All plants growing in low resource
 2     environments (e.g., infertile soil, shaded understory, deserts, tundra) have been observed to have
 3     certain similar characteristics:  a slow growth rate, low photosynthetic rate, and low capacity for
 4     nutrient uptake.  An important feature to plants adapted to low-resource environments is that they
 5     grow slowly and tend to respond less, even when provided with an optimal supply and balance of
 6     resources (Pearcy et al., 1987; Chapin, 1991).  Plants adapted to cold, moist environments grow
 7     more leaves than roots as the relative availability to nitrogen increases; however, other nutrients
 8     may soon be limiting.  The capacity of gymnosperms in general, and in subalpine and boreal
 9     species in particular, to reduce nitrates in either roots or leaves appears to be limited. In addition,
10     the ability of trees to use nitrogen varies with the age of the tree and the density of the stand
11     (Waring, 1987).
12           Because the  competitive equilibrium of plants in any community is finely balanced, the
13     alteration of one of a number of environmental parameters, (e.g., continued nitrogen additions),
14     can change the vegetation structure of an ecosystem (Bobbink, 1998; Skeffington and Wilson,
15     1988). Increases in soil nitrogen play a selective role. When nitrogen becomes more readily
16     available, plants adapted to living in an environment of low nitrogen availability will be replaced
17     by plants capable of using increased nitrogen because they have a competitive advantage.
1 g           The long-term impacts of increased nitrogen deposition have been studied in several
19     western and central European plant communities: lowland heaths, species-rich grasslands,
20     mesotrophic fens, ombrotrophic bogs, upland moors, forest-floor vegetation, and freshwater
21     lakes (Bobbink, 1998). Large changes in species composition have been observed in regions
22     with high nitrogen loadings or infield experiments after years of nitrogen addition (Bobbink
23     et al., 1998). The increased input of nitrogen gradually increased availability of nitrogen in the
24      soil, and its retention because of low rates  of leaching and denitrification resulted in faster litter
25      decomposition and rate of mineralization.  Faster growth and greater height of nitrophilic species
26      enables these plants to shade out the slower growing species, particularly those in oligotrophic or
27      mesotrophic conditions (Bobbink, 1998; Bobbink et al., 1998). Excess nitrogen inputs to
28      unmanaged heathlands in the Netherlands has resulted in nitrophilous grass species replacing
29      slower growing heath species (Roelofs et al., 1987; Garner, 1994). Van Breemen and Van Dijk
30      (1988) noted that  over the past several decades the composition of plants in the forest herb layers
        March 2001
4-46
DRAFT-DO NOT QUOTE OR CITE

-------
  1      has been shifting toward species commonly found on nitrogen-rich areas. It also was observed
  2      that the fruiting bodies of mycorrhizal fungi had decreased in number.
  3           Other studies in Europe point out the effects of excessive nitrogen deposition on mixed-oak
  4      forest vegetation along a deposition gradient largely controlled by soil acidity, nitrogen supply,
  5      canopy composition, and location of sample plots (Brunet et al., 1998; Falkengren-Grerup,
  6      1998).  Results of the study, using multivariate methods, suggest that nitrogen deposition has
  7      affected the field-layer vegetation directly by increased nitrogen availability and, indirectly, by
  8      accelerating soil acidity. Time series studies indicate that 20 of the 30 field-layer species
  9      (nonwoody plants) that were associated most closely with high nitrogen deposition increased in
 10      frequency in areas with high nitrogen deposition during the past decades.  Included in the field-
 11      layer species were many generally considered nitrophilous; however, there were several acid
 12      tolerant species (Brunet et al,  1998). Falkengren-Grerup (1998), in an experimental study
 13      involving 15 herbs and 13 grasses, observed that species with a high nitrogen demand and a
 14      lesser demand for other nutrients were particularly competitive in areas with acidic soils and high
 15      nitrogen deposition. The grasses grew better than herbs with the addition of nitrogen. It was
 16      concluded that, at the highest nitrogen deposition, growth was limited for most species by the
 17      supply of other nutrients; and, at the intermediate nitrogen concentration, the grasses were more
 18      efficient than the herbs in utilizing nitrogen. Nihlgard (1985) suggested that excessive nitrogen
 19      deposition may contribute to forest decline in other specific regions of Europe. Also, Schulze
20      (1989), Heinsdorf (1993), and Lamersdorf and Meyer (1993) attribute magnesium deficiencies in
21      German forests, in part, to excessive nitrogen deposition.
22           Plant succession patterns and biodiversity are affected significantly by chronic nitrogen
23      additions in some North American ecosystems (Figure 4-3). Fenn et al. (1998) report that
24      long-term nitrogen fertilization studies in both New England and Europe, as well, suggest that
25      some forests receiving chronic inputs of nitrogen may decline in productivity and experience
26      greater mortality. Long-term fertilization experiments at Mount Ascutney, Vermont, suggest that
27      declining coniferous forest stands with slow nitrogen cycling may be replaced by deciduous
28      fast-growing forests that cycle nitrogen rapidly (Fenn et al., 1998).
29           In experimental studies of nitrogen deposition conducted by Wedin and Tilman (1996) over
30      a 12-year period on Minnesota grasslands, plots dominated by native warm-season grasses
31      shifted to low-diversity mixtures dominated by cool-season grasses at all but the  lowest rates of
        March 2001
4-47
DRAFT-DO NOT QUOTE OR CITE

-------
                                                                            N-Saturated Ecosystems
                                                                            in North America
Review of Ecosystem Effects
and Responses to Excess N
                               1. Nitrogen Inputs:                     •
                                 > Atmospheric deposition, N2 fixation, fertilization
                                 Nitrogen Retention:
                                 >• In plant biomass and soil organic matter
                                 > The role of soil microbes and woody residues
                                 > Abiotic retention
                                 Nitrogen Outputs:
                                 >-Hydrologic transport, gaseous emissions from soil
                                 >Removal in harvest, fire emissions, and soil erosion
           ' 2. Characteristics Predisposing Forests
              to N Saturation:
              >• Stand vigor and succession, forest type
              >• Previous land use-stand history
              ••Soil N accumulation
              > Topography and climate
              > Nitrogen deposition
3. Ecosystem Responses to Excess Nitrogen:
  >• Nitrate leaching and export
  > Eutrophicationof estuaries
  >• Toxicity of surface waters
  > Foliar nutrient responses
  >• Nitrogen mineralization and nitrification
  >• Effects on soil organic matter
  >• Soil acidification, cation depletion, Al toxicity
  >• Foliar nutrient responses
  > Greenhouse gas fluxes
                                4. Regional N Saturation Conceptual Models:
                                  >• New England forests
                                  >• California forests
                                  > Colorado alpine ecosystems
Figure 4-3. Diagrammatic overview of excess nitrogen (N) in North America.


Source: Fennelal.(1998).
March 2001
4-48
DRAFT-DO NOT QUOTE OR CITE

-------
  1      nitrogen addition.  Grasslands with high nitrogen retention and carbon storage rates were the
  2      most vulnerable to loss of species and major shifts in nitrogen cycling.  The shift to low-diversity
  3      mixtures was associated with the decrease in biomass carbon to nitrogen (C:N) ratios, increased
  4      nitrogen mineralization, increased soil nitrate, high nitrogen losses, and low carbon storage
  5      (Wedin and Tilman, 1996). Naeem et al. (1994) experimentally demonstrated (under controlled
  6      environmental conditions) that loss of biodiversity, in addition to loss of genetic resources, loss
  7      of productivity, loss of ecosystem buffering against ecological perturbation, and loss of aesthetic
  8      and commercially valuable resources, also may alter or impair ecosystems services.
  9           The C:N ratio of the forest floor can be changed by nitrogen deposition over time. This
 10      change appears to occur when the ecosystem becomes nitration saturated (Gundersen et al.,
 11      1998a). Long-term changes in C:N status have been documented in Central Europe and indicate
 12      that nitrogen deposition has changed the forest floor. In Europe, low C:N ratios coincide with
 13      high deposition regions (Gundersen et al,, 1998a).  A strong decrease in forest floor root biomass
 14      has been observed with increased nitrogen availability. Roots and the associated rnycorrhizae
 15      appear to be an important factor in the accumulation of organic matter in the forest floor at
 16      nitrogen limited sites.  If root growth and mycorrhizal formation are impaired by nitrogen
 17      deposition, the stability of the forest floor may be affected by stimulating turnover and decreasing
 18      the root litter input to the forest floor and thus decrease the nitrogen that can be stored in the
 19      forest floor pool (Gundersen et al., 1998b). Nitrogen-limited forests have a high capacity for
20      deposited nitrogen to be retained by the plants and microorganisms competing for available
21      nitrogen (Gundersen et al., 1998b).  Nitrate leaching has been correlated significantly with nitrate
22      status but not with nitrate depositions. Forest floor C:N ratio has been used as a rough indicator
23      of ecosystem nitrogen status in mature coniferous forests and the risk of nitrate leaching;
24      analyses of European databases indicated an empirical relationship between forest floor C:N ratio
25      and nitrate leaching (Gundersen et al., 1998a).  Nitrate leaching was observed when the
26      deposition received was more than 10kg N/ha.  All of the data sets supported a threshold at
27      which nitrate leaching seems to increase at a C:N ratio of 25. Therefore, to predict the rate of
28      changes in nitrate leaching it is necessary to be able to predict the rate of changes in the forest
29      floor C:N ratio. Understanding the variability in forest ecosystem response to nitrogen input is
30      essential in assessing pollution risks (Gundersen et al., 1998a).
        March 2001
4-49
DRAFT-DO NOT QUOTE OR CITE

-------
 1          The plant root is an important region of nutrient dynamics. The rhizosphere includes the
 2     soil that surrounds and is influenced by plant roots (Wall and Moore, 1999).  The mutualistic
 3     relationship between plant roots, fungi, and microbes is critical for the growth of the organisms
 4     involved. The plant provides shelter and carbon, whereas the symbiont provides access to a
 5     limiting nutrient such as nitrogen and phosphorus. As indicated above, changes in soil nitrogen
 6     influence the mycorrhizal-plant relationship. Mycorrhizal fungal diversity is associated with
 7     above-ground plant biodiversity, ecosystem variability, and productivity (Wall and Moore, 1999).
 8     Aber et al. (1998) showed a close relationship between mycorrhizal fungi and the conversion of
 9     dissolved inorganic nitrogen to soil nitrogen. During nitrogen saturation, soil microbial
10     communities change from being fungal, and probably being dominated by mycorrhizae, to being
11     dominated by bacteria. The loss of mycorrhizal function has been hypothesized as the key
12     process leading to increased nitrification and nitrate mobility.  Increased nitrate mobility leads to
13     increased cation leaching and soil acidification (Aber et al., 1998).
14           The interrelationship of above- and below-ground flora is illustrated by the natural invasion
15     of heath lands by oaks (Quercus robur).  Soils are dynamic entities, the features of which can
16     change like the rest of the ecosystem with age and management. The soil-forming factors under
17     the heath have been vegetation type during the last 2000 years, whereas the invasion by oaks has
18     been taking place for only a few decades. Clear changes in the ground floor and soil morphology
19     takes place when trees colonize heath (Nielsen et al., 1999). The distribution of roots also
20     changed under the three different vegetation types. Under both heather and the Sitka spruce
21     plantation, the majority of roots are confined to the uppermost horizons, whereas under oak, the
22     roots are distributed more homogeneously. There was also a change in the C:N ratio when
23     heather was replaced by oaks.  Also, the spontaneous succession of the heath by oaks changed
24     the biological nutrient cycle into a deeper vertical cycle, when compared to the heath where the
25     cycle is confined to the upper soil horizons. Soils similar to those described in this study
26     (Jutland, Denmark), with mainly an organic buffer system, seem to respond quickly to changes in
27     vegetation (Nielsen et al., 1999).
28           In addition to excess nitrogen deposition effects on terrestrial ecosystems of the types noted
29     above (e.g., dominant species shifts and other biodiversity impacts), direct atmospheric nitrogen
30     deposition and increased nitrogen inputs via runoff into streams, rivers, lakes, and oceans can
31     have notable impacts on aquatic ecosystems as well. One illustrative example is recently
        March 2001
4-50
DRAFT-DO NOT QUOTE OR CITE

-------
  1      reported research (summarized by Paerl et al., in press) characterizing impacts of nitrogen
  2      deposition on the Pamlico Sound, NC, estuarine complex, which serves as a key fisheries nursery
  3      supporting an estimated 80% of commercial and recreational finfish and shellfish catches in the
  4      southeastern U.S. Atlantic coastal region.  Such direct atmospheric nitrogen deposition onto
  5      waterways feeding into the Pamlico Sound or onto the sound itself and indirect nitrogen inputs
  6      via runoff from upstream watersheds contribute to conditions of severe water oxygen depletion,
  7      formation of algae blooms in portions of the Pamlico Sound estuarine complex, and altered fish
  8      distributions, catches, and physiological states and incidence of disease.  Under extreme
  9      conditions of especially high rainfall rate events (e.g., hurricanes) affecting watershed areas
 10      feeding into the sound, the effects of nitrogen runoff (in combination with excess loadings of
 11      metals or other nutrients) can be massive (e.g., creation of the widespread "dead-zone" affecting
 12      large areas of the Pamlico Sound for many months after hurricane Fran in 1996 and hurricanes
 13      Dennis, Floyd, and Irene in 1999 impacted eastern North Carolina).
 14
 15      Sulfur Deposition Effects
 16           Sulfur is an essential plant nutrient and, as such, is a major component of plant proteins.
 17      The most important source of sulfur is sulfate taken up from the soil by plant roots even though
 18      plants can utilize atmospheric SO2 (Marschner, 1995). The availability of organically bound
 19      sulfur in soils depends largely on microbial decomposition, a relatively slow process. The major
20      factor controlling the movement of sulfur from the soil into vegetation is the rate of release from
21      the organic to the inorganic compartment (May et al., 1972; U. S. Environmental Protection
22      Agency, 1982; Marschner, 1995).  Sulfur plays a critical role in agriculture as an essential
23      component of the balanced fertilizers needed to grow and increase worldwide food production
24      (Ceccotti and Messick, 1997). Atmospheric deposition is an important component of the sulfur
25      cycle. This is true not only in polluted areas where atmospheric deposition is very high, but also
26      in areas of low sulfur input.  Additions of sulfur into the soil in the form of SO4 2" could alter the
27      important organic-sulfur/organic-nitrogen relationship involved in protein formation in plants.
28      The biochemical relationship between sulfur and nitrogen in plant proteins indicates that neither
29      element can be assessed adequately without reference to the other. There is a regulatory coupling
30      of sulfur and nitrogen metabolism.  Sulfur deficiency reduces nitrate reductase and, to a similar
31      extent, also glutamine synthetase activity. Nitrogen uptake in forests, therefore, may be loosely
       March 2001
4-51
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
regulated by sulfur availability, but sulfate additions in excess of needs do not necessarily lead to
injury (Turner and Lambert, 1980; Hogan et al., 1998).
     Only two decades ago, there was little information comparing sulfur cycling in forests with
nutrients, especially nitrogen.  With the discovery of deficiencies in some unpolluted regions
(Kelly and Lambert, 1972; Humphreys et al., 1975; Turner et al., 1977; Schnug, 1997) and
excesses associated with acidic deposition in other regions of the world (Meiwes and Khanna,
1981; Shriner and Henderson, 1978; Johnson et al., 1982a,b), interest in sulfur nutrition and
cycling in forests has heightened. General reviews of sulfur cycling in forests have been written
by Turner and Lambert (1980), Johnson (1984), Mitchell et al. (1992a,b), and Hogan et al.
(1998). The salient elements of the sulfur cycle as it may be affected by changing atmospheric
deposition are summarized by Johnson and Mitchell (1988). Sulfur has become the most
important limiting factor in European agriculture because of the desulfurization of industrial
emissions (Schnug, 1997).
     Most studies dealing with the impacts of sulfur deposition on plant communities have been
conducted in the vicinity of point sources and have investigated above-ground effects of SO2 or
acidifying effects of sulfate on soils (Krupa and Legge, 1998; Dreisinger and McGovern, 1970;
Legge, 1980; Winner and Bewley, 1998a,b; Laurenroth and Michunas, 1985; U.S. Environmental
Protection Agency, 1982). Krupa and Legge (1986), however, observed a pronounced increase
in foliar sulfur concentrations in all age classes of needles of the hybrid pine lodgepole x jack
pine (Finns contorta x P. banksiana). This vegetation had been exposed to chronic low
concentrations of sulfur gas pollution (SO2), hydrogen sulfide (H2S), and fugitive sulfur aerosol
for more than 20 years.  Observations under the microscope showed no sulfur deposits on the
needle surfaces and led to the conclusion that the sulfur was derived from the soil.  The oxidation
of elemental sulfur and the generation of protons is well known for the soils of Alberta, CN.
This process is mediated by bacteria of the Thiobacillus sp. As elemental sulfur gradually is
converted to protonated SO4, it can be leached downward and readily taken up by plant roots.
The activity of Thiobacillus sp. is stimulated by elemental sulfur additions (Krupa and Legge,
 1986).
        March 2001
                                          4-52
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 Effects of Acidic Deposition on Forest Soils
      Substantial and previously unsuspected changes in soils are occurring both in polluted areas
 of eastern North America, the United Kingdom, Sweden, and Central Europe and in less polluted
 regions of Australia and western North America (reviewed by Johnson et al., 1991b; see review
 by Huntington, 2000).  In some cases, trends are toward more acidic soils (e.g., Markewitz et al.,
 1998), and, in others, there are no consistent trends, with some soils showing increases and some
 showing decreases at different sampling times, and some showing no change (e.g., Johnson and
 Todd, 1998; Trettin et al.,  1999; Yanai et al., 1999).
      Significant changes have occurred at many sites in the eastern United States during recent
 decades.  Temporal trends in tree ring chemistry were examined as indicators of historical
 changes in the chemical environmental of red spruce. Chemical changes in tree-ring chemistry
 reflect changing inputs of regional pollutants to forests. If significant base cation mobilization
 and depletion of base cations from eastern forest soils has occurred, a temporal sequence of
 changes in uptake patterns and possibly in tree growth would be expected. Patterns of tree-ring
 chemistry principally at high-elevation sites in the eastern United States, leads to the conclusion
 that significant changes in  soil chemistry have occurred in many of these sites during recent
 decades leading to changes in growth (Bondietti and McLaughlin, 1992).  These changes are
 spatially and temporally consistent with emissions of SO2 and NO2 across the region, suggesting
 that increased acidification of forest soils has occurred.
      Increases in levels of Al and Fe typically occur as base cations are removed from soils by
 tree uptake. A region-wide Ca increase above expected levels followed by a  decrease suggests
 that increased mobilization began perhaps  30 to 40 years ago (Bondietti and McLaughlin, 1992).
 The period of Ca mobilization coincides with a region-wide increase in growth rate of red spruce,
 whereas the period of decreasing levels of Ca in wood corresponds temporally with patterns of
 decreasing radial growth at high elevation sites throughout the region during the past 20 to 30
years. The decline in wood Ca suggests that Ca  loss may have been increased to the point at
which base saturation of soils has, been reduced (Bondietti and McLaughlin, 1992).
      Studies by Shortle and Bondietti, (1992) support the view that changes in soil chemistry in
eastern North America forest sites occurred many decades ago "before anybody was looking".
Sulfur and nitrogen emissions began increasing in eastern North America in the 1920s and
continued to increase into the 1980s, when sulfur began to decrease but nitrogen emissions have
       March 2001
                                          4-53
DRAFT-DO NOT QUOTE OR CITE

-------
 1      not (Garner et al., 1989).  Shortle and Bondietti (1992) present evidence that, from the late 1940s
 2      into the 1960s, the mor humus layer of acid-sensitive forest sites in eastern North America
 3      underwent a significant change that resulted in the loss of exchangeable essential base cations
 4      and interrupted the critical base nutrient cycles between mature trees and the root-humus
 5      complex. The timing of the impact appears to have coincided with the period when the SOX and
 6      NOX emissions in eastern North America subject to long-range transport were increasing the most
 7      rapidly (See above; Shortle and Bondietti, 1992). Although forest ecosystems other than the
 8      high-elevation spruce-fir  forests are not currently manifesting symptoms of injury directly
 9      attributable to acid deposition, less sensitive forests throughout the United States are
10     experiencing gradual losses of base cation nutrient, which in many cases will reduce the quality
11      of forest nutrition over the long term (National  Science and Technology Council, 1998). In some
12     cases, it may not even take decades, because these forests already have been receiving sulfur and
13     nitrogen deposition for many years. The current status of forest ecosystems in different U.S.
14     geographic regions varies, as does their sensitivity to nitrogen and sulfur deposition.  Variation in
15     potential future forest responses or sensitivity are caused, in part, by differences in deposition of
16     sulfur and nitrogen, ecosystem sensitivities to sulfur and nitrogen additions, and responses of
17     soils to sulfur and nitrogen inputs (National Science and Technology Council, 1998).
lg           Acidic deposition has played a major role in recent soil acidification in some areas of
19     Europe and, to a more limited extent, eastern North America. Examples include the study by
20     Hauhs (1989) at Lange Bramke, Germany, which indicated that leaching was of major
21     importance in causing substantial reduction in soil-exchangeable base cations over a  10-year
22     period (1974-1984). Soil acidification and its effects result from the deposition of nitrate (NO3')
23     and sulfate (SO4 *') and the associated hydrogen (H +) ion.  The effects of excessive nitrogen
24     deposition on soil acidification and nutrient imbalances have been well established in Dutch
25     forests (Van Breemen et al., 1982; Roelofs et al., 1985; Van Dijk and Roelofs, 1988).
26     For example, Roelofs et  al. (1987) proposed that NH3 /NH4+ deposition leads to heathland
27      changes via two modes:  (1) acidification of the soil and the loss of cations K+, Ca2+,  and Mg2+;
28      and (2) nitrogen enrichment that results in "abnormal" plant growth rates and altered competitive
29      relationships. Nihlgard (1985) suggested that excessive nitrogen deposition may contribute to
30      forest decline in other specific regions of Europe.  Falkengren-Grerup (1987) noted that, during
        March 2001
4-54
DRAFT-DO NOT QUOTE OR CITE

-------
  1      about 50 years, unexpectedly large increases in growth of beech (Fagus sylvatica Z.) were
  2      associated with decreases in pH and exchangeable cations in some sites in southernmost Sweden.
  3           Likens et al. (1996) suggested that soils are changing at the Hubbard Brook Watershed,
  4      NH, because of a combination of acidic deposition and reduced base cation deposition. They
  5      surmised, based on long-term trends in stream-water data, that large amounts of Ca and Mg have
  6      been lost from the soil-exchange complex over a 30-year period from approximately 1960 to
  7      1990. The authors speculate that the declines in base cations in soils may be the cause of recent
  8      slowdowns in forest growth at Hubbard Brook. In a follow-up study, however, Yanai et al.
  9      (1999) found no significant decline in Ca and Mg concentrations in forest floors at Hubbard
10      Brook over the period 1976 to 1997.  They also found both gains and losses in forest floor Ca
11      and Mg between 1980 and 1990 in a regional survey.  Thus, they concluded that "Forest floors in
12      the region are not currently experiencing rapid losses of base cations, although losses may have
13      preceded the onset of these three studies."
14           Hydrogen ions entering a forest ecosystem first encounter the forest canopy, where they are
15      often exchanged for base cations that then appear in throughfall (Figure 4-4 depicts a model of
16      H+ sources and sinks). Base cations leached from the foliage must be replaced through uptake
17      from the soil, or foliage cations will be reduced by the amounts leached. In the former case, the
18      acidification effect is transferred to the soil, where H + is exchanged for a base cation at the
19      root-soil interface. Uptake of base cations or NH4 + by vegetation or soil microorganisms causes
20      the release of H + in order to maintain charge balance. Uptake of nutrients in anionic form (NO3',
21      SO4 2", PO4 3") causes the release of OH" in order to maintain charge balance. Thus, the net
22      acidifying effect of uptake is the difference between cation and anion uptake. The form of ions
23      taken up is known for all nutrients but nitrogen, where either NH4+ or NO3" can be taken up.
24      In that, nitrogen is a nutrient taken up in great quantities, the uncertainty in the ionic form of
25      nitrogen taken up creates great uncertainty in the overall H+ budget for soils (Johnson 1992).
26           The cycles of base cations differ from those of N, P, and S in several respects.  The fact that
27      Ca, K, and Mg exist primarily as cations in solution whereas N, P, and S exist primarily as anions
28      has major implications for the cycling of the nutrients and the effects of acid deposition on these
29      cycles. The most commonly accepted model of base cation cycling in soils is one in which base
30      cations are released by weathering of primary minerals to  cation exchange sites,  where they are
31      then available for either plant uptake or leaching (Figure 4-4).  The introduction  of H + by
       March 2001
4-55
DRAFT-DO NOT QUOTE OR CITE

-------
                                           Deposition
                                                                           so?
              Soil
            Organism
             Uptake
                                  2H++NO3
                                  Nitrification

                        CO2 + H2O
                        Carbonic Acid Formation


                        R-COOH
                        Organic Acid Formation
                                     2OH
                                     Soil
                                  Organism
                                   Uptake
                                                                        20H
                                                      Leaching


      Figure 4-4.  Schematic of sources and sinks of hydrogen ions in a forest (from Taylor et al.,
                  1994).
1     atmospheric deposition or by internal processes will impact directly the fluxes of Ca, K, and

2     Mg via cation exchange or weathering processes. Therefore, soil leaching is often of major

3     importance in cation cycles, and many forest ecosystems show a net loss of base cations

4     (Johnson, 1992a).

5          Two basic types of soil change are involved:  (1) a short-term intensity type change

6     resulting from the concentrations of chemicals in soil water, and (2) a long-term capacity change
      March 2001
4-56
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 based on the total content of bases, aluminum and iron stored in the soil (Reuss and Johnson,
 1986; Van Breemen, 1983). Changes in intensity factors can have a rapid impact on the
 chemistry of soil solutions. Increases in the amounts of sulfur and nitrogen in acidic deposition
 can cause immediate increases in acidity and mobilization of aluminum in soil solutions.
 Increased aluminum concentrations and an increase in the Ca/Al ratio in soil solution have been
 linked to a significant reduction in the availability of essential base cations to plants, an increase
 in plant respiration, and increased biochemical stress (National Science and Technology Council,
 1998).
      Rapid changes in intensity, resulting from the addition of increased amounts of nitrogen or
 sulfur in acidic deposition, can have a rapid impact on the chemistry of soil solutions by
 increasing the acidity and mobilizing aluminum. Increased concentrations of aluminum and an
 increase in the ratio of calcium-to-aluminum in soil solution have been linked to significantly
 reduced availability of essential cations to plants.
      Capacity changes are the result of many factors acting over long time periods. The content
 of base cations (calcium, magnesium, sodium, and potassium) in soils result from additions from
 the atmospheric deposition, decomposition of vegetation and geologic weathering. Loss of base
 cations may occur through plant uptake and leaching. Increased leaching of base cations may
 result in nutrient deficiencies in soils as has been happening in some sensitive forest ecosystems
 (National Science and Technology Council, 1998).
      A major concern has been that soil acidity would lead to nutrient deficiency.  Calcium is
 essential in the formation of wood and the maintenance of cells, the primary plant tissues
 necessary for tree growth. Trees obtain Ca from the soil, but to be taken up by roots, the Ca
 (a positively charged ion) must be dissolved in soil water (Lawrence and Huntington, 1999).
 Tree species may be adversely affected if high Al to nutrient ratios limit uptake of Ca and Mg
 and create a nutrient deficiency (Shortle and Smith, 1988; Garner, 1994). Acid deposition by
 lowering the pH of aluminum-rich soil can increase aluminum concentrations in soil water
through dissolution and ion-exchange processes. When in solution, aluminum can be taken up
by roots, transported through the tree and, eventually, deposited on the forest floor in leaves and
branches.  Aluminum is more readily taken up than is Ca because it has a higher affinity for
negatively charged surfaces than does Ca. When present in the forest floor, Al tends to displace
adsorbed Ca and causes it to be more readily leached. The continued buildup of Al in the forest
       March 2001
                                         4-57
DRAFT-DO NOT QUOTE OR CITE

-------
 1      floor layer, where nutrient uptake is greatest, can decrease the availability of Ca to the roots
 2      (Lawrence et al., 1995) and lower the efficiency of Ca uptake because Al is more readily taken up
 3      than is Ca 2* when the ratio of Ca to Al in soil water is less than one (Lawrence and Huntington,
 4      1999). A 1968 Swedish report to the United Nations postulated a decrease in forest growth of ca.
 5      1.5%/year as result of Ca 2+ loss by leaching (Johnson and Taylor, 1989). The concern that soil
 6      acidification and nutrient deficiency may result in forest decline remains extant today.
 7           Aluminum toxicity is a possibility in acidified soils. Atmospheric deposition (or any other
 8      source of mineral anions) can increase the concentration of Al, especially A13+, in soil solution
 9      without causing significant soil acidification (Johnson and Taylor, 1989). Aluminum can be
10     brought into soil solution in two ways: (1) by acidification of the soil and (2) by an increase in
11      the total anion and cation concentration of the soil solution. The introduction of mobile, mineral
12     acid anions to an acid soil will cause increases in the concentration of aluminum in the soil
13     solution, but extremely acid soils in the absence of mineral acid anions will not produce a
14     solution high in aluminum. An excellent review of the relationships among the most widely used
15     cation-exchange equations and their implications for the mobilization of aluminum into soil
16     solution is provided by Reuss (1983).
17  .         Aluminum toxicity may influence forest tree growth, where acid deposition and natural
18     acidifying processes increase soil acidity. Aluminum concentrations have been observed to
19     exhibit a strongly descending gradient from bulk soil through the rhizosphere to the root (Smith,
20      1990a).  Once it enters the forest tree roots, Al accumulates in root tissue (Thornton et al., 1987;
21     Vogt et al., 1987a,b). There is abundant evidence that Al is toxic to plants. Reductions of Ca
22     uptake by roots has been associated with increases in Al uptake (Clarkson and Sanderson, 1971).
23     Calcium plays a major role in cell membrane integrity and cell wall structure. A number of
24     studies have suggested that the toxic effect of aluminum on forest trees could be caused by Ca2*
25      deficiency (Shortle and Smith, 1988; Smith, 1990a).  Mature trees have a high Ca2+ requirement
26     relative to agricultural crops (Rennie,  1955). Shortle and Smith (1988) attributed the decline of
27      red spruce in eight stands across northern New England from Vermont to Maine to an imbalance
28      of Al3* and Ca2+ in the fine root environment. Aluminum in the soil solution reduces Ca uptake
29      by competing for binding sites in the cortex of fine roots. Reduction in Ca uptake suppresses
30      cambial growth and reduces the rate of wood formation (annual ring formation), decreases the
31      amount of functional sapwood and live crown, and predisposes trees to disease and injury from
        March 2001
4-58
DRAFT-DO NOT QUOTE OR CITE

-------
  1      stress agents when the functional sapwood becomes less than 25% of cross-sectional stem area
  2      (Smith, 1990a).
  3           Air pollution is not the sole cause of soil change. High rates of acidification are occurring
  4      in less polluted regions of the western United States and Australia because of internal soil
  5      processes, such as tree uptake of nitrate and nitrification associated with excessive nitrogen
  6      fixation (Johnson et al., 1991b). Many studies have shown that acidic deposition is not a
  7      necessary condition for the presence of extremely acid soils, as evidenced by their presence in
  8      unpolluted, even pristine forests of the northwestern United States and Alaska (Johnson et al.,
  9      1991b).  The soil becomes acidic when H+ ions attached to NH4+ or HNO3 remain in the soil after
10      nitrogen is taken up by plants.  For example, Johnson et al. (1982b) found significant reductions
11      in exchangeable K + over a period of only 14 years in a relatively unpolluted Douglas fir
12      Integrated Forest Study (IPS) site in the Washington Cascades. The effects of acid deposition at
13      this site were negligible relative to the effects of natural leaching (primarily carbonic acid) and
14      nitrogen tree uptake (Cole and Johnson, 1977). Even in polluted regions, numerous studies have
15      shown the importance of tree uptake of NH4+ and NCy in soil acidification. Binkley et al. (1989)
16      attributed the marked acidification (pH decline of 0.3 to 0.8 units and base saturation declines of
17      30 to 80%) of abandoned agricultural soil in South Carolina over a 20-year period to NH4+ and
18      NO3" uptake by a loblolly pine plantation.
19           An interesting example of uptake effects on soil acidification is that of Al uptake and
20      cycling (Johnson et al., 1991b). Aluminum accumulation in the leaves of coachwood
21      (Ceratopetalum apetalum) in Australia has been found to have a major impact on the distribution
22      and cycling of base cations (Turner and Kelly, 1981).  The presence of C. apetalum as a
23      secondary tree layer beneath brush cox (Lophostemon confertus) was found to lead to increased
24      soil exchangeable Al3+ and decreased soil exchangeable Ca 2+ (Turner and Kelly,  1981). The
25      constant addition of aluminum-rich litter fall obviously has had a substantial effect on soil
26      acidification, even if base cation uptake is not involved directly.
27           Given the potential importance of particulate deposition for base cation status of forest
28      ecosystems, the findings of Driscollet al. (1989) and Hedin et al. (1994) are especially relevant.
29      Driscoll et al. (1989) noted a decline in both SO42' and base cations in both atmospheric
30      deposition and stream water over the past two decades at Hubbard Brook Watershed, NH.  The
        March 2001
4-59
DRAFT-DO NOT QUOTE OR CITE

-------
 1     decline in SO42" deposition was attributed to a decline in emissions, and the decline in stream
 2     water SO42" was attributed to the decline in sulfur deposition.
 3          Hedin et al. (1994) reported a steep decline in atmospheric base cation concentrations in
 4     both Europe and North America over the past 10 to 20 years. The reductions in SO 2 emissions
 5     in Europe and North America in recent years have not been accompanied by equivalent declines
 6     in net acidity related to sulfate in precipitation. These current declines in sulfur deposition have,
 7     in varying degrees, been offset by declines in base cations and may be contributing "to the
 8     increased sensitivity of poorly buffered systems." Analysis of the data from the Integrated Forest
 9     Studies (IFS) supports the authors' contention that atmospheric base cation inputs may seriously
10     affect ecosystem processes. Johnson et al. (1994a) analyzed base cation cycles at the Whiteface
11     Mountain IFS site in detail and concluded that Ca losses from the forest floor were much greater
12     than historical losses, based on historical changes in forest floor Ca observed in an earlier study
13     (Johnson et al., 1994b). Further, the authors suggest that the difference between historical and
14     current net loss rates of forest floor Ca may be caused by sharply reduced atmospheric inputs of
15     calcium after about 1970 and exacerbated by sulfate leaching (U.S. Environmental Protection
16     Agency, 1999).
17  .        The calcium/aluminum molar ratio has been suggested as a valuable ecological indicator of
18     an approximate threshold beyond which the risk of forest injury from Al stress and nutrient
19     imbalances increases (Cronan and Grigal, 1995). The Ca/Al ratio also can be used as an
20     indicator to assess forest ecosystem changes over time in response to acidic deposition, forest
21     harvesting, or other process that contribute to acid soil infertility.  This ratio, however, may not
22     be a reliable indicator of stress in areas with both high atmospheric deposition of ammonium and
23     magnesium deficiency via antagonism involving ammonium rather than aluminum, and in areas
24     with soil solutions with calcium concentrations greater than 500 micromoles per liter (National
25     Science and Technology Council, 1998).  Cronan and Grigal (1995) based on a review of the
26     literature have made the following estimates for determining the adverse impact of acidic
27     deposition on tree growth or nutrition:
28           • forests have a 50% risk of adverse impacts if the Ca/Al ration is  1.0,
29           • the risk is 75% if the ratio is 0.5, and
30           • the risk approaches 100% if the ratio is 0.2.
        March 2001
4-60
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 The Ca/Al ratio of soil solution provides only an index of the potential for Al stress. Cronan and
 Grigal (1995), state that the overall uncertainty of the Ca/Al ratio associated with a given
 probability ratio is considered to be approximately ±50%. Determination of thresholds for
 potential forest impacts requires the use of the four successive measurement endpoints in the soil,
 soil solution, and plant tissue listed below.
      (1) Soil base saturation less than 15% of effective cation exchange capacity
      (2) Soil solution Ca/Al molar ratio less than 1.0 for 50% risk
      (3) Fine roots tissue Ca/Al molar ratio less than 0.2 for 50% risk
      (4) Foliar tissue Ca/Al molar ratio less than 12.5 for 50% risk
 The application of the Ca/Al ratio indicator for assessment and monitoring of forest health risks
 has been recommended for sites or in geographic regions where the soil base saturation <15%.

 Critical Loads
      In Europe, the critical load concept generally has been accepted as the basis for abatement
 strategies to reduce or prevent injury to the functioning and vitality of forest ecosystems caused
 by long-range transboundary .acidic deposition (Lokke, et al., 1996). The critical load has been
 defined as a "quantitative estimate of an exposure to one or more pollutants below which
 significant harmful effects on specified sensitive elements of the environment do not occur
 according to present knowledge" (Lokke et al., 1996).  A biological indicator, a chemical
 criterion, and a critical value are the elements used in the critical load concept. The biological
 indicator is the organism used to indicate the status of the receptor ecosystem, the chemical
 criterion is the parameter that results in harm to the biological indicator, and the critical value is
 the value of the chemical criterion below which no significant harmful response occurs to the
 biological indicator (Lokke et al., 1996). Trees, and sometimes other plants, are used as the
 biological indicators in the case of critical loads for forests.  The critical load calculation using
 the current methodology, is essentially an acidity/alkalinity mass balance calculation. The
 chemical criterion must be expressible in terms of alkalinity. Initially, the Ca/Al ratio was used,
but, recently, the (Ca+Mg+K)/Al ratio has been used (Lokke et al., 1996).
     Ideally, changes in acidic deposition should result in changes in the status of the biological
indicator used in the critical load calculation. However, the biological indicator is the integrated
response to a number of different stresses. Furthermore, there are other organisms more sensitive
        March 2001
                                          4-61
DRAFT-DO NOT QUOTE OR CITE

-------
 1     to acid deposition than trees. At high concentrations, Al3+ is known to be toxic to plants,
 2     inhibiting root growth and, ultimately, plant growth and performance (Lokke et al., 1996;
 3     National Science and Technology Council, 1998). Sensitivity to Al varies considerably between
 4     species and within species because of changes in nutritional demands and physiological status,
 5     which are related to age and climate. Experiments have shown that there are large variations in
 6     Al sensitivity, even among ecotypes.
 7           Mycorrhizal fungi as possible biological indicators have been suggested by Lokke et al.
 8     (1996) because they are intimately associated with tree roots, depend on plant assimilates, and
 9     play an essential role in plant nutrient uptake, influencing the ability of their host plants to
10     tolerate different anthropogenically generated stresses. Mycorrhizas and fine roots are an
11     extremely dynamic component of below-ground ecosystems and can respond rapidly to stress.
12     They have a relatively short life span, and their turnover appears to be strongly controlled by
13     environmental factors. Changes in mycorrhizal species composition or the loss of dominant
14     mycorrhizal species in areas where diversity is already low may lead to increased susceptibility of
15     plant to stress (Lokke et al., 1996). Stress affects the total amount of carbon fixed by plants and
16     modifies carbon allocation to biomass, symbionts and secondary metabolites. Because
17  .   mycorrhizal fungi are dependent for their growth on the supply of assimilates from the host
18     plants, stresses that shift the allocation of carbon reserves to the production of new leaves at the
19     expense of supporting tissues will be reflected rapidly in decreased fine root and mycorrhizzal
20     biomass (Winner and Atkinson, 1986).  The physiology of carbon allocation has also been
21     suggested as an indicator of anthropogenic stress (Andersen and Rygiewicz, 1991). Soil
22     dwelling animals are important for decomposition, soil aeration, and nutrient redistribution in the
23     soil. They contribute to decomposition and nutrient availability mainly by increasing the
24     accessibility of dead plant material to microorganisms.  Earthworms decrease in abundance and
25      in species number in acidified soils Lokke et al., 1996).
26
27     Biogeochemical Cycling—The Integrated Forest Study
28           The Integrated Forest Study (IPS) (Johnson and Lindberg, 1992a) has provided the most
29      extensive data set available on wet and dry deposition and the effects of deposition on the cycling
30      of elements in forest ecosystems. The overall patterns of deposition and cycling have been
31      summarized by Johnson and Lindberg (1992a), and the reader is referred to that reference for
        March 2001
4-62
DRAFT-DO NOT QUOTE OR CITE

-------
 1      details. The following is a summary of particulate deposition, total deposition, and leaching in
 2      the IPS sites.
 3           Particulate deposition in the IPS was separated at the 2-fj.m level; a decision was made to
 4      include total particulate deposition in this analysis and may include the deposition of particles
 5      larger than 10 //m.
 6           Particulate deposition contributes considerably to the total impact of base cations to most of
 7      the IPS sites.  On average, particulate deposition contributes 47% to total calcium deposition
 8      (range: 4 to 88%), 49% of total potassium deposition (range: 7 to 77%), 41% to total magnesium
 9      deposition (range: 20 to 88%), 36% to total sodium deposition (range:  11 to 63%), and 43% to
10      total base cation deposition (range: 16 to 62%). Of the total particulate deposition, the vast
11      majority (>90%) is >2 //m.
12           Figures 4-5 through 4-8 summarize the deposition and leaching of calcium, magnesium,
13      potassium, and total base cations for the IPS sites. As noted in the original synthesis (Johnson
14      and Lindberg, 1992a), some sites show net annual gains of base cations (i.e., total deposition
15      > leaching), some show losses (total deposition < leaching), and some are approximately in
16      balance. Not all cations follow the same pattern at each site.  For example, calcium shows net
17      accumulation  at the Coweeta, TN; Durham (Duke), NC; and Florida sites (Figure 4-5), potassium
18      shows accumulation at the Duke; Florida; Douglas-fir; red alder; Thompson, WA; Huntingdon
19      Forest, NY; and Whiteface Mountain, NY, sites (Figure 4-7),  and magnesium accumulated only
20      at the Florida sites (Figure 4-6). Only at the Florida site is there a clear net accumulation of total
21      base cations (Figure 4-8).
22           The factors affecting net calcium accumulation or loss include the soil-exchangeable cation
23      composition, as noted previously; base cation deposition rate; the total leaching pressure because
24      of atmospheric sulfur and nitrogen inputs, as well as natural (carbonic and organic) acids; and
25      biological demand (especially for potassium). At the Florida site, which has a very cation-poor,
26      sandy soil (an Ultic Haploquod derived from marine sand), the combination of all these factors
27      leads to net base cation accumulation from atmospheric deposition (Johnson and Lindberg,
28      1992a). The site showing the greatest net base cation losses, the red alder stand in Washington
29      state, is one that is under extreme leaching pressure by nitrate produced because of excessive
30      fixation by that species (Van Miegroet and Cole, 1984). hi the red spruce site in the Smokies,
31      the combined  effects of SO42" and NO3" leaching are even greater than hi the red alder site
        March 2001
4-63
DRAFT-DO NOT QUOTE OR CITE

-------
          1,000
           500
       CO
       s,
           -500
       JD
       1
       g. -1,000
       01
         -1,500 -


         -2,000
63%




41%
E
H




K < 2|jm
HI Wet
^ Leaching
Pe
55%





rcent c
65%




I
I





)f total
40%


1
[zJ





depos
26%
R
M





tion a:
25%





fl
I
I
I






> partic
49%

m



;les:
58%

I



88%

0



4%
IP
i





i



[69
|
1

>





CP   DL   GS    LP    FS
-   Warmer Sites
                                                DF   RA   NS
                                                     -> •* — —
         FF   MS   WF
        - Colder Sites—
ST
      Figure 4-5. Calcium deposition in >2-/^m particles, <2-/zm particles, and wet forms (upper
                 bars) and leaching (lower bars) in the Integrated Forest Study sites.
                 CP = Pinus strobus, Coweeta, TN; DL = Pinus taeda, Durham (Duke), NC;
                 GS = Pinus taeda, B. F. Grant Forest, GA; LP = Pinus taeda, Oak Ridge, TN;
                 FS = Pinus ettottii, Bradford Forest, FL; DF = Psuedotsuga menziesii,
                 Thompson, WA; RA = Alnus rubra; Thompson WA; NS = Picea abies,
                 Nordmoen, Norway; HF = northern hardwood, Huntington Forest, NY;
                 MS = Picea rubens, Rowland, ME; WF = Picea rubens, Whiteface Mountain,
                 NY; and ST = Picea rubens, Clingman's Dome, NC.
1     (Figure 4-9), but a considerable proportion of the cations leached from this extremely acid soil

2     consist of H1" and A13+ rather than of base cations (Johnson and Lindberg, 1992a). Thus, the red
3     spruce site in the Smokies is approximately in balance with respect to calcium and total base

4     cations, despite the very high leaching pressure at this site (Figures 4-5 and 4-8).
5  !        The relative importance of particulate base cation deposition varies widely with site and

6     cation and is not always related to the total deposition rate.  The proportion of calcium deposition

7     in particulate form ranges from a low of 4% at the Whiteface Mountain site to a high of 88% at

8     the Maine site (Figure 4-5). The proportion of potassium deposition as particles ranges from

9     7% at the Smokies site to 77% at the Coweeta  site (Figure 4-7), and the proportion of total base
      March 2001
                             4-64
DRAFT-DO NOT QUOTE OR CITE

-------
            1,000
             400
        v-   200
         CO
         CO
            -200
         CT
         UJ  -400
            -600 -
            -800
40%




^
1
k^d



49%

PS
I



[ | > 2 |jm
| < 2 |jm
^j V\fet
^ Leaching
P
44%






srcent
48%

Essl


I
I




of tota
36%




••M
y






depo
27%

PI
1



sition a
27%





PP
1
i
I





s parti
39%


1



cles:
46%


0



88%


i



20%


I
m




26%





§
i





                   CP    DL    GS    LP    FS    DF   RA    NS    FF    MS    WF   ST
                   —	  Warmer Sites   	*• *	Colder Sites	
        Figure 4-6. Magnesium deposition in >2-/an particles, <2-//m particles, and wet forms
                   (upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
                   See Figure 4-5 for legend.
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
cation deposition ranges from 16% at the Whiteface site to 62% at the Maine site (Figure 4-8).
Overall, participate deposition at the site in Maine accounted for the greatest proportion of
calcium, magnesium, potassium, and base cation deposition (88, 88,57, and 62%, respectively),
even though total deposition was relatively low. At some sites, the relative importance of
particulate deposition varies considerably by cation. At the Whiteface Mountain site, particulate
deposition accounts for 4, 20, and 40% of calcium, magnesium, and potassium deposition,
respectively. At the red spruce site in the Smokies, particulate deposition accounts for 46, 26%,
7% of calcium, magnesium, and potassium deposition, respectively.
     As observed in the IPS synthesis, SO42' and NOj leaching often are dominated by
atmospheric sulfur and nitrogen (Johnson and Lindberg, 1992a). The exceptions to this are in
cases where natural nitrogen inputs are high (i.e.,the nitrogen-fixing red alder stand), as are NOj
leaching rates, even though nitrogen deposition is low, and where soils adsorb much of the
       March 2001
                                         4-65
DRAFT-DO NOT QUOTE OR CITE

-------
        co
        Q>
       1
       S
400-
300-
200-
100-
0-
-100-
-200-
-300-
-400-
-500.
77%






76%

m
|
1





H< 2(jm
H! Wet
^ Leaching
Pe
61%


F1
^




jrcent
48%



i







3f total
54%


H





depos
40%


H
E




it on a:
40%





•
^
%






s partic
68%


N





;ulates
23%


n





57%


n
^




40%


H





7%







1
1
^
X^
^
^
I







                   CP
DL
 GS    LP    FS    DF   RA   NS    FF    MS    WF
Warmer Sites   	^ +—	Colder Sites-
                                                                                        ST
       Figure 4-7.  Potassium deposition in >2-/zm particles, <2-Atm particles, and wet forms
                   (upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
                   See Figure 4-5 for legend.
 1      atmospherically deposited SO42', thus reducing SO42" leaching compared to atmospheric sulfur
 2      input.
 3           Sulfate and NO3" leaching have a major effect on cation leaching in many of the IPS sites
 4      (Johnson and Lindberg, 1992a). Figure 4-9 shows the total cation leaching rates of the IPS sites
 5      and the degree to which cation leaching is balanced by SO42' + NO3" deposition. The SO42' and
 6      NO3' fluxes are subdivided further into that proportion potentially derived from particulate sulfur
 7      and nitrogen deposition (assuming no ecosystem retention, a maximum effect) and other sulfur
 8      and nitrogen sources (wet and gaseous deposition, internal production).
 9           As noted in the IPS synthesis, total SO42' and NO3" inputs account for a large proportion
10      (28 to 88%) total cation leaching in most sites. The exception is the Georgia loblolly pine site,
11      where there were high rates of HCO{ and Cl" leaching (Johnson and Lindberg, 1992a).  The role
12     of particulate sulfur and nitrogen deposition in this leaching is generally very small (<10%),
13      however, even if it is assumed that there is no ecosystem sulfur or nitrogen retention.
       March 2001
                       4-66
                              DRAFT-DO NOT QUOTE OR CITE

-------
             3,000

             2,000
        L    1,000
        OJ
            -1,000
        cr
            -2,000

            -3,000

            -4,000
53%


1

47%

S
1



| 	 | > 2 |jm
Jj < 2|jm
H Wet
^ Leaching
P
48%


1





ercent
62%

5S
I
n





of tote
49%
^^ ^^


1
1





1 depo
28%

H
1





sition e
28%





i
I
I
I





is part
47%

H
1



culate
44%


i



s:
62%


1



16%


1





31%





1
I
i






                    CP    DL    GS    LP    FS    DF
                 "*            Warmer Sites    —	
                                                         NS     FF   MS   WF    ST
                                                        	Colder Sites	>•
       Figure 4-8. Base cation deposition in >2-yum particles, <2-//m particles, and wet forms
                   (upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
                   See Figure 4-5 for legend.
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
     It was noted previously in this chapter that the contribution of particles to total deposition
of nitrogen and sulfur at the IPS sites is lower than is the case for base cations. On average,
particulate deposition contributes 18% to total nitrogen deposition (range: 1 to 33%) and 17%
to total sulfur deposition (range:  1 to 30%). Particulate deposition contributes only a small
amount to total ET deposition (average = 1%; range:  0 to 2%). (It should be noted, however,
that particulate H+ deposition in the >2-jum fraction was neglected.)
     Based on the IPS data, it appears that the particulate deposition has a greater effect on base
cation inputs to soils than on base cation losses associated with inputs of sulfur, nitrogen, and H+.
It cannot be determined what fraction of the mass of these particles are <10 yum, but only a very
small fraction is <2 /zm. These inputs of base cations have considerable significance, not only to
the base cation status of these ecosystems, but also to the potential of incoming precipitation to
acidify or alkalize the soils in these ecosystems. As noted above, the potential of precipitation to
       March 2001
                                         4-67
DRAFT-DO NOT QUOTE OR CITE

-------
        (0
            7,000
            6,000
            5,000
4,000 - •
                        % of total cation leaching balanced by SOf and NOs from Particles (P) and other (O) sources
                   P: 4 %
                   O:28%
               7%
              55%
1%
6%
 8%
78%
 9%
30%
10%
42%
                       Other Anions
                       Particulate Sulphur and Nitrogen
                       Other Sulphur and Nitrogen Sources
10%
88%
 8%
55%
 6%
73%
18%
69%
 1%
77%
 9%
69%
                                                                            MS    WF    ST
       Figure 4-9.  Total cation leaching (total height of bar) balanced by sulfate and nitrate
                   estimated from particulate deposition (assuming no ecosystem retention,
                   particulate sulfur and nitrogen) and by other sources (both deposition and
                   internal) of sulfate and nitrate (other sulfur and nitrogen sources) and by
                   other anions in the Integrated Forest Study sites. See Figure 4-5 for legend.
 1      acidify or alkalize soils depends on the ratio of base cations to H+ in deposition, rather than

 2     simply on the inputs of H+ alone.  In the case of calcium, the term "lime potential" has been

 3     applied to describe this ratio; the principle is the same with respect to magnesium and potassium.

 4     Sodium is a rather special case, in that it is a poorly absorbing cation, and leaching tends to

 5     balance input over a relatively short term.

 6          Net balances of base cations tell only part of the story as to potential effects on soils; these

 7     net losses or gains must be placed in the perspective of the soil pool size. One way to express

 8     this perspective is to simply compare soil pool sizes with the net balances. This comparison is

 9     made for exchangeable pools and net balances for a 25-year period in Figures 4-10 to 4-12.

10     It readily is seen that net leaching losses of cations pose no threat in terms of depleting

11     soil-exchangeable Ca2"1", K+, or magnesium ion within 25 years at the Coweeta, Duke, Georgia,
       March 2001
                                     4-68
                            DRAFT-DO NOT QUOTE OR CITE

-------
               350,000
                                                                            Soil Exchangeable
                                                                            (Dep-Leaching)*25
                       DL   GS   LP    FS
                             Warmer Sites
                                                                       HF   MS   WF   ST
                                                                         Colder Sites - >•
        Figure 4-10.  Soil exchangeable Ca2+ pools and net annual export of Ca2+ (deposition minus
                     leaching times 25 years) in the Integrated Forest Study sites. See Figure 4-5
                     for legend.
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
Oak Ridge, or Douglas-fir sites.  There, however, is a potential for significant depletion at the red
alder, Whiteface Mountain (magnesium), and Smokies red spruce sites.
     The range of values for soil-exchangeable turnover is very large, reflecting variations in
both the size of the exchangeable pool and the net balance of the system. Soils with the highest
turnover rates are those most likely to experience changes in the shortest time interval, other
things being equal. Thus, the Whiteface Mountain, Smokies, and Maine red spruce sites; the
Thompson red alder site; and the Huntington Forest northern hardwood site appear to be most
sensitive to change.  The actual rates, directions, and magnitudes of changes that may occur in
these soils (if any) will depend on weathering inputs and vegetation outputs, in addition to
deposition and leaching.  It is noteworthy that each of the sites listed above as sensitive has a
large store of weatherable minerals, whereas many of the other soils, with larger exchangeable
cation reserves, have a small store of weatherable minerals (e.g., Coweeta white pine, Duke
       March 2001
                                         4-69
DRAFT-DO NOT QUOTE OR CITE

-------
                                                                        I  I Soil Exchangeable
                                                                        • (Dep - Leaching)*25
                                  GS   LP    FS
                                  Warmer Sites
                     HF   MS   WF
                       Colder Sites
       Figure 4-11.  Soil exchangeable Mg2+ pools and net annual export of Mg2"1" (deposition
                    minus leaching times 25 years) in the Integrated Forest Study sites.
                    See Figure 4-5 for legend.
 1      loblolly pine, Georgia loblolly pine, and Oak Ridge loblolly pine) (Johnson and Lindberg, 1992a;
 2      April and Newton, 1992).
 3           Base cation inputs are especially important to the Smokies red spruce site because of
 4      potential aluminum toxicity and calcium and magnesium deficiencies.  Johnson et al. (1991a)
 5      found that soil solution aluminum concentrations occasionally reached levels found to inhibit
 6      calcium uptake and cause changes in root morphology in solution culture studies of red spruce
 7      (Raynal et al.,  1990). In a follow-up study, Van Miegroet et al. (1993) found a slight but
 8      significant growth response to calcium and magnesium fertilizer in red spruce saplings near the
 9      Smokies red spruce site. Joslin et al. (1992) reviewed soil and solution characteristics of red
10     spruce in the southern Appalachians, and it would appear that the IPS site is rather typical.
11           Wesselink et al. (1995) reported on the complicated interactions among changing
12     deposition and soils at this site (including repeated sampling of soil exchangeable base cation
13      pools) from 1969 to 1991 and compared these results with those of a simulation model. They
       March 2001
4-70
DRAFT-DO NOT QUOTE OR CITE

-------
160000 •} 	 • 	
140 000 •
<0
9J, 120 000 •
x: 100000 •
8"
o 80000 •
^ 60 000 •
cr

-------
 1     variety of forest ecosystems and to determine if these effects are in any way related to current or
 2     potential forest decline. Acidic deposition is having a significant effect on nutrient cycling in
 3     most of the forest ecosystems studied in the EPS project. The exceptions were the relatively
 4     unpolluted Douglas fir, red alder, and Findley Lakes in Washington state. The nature of the
 5     effects, however, varies from one location to another (Johnson, 1992).  hi all but the relatively
 6     unpolluted Washington sites, atmospheric deposition was having a significant, often
 7     overwhelming effect on cation leaching from the soils. In general, nutrient budget data from IPS
 8     and literature suggest that the susceptibility of southeastern sites to base cation depletion from
 9     soils and the development of cation deficiencies by that mechanism appears to be greater than in
10     northern sites (Johnson, 1992).
11           Atmospheric deposition may have affected significantly the nutrient status of some IPS
12     sites through the mobilization of Al.  Soil solution Al levels in the Smokies sites approach and
13     sometimes exceed levels noted to impede cation uptake in solution culture studies. It is therefore
14     possible that the rates of base cation uptake and cycling in these sites have been reduced because
15     of soil solution Al.  To the extent that atmospheric deposition has contributed to these elevated
16     soil solution Al levels, it has likely caused a reduction in base cation uptake and cycling rates at
17  '  these sites. Nitrate and sulfate are the dominant anions in the Smokies sites, and nitrate pulses
18     are the major cause of Al pulses in soil solution (Johnson, 1992).  The connection between Al
19     mobilization and forest decline is not clear. The decline in red spruce certainly has been more
20     severe in the Northeast than in the Southeast, yet all evidence indicates that Al mobilization is
21     most pronounced in the southern Appalachians. However, at the Whiteface Mountain site
22     selected for study because it was in a state of decline, soil solution levels there are lower than in
23     the Smokies, which are in a visibly obvious state of decline (e.g., no dieback other than the fir
24     killed by the balsam wooly adelgid, no needle yellowing, etc). Thus, Al mobilization constitutes
25      a situation worthy of further study (Johnson, 1992).
26           The simple calculations shown above give some idea of the importance of particulate
27      deposition in these forest ecosystems, but they cannot account for the numerous potential
28      feedbacks between vegetation and soils nor for the dynamics through time that can influence the
29      ultimate response.  One way to examine some of these interactions and dynamics is to use
30      simulation modeling.  The nutrient cycling model (NuCM) has been developed specifically for
31      this purpose and has been used to explore the effects of atmospheric deposition, fertilization, and
        March 2001
4-72
DRAFT-DO NOT QUOTE OR CITE

-------
  1     harvesting on some of the IPS sites (Johnson et al., 1993). The NuCM model is a stand-level
  2     model that incorporates all major nutrient cycling processes (uptake, translocation, leaching,
  3     weathering, organic matter decay, and accumulation).
  4          Johnson et al. (1999) used the NuCM model to simulate the effects of reduced S, N, and
  5     base cation (CB) deposition on nutrient pools, fluxes, soil, and soil solution chemistry in two
  6     contrasting southern Appalachian forest ecosystems: (1) the red spruce and (2) Coweeta
  7     Hardwood sites from the EPS project. The scenarios chosen for these simulations included
  8     "no change", 50% N and S deposition,  50% CB deposition, and 50% N, S, and CB deposition
  9     (50% N, S, CB). The NuCM simulations suggested that, for the extremely acid red spruce site,
 10     S and N deposition is the major factor affecting soil solution Al concentrations and CB deposition
 11     is the major factor affecting soil solution CB concentrations.  The effects of S and N deposition
 12     were largely through changes in soil solution SO42' and NO3" and, consequently, mineral acid
 13     anion (MAA) concentrations rather than through changes in soils. This is illustrated in
 14     Figures 4-13 and 4-14, which shows simulated soil solution mineral acid anions, base cations,
 15     Al, and soil base saturation in B horizon from in the red spruce site. The 50% S and N scenario
 16     caused reductions in soil solution SO42', NO3" and, therefore, MAA concentrations, as  expected,
 17     and this, in turn, caused short-term reductions in base cation concentrations. However, by the
 18     end of the 24-year simulation, base cations in the 50% S, N scenario were nearly as high as in the
 19     no change scenario because base saturation had increased and the proportion of cations as Al
20     decreased. The 50% CB scenario had virtually no effect on soil solution SO42", NO3" and,
21     therefore, MAA concentrations, as expected, but did cause a long-term reduction in base cation
22     concentrations.  This was caused by a long-term reduction in base saturation (Figure 14).  Thus,
23     the effects of CB deposition were solely through changes in soils rather than through changes in
24     soil solution MAA, as postulated by Driscoll et al. (1989). In the less acid Coweeta soil, base
25     saturation was high and little affected by scenario (not shown), Al was unimportant, and S and
26     N deposition had a much greater effect than CB deposition in all respects (Figure 15).
27          In summary, Johnson et al. (1999) found that the results of the red spruce simulations
28     support the hypothesis of Driscoll et al.  (1989) in part:  CB deposition can have a major effect on
29     CB leaching through time in an extremely acid system. This effect occurred through changes in
30     the soil exchanger and not through changes in soil solution MAA concentration.  On the other
31      hand, S and N deposition had a major effect on Al leaching at the Noland Divide site.  This
        March 2001
4-73
DRAFT-DO NOT QUOTE OR CITE

-------
                              Red Spruce
      500 —
      400
      300
      200 —
      100
                             Mineral Acid Anions
No Change

50%N,S

50% BC
      200
           1  23  4  5  6  7  8  9  10 11  12 13 14  15 16 17 18 19 20 21 22 23 24
                                        Year
           1  2  3  4  5  6  7  8  9 10 11  12 13 14 15 16 17 18 19 20 21 22 23  24
Figure 4-13. Simulated soil solution mineral acid anions and base cations in the red spruce
           site with no change, 50% N and S deposition, and 50% base cation
           deposition. Redrawn from Johnson et al. (1999).
March 2001
                   4-74
DRAFT-DO NOT QUOTE OR CITE

-------
                               Red Spruce
       500
       400
       300-
                  No Change

                  50%N,S

                  50% BC
       200 —
100 —
           1  2  3  4  5  6  7  8  9 10 11  12 13 14 15 16 17 18 19  20 21 22 23 24
                                         Year
        5 —
                               Base Saturation
           1  2  3  4  5  6 7  8  9  10  11 12 13 14 15  16 17 18  19 20 21 22 23 24
Figure 4-14.  Simulated soil solution Al and soil base saturation in the red spruce site with
            no change, 50% N and S deposition, and 50% base cation deposition.
            Redrawn from Johnson et al. (1999).
March 2001
                              4-75
DRAFT-DO NOT QUOTE OR CITE

-------
                                 Coweeta
     100
                            Mineral Acid Anions
                                                              No Change

                                                              50% S,N

                                                              50% BC
          1  2  3 4  5  6 7  8  9  10 11 12  13 14 15 16 17 18 19 20 21 22 23 24
                                Base Cations
          1  2  3  4  5  6 7  8  9  10 11 12  13 14 15 16 17 18 19 20 21 22 23 24
                                       Year
Figure 4-15. Simulated soil solution mineral acid anions and base cations in the Coweeta
           site with no change, 50% N and S deposition, and 50% base cation
           deposition. Redrawn from Johnson et al. (1999).
March 2001
4-76
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 occurred primarily because of changes in soil solution MAA concentration. At the less acidic
 Coweeta site, CB deposition had a minor effect on soils and soil solutions, whereas S and N
 deposition had delayed but major effects on CB leaching because of changes in SO42~ and MAA
 concentrations.

 Trace Element Effects
      Trace metals are natural elements that are ubiquitous in small (trace) amounts in soils,
 ground water and vegetation.  Many are essential elements required for growth by plants and
 animals as micronutrients. Naturally occurring surface mineralizations can produce metal
 concentrations in soils and vegetation that are as high, or higher, than those in the air and
 deposited near man-made sources (Freedman and Hutchinson, 1981). The occurrence and
 concentration of trace metals in any ecosystem component depend on the sources of the metal via
 the soil or as particulate. Even when air pollution is the primary source, continued deposition
 can result in the accumulation of trace metals in the soil (Martin and  Coughtrey, 1981). Many
 metals are deposited into soils by chemical processes and are not available to plants (Saunders
 andGodzik, 1986).
      When aerial deposition is the primary source of metal particles, both the chemical form and
 particle size deposited determine the heavy metal concentration in the various ecosystem
 components (Martin and Coughtrey, 1981). Human activities introduce heavy metals into the
 atmosphere and have resulted in the deposition of antimony, cadmium, chromium, copper, lead,
 molybdenum, nickel, silver, tin, vanadium, and zinc (Smith, 1990c).  Extensive evidence
 indicates that heavy metals deposited from the atmosphere to forests accumulate either in the
 richly organic forest floor or in the soil layers immediately below, areas where the activity in
 roots and soil is greatest. The greater the depth of soil, the lower the metal concentration. The
 accumulation of metal in the soil layers where the biological activity is greatest, therefore, has the
 potential for being toxic to roots and soil organisms and interfering with nutrient cycling (Smith,
 1990e). Though all metals can be directly toxic at high levels, only toxicity from copper, nickel,
 and zinc have been documented frequently. Toxicity of cadmium, cobalt, and lead has been seen
 only under unusual conditions (Smith, 1990e).  Exposures at lower concentrations have the
potential, over the long-term, for interfering with the nutrient-cycling processes when they affect
mycorrhizal function.
        March 2001
                                         4-77
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Accumulation of heavy metals in litter presents the greatest potential for interference with
 2      nutrient cycling.  Accumulation of metals in the litter occurs chiefly around brass works and lead
 3      and zinc smelters. There is some evidence that invertebrates inhabiting soil litter do accumulate
 4      metals.  Earthworms from roadsides were shown to contain elevated concentrations of cadmium,
 5      nickel, lead, and zinc; however, interference with earthworm activity was not cited (Martin and
 6      Coughtrey, 1981). It has been shown, however, that when soils are acidic, earthworm abundance
 7      decreases and bioaccumulation of metals from soil may increase exponentially with decreasing
 8      pH (Lokke et al, 1996). Organisms that feed on earthworms living in soils with elevated levels
 9      of Cd, Ni, Pb, and Z for extended periods could accumulate lead and zinc to toxic levels (Martin
10      and Coughtrey, 1981).  Increased concentrations of heavy metals have been found in a variety of
11      small mammals living in areas with elevated heavy metal concentrations in the soils.
12           Studies by Babich and Stotsky (1978) support the concept that increased accumulation of
13      litter in metal-contaminated areas is the result of effects on the microorganismal populations.
14     Cadmium toxicity to microbial populations was observed to decrease and prolong logarithmic
15     rates of microbial increase, to reduce microbial respiration and fungal spore formation and
16     germination, to inhibit bacterial transformation, and to induce abnormal morphologies. Also, the
17     effects of cadmium, copper, nickel, and zinc on the symbiotic activity of fungi, bacteria, and
18     actinomycetes were reported by Smith (1991). The formation of mycorrhizae by Glomus
19     mosseae with onions was reduced when zinc, copper,  nickel, or cadmium was added to the soil.
20     The relationship of the  fungus with white clover, however, was not changed.  It was suggested
21     that the effect of heavy metals on vesicular-arbuscular mycorrhizal fungi will vary from host to
22     host (Gildon and Tinker, 1983). Studies with ericoid  plants indicated that, in addition to Calluna
23     vulgaris, mycorrhizae also protect Vaccinium macrocarpa and Rhodendron ponticum from heavy
24     metals  (Bradley et al., 1981).  Heavy metals tend to accumulate in the roots, and shoot toxicity is
25     prevented.
26           The effects of sulfur deposition on litter decomposition in the vicinity of smelters also must
27     be considered.  Metal smelters emit SO2 as well as heavy metals. Altered litter decomposition
28     rates have been well documented near SO2 sources (Prescott and Parkinson, 1985). The presence
29      of sulfur in litter has been associated with reduced microbial activity (Bewley and Parkinson,
30      1984).  Additionally, the effects on symbiotic activity of fungi, bacteria and actinomycetes were
 31      reported by Smith (1990b).
        March 2001
4-78
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
      The potential pathways of accumulation of trace metals in terrestrial ecosystems, as well as
the possible consequences of trace metal deposition on ecosystem functions is summarized in
Figure 4-16.  The generalized trophic levels found in an ecosystem and the various physiological
and biological processes that could be affected by trace metals are shown in the figure.
Reduction in physiological processes can affect productivity, fecundity, and mortality (Martin
and Coughtrey, 1981). Therefore, any effects on structure and function of an ecosystem are
likely to occur through the soil and litter (Tyler, 1972).
      Trace metals deposited from the atmosphere to forests accumulate either in the richly
organic forest floor or in the soil layers immediately below, layers where greatest biological
activity occurs. The shallow-rooted species plant species are those most likely to take up metals
from the soil (Martin and Coughtrey, 1981).
      Certain species of plants are tolerant of metal contaminated soils (e.g., soils from mining
activities) (Antonovics et al., 1971).  Certain species of plants also have been used as
bioindicators of metals (e.g., Astragalius is an accumulator of selenium).  The sources of both
macroelements and trace metals in the soil of the Botanical Garden of the town of Wroclow,
Poland, were determined by measuring the concentrations of the metals in Rhododendron
catawbiense, Ilex aquifolium, and Mahonia aquifolium growing in the garden and comparing the
results with the same plant species growing in two other botanical gardens in nonpolluted areas.
Air pollution deposition was determined as the source of metals in plants rather than the soil
(Samecka-Cymerman and Kempers, 1999).
     Biological accumulation of metals through the plant-herbivore and litter-detrivore chains
can occur. A study of the accumulation of cadmium, lead, and zinc concentrations in
earthworms suggested that cadmium and zinc were concentrated, but not lead. Studies indicate
that heavy metal deposition onto the soil, via food chain accumulation, can cause excessive
levels and toxic effects in certain animals. Cadmium appears to be relatively mobile within
terrestrial food chains; however, the subsequent mobility of any metal after it is ingested by a
herbivorous animal depends on the site of accumulation within body tissues. Although food
chain accumulation may not in itself cause death, it can reduce the breeding potential in a
population (Martin and Coughtrey, 1981).
     In actual case studies, it was observed that the deposition of copper and zinc particles
around a brassworks resulted in an accumulation of incompletely decomposed litter. In one
        March 2001
                                         4-79
DRAFT-DO NOT QUOTE OR CITE

-------
                                                                       vo
                                                                       1—I
                                                                       4

                                                                        I
March 2001
4-80
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 study, litter accumulation was reported up to 7.4 km from the stack of a primary smelter in
 southeastern Missouri.  Similar results were reported around a metal smelter at Avonmouth,
 England.  In the latter case, litter accumulation was associated closely with concentrations
 specifically of cadmium, as well as with those of lead, copper, and zinc (Martin and Coughtrey,
 1981). Experimental data (using mesh bags containing litter) supports the hypothesis that
 reduced decomposition occurs close to heavy metal sources.
      Accumulations of metals emitted in particulate matter also were reported in soil litter close
 to a metal smelter at Palmerton, PA, in 1975 and 1978. The continued presence of cadmium,
 lead,  zinc, and copper in the upper soil horizons (layers) were observed 6 years after the smelter
 terminated operation in 1980.  Metal levels were highest near the smelter.  The relationship  of
 decreasing amounts of metal in body tissues also held true for amphibians and mammals. Levels
 of cadmium in kidneys and liver of white-tailed deer (Odocoileus virginaus) were five times
 higher at Palmerton than in those collected 180 km southwest downwind. The abnormal
 amounts of metal in the tissues of terrestrial vertebrates and the absence or low abundance of
 wildlife at Palmerton indicated that ecological processes within 5 km of the zinc smelter
 continued to be markedly influenced even 6 years after its closing (Storm et al., 1994).
      The effects of lead in ecosystems are discussed in fo&'Air Quality Criteria for Lead
 (U.S. Environmental Protection Agency, 1986b).  Studies have shown that there is cause for
 concern in three areas where ecosystems may be extremely sensitive to lead: (1) delay of
 decomposition because the activity of some decomposer microorganisms and invertebrates is
 inhibited by lead, (2) subtle shifts toward plant populations tolerant of lead, and (3) lead in the
 soil and on the surfaces of vegetation circumvent the processes of biopurification.  The problems
 cited above arise because lead is deposited on the surface of vegetation, accumulates in the soil,
 and is not removed by the  surface and ground water of the ecosystem (U.S. Environmental
 Protection Agency, 1986b).

 4.2.3 Ecosystem Goods and Services and Their Economic Valuation
     Human existence on this planet depends on ecosystems  and the services and products they
provide. The essential services and products provided by the  planet's collective biodiversity (the
earth's flora, fauna, and microorganisms) are clean air, clean water, clean soil, and clean energy
(Table 4-6). Today, governments around the world pursue a "bottom line" that driven is by an
March 2001                              4-81        DRAFT-DO NOT QUOTE OR CITE

-------
   TABLE 4-6.  PRIMARY GOODS AND SERVICES PROVIDED BY ECOSYSTEMS

 Ecosystem                       Goods   	         Services	
 Agroecosvstems
Coastal ecosystems
 Forest ecosystems
 Freshwater
 Grassland
 ecosystems
                    • Food crops
                    • Fiber corps
                    • Crop genetic resources
                     • Fish and shellfish
                     • Fishmeal (animal feed)
                     • Seaweeds (for food and industrial
                      use)
                     •Salt
                     • Genetic resources
                    • Timber
                    • Fuelwood
                    • Drinking and irrigation water
                    • Fodder
                    • Nontimber products (vines,
                     bamboos, leaves, etc.)
                    • Food (honey, mushrooms fruit,
                     and other edible plants; game)
                    • Genetic resources
                    • Drinking and irrigation water
                    •Fish
                    • Hydroelectricity
                    • Genetic resources
                     • Livestock (food, game, hides, and
                     fiber)
                     • Drinking and irrigation water
                     • Genetic resources
• Maintain limited watershed functions (infiltration,
 flow control, and partial soil protection)
• Pro vide-habitat for birds, pollinators, and soil
 organisms important to agriculture
• Sequester atmospheric carbon
• Provide employment

• Moderate storm impacts (mangroves, barrier
 islands)
• Provide wildlife (marine and terrestrial) habitat
 and breeding areas/hatcheries/nurseries
• Maintain biodiversity
• Dilute and treat wastes
• Provide harbors and transportation routes
• Provide human and wildlife habitat
• Provide employment
• Contribute aesthetic beauty and provide recreation

• Remove air pollutants, emit oxygen
• Cycle nutrients
• Maintain array of watershed functions (infiltration,
 purification, flow control, soil stabilization)
• Maintain biodiversity
• Sequester atmospheric carbon
• Moderate weather extremes and impacts
• Generate soil
• Provide employment
• Provide human and wildlife habitat
• Contribute aesthetic beauty and provide recreation

•Buffer water flow (control timing and volume)
• Dilute and carry away wastes
• Cycle nutrients
• Maintain biodiversity
• Provide aquatic habitat
• Provide transportation corridor
• Provide employment
• Contribute aethetic beauty and provide recreation

•Maintain array of watershed functions (infiltration,
 purification, flow control, and soil stabilization)
• Cycle nutrients
• Remove air pollutants and emit oxygen
• Maintain biodiversity
• Generate soil
• Sequester atmospheric carbon
• Provide human and wildlife habitat
• Provide employment
• Contribute aesthetic beauty and provide recreation
 Source: World Resources (2000-2001).
March 2001
                                                4-82
        DRAFT-DO NOT QUOTE OR CITE

-------
  1      economy that is disconnected from the natural world and is fundamentally destructive of local
  2      ecosystems (Suzuki, 1997). For this reason, human society needs to be reconnected to the
  3      biologically diverse ecosystems and the natural world of which they are a part (Suzuki, 1997).
  4      There is a need to understand the biodiversity that encompasses all levels of biological
  5      organization, including populations, individuals, species and ecosystems (Wilson, 1997).
  6      Populations, geographical entities within a species of organisms, usually distinguished
  7      ecologically or genetically, are essential to the conservation of species diversity. Their number
  8      and size influence the probability of the future existence of the entire species (Hughes et al.,
  9      1997). The number, biodiversity, structure, and functions of ecosystem populations, provide
 10      ecosystem products (goods) and services. For any given population, the number of individuals,
 11      the genetic variation between individuals, and the  area occupied affects ecosystem functioning
 12      and the delivery of ecosystem services and other benefits provided by that population (Hughes,
 13      et al., 1997). Loss of population diversity means loss of the benefits described in Table 4-6 and,
 14      in particular, with time, the loss of the life-support systems on which humanity relies (Hughes
 15      etal., 1997).
 16           Attempts have been made to value biodiversity and the world's ecosystem services and
 17      natural capital (Pimentel et al., 1997; Costanza et al., 1997).  Pimentel et al. (1997) estimated
 18      economic and environmental benefits for services  contributed from all biota (biodiversity) in the
 19      United States, including their genes, at $319 billion per year. Costanza et al. (1997) have
20      estimated the total value of ecosystem services by biome for the entire bioshere.  Ecosystems
21      provide at least $33 trillion worth of services annually. Approximately, 63% of the estimated
22      value is contributed by marine ecosystems ($20.9 trillion per year), most of which comes from
23      coastal ecosystems ($10.6 trillion per year). About 38% of the estimated value comes from
24      terrestrial ecosystems, mainly from forests ($4.7 trillion per year) and wetlands ( $4.9 trillion per
25      year). Costanza et al. (1997) state that it may never be possible to make a precise estimate of the
26      services provided by ecosystems. Their estimates,  however, indicate the relative importance of
27      ecosystem services.
28           Heal (2000), however, feels that attempts to value ecosystems and their services are
29      probably misplaced. "Economics cannot estimate the importance of natural environments to
30      society: only biology can do that" (Heal, 2000).  The role of economics is to help design
31      institutions that will provide incentives to the public and policy makers for the conservation of
       March 2001
4-83
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
important natural systems and for mediating human impacts on the biologically diverse
ecosystems and the biosphere so that they are sustainable. The approach of Harwell et al. (1999)
also deals with the need to understand human impacts on ecosystems so that ecosystem
management can define what ecological conditions are desired. Further, they state that the
establishment of ecological goals involves a close linkage between scientists and decision
makers, in which science informs decision makers and the public by characterizing the ecological
conditions that are achievable under particular management regimes.  Decision makers then can
make choices that reflect societal values, including issues of economics, politics, and culture.
For management to achieve their goals, the general public, scientific community, resource
managers, and decision makers need to be routinely apprised of the condition or integrity of
ecosystems, so that ecological goals may be established (Harwell et al., 1999).
      The above assessment of new information leads to the clear conclusion that atmospheric
PM at levels currently found in the United States has the potential to alter ecosystem structure
and function in ways that may reduce their ability to meet societal needs. The possible direct
effects of airborne PM on individual plants were discussed in Section 4.2.1 above. The major
impacts of airborne PM on ecosystems, however, are the indirect effects on plant populations that
occur through the soil and affect the cycling of nutrients necessary for plant growth and vigor, as
discussed in Section 4.2.2. By altering the cycling of nitrogen, nitrogen deposition changes the
biodiversity of ecosystems and their functioning and, by altering the vigor of forest tree stands,
alters forest succession. Also, nitrogen deposition in combination with the deposition of sulfur in
the form of acid rain alters the biogeochemical cycling of soil mineral nutrients and changes the
biodiversity and functioning of forest ecosystems. The changes in the ability of forest vegetation
and soil microorganisms to utilize nutrients results in the leaching of nitrates and other minerals
from the soils. The nitrate and mineral runoff impacts coastal and aquatic ecosystems and, thus,
influences the services important to human life provided by these ecosystems as well (Table 4-6).
        March 2001
                                           4-84
DRAFT-DO NOT QUOTE OR CITE

-------
  1      4.3  EFFECTS ON VISIBILITY
  2      4.3.1 Introduction
  3           Visibility is defined as the degree to which the atmosphere is transparent to visible light and
  4      the clarity (transparency) and color fidelity of the atmosphere (National Research Council, 1993).
  5      Visibility impairment is defined as any humanly perceptible change in visibility (light extinction,
  6      visual range, contrast, or coloration). Visual range is described as the farthest distance at which a
  7      large black object can be distinquished against the horizontal sky (U.S. Environmental Protection
  8      Agency, 1979). For regulatory purposes, visibility impairment is classified into two principal
  9      forms:  (1) "reasonably attributable" impairment, attributable to a single source or small group of
 10      sources and (2) regional haze, described as any perceivable change in visibility (light extinction,
 11      visual range, contrast, or coloration) from which would have existed under natural conditions
 12      that is caused predominantly by a combination of many sources over a wide geographical area
 13      (U.S. Environmental Protection Agency, 1999).
 14           The objective of the visibility discussion in this section is to summarize the linkage
 15      between air pollution, in particular particulate matter, and visibility.  This section summarizes the
 16      information discussed in the previous 1996 PM air quality criteria document (PM AQCD) and
 17      includes additional relevant information available since publication of that document. For a
 18      more detailed discussion on visibility, the reader is referred to the earlier PM AQCD entitled, Air
 19      Quality Criteria for Particulate Matter (U.S. Environmental Protection Agency, 1996a), the
20      Recommendations of the Grand Canyon Visibility Transport Commission (Grand Canyon
21      Visibility Transport Commission, 1996), the National Research Council (National Research
22      Council, 1993), the National Acid Precipitation Assessment Program (Trijonis et al.,  1991), and
23      the U.S. Environmental Protection Agency (1995a).
24
25      4.3.2 Factors Affecting Atmospheric Visibility
26      4.3.2.1  Anthropogenic Pollutants
27'          Visibility impairment may be connected to air pollutant properties, including size
28      distribution, aerosol chemical composition, and relative humidity,  hi the United States, visibility
29      impairment is caused by sulfate and nitrate particles in the 0.1- to 1.0-micron (ju.m) range, and
30      organic aerosols, carbon soot, and crustal dust. Generally, sulfates are responsible for most of
        March 2001
4-85
DRAFT-DO NOT QUOTE OR CITE

-------
 1      the visibility impairment in the United States, as measured by light extinction, accounting for
 2      approximately two-thirds of the light extinction in the eastern United States.  Sulfate
 3      concentrations are higher in summer months than in the wintertime (Malm et al., 1994).
 4      Exceptions to the sulfate-related effects on visibility include California, where the primary cause
 5      of visibility effects is ambient nitrate, and Alaska, where visibility impairment is caused by fine
 6      soil plus coarse mass (classified as coarse extinction) or organics, thought to be from natural
 7      sources (Sisler and Cahill, 1993).
 8
 9      4.3.2.2 Human Vision
10           Human vision is one of the factors that affects the way an object is viewed.  Vision is the
11      response to the electromagnetic radiation that enters the eye between wavelengths of 400 and
12     700 nm. The cones, a receptor cell in the retina, govern visibility interpretations.
13           The eye perceives the lightest and brightest object in a scene as white, and determines the
14     color of other objects by comparison!  The ability of the eye to perceive contrasts, the degree of
15     color difference between the lightest and darkest object in a scene, changes in response to the
16     illumination and setting. The effects of illumination on visibility are discussed in the following
17     subsection. At increasing distances the brightness of a target or object will approach the
18     brightness of the horizon making the target indistinquishable from the horizon, hence, visual •
19     range.
20
21     4.3.2.3  Characteristics of the Atmosphere
22           The appearance of a distant object is determined by illumination of the sight path by the
23      direct rays of the sun, diffused skylight, light that has been reflected from the surface of the Earth
24      (path radiance or air light), and the light reflected from the object itself.  Some of the light in the
25      sight path is absorbed or scattered towards the observer.  The remaining light is absorbed or
26      scattered in other directions.  The portion of scattered light from the object being viewed that
27      reaches the observer is the transmitted radiance.  The radiance seen by the observer looking at a
28      distant object is the sum of the transmitted radiance and the path radiance. Figure 4-17
29      demonstrates light being absorbed and scattered by the atmosphere and a target object.
30           On a clear day when the sun is high in the sky, 80 to  90% of the visible solar radiation
31      reaches the surface of the Earth without being scattered or absorbed. Rayleigh scattering by
        March 2001
4-86
DRAFT-DO NOT QUOTE OR CITE

-------
       Figure 4-17. Light reflected from a target toward an observer. The intervening
                   atmosphere scatters a portion of this light out of the sight path and scatters
                   light from the sun into the sight path. Some particles and gases also absorb a
                   portion of the light from the target. The light scattered into the sight path
                   increases with distance from the target, whereas the light transmitted from
                   the target decreases with distance from the target. The visual range is the
                   closest distance between the target and the observer at which the transmitted
                   light no longer can be distinguished from the light scattered into the sight
                   path.
       Source: Watson and Chow (1994).
1     gases is the major component of light extinction in relatively unpolluted areas. Mie scattering is
2     the scattering of all visible wavelengths equally (Shodor Education Foundation, Inc., 1996).  It is
3     the attenuation of light in the atmosphere by scattering because of particles of a size comparable
4     to the wavelength of the incident light (National Acid Precipitation Assessment Program, 1991).
5     The term, multiple scattering, is used when light is scattered more than once in a turbid medium.
6     The great majority of light absorption by particles is caused by black carbonaceous particles,
7     assumed to be elemental carbon, that are products of incomplete combustion (Rosen et al., 1978;
8     Japar et al., 1986; Watson and Chow, 1994). Malm et al. (1996) suggested that organic carbon
      March 2001
4-87
DRAFT-DO NOT QUOTE OR CITE

-------
 1     also acts to scatter and absorb light. The estimated natural visibility for the east and west is 60 to
 2     90 mi and up to 140 mi, respectively. Current visibility conditions range from 18 to 40 mi in the
 3     rural east to 35 to 90 mi in the rural west (U.S. Environmental Protection Agency, 2000b).
 4           At the surface, a variable fraction of the solar radiation is reflected back upwards, referred
 5     to as surface reflectance or the albedo, illuminating the atmosphere from above and below.  The
 6     amount of solar radiation reflected depends on the color of the terrain. Dark-colored terrain
 7     reflects less radiation than light-colored terrain.
 8           Visibility within a sight path longer than approximately 100 km (60 mi) is affected by
 9     changes in the properties of the atmosphere over the length of the sight path. The atmosphere
10     generally will not have uniform optical properties over distances greater than a few tens of
11     kilometers.  Air quality within a sight path can affect the illumination of the sight path by
12     scattering or absorbing solar radiation before it reaches the Earth's surface. The light-extinction
13     coefficient,  oext, is a measure of the fraction of light that is lost as it travels through the
14     atmosphere. The light-extinction coefficient is the sum of the light-scattering coefficient, oscat,
15     and the light-absorption coefficient, oabs, expressed in units of inverse lengths of the atmosphere
16     (megameters ; Mm'1). Typical extinction coefficients range from 0.01 km'1 (10 Mm'1) in
17     relatively clean air to ~ 1000 Mm'1 in highly polluted areas (Watson and Chow, 1994).
18           The light-extinction coefficient can be divided into coefficients for the following
19     components:
20           aag, light absorption by gases,
21           asg, light scattering by gases (Rayleigh scattering),
22           aap, light absorption by particles,  and
23           Ojp, light scattering by particles.
24     Light scattering by particles, a^, can be divided to indicate scattering by coarse and fine particles:
25      asfp, light scattering by fine particles and oscp, light scattering by coarse particles.
26
27      4.3.3  Optical Properties  of Particles
28           Visibility impairment is typically caused by fine particles. Fine particles are small enough
29      in comparison with the wavelength of visible light that their optical properties are nearly the
30      same as those of homogeneous spheres of the same volume and average index of refraction.
31      Accordingly, Mie equations (Mie, 1908; Kerker, 1969), for calculating the optical properties of
        March 2001
4-88
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
25
26
27
28
29
30
31
 homogeneous spheres also may be used to calculate the optical properties of fine particles with
 the only uncertainties being in the fine particle size distribution and index of refraction (Richards,
 1973). However, within the range of indices of refraction that most commonly occur in
 atmospheric fine particles, the results of Mie calculations can be scaled to account for the effect
 of the index of refraction.  Coarse particles have less of an impact on visibility than do fine
 particles. However, in most actual cases, the dominant uncertainty in using the optical properties
 for coarse particles calculated with Mie equations is the uncertainty in the particle size
 distribution. Uncertainties exist in the use of Mie calculations for calculating light absorption for
 course particles because the refractive index of the particle is generally not known, and the
 light-absorbing particles are not spherical in shape, making the calculated light absorption
 efficiency factor less reliable.  Also, light absorption by elemental carbon particles can be
 reduced when the particle is covered by some chemical species (Dobbins et al, 1994).
 Conversely, light absorption by carbon particles can be enhanced when coated with a
 nonabsorbing refractive material such as ammonium sulfate (Fuller et al., 1999).
      The output of the Mie calculations includes efficiency factors for extinction, Qext,
 scattering, Qscat, and absorption,  Qabs. The Qext, Qscat, and Qabs give the fraction of the incident
 radiation falling on a circle with the same diameter as the particle that is either scattered or
 absorbed. The light scattering or absorption efficiency factor (in units of m2/g) is the change in
 the light scattering or absorption efficiencies per unit change in mass of the fine particle
 constituent. The scattering and absorption efficiencies  are determined by estimating the size
 distribution of each particle. The results of the calculations for the light absorption efficiencies
 contains significant uncertainties because the components of the index of refraction is generally
 unknown and the light-absorbing particles are frequently chained agglomerates that do not have a
 spherical shape.  Multiplying the values of the light-scattering efficiency factor by the aerosol
 volume concentration (in units of ^nrVcm3) gives the value of the light-scattering coefficient, asp,
 (in units of Mm"1) for these particles.
      Richards et al. (1991) reported a scattering efficiency for fine particles of ammonium
 sulfate of 1.2 m2/g based on Mie calculations.  The value was in agreement with the value
determined using the integrating nephelometer readings and the sulfate concentrations. Sulfate
scattering efficiencies have been reported to increase by a factor of two when the size distribution
went from 0.15 to 0.5 /u.m (McMurry et al., 1996). The calculated scattering efficiencies for
        March 2001
                                           4-89
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
 30
sulfates were 4.1 m2/g for 100% mass removal and 3.4 and 5.6 m2/g for 25% mass removal.
Calculated scattering efficiencies for carbon particles ranged from 0.9 to 8.1 m2/g (Zhang et al.,
1994; Sisler and Malm, 2000; Sloane et al., 1991). A scattering efficiency of 1.0 and 0.6 m2/g
was reported for soil and coarse mass, respectively (Trijonis et al., 1987).
     Scattering efficiencies of 2.4 and 3.1 m2/g for fine particles were reported by White et al.
(1994) and Waggoner et al. (1981), using an integrating nephelometer.  Coarse particle scatter
less light, resulting in lower scattering efficiencies. Scattering efficiencies for coarse particles
ranged from 0.4 to 0.6 m2/g, based on integrating nephelometer readings (White et al., 1994;
Trijonis et al., 1987; White and Macias, 1990; Watson et al., 1991).
     Absorption efficiencies for elemental carbon particles have been reported to range from
9 to 10 m2/g (Japar et al., 1984; Adams et al., 1989; Sloane et al., 1991). Based on a review of
the available data, Horvath (1993) reported that measured light absorption efficiencies for light
absorbing carbon ranges from 3.8 to 17 m2/g. According to Horvath (1993), calculated
absorption efficiencies are too high, ranging from 8 to 12 m2/g for monodispersed carbon
particles. Fuller et al. (1999)  suggested that isolated spheres of light absorbing carbon have a
specific absorption of less than 10 m2/g.  Light absorption by carbon particles only will be greater
than 10 m2/g if the particles are internally mixed and the occluding particles are sufficiently large.
Absorption  values for graphitic and amorphous carbon spheres for primary sizes typical of diesel
soot are around 5 m2/g. Light absorption by aggravated carbon at visible wavelengths is
enhanced by no more than 30% and diminishes if encapsulated by a nonabsorbing aerosol.
Malm et al. (1996) suggested a combined scattering and absorption efficiency of 10 m2/g for
organic carbon.
      Light-extinction budgets may be estimated using the light extinction efficiency and the
measured species concentrations.  Light-extinction budgets estimate the fraction of the total light
 extinction contributed by each chemical species in the sight path; however, the values obtained
 will depend on the assumptions used (Malm et al., 1996; Lowenthal et al., 1995; Sisler and
 Malm, 1994).
        March 2001
                                           4-90
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 4.3.4 Effect off Relative Humidity on Particle Size and Light-Scattering
       Properties
      Ambient particles contain water, even on relatively dry days.  As the relative humidity
 increases, the particle absorbs more water and increases in size and volume. It is the increase in
 particle size and volume that acts to increase the light scattering properties of most particles
 (Malm etal., 1996).
      Ambient particles are a mixture of chemical compounds. The amount of increase in
 particle size with increasing relative humidity is dependent on the particle composition (Zhang
 et al., 1993). Available data indicate that particles containing ammonium salts are in a liquid
 solution at relative humidities above 80%.  Particles containing inorganic salts and acids are
 more hygroscopic than particles composed primarily of organic species (Day et al., 1996;
 McMurry and Stolzenburg, 1989; Saxena et al., 1995; Zhang et al.,  1993, 1994; Sloane et al.,
 1991). Particles containing the more hygroscopic salts  and acid species deliquesce and undergo
 changes in particle size in response to changes in relative  humidity. For sulfate and nitrate
 aerosols, light-scattering properties are similar for all mixture types and compositions, as long as
 there is the same particle size distribution (Tang, 1997). Saxena et al. (1995) found that the
 hygroscopic properties of inorganic particles can be altered positively or negatively in the
 presence of organics.  Based on limited data, nonurban  organics were found to add to water
 absorption by inorganics, whereas the urban organics diminished the absorption of water by
 inorganic particles at relative humidities of 80 to 93%.  Figure 4-18 demonstrates the humidity
 effect on the scattering coefficients for several internally mixed (individual particles  containing
 one or more species) and externally mixed (species that co-exist as separate particles) aerosols.
 The total scattering computed for an aerosol is relatively insensitive to whether the sample is
 internally or externally mixed (Malm et al., 1997). Figure 4-19 demonstrates changes in the
 scattering coefficient ratio, ospw/aspd, where aspw is the scattering coefficient under humid
 conditions, and ospd is the scattering coefficient under dry conditions. The figure demonstrates
that light scattering is a function of relative humidity and chemical composition. The monitoring
data were generated as part of the Southeastern Aerosol and Visibility Study (Day et al., 2000).
     There is also a relative humidity-related effect on the scattering efficiency. Ammonium
sulfate fine-particle-scattering efficiency varied from 1.5 to 4.5 m2/g, with low relative humidity
and median particle sizes ranging from 0.07 to 0.66 /^m (McMurry et al., 1996). Sloane et al.
        March 2001
                                          4-91
DRAFT-DO NOT QUOTE OR CITE

-------
          0.1
      Q.

     b
     0
     O
     o
     D)
     C
     to
     o
     CO
        0.001-
               30
                     	(NH4)2S04
                     	Na2SO4
                     o   Internal Mixture
                     	External Mixture
                                            (0.3 |jm, 1.5)   j
                         (0.6 |jm, 2.0)  ../a
T"
 40
50
60      70
    %RH
T"
 80
90
100
Figure 4-18. Humidity effect on scattering coefficients computed for internal and external
            mixtures of the mixed-salt aerosol: Na2SO4 (x2 = 0.5)-(NH4)2 SO4 (x3 = 0.5),
            for two dry-salt particle size distributions, where x is the mass fraction of the
            dry solutes. Particle size distributions are stated in the parenthesis.

Source: Tang (1997).
March 2001
                4-92
                   DRAFT-DO NOT QUOTE OR CITE

-------
            4-
        I
       S.
        o
       f •
       S   2'
        O)
        0)
       1   1.
            0
                  A Day 207  (32.6% Sol. Inorg.; 25.8% Org.; 41.6% Soil)
                     Day 211 (42.7% Sol. Inorg.; 51.7% Org.; 5.6% Soil)
            o Day 224 (63.5% Sol. Inorg.; 32.0% Org.; 4.5% Soil)
                     10     20     30     40     50     60     70
                                           Relative Humidity (%)
                                                                     80      90
                              100
       Figure 4-19.  Scattering ratios, ospw/ospd, for different chemical compositions as a function
                    of relative humidity.
       Source: Day et al. (2000).
1
2
3
4
5
6
7
8
(1991) reported scattering efficiencies of 7.1 to 8.2 m2/g for sulfate at 74% relative humidity and
2.1 to 2.9 m2/g at 38% relative humidity. Average dry scattering efficiencies for sulfate ranged
from 2.03 to 2.23 m2/g for two western sites and one eastern site (Malm and Pitchford, 1997).
The dry scattering efficiency increased with increasing particle size.  Dry specific scattering
efficiencies of 3 m2/g were reported for sulfates and nitrates (Sisler and Malm, 2000). Omar
et al. (1999) reported a calculated scattering efficiency range of 1.23 m2/g for sulfate when the
relative humidity was <63% to 5.78 m2/g when the relative humidity was >75%. The calculated
scattering efficiencies for organic carbon ranged from 3.81 m2/g when the relative humidity was
      March 2001
                                         4-93
DRAFT-DO NOT QUOTE OR CITE

-------
 1      <63% to 6.9 m2/g at relative humidities above 75% (Omar et al., 1999). A more detailed
 2      discussion of the effects of relative humidity on the size distribution of ambient particles appears
 3      in Chapter 2 of this document.
 4
 5      4.3.5  Measures of Visibility
 6      4.3.5.1 Human Observations
 7           The National Weather Service has in recent decades recorded hourly visibility readings at
 8      all major airports in the United States based on human observations of the most distant targeted
 9      object's perceivability. Human observation of visibility, although providing a historical record of
10     visibility readings in the United States, are dependent on the individual and the availability of a
11      target and generally are related poorly to air quality.
12
13     4.3.5.2 Light-Extinction Coefficient and Parameters Related to the Light-Extinction
14             Coefficient
15          The most frequently used indicator for visibility characterization for air quality is the
16     light-extinction coefficient because it is closely linked to air quality (U.S. Environmental
17     Protection Agency, 1996a). Various meteorological conditions (moisture and cloud cover) can
18     affect the light-extinction coefficient; however, these effects can be minimized (Husar et al.,
19      1994;  Blandford, 1994; Mercer, 1994).  The light-extinction coefficient can be measured directly
20     using a transmissometer (Molenar et al., 1990,1992) or can be estimated by measuring the
21     components of light extinction (scattering and absorption) and calculating the sum (Malm et al.,
22      1994).
23           The light-extinction coefficient is the quantitative measure of haziness, defined as
24      0«a= K/visual range, where K is the Koschmieder constant. The value of K is determined both
25     by the threshold sensitivity of the human eye and the initial contrast of the visible object against
26     the horizon sky.
27           The visual range may be calculated from the light-extinction coefficient using the
28      Koschmieder equation by assuming the atmosphere and the illumination over a sight path in the
29      daytime is uniform, and that the threshold contrast is 2% (Katsev and Zege, 1994; Koschmieder,
30      1924). These assumptions are, however, invalid for visual ranges greater than 100 km (U.S.
31      Environmental Protection Agency, 1996a). Visual range is an understandable, and for most
        March 2001
4-94
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
purposes, an appropriate measure of the optical environment. It has the disadvantage of being
related inversely to aerosol concentration.
                                  Visual Range = 3.
      The deciview index is an atmospheric haze index that expresses uniform changes in
haziness in common increments from pristine conditions to extremely visibility impaired
environments. The deciview scale is linear with perceived visual changes, starting near zero for
a pristine atmosphere (particle-free) at a 1.8-km elevation, and increases with increasing
haziness. The deciview index may be calculated from the light-extinction coefficient for green
light. For consistency, a Rayleigh scattering value of 10 Mm"1 is used.

                                dv = 10 Iogj0 (a^ /10 Mm -1)

      Under ideal conditions, a just noticeable change in the light-extinction coefficient should
represent a one or two deciview change in the deciview scale, about a 10 to 20% change in the
extinction coefficient. Any change in the deciview scale should have a change of similar
magnitude in the visual appearance of the scene in cases where the assumptions used to develop
the deciview scale are met (Pitchford and Malm, 1994; Sisler and Malm, 2000). Figure 4-20
illustrates the relationship of light extinction in Mn"1, deciview index, and visual range in
kilometers. Although the deciview is related to extinction, it is scaled in such a way that is
perceptually correct (Fox et al., 1999).
               Extinction (Mm'1)   _J°
              Deciviews   (dv)
                               20
                                     30
                                          40   50   70 100     200    300   400  500  700 1000
           Visual Range   (km)
I
0
I
I
7
I
I
11
I
I I
14 16
I I
I Illl
19 23
I Illl
I
30
I
1
34
1
1
37
1
1
39
1
1 Illl
42 46
1 Illl
                                       200
                                             130  100  80  60 40
                                                                   20
                                                                         13
        Figure 4-20.  Comparison of extinction (Mn"1) and visual range (km).
        Source: Fox et al. (1999).

        March 2001                                4-95        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Figures 4-2 la,b illustrate a change in deciview scale based on reconstructed extinction
 2     coefficients for the Great Plains Region (Badlands) using data from the Interagency Monitoring
 3     of Protected Visual Environments Network (IMPROVE). Details about the IMPROVE network
 4     appears in Section 4.5.6. The data are sorted by year into three groups based on the cumulative
 5     frequency of occurrence of PM2 5: best visibility days (1 Oth percentile), median (50th percentile),
 6     and worst visibility days (90th percentile) (Sisler and Malm, 2000).
 7           Richards (1999) suggests that the deciview index may not be a good tool for measuring
 8     visibility impairment in areas restricted by boundaries. The deciview index is, however, suitable
 9     for measuring visibility conditions over a broad geographic region, which is consistent with the
10     definition of regional haze, uniform haze caused by pollutant sources over broad areas (U.S.
11     Environmental Protection Agency, 1999).
12
13     4.3.5.3 Light-Scattering Coefficient
14           Light-scattering by particles has been reported to account for 68 to 86%  of the total
15     extinction coefficient in several cities in California (Eldering et al., 1994). The light-scattering
16     coefficient is closely linked to fine particle concentrations, making it a good tool for determining
17     small particle-related effects on visibility.  When the light-scattering coefficient is increased,
18     visibility is impaired because the transmitted radiance is decreased and the path radiance is
19     increased. (See discussion in the previous sections on transmitted radiance and path radiance.)
20     The light-scattering coefficient can be measured directly with an open and enclosed integrating
21     nephelometer and a forward scatter visibility monitor (Molenar et al., 1992; National Oceanic
22     and Atmospheric Administration, 1992). The light-scattering coefficient also may be calculated
23     using analytical approximations of the particle size distributions, log normal size distributions, or
24      sectional particle size distributions. In the sectional approach, the size composition distribution
25      is represented by a set of particle size sections. The chemical composition of each size section is
26      assumed to be the same (Wu et al., 1996).
27
28     4.3.5.4  Fine Particulate Matter Concentrations
29           The influence of particles on visibility degradation is dependent on the particle
30      composition, solubility, and size (Pryor and Steyn, 1994). Fine particle species have been
31      classified into five major types:  (1) sulfates, (2) nitrates, (3) organics, (4) light absorbing carbon,
        March 2001
4-96
DRAFT-DO NOT QUOTE OR CITE

-------
                            Fine Mass PM
                                            2.5
          Q.
6-
5-
4-
3-
2-
1-
               88    89   90
                  91    92   93    94
                  Sample Year
         95    96
                          Visibility Impairment
                88   89    90   91   92    93   94   95   96
                                Sample Year
                  •10th Percentile
                    • 50th Percentile
                                                   • 90th Percentile
Figure 4-21a,b. Plots of the 10th, 50th, and 90th percentile groups for PM25 and deciview
             at the Badlands National Park. The sample year began in March of each
             year.

Source: Sisler and Malm (2000).
March 2001
                     4-97
DRAFT-DO NOT QUOTE OR CITE

-------
1      and (5) soil (Malm et al., 1994). The coefficient of light-scattering by fine particles is primarily
2      responsible for visibility impairment making fine particle concentration a suitable indicator of.
3      particle related effects on visibility. Several studies have demonstrated a relationship between
4      the coefficient for light-scattering by particles, measured using an integrating nephelometer, and
5      fine particle concentrations (Dattner, 1995; Waggoner and Weiss, 1980; Waggoner et al., 1981;
6      White et al., 1994). Figure 4-22 demonstrates visual range based on particle concentrations and
7      extinction efficiencies for road dust and sulfate.
8
               400
               300 -
             0)
             D)
             i 200
             —
             15
             CO

               100 -
                                                                 • Sulfate A Road Dust
                                          change in visibility based on 1 ug/m3
                                          difference in concentration
                         —i	—i	1	—i	1	1	i—
                          10     20      30      40      50      60      70
                                             Concentration (pg/m3)
                        so
                   go
100
       Figure 4-22.  Reduction in visual range as a function of increasing fine (sulfate) and coarse
                     (dust) particle concentrations.
       Source: Watson and Chow (1994).
 1      4.3.5.5 Discoloration
 2           Discoloration may be used as a quantitative measurement of atmospheric color changes in
 3      urban hazes. Atmospheric color changes is a component of plume visibility models. The color
        March 2001
4-98
DRAFT-DO NOT QUOTE OR CITE

-------
 1      of haze will primarily depend on the scene used and human vision. For plume visibility, the
 2      threshold for perception of color differences depend on the apparent width of the plume and is
 3      greater for color patches separated by sharp edges.  Methods for specifying the colors of hazes
 4      include the CIE XYZ system of color matching, the Hunt94 color-appearance model, and the
 5      visual colorimeter, VISUAL colorimeter for Atmospheric Research (Trijonis et al., 1991;
 6      Mahadev and Henry, 1999).
 7
 8      4.3.6 Visibility Monitoring Methods and Networks
 9           Visibility monitoring studies measure the properties of the atmosphere either at the sampler
10      inlets (point measurements), as is the case with air quality measurements, or by determining the
11      optical properties of a sight path through the atmosphere (path measurements).  Instrumental
12      methods for measuring visibility are generally of three types:  (1) direct measurement of light
13      extinction of a sight path using a transmissometer, (2) measurement of light scattering at one
14      location using an integrating nephelometer, and (3) measurement of ambient aerosol mass
15      concentration and composition (Mathai, 1995).
16           The largest instrumental visibility monitoring network in the United States is designed to
17      provide real-time data for runway visibility to aid in controlling airport operations.
18      An automated observing system, Automated Surface Observing System (ASOS), is being placed
19      at airports around the country.  This monitoring network is sponsored by the National Weather
20      Service, the Federal Aviation Administration, and the Department of Defense.  More than
21      500 airports are currently commissioned and an additional 500 are expected to come online in the
22      next few years.
23           The visibility sensor, instead of measuring how far one can see, measures the clarity of the
24      air using a  forward scatter visibility meter. The forward scatter meter was found to correlate
25      fairly well with extinction coefficient measurements from the Optec Transmissometer. The
26      clarity is then converted to what would be perceived by the human eye using a value called
27      Sensor Equivalent Visibility (SEV).  Values derived from the sensor are not affected by terrain,
28      location, buildings, trees, lights, or cloud layers near the surface. The sensor transmits an
29      average 1-min value for a 10-min period. The sensor only samples 0.75 ft of the atmosphere.
30      An algorithm processes the air passing through the sensor over the 10-min measurement period
31      to provide a generally accurate visibility measurement for within 2 to 3 mi of the site. Moisture,
       March 2001
4-99
DRAFT-DO NOT QUOTE OR CITE

-------
 1     dust, snow, rain, or particles in the light beam affect the amount of light scattered (National
 2     Weather Service, 1998).
 3          Visibility data from the ASOS network is typically reported in small increments, up to
 4     10 mi, for the purposes of airport operations. However, beginning in 1998, the raw visibility
 5     data, including light extinction measurements equaling to visual ranges exceeding 10 mi, have
 6     been archived in databases available from the National Climatic Data Center. Data for visibility
 7     at larger distances from ASOS sites are available at the sensors for only a short period of time.
 8     The data can be directly downloaded from the site. The ASOS data may be useful for
 9     characterizing visibility in urban and suburban areas across the country. It also may be used in
10     future analyses to better understand the effects of fine PM on visibility in non-class I areas.
11           The largest monitoring network that includes both visibility and air quality measurements is
12     the Interagency Monitoring of Protected Visual Environments (IMPROVE) network.  The
13     IMPROVE network was formed as a collaborative effort between the EPA and federal land
14     management agencies (National Park Service, U.S. Forest Service, Bureau of Land Management,
15     and Fish and Wildlife Service) responsible for Class I areas and the land around them (National
16     Park Service, 1998; Malm et al., 1994; Sisler et al., 1993; U.S. Environmental Protection
17     Agency, 1995a; Eldred et al., 1997; Perry et al., 1997). The primary monitoring objectives of the
18     IMPROVE program are to establish visibility levels, identify anthropogenic sources of
19     impairment, document progress towards elimination of visibility impairment in protected areas
20     from anthropogenic sources, and promote the development of visibility monitoring equipment
21     and the collection of comparable visibility data (National Park Service, 1998; Evans and
22     Pitchford, 1991).  Presently over 70 sites employ the IMPROVE program monitoring methods.
23     It is anticipated that an additional 80 sites will be added in 2000.
24           Table 4-7 contains PM25 monitoring data from 30 IMPROVE sites  for the years 1988 to
25      1996. The data includes averaged PM25 mass and specific species contributions. The data are
26     divided into eastern and western regions. The eastern regions, in addition to Washington, DC,
27     include Acadia National Park and Appalachia and consist of data from Shenandoah and the
28     Great Smoky Mountains National Parks.  The western regions include the Northern Great Plains,
29     West Texas, Sonora, the Colorado Plateau, Central Rockies, Cascade, Sierra Humbolt, West
30     Coast, Sierra Nevada, Southern California, and Alaska (Sisler and Malm, 2000).
31
        March 2001
4-100
DRAFT-DO NOT QUOTE OR CITE

-------
        TABLE 4-7. AVERAGED REGIONAL PM2.S MASS AND EXTINCTION
                   SUMMARIES FOR THE YEARS 1988 TO 1996"
Region
Alaska

Appalachia

Cascades

Colorado Plateau

Central Rockies

Coastal

Northeast

Northern Great Plains

Northern Rockies

Southern California

Sonora

Sierra Nevada

Sierra Humbolt

Washington, DC

West Texas

PM2.5
1.71
(11-9)
10.81
(97.6)
4.67
(50.6)
3.15
(17.3)
2.87
(15.8)
4.40
(43.5)
6.13
'(59.3)
4.26
(30.3)
5.15
(39.5)
8.64
(51.7)
4.09
(21.3)
4.40
(25.2)
2.67
(16.7)
16.90
(132.8)
5.11
(27.0)
Sulfate
0.55
(5.1)
6.53
(71.7)
1.30
(29.1)
1.06
(6.7)
0.80
(5.5)
1.35
(18.4)
3.32
(40.6)
1.61
(14.6)
0.98
(15.0)
1.45
(9.3)
1.52
(8.3)
0.96
(7.0)
0.52
(5.2)
7.91
(73.2)
2.13
(12.9)
Nitrate
Organics
0.06
(0.06)
0.60
(6-9)
0.23
(5.0)
0.21
(1.3)
0.18
(1.2)
0.90
(10.9)
0.40
(4.8)
0.51
(4.7)
0.31
(4.7)
3.53
(22.6)
0.24
(1.3)
0.47
(3.5)
0.16
(1.5)
2.16
(19.9)
0.25
(1.5)
Organics
0.77
(3.1)
2.73
(10.9)
2.51
(10.0)
1.08
(4.3)
1.11
(4.4)
1.65
(6.6)
1.84
(7.3)
1.35
(5.4)
2.88
(11.5)
2.29
(9.2)
1.28
(5.1)
2.16
(8.6)
1.36
(5.5)
4.44
(17.8)
1.29
(5.2)
Fine Soil
0.22
(1.0)
0.52
(4.3)
0.22
(4.1)
0.64
(1.7)
0.64
(1.4)
0.25
(2.5)
0.23
(3.4)
0.63
(1.6)
0.57
(4.1)
0.94
(4.2)
0.84
(2.0)
0.55
(2.6)
0.42
(2.0)
0.82
(15.6)
1.27
(1 7)
Elemental
Carbon
0.10
(2.2)
0.43
(3.8)
0.41
(2.3)
0.17
(3.3)
0.14
(3.2)
0.25
(5-1)
0.34
(3.0)
0.16
(4.0)
0.41
(4.1)
0.42
(6.3)
0.20
(4.6)
0.26
(3.5)
0.20
(2.5)
1.56
(6.3)
0.17
(5 7)
"Mass is in fj.g/m3. Extinction summaries in parenthesis are in Mm.

Adapted:  Sisler and Malm (2000).
March 2001
4-101
DRAFT-DO NOT QUOTE OR CITE

-------
 1          The U.S. Environmental Protection Agency is currently in the process of establishing a
 2     national PM2.S monitoring network of approximately 1,700 monitors at over 1,100 sites. The
 3     PM^s monitoring effort will be coordinated with visibility monitoring efforts currently in place,
 4     such as IMPROVE, to maximize benefits of both programs.  The monitoring network is expected
 5     to be fully implemented by the end of 2000 or shortly thereafter (U.S. Environmental Protection
 6     Agency, 1997b; U.S. Environmental Protection Agency, 2000b).
 7
 8     4.3.7 Visibility Modeling
 9           There are several types of models available for the evaluation of pollution-related effects on
10     visibility. Plume visibility models and regional haze models are source models that simulate the
11     transport, dispersion, and transformation of chemical species in the atmosphere. Plume models
12     use the resulting air quality data to calculate the values of parameters related to human
13     perception, such as contrast and color differences.  Regional haze models calculate aerosol
14     species concentrations and the light-extinction coefficient. Models for the photographic
15     representation of haze use air quality data as an input and perform the optical calculations
16     required to create images that represent the visual effects of the air quality.
17
18     4.3.7.1 Regional Haze
19           Regional haze models may be used to assess the impact of pollutant sources on an
20      identified area or region, in most cases identified class I wilderness areas, or to evaluate the
21      impact of new or existing air quality regulations. Light extinction by fine particles is used to
22      determine the effect of anthropogenic pollutants on regional visibility degradation (regional
23      haze). In the United States, these anthropogenic particles are composed primarily of sulfate
24      compounds, organic compounds, and, to a much lesser extent, nitrate compounds, with the
25      exception of California, where nitrates are the largest single contributor to light extinction. The
26      contribution to light extinction by these compounds will vary based on the particle composition
27      and size distribution. Once the particles are formed, then- size can change, resulting in a  change
 28      in their light extinction efficiency. Model calculations take into consideration the mass of the
 29      particulate constituents and the relative humidity.
 30            The model requirements for regional-scale, multiple-source haze models are nearly
 31      identical to the model requirements for simulations of regional-scale, multiple-source
        March 2001
4-102
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 fine-particle impacts. Hence, the Eulerian-based grid models currently under development to
 support fine particle impact assessments will be relied on to provide a means for assessing
 large-scale, multiple-source haze impacts.
      Middleton (1997) described the findings of a Eulerian-based grid model, the Denver Air
 Quality Model (DAQM). The DAQM is the principal component of the Brown Cloud H study
 that is part of earlier work investigating visibility in Denver over the last 20 years. The DAQM
 is derived from the Regional Acid Deposition Model (RADM) and includes aerosol processes,
 meteorological modeling analysis, and visibility analysis procedures. The DAQM has been used
 to determine the relationship between emissions and concentrations of fine and coarse particles
 and all major gaseous pollutants under various emission scenarios and meteorological
 conditions.  The results of the study demonstrated an association between visibility and air
 quality issues in the Colorado Front Range area.
      Neff (1997), in his evaluation of the DAQM model, suggested that the meteorological
 model does not address adequately mesoscale structures responsible for the initiation and
 maintenance of the brown cloud episodes or cloud systems and surface moisture fluxes. Given
 these model uncertainties, it was  suggested that there may be errors in the quantification of
 emissions and in the calculated optical extinction and scattering.
      The Visibility Assessment Scoping Model (VASM) uses Monte Carlo techniques to
 generate multiple realizations of daily concentrations of sulfates, nitrates, elemental carbon,
 organic carbon, fine and coarse dust, and the relative humidity to determine particle effects on
 regional haze.  Species-specific light attenuation is calculated based on particle concentration and
 relative humidity, producing short-term haze intensity or visual range information (Shannon
 etal., 1997).
      The Elastic Light Scattering and Interactive Efficiency (ELSIE) model was used by Omar
 et al. (1999) to determine the species concentrations and to relate apportionment to the extinction
 coefficient in an aerosol mixture.  The model assumes the aerosol is an internal inhomogeneous
mixture of chemical species and size distributions. Model input parameters included the size
distributions, prevailing relative humidity, refractive indices of the constituents, percent
solubility of the aerosol components, and the growth function of the aerosol particles.  The model
assumes that the particles grow with increasing relative humidity according to a predetermined
growth function.
       March 2001
                                         4-103
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Several source-oriented models have been developed to evaluate the effects of pollutants on
 2      regional haze. The U.S. Environmental Protection Agency, in cooperation with the U.S. Forest
 3      Service, the Fish and Wildlife Service, and the National Park Service (the Interagency
 4      Workgroup for Air Quality Modeling), developed the MESOPUFFII system of assessing
 5      regional haze impacts.  The MESOPUFF E system uses the light extinction for sulfates and
 6      nitrates for an estimated 3- to 24-h average concentration (U.S. Environmental Protection
 7      Agency, 1995b). The CALPUFF modeling system can process mesoscale meteorological data
 8      and address dispersive processes of a regional nature. Simulated long-range pollutant trajectories
 9      have been compared successfully to results from a field study involving transport to 1000 km
10     downwind (U.S. Environmental Protection Agency, 1995c). However, Lagrangian puff
11  '    dispersion modeling involving transport of 200 km or more tend to underestimate the horizontal
12     extent of the dispersion, causing the surface concentration to be overestimated (Moran and
13      Pielke, 1994). Another source-oriented Lagrangian trajectory model capable of computing light
14     extinction and scattering and estimating visual range from gas phase and primary particle phase
15     air pollutant emissions directly from sources was reported by Eldering and Cass (1996). The
16     model is comprised of several modules that take into  consideration particle size distribution and
17     chemical composition, the speciation of organic vapor emissions, atmospheric chemical
18     reactions, transport of condensible material between the gas and particle phase, fog chemistry,
19     dry deposition, and light scattering and absorption.  The model is, however, not suitable for
20     predicting visibility over great distances through nonuniform hazes and for visualization of
21     pollutant effects of isolated major point source plumes. Single line Lagrangian trajectory models
22     cannot represent horizontal turbulent diffusion, the effects of wind shear, and  advection by
23     turbulent wind components.  Error in transport calculations have been reported of up to ± 50%
24     (Eldering and Cass, 1996).
25           Gray and Cass (1998) developed a lagrangian particle-in-cell model for predicting source
26     class contributions of fine particle total carbon and elemental carbon. The model simulates the
27     motion and deposition of pollutants in an air basin with varying meteorological conditions. The
28     model also takes into consideration the vertical mixing characteristics of pollutants in areas
29      located near the source. The model is useful in determining changes in long-term average
30     pollutant concentrations from implementing specific  emission control measures.
        March 2001
4-104
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
      The Regional Particulate Model (RPM) simulates secondary fine particulate matter (PM25)
 formation and long-range transport.  The RPM is used with the Regional Acid Deposition Model
 (RADM), a comprehensive acid rain model. Predictions from the RADM are used to simulate
 the formation of sulfate and nitrate, ammonium particles, and secondary organic aerosols. The
 external RADM includes particle physics from the RPM and operates at an 80- and 20-km
 resolution. Additional work currently is being done that will incorporate the RADM/RPM and
 external RADM models into a more comprehensive air quality modeling system,
 Models-3/Community Multi-Scale Air Quality (CMAQ). This modeling system simulates the
 processes involved in primary and secondary PM!0 and PM2 5 and ozone formation, regional haze,
 acid deposition, and nutrient deposition.  The modeling system includes a mesoscale
 meteorological model, emission model, and a version of the CMAQ.
      The Regulatory Modeling System for Aerosols and Deposition (REMSAD) also simulates
 PM2 5 formation. The REMSAD was derived from the Urban Airshed Model Version V
 (UAM-V) for primary and secondary PM2 5 and PM10 formation, and acid nutrient and toxic
 deposition. The REMSAD system consists of a meteorological data preprocessor, the core
 aerosol and toxic deposition model (ATOM), and postprocessing programs. The ATOM is a
 three-dimensional Eulerian grid model designed to calculate the concentrations of both inert and
 chemically reactive pollutants  by simulating the physical and chemical processes in the
 atmosphere that affect pollutant concentrations. The basis for the model is the atmospheric
 diffusion or species continuity equation. This equation represents a mass balance in which all of
 the relevant emissions, transport, diffusion, chemical reactions, and  removal processes are
 expressed in mathematical terms (Systems Applications International, Inc., 1998).
      Zannetti et al. (1990, 1993) and Fox et al. (1997) described a semi-empirical model that
 could be used to estimate the visibility impact on one region resulting from sulfur dioxide
 emission controls in a different region.  The model combined four different input parameters:
 (1) chemical transport; (2) possible nonlinearity of pollutant chemical transformation; (3) sulfate
 fraction of fine particulate matter, including the amount of water absorbed by the fine particles;
 and (4) the fraction of light extinction caused by fine particles.  The model uses physically
realistic concepts of atmospheric transport, chemical transformation, and physical effects.
However, actual data sets, mathematical constructs, or expert opinions also maybe used.  Models
also have been developed that predict the downwind concentration of smoke particulate and other
       March 2001
                                        4-105
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1      combustion products from the burning of crude oil from accidental spills (McGrattan et al., 1995,
 2      1996).
 3
 4      4.3.7.2 Plume Models
 5           Several plume visibility models are currently available. Plume visibility models estimate
 6      the value of optical parameters related to human perception, such as contrast and color
 7      differences, and compare these values with perception thresholds to determine whether the plume
 8      is likely to be perceptible under various simulated conditions (U.S. Environmental Protection
 9      Agency, 1988; Larimer, 1988).  An empirical algorithm, Probability of Detection Algorithm
10      (PROBDET), allows the prediction of the lower limit of plume contrast that can be detected
11      visually. The PROBDET can be used to estimate the detection level for plumes that fall within
12     the bounds defined by the full-length, oval, and circular plume stimuli (Ross et  al., 1997).
13           A simplified dispersion model using a second-order turbulence closure scheme to account
14     for averaging time effects on the dispersion rate was described by Sykes and Gabruk (1997).  The
15     lateral and vertical spread is estimated using a Gaussian plume framework.  A simplified
16     representation of the turbulence spectrum is used to predict the reduced spread rate for short
17     averaging tunes.
18           Earlier plume models included PLUVUEI and H, used during the preparation of a permit
19     application to determine whether or not a proposed new facility would cause visibility
20     impairment in a Class I area (Larimer et al., 1978; Johnson et al., 1980; White et al., 1985; U.S.
21     Environmental Protection Agency, 1992). Seigneur et al. (1997) developed a plume visibility
22     model, the Reactive and Optics Model Emissions (ROME), that improves on the existing plume
23     visibility models. The model simulates the momentum and buoyancy forces of the plume rise,
24     the dispersion and chemistry, and condensation and evaporation of the aqueous phase.
25     A second-order closure algorithm is used to estimate instantaneous plume concentrations, or the
26     time-averaged plume concentration may be estimated using a first-order closure algorithm.
27      A comprehensive chemical kinetic mechanism simulates chemical transformation processes in
28      the gas, aqueous, and particle phases. Particle dynamics and chemical composition is based on
 29      sectional representation of the particle size distribution.  The model includes a  radioactive
 30      transfer module that provides optical properties using sectional particle size distributions.
 31      Deposition velocities based on atmospheric stability, surface type, chemical type, and particle
        March 2001
4-106
DRAFT-DO NOT QUOTE OR CITE

-------
  1      size are derived using a resistance-based, dry deposition algorithm. The ROME can be used with
  2      other models to estimate a stack plume opacity, the percentage of light intensity attenuated by the
  3      plume near the stack after any condensed water has evaporated (Meng et al., 2000).  When
  4      compared with the PLUVUE II, the ROME, with the second-order dispersion algorithm, was
  5      found to present a more accurate estimate of plume height, width, nitrogen oxide concentration,
  6      nitrogen dioxide/nitrogen oxide ratio, and visibility. Error, bias, correlation coefficients, and
  7      simulations were within a factor of two of that observed (Gabruk et al., 1999).
  8
  9      4.3.7.3 Photographs
10           Computer-generated photographs are sometimes used to illustrate the effects of pollution
11      on visibility. To begin, a photograph is taken on a very clean, cloud-free day to serve as the
12      initial scene image.  As previously indicated, the appearance of an object is determined by the
13      path radiance and the transmitted radiance.  To determine the transmitted radiance, an estimate of
14      the light-extinction coefficient from the photograph is used to determine the initial radiance for
15      each element in the scene. The transmitted radiance is  equal to the initial  radiance of the
16      element in the scene multiplied by the transmittance of the atmosphere in the sight path.  Because
17      the path radiance changes over the distance of the sight path, the source function, the rate of
18      change over the distance of the sight path, also must be determined.
19           Eldering et al. (1996) proposed the use of a model that uses simulated photographs  from
20      satellite and topographic images to evaluate the effect of atmospheric aerosols and gases  on
21      visibility. Use of this model requires ground-based photography and size distribution and
22      chemical composition of atmospheric aerosols, NO2 concentration, temperature, and relative
23      humidity for a clear day, for comparison purposes. Light extinction and sky color are then
24      calculated based on differences in aerosol size distribution, NO2 concentration, temperature, and
25      relative humidity.  The images created represent natural landscape elements.
26           Molenar et al. (1994) provides a discussion of existing visual air quality simulation
27      methods based on techniques under development for the past 20 years.  The WinHaze visual air
28      quality modeling system is one tool that has been developed using techniques to simulate
29      changes in visibility due to changes in air quality.
30           One of the limitations in using photographic models for representation of haze is that haze
31      is assumed to be uniformly distributed throughout the scene and selected conditions are
        March 2001
4-107
DRAFT-DO NOT QUOTE OR CITE

-------
 1     idealized, so the full range of conditions that occur in a scene are not represented. Photographs
 2     are also expensive to produce. More detailed information on the use of photographic
 3     representation of haze may be found in the U.S. Environmental Protection Agency (1996b),
 4     Trijonisetal. (1991), Molenaretal. (1994), and Elderingetal. (1993).
 5
 6     4.3.8 Trends in Visibility Impairment
 7           Trends in visibility impairment or haziness, visual range, often are associated with fine
 8     mass concentrations (^2.5 //g/m3). Observations of visual range, obtained by the National
 9     Weather Service and available through the National Climatic Data Center of the National
10     Oceanic and Atmospheric Administration, provide one of the few truly long-term, daily records
11     of any parameter related to air pollution.  After some manipulation, the visual range data can be
12     used as an indicator of fine mode particle pollution.  The data reduction process and analyses of
13     resulting trends have been reported by Husar et al. (1994), Husar and Wilson (1993), and Husar
14     etal. (1981).
15           Generally, visibility impairment is greatest in the eastern United  States and southern
16     California. Haziness in the southeastern United States is greatest in the humid summer months
17     because of its affinity to atmospheric water vapor, followed by the spring and fall, and winter.
18     Summer haziness in the southeastern United States has increased by approximately 80% since
19     the 1950s (Husar and Wilson, 1993) because of increased sulfate from increased SO2 emissions
20     (Husar et al., 1994). The resulting sulfate, considered to be ammonium sulfate, accounts for
21     40 to 70% of the fine particle mass (Husar and Wilson, 1993). Sulfate-related effects on
22     visibility in the southeast is a factor of 20 higher than the Great Basin area and 10 higher than the
23     desert southwest, central  Rocky Mountains, and Sierra Mountains (Malm et al., 1994).  For most
24     rural eastern sites, sulfates accounts for >60%  of the annual average light extinction on the best
25     days and >75% of the light extinction on the worst days. A statistically significant increase in
26     summer sulfate concentrations was noted in two class I areas in the eastern United States
27     (Shenandoah and the Great Smoky Mountains) from 1982 to 1992 (Eldred et al., 1993; Cahill
28     et al., 1996). The increase was largest in the summer and decreased in the winter.  The majority
29     of the southwest showed  decreasing sulfur (Eldred et al., 1993; Eldred and Cahill,  1994). White
30     (1997) suggested that the increase in fine-particle sulfur may be the result of the measurement
31     method and not an upward trend in fine particle concentrations in those Class I areas. However,
       March 2001                              4-108       DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
27
28
29
30
31
 Iyer et al. (2000), using the Spearman correlation of trend, reported an increased trend in hazy
 days during the summer months in Shenandoah and the Great Smoky Mountains based on
 monitoring data for the period 1979 to 1996 showing high sulfur concentrations.
      Sulfates also may be a significant contributor to total light extinction in the rural western
 United States, accounting for 30 to 40% of the total light extinction on the best days and 35 to
 45% of the total light extinction on the haziest days. In several areas of the west, sulfates account
 for over 50% of the annual average aerosol extinction (U.S. Environmental Protection Agency,
 2000b).
      Organics are the second largest contributor to light extinction in most areas in the United
 States. Extinction caused by organic carbon is greatest in the Pacific Northwest, Oregon, Idaho,
 and Montana, accounting for 40 to 45% of the total extinction. Organic carbon can contribute
 between 20 to 30% to the total extinction in most of the western United States and 10 to 15% in
 the remaining areas of the United States. Light absorption by carbon is relatively insignificant
 but is highest in the Pacific Northwest (up to  15%) and in the eastern United States (up to 6%)
 (Malm et al., 1994; U.S. Environmental Protection Agency, 2000b).
      Some of the visibility impairment in northern California and Nevada, including Oregon,
 southern Idaho and western Wyoming, results from coarse mass and soil, primarily considered
 natural extinction.  In some areas of the United States, extinction from coarse mass is almost
 negligible because the overall extinction is so high.  High dust concentrations from southern
 California have contributed to regional haze in the Grand Canyon and other class I areas in the
 southwestern United States (Vasconcelos et al., 1996).  White et al. (1999) reported that  some of
 the worst haze near the  Grand Canyon is associated with pollutant transport from southern
 California and the subtropics.
     Visibility impairment in southern California is primarily caused by light extinction by
 nitrates. Nitrates contribute about 40% to the total light extinction in Southern California and
 10 to 20% of the total extinction in other areas of the United States.
     The average haze  patterns across the continental United States,  for five-season averages for
the years 1980 to 1985 and 1990 to 1995  are shown in Figure 4-23. Haze is indicated by the
75th percentile of the extinction coefficient that is calculated from the visual range, corrected to
60% relative humidity by the Koschmeider relationship.
        March 2001
                                         4-109
DRAFT-DO NOT QUOTE OR CITE

-------
                                                                               o
                                                                               o
                                                                               o
                                                                               r-l
                                                                               a

                                                                               ts
                                                                               g
March 2001
4-110
DRAFT-DO NOT QUOTE OR CITE

-------
  1           The trends graphs in Figure 4-24 for regions in the United States represent the 75th
  2      percentile of the light extinction coefficient for the stations located within the designated region
  3      over a 30-year period (1940 to 1990). The trends are presented for quarters 1 (winter) and
  4      3 (summer). The northeastern United States exhibited an increase in haze during quarter 3
  5      between 1960 and 1970 and a steady decline between 1973 (0.22 km'1) and 1992 (0.12 km'1).
  6      In quarter 1 the haziness steadily declined from 0.15 to 0.10 km"1 in the 30-year period. The
  7      Mid-Atlantic region, the Virginias and Carolinas, shows a strong increase in haziness in quarter 3
  8      between 1960 and 1973, followed by a decline. The winter haze was virtually unchanged over
  9      the 30-year period. The haziness over the Gulf states increased between 1960 and 1970 and
10      remained virtually unchanged since then. The central Midwest, including Missouri and
11      Arkansas, exhibit virtually no change during the winter season and a slight increase in the
12      summer (1960 to 1970). The upper Midwest shows an opposing trend for summer and winter.
13      Although summer haze has increased, mostly from 1960 to 1973, the winter haze has declined.
14           Based on PM2 5 concentrations and changes in the deciview scale, calculated from
15      reconstructed extinction coefficients, Sisler and Malm (2000) reported no significant
16      deterioration in air quality and visibility conditions at 30 IMPROVE network sites for the years
17      1988 to 1996. The sites were divided into eastern and western regions. Averaged PM2 5 mass
18      and extinction summaries for the sites appear in Table 4-7. The annual best visibility
19      (10th percentile) and median visibility days (50th percentile) are improving at approximately
20      70% of the sites.  However, several sites are not showing steady improvements  in either visibility
21      or PM2 5, particularly in the number of worst visibility days (90th percentile). The sites included
22      the Badlands, Big Bend, Crater Lake, Great Smoky Mountains, Mesa Verde, Shenandoah and
23      Yosemite National Parks, Chiricahua National Monument, and the District of Columbia.
24
25      4.3.9 Economics of Particulate Matter Visibility Effects
26           Given the evidence of potential economically significant effects of visibility impairtment,
27      economic analysis proceeds by quantifying in monetary terms the costs associated with different
28      ambient levels of PM. Where possible, direct economic valuation can take place using prices
29      that are determined in the marketplace. There are a variety of ways to estimate  costs/benefits.
30      Avoided cost methods estimate the costs of pollution by using the expenditures  that are made
31      necessary by pollution damage. For example, if ambient levels of particulate matter results in
       March 2001
4-111
DRAFT-DO NOT QUOTE OR CITE

-------
                                           e japeno
                                                                             I
                                                                             C5
                                                                             03

                                                                             5

                                                                             I
                                                                             I
                                                                             in
                                                                             C
                                                                             O
                                                                             •a
                                                                             I
                                                                             N  fl
                                                                             S3  4)

                                                                             •2  "
                                                                             b  &
                                                                             £  a.
                                                                             s  js
                                                                             U  *3
                                                                             
-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
31
 increased frequency of building cleaning or repainting, then the appropriately calculated increase
 in these costs is a reasonable estimate of true economic damage.  Benefits associated with
 reductions in the pollution levels then are represented by the avoided costs of these damages.
      Estimating benefits for visibility is a more difficult and less precise exercise because the
 effects are not valued in the marketplace.  There are several methods that economists have
 developed to estimate changes in environmental effects that are not valued in the marketplace
 (Freeman, 1993). These include hedonic price analysis, stated preference models (including
 contingent valuation, contingent choice, and contingent ranking), and travel cost models.
 Hedonic price analysis works by analyzing the way that market prices change when an associated
 environmental effect changes.  Part of the economic costs imposed by the reduced visibility
 caused by PM can be estimated by looking at the differences in sales price between otherwise
 identical houses that have different degrees of visibility impairment.
      The contingent valuation method (CVM) has been used to determine estimated value
 changes in both visibility and ecosystem functions (Hanley and Spash,  1993; Chestnut, 1997).
 The CVM determines pollutant-related effects by using carefully structured surveys to estimate
 the amount of compensation equivalent to a given change in environmental quality or
 equivalently, how much they would be willing to pay to obtain a given change in environmental
 quality. There is an extensive scientific literature and body of practice  on both this theory and
 technique.
      Other valuation methods include stated preference models, including contingent choice and
 contingent ranking (also known as conjoint analysis), as well as travel cost models (Johnson and
 Desvousges,  1997; Hanley and Spash, 1993), However, the primary methods used to date for
 valuation of visibility have been the hedonic price and contingent valuation methods (Hanley and
 Spash, 1993).
      The effects of PM on visibility may differ widely between urban residential and
recreational areas.  Separate estimates are needed to account for welfare changes associated with
improvements in visibility in class I areas. Chestnut and Dennis (1997) developed a method for
estimating the value to the public of visibility improvements in class I areas using the results of a
 1990  cooperative agreement project jointly funded by the EPA and the National  Park Service:
"Preservation Values For Visibility Protection at the National Parks." Using the contingent
valuation method, Chestnut and Davis calculated a household willingness to pay for visibility
        March 2001
                                         4-113
DRAFT-DO NOT QUOTE OR CITE

-------
 1     improvements in class I areas, capturing both use and nonuse recreational values.  This analysis
 2     also accounts for geographic variations in the willingness to pay. The results indicate a
 3     willingness to pay per deciview improvement in visibility of between $5 and $ 17 per household.
 4
 5
 6     4.4  EFFECTS ON MATERIALS
 7           Effects of air pollution on materials are related to both aesthetic appeal and physical
 8     damage. Studies have demonstrated that particles, primarily consisting of carbonaceous
 9     compounds, cause soiling of commonly used building materials and culturally important items,
10     such as statutes and works of art. Physical damage from the dry deposition of air pollutants, such
11     as PM (especially sulfates and nitrates) and SO2, and the absorption or adsorption of corrosive
12     agents on deposited particles also can result hi the acceleration of naturally occurring weathering
13     processes of man-made building and cultural materials.
14           In the atmosphere, PM may be "primary", existing in the same form in which it was
15     emitted, or "secondary", formed by the chemical reactions of free, absorbed, or dissolved gases.
16     The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen ions;
17     particle-bound water; elemental carbon; a great variety of organic compounds; and crustal
18     material.  A substantial fraction of the fine particle  mass, particularly during the warmer months,
19     is secondary sulfate and nitrate. Sulfates may be formed by the gas-phase conversion of SO2 to
20     H2SO4 by OH radicals and aqueous-phase reactions of SO2 with H2O2, O3, or O2. During the day,
21     NO2 may be converted to nitric acid (HNO3) by reacting with OH radicals. Nitrogen dioxide also
22     can be oxidized to HNO3 by a sequence of reactions initiated by O3. A more detailed discussion
23     of the atmospheric chemistry of PM appears in Chapter 2 of this document.
24           Limited new studies have been published that better define the role of air pollution in
25     materials damage. This section briefly summarizes information on particle and sulfur-containing
26     pollutants (formed by the chemical reactions of SO2 with other atmospheric pollutants) exposure-
27     related effects on materials addressed in the 1996 PM AQCD (U.S. Environmental Protection
28     Agency, 1996a) and presents relevant information published since  completion of that document.
29     The effects of nitrates on manmade building materials and naturally occurring cultural materials
30     was discussed in the criteria document on nitrogen oxides (U.S. Environmental Protection
31     Agency, 1993).
        March 2001
4-114
DRAFT-DO NOT QUOTE OR CITE

-------
  1      4.4.1  Effects of Particles and Sulfur Dioxide on Man-Made Surfaces
  2      4.4.1.1 Metals
  3           Metals under go natural weathering processes in the absence of environmental pollutants.
  4      The additive effect of pollutants on the natural weathering processes will depend on the nature of
  5      the pollutant and the deposition rate (the uptake of a pollutant by the material's surface), and the
  6      presence of moisture. The influence of the metal protective corrosion film, the presence of other
  7      surface electrolytes, the orientation of the metal surface, the presence of surface moisture, and the
  8      variability in the electrochemical reactions will also contribute to the affect of pollutant exposure
  9      on metal surfaces.
10           Several studies demonstrate the importance of tune of surface wetness (caused by dew and
11      fog condensation and rain) on metals.  Surface moisture facilitates the deposition of pollutants,
12      especially SO2, and promotes corrosive electrochemical reactions on metals (Haynie and Upham,
13      1974; Sydberger and Ericsson, 1977).  Of critical importance is the formation of hygroscopic
14      salts on the metal that increases the time of surface wetness and, thereby, enhances the corrosion
15      process.
16           Pitchford and McMurry (1994) and Zhang et al. (1993) demonstrated particle size-related
17      effects of relative humidity.  The effect of temperature on the rate of corrosion is complex.
18      Under normal temperature conditions, temperature would not have an affect on the rate of
19      corrosion. When the temperature decreases the relative humidity increases and the diffusivity
20      decreases. The corrosion rate decreases as the temperature approaches freezing because ice
21      prohibits the diffusion of SO2 to the metal surface and minimizes electrochemical processes
22      (Haynie, 1980; Biefer, 1981; Sereda, 1974).
23           The metal protective corrosion film (i.e., the rust layer on metal surfaces) provides some
24      protection against further corrosion. The effectiveness of the corrosion film in slowing down the
25      corrosion process is affected by the solubility of the corrosion layer, and the concentration and
26      deposition rate of pollutants. If the metal protective corrosion film is insoluble, it may add some
27      protection against acidic pollutants. An atmospheric corrosion model that considers the
28      formation and dissolution of the corrosion film on galvanized steel was proposed by Spence et al.
29      (1992). The model considers the effects of SO2, rain acidity, and the time of wetness on the rate
30      of corrosion. Although the model does  not characterize specifically particle effects, the
31      contribution of particulate sulfate was considered in model development.
        March 2001
4-115
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Whether suspended particles actually impact on the corrosion of metals is not clear.
 2      Several studies suggest that suspended particles will promote the corrosion of metals (Goodwin
 3      et al., 1969; Barton, 1958; Sanyal and Singhania, 1956; Baedecker et al., 1991); however, other
 4      studies have not demonstrated a correlation between particle exposure and metal corrosion
 5      (Mansfeld, 1980; Edney et al., 1989). Walton et al. (1982) suggested that catalytic species within
 6      several species in fly ash promote the oxidation of SOX to a corrosive state. Still other
 7      researchers indicate that the catalytic effect of particles is not significant, and that the corrosion
 8      rate is dependent on the conductance of the thin-film surface electrolytes during periods of
 9      wetness.  Soluble particles likely increase the solution conductance (Skerry et al., 1988; Askey
10      etal., 1993).
11           The corrosion of most ferrous metals (iron, steel, and steel alloys) is increased by
12      increasing SO2 exposure.  Steels are susceptible to corrosion when exposed to SO2 in the absence
13      of protective organic or metallic coatings. Studies on the corrosive effects of SO2  on steel
14      indicate that the rate of corrosion increases with increasing SO2 and is dependent on the
15      deposition rate of the SO2 (Baedecker et al., 1991; Butlin et al., 1992a). The corrosive effects of
16      SO2 on aluminum is exposure-dependent, but appears to be insignificant (Haynie,  1976; Fink
17      et al., 1971; Butlin  et al., 1992a).  The rate of formation of the patina on copper (protective
18      covering) can take as long as 5 years and is dependent on the SO2 concentration, deposition rate,
19      temperature, and relative humidity (Simpson and Horrobin, 1970). Further corrosion is
20      controlled by the availability of copper to react with deposited pollutants (Graedel et al., 1987).
21      Butlin et  al. (1992a), Baedecker et al. (1991), and Cramer et al. (1989) reported an average
22      corrosion rate of 1 //in/year for copper; however, less than a third of the corrosion was attributed
23      to SO2 exposure, suggesting that the rate of patina formation was more dependent on factors
24      other than SO2. A recent report by Strandberg and Johansson (1997) showed relative humidity to
25      be the primary factor in copper corrosion and patina formation. The results of the studies on
26      particles  and SO2 corrosion of metals are summarized in Table 4-8.
27
28      4.4.1.2 Painted Finishes
29           Exposure to air pollutants affect the durability of paint finishes by promoting discoloration,
30      chalking, loss of gloss, erosion, blistering, and peeling. Evidence exists that indicates particles
31      can damage painted finishes by serving as carriers for corrosive pollutants (Cowling and Roberts,
        March 2001
4-116
DRAFT-DO NOT QUOTE OR CITE

-------
         ^ .
CB §
CQ U

JJ 0§
tl"^
Average corrosion rate ranged from 0.63 to
1 .33 //rn/y. The highest corrosion was noted i
most industrialized area. However, the corros
rates did not differ significant regardless of th
concentration, suggesting that SO2 exposure n
be the dominant factor in zinc corrosion.
0
illil
•1311
L. on -H --, „
Specimens exposed at 5 sites fo:
Average S02 concentrations ran
15 ±17 ppb (5.2 ±10.4 to 39.3
PM concentrations ranged from
Highest pollutant concentratiom
1 year exposure site.




u
B
N
Is

^ OS

W ^

-------
CO
H*j
^7
H
w
1
fy^
Q
O
Q
1
&
1
1
S
1
1
5*
CO
Q
S
QS
O
U

_y
•s
"S
o
oo
4
H
I


>,
43
CX
ob
2
3



Comments

_o
•3
1
3
VI
a
UJ











u
S
5\
o\

s~~/
"3
ts
•a
C4
ffl
8(5" K
M P t/3 t^
Average corrosion rate for 3- and 5-year exposure
was about 1 /an/year but the soluble portion was 1
than a third of that which could be contributed to
exposure. Dry deposition of S02 was not as
important in patina formation as wet deposition oi
S
=>
g
1
o
•§
Ov
^ — >
"3
t>
Ui
0
^
•§ el

^3
en

Majority of test sites showed a corrosion rate of
1 ± 0.2 //m/year. The corrosion rate was
1.48 /urn/year at the site receiving the most rainfal
The lowest corrosion rate, 0.66 //m/year, was
associated with low rainfall, low S02.

§
'S
ex
1
Sf

ON
^
"3
u
1.8
*5 £3
CQ is
4> C
U O
C/l 0





u
1
u


£^*
t3 O\
CO ^
s°§
11
•fc* *n
OO i-^
^>
S02 had no effect on copper when relative humidi
was «75%. Increasing relative humidity increase:
patina formation in presence of trace S02.
No S02-related effects were noted on copper
specimens exposed to high S02 regardless of the
percent relative humidity.

•2-C »
!!l
®5I
3 P —
Ifi
•§-••§

8* ts '
^ 'T-« _c
ai S O
I'll 7!
tt ""^ -O
*S ^

S r~- '-e

-------
  1      1954) or by staining and pitting of the painted surfaces (Fochtman and Langer, 1957; Wolff et al.,
  2      1990).
  3           The erosion rate of oil-base house paint has been reported to be enhanced by exposure to
  4      SO2 and high humidity. In a study by Spence et al. (1975), an erosion rate of 36.71 ±
  5      8.03 yum/year was noted for oil-base house paint samples exposed to SO2 (78.6 Aig/m3), O3
  6      (156.8 Aig/m3), and NO2 (94 Afg/m3) and low humidity (50%). The erosion rate increased with
  7      increased SO2 and humidity. The authors concluded that SO2 and humidity accounted for 61% of
  8      the erosion.  Acrylic coil coating and vinyl coil coating shows less pollutant-related erosion.
  9      Erosion rates range from 0.7 to 1.3 yum/year and 1.4 to 5.3 /zm/year, respectively.  Similar
10      findings on SO2-related erosion of oil-base house paints and coil coatings have been reported by
11      other researchers (Davis et al., 1990; Yocom and Grappone, 1976; Yocom and Upham, 1977;
12      Campbell et  al., 1974).  Several studies suggest that the effect of SO2 is caused by its reaction
13      with extender pigments such as calcium carbonate and zinc oxide (Campbell et al., 1974; Xu and
14      Balik, 1989;  Edney, 1989; Edney et al., 1988, 1989). However, Miller et al. (1992) suggested
15      that calcium  carbonate acts to protect paint substrates. Another study indicated that exposure to
16      SO2 can increase the drying time of some paints by reacting with certain drying oils and will
17      compete with the auto-oxidative curing mechanism responsible for crosslinking the binder
18      (Holbrow, 1962).
19
20      4.4.1.3 Stone and Concrete
21           Numerous studies suggest that air pollutants can enhance the natural weathering processes
22      on building stone. The development of crusts on stone monuments have been attributed to the
23      interaction of the stone's surface with sulfur-containing pollutants, wet or dry deposition of
24      atmospheric particles, and dry deposition of gypsum particles from the atmosphere. Because of a
25      greater porosity and specific surface, mortars have a greater potential for reacting with
26      environmental pollutants (Zappia et al., 1998). Details on these studies are discussed hi
27      Table 4-9.  The stones most susceptible to the deteriorating effects of sulfur-containing pollutants
28      are the calcareous stones (limestone, marble,  and carbonated cement). Exposure-related damage
29      to building stones result from the formation of salts in the stone that are subsequently washed
30      away during rain events leaving the stone surface more susceptible to the effects of pollutants.
31      Dry deposition of sulfur-containing pollutants promotes the formation of gypsum on the stone's
        March 2001
4-119
DRAFT-DO NOT QUOTE OR CITE

-------
       i_i

      .2
Exposure Co

C3


tj
                  S

                £«s i o"
                x S= ,2 S O
          OQ
          II
C  o
o^

S  2

c •«
i  s
          tl
             a

          >

                              i
   O
   



                                                                      6 ts
                                                            C
-§ a



                                                                '
             HI


             |oal
             <; oo o

                                      1   s-si
March 2001
                              4-120
                                    DRAFT-DO NOT QUOTE OR CITE

-------





E MATTER AND SULFUR DIOXIDE ON STOP
H
; 4-9 (cont'd). EFFECTS OF PARTICULA
w
1
H



>>
•s.
1
1

I
0

Exposure Conditions




ID
§
1

"cS
•s
IS
H O\
^^ ^^
In the absence of moisture, little reaction is seen. SO2 is
oxidized to sulfates in the presence of moisture. The effect
is enhanced in the presence of O3. Massangis Jaune Roche
limestone was the least affected by the pollutant exposure.
Crust lined pores of specimens exposed to SO2.

Samples exposed to S02, N02, and NO at 10 ppmv,
both with and without 03 and under dry (coming to
equilibrium with the 84% RH) or wetted with
C02-equilibrated deionized water conditions.
Exposure was for 30 days.

u
I g g2
to 3 B u
§ i 3 IB
g 1— > CO M
i^ '§>.§ 5! .2
| |||J
«S-S«§ll

o^
OS
OS
CO
•>"
Significant amounts of gypsum were noted on the Portland
stone. Sheltered stones also showed soiling by
carbonaceous particles and other combustion products.
Etch holes and deep etching was noted in some of the
exposed unsheltered samples.

Samples exposed for 2 mo under both sheltered and
unsheltered conditions. Mean daily atmospheric S02
concentration was 68.7 /ug/m3 and several heavy
rainfalls.

!
CO
S3 3s
T3
£ §
§|

*
j_.
'I
O s~**
•§ Os
"a °*
oo C-
Exposure to particles from combustion processes enhanced
sulfation of calcareous materials by S02 because of metal
content of particles.

Sample exposed in laboratory to 3 ppm SO2 and 95%
RH at 25 °C for 150 days. Samples were coated with
three carbonaceous particle samples from combustion
sources, and with activated carbon and graphite.


JU
1 8 8
ail '

UEHft.
a
•§
!| K1
oi O
Pollutant exposed samples showed increased weight gain
over that expected from natural weathering processes.
There was a blackening of stone samples exposed to
carbonaceous rich particulate matter.

Samples exposed for 6 mo (cold and hot conditions)
in ambient environment. PM concentrations ranged
from 57.3 to 1 16.7 yug/m3 (site 1) and 88 to
189.8 /ug/m3 (site 2). Some exposures also were
associated with high SO2, NO, and N02.


J2
i
1
u

s /~s
la
vG
•g a
35 t3
Exposure to fly-ash did not enhance oxidation of SO2 to
sulfates. Mineral oxides in fly ash contributed to sulphation
ofCaC03.

Samples artificially exposed to fly-ash containing
1309.3 Mg/m3 SO2 (0.5 ppm), at 95% RH and 25 °C
for 81 or 140 days. Fly-ash samples from five
different sources were used in study.

I
"w
OH O *^
j | 8
J-§l
•=; s ftn
March 2001
4-121
DRAFT-DO NOT QUOTE OR CITE

-------
04
CTS OF P
E
 a
 o
 u

5
6d
ffi
H
J?
1
3
S







Comments









CO
asure Condition
g-
u






g
I

*C3
'I.?
*!
i)
ji "E.
^O . *. "*"* g
CO *+-» C3
CO c; O co
'•C .S o **3
C O *O '| i
« 
O 42 13
oo "E-3
tlj
Q. OT CO
,856Aig/m3(3p;
60, or 90 days;
every 7 days to
oSS
* V. "
Igl
c? <_) u
P<° S3
Isl
fl*
C3 ^^ *^
00 K f
Ib
1 §1
1 c E
g -2 'c
3 (S o
                    u
                    o
                    JO
                         CO
                    g    S
                    o   eg
                    M =^ S
                    ^•3-
                    B '-g .g «*
                    §• g_ i £*
                    g §-33 o
                    cfc g    (D
                    on M« co !2
                    S "S  S §
                    P Ji « o
                    Jl^JJ
                    ^3 u  8 S3
                    "s  i-i
                      iln>
                      g .g> g)

                      HI
-  £
   T3
   S
                       « J§
n
fo
a

l
to
xp
Samples from structures exposed for varying periods of Black layers were found to be primarily comprised of iron Nord and
sited time under ambient air conditions. Samples selected compounds, quartz, silicate, soot, and dirt. Ericsson (1993)
because of black layer on surface.
ented

Limestone
5
§
3
B
sandstone
Calcite-cei

sandstone
Granite
Brick
Samples of ancient grey crust formed between 1 1 80 Crust samples contained calcite, soil dust, carbonaceous Ausset et al.
and 1636 on the Church of Saint Trophime in Arks and particles, and gypsum crystals. (1998)
formed between 1 530 and 1 1 87 on the Palazz
d'Accursio in Bolonga.
ble Samples of the stones and mortars were representative Mortars were more reactive than the stones. Of the Zappia et al.
narble of those used in the past and currently for new mortars, cement and pozzolan mortar were more reactive (1998)
tone construction and restorations. Samples were exposed than the lime mortar. Carrara marble was the least
lestone for 6, 12, and 24 mo under ambient conditions in reactive of the stones. The maximum amount of
r Milan. degradation was found in areas sheltered from rain.

Limestone

Sandstone
ca
1
U
1—
Travertine
W5
1
c
Portland li
CO
Lime mort
1 |
C o
Pozzolan t
Cement m
March 2001
                      4-122
DRAFT-DO NOT QUOTE OR CITE

-------




rV]
^
O
H
O3
'j^.
LTE MATTER AND SULFUR DIOXIDE O
^
ja
(j
6
^
c^
te
O
«3
H
O
w
is
w
•
^^s
•^_)
a
o

9s
4
2
H




1
_o

1
n





Comments






CA
g
•s
••3
§
u
E
3
1
w











I
CO
"ca
ts
'S
o /*"••*

•S o^
co C-


S
3
. Exposure to environmental pollutants caused the
formation of two separate layers on the mortar: an o
thin surface black crust composed of gypsum and
carbonaceous particles and the inner composed of
products from the dissolution and sulphation of the
carbonate matrix in the mortar.
ts
O
fe
£
'5
2
g
N
e
g
1
S
g
^
C3
2)
Cl_,
O
*p,

CO

1
e
1
3
"w
t3

.2 »
S'o
OQ Ci

t*-,
o
'S 2
Gypsum main component of crust followed by
carbonaceous particles and iron oxides. Estimated r
crust formation was 2-5 ^m/year. Total amount of
gypsum formed over the lifetime of exposure was 5
13 mg/cm2, an estimated 0.2 mg/cm2/year.



1
 -^
1 "S
S o
to OS
S 4i
° "3
o *^
3^"

1 o
coK
_o
to
03
1
U
March 2001
4-123
DRAFT-DO NOT QUOTE OR CITE

-------
 1     surface. Gypsum is a gray to black crusty material comprised mainly of calcium sulfate
 2     dihydrate from the reaction of calcium carbonate (calcite) in the stone with atmospheric SO2 and
 3     moisture (relative humidities exceeding 65%). Approximately 99% of the sulfur in gypsum is
 4     sulfate because of the sulphation process caused by the deposition of SO2 aerosol.  Sulphites also
 5     are present in the gypsum layer as an intermediate product (Sabbioni et al., 19961; Ghedini et al.,
 6     2000; Gobbi et al.,  1998; Zappia et al., 1998). Gypsum is more soluble than calcite and is known
 7     to form on limestone, sandstones, and marble when exposed to SO2. Gypsum also has been
 8     reported to form on granite stone by replacing silicate minerals with calcite (Schiavon et al.,
 9     1995). Gypsum occupies a larger volume than the original stone, causing the stone's surface to
10     become cracked and pitted.  The rough surface serves as a site for deposition of airborne
11     particles.
12          The dark colored gypsum is caused by surface deposition of carbonaceous particles
13     (noncarbonate carbon) from combustion processes occurring in the area (Sabbioni, 1995;
14     Saiz-Jimenez, 1993; Ausset et al., 1998), trace metals contained in the stone, dust, and numerous
15     other anthropogenic pollutants. After analyzing damaged layers of several stone monuments,
16     Zappia et al. (1993) found that the dark-colored damaged surfaces contained 70% gypsum and
17     20% noncarbonate carbon. The lighter colored damaged layers were exposed to rain and
18     contained 1 % gypsum and 4% noncarbonate carbon. It is assumed that rain removes reaction
19     products, permitting further pollutant attack of the stone monument, and likely redeposits some
20     of the reaction products at rain runoffs sites on the stone. Following sulfur compounds, carbon
21     was reported to be  the next highest element in dark crust on historical monuments in Rome.
22     Elemental carbon and organic carbon accounted for 8 and 39% of the total carbon in the black
23     crust samples. The highest percentage of carbon, carbonate carbon, was caused by the  carbonate
24     matrix in the stones.  The high ratio of organic carbon to elemental carbon indicates the presence
25     of a carbon source other than combustion processes (Ghedini et al., 2000). Cooke and  Gibbs
26     (1994) suggested that stones damaged during times of higher ambient pollution exposure likely
27     would continue to exhibit a higher rate of decay, termed the "memory effect", than newer stones
28     exposed under lower pollution conditions. Increased stone damage also has been associated with
29     the presence of sulfur oxidizing bacteria and fungi on stone surfaces (Garcia-Valles et al., 1998;
30     Young, 1996; Saiz-Jimenez, 1993; Diakumaku et al., 1995).
        March 2001
4-124
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
      Dissolution of gypsum on the stone's surface initiates structural changes in the crust layer.
 Garica-Valles et al. (1998) proposed a double mechanism; the dissolution of the gypsum, in the
 presence of sufficient moisture, followed by recrystallization inside fissures or pores. In the
 event of limited moisture, the gypsum in dissolved and recrystallizes at its original location.
 According to the authors, this would explain the gypsum-rich crustal materials on stone surfaces
 sheltered from precipitation.
      Moisture was found to be the dominant factor in stone deterioration for several sandstones
 (Petuskey et al., 1995). Dolske (1995) reported that the deteriorative effects of sulfur-containing
 rain events, sulfates, and SO2 on marble were largely dependent on the shape of the monument or
 structure rather than the type of marble. The author attributed the increased fluid turbulence over
 a nonflat vertical surface versus a flat surface to the increased erosion. Sulfur-containing
 particles also have been reported to enhance the reactivity of Carrara marble and Travertine and
 Trani stone to SO2 (Sabbioni et al., 1992). Particles with the highest carbon content had the
 lowest reactivity.
      The rate of stone deterioration is determined by the pollutant and the pollutant
 concentration, the stone's permeability and moisture content, and the pollutant deposition
 velocity. Dry deposition of SO2 between rain events has been reported to be a major causative
 factor in pollutant-related erosion of calcareous stones (Baedecker et al., 1991; Dolske, 1995;
 Cooke and Gibbs, 1994; Schuster et  al., 1994; Hamilton et al., 1995; Webb et al., 1992). Sulfur
 dioxide deposition increases with increasing relative humidity (Spiker et al., 1992), but the
pollutant deposition velocity is dependent on the stone type (Wittenburg and Dannecker, 1992),
the porosity of the stone, and the presence of hygroscopic contaminants.
     Although it is clear from the available information that gaseous pollutants, in particular dry
deposition of SO2 will promote the decay of some types of stones under the specific conditions,
carboneous particles (noncarbonate carbon) may help to promote the decay process by aiding in
the transformation of SO2 to a more acidic species (Del Monte and Vittori, 1985). Several
authors have reported enhanced sulfation of calcareous material by SO2 in the presence of
particles containing metal oxides (Sabbioni et al., 1996; Hutchinson et al., 1992).
        March 2001
                                         4-125
DRAFT-DO NOT QUOTE OR CITE

-------
 1     4.4.2  Soiling and Discoloration of Man-Made Surfaces
 2          Ambient particles can cause soiling of man-made surfaces.  Soiling has been defined as the
 3     deposition of particles of less than 10 //m on surfaces by impingement. Soiling generally is
 4     considered an optical effect, that is, soiling changes the reflectance from opaque materials and
 5     reduces the transmissions of light through transparent materials. Soiling can represent a
 6     significant detrimental effect requiring increased frequency of cleaning of glass windows and
 7     concrete structures, washing and repainting of structures, and, in some cases, reduction in the
 8     useful life of the object. Particles, in particular carbon, also may help catalyze chemical reactions
 9     that result in the deterioration of materials during exposure.
10          It is difficult to determine the accumulated particle levels that cause an increase in soiling;
11     however, soiling is dependent on the particle concentration in the ambient environment, particle
12     size distribution, and the deposition rate and the horizontal or vertical orientation and texture of
13     the surface being exposed (Haynie, 1986).  The chemical composition and morphology of the
14     particles and the optical properties of the surface being soiled will determine the time at which
15     soiling is perceived (Nazaroff and Cass, 1991). Carey (1959) reported that the average observer
16     could observe a 0.2% surface coverage of black particles on a white background. A recent study
17     suggest that it would take a 12% surface coverage by black particles before there is 100%
18     accuracy in identifying soiling (Bellan et al., 2000).  The rate at which an object is soiled
19     increases linearly with time; however, as the soiling level increases, the rate of soiling decreases.
20     The buildup of particles on a horizontal surface is counterbalanced by an equal and opposite
21     depletion process. The depletion process is based on the scouring and washing effect of wind
22     and rain (Schwar, 1998).
23
24     4.4.2.1 Stones and Concrete
25           Most of the research evaluating the effects of air pollutants on stone structures have
26     concentrated on gaseous pollutants. The deposition of the sulfur-containing pollutants are
27     associated with the formation of gypsum on the stone (see  Section 4.4.1.3). The dark color of
28     gypsum is attributed to soiling by carbonaceous particles from nearby combustion processes.
29     A lighter gray colored crust is attributed to soil dust and metal deposits (Ausset et al., 1998;
30     Camuffo, 1995; Moropoulou et al., 1998).  Realini et al. (1995) found the formation of a dark
31     gypsum layer and a loss of luminous reflection in Carrara marble structures exposed for 1 year
        March 2001
4-126
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 under ambient air conditions. Dark areas of gypsum were found by McGee and Mossitti (1992)
 on limestone and marble specimens exposed under ambient air conditions for several years. The
 black layers of gypsum were located in areas shielded from rainfall. Particles of dirt were
 concentrated around the edges of the gypsum formations.  Lorusso et al. (1997) attributed the
 need for frequent cleaning and restoration of historic monuments in Rome to exposure to total
 suspended particulates. They also concluded that, based on a decrease in brightness (graying),
 surfaces are soiled proportionately over time; however, graying is higher on horizontal surfaces
 because of sedimented particles. Davidson et al. (2000) evaluated the effects of air pollution
 exposure on a limestone structure on the University of Pittsburgh campus using estimated
 average TSP levels in the 1930s and 1940s and actual values for the years 1957 to 1997.
 Monitored levels of SO2 were available for the years 1980 to 1998. Based on the available data
 on pollutant levels and photographs, it was thought that soiling began while the structure was
 under construction. With decreasing levels of pollution, the soiled areas have been slowly
 washed away, the process taking several decades, leaving a white, eroded surface. Studies
 describing the effects of particles on stone surfaces are discussed in Table 4-9.

 4.4.2.2  Household and Industrial Paints
      Few studies are available that evaluate the soiling effects of particles on painted surfaces.
 Particles composed of elemental carbon, tarry acids, and various other constituents are
 responsible for soiling of structural painted surfaces. Coarse-mode particles (>2.5 //m) initially
 contribute more soiling of horizontal and vertical painted surfaces than do fine-mode particles
 (<2.5 /^m), but are more easily removed by rain (Haynie and Lemmons, 1990). The
 accumulation of fine particles likely promotes remedial action (i.e., cleaning of the painted
 surfaces). Coarse-mode particles are primarily responsible for soiling of horizontal surfaces.
 Rain interacts with coarse particles, dissolving the particle and leaving stains on the painted
 surface (Creighton et al., 1990; Haynie and Lemmons, 1990). Haynie and Lemmons (1990)
proposed empirical predictive equations for changes in surface reflectance of gloss-painted
 surfaces that were exposed protected and unprotected from rain and oriented horizontally and
vertically.
     Early studies by Parker (1955) and Spence and Haynie (1972) demonstrated an association
between particle exposure and increased frequency of cleaning of painted surfaces.  Particle
        March 2001
                                         4-127
DRAFT-DO NOT QUOTE OR CITE

-------
 1     exposures also caused physical damage to the painted surface (Parker, 1955). Unsheltered
 2     painted surfaces are initially more soiled by particles than sheltered surfaces but the effect is
 3     reduced by rain washing.  Reflectivity is decreased more rapidly on glossy paint than on flat paint
 4     (Haynie and Lemmons, 1990). However,  surface chalking of the flat paint was reported during
 5     the exposure.  The chalking interfered with the reflectance measurements for particle soiling.
 6     Particle composition measurements that were taken during exposure of the painted surfaces
 7     indicated sulfates to be a large fraction of the fine mode and only a small fraction of the coarse
 8     mode.  Although no direct measurements were taken, fine mode particles likely also contained
 9     large amounts of carbon and possibly nitrogen or hydrogen (Haynie and Lemmons, 1990).
10
11
12     4.5  EFFECTS OF ATMOSPHERIC PARTICIPATE MATTER ON
13          CLIMATE CHANGE PROCESSES AND THEIR POTENTIAL
14          HUMAN HEALTH AND ENVIRONMENTAL IMPACTS
15          Global climate change processes and their potential human health and environmental
16     impacts have been accorded extensive attention during the past several decades, and they still
17     continue to be of broad national and international concern.  This is reflected by extensive
18     research and assessment efforts undertaken since the mid-1970s by U.S. Federal Government
19     Agencies (e.g., NOAA, EPA, CDC, etc.) or via U.S. Federal hiteragency programs (e.g., the U.S.
20     Global Climate Change Research Program [USGCRP]) and by analogous extensive research and
21     assessment efforts undertaken by numerous other national governments or international
22     collaborative activities (e.g., those coordinated by the Intergovernmental Panel on Climate
23     Change [IPCC], established in the  1980s under the joint auspices of the World Meteorological
24     Organization [WMO], and the United Nations Environment Programme [UNEP]).
25          Atmospheric particles play important roles in two key types of global climate change
26     processes or phenomena: (1) alterations in the amount of solar radiation in the ultraviolet range
27     (especially UV-B) penetrating through the Earth's atmosphere and reaching its surface, where it
28     can exert a variety of effects on human health, plant and animal biota, and other environmental
29     components; and (2) alterations in the amount of solar radiation in the visible range being
30     transmitted through Earth's atmosphere and either being reflected back into space or absorbed
31     (together with trapping of infrared radiation emitted by the Earth's surface by certain gases),
       March 2001
4-128
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28'
29
30
31
which enhances heating of the Earth's surface and lower atmosphere (i.e., the widely-known
"greenhouse effect") and leads to consequent "global warming" impacts on human health and the
environment. Atmospheric particles also play a lesser role by absorbing infrared radiation
emitted by the Earth's surface.
      The effects of atmospheric PM on the transmission of electromagnetic radiation emitted by
the sun at ultraviolet and visible wavelengths and by the earth at infrared wavelengths depend on
the radiative properties (extinction efficiency, single scattering albedo, and asymmetry
parameter) of the particles, which are, in turn, dependent on the size and shape of the particles,
the composition of the particles arid the distribution of components within individual particles.
In general, the radiative properties of particles are size and wavelength dependent.  In addition,
the extinction cross-section tends to be at a maximum when the particle radius is similar to the
wavelength of the incident radiation. Thus, fine particles present mainly in the accumulation
mode would be expected to exert a greater influence on the transmission of electromagnetic
radiation than would coarse particles. The composition of particles can be crudely summarized
in terms of the broad classes identified in Chapter 6 of the  1996 PM AQCD and recapitulated in
Chapter 2 of this document (e.g., fine particles mainly consisting of nitrate, sulfate, mineral dust,
elemental carbon,  organic carbon compounds [e.g., PAHs], and metals derived from high
temperature  combustion or smelting processes).  The major sources of these components are
shown in Table  2.1 of Chapter 2 in this document.
     Knowledge of the factors controlling the transfer of solar radiation in the ultraviolet
spectral region is needed for assessing the potential biological and environmental impacts
associated with  exposure to UV-B radiation (290 to 315 nm). Knowledge of the effects of PM
on the transfer of radiation in the visible and infrared spectral regions is needed for assessing the
relation between particles and global warming and its environmental and biological impacts.
Key information regarding important conceptual aspects and factors related to solar ultraviolet
radiation processes and effects is summarized first below and atmospheric PM roles noted,
followed by summarization of global wanning processes, their potential human health and
environmental impacts, and potential relationships to atmospheric PM.
       March 2001
                                        4-129
DRAFT-DO NOT QUOTE OR CITE

-------
 1      4.5.1 Solar Ultraviolet Radiation Transmission Impacts on Human Health
 2            and the Environment:  Atmospheric Particulate Matter Effects
 3      4.5.1.1 Bases for Concern Regarding Increased Ultraviolet Radiation Transmission
 4           The transmission of solar UV-B radiation through the earth's atmosphere is controlled by
 5      ozone, clouds, and particles. The depletion of stratospheric ozone caused by the release of
 6      anthropogenically produced chlorine (Cl)-and bromine (Br)-containing compounds has resulted
 7      in heightened concern over potentially serious increases in the amount of solar UV-B radiation
 8      (SUVB) reaching the Earth's surface. SUVB is also responsible for initiating the production of
 9      OH radicals that oxidize a wide variety of volatile organic compounds, some of which can
10      deplete stratospheric ozone (e.g., CH3C1, CH3Br), absorb terrestrial infrared radiation (e.g., CH4 ),
11      and contribute to photochemical smog formation (e.g., C2H4  , C5H8 ).
12          Increased penetration of SUVB to the Earth's surface as the result of stratospheric ozone
13      depletion continues to be of much concern because of projections of consequent increased
14     surface-level SUVB exposure and associated potential negative impacts on human health, plant
15     and animal biota, and man-made materials. Several summary overviews (Kripke, 1989; Grant,
16     1989; Kodama and Lee, 1993; Van der Leun et al., 1995,1998) of salient points related to
17     stratospheric ozone depletion processes and bases for concern provide a concise introduction to
18     the subject, as does Figure 4-25. As shown to the left in the  figure, stratophospheric ozone
19     depletion results from:  (a) anthropogenic production and associated emission into the lower
20     atmosphere of certain trace gases having long atmospheric residence times (e.g.,
21     chlorofluorocarbons [CFCs], carbon tetrachloride [CC14], and Halon 1211 [CF2C1 Br] and 1301
22     [CF3Br], which have atmospheric residence times of 75 to 100 years, 50 years, 25 years, and
23     110 years, respectively); (b) their tropospheric accumulation and gradual transport, over decades,
24     up to the stratosphere, where (c) photodissociation processes release Cl and Br, that (especially
25     under very cold subzero upper atmospheric conditions) catalyze ozone reduction; leading to
26     (d) stratospheric ozone depletion that is most marked over Antarctica during Southern
27     Hemisphere wintertime, to a less marked but still significant extent over the Arctic Polar Region
28     during Northern Hemisphere wintertime, and to a lesser extent over mid-latitude regions during
29     any season.
30          Given the long tune involved in transport of such gases to the stratosphere and their long
31     residence times there, any effects already seen on stratospheric ozone are likely caused by the
        March 2001
4-130
DRAFT-DO NOT QUOTE OR CITE

-------
                 BASES FOR CONCERN ABOUT STRATOSPHERIC OZONE DEPLETION
                         DUE TO CFC's, HALONS, AND OTHER TRACE GASES
             Stratospheric
           Ozone Depletion
            Cl, Br Catalyze
           Ozone Reduction
           Photodissociation
          Releases Cl and Br
            Slow Transport
            to Stratosphere
            Tropospheric
            Accumulation
           Air Emissions of
          CFC's, Hajons, etc.
             OZONE DEPLETION EFFECTS
[CFC's & O3 Column
  Reorganization
                              Climate Changes:
                              Temp., Winds, eta
                               Increase of Air
                             Stagnation Periods
 Accumulation of
 Tropospheric O3
and Acid Aerosols
               Increased UV-B Light
               Penetration to Surface
 Man's Production
 of CFC's, Halons,
Other Trace Gasses
                           Environmental
                           Effects: Crop,
                           Forest Damage
           Human
           Health
           Effects
Infectious
Diseases
Increased
                                                      Altered Bio-
                                                      geochemical
                                                        Cycling
                                                          UV-B Radiation Direct
                                                          Human Health Impacts
                                Natural Ecosystem
                                  and Agriculture
                                    Impacts
                              Skin Damage
                               (Sunburn)
                       Damage
                        to Eye
Skin Cancer
 Premature
 Skin Aging
                        Terrestrial
                      Ecosystem Shifts
                     Lower Crop Yields
                                                                                      AND
Cataracts
Incidence
Increased
   Aquatic
Ecosystem Shifts
Less Plankton &
   Seafood
       Figure 4-25. Processes involved in stratospheric ozone depletion because of man's
                    production of CFCs, halons, and other trace gases are shown to the left.  The
                    types of effects caused by stratospheric ozone depletion and consequent
                    increased UV-B penetration to the Earth's surface are hypothesized to include
                    both direct effects on human health (e.g., increased cancer rates, immune
                    suppression, etc.) and other terrestrial and aquatic ecological effects resulting
                    from increased UV-B alterations of biogeochemical cycles.

       Source: Adapted from Grant (1989).
1      atmospheric loadings of trace gases from anthropogenic emissions several decades ago, and those

2      gases already in the atmosphere may continue to exert stratospheric ozone depletion effects well

3      into the 21st century. Shorter lived gases, such as CH3Br, also exert significant ozone depletion

4      effects.
       March 2001
                   4-131
             DRAFT-DO NOT QUOTE OR CITE

-------
 1           The main types of effects hypothesized as likely to result from stratospheric ozone
 2     depletion and consequent increased SUVB penetration through the Earth's atmosphere include
 3     the following.
 4     (1) Direct Human Health Effects, such as skin damage (sunburn), leading to more rapid aging
 5         and increased incidence of skin cancer; ocular effects (retinal damage and increased cataract
 6         formation possibly leading to blindness); and suppression of some immune system
 7         components (contributing to skin cancer induction and spread to nonirradiated skin areas, as
 8         well as possibly increasing susceptibility to certain infectious diseases or decreasing
 9         effectiveness of vacinations).
10     (2) Agricultural/Ecological Effects, mediated largely through altered biogeochemical cycling
11         resulting in consequent damaging impacts on terrestrial plants (leading to possible reduced
12         yields of rice, other food crops, and commercially important trees, as well as to biodiversity
13         shifts in natural terrestrial ecosystems); and deleterious effects on aquatic life (including
14         reduced ocean zooplankton and phytoplankton, as important base components of marine
15         food-chains supporting the existence of commercially important, edible fish and other
16         seafood, as well as to other aquatic ecosystem shifts).
17     (3) Indirect Human Health and Ecological Effects, mediated through increased tropospheric
1 g         ozone formation (and consequent exacerbation of surface-level, ozone-related health and
19         ecological impacts) and alterations in the concentrations of other important trace species,
20         most notably the hydroxyl radical and acidic aerosols.
21     (4) Other Types of Effects, such as faster rates of polymer weathering because of increased
22         UV-B radiation and other effects on man-made commercial materials and cultural artifacts,
23         secondary to climate change or exacerbation of air pollution problems.
24           Extensive qualitative and quantitative  characterizations of stratospheric ozone depletion
25     processes and projections of their likely potential impacts on human health and the environment
26     have been the subjects of periodic (1988,1989,1991,1994, 1998) international assessments
27     carried out under WMO and UNEP auspices since the 1987 signing of the Montreal Protocol on
28     Substances that Deplete the Ozone Layer. The  reader is referred for more detailed up-to-date
29     information to the two most recently completed international assessments of processes
30     contributing to stratospheric ozone depletion and the status of progress towards ameliorating the
31     problem (WMO, 1999) and revised qualitative and quantitative projections of likely consequent
        March 2001
4-132
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 human health and environmental effects (UNEP, 1998).  (See Appendices 4A and 4B for
 synopses of key points abstracted from the executive summaries of these assessments).
      Of considerable importance is the growing recognition, as reflected in these newer
 assessments, of impacts of enhanced solar radiation on biogeochemical cycles (see, for example,
 Zepp et al., 1998, and earlier discussions in this chapter [Sections 4.2.2.1 and 4.2.2.2]). As noted
 in the Zepp et al. paper, the effects of UV-B radiation (both in magnitude and direction) on trace
 gas (e.g., CO) emissions and mineral nutrient cycling are species specific and can affect a variety
 of processes. These include, for example, changes in the chemical composition of living plant
 tissue, photodegradation of dead plant matter (e.g., ground litter), release of CO from vegetation
 previously charred by fire, changes in microbial decomposer communities, and effects on
 nitrogen-fixing microorganisms and plants. Also, studies of natural acquatic ecosystems indicate
 that organic matter is the primary determinant of UV-B penetration through water. Organic
 matter changes, caused by enhanced UV-B penetration and augmented by acidification and
 climate change, contribute to clarification of water and changes in light quality that broadly
 impact the effects of UV-B on aquatic biogeochemical cycles.  Enhanced UV-B levels have both
 positive and negative impacts on aquatic ecosystem microbial activities that can affect nutrient
 cycling and the uptake or release of greenhouse gases. Thus, there are emerging complex issues
 regarding interactions and feedbacks between climate change and changes in terrestrial and
 marine biogeochemical cycles because of increased UV-B penetration to the Earth's surface.
      As noted in the above detailed assessments, since the signing of the Montreal Protocol,
 much progress has been made in reducing emissions of ozone depleting gases, leading to
 estimates of the maximum extent of stratospheric ozone depletion as likely having been reached
 in the year 2000, to be followed by gradual lessening of the problem and its impacts during the
 next half-century.  However, the assessments also note that the modeled projections are subject
 to considerable uncertainty.  The role of atmospheric particles, discussed below, is one of
 numerous salient factors complicating modeling efforts.

4.5.1.2 Airborne Particle Impacts on Atmospheric Ultraviolet Radiation Transmission
      A given amount of ozone  in the lower troposphere has been shown to absorb more solar
radiation than an equal amount of ozone in the stratosphere because of the increase in its
effective optical path produced by Rayleigh scattering in the lower atmosphere (Bruehl and
       March 2001
                                         4-133
DRAFT-DO NOT QUOTE OR CITE

-------
 1     Crutzen, 1988). The effects of particles are more complex. The impact of particles on the SUVB
 2     flux throughout the boundary layer are highly sensitive to the altitude of the particles and to their
 3     single scattering albedo. Even the sign of the effect can reverse as the composition of the particle
 4     mix changes from scattering to absorbing types (e.g., from sulfate to elemental carbon or PAHs)
 5     (Dickerson et al.,  1997). In addition, scattering by particles also may increase the effective
 6     optical path of absorbing molecules, such as ozone, in the lower atmosphere.
 7           The effects of particles present in the lower troposphere on the transmission of SUVB have
 8     been examined both by field measurements and by radiative transfer model calculations. The
 9     presence of particles in urban areas modifies the spectral distribution of solar irradiance at the
10     surface.  Shorter wavelength radiation (i.e., in the ultraviolet) is attenuated more than visible
11     radiation (e.g., Peterson et al., 1978; Jacobson, 1999).  Wenny et al. (1998) also found greater
12     attenuation of SUVB than SUVA (315 to 400 nm). However, this effect depends on the nature
13     of the specific particles involved and, therefore, is expected to depend strongly on location.
14     Lorente et al. (1994) observed an attenuation of SUVB ranging from 14 to 37%, for solar zenith
15     angles ranging from about 30° to about 60°, in the total (direct and diffuse) SUVB reaching the
16     surface hi Barcelona during cloudless conditions on very polluted days (aerosol scattering optical
17     depth at 500 nm, 0.46 s TSOO „„, 5 1.15) compared to days on which the turbidity of urban air was
18     similar to that for rural air (i;soo nm £ 0.23). Particle concentrations that can account for these
19     observations can be estimated roughly by combining Koschmeider's relation for expressing
20     visual range in terms of extinction coefficient with one for expressing the mass of PM2 5 particles
21     in terms of visual range (Stevens et al., 1984). By assuming a scale height (i.e., the height at
22     which the concentration of a substance falls off to 1/e of its value at the surface) of 1 km for
23     PM2 5, an upper limit of 30 fig/ m3 can be derived for the clear case and between 60 and
24      150 Aig/m3 for the polluted case. Estupinan et al. (1996) found that summertime haze under clear
25      sky conditions attenuates SUVB between  5 and 23% for a solar zenith angle of 34 °, compared to
26      a clear sky day in autumn. Mims (1996) measured a decrease in SUVB by about 80% downwind
27      of major biomass burning areas in Amazonia in 1995. This decrease in transmission
28      corresponded to optical depths at 340 nm  ranging from three to four. Justus and Murphey (1994)
29      found that SUVB reaching the surface decreased by about 10% because of changes in aerosol
30      loading in Atlanta, GA, from 1980  to 1984. Also, higher particle levels in Germany (48 °N) may
        March 2001
4-134
DRAFT-DO NOT QUOTE OR CITE

-------
  1      be responsible for greater attenuation of SUVB than in New Zealand (Seckmeyer and McKenzie,
  2      1992).
  3           In a study of the effects of nonurban haze on SUVB transmission, Wenny et al. (1998)
  4      derived a very simple regression relation between the measured aerosol optical depth at 312 nm,
  5
  6                ln( SUVB transmission at solar noon) = -0.1422 T312 „„ - 0.138, R2 = 0.90,
  7
  8      and the transmission of SUVB to the surface. In principle, values of T3I2 nm could be found from
  9      knowledge of the aerosol optical properties and visual range values.  Wenny et al. (1998) also
10      found that absorption by particles accounted for 7 to 25% of the total (scattering + absorption)
11      extinction. Relations such as the above one are strongly dependent on local conditions and
12      should not be used in other areas without knowledge of the differences in aerosol properties.
13      Although all of the above studies reinforce the idea that particles play a maj or role in modulating
14      the attenuation of SUVB, none included measurements of ambient PM concentrations, so direct
15      relations between PM levels and SUVB transmission could not be determined.
16           Liu et al. (1991) estimated, roughly, overall effects of increases of anthropogenic airborne
17      particles that have occurred since the beginning of the industrial revolution on atmospheric
18      transmission of SUVB.  Based on (a) estimates of the reduction in visibility from about 95 km to
19      about 20 km over nonurban areas in the eastern United States and in Europe, (b) calculations of
20      optical properties of airborne particles found in rural areas to extrapolate the increase in
21      extinction at 550 to 310 nm, and (c) radiative transfer model calculations, Liu et al. concluded
22      that the amount of SUVB reaching Earth's the surface likely has decreased from 5 to 18% since
23      the beginning of the industrial revolution. This was attributed mainly to scattering of SUVB
24      back to space by sulfate containing particles. Radiative transfer model calculations have not
25      been done for urban particles.
26           Although aerosols are expected to decrease the flux of SUVB reaching the surface,
27      scattering by particles is expected to result in an increase in the actinic flux within and above the
28      aerosol layer. However, when the particles  significantly absorb SUVB, a decrease in the actinic
29      flux is expected. Actinic flux is the radiant  energy integrated over all directions at a given
30      wavelength incident on a point in the atmosphere, and is the quantity needed to calculate rates of
31      photolytic reactions in the atmosphere.  Blackburn et al. (1992) measured attenuation of the
        March 2001
4-135
DRAFT-DO NOT QUOTE OR CITE

-------
 1     photolysis rate of ozone and found that aerosol optical depths near unity at 500 nm reduced
 2     ozone photolysis rate by as much as a factor of two.  Dickerson et al. (1997) showed that the
 3     photolysis rate for NO2 , a key parameter for calculating the overall intensity of photochemical
 4     activity, could be increased within and above a scattering aerosol layer extending from the
 5     surface, although it would be decreased at the surface. This effect is qualitatively similar to what
 6     is seen in clouds, where photolysis rates are increased in the upper layers of a cloud and above
 7     the cloud (Madronich, 1987). For a simulation of an ozone episode that occurred during July
 8      1995 in the Mid-Atlantic region, Dickerson et al. (1997) calculated ozone increases of up to
 9     20 ppb  compared to cases that did not include the radiative effects of particles in urban airshed
10     model (UAM-IV) simulations. In contrast, Jacobson (1998) found that particles may have
11     caused a 5 to 8% decrease in O3 levels during the Southern California Air Quality Study in 1987.
12     Absorption by organic compounds and nitrated inorganic compounds was hypothesized to
13     account for the reductions in UV radiation intensity.
14           The photolysis of ozone in the Hartley bands also leads to production of electronically
15     excited oxygen atoms, O('D) that then react with water vapor to form OH radicals. Thus,
16     enhanced photochemical production of ozone is accompanied by the scavenging of species
17     involved  in greenhouse warming and stratospheric depletion. However, these effects may be
18     neutralized or even reversed by the presence of absorbing material in the particles.  Any
19     evaluation of the effects of particles on photochemical activity therefore will depend on the
20     composition of the particles and also will be location-specific.
21           Also complicating any straightforward evaluation of UV-B penetration to specific areas of
22     the Earth's surface are the influences of clouds, as discussed by Erlick et al. (1998), Frederick
23      et al. (1998), and Soulen and Fredrick (1999). Varying estimations of atmospheric transmission
24      of UV and visible spectrum light are obtained for cloudy atmospheres, depending on presence of
25      aerosols and the extent of their external or internal mixing with cloud droplets. Even in
26      situations of very low atmospheric PM (e.g., over Antarctica), interannual variations in
27      cloudiness over specific areas can be as important as ozone levels in determining UV surface
28      irradiation, with net impacts varying from a month or season to another (Soulen and Fredrick,
29      1999).
30           Given the above considerations, quantitation of projected effects of variations in
31     atmospheric PM on human health or the environment because of particle impacts on transmission
        March 2001
4-136
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
26
27
28
29
30
31
 of solar UV-B would require location-specific evaluations, taking into account composition,
 concentration, and internal structure of the particles; temporal variations in atmospheric mixing
 heights and depths of layers containing the particles; and consequent impacts on surface-level
 exposures of humans, ecosystem constituents, or man-made materials.  The outcome of such
 modeling effects would likely vary from location to location in terms of increased or decreased
 surface level UV-B exposures because of location-specific changes in atmospheric PM
 concentrations or composition.  For example, to the extent that any location-specific scattering by
 airborne PM were to affect the directional characteristics of UV radiation at ground level, and
 thereby enhance radiation incident from low angles (Dickerson, 1997), the biological
 effectiveness of resulting ground-level UV-B exposures could be enhanced. Airborne PM also
 can reduce the ground-level ratio of photorepairing radiation (UV-A and short-wavelength
 visible) to damaging UV-B radiation. Lastly, PM deposition is a major source of PAH in certain
 freshwater lakes and coastal areas, and the adverse effects of solar UV are enhanced by uptake of
 PAH by aquatic organisms.  Thus, although airborne PM may, in general, tend to reduce ground-
 level UV-B, its net effect in some locations may be to increase UV damage to certain aquatic and
 terrestrial organisms, as discussed by Cullen and Neale (1997).

 4.5.2  Global Warming Processes, Human Health and Environmental
       Impacts, and Atmospheric Particle Roles
 4.5.2.1 Bases for Concern Regarding Global Warming and Climate Change
     Various trace gases emitted because of man's activities, including several noted above as
 contributing to stratospheric ozone depletion, can act as "greenhouse gases" (GHG). That is, as
 their tropospheric concentrations increase, they retard the escape of infrared radiation from the
 earth's surface and thereby contribute to the trapping of heat near the surface (the "greenhouse
 effect") and, ultimately, to consequent global warming and climate change. Much concern has
 evolved,with regard to increases in the naturally very low concentrations in the atmosphere of
 some of these gases, especially carbon dioxide (CO2), nitrous oxide (N2O), methane (CH4),
 chloroflurocarbons (CFCs), and tropospheric ozone (O3).
     Atmospheric processes involved in mediating global warming and its  likely consequent
effects have been reviewed extensively previously (United Nations Environment Programme,
 1986; World Meteorological Organization, 1988; U.S. Environmental Protection Agency, 1987;
       March 2001
                                        4-137
DRAFT-DO NOT QUOTE OR CITE

-------
 1      IPCC, 1996, 1998; US GCRP, 2000) and more concisely summarized by others (e.g., Grant
 2      1989; Patz et al., 2000a; Patz et al., 2000b).  The main focus here is (a) to provide first a very
 3      brief summary of key points regarding processes involved and types of effects projected as likely
 4      to be associated with global warming and climate change and, then, (b) to undertake discussion
 5      of salient considerations regarding potential impacts of atmospheric PM on such processes and
 6      effects.
 7           All of the above noted assessments and summaries emphasize that estimating likely future
 8      global warming trends and associated climate change caused by greenhouse gases is extremely
 9      complex, with modeling results being highly dependent on key assumptions about the rates of
10      future increases in various gases and numerous other factors (including particle effects).
11      Modeling of the magnitude of the warming directly associated with radiative forcing by
12     greenhouse gases  (without feedback enhancement) projects temperature increases, for example,
13     of about 1.2 °C for a doubling of CO2; another 0.45 °C for a simultaneous doubling of N2O and
14     CH4; and an additional 0.15 °C from a uniform 1-ppb increase in atmospheric concentrations of
15     CFC-11 and CFC-12. Indirect effects  (feedbacks) that likely would increase temperatures further
16     are expected to occur. Increased water vapor (trapping heat) and snow and ice melting (reducing
17     reflection of radiation back into space) are two examples of such feedback factors expected to
18     increase temperatures. However, major uncertainties exist with regard to feedbacks between
19     global wanning and clouds, which could either amplify or, perhaps, reduce a temperature rise.
20     Taking assumptions about rates of increase  (or decrease) in GHG concentrations, consequent
21     initial warming effects, feedback effects, and accompanying uncertainties into account, numerous
22     modeling efforts have attempted to project likely future trends in global warming. Despite the
23     complexity and uncertainties inherent in such modeling efforts, all typically agree that some
24     global warming has occurred and will continue to occur during the coming decades, but the
25     ranges of quantitative estimates vary considerably depending on specific assumptions
26     incorporated into the models. Thus,  for example, "low" scenarios assuming stabilization or
27     reductions in GHG emissions (resulting from implementation of the  1987 Montreal Protocol)
28     project lower temperature changes than other scenarios assuming higher rates of increase in GHG
29      emissions or differing feedback-effect patterns.
 3 o           Given the wide range of estimates of global warming trends and patterns of associated
 31      climate change emerging from modeling efforts, the estimation of likely human health and
        March 2001
4-138
DRAFT-DO NOT QUOTE OR CITE

-------
  1      ecological effects associated with global warming on any quantitative basis is extremely difficult.
  2      The onset of any notable global warming effect is also important, with various analyses
  3      indicating that global temperatures for the past century have been rising (and now appear to be
  4      beyond average levels within the range of variation seen with cycles of global warming or
  5      cooling over the past several centuries before marked anthropogenic emissions of greenhouse
  6      gases occurred). Also posing difficulties for the quantitative estimation of human health and
  7      other effects are expected wide regional variations in temperature and climate characteristics
  8      (e.g., rain and snowfall amounts) that may be projected reasonably to result from various global
  9      warming trend scenarios. Lastly, it should be noted that, despite general warming trends in
 10      long-term average temperatures, wide extremes in both high and low temperatures also are
 11      expected to occur more frequently in some areas.
 12           A Special Report of the IPCC Working Group II on Regional Impacts of Climate Change:
 13      An Assessment of Vulnerabilities (IPCC, 1998) assesses global wanning processes and identifies
 14      several types of vulnerabilities likely to occur because of climate change resulting from global
 15      warming. Such general types of vulnerabilities include impacts on terrestrial and aquatic
 16      ecosystems, hydrology and water resources, food and fiber production, coastal systems, and
 17      human health. Appendix 4C provides excerpts of materials from the executive summary of the
 18      IPCC (1998) report that comprise a helpful overview of key points regarding projected global
 19      warming processes, likely climate change patterns, and their consequent impacts in terms of the
20      types of vulnerabilities noted above.
21           The IPCC (1998) report notes that human activities resulting in emissions of long-lived
22      GHCs are projected by General Circulation Models (GCMs) to lead to global and regional
23      changes in temperature, precipitation and other climate variables—resulting in increases in
24      global mean sea level; prospects for more extreme weather events, floods, and droughts in some
25      areas; and consequent changes in soil moisture.  Based on various scenarios of current and
26      plausible future emissions of GHGs and aerosols and the range of sensitivities of climate change
27      to atmospheric levels (and residence time) of GHGs, GCMs project mean annual global surface
28      temperature increases in the range of 1 to 3.5 °C by 2100, a global mean sea level rise of 15 to
29      95 cm, and significant changes in spatial and temporal patterns of precipitation.  The average rate
30      of warming will be more rapid than any seen hi the past 10,000 years, although regional changes
31      could differ substantially from mean global rates.
       March 2001
4-139
DRAFT-DO NOT QUOTE OR CITE

-------
 1          Human health, ecosystems, and socioeconomic sectors (e.g., hydrology and water
 2     resources, food and fiber production, etc.) are projected to be vulnerable to the magnitude and
 3     rate of climate change, as well as increased climate variability. Wide variations in the courses
 4     and net impacts of climate change in different geographic areas can be expected, and, although
 5     many regions are likely to experience severe adverse impacts (some possibly irreversible) of
 6     climate change, some climate change impacts may be locally beneficial in some regions.
 7     In general, projected climate change impacts can be expected to represent additional stresses on
 8     those natural ecosystems and human societal systems already impacted by increasing resource
 9     demands, unsustainable resource management practices, and pollution, with wide variation likely
10     across regions and nations in their ability to cope with consequent alterations in ecological
11     balances, in availability of adequate food, water, and clean air, and in human health and safety.
12     Appendix 4C also includes excerpts from the executive summary of the IPCC 1998 special report
13     regarding the assessment of different types of vulnerabilities to climate change projected for each
14     of 10 different geographic regions of the Earth, with emphasis being placed in Appendix 4C on
15     those projected for two regions (North America and Polar) of most relevance to the continental
16     United States and Alaska.
17           Appendix 4C notes that (a) the characteristics of subregions and sectors of North America
18     suggest that neither impacts of climate change nor response options will be uniform, and (b)
19     many systems of North America are moderately to highly sensitive  to climate change, with the
20     range of estimated effects including the potential for substantial damage or, conversely, the
21     potential for some beneficial outcomes.  The most vulnerable continental United States sectors
22     and regions include long-lived natural forest ecosystems in the East and ulterior West, water
23     resources in the southern plains, agriculture in the Southeast and southern plains, northern
24     ecosystems and habitats, estuaries and beaches in developed areas,  and low-latitude cool and cold
25     water fisheries. Other sectors or subregions may benefit from warmer temperatures or increased
26      CO2 fertilization (e.g., west coast coniferous forests; some western rangelands; reduced energy
27      costs for heating in northern latitudes; reduced road salting and snow-clearance costs; longer
28      open-water seasons in norther channels and ports; and agriculture in the northern latitudes, the
29      interior West, and the west coast). For Alaska, substantial shifts in ecosystems (with possible
30      major declines or loss of some sensitive species like bear and caribou or of other ice-dependent
31      animals) may occur in parallel to beneficial effects  such as opening of ice-bound water
        March 2001
4-140
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
transportation routes or possible expanded agricultural viability secondary to longer growing
seasons.  On the other hand, for North America, the potential for mainly deleterious direct or
indirect effects on human health is likely to increase (e.g., increased mortality directly linked to
temperature extremes, increases in incidence and spread of vector-borne infectious diseases,
impacts secondary to sea-level rise, and impacts secondary to increased tropospheric air pollution
[as depicted in Figure 4-26]).
     More detailed evaluations of possible global climate change impacts on various U.S.
geographic areas are being conducted by the United States Global Change Research Program
(USGCRP). An overview report on the assessment results and key findings from a series of
workshops convened by the USGCRP National Assessment Synthesis team (NAST) has been
prepared  (USGCRP, 2000).  Selected highly salient key points from the report and subsidiary
regional assessments are presented in Appendix 4D. Overall key findings from the USGCRP
(2000) report are noted below.
 (1)  Increased Warming. Assuming continued growth in world GHG emissions, the primary
     climate models used in the USGCRP assessment project that temperatures in the United
     States will rise by 5 to 10 °F (3 to 6 °C) on average during the next 100 years.
 (2)  Differing Regional Impacts.  Climate change will vary widely across the United States.
     Temperature increases will vary somewhat from region to region. Heavy and extreme
     precipitation events are likely to become more frequent, yet some regions will get drier.
     The potential impacts of climate change will vary widely across the nation.
 (3)  Vulnerable Ecosystems. Many ecosystems are highly vulnerable to the projected rate and
     magnitude of climate change. A few,  such as  alpine meadows in the Rocky Mountains and
     some barrier islands, are likely to disappear entirely in some areas, with others, such as
     some forests of the Southeast, being likely to experience major species shifts or break up.
     Goods and services lost through disappearance or fragmentation of certain ecosystems are
     likely to be costly or impossible to replace.
 (4)  Widespread Water Concerns. Water is an issue in every region, but the nature of the
     vulnerabilities varies, with different nuances in each. Drought is an important concern in
     every region.  Floods and water quality are concerns in many regions. Snowpack changes
     are especially important in the West, the Pacific Northwest, and Alaska.
       March 2001
                                        4-141
DRAFT-DO NOT QUOTE OR CITE

-------
                    BASES FOR CONCERN ABOUT GLOBAL WARMING AND CLIMATE
                    CHANGE EFFECTS ON THE ENVIRONMENT AND HUMAN HEALTH
                                       GLOBAL WARMING
                                     AND CLIMATE CHANGE
                               Geographic Variations in
                               Temperature Increases
                              (Avg. & Extremes), Rainfall,
      I Sea-Level Rise I
           / Increased Frequency \
                   of
           \ Air Stagnation Periods /
1
f >
Elderly (> 65 yrs) Most at
Risk — Also Infants
Key: Acclimatization



Geographic Distribution &
Abundance of Vectors/Hosts
Depend on Temperature,
Moisture, Habitat effects
/ Y / \
HEAT-STRESS
MORTALITY
Lower
threshold
temperatures
in North than
in South
If only partial
acclimatization
then heat
deaths rise
COLD-INDUCED
MORTALITY
Higher
threshold
temperatures
in South
(>0°C) than
North (<0'C)
Prob. drop in
cold-related
mortality








TICK-BORNE
DISEASES
USA
EXAMPLES:

Lyme Disease
Rocky
Mountain
Spotted
Fever

MOSQUITO-
BORNE
DISEASES
USA
EXAMPLES:

Malaria
Dengue Fever
Arbovirus-
Related
Encephalitis


Near-term: Storm Surges,
Costal Flooding
Long Term: Inland Advance
of Saltwater Oceans/Seas
/
HUMAN
HEALTH
IMPACTS
Loss of Life

Nutrition
Vector-Borne
Diseases
Other
Communicable
Diseases
\
OTHER
TYPES OF
IMPACTS
Damage to:

Industries
Agriculture
Aquatic and
Land
Ecosystems

i

Increased Tropospheric
Air Pollution
(PM, O3, CO, etc)
/
HUMAN
HEALTH
IMPACTS
Acute Pulmon.
Function
Decrements
Impaired Lung
Defenses
Increased
Respiratory
Disease
Susceptibility
\
OTHER
TYPES OF
IMPACTS
Forest and
Agriculture
Damage
Ecosystem
Effects
Materials
Damage

      Figure 4-26.  Bases for concern about global warming and climate change effects on the
                   environment and human health. Types of hypothesized likely human health
                   effects include (1) increases in mortality directly linked to temperature
                   extremes, (2) increases in incidence and spread of vector-borne infectious
                   diseases, (3) impacts secondary to projected sea-level rise, and (4) impacts
                   secondary to increased tropospheric air pollution. Additional impacts can be
                   expected because of shifting agricultural sustainability in various U.S. regions
                   consequent to extreme weather patterns leading to inland flooding or
                   droughts.

      Source: Adapted from Grant (1989).
1      (5) Secure Food Supply. At the national level, the U.S. agriculture sector is likely to be able to
2          adapt to climate change. Overall, U.S. crop productivity is very likely to increase over the
3          next few decades, but the gains will not be uniform across the nation. Falling prices and
4          competitive pressures are very likely to stress some farmers, while benefiting consumers.
      March 2001
4-142
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
  (6) Near-Term Increases in Forest Growth. Forest productivity is likely to increase over the
     next several decades in some areas as trees respond to higher CO2 levels. Over the longer
     term, changes in larger scale processes such as fire, insects, droughts, and disease will
     possibly decrease forest productivity. Also, climate change is likely to cause long-term
     shifts in forest species (e.g., distribution of sugar maple stands more northward, out of the
     United States).
  (7) Increased Damage in Coastal and Permafrost Areas.  Climate change and the resulting rise
     in sea level are likely to exacerbate threats to building, roads, powerlines, and other
     infrastructure in climatically sensitive places, such as low-lying coastlines and the
     permafrost regions of Alaska.
  (8) Other Stresses Magnified by Climate Change. Climate change will very likely magnify the
     cumulative impacts of other stresses, such as air and water pollution and habitat destruction
     caused by human development patterns.  For some systems, such as coral reefs, the
     combined effects of climate change and other stresses are very likely to exceed a critical
     threshold, bringing large, possibly irreversible impacts.
  (9) Surprises Expected. It is likely that some aspects and impacts of climate change will be
     totally unanticipated as complex systems respond to ongoing climate change in
     unforeseeable ways.
(10) Uncertainties Remain.  Significant uncertainties remain in the science underlying regional
     climate changes and their impacts. Further research is needed to improve understanding
     and predictive ability about societal and ecosystem impacts and to provide the public with
     additional useful information about adaptation strategies.
     The selected findings highlighted in Appendix 4D from the USGCRP (2000) report and
subsidiary regional reports illustrate well the considerable uncertainties and difficulties in
projecting likely climate change impacts on regional or local scales. The findings presented in
Appendix 4D also reflect well the mixed nature of projected potential climate change impacts
(combinations of mostly deleterious, but other possible beneficial effects) for U.S. regions and
their variation across the different regions. Difficulties in assessing regional-specific potential
impacts also can be illustrated by discussion below of determinants of the potential extent of
direct or indirect impacts of global warming on human health, as abstracted from various
published assessments cited above or alluded to in Appendices 4C, 4D, and 4E.
        March 2001
                                         4-143
DRAFT-DO NOT QUOTE OR CITE

-------
 1          Modeling efforts and published analyses by Kalkstein and others have helped to identify
 2     important factors that affect the magnitude of temperature-dependent mortality and provide
 3     bases for projecting future temperature-related mortality trends (see Appendix 4E for recently
 4     published projections for U.S. cities by Kalkstein and Greene, 1997). Examples of key
 5     determinants of temperature-related mortality include (1) weather-sensitive mortality occurs
 6     mainly as a function of extremes of temperature beyond certain threshold points (for increasing
 7     or decreasing temperatures) that are characteristic of any particular city; (2) the  extent of the
 8     mortality is generally more dependent on the duration of the periods (days) during which
 9     threshold points are exceeded than on maximum temperatures and also varies as a function of
10     combined relative humidity, temperature, and barometric pressure conditions that constitute
11     "oppressive" weather events that vary for different locales; and (3) the major population segment
12     typically most severely affected are the elderly (2:65 years old).
13          Threshold temperature findings for summer and winter in U.S. cities suggest that weather
14     effects on mortality are relative (i.e., they vary in relation to the typical conditions to which local
15     residents have become acclimatized). Thus, the highest summer threshold temperatures for
16     mortality are found for the South and Southeast and the lowest hi the Pacific and Northeast U.S.
17     regions. Conversely, lowest threshold temperatures for winter mortality are found for cities in
18     the coldest regions, whereas notably higher thresholds for cold-associated deaths occur for
19     warmer region cities, with threshold values for some being well above the freezing point. Also,
20     the total accumulated times of occurrence in the season of particular oppressive weather events
21     are important determinants of mortality levels (e.g., hot conditions early in the spring and
22     summer have a larger impact than similar conditions later in the summer, and length of a
23     heat-stress period also has a larger impact than maximum temperatures reached).
24          Acclimatization is a key determinant of weather-related mortality, and the greatest initial
25     increases in heat-related mortality might be expected in cities where temperatures are normally
26     cooler or in areas where global-warming-induced climate changes lead to increased frequency
27     and durations of high-temperature episodes.  Eventual acclimatization may occur over the years,
28     however, when higher-than-usual temperatures become the new norm for cities in currently
29     cooler, more northern regions. As for cold-associated deaths, if average winter temperatures
30     were not to drop as low as usual in various regions, then whiter mortality might generally
31     decrease because of fewer days falling below existing winter threshold levels for many cities.
        March 2001
4-144
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 But, if acclimatization occurs to higher average winter temperatures and wider variations in
 temperature extremes occur in some areas because of global-warming-induced weather changes,
 then those periods of lower maximum temperatures (especially of several days duration) could
 cause even higher than past mortality rates previously observed with comparable winter
 conditions. More sophisticated modeling also is needed to take into account combined effects of
 temperature extremes and weather-related increases in air pollutants as possible mortality
 determinants; for example, increased mortality or morbidity effects because of temperature
 extremes may be exacerbated or added to by higher surface level atmospheric PM derived from
 increased coal or oil combustion to generate more heat (in winter) or electricity (in summer for
 air conditioning) during extreme temperature periods.
     In addition to concern about possible mortality increases because of temperature extremes,
 global warming, and consequent climate change also may impact human health through increases
 in some infectious diseases. For many parts of the world, infectious diseases remain among the
 leading causes of death, as occurred earlier in industrialized or "developed" countries (where
 diseases such as influenza, pneumonia, and tuberculosis were among the leading causes of death
 in 1900).  Since then, the incidence and associated mortality for these and other infectious
 diseases such as diphtheria,  typhus, and polio have been reduced dramatically in industrialized
 countries, hi developed countries, it is not clear to what extent global-warming-induced climate
 change may cause general increases in the incidence of such diseases, unless serious disruptions
 of social structures occur or, in some coastal areas, breakdowns in sanitation systems happen as a
 consequence of sea-level rise. The spread of infectious diseases is likely of greater concern for
 many less developed countries, where inadequate medical care systems, immunization programs,
 housing conditions, and nutrition make them more vulnerable to the spreading of such diseases.
     Of particular shared concern for both developed and less developed countries with regard to
potential global warming impacts are infectious diseases spread by climate-dependent vectors.
Vector-borne diseases are those for which the infectious microbial agent is transmitted to humans
via another agent (the vector), such as the flea, tick, or mosquito. Well known examples of
vector-borne diseases are malaria (transmitted to humans via mosquitos) and bubonic plague
(transmitted via fleas or, at times, via animals directly to man as a respiratory disease).  Climate
change can affect vector-borne diseases by various direct impacts on the infectious agent, the
vector, or intermediate hosts through variations in temperature, humidity, rainfall, or storm
        March 2001
                                         4-145
DRAFT-DO NOT QUOTE OR CITE

-------
 1      patterns that alter (1) the multiplication rates of the infectious agent or the vector, (2) the biting
 2      rate of the vector, (3) the geographic distribution of the intermediate animal hosts, or (4) the
 3      amount of time that intermediate hosts or human hosts are exposed to the vector.  Climate change
 4      also can affect indirectly the rates or incidences of vector-borne diseases via impacts on
 5      agricultural practices, ecosystem mixes (of grasses, trees, underbrush, etc.), surface water levels,
 6      or other factors that determine intermediate host or vector distribution or survival. The variety of
 7      vector-borne diseases is considerable, with some being of more concern than others for particular
 8      countries, depending on specific climatic conditions and existing pools of infected hosts (both
 9      human and intermediate animal hosts). Examples of vector-borne diseases illustrative of
10     concerns that apply to the United States for potential spread of vector-borne diseases include
11      Lyme disease, Rocky Mountain spotted fever, dengue fever, malaria, and viral encephalitis.
12           Lyme disease (initially recognized in Lyme, CT) is an inflammatory disease caused by a
13     spirochete, Borrelia burgdorferi, transmitted by several subspecies ofLcodes racinus ticks.
14     Numerous species of birds and mammals can be hosts for various subspecies of the tick vector,
15     with  varying geographic distributions.  Lyme disease has four major U.S. foci, is spreading
16     rapidly, and has been found in Europe (Germany, Switzerland, France, and Austria). The U.S.
17     distribution of human cases of the disease tends to match areas where the tick vector is abundant,
18     and deer populations, along with factors such as temperature, humidity and local vegetation,
19     represent key determinants of tick abundance.  The precise impact of global warming and climate
20      change on the distribution of Lyme disease is difficult to estimate. Lengthening of warm weather
21     periods and shortening of winter weather could enhance tick vector abundance and its potential
22      spread into adjoining areas if the weather changes (temperature, precipitation, etc.) were to favor
23      wider distribution of deer or other animal or bird hosts. Shifts of human populations into or out
24      of affected areas in response to changes in local climate also would help determine location-
25      specific alterations in Lyme disease rates.
26           Rocky Mountain spotted fever (initially identified in western mountain areas but actually
27      much more prevalent in southeastern U.S. states) is a highly fatal disease if not promptly
28      diagnosed and treated.  Caused by the occobacillus, Rickettsiae rickettsii, the disease is spread by
29      ticks and is also known as tick fever, with analogous diseases occurring in many other countries.
30      The  main North America vectors are the dog tick, D. variabilis, and the wood ticks, D. andersoni
31      and D. occidentalis, with varying geographic distributions. Geographic tick distributions parallel
        March 2001
4-146
DRAFT-DO NOT QUOTE OR CITE

-------
  1      closely the typical U.S. distribution of disease cases (highest incidence across the South).
  2      Crucial for the spread of Rocky Mountain spotted fever is the wide variety of intermediate hosts
  3      available to the ticks (i.e., many woodland mammals and birds) and temperature. Certain
  4      optimum ranges of high temperatures (24 to 30 °C) likely speed the rickettsial growth in the
  5      ticks, and ambient temperatures are important in determining tick breeding season length and
  6      cycles, as well as their activity levels and biting rates. Each are enhanced by higher temperatures,
  7      and the abundance of the vector is held in check, in part, by frequency and length of time that
  8      winter temperatures drop well below freezing, thus killing overwintering adults.  Lastly, relative
  9      humidity conditions and rainfall are important as well, hi that hot dry weather results in
10      desiccation of ticks and their eggs, reducing reproduction rates. Global warming-induced climate
11      change might increase the range of Rocky Mountain spotted fever tick vectors into more
12      northward U.S. areas and, possibly, into Canada, assuming the climate change includes  sufficient
13      rainfall to sustain adequate habitats for host species and adequate moisture for survival of ticks
14      and eggs. Hot, dry periods caused by any prolonged drought conditions in the United States or
15      Canada predicted by some global warming scenarios, conversely, would not be conducive to
16      increased incidence of the disease in drought-affected areas.
17          Malaria, once widespread in the southern United States, remains endemic in many areas of
18      the world and is caused by four agents: (1) Plasmodium vivax, (2) P. malariae, (3) P. ovale, and
19      (4) P. falciparum. The agents cause clinical syndromes of varying severity, the most serious
20      being caused by P. falciparum, which can progress to death (>10% fatality in untreated children
21      and nonimmune adults). The other forms, although less severe, are still debilitating and are
22      typified by recurring episodes of fever, chills, and sweating. Malarial agents are transmitted from
23      infected humans, as the main host pool, by the bite of various subspecies of anopheles mosquitos.
24      Ambient temperatures of at least 15 to 18 °C are crucial for development of the malarial agents
25      within the mosquitos, and ambient temperature levels determine breeding season length and
26      survival rates (higher tropical temperatures being most favorable). Man's agricultural activities,
27      in providing irrigation ditches and more stagnant water habitats, has contributed to spread and
28      abundance of the anopheles mosquito in many areas of the world. Malaria is now rarely
29      endogenously transmitted in the United States, the pool of infected humans as hosts having been
30      reduced very substantially, owing to mosquito  eradication programs.  Prior to such programs, the
31      disease was endemic in widespread southern U.S. areas up to the 1940s, but, since then,
        March 2001
4-147
DRAFT-DO NOT QUOTE OR CITE

-------
 1     outbreaks mainly have occurred because of infected immigrants entering the country or U.S.
 2     military veterans returning from overseas endemic areas. Global warming leading to higher
 3     temperatures in more northerly U.S. areas and Europe could enhance conditions for the spread of
 4     the disease. Both the range and abundance of competent vectors (various anopheles subspecies)
 5     likely would be increased, especially if increased irrigation were required to support agriculture,
 6     owing to higher temperatures. Also, higher temperatures in more northerly areas could extend
 7     the range of adequate temperatures (>15 to 18 °C) needed for development of malarial agents in
 8     the mosquitos.  The remaining key factor in determining the likelihood of the spread of malaria,
 9     however, is the infected host pool, with numbers of infected human hosts moving into or out of
10     areas of enhanced vulnerability being of crucial importance, as emphasized by Longstreth (1999).
11          Dengue fever, another mosquito-borne disease, is caused by four serotypes of a Group B
12     arbovirus.  Fever, general muscle ache, severe headache, and retroorbital pain typify dengue fever
13     (usually not fatal); but it can progress to dengue haemorrhagic fever or dengue shock syndrome
14     (often fatal). Once endemic along the U.S. Gulf and South Atlantic coasts, dengue fever is now
15     rarely endogenously transmitted in the United States. The Aedis aegypti mosquito is the primary
16     vector, with wide southern U.S. distribution.  The breeding season of the A. aegypti mosquito is
17     temperature-dependent, with breeding year-round in southern Florida, nearly year-round
18     elsewhere in Florida and along the Gulf Coast, and much shorter for successively more
19     northward bands of geographic distribution.  Another potential vector, Aedes triseriatus, is
20     endogenous to states east of the Mississippi, and Aedes albopictus, a proven dengue vector
21     introduced from northern Asia, has been found hi scattered U.S. sites. Higher temperatures are
22     also crucial for dengue transmission; transmission of dengue occurred experimentally only if
23     A. aegypti mosquitos were kept at 30 °C, and the incubation period for the virus to develop in the
24     mosquitos was shortened at 32 to 35 °C. Consistent with this, cases  of dengue haemorrhagic
25     fever increased at non-U.S. sites when daily mean temperatures were 28 to 30 °C during hot
26     seasons, but decreased at the sites during cooler seasons with 25 to 28 °C temperatures.
27     Temperature increases in temperate ozone areas with A. Aegypti or A. albopictus present would
28     tend to expand the range of these dengue fever vectors, including potential spread especially of
29     A. albopictus farther north in the United States and, perhaps, into Canada, in view of its
30     adaptation to cold weather as well. Whether or not increases in dengue will actually occur,
31     however, likely will depend on the distribution of rainfall and moisture content, the effects of
        March 2001
4-148
DRAFT-DO NOT QUOTE OR CITE

-------
  1     agricultural practices (e.g., increased irrigation), and movements of infected human hosts into or
  2     out of areas with increased vector density.
  3          Arbovirus-induced encephalitis syndromes vary in severity but include several that can be
  4     highly fatal and are related to several other types of arbovirus-related syndromes (e.g., yellow
  5     fever, dengue and other haemorrhagic fevers, hepatitis, arthritis, rashes, various tropical fevers).
  6     Different types of mosquitos that serve as competent vectors for various types of
  7     arbovirus-induced encephalitis of concern for the United States display different patterns of
  8     distribution and differentially infect other hosts besides man (e.g., birds and large vertebrates
  9     [horses, etc.] for some, birds and swine for another, and small woodland animals for others).
 10     All have temperature-dependent components involved in development or transmission of the
 11     viruses, but specific effects vary for different types. For example, the maximum temperatures
 12     allowing the western equine encephalitis (WEE) vector to transmit the  virus effectively are below
 13     25 °C, and this allows for earlier spread of the disease in warm periods and the possible more
 14     northern spread of the disease.  In contrast, St. Louis encephalitis (SLE) arbovirus development
 15     and transmission are markedly enhanced by temperatures exceeding 25 °C. Rainfall and
 16     moisture patterns are also important, with most vectors (e.g., Cx tarsalis) benefitting from higher
 17     rainfall; but at least one (Cxpipiens) is enhanced by less rainfall, with outbreaks of its
 18     encephalitis syndrome being more common during high-temperature drought periods. Thus,
 19     effects of global warming and climate change on the incidence and spread of arbovirus-related
20     encephalitis syndromes are difficult to predict. However, it generally appears  that higher
21      temperatures should enhance the abundance and wider U.S. geographic distribution of most of
22     the competent mosquito vectors.  All of this again assumes that higher temperatures and rainfall
23      patterns will be such to allow adequate habitats for other hosts besides humans in the potential
24     new range areas. Lastly, as noted before for the other infectious diseases discussed, the
25      movement of populations into or out of the affected areas also will be important in determining
26      any location-specific increased (or decreased) incidence of arbovirus-related encephalitises.
27      Of special concern would be the introduction of any new arboviruses not now  currently endemic
28      in the United States (e.g., Japanese B encephalitis [VBE], not currently found hi the United
29      States but closely related to SLE in terms of involving Culex mosquitos and birds, with several
30      Culex subspecies in the United States found to be effective vectors for the virus).
        March 2001
4-149
DRAFT-DO NOT QUOTE OR CITE

-------
 1           The above discussion of potential effects of global warming and climate change on the
 2      incidence and spread of infectious diseases is further complicated by considerations of possible
 3      impacts of expected sea-level rise in response to global warming. Some low-lying coastal areas
 4      now serving as excellent habitats for certain mosquitos, for example, might come to be inundated
 5      by seawater and no longer be available breeding areas.  Or, increased storm surges or expansion
 6      of marshy areas reaching farther inland might contribute to creation of conditions in some areas
 7      more favourable to enhance mosquito breeding.  Also of concern is the potential for disruption of
 8      sanitation systems. The spread of infectious diseases, besides the vector-borne types discussed
 9     above, could be increased because of flooding of coastal cities secondary to heavy precipitation
10     events (e.g., hurricanes).  Inundation of sewage treatment facilities and sewage lines might not
11      only result in immediate spread of disease-containing fecal or other material, but damage to such
12     sanitation system components could result in longer term disruption of waste-removal
13     capabilities and the spread of disease.
14           Lastly, another concern with climate-induced heavy precipitation events or sea-level rise is
15     the potential for flooding of inland or coastal waste disposal sites. This could result in increased
16     spread of waterborne infectious diseases, depending on the specific materials present in such
17     dumps and the extent of their dispersal caused by flooding. The flooding of dump sites
18     containing hazardous chemical wastes represents yet another potential concern associated with
19     sea-level rise.  The spread of various toxic chemicals from such waste disposal sites could carry
20     with it increased threats of many types of possible health effects, as well as potential
21     environmental effects (natural vegetation and ecosystem damage, contamination of crop lands by
22     toxic chemicals, etc.).
23           Difficulties in projecting region-specific climate change impacts are complicated further by
24     the need to evaluate potential effects of local- or regional-scale changes in key air pollutants not
25     only on global scale temperature trends but also in terms of potentially more local- or regional-
26      scale impacts on temperature and precipitation patterns. Of much importance for this are varying
27     roles played by atmospheric particles.
28
29      4.5.2.2  Airborne Particle Relationships to Global Warming and Climate Change
30           Atmospheric particles both scatter and absorb incoming solar radiation at visible light
31     wavelengths. The scattering of solar radiation back to space leads to a decrease in transmission
        March 2001
4-150
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 of visible radiation to the Earth's surface and, hence, to a decrease in the heating rate of the
 surface and the atmosphere. The absorption of either incoming solar radiation or outgoing
 terrestrial infrared radiation by atmospheric particles results in heating of the lower atmosphere.
 Interactions of atmospheric particles with electromagnetic radiation from the visible through the
 infrared spectral regions are responsible for their direct effects on climate, which are the result of
 the same physical processes responsible for visibility degradation. Visibility reduction is caused
 by particle scattering in all directions, whereas climate effects result mainly from scattering in the
 upward direction.  The net effect of the above processes can be expressed as a radiative forcing,
 which is the change in the average net radiation at the top of the troposphere because of a change
 in solar (shortwave, or visible) or terrestrial (longwave, or infrared) radiation (Houghton et al.,
 1990). The radiative forcing drives the climate to respond, but because of uncertainties in a
 number of feedback mechanisms involving climate response, radiative forcing is used as a first-
 order estimate of the potential importance of various substances.  Sulfate particles scatter solar
 radiation effectively and do not absorb at visible wavelengths, whereas they absorb weakly at
 infrared wavelengths (IPCC, 1995). Nitrate particles exhibit grossly similar properties. The
 effects of mineral dust particles are complex; they weakly absorb solar radiation but their overall
 effect on solar radiation depends on particle size and the reflectivity of the underlying surface.
 They absorb infrared radiation and thus contribute to greenhouse warming (Tegen et al., 1996).
 Organic carbon particles mainly reflect solar radiation, whereas elemental carbon and other black
 carbon particles (e.g., PAHs with H:C ratios of <0.3) are strong absorbers of solar radiation
 (IPCC, 1995). However, the optical properties of carbonaceous particles are modified if they
 become coated with water or sulfuric acid. Particles containing black carbon also can exert  a
 direct effect after deposition onto surfaces that are more reflective (e.g., snow and ice).  In this
 case, additional solar radiation is absorbed by the surface; conversely, more reflective particles
 deposited on a dark surface result in additional solar radiation being reflected back to space.
     Anthropogenic (Twomey, 1974; Twomey, 1977) and biogenic (Charlson et al., 1987)
 sulfate particles also exert indirect effects on climate by serving as cloud condensation nuclei,
which results in changes in the size distribution of cloud droplets by producing more particles
with smaller sizes. The same mass  of liquid water in smaller particles leads to an increase in
amount of solar radiation that clouds reflect back to space because the total surface area of the
cloud droplets is increased.  This has been supported by satellite observations indicating that the
        March 2001
                                         4-151
DRAFT-DO NOT QUOTE OR CITE

-------
 1     effective radius of cloud droplets is smaller in the Northern Hemisphere than in the Southern
 2     Hemisphere (Han et al., 1994).  Smaller cloud droplets also have a lower probability of
 3     precipitating and, thus, have a longer lifetime than larger ones. Although the effects of sulfate
 4     have been considered most widely, interactions with other aerosol components also may be
 5     important.  Novakov and Penner (1993) have provided evidence that carbonaceous particles can
 6     modify the nucleation properties of sulfate particles.
 7          The amount of solar radiation incident on the earth-atmosphere system, or the solar
 8     constant, is 1370 W m'2, or 342.5 W rn2 on a globally averaged basis (calculated by dividing the
 9     solar constant by 4). The addition of sulfate and organic carbon as airborne PM results in
10     enhanced scattering and net cooling, whereas the addition of particles containing elemental
11     carbon results in absorption of solar and terrestrial radiation and net heating. The estimated
12     raditive forcing because of the scattering of solar radiation back to space caused mainly by
13     sulfate particles is - 0.4 W m"2 (IPCC,  1995), with an uncertainty range of a factor of two. The
14     uncertainty range reflects uncertainties in the emissions of SO2, the amount of SO2 that is
15     oxidized to sulfate, the atmospheric lifetime of sulfate, and the optical properties of the sulfate
16     particles. These values may be compared to the radiative forcing exerted by greenhouse gases of
17     about + 2.4 W m'2, with an uncertainty factor of 1.15 from the preindustrial era (ca.  1800) to
18      1994. Since the latter part of the 19th century, the mean surface temperature of the earth has
19     increased from 0.3 to 0.6 °C according to the IPCC (1995) assessment. Estimates of the indirect
20     effects of particles range from 0 to -1.5 W m"2 (IPCC, 1995). Because of a lack of quantitative
21     knowledge, no central value could be given.  Therefore, on a globally averaged basis, the direct
22     and indirect effects of anthropogenic sulfate particles likely have offset partially the warming
23     effects caused by increases in levels of greenhouse gases (Charlson et al., 1992).
24           Much of the work investigating the effects of particles on climate has focused on sulfate
25     particles. However, particles containing elemental carbon (EC) from fossil fuel combustion and
26     biomass burning or mineral  dust may exert radiative forcing, with spatial distributions very
27      different than for sulfate. Tegen et al. (1996) and Tegen and Lacis (1996) used a global scale
28     three-dimensional model to  evaluate the radiative forcing caused by mineral dust particles.
29      Tegen and Lacis (1996) found that the sign and the magnitude of the radiative forcing depends on
30     the height distribution of the dust and  the effective radius of the particles. In particular,  for a dust
31      layer extending from 0 km to 3 km, positive radiative forcing at visible wavelengths is found for
        March 2001
4-152
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
particle radii greater than 1.8 ^m, whereas negative forcing is found for smaller particles. They
calculated a global mean radiative forcing caused by mineral dust from all sources of 0.14 W m"2
and from mineral dust from lands disturbed by human activity of 0.09 W m"2.  This value
represents a near cancellation between a much larger solar forcing of-0.25 W m"2 and a thermal
forcing of 0.34 W m"2.  Uncertainty factors could not be estimated for these calculations because
they were judged to be largely unknown. Haywood and Shine (1995) estimated a global mean
radiative forcing of 0.1 W m"2, with an uncertainty factor >3, caused by the absorption of solar
radiation by EC released by fossil fuel combustion. The IPCC (1995) estimated a global mean
radiative forcing of - 0.1 W m"2 caused by particles produced by biomass burning, with an
uncertainty factor of three. The global mean radiative forcing exerted by particles would then be
-0.5 W m'2, with an uncertainty of about a factor of 2.4.  Figure 4-27 summarizes estimates of
global mean radiative forcing exerted by greenhouse gases and various types of particles.
     Deviations from the global mean values can be very large on the regional scale. For
instance, Tegen et al. (1996) found that local radiative forcing exerted by dust  raised from
disturbed lands ranges from -2.1 W m"2 to 5.5 W m"2 over desert areas and their adjacent seas.
The largest regional values of radiative forcing caused by anthropogenic sulfate are about
-3 W m'2 in the eastern United States, south central Europe,  and eastern China (Kiehl and
Briegleb, 1993). These regional maxima in aerosol forcing are at least a factor  of 10 greater than
their global mean values shown in Figure 4-27. By comparison, regional maxima in forcing by
the well-mixed greenhouse gases are only about 50% greater than their global mean value (Kiehl
and Briegleb, 1993). Thus, the estimates of local radiative forcing by particles also are large
enough to completely cancel the effects of greenhouse gases  in many regions and to cause a
number of changes in the dynamic structure of the atmosphere that still need to be evaluated.
A number of anthropogenic pollutants whose distributions are highly variable are also effective
greenhouse absorbers.  These gases include O3 and, possibly, HNO3, C2H4 , NH3, and SO2, all of
which are not commonly considered in radiative forcing calculations (Wang et  al. 1976). High
ozone values are found downwind of urban areas and areas where there is biomass burning.
However, Van Borland et al. (1997) found that there may not be much cancellation between the
radiative effects for ozone and for sulfate, because both species have different seasonal cycles
and show significant differences in their spatial distribution.
       March 2001
                                         4-153
DRAFT-DO NOT QUOTE OR CITE

-------
adiative forcing (W m"2)
J _* M W
III III
c.
CO
f "
15-1-
JQ
0
o -


T
1
JL.


•
— Halocarbons
— N2O
*~~" CH^
— c°2 Sulfate
T ^
T /\
I Tropospheric -.-
Ozone 1
Direct
Effect
In
E
Fossil Solar
Fuel Biomass Mineral Variability
Soot Burning QUSt "p
v .x, _ m
T
direct
!ffect
Confidence level
High Low Low Low Very Very Very Very Very
low low low low low
       Figure 4-27. Estimated global mean radiative forcing exerted by gas and various particle
                    phase species from 1850 to 1950.
       Source: Adapted from IPCC (1995) and Tegen and Lacis (1996).
 1          Observational evidence for the climatic effects of particles is sparse.  Haywood et al. (1999)
 2      found that the inclusion of anthropogenic aerosols results in a significant improvement between
 3      calculations of reflected sunlight at the top of the atmosphere and satellite observations in
 4      oceanic regions close to sources of anthropogenic PM.
 5          Uncertainties in calculating the direct effect of airborne particles arise from a lack of
 6      knowledge of their vertical and horizontal variability, their size distribution, chemical
 7      composition and the distribution of components within individual particles. For instance,
 8      gas-phase sulfur species may be oxidized to form a layer of sulfate around existing particles in
 9      continental environments, or they may be incorporated in sea-salt particles (e.g., Li-Jones and
10      Prospero, 1998).  In either case, the radiative effects of a given mass of the sulfate will be much
        March 2001
4-154
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
lower than if pure sulfate particles were formed. It also must be stressed that the overall radiative
effect of particles at a given location is not simply determined by the sum of effects caused by
individual classes of particles because of interactions between particles with different radiative
characteristics and with gases.
     Calculations of the indirect effects of particles on climate are subject to much larger
uncertainties than are calculations of their direct effects, reflecting uncertainties in a large
number of chemical and microphysical processes in describing the effects of sulfate on the size
distribution and number of droplets within a cloud. A complete assessment of the radiative
effects of PM will require supercomputer calculations that incorporate the spatial and temporal
behavior of particles of varying composition that have been emitted or formed from precursors
emitted from different sources. Refining values of model input parameters (such as improving
emissions estimates) may be as important as improving the models per se in calculations of direct
radiative forcing (Pan et al., 1997) and indirect radiative forcing (Pan et al., 1998) caused by
sulfate. However, uncertainties associated with the calculation of radiative effects of particles
likely will remain much larger than those associated with well-mixed greenhouse gases.
     This means that, although on a global scale atmospheric particles likely exert an overall net
effect of slowing global warming, much uncertainty would apply to any modeling efforts aimed
at projecting net effects on global warming processes, resulting climate change, and any
consequent human health or environmental effects because of location-specific increases or
decreases in anthropogenic emissions of atmospheric particles or their precursors. For example,
any net impacts of regional sulfates in reducing global-climate-change-induced increases in local
temperatures may well be offset partially by local surface level heating because of carbonaceous
particles from diesel emissions or coal combustion energy generation being deposited on snow or
ice covered surfaces or contributing to more rapid evaporation or rainout of water from overhead
clouds.
        March 2001
                                         4-155
DRAFT-DO NOT QUOTE OR CITE

-------
 1     4.6  SUMMARY
 2     4.6.1  Particulate Matter Effects on Vegetation and Ecosystems
 3          Human existence on this planet depends on ecosystems and the services and products they
 4     provide. Both ecosystem structure and function play an essential role in providing societal
 5     benefits. Society derives two types of benefits from the structural aspects of an ecosystem:
 6     (1) products with market value such as fish, minerals, forage, forest products, biomass fuels,
 7     natural fiber, and many pharmaceuticals, and the genetic resources of valuable species (e.g.,
 8     plants  for crops and timber and animals for domestication); and (2) the use and appreciation of
 9     ecosystem for recreation, aesthetic enjoyment, and study.
10          Ecosystem functions that maintain clean water, pure air, a green earth, and a balance of
11     creatures, are functions that enable humans to survive.  They are the dynamics of ecosystems.
12     The benefits they impart include absorption and breakdown of pollutants, cycling of nutrients,
13     binding of soil, degradation of organic waste, maintenance of a balance of gases in the air,
14     regulation of radiation balance, climate, and the fixation of solar energy. Concern has risen in
15     recent years concerning the integrity of ecosystems because there are few ecosystems on the
16     Earth today that are not influenced by humans.  For this reason, the deposition of PM and its
17     impact on vegetation and ecosystems is of great importance.
18          The PM whose effects on vegetation and ecosystems are considered in this chapter is not a
19     single  pollutant but represents a heterogeneous mixture of particles differing in origin, size, and
20     chemical constituents. The effects of exposure to a given mass concentration of PM of particular
21     size (measured as PM10; PM2 5, etc.) may, depending on the particular mix of deposited particles,
22     lead to widely differing phytotoxic responses. This has not been characterized adequately.
23          Atmospheric deposition of particles to ecosystems takes place via both wet and dry
24     processes through the three major routes indicated below.
25          (1) Precipitation scavenging, in which particles are deposited in rain and snow
26          (2) Fog, cloud water, and mist interception
27          (3) Dry deposition, a much slower, yet more continuous removal to surfaces
28          Deposition of heavy metal particles to ecosystems occurs by wet and dry processes. Dry
29     deposition is considered more effective for coarse particles of natural origin and elements such as
30     iron and manganese, whereas wet deposition generally is more effective for fine particles of
       March 2001
4-156
DRAFT-DO NOT QUOTE OR CITE

-------
  1      atmospheric origin and elements such as cadmium, chromium, lead, nickel, and vanadium. The
  2      actual importance of wet versus dry deposition, however, is highly variable, depending on the
  3      type of ecosystem, location, and elevation.
  4           Deposition of PM on above-ground plant parts can have either a physical and or chemical
  5      impact, or both.  Particles transferred from the atmosphere to plant surfaces may cause direct
  6      effects if they (1) reside on the leaf, twig, or bark surface for an extended period; (2) be taken up
  7      through the leaf surface; or (3) are removed from the plant via resuspension to the atmosphere,
  8      washing by rainfall, or litter-fall with subsequent transfer to the soil.
  9           Chemical effects include excessive alkalinity or acidity. The effects of "inert" PM are
10      mainly physical, whereas the effects of toxic particles are both chemical and physical.  The
11      effects of dust deposited on plant surfaces or on soil are more likely to be associated with their
12      chemistry than with the mass of deposited particles and are usually of more importance than any
13      physical effects.  The majority of the easily identifiable direct and indirect effects,  other than
14      climate-change impacts, occur in severely polluted areas around heavily industrialized point
15      sources such as limestone quarries; cement kilns; and iron; lead, and various smelting factories.
16      Studies of the direct effects of chemical additions to foliage in particulate deposition have found
17      little or no effects of PM on foliar processes; however, both conifers and deciduous species have
18      shown significant effects on leaf surface structures after exposure to  simulated acid rain or mist
19      at pH 3.5.  Many experimental studies indicate that epicuticular waxes (which function to prevent
20      water loss from plant leaves) can be destroyed by acid rain in a few weeks. This function is
21      particularly crucial in conifers because of the longevity of evergreen foliage.
22           Though there has been no direct evidence of a physiological association between tree injury
23      and exposure to metals, heavy metals  have been implicated because their deposition pattern is
24      correlated with forest decline. The role of heavy metals has been indicated by phytochelatin
25      measurements. Phytochelatins are intracellular metal-binding peptides that act as indicator of
26      metal stress. Because they are produced by plants as a response to sublethal concentrations of
27      heavy metals, they can be used to indicate that heavy metals are involved in forest decline.
28      Concentrations of the phytochelatins increased with altitude, as did forest decline, and they also
29      increased across regions showing increased levels of forest injury.
30           Secondary organics formed in the atmosphere have been referred to under the following
31      terms:  toxic substances, pesticides, hazardous air pollutants (HAPS), air toxics, semivolatile
        March 2001
4-157
DRAFT-DO NOT QUOTE OR CITE

-------
 1     organic compounds (SOCs), and persistent organic pollutants (POPS). The chemical substances
 2     listed under the above headings are not criteria pollutants controlled by NAAQS as cited under
 3     CAA Sections 108 and 109 (U.S. Code, 1991), but rather are controlled under CAA Sect. 112,
 4     Hazardous Air Pollutants.  Their possible effects in the environment on humans and ecosystems
 5     are discussed in many other government documents and publications.  They are mentioned in this
 6     chapter because, in the atmosphere many of the chemical compounds are partitioned between gas
 7     and particle phases and are deposited as particulate matter. As particles, they become airborne
 8     and can be distributed over a wide area and impact remote ecosystems. Some of the chemical
 9     compounds are of concern to humans because they may reach toxic levels in food chains of both
10     animals and humans, whereas others tend to decrease or maintain the same toxicity as they move
11     through the food chain.
12          An important characteristic of fine particles is their ability to affect the flux of solar
13     radiation passing through the atmosphere directly, by scattering and absorbing solar radiation,
14     and indirectly, by acting as cloud condensation nuclei that, in turn,  influence the optical
15     properties of clouds. Regional haze has been estimated to diminish surface solar visible radiation
16     by approximately 8%. Crop yields have been reported as being sensitive to the amount of
17     sunlight received, and crop losses have been attributed to increased airborne particle levels in
18     some areas of the world.
19          The transmission of solar UV-B radiation through the Earth's atmosphere is controlled by
20     ozone, clouds, and particles. The depletion of stratospheric ozone caused by the release of
21     chlorofluorcarbons and other ozone-depleting substances has resulted in heightened concern
22     regarding potentially serious increases in the amount of solar UV-B (SUVB) radiation reaching
23     the Earth's surface. Plant  species vary enormously in their response to UV-B exposures, and
24     large differences in response also occur among different genotypes within a species.  In general,
25     dicotyledonous plants are more sensitive than monocotyledons from similar environments.
26     In addition, plant responses may differ depending on stage of development. Because plants
27     evolved under'the selective pressure of ambient UV-B radiation in sunlight, they have developed
28     adaptive mechanisms.  Although inhibition of photosynthesis is a detrimental growth effect,
29     flavonoid synthesis represents acclimation. Plants growing under full light have been shown to
30     be protected against UV-B effects but not when growing under weak visible light. A common
        March 2001
4-158
DRAFT-DO NOT QUOTE OR CITE

-------
  1      adaptation is alteration in leaf transmission properties, which results in attenuation of UV-B in
  2      the epidermis before it can reach the leaf interior.
  3           Indirect effects of PM on plants are usually the most significant because they can alter
  4      nutrient cycling in ecosystems and inhibit plant uptake of nutrients and, therefore, have a great
  5      impact on ecosystem biodiversity. Indirect effects occur through the soil and result from the
  6      deposition of heavy metals, nitrates, sulfates, or acidic precipitation and their impact on the soil
  7      microbial community. The soil environment is one of the most dynamic sites of biological
  8      interaction in nature. Bacteria in the soil are essential components of the nitrogen and sulfur
  9      cycles that make these elements  available for plant uptake. Fungi form mycorrhizae,
10      a mutualistic symbiotic relationship, that is integral in mediating plant uptake of mineral
11      nutrients. Changes in the soil environment that influence the role of the bacteria and fungi in
12      nutrient cycling and availability determine plant and ecosystem response.
13           Major impacts of PM on soil environments occur through deposition of nitrates and sulfates
14      and the acidifying effect of the ET ion associated with these compounds in wet and dry
15      deposition.  Although the soils of most of North American forest ecosystems are nitrogen
16      limited, there are some forests that exhibit severe symptoms of nitrogen saturation. They include
17      the high-elevation, spruce-fir ecosystems in the Appalachian Mountains; the eastern hardwood
18      watersheds at the Femow Experimental Forest near Parsons, WV; the mixed conifer forest and
19      chaparral watershed with high smog exposure in the Los Angeles Air Basin; the high-elevation
20      alpine watersheds in the Colorado Front Range; and a deciduous forest in Ontario, Canada.
21           Nitrogen saturation results when additions to soil background nitrogen (nitrogen loading)
22      exceed the capacity of plants and soil microorganisms to utilize and retain nitrogen. An
23      ecosystem no longer functions as a sink under these circumstances.  Possible ecosystem
24      responses to nitrate saturation, as postulated by Aber and his coworkers, include (1) a permanent
25      increase in foliar nitrogen and reduced foliar phosphorus and lignin because of the lower
26      availability of carbon, phosphorus, and water; (2) reduced productivity in conifer stands caused
27      by disruptions of physiological function; (3) decreased root biomass and increased nitrification
28      and nitrate leaching; (4) reduced soil fertility, the results of increased cation leaching, increased
29      nitrate and aluminum concentrations in streams, and decreased water quality.  Saturation implies
30      that some resource other than nitrogen is limiting biotic function.  Water and phosphorus for
31      plants and carbon for microorganisms are the resources most likely to be the secondary limiting
        March 2001
4-159
DRAFT-DO NOT QUOTE OR CITE

-------
 1      factors.  The appearance of nitrogen in soil solution is an early symptom of excess nitrogen. In
 2      the final stage, disruption of forest structure becomes visible.
 3           Changes in nitrogen supply can have a considerable impact on an ecosystem's nutrient
 4      balance. Increases in soil nitrogen play a selective role. Plant succession patterns and
 5      biodiversity are affected significantly by chronic nitrogen additions in some ecosystems.
 6      Long-term nitrogen fertilization studies in both New England and Europe suggest that some
 7      forests receiving chronic inputs of nitrogen may decline in productivity and experience greater
 8      mortality.  For example, long-term fertilization experiments at Mount Ascutney, VT, suggest that
 9      declining coniferous forest stands with slow nitrogen cycling may be replaced by deciduous
10      fast-growing forests that cycle nitrogen rapidly. Excess nitrogen inputs to unmanaged heathlands
11      in the Netherlands also have been found to result in nitrophilous grass species replacing slower
12      growing heath species.  Over the past several decades, the  composition of plants in the forest
13      herb layers had been shifting toward species commonly found on nitrogen-rich areas.  It also was
14      observed that the fruiting bodies of mycorrhizal fungi had  decreased in number.
15           Notable impacts of excess nitrogen deposition also have been observed with regard to
16      aquatic systems.  For example, atmospheric nitrogen deposition into soils in watershed areas
17      feeding into estuarine sound complexes (e.g., the Pamlico  Sound of North Carolina) appear to
18      contribute to excess nitrogen flows in runoff (especially during and after heavy rainfall events
19      such as hurricanes).  Together with excess nitrogen runoff from agricultural practices or other
20      uses (e.g., fertilization of lawns or gardens), massive influxes of such nitrogen into watersheds
21      and sounds can lead to dramatic decreases in water oxygen and increases in algae blooms that
22      can cause extensive fish kills and damage to commercial fish and sea food harvesting.
23           Acidic deposition has played a major role in soil acidification in some areas of Sweden,
24      elsewhere in Europe, and in eastern North America.  Soil acidification and its effects result from
25      deposition of nitrates, sulfates, and associated H+ ion. A major concern is that soil acidity will
26      lead to nutrient deficiency. Growth of tree species can be affected when high aluminum-to-
27      nutrient ratios limit uptake of calcium and magnesium and create a nutrient deficiency. Calcium
28      is essential in the formation of wood and the maintenance of cells (the primary plant tissues
29      necessary for tree growth), and it must be dissolved in soil water to be taken up by plants. Acidic
30      deposition can increase aluminum concentrations in soil water by lowering the pH in aluminum-
31      rich soils through dissolution and ion-exchange processes.  Aluminum in soil can then be taken
        March 2001
4-160
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 up by roots more readily than calcium because of its greater affinity for negatively charged
 surfaces. Tree species can be adversely affected if altered Ca/Al ratios impair Ca or Mg uptake.
      Overall, then, PM produced by human activities has the potential to cause the loss of
 ecosystem biodiversity in ways that reduces the ability of ecosystems to provide the services that
 society requires to sustain life.  The major impacts of PM on ecosystems are the indirect effects
 that occur through the soil and affect plant growth, vigor, and reproduction. Mineral nutrient
 cycling can be altered by the deposition of heavy metals. The deposition of nitrogen and sulfur
 and the acidifying effects of the two in association with the H+ ion in precipitation also alter
 biogeochemical cycling, cause soil acidification, alter the Ca/Al ratio,  and impact the growth of
 vegetation and forest trees, in particular. Leaching of nitrates and other minerals through runoff
 can impact coastal and aquatic wetlands and, thus,  influence their ability to produce the products
 and services necessary for human society.

 4.6.2 Particulate Matter-Related Effects on Visibility
      Visibility is defined as the degree to which the atmosphere is transparent to visible light and
 the clarity and color fidelity of the atmosphere. Visual range is the farthest distance a black
 object can be distinquished against the horizontal sky. Visibility impairment is any humanly
 perceptible change in visibility. For regulatory purposes, visibility impairment, characterized by
 light extinction, visual range, contrast, and coloration, is classified into two principal forms:
 (1) "reasonably attributable" impairment, attributable to a single source or small group of
 sources, and (2) regional haze, any perceivable change in visibility caused by a combination of
 many sources  over a wide geographical area.
     Visibility is measured by human observation,  light scattering by particles, the light
 extinction-coefficient and parameters related to the light-extinction coefficient (visual range and
 deciview scale), the light scattering coefficient, and fine PM concentrations. The air quality
 within a sight  path will affect the illumination of the sight path by scattering or absorbing solar
 radiation before  it reaches the Earth's surface.  The rate of energy loss with distance from a beam
 of light is the light extinction coefficient. The  light extinction coefficient is the sum of the
 coefficients for light absorption by gases (oag),  light scattering by gases (asg), light absorption by
particles (aap), and light scattering by particles (asp). Atmospheric particles are frequently divided
into fine and coarse particles. Corresponding coefficients for light scattering and absorption by
March 2001                               4-161        DRAFT-DO NOT QUOTE OR CITE

-------
 1     fine and coarse particles are osfp and oafp and oscp and oacp, respectively.  Visibility within a sight
 2     path longer than approximately 100 km (60 mi) is affected by change in the optical properties of
 3     the atmosphere over the length of the sight path.
 4          Visibility impairment is associated with airborne particle properties, including size
 5     distributions (i.e., fine particles in the 0.1- to 1.0-/zm size range) and aerosol chemical
 6     composition, and with relative humidity.  With increasing relative humidity, the amount of
 7     moisture available for absorption by particles increases, thus causing the particles to increase in
 8     both size and volume. As the particles increase in size and volume, the light scattering potential
 9     of the particles also generally increases. Visibility impairment is greatest in the eastern United
10     States and Southern California. In the eastern United States, visibility impairment is caused
11     primarily by light scattering by sulfate aerosols and, to a lesser extent, by nitrate particles and
12     organic aerosols, carbon soot, and crustal dust. Haziness in the  southeastern United States,
13     caused by increased atmospheric sulfate, has increased by ca. 80% since the 1950s and is greatest
14     in the summer months, followed by the spring and fall, and winter. Light scattering by nitrate
15     aerosols is the major cause of visibility impairment in Southern California. Nitrates contribute
16     about 40% to the total light extinction in Southern California and accounts for 10 to 20% of the
17     total extinction in other U.S. areas.
18           Organic particles are the second largest contributors to light extinction in most U.S. areas.
19     Organic carbon is the greatest cause of light extinction in the Pacific Northwest, Oregon, Idaho,
20     and Montana, accounting  for 40 to 45% of the total extinction.  Also, organic carbon contributes
21     between 15 to 20% to the  total extinction in most of the western United States and 20 to 30% in
22     the remaining U.S. areas.
23           Coarse mass and soil, primarily considered "natural extinction", is responsible for some of
24     the visibility impairment in northern California and Nevada, Oregon, southern Idaho, and
25     western, Wyoming. Dust transported from Southern California and the subtropics has been
26     associated with regional haze in the Grand Canyon and other southwestern U.S. class I areas.
27
28     4.6.3 Particulate Matter-Related Effects on Materials
29           Building materials (metals, stones, cements, and paints) undergo natural weathering
30     processes from exposure to environmental elements (wind, moisture, temperature fluctuations,
31     sun light, etc.). Metals form a protective  film that protects against environmentally induced
        March 2001
4-162
DRAFT-DO NOT QUOTE OR CITE

-------
 1      corrosion. The natural process of metal corrosion from exposure to natural environmental
 2      elements is enhanced by exposure to anthropogenic pollutants, in particular SO2, rendering the
 3      protective film less effective.
 4           Dry deposition of SO2 enhances the effects of environmental elements on calcereous stones
 5      (limestone, marble, and cement) by converting calcium carbonate (calcite) to calcium sulfate
 6      dihydrate (gypsum).  The rate of deterioration is determined by the SO2 concentration, the stone's
 7      permeability and moisture content, and the deposition rate; however, the extent of the damage to
 8      stones produced by the pollutant species apart from the natural weathering processes is uncertain.
 9      Sulfur dioxide also has been found to limit the life expectancy of paints by causing discoloration
10      and loss of gloss and thickness of the paint film layer.
11           A significant detrimental effect of particle pollution is the soiling of painted surfaces and
12      other building materials. Soiling changes the reflectance of a material from opaque and reduces
13      the transmission of light through transparent materials.  Soiling is a degradation process that
14      requires remediation by cleaning or washing, and, depending on the soiled surface, repainting.
15      Available data on pollution exposure indicates that particles can result in increased cleaning
16      frequency of the exposed surface and may reduce the life usefulness of the material soiled.
17      Attempts have been made to quantify the pollutants exposure levels at which materials damage
18      and soiling have been perceived.  However, to date, insufficient data are available to advance our
19      knowledge regarding perception thresholds with respect to pollutant concentration, particle size,
20      and chemical composition.
21
22      4.6.4 Effects  of Participate Matter on the Transmission of Solar Ultraviolet
23            Radiation and Global Warming Processes
24           Extensive potential future impacts on human health and the environment are projected to
25      occur because of increased transmission of solar ultraviolet radiation (UV-B) through the Earth's
26      atmosphere, secondary to stratospheric ozone depletion resulting from anthropogenic emissions
27      of chlorofluorcarbons (CFCs), halons, and certain other gases. However, the estimation of the
28      likely future extent of detrimental effects caused by increased penetration of solar UV-B to the
29      Earth's surface is complicated by atmospheric particle effects, which vary depending on size and
30      composition of particles that can differ substantially over different geographic areas and from
31      season to season over the same area. Also, atmospheric particles greatly complicate projections
        March 2001
4-163
DRAFT-DO NOT QUOTE OR CITE

-------
 1      of future trends in global warming processes because of emissions of greenhouse gases;
 2      consequent increases in global mean temperature, and resulting changes in regional and local
 3      weather patterns; and mainly deleterious (but some beneficial) location-specific human health
 4      and environmental impacts.
 5          The physical processes (i.e., scattering and absorption) responsible for airborne particle
 6      effects on transmission of solar ultraviolet and visible radiation are the same as those responsible
 7      for visibility degradation. Scattering of solar radiation back to space and absorption of solar
 8      radiation  determine the effects of an aerosol layer on solar radiation. The transmission of solar
 9      UV-B radiation is affected strongly by atmospheric particles.  Measured attenuations of UV-B
10      under hazy conditions range up to 37% of the incoming solar radiation. Measurements relating
11      variations in PM mass directly to UV-B transmission are lacking. Particles also can affect the
12      rates of photochemical reactions occurring in the atmosphere.  Depending on the amount of
13      absorbing substances in the particles, photolysis rates either can be increased or decreased.
14          In addition to direct climate effects through the scattering and absorption of solar radiation,
15      particles also exert indirect effects on climate by serving  as cloud condensation nuclei, thus
16      affecting  the abundance and vertical distribution of clouds. The direct and indirect effects of
17      particles appear to have significantly offset the global warming effects caused by the buildup of
18      greenhouse gases because the onset  of the Industrial Revolution, on a globally averaged basis.
19      However, because the lifetime of particles is much shorter than that required for complete mixing
20      within the Northern Hemisphere, the climate effects of particles generally are felt much less
21      homogeneously than are the effects of long-lived greenhouse gases.
22          Any effort to model the impacts of local alterations  in particle concentrations on projected
23      global climate change or consequent local and regional weather patterns would be subject to
24      considerable uncertainty.  This also would be the case for any projections of impacts of location-
25      specific airborne PM alterations on potential human health or environmental effects associated
26      with either increased atmospheric transmission of solar UV radiation or global warming
27      secondary to accumulation of stratospheric ozone-deleting substances or "greenhouse gases."
        March 2001
4-164
DRAFT-DO NOT QUOTE OR CITE

-------
  1
 REFERENCES
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
45
46
47
48
49
 50
51
52
53
54
 Abdullah, U. M.; Iqbal, M. Z. (1991) Response of automobile, stone and cement particulate matters on stomatal
       clogging of plants. Geobios (Jodhpur, India) 18:196-202.
 Aber, J. D.; Nadelhoffer, K. J.; Steudler, P.; Melillo, J. M. (1989) Nitrogen saturation in northern forest ecosystems:
       excess nitrogen from fossil fuel combustion may stress the biosphere. Bioscience 39: 378-386.
 Aber, J.; McDowell, W.; Nadelhoffer, K.; Magill, A.; Berntson, G.; Kamakea, M.; McNulty, S.; Currie, W.;
       Rustad, L.; Fernandez, I. (1998) Nitrogen saturation in temperate forest ecosystems. BioScience 48: 921-934.
 Adams, K. M.; Kavis, L. I., Jr.; Japar, S. M.; Pierson, W. R. (1989) Real-time, in situ measurements of atmospheic
       optical absorption in the visible via photoacoustic spectroscopy - II. validation for atmospheric elemental
       carbon aerosol. Atmos. Environ. 23:  693-700.
 Alexander, M. (1977) Introduction to soil microbiology. 2nd ed. New York, NY: John Wiley & Sons- pp 225-230-
       239-246; 272-286; 355-356.
 Allen, M. F. (1991) The ecology of mycorrhizae. Cambridge, MA: Cambridge University Press.
 Allen, E. R.; Cabrera, N.; Kim, J.-C.  (1994) Atmospheric deposition studies in rural forested areas of the
       southeastern United States. Presented at: 87th annual meeting & exhibition of the Air & Waste Management
       Association; June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association- paper no
       94-WP88.02.
 Altshuller, A. P.;  Linthurst, R. A., eds. (1984) Acidic deposition phenomenon and its effects: critical assessment
       review papers; volume I, atmospheric sciences; volume II, effects sciences. Washington, DC: U.S.
       Environmental Protection Agency, Office of Research and Development; EPA report nos.
       EPA/600/8-83/016aF-bF. Available from: NTIS, Springfield, VA; PB85-100030 and PB85-100048.
 Andersen, C. P.; Rygiewicz, P. T. (1991) Stress interactions and mycorrhizal plant response: understanding carbon
       allocation priorities. Environ. Pollut.  73: 217-244.
 Antonovics, J.; Bradshaw, A. D.; Turner, R. G. (1971) Heavy metal tolerance in plants. In: Cragg, J. B., ed.
       Advances in ecological research. Volume 7. London, United Kingdom: Academic Press; pp.  1-85.
 April, R.; Newton, R. (1992) Mineralogy and mineral weathering. In: Johnson, D. W.; Lindberg, S.  E., eds.
       Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
       Springer-Verlag; pp. 378-425. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
      Ecological  studies: analysis and synthesis: v. 91).
 Askey, A.; Lyon, S. B.; Thompson, G. E.; Johnson, J. B.; Wood, G. C.; Sage, P. W.; Cooke, M. J. (1993) The effect
      of fly-ash particulates on the atmospheric corrosion of zinc and mild steel. Corros. Sci. 34: 1055-1081.
 Ausset, P.; Bannery, F.; Del Monte, M.; Lefevre, R. A. (1998) Recording of pre-industrial atmospheric environment
      by ancient crusts on  stone monuments. Atmos. Environ. 32: 2859-2863.
 Ayensu, E.; Van R. Claasen, D.; Collins, M.; Dealing, A.; Fresco, L.; Gadgil, M.; Gitay, H.; Glaser, G.; Juma, C.;
      Krebs, J.; lenton, R.; Lubchenco, J.; McNeely, J. A.; Mooney, H. A.;  Pinstrup-Andersen, P.; Ramos, M.;
      Raven, P.; Reid, W. V.; Samper, C.; Sarukhan, J.; Schei, P.; Tundisi,  J. G.; Watson, R. T.; Guanhua, X.;
      Zakri, A. H. (1999) International ecosystem assessment. Science 286: 685-686.
 Babich, H.; Stotzky, G. (1978) Effects of cadmium on the biota: influence of environmental factors. In: Perlman, D.,
      ed. Advances in applied microbiology, volume 23. New York, NY: Academic Press; pp.  55-117.
 Baedecker, P. A.;  Edney, E. O.; Moran, P. J.; Simpson, T. C.; Williams, R. S. (1991) Effects of acidic deposition on
      materials. In: Irving, P. M., ed. Acidic deposition: state of science and technology, volume III: terrestrial,
      materials, health and visibility effects. Washington, DC: The U.S. National Acid Precipitation Assessment
      Program. (State of science and technology report no. 19).
 Baedecker, P. A.; Reddy, M. M.; Reimann, K. J.; Sciammarella, C. A. (1992) Effects of acidic deposition on the
      erosion of carbonate stone-experimental results from the U.S. National Acid Precipitation Assessment
      Program (NAPAP). Atmos. Environ. Part B 26: 147-158.
 Barone, J. B.; Cahill,  T. A.; Eldred, R. A.; Flocchini, R. G.; Shadoan, D. J.; Dietz, T. M. (1978) A multivariate
      statistical analysis of visibility degradation at four California cities. Atmos. Environ. 12: 2213-2221.
Barton, K. (1958)  The influence of dust on atmospheric corrosion of metals. Werkst. Korros. 8/9: 547-549.
Bellan, L. M.; Salmon, L. G.; Cass, G. R. (2000) A study on the human ability to detect soot deposition onto works
      of art. Environ. Sci. Technol. 34:  1946-1952.
Bewley, R. J. F.; Parkinson, D. (1984) Effects of sulphur dioxide pollution on forest soil microorganisms. Can J
      Microbiol. 30:  179-185.
Biefer, G. J. (1981) Atmospheric corrosion of steel in the Canadian Arctic. Mater. Perform. 20:  16-19.
         March 2001
                                                4-165
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Billings, W. D. (1978) Plants and the ecosystem. 3rd ed. Belmont, CA: Wadsworth Publishing Company, Inc.;
 2             pp. 1-62,83-108.
 3       Binenko, V. I.; Harshvardhan, H. (1993) Aerosol effects in radiation transfer. In: Jennings, S. G., ed.
 4             Aerosol effects on climate. Tucson, AZ: University of Arizona Press; pp. 190-232.
 5       Binkley, D.; Valentine, D.; Wells, C.; Valentine, U. (1989) An empirical analysis of the factors contributing to
 6             20-year decrease in soil pH in an old-field plantation of loblolly pine. Biogeochemistry 8: 39-54.
 7       Bjom, L. O. (1996) Effects of ozone depletion and increased UV-B on terrestrial ecosystems. Int. J. Environ. Stud.
 8             51:217-243.
 9       Blackburn, T. E. (1992) Solar photolysis of ozone to singlet D oxygen atoms. J. Geophys.  Res. 97: 10,109-10,117.
10       Blandford, J. C. (1994) Transmissometer data reduction and validation (IMPROVE protocol). Fort Collins, CO:
11             Air Resource Specialists, Inc.; technical instruction no. 4400-5000.
12       Blaschke, H. (1990) Mycorrhizal populations and fine root development on Norway spruce exposed to controlled
13             doses of gaseous pollutants and simulated acidic rain treatments. Environ. Pollut. 68: 409-418.
14       Bobbink, R. (1998) Impacts of tropospheric ozone and airborne nitrogenous pollutants on  natural and semi-natural
15             ecosystems: a commentary. New Phytol.  139:161-168.
16       Bobbink, R.; Homung, M; Roelofs, J. G. M. (1998) The effects of air-borne nitrogen pollutants on species diversity
17             in natural and semi-natural European vegetation. J. Ecol. 86: 717-738.
18       Bolle, H.-J.; Seiler, W.; Bolin, B. (1986) Other greenhouse gases and aerosols: assessing their role for atmospheric
19             radiative transfer. In: Bolin, B.; Doos, B. R.; Jager, J.; Warrick, R. A., eds. The greenhouse effect, climatic
20             change, and ecosystems. Chichester, United Kingdom: John Wiley and Sons; pp. 157-203. (SCOPE 29).
21       Bondietti, E. A.; McLaughlin, S. B. (1992) Evidence of historical influences of acidic deposition on wood and soil
22             chemistry. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric deposition and forest nutrient cycling:
23             a synthesis of the integrated forest study. New York, NY: Springer-Verlag: pp. 358-377. (Billings, W. D.;
24             Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds. Ecological studies analysis and synthesis: v. 91).
25       Bowden, R. D.; Geballe, G. T.; Bowden, W. B. (1989) Foliar uptake of 15N from simulated cloud water by red
26             spruce (Picea rubens) seedlings. Can. J. For. Res.  19: 382-386.
27       Bradley, R.; Burt, A. J.; Read, D. J. (1981) Mycorrhizal infection and resistance to heavy metal toxicity in Calluna
28             vulgaris. Nature (London) 292: 335-337.
29       Brandt, C. J.; Rhoades, R. W. (1972) Effects of limestone dust accumulation on composition of a forest community.
30             Environ. Pollut. 3:217-225.
31       Brandt, C. J.; Rhoades, R. W. (1973) Effects of limestone dust accumulation on lateral growth of forest trees.
32             Environ. Pollut. 4: 207-213.
33       Briihl, C.; Crutzen, P. J. (1988) On the disproportionate role of tropospheric ozone as a filter against solar UV-B
34            radiation. Geophys. Res. Lett. 16: 703-706.
35       Bruhn, J. N. (1980) Effects of oxidant air pollution on Ponderosa and Jeffrey pine foliage decomposition
36             [dissertation]. Berkeley, CA: University of California; Dissertation Abstracts International v. 42-07B.
37      Brumme, R.;  Leimcke, U.; Matzner, E. (1992) Interception and uptake of NH,, and NO3 from wet deposition by
38            above-ground parts of young beech (Fagus silvatica L.) trees. Plant Soil 142: 273-279.
39      Brunei, J.; Diekmann, M.; Falkengren-Grerup, U. (1998) Effects of nitrogen deposition on field layer vegetation in
40 •           south Swedish oak forests. Environ. Pollut. (Oxford, U.K.) 102(suppl. 1): 35-40.
41       Bugini, R.; Tabasso, M. L.; Realini, M. (2000) Rate of formation of black crusts on marble.  A case study. J. Cult.
42            Heritage 1:111-116.
43      Butlin, R. N.; Coote, A. T.; Devenish, M.; Hughes, I. S. C.; Hutchens, C. M.; Irwin, J. G.; Lloyd, G. O.; Massey,
44            S. W.; Webb, A. H.; Yates, T. J. S. (1992a) Preliminary results from the analysis of metal samples from the
45            National Materials Exposure Programme (NMEP). Atmos. Environ. Part B 26: 199-206.
46      Butlin, R. N.; Coote, A. T.; Devenish, M.; Hughes, I. S. C.; Hutchens, C. M.; Irwin, J. G.; Lloyd, G. O.; Massey,
47            s. W.; Webb, A. H.; Yates, T. J. S. (1992b) Preliminary results from the analysis of stone tablets from the
48            National Materials Exposure Programme (NMEP). Atmos. Environ. Part B 26:189-198.
49      Bytnerowicz, A.; Fenn, M. E. (1996) Nitrogen deposition in California forests: a review. Environ. Pollut.
50            92:127-146.
51      Cahill, T. A.; Eldred, R. A.; Wakabayashi, P. H. (1996) Trends in fine particle concentrations at Great Smoky
52            Mountains National Park. For presentation at: Annual meeting of the Air & Waste  Management Association;
53            June; Nashville, TN. Pittsburgh, PA: Air & Waste Management Association; submitted paper no.
54            96-MP1A.05.
          March 2001
4-166
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
 Campbell, G. G.; Schurr, G. G.; Slawikowski, D. E.; Spence, J. W. (1974) Assessing air pollution damage to
       coatings. J. Paint Technol. 46: 59-71.
 Camuffo, D. (1995) Physical weathering of stones. In: Saiz-Jimenez, C., ed. The deterioration of monuments:
       proceedings of the 2nd international symposium on biodeterioration and biodegradation; January 1994;
       Sevilla, Spain. Sci. Total Environ. 167: 1-14.
 Cape, J. N. (1993) Direct damage to vegetation caused by acid rain and polluted cloud: definition of critical levels
       for forest trees. Environ. Pollut. 82: 167-180.
 Cape, J. N.; Sheppard, L. J.; Fowler, D.; Harrison, A. F.; Parkinson, J. A.; Dao, P.; Paterson, I. S. (1992)
       Contribution of canopy leaching to sulphate deposition in a Scots pine forest. Environ. Pollut. 75: 229-236.
 Carey, W. F. (1959) Atmospheric deposits in Britain: a study of dinginess. Int. J. Air Pollut. 2: 1-26.
 Ceccotti, S. P.; Messick, D. L. (1997) A global review of crop requirements, supply, and environmental impact on
       nutrient sulphur balance. In: Cram, W. J.; De Kok, L. J.; Stulen, I.; Brunold, C.; Rennenberg, H., eds.
       Sulphur metabolism in higher plants. Leiden, The Netherlands: Backhuys Publishers; pp. 155-163.
 Chameides, W. L.; Yu, H.; Liu, S. C.; Bergin, M.; Zhou, X.; Meams, L.; Wang, G.; Kiang, C. S.; Saylor, R. D.;
       Luo, C.; Huang,  Y.; Steiner, A.; Giorgi, F. (1999) Case study of the effects of atmospheric aerosols and
       regional haze on agriculture:  an opportunity to enhance crop yields in China through emission controls?
       Proc. Natl. Acad. Sci. U. S. A. 96: 13626-13633.
 Chapin, F. S., III. (1991) Integrated responses of plants to stress: a centralized system of physiological responses.
       Bioscience 41: 29-36.
 Chapin, F. S., Ill; Bloom,.A. J.; Field, C. B.; Waring, R. H. (1987) Plant responses to multiple environmental
       factors. BioScience 37: 49-57.
 Chapin, F. S., Ill; Sala,  O. E.; Burke, I. C.; Grime, J. P.; Hooper, D. U.; Lauenroth, W. K.; Lombard, A.; Mooney,
       H. A.; Mosier, A. R.; Naeem, S.; Pacala, S. W.; Roy, J.; Steffen, W. L.; Tilman, D. (1998) Ecosystem
       consequences of changing biodiversity. BioScience 48:45-52.
 Charlson, R. J.; Lovelock, J. E.; Andreae, M. O.; Warren, S. G. (1987) Oceanic phytoplankton, atmospheric
       sulfphur, cloud albedo and climate. Nature (London) 326: 655-661.
 Charlson, R. J.; Schwartz, S. E.; Hales, J. M.; Cess, R. D.; Coakley, J. A., Jr.; Hansen, J. E.; Hofmann, D. J. (1992)
      Climate forcing by anthropogenic aerosols. Science (Washington, DC) 255: 423-430.
 Chestnut, L. (1997) Methodology for estimating values for changes in visibility at national parks [draft
      memorandum]. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
      Planning and Standards; contract no. 68-D-98-001; EPA air docket A-96-56, document no. VI-B-09-(ooo).
 Chestnut, L. G.; Dennis, R. L. (1997) Economic benefits of improvements in visibility: acid rain. Provisions of the
       1990 clean air act amendments. J. Air Waste Manage. Assoc. 47: 395-402.
 Chevone, B. I.; Herzfeld, D. E.; Krupa, S. V.; Chappelka, A. H. (1986) Direct effects of atmospheric sulfate
      deposition on vegetation. J. Air Pollut. Control Assoc. 36: 813-815.
 Choudhury, B. J. (1987) Relationships between vegetation indices, radiation absorption, and net photosynthesis
      evaluated by a sensitivity analysis. Remote Sens. Environ. 22: 209-233.
 Clarkson, D. T.; Sanderson, J. (1971) Inhibition of the uptake and long-distance transport of calcium by aluminum
      and other polyvalent cations. J. Exp. Bot. 22: 837-851.
 Clevenger, T. E.; Saiwan, C.;  Koirtyohann, S. R. (1991) Lead speciation of particles on air filters collected in the
      vicinity of a lead smelter. Environ. Sci. Technol. 25: 1128-1133.
 Cole, D. W.; Johnson, D. W. (1977) Atmospheric sulfate additions and cation leaching in a Douglas fir ecosystem.
      Water Resour. Res. 13: 313-317.
 Cooke, R. U.; Gibbs, G. B. (1994) Crumbling heritage? Studies of stone weathering in polluted atmospheres.
      Atmos. Environ. 28: 1355-1356.
Costanza, R.; d'Arge, R.; De Groot, R.; Farber, S.; Grasso, M.; Hannon, B.; Limburg, K.; Naeem, S.; O'Neill, R. V.;
      Paruelo, J.; Raskin, R. G.; Sutton, P.; Van Den Belt, M. (1997) The value of the world's ecosystem services
      and natural capital. Nature (London) 387: 253-259.
Cotrufo, M. F.; De Santo, A. V.; Alfani, A.; Bartoli, G.; De Cristofaro, A. (1995) Effects of urban heavy metal
      pollution on organic matter decomposition in Quercus ilex L. woods. Environ. Pollut. 89: 81-87.
Cowling, J. E.; Roberts, M. E. (1954) Paints, varnishes, enamels, and lacquers. In: Greathouse, G. A.; Wessel, C. J.,
      eds. Deterioration of materials: causes and preventive techniques. New York, NY: Reinhold Publishing
      Corporation; pp. 596-645.
         March 2001
                                                4-167
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Cramer, S. D.; McDonald, L. G.; Bhagia, G.; Flinn, D. R.; Linstrom, P. J.; Carter, J. P. (1989) Effects of acid
 2             deposition on the atmospheric corrosion of structural metals. Research Triangle Park, NC: U.S.
 3             Environmental Protection Agency; internal report.
 4       Creighton, P. J.; Lioy, P. J.; Haynie, F. H.; Lemmons, T. J.; Miller, J. L.; Gerhart, J. (1990) Soiling by atmospheric
 5             aerosols in an urban industrial area. J. Air Waste Manage. Assoc. 40: 1285-1289.
 6       Cronan, C. S.; Grigal, D. F. (1995) Use of calcium/aluminum ratios as indicators of stress in forest ecosystems.
 7             J. Environ. Qual. 24:209-226.
 8       Cullen, J. J.; Neale, P. J. (1997) Biological weighting functions for describing the effects of ultraviolet radiation on
 9             aquatic ecosystems. In: Hader, D.-P., ed. The effects of ozone depletion on aquatic ecosystems. Austin, TX:
10             Academic Press, pp. 97-118.
11       Dahlgren, R. A. (1994) Soil acidification and nitrogen saturation from weathering of ammonium-bearing rock.
12             Nature 368: 838-841.
13       Daily, G. C. (1997) Introduction: what are ecosystem services?  In: Daily, G. C., ed. Nature's services: societal
14             dependence on natural ecosystems. Washington, DC: Island Press; pp. 1-10.
15       Daily, G. C.; Ehrlich, P. R. (1999) Managing Earth's ecosystems: an interdisciplinary challenge. Ecosystems
16             2:277-280.
17       Darley, E. F. (1966) Studies on the effect of cement-kiln dust on vegetation. J. Air Pollut. Control Assoc.
18             16:145-150.
19       Dassler, H.-G.; Ranft, H.; Rehn, K.-H. (1972) Zur Widerstandsfahigkeit von Geholzen gegeniiber
20             Fluorverbindungen und Schwefeldioxid [The susceptibility of woody plants exposed to fluorine compounds
21             and SOJ. Flora (Jena) 161:289-302.
22       Dattner, S. L. (1995) Urban visibility in the El Paso del Norte airshed. Presented at: 88th annual meeting &
23             exhibition of the Air & Waste Management Association, volume 3B. Basic sciences: air quality and noise;
24             June; San Antonio, TX. Pittsburgh,  PA: Air & Waste Management Association; paper no. 95-WA76.02.
25       Davidson, C. L; Tang, W.; Finger, S.; Etyemezian, V.; Striegel, M. F.; Sherwood, S. I. (2000) Soiling patterns on a
26             tall limestone building: changes over 60 years. Environ.  Sci. Technol. 34: 560-565.
27       Davis, G. D.; Shaw, B. A.; Arah, C. O.; Fritz, T. L.; Moshier, W. C.; Simpson, T. C.; Moran, P. J.; Zankel, K. L.
28             (1990) Effects of SO2 deposition on painted steel surfaces. Surf. Interface Anal. 15: 107-112.
29       Day, D.; Malm, W. C.; Kreidenweis, S. M. (1996) Aerosol light scattering measurements as a function of relative
30             humidity at Great Smoky Mountains National Park. In: 89th annual meeting & exhibition of the Air & Waste
31             Management Association; June; Nashville, TN. Pittsburgh, PA: Air & Waste Management Association;
32             paper no. 96-TA1B.06.
33       Day, D. E.; Malm, W. C.; Kreidenweis, S. M. (2000) Aerosol light scattering measurements as a function of relative
34             humidity. J. Air Waste Manage. Assoc. 50: 710-716.
35       Del Monte, M.; Vittori, O. (1985) Air pollution and stone decay: the case of Venice. Endeavour 9: 117-122.
36       Diakumaku, E.; Gorbushina, A. A.; Krumbein, W. E.; Panina, L.; Soukharjevski, S. (1995) Black fungi in marble
37             and limestones-an aesthetical, chemical and physical problem for the conservation of monuments.
38             In: Saiz-Jimenez, C., ed. The deterioration of monuments: proceedings of the 2nd international symposium
39             on biodeterioration and biodegradation; January 1994; Sevilla, Spain. Sci. Total Environ. 167: 295-304.
40       Dickerson, R. R.; Kondragunta, S.; Stenchikov, G.; Civerolo, K. L.; Doddridge, B. G.; Holben, B. N. (1997)
41             The impact of aerosols on solar ultraviolet radiation and photochemical smog. Science (Washington, DC)
42             278:827-830.
43       Dobbins, R. A.; Mulholland, G. W.; Bryner, N. P. (1994) Comparison of a fractal smoke optics model with light
44             extinction measurements. Atmos. Environ. 28: 889-897.
45       Dolske, D. A. (1995) Deposition of atmospheric pollutants to monuments, statues, and buildings.  Sci. Total
46             Environ. 167: 15-31.
47       Dreisinger, B. R.; McGovern, P. C. (1970) Monitoring atmospheric sulphur dioxide and correlating its effects on
48             crops and forests in the Sudbury area. In: Linzon, S. N.,  ed. Impact of air pollution on vegetation conference;
49             April; Toronto, ON, Canada. Pittsburgh, PA: Air Pollution Control Association.
50       Driscoll, C. T.; Van Dreason, R. (1993) Seasonal and long-term temporal patterns in the chemistry of Adirondack
51             lakes. Water Air Soil Pollut. 67: 319-344.
52       Driscoll, C. T.; Wyskowski, B. J.; DeStaffan, P.; Newton, R. M. (1989) Chemistry and transfer of aluminum in a
53             forested watershed in the Adirondack region of New York, USA. In: Lewis,  T. E., ed. Environmental
54             chemistry and toxicology of aluminum. Chelsea, MI: Lewis Publishers, Inc.; pp. 83-105.
         March 2001
4-168
DRAFT-DO NOT QUOTE OR CITE

-------
  1       Driscoll, C. T.; Newton, R. M.; Gubala, C. P.; Baker, J. P.; Christensen, S. W. (1991) Adirondack mountains.
  2             In: Charles, D. F., ed. Acidic deposition and aquatic ecosystems: regional case studies. New York, NY:
  3             Springer-Verlag; pp. 133-202.
  4       Edney, E. O. (1989) Paint coatings: controlled field and chamber experiments. Research Triangle Park, NC:
  5             U.S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment Laboratory; report
  6             no. EPA/600/3-89/032. Available from: NTIS, Springfield, VA; PB89-189849.
  7       Edney, E. O.; Stiles, D. C.; Corse, E. W.; Wheeler, M. L.; Spence, J. W.; Haynie, F. H.; Wilson, W. E., Jr. (1988)
  8             Field study to determine the impact of air pollutants on the corrosion of galvanized steel. Mater. Perform.
  9             27:47-50.
 10       Edney, E. O.; Cheek, S. F.; Corse, E. W.; Spence, J. W.; Haynie, F.  H. (1989) Atmospheric weathering caused by
 11             dry deposition of acidic compounds. J. Environ. Sci. Health Part A 24: 439-457.
 12       Eldering, A.; Cass, G. R. (1996) Source-oriented model for air pollutant  effects on visibility. J. Geophys. Res.
 13             [Atmos.]: 101:19,343-19,369.
 14       Eldering, A.; Larson, S. M.; Hall, J. R.; Hussey, K. J.; Cass, G. R. (1993) Development of an improved image
 15             processing based visibility model. Environ. Sci. Technol. 27:  626-635.
 16       Eldering, A.; Cass, G. R.; Moon, K. C. (1994) An air monitoring network using continuous particle size distribution
 17             monitors: connecting pollutant properties to visibility via Mie scattering calculations. Atmos. Environ.
 18             28:2733-2749.
 19       Eldering, A.; Hall, J. R.; Hussey, K. J.; Cass, G. R. (1996) Visibility model based on satellite-generated landscape
 20             data. Environ. Sci. Tech. 30: 361 -370.
 21       Eldred, R. A.; Cahill, T. A. (1994) Trends in elemental concentrations of fine particles at remote sites in the
 22             United States of America. Atmos. Environ. 28: 1009-1019.
 23       Eldred, R. A.; Cahill, T. A.; Feeney, P. J. (1993) Comparison of independent measurements of sulfur and sulfate in
 24             the IMPROVE network. Presented at: 86th annual meeting and exhibition of the Air & Waste Management
 25             Association; June; Denver, CO. Pittsburgh, PA: Air & Waste Management Association; paper no.
 26             93-RA-110.02.
 27       Eldred, R. A.; Cahill, T. A.; Flocchini, R. G. (1997) Composition of PM25 and PM10 aerosols in the IMPROVE
 28             network. In: Special issue on aerosols and atmospheric optics-part I, [published papers from the international
 29             specialty conference; September  1994; Snowbird, UT]. J. Air Waste Manage. Assoc. 47: 194-203.
 30       Ellenberg, H. (1987) Floristic changes due to eutrophication. In: Asman, W. A. H.; Diederen, S. M. A., eds.
 31             Ammonia and acidification: proceedings of a symposium of the European Association for the Science of Air
 32             Pollution (EURASAP); April; Bilthoven, The Netherlands. European Association for the Science of Air
 33             Pollution; pp. 301-308.
 34       Eller, B. M. (1977) Road dust induced increase of leaf temperature. Environ. Pollut. 13: 99-107.
 35       Eriksson, P.; Johansson, L.-G.; Strandberg, H. (1993) Initial stages of copper corrosion in humid air containing SO2
 36             and NO2. J. Electrochem. Soc. 140: 53-59.
 37       Erlick, C.; Frederick, J. E.; Saxena, V. K.; Wenny, B. N. (1998) Atmospheric transmission in the ultraviolet and
 38            visible: aerosols in cloudy atmospheres. J. Geophys. Res. 103: 31,541-31,556.
 39      Estupinan, J. B.; Raman, S.; Crescenti, G. H.; Streicher, J. J.; Barnard, W. F. (1996) Effects of clouds and haze on
40            UV-B radiation. J. Geophys. Res. 101: 16,807-16,816.
41      Evans, E. G.; Pitchford, M. (1991) Federal visibility monitoring in the eastern United States. Presented at:
42             84th annual meeting and exhibition of the Air & Waste Management Association; June; Vancouver, BC,
43            Canada. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-49.6.
44      Falkengren-Grerup, U. (1987) Long-term changes in pH of forest soils in southern Sweden. Environ. Pollut.
45            43:79-90.
46      Falkengren-Grerup, U. (1998) Nitrogen response of herbs and graminoids in experiments with simulated acid soil
47            solution. Environ. Pollut. 102(suppl. 1): 93-99.
48      Farber, R. J.; Welsing, P. R.; Rozzi, C. (1994) PM10 and ozone control strategy to improve visibility in the
49            Los Angeles Basin. Atmos. Environ. 28: 3277-3283.
50      Farmer, A. M. (1993) The effects of dust on vegetation-a review. Environ. Pollut. 79: 63-75.
51       Fenn, M. E.; Poth, M. A.; Johnson, D. W. (1996) Evidence for nitrogen saturation in the San Bernardino Mountains
52            in southern California. For. Ecol. Manage. 82: 211-230
53       Fenn, M. E.; Poth, M. A.; Aber, J. D.; Baron, J. S.; Bormann, B. T.; Johnson, D. W.; Lemly, A. D.; McNulty, S. G.;
54            Ryan, D. F.; Stottlemyer, R. (1998) Nitrogen excess in North American ecosystems: predisposing factors,
55            ecosystem responses, and management strategies. Ecol. Appl.  8: 706-733.
         March 2001
4-169
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Fink, F. W.; Buttner, F. H.; Boyd, W. K. (1971) Final report on technical-economic evaluation of air-pollution
 2             corrosion costs on metals in the U.S. Columbus, OH: Battelle Memorial Institute, Columbus Laboratories;
 3             EPA contract no. CPA 70-86; EPA report no. APTD-0654. Available from: NTIS, Springfield, VA;
 4             PB-198-453.
 5       Flowers, E. C.; McCormick, R. A.; Kurfis, K. R. (1969) Atmospheric turbidity over the United States, 1961-1966.
 6             J. Appl. Meteorol. 8: 955-962.
 7       Fochtman, E. G.; Langer, G. (1957) Automobile paint damaged by airborne iron particles. J. Air Pollut. Control
 8             Assoc. 6: 243-247.
 9       Foster, J. R. (1991) Effects of organic chemicals in the atmosphere on terrestrial plants. In: Moser, T. J.; Barker,
10             J. R.; Tingey, D. T., eds. Ecological exposure and effects of airborne toxic chemicals: an overview. Corvallis,
11             OR: U.S. Environmental Protection Agency, Environmental Research Laboratory; pp. 60-89; EPA report no.
12             EPA-600/3-91-001. Available from: NTIS, Springfield, VA; PB91-148460.
13       Foster, N. W.; Nicolson, J. A.; Hazlett, P. W. (1989) Temporal variation in nitrate and nutrient cations in drainage
14             waters from a deciduous forest. J. Environ. Qual. 18: 238-244.
15       Fox, D.; McDonald, K.; Zannetti, P.; Nejedley, Z. (1997) Impact of north-western emission changes on visibility in
16             the Rocky Mountain parks. Presented at: Air & Waste Management Association's 90th annual meeting &
17             exhibition; June; Toronto, Ontario, Canada. Pittsburgh, PA: Air & Waste Management Association; paper
18             no. 97-WA70.03.
19       Fox, D. G.; Malm, W. C.; Mitchell, B.; Fisher, R. W. (1999) Where there's fire, there's smoke: fine particulate and
20             regional haze. EM November: 15-24.
21       Frederick, J. E.; Qu, Z.; Booth, C. R. (1998) Ultraviolet radiation at sites on the Antarctic coast. J. Photochem.
22             Photobiol. 668:183-190.
23       Frcedman, B.; Hutchinson, T. C. (1981)  Sources of metal and elemental contamination of terrestrial environments.
24             In: Lepp, N. W., ed. Effect of heavy metal pollution on plants. Volume 2: metals in the environment. London,
25             United Kingdom: Applied Science Publishers; pp. 35-94.
26       Freeman, A. M., III. (1993) The measurement of environmental and resource values: theory and methods.
27             Washington, DC: Resources for the Future; pp. 1-27.
28       Freudenburg, W. R.; Alario, M. (1999) What ecologists can learn from nuclear scientists. Ecosystems 2: 286-291.
29       Fuller, K. A.; Malm, W. C.; Kreidenweis, S. M. (1999) Effects of mixing on extinction by carbonaceous particles.
30             J. Geophys. Res. 104:15,941-15,954.
31       Furst, P.; Krause, G. H. M.; Hein, D.;  Delschen, T.; Wilmers, K. (1993) PCDD/PCDF in cow's milk in relation to
32             their levels in grass and soil. Chemosphere 27: 1349-1357.
33       Gabruk, R. S.; Sykes, R. I.; Seigneur, C.; Pai, P.; Gillespie, P.; Bergstrom, R. W.; Saxena, P. (1999) Evaluation of
34             the  reactive and optics model of emissions (ROME). Atmos. Environ. 33: 383-399.
35       Gaggi, C.; Bacci, E.; Calamari, D.; Fanelli, R. (1985) Chlorinated hydrocarbons in plant foliage: an indication of the
36             tropospheric contamination level. Chemosphere 14:1673-1686.
37       Garcia-Valles, M.; Vendrell-Saz, M.; Molera, J.; Blazquez, F. (1998) Interaction of rock and atmosphere: patinas on
38             Mediterranean monuments. Environ. Geol. 36(1-2): 137-149.
39       Garner, J.  H. B. (1991) Ozone exposure and nitrogen loading in the forests of eastern North America. In: Berglund,
40             R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment: papers from an
41             international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
42            Association; pp. 289-310. (A&WMA transactions series no. TR-19).
43       Garner, J.  H. B. (1994) Nitrogen oxides, plant metabolism, and forest ecosystem response. In: Alscher, R. G.;
44            Wellburn, A. R., eds. Plant responses to the gaseous environment: molecular, metabolic and physiological
45             aspects, [3rd international symposium on air pollutants and plant metabolism]; June 1992; Blacksburg, VA.
46            London, United Kingdom: Chapman & Hall; pp. 301 -314.
47      Gamer, J.  H. B.; Pagano, T.; Cowling, E. B. (1989) An evaluation of the role of ozone, acid deposition, and other
48             airborne pollutants in the forests of eastern North America. Asheville, NC: U.S. Department of Agriculture,
49            Forest Service, Southeastern Forest Experiment Station; general technical report SE-59.
50      Garten, C. T., Jr. (1988) Fate and distribution of sulfur-35  in yellow poplar and red maple trees. Oecologia
51             76:43-50.
52      Garten, C. T., Jr.; Hanson, P. J. (1990) Foliar retention of 15N-nitrate and 15N-ammonium by red maple
53             (Acer rubrum) and white oak (Quercus alba) leaves from simulated rain. Environ. Exp. Bot. 30: 333-342.
54      Gawel, J. E.; Ahner, B. A.; Friedland, A. J.; Morel, F. M. M. (1996) Role for heavy metals in forest decline
55             indicated by phytochelatin measurements. Nature 381: 64-65.
         March 2001
4-170
DRAFT-DO NOT QUOTE OR CITE

-------
  1      Ghedini, N.; Gobbi, G.; Sabbioni, C.; Zappia, G. (2000) Determination of elemental and organic carbon on
  2            damaged stone monuments. Atmos. Environ. 34: 4383-4391.
  3 .      Gildon, A.; Tinker, P. B. (1983) Interactions of vesicular-arbuscular mycorrhizal infection and heavy metals in
  4            plants: I. the effects of heavy metals on the development of vesicular-arbuscular mycorrhizas. New Phytol.
  5            95:247-261.
  6      Gilliam, F. S.; Adams, M. B.; Yurish, B. M. (1996) Ecosystem nutrient responses to chronic nitrogen inputs at
  7            Fernow Experimental Forest, West Virginia. Can. J. For. Res. 26: 196-205.
  8      Gobbi, G.; Zappia, G.; Sabbioni, C. (1998) Sulphite quantification on damaged stones and mortars. Atmos. Environ.
  9            32:783-789.
10      Goodwin, J. E.; Sage, W.; Tilly, G. P. (1969) Study of erosion by solid particles. Proc. Inst. Mech. Eng.
11            184:279-292.
12      Graedel, T. E.; Nassau, K.; Franey, J. P. (1987) Copper patinas formed in the atmosphere—I. introduction.
13            Corros. Sci. 27: 639-657.
14      Grand Canyon Visibility Transport Commission. (1996) Report of the Grand Canyon Visibility Transport
15            Commission to the United States Environmental Protection Agency.
16      Grant, L. D. (1989) Health effects issues associated with regional and global air pollution  problems. In: Proceedings
17            of the changing atmosphere: implications for global security; June, 1988; Toronto, Canada. Geneva,
18            Switzerland: World Meteorological Organization; report no. WMO-No. 710.
19      Gray, H. A.; Cass, G. R. (1998) Source contributions to atmospheric fine carbon particle concentrations.
20            Atmos. Environ. 32: 3805-3825.
21      Greenberg, B. M.; Wilson, M. I.; Huang, X.-D.; Duxbury, C. L.; Gerhardt, K. E.; Gensemer, R. W. (1997)
22            The effects of ultraviolet-B radiation on higher plants. In: Wang, W.; Gorsuch, J. W.; Hughes, J. S., eds.
23            Plants for Environmental Studies. New York, NY: Lewis Publishers; pp. 1-35.
24      Guderian, R. (1977) Accumulation of pollutants in plant organs. In: Air pollution: phytotoxicity of acidic gases and
25            its significance in air pollution control. Berlin, Germany: Springer-Verlag; pp. 66-74. (Ecological studies:
26            v. 22).
27      Guderian, R., ed. (1985) Air pollution by photochemical oxidants: formation, transport, control, and effects on
28            plants. New York, NY: Springer-Verlag. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.;
29            Remmert, H., eds. Ecological studies: analysis and synthesis: v. 52).
30      Guderian, R. (1986) Terrestrial ecosystems: particulate deposition.  In: Legge, A. H.; Krupa, S. V., eds.
31            Air pollutants and their effects on the terrestrial ecosystem. New York, NY: John Wiley & Sons; pp.
32            339-363. (Advances in environmental science and technology: v. 18).
33      Gundersen, P.; Callesen, I.; De Vries, W. (1998a) Nitrate leaching in forest ecosystems is related to forest floor C/N
34            ratios. Environ. Pollut. 102(suppl. 1): 403-407.
35      Gundersen, P.; Emmett, B. A.; Kjonaas, O. J.; Koopmans, C. J.; Tietema, A. (1998b) Impact of nitrogen deposition
36            on nitrogen cycling in forests: a synthesis of NITREX data. For. Ecol. Manage. 101: 37-55.
37      Hamilton, R. S.; Revitt, D. M.; Vincent, K. J.; Butlin, R. N. (1995) Sulphur and nitrogen particulate pollutant
38            deposition on to building surfaces. In: Saiz-Jimenez, C., ed. The deterioration of monuments: proceedings of
39            the 2nd international symposium on biodeterioration and biodegradation; January 1994; Sevilla, Spain. Sci.
40            Total Environ. 167: 57-66.
41      Han, Q.; Rossow, W. B.; Lacis, A. A. (1994) Near-global survey of effective droplet radii  in liquid water clouds
42            using ISCCP data. J. Clim. 7: 465-497.
43      Haneef, S. J.; Johnson, J. B.; Thompson, G. E.; Wood, G. C. (1993) Simulation of the degradation of limestones
44            and dolomitic sandstone under dry deposition conditions. In: Corrosion control for low-cost reliability:
45            preceedings [of the] 12th international corrosion congress. Volume 2: process industries plant operation.
46            Houston, TX: National Association of Corrosion Engineers; pp. 700-710.
47      Hanley, N.; Spash, C. L. (1993) Cost-benefit analysis and the environment. Hants, United Kingdom: Edward Elgar.
48      Harwell, M. A.; Myers, V.; Young, T.; Bartuska, A.; Gassman, N.;  Gentile, J. H.; Harwell, C. C.; Appelbaum, S.;
49            Barko, J.; Causey, B.; Johnson, C.; McLean, A.; Smola, R.; Templet, P.; Tosini, S. (1999) A framework for
50            an ecosystem integrity report card. BioScience 49: 543-556.
51      Hauhs, M. (1989) Lange Bramke: an ecosystem study of a forested catchment. In: Adriano, D. C.; Havas, M., eds.
52            Acidic precipitation: volume 1, case  studies. New York, NY: Springer-Verlag; pp. 275-305. (Adriano, D. C.;
53            Salomons, W., eds. Advances in environmental science).
         March 2001
4-171
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Haynie, F. H. (1976) Air pollution effects on stress induced intergranular corrosion of 7005-T53 aluminum alloy.
 2             In: Stress corrosion-new approaches. Philadelphia, PA: American Society for Testing and Materials;
 3             pp. 32-43. (ASTM special technical publication 610).
 4       Haynie, F. H. (1980) Theoretical air pollution and climate effects on materials confirmed by zinc corrosion data.
 5             In: Sereda, P. J.; Litvan, G. G., eds. Durability of building materials and components: proceedings of the
 6             1 st international conference; August 1978; Ottawa, ON, Canada. Philadelphia, PA: American Society for
 7             Testing and Materials; pp. 157-175. (ASTM special technical publication 691).
 8       Haynie, F. H. (1986) Environmental factors affecting corrosion of weathering steel. In: Baboian, R., ed. Materials
 9             degradation caused by acid rain: developed from the 20th state-of-the-art symposium of the American
10             Chemical Society; June 1985; Arlington, VA. Washington, DC: American Chemical Society; pp. 163-171.
11             (ACS symposium series 318).
12       Haynie, F. H.; Lemmons, T. J. (1990) Particulate matter soiling of exterior paints at a rural site. Aerosol Sci.
13             Technol. 13: 356-367.
14       Haynie, F. H.; Upham, J. B. (1974) Correlation between corrosion behavior of steel and atmospheric pollution data.
15             In: Cobum, S. K., ed. Corrosion in natural environments: presented at the 76th annual meeting American
16             Society for Testing and Materials; June 1973; Philadelphia, PA. Philadelphia, PA: American Society for
17             Testing and Materials; pp. 33-51. (ASTM special technical publication 558).
18       Haywood, J. M.; Shine, K. P. (1995) The effect of anthropogenic sulfate and  soot aerosol on the clear sky planetary
19             radiation budget. Geophys. Res. Lett. 22: 603-606.
20       Haywood, J. M.; Ramaswamy, V.; Soden, B. J. (1999) Tropospheric aerosol climate forcing in clear-sky satellite
21             observations over the oceans. Science (Washington, DC) 283: 1299-1303.
22       Heal, G. (2000) Valuing ecosystem services. Ecosystems 3: 24-30.
23       Hedin, L. O.; Granat, L.; Likens, G. E.; Buishand, T. A.; Galloway, J. N.; Butler, T. J.; Rodhe, H. (1994)
24             Steep declines in atmospheric base cations in regions of Europe and North America. Nature (London)
25             367:351-354.
26       Heinsdorf, D. (1993) The role of nitrogen in declining Scots pine forests (Pinus sylvestris) in the lowland of
27             East Germany. Water Air Soil Pollut. 69: 21-35.
28       Herbert, D. A.; Rastetter, E. B.; Shaver, G. R.; Agren, G. I. (1999) Effects of plant growth characteristics on
29             biogeochemistry and community composition in a changing climate. Ecosystems 2: 367-382.
30       Hicks, B. B. (1986) Differences in wet and dry particle deposition parameters between North America and Europe.
31             In: Lee, S. D.; Schneider, T.; Grant, L. D.; Verkerk, P. J., eds. Aerosols: research, risk assessment and
32             control strategies; proceedings of the second U.S.-Dutch international  symposium; May 1985; Williamsburg,
33             VA. Chelsea, MI: Lewis Publishers; pp. 973-982.
34       Hicks, B. B.; Baldocchi, D. D.; Meyers, T. P.; Hosker, R. P., Jr.; Matt, D. R. (1987) A preliminary multiple
35             resistance routine for deriving dry deposition velocities from measured quantities. Water Air Soil Pollut.
36             36:311-330.
37       Hinckley, T.; Ford, D.; Segura, G.; Sprugel, D. (1992) Key processes from tree to stand level. In: Wall, G., ed.
38             Implications of climate change for Pacific Northwest forest management. Waterloo, ON, Canada: University
39             of Waterloo, Department of Geography; pp. 33-43. (Department of Geography publication series: occasional
40             paper no. 15).
41       Hogan, G. D.; Rennenberg, H.; Fink, S. (1998) Role and effect of sulfur in tree biology. In: Maynard, D. G., ed.
42             Sulfur in the environment. New York, NY: Marcel Dekker, Inc.; pp. 173-217. (Books in soils, plants, and the
43             environment: v. 67).
44       Holbrow, G. L. (1962) Atmospheric pollution: its measurement and some effects on paint. J. Oil Colour Chem.
45             Assoc. 45:701-718.
46       Holling, C. S. (1986) The resilience of terrestrial ecosystems: local surprise and global change. In: Clark, W. C.;
47             Munn, R. E., eds. Sustainable development of the biosphere. Cambridge, UK: Cambridge University;
48             pp. 292-317.
49       Hooper, D. U.; Vitousek, P. M. (1997) The effects of plant composition and diversity on ecosystem processes.
50             Science 277:1302-1305.
51       Hornung, M.; Langan, S. J. (1999) Nitrogen deposition, sources, impacts and responses in natural and semi-natural
52             ecosystems. In: Bangan, S. J., ed. Impact of nitrogen deposition on natural ecosystems and semi-natural
53             ecosystems. Dordrect, Netherlands: Kluwer Academic Publishers; pp.  1-14. [Environmental Pollution,  no.3].
54       Horvath, H. (1993) Atmospheric light absorption~a review. Atmos.  Environ. Part A 27: 293-317.
         March 2001
4-172
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
45
46
47
48
49
 50
 51
 52
53
54
 Houghton, J. T.; Jenkins, G. J.; Ephraums, J. J., eds. (1990) Climate change: the IPCC scientific assessment.
       Cambridge, MA: Cambridge University Press; p. 55.
 Hoyt, D. V. (1978) Amodel for the calculation of solar global insolation. Sol. Energy 21: 27-35.
 Hughes, M. K. (1981) Cycling of trace metals in ecosystems. In: Lepp, N. W., ed. Effect of heavy metal pollution
       on plants. Volume 2: metals in the environment. London, United Kingdom: Applied Science Publishers-
       pp. 95-118.
 Hughes, J. B.; Daily, G. C.; Ehrlich, P. R. (1997) The loss of population diversity and why it matters. In: Raven,
       P. H., ed. Nature and Human Society: the quest for a sustainable world. Washington, DC: National Academy
       Press; pp. 71-83.
 Humphreys, F. R.; Lambert, M. J.; Kelly, J. (1975) The occurrence of sulphur deficiency in forests. In: McLachlan,
       K. D., ed. Sulphur in Australasian agriculture. Sydney, Australia: Sydney University Press; pp.  154-162.
 Huntington, T.G. (2000) The potential for calcium depletion in forest ecosystems of southeastern United States:
       review and analysis. Global Biogeochem. Cycles 14: 623-638.
 Husar, R. B.; Wilson, W. E. (1993) Haze and sulfur emission trends in the eastern United States Environ Sci
       Technol. 27: 12-16.
 Husar, R. B.; Holloway, J. M.; Patterson, D. E.; Wilson, W. E. (1981) Spatial and temporal pattern of eastern U.S.
       haziness: a summary. Atmos. Environ. 15: 1919-1928.
 Husar, R. B.; Elkins,  J. B.; Wilson, W. E. (1994) U.S. visibility trends,  1960-1992, regional and national. Presented
       at: 87th annual meeting of the Air & Waste Management Association; June; Cincinnati, OH. Pittsburgh, PA:
      Air & Waste Management Association; paper no. 94-MP3.05.
 Husar, R. B.; Schichtel, B. A.; Falke, S, R.; Wilson, W. E. (2000) Haze trends over the United States, 1980-1995.
      J. Geophys. Res. Lett.: In Press.
 Hutchinson, A. J.; Johnson, J. B.; Thompson, G. E.; Wood, G. C.; Sage, P. W.; Cooke, M. J. (1992) The role of
      fly-ash particulate material and oxide catalysts in stone degradation. Atmos. Environ. 26A: 2795-2803.
 Huttunen, S. (1994) Effects of air pollutants on epicuticular wax structure. In: Percy, K. E.; Cape, J. N.; Jagels, R.;
      Simpson, C. J., eds. Air pollutants and the leaf cuticle. New York, NY: Springer-Verlag; pp 81-96
      (NATO ASI series).
 Intergovernmental Panel on Climate Change (IPCC). (1995) Climate change 1994: radiative forcing of climate
      change and an  evaluation of the IPCC IS92 emission scenarios. Cambridge, United Kingdom: University
      Press.
 Intergovernmental Panel on Climate Change (IPCC). (1996) Climate change 1996: contribution of working group
      I to the second assessment of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom:
      Cambridge University Press; p. 572. [IPCC second assessment report].
 Intergovernmental Panel on Climate Change (IPCC). (1998) The regional impacts of climate change: an assessment
      of vulnerability. Cambridge, United Kingdom: Cambridge University Press.
 Irving, P. M., ed. (1991) Acidic deposition: state of science and technology: volumes I-IV. Washington, DC:
      The U.S. National Acid Precipitation Assessment Program.
 Iyer, H.; Patterson, P.; Malm, W. C. (2000) Trends in the extremes of sulfur concentration distributions. J. Air
      Waste Manage. Assoc. 50: 802-808.
 Jacobson, M. Z. (1998) Studying the effects of aerosols on vertical photolysis rate coefficient and temperature
      profiles  over an urban airshed. J. Geophys. Res. [Atmos.] 103: 10,593-10,604.
 Jacobson, M. Z. (1999) Isolating nitrated and aromatic aerosols and nitrated aromatic gases as sources of ultraviolet
      light absorption. J. Geophys. Res. 104 (D3): 3527-3542.
 Janzen, J. (1980) Extinction of light by highly nonspherical strongly absorbing colloidal particles:
      spectrophotometric determination of volume distributions for carbon blacks. Appl. Opt. 19: 2977-2985.
 Japar, S. M.; Szkarlat, A. C.; Pierson, W. R. (1984) The determination of the optical properties of airborne particle
      emissions from diesel vehicles. Sci. Total Environ. 36: 121-130.
Japar, S. M.; Brachaczek, W. W.; Gorse, R. A.,  Jr.; Norbeck, J. M.; Pierson, W. R. (1986) The contribution of
      elemental carbon to the optical properties of rural atmospheric aerosols. Atmos. Environ. 20: 1281-1289.
Jensen, V. (1974) Decomposition of angiosperm tree leaf litter. In: Dickinson, C. H.; Pugh, G. J. F., eds. Biology of
      plant litter decomposition: v. I. London, United Kingdom: Academic Press; pp. 69-104.
Johnson, D. W. (1984) Sulfur cycling in forests. Biogeochemistry 1: 29-43.
         March 2001
                                                4-173
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Johnson, D. W. (1992) Base cation distribution and cycling. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric
 2             deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
 3             Springer-Verlag, Inc.; pp. 275-333. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
 4             Ecological studies, analysis and synthesis: v. 91).
 5       Johnson, F. R.; Desvousges, W. H. (1997) Estimating stated preferences with rated-pair data: environmental, health,
 6             and employment effects of energy programs. J. Environ. Econ. Manage. 34: 79-99.
 7       Johnson, D. W.; Lindberg, S. E., eds. (1992a) Atmospheric deposition and forest nutrient cycling: a synthesis of the
 8             integrated forest study. New York, NY: Springer-Verlag, Inc. (Billings, W. D.; Golley, F.; Lange, O. L.;
 9             Olson, J. S.; Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
10       Johnson, D. W.; Lindberg, S. E., eds. (1992b) Nitrogen chemistry, deposition, and cycling in forests. In: Johnson,
H             D. W.; Lindberg, S. E., eds. Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated
12             forest study. New York, NY: Springer-Verlag, Inc.; pp. 150-213. (Billings, W. D.; Golley, F.; Lange, O. L.;
13             Olson, J. S.; Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
14       Johnson, D. W.; Mitchell, M. J. (1988) Responses of forest ecosystems to changing sulfur inputs. In: Maynard,
15             D. G., ed. Sulfur in the environment. New York, NY: Marcel Dekker, Inc.; pp. 219-262. [Books in Soils,
16             Plants, and the Environment, v. 67].
17       Johnson, D. W.; Taylor, G. E. (1989) Role of air pollution in forest decline in eastern North America. Water Air
18             SoilPollut.48:21-43.
19       Johnson, D. W.; Todd, D. E., Jr. (1998) Harvesting effects on long-term changes in nutrient pools of mixed oak
20             forest. Soil Sci. Soc. Amer. J. 62: 1725-1735.
21       Johnson, C. D.; Latimer, D. A.; Bergstrom, R. W.; Hogo, H. (1980) User's manual for the plume visibility model
22             (PLUVUE). Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
23             Planning and Standards; report no. EPA-450/4-80-032. Available from: NTIS, Springfield, VA;
24             PB81-163297.
25       Johnson, D. W.; Henderson, G. S.; Huff, D. D.; Lindberg, S. E.; Richter, D. D.; Shriner, D. S.; Todd, D. E.;
26             Turner, J. (1982a) Cycling of organic and inorganic sulphur in a chestnut oak forest. Oecologia 54: 141-148.
27       Johnson, D. W.; Cole, D. W.; Bledsoe, C. S.; Cromack, K.; Edmonds, R. L.; Gessel, S. P.; Grier, C. C.; Richards,
28             B. N.; Vogt, K. A. (1982b) Nutrient cycling in forests of the Pacific Northwest. In: Edmonds, R. L., ed.
29             Analysis of coniferous forest ecosystems in the Western United States. Stroudsburg, PA: Hutchinson Ross
30             Publishing Company; pp. 186-232. (US/IBP synthesis  series: v. 14).
31       Johnson, D. W.; Van Miegroet, H.; Lindberg, S. E.; Todd, D. E.; Harrison, R. B. (1991a) Nutrient cycling in red
32             spruce forests of the Great Smoky Mountains. Can. J. For. Res. 21: 769-787.
33       Johnson, D. W.; Cresser, M. S.; Nilsson, S. I.; Turner, J.; Ulrich, B.; Binkley, D.; Cole, D. W. (199 Ib) Soil changes
34            in forest ecosystems: evidence for and probable causes. Proc. R.  Soc. Edinburgh Sect. B: Biol. Sci.
35             97B:81-116.
36      Johnson, D. W.; Swank, W. T.; Vose, J. M. (1993) Simulated effects of atmospheric sulfur deposition on nutrient
37            cycling in a mixed deciduous forest. Biogeochemistry 23:169-196.
38      Johnson, A. H.; Friedland, A. J.; Miller, E. K.; Siccama, T. G. (1994a) Acid rain and soils of the Adirondacks. III.
39            Rates of soil acidification in a montane spruce-fir forest at Whiteface Mountain, New York. Can. J. For. Res.
40            24:663-669.
41       Johnson, A. H.; Andersen, S. B.; Siccama, T. G. (1994b) Acid rain and soils of the Adirondacks. I. Changes in pH
42            and available calcium, 1930-1984. Can. J. For. Res. 24: 39-45.
43      Johnson, D. W.; Susfalk, R. B.; Brewer, P. F. (1996) Simulated responses of red spruce forest soils to reduced
44            sulfur and nitrogen deposition. J. Environ. Qual. 25: 1300-1309.
45      Johnson, D.W.; Susfalk, R.B.; Brewer, P.,F.; Swank, W.T. (1999) Simulated effects of reduced sulfur, nitrogen, and
46            base cation deposition on soils and solutions in southern Appalachian forests. J. Environ. Qual
47            28:1336-1346.
48      Joslin, J. D.; Wolfe, M. H. (1992) Red spruce soil chemistry  and root distribution across a cloud water deposition
49            'gradient. Can. J. For. Res. 22: 893-904.
 50      Joslin, J. D.; Wolfe, M. H. (1994) Foliar deficiencies of mature Southern Appalachian red spruce determined from
 51            fertilizer trials. Soil Sci. Soc. Am. J. 58:1572-1579.
 52      Joslin, J. D.; Kelly, J. M.; Van Miegroet, H. (1992) Soil chemistry and nutrition of North American spruce-fir
 53            stands: evidence for recent change. J. Environ. Qual. 21: 12-30.
 54      Justus, C. G.; Murphey, B. B. (1994) Temporal trends in surface irradiance at ultraviolet wavelengths. J. Geophys.
 55            Res. 99 (Dl): 1389-1394.
          March 2001
4-174
DRAFT-DO NOT QUOTE OR CITE

-------
  1      Kalkstein, L. S.; Greene, J. S. (1997) An evaluation of climate/mortality relationships in large U.S. cities and the
  2            possible impacts of a climate change. Environ. Health Perspect. 105: 84-93.
  3      Karmoker, J. L.; Clarkson, D. T.; Saker, L. R.; Rooney, J. M.; Purves, J. V. (1991) Sulphate deprivation depresses
  4            the transport of nitrogen to the xylem and the hydraulic conductivity of barley (Hordeum vulgare L.) roots.
  5            Planta 185: 269-278.
  6      Katsev, I. L.; Zege, E. P. (1994) The modern theory of black object visibility and meteorological visibility range.
  7            Atmos. Environ. 28: 763-767.
  8      Kelly, J.; Lambert, M. J. (1972) The relationship between sulphur and nitrogen in the foliage ofPinus radiata.
  9            Plant Soil 37: 395-407.
 10      Kenk, G.; Fischer,  H. (1988) Evidence from nitrogen fertilisation in the forests of Germany. Environ. Pollut.
 11            54:199-218.
 12      Kerker, M. (1969) The scattering of light and other electromagnetic radiation. New York, NY: Academic Press, Inc.
 13            (Loebl, E. M., ed. Physical chemistry: a series of monographs, no. 16).
 14      Kiehl, J. T.;  Briegleb, B. P. (1993) The relative roles of sulfate aerosols and greenhouse gases in climate forcing.
 15            Science (Washington, DC) 260: 311-314.
 16      Kodama, Y.; Lee, S. D., eds. (1994) The 13th UOEH international symposium and the 2nd pan Pacific cooperative
 17            symposium on impact of increased UV-B exposure on human health and ecosystem;  October 1993;
 18            Kitakyushu, Japan. Kitakyushu, Japan: University of Occupational and Environmental Health.
 19      Koschmieder, H. (1924) Theorie der horizontalen Sichtweite II: Kontrast und Sichtweite [Theory of horizontal
20            visibility]. Beitr. Phys. Atmos. 12: 171-181.
21      Koslow, E. E.; Smith, W. H.; Staskawicz, B. J. (1977) Lead-containing particles on urban leaf surfaces.
22            Environ. Sci. Technol. 11: 1019-1021.
23      Krajickova, A.; Mejstfik, V. (1984)  The effect of fly-ash particles on the plugging of stomata. Environ. Pollut.
24            36:83-93.
25      Kripke, M. L. (1989) Health effects  of stratospheric ozone depletion: an overview. In: Schneider, T.; Lee, S. D.;
26            Wolters, G. J. R.; Grant, L. D. eds. Atmospheric ozone research and its policy implications: proceedings of
27            the Third US-Dutch International Symposium; May, 1988; Nijmegen, The Netherlands. Amsterdam,  The
28            Netherlands: Elsevier, pp. 795-802. [Studies in Environmental Science 35].
29      Krupa, S. V.; Legge, A. H. (1986) Single or joint effects of coarse and fine particulate sulfur aerosols and ozone on
30            vegetation. In: Lee, S. D.;  Schneider, T.; Grant, L. D.; Verkerk, P. J., eds. Aerosols: research, risk assessment
31            and control strategies, proceedings of the second U.S.-Dutch international symposium; May 1985;
32            Williamsburg, VA. Chelsea, MI: Lewis Publishers, Inc.; pp. 879-887.
33      Krupa, S. V.; Legge, A. H. (1998) Sulphur dioxide, particulate sulphur and its impacts on a  boreal forest ecosystem.
34            In: Ambasht, R. S., ed. Modern trends in ecology and environment.  Leiden, The Netherlands: Backhuys
35            Publishers; pp. 285-306.
36      Kylin, H.; Grimvall, E.; Ostman,  C. (1994) Environmental monitoring of polychlorinated biphenyls using pine
37            needles as passive samplers. Environ. Sci. Technol. 28: 1320-1324.
38      Lamersdorf,  N. P.; Meyer, M. (1993) Nutrient cycling and acidification of a northwest German forest site with high
39            atmospheric nitrogen deposition. For. Ecol. Manage. 62: 323-354.
40      Latimer, D. A. (1988) Plume perceptibility thresholds. Presented at: 81st annual meeting of the Air Pollution
41            Control Association; June; Dallas, TX. Pittsburgh, PA: Air Pollution Control Association; paper no. 88-56.6.
42      Latimer, D. A.; Bergstrom, R. W.; Hayes, S. R.; Liu, M.-K.; Seinfeld, J. H.; Whitten, G. Z.;  Wojcik, M. A.;  Hillyer,
43            M. J. (1978) The development of mathematical models for the prediction of anthropogenic visibility
44            impairment:  volumes I-III. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of
45            Air Quality Planning and Standards; EPA report no. EPA-450/3-78-11 Oa-c. Available from: NTIS,
46            Springfield, VA; PB-293118-SET.
47      Laurenroth, W. K.; Milchunas, D. G. (1985) SO2 effects on plant community function. In: Winner, W. E.; Mooney,
48            H. A.; Goldstein, R. A., eds. Sulfur dioxide and vegetation: physiology, ecology, and policy issues. Stanford,
49            CA: Stanford University Press; pp. 454-477.
50      Lawrence, G. B.; Huntington, T. G. (1999) Soil-Calcium depletion linked to acid rain and forest growth in the
51            eastern United States. Reston, VA: U.S. Geological Survey. Available: http://btdqs.usgs.gov/acidrain [1999,
52            November 23].
53      Lawrence, G. B.; David, M. B.; Shortle, W. C. (1995) A new mechanism for calcium loss in forest-floor soils.
54            Nature 378:  162-164.
         March 2001
4-175
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Legge, A. H. (1980) Primary productivity, sulfur dioxide, and the forest ecosystem: an overview of a case study.
 2             In: Miller, P. R., ed. Proceedings of symposium on effects of air pollutants on Mediterranean and temperate
 3             forest ecosystems; June; Riverside, CA. Berkeley, CA: U.S. Department of Agriculture, Pacific Southwest
 4             Forest and Range Experiment Station; pp. 51-62; Forest Service general technical report no. PSW-43.
 5             Available from: NTIS, Springfield, VA; PB81-133720.
 6       Lcrman, S. L.; Darley, E. F. (1975) Particulates. In: Mudd, J. B.; Kozlowski, T. T., eds. Responses of plants to air
 7             pollution. New York, NY: Academic Press; pp. 141-158.
 8       Levin, S. A. (1998) Ecosystems and the biosphere as complex adaptive systems. Ecosystems 1: 431-436.
 9       Li-Jones, X.; Prospero, J. M. (1998) Variations in the size distribution of non-sea-salt sulfate aerosol in the marine
10             boundary layer at Barbados: impact of African dust. J. Geophys. Res. 103: 16,073-16,084.
11       Likens, G. E.; Driscoll, C. T.; Buso, D. C. (1996) Long-term effects of acid rain: response and recovery of a forest
12             ecosystem. Science (Washington, DC) 272: 244-246.
13       Lindberg, S. E.; Harriss, R. C. (1981) The role of atmospheric deposition in an eastern U.S. deciduous forest.
14             Water Air Soil Pollut. 16:13-31.
15       Lindberg, S. E.; Lovett, G. M.; Meiwes, K.-J. (1987) Deposition and forest canopy interactions of airborne nitrate.
16             In: Hutchinson, T. C.; Meema, K. M., eds. Effects of atmospheric pollutants on forests, wetlands  and
17             agricultural ecosystems. Berlin, Federal Republic of Germany: Springer-Verlag; pp. 117-130. (NATO as
18             1 series: v. Gl 6).
19       Liu, S. C.; McKeen, S. A.; Madronich, S. (1991) Effect of anthropogenic aerosols on biologically active ultraviolet
20             radiation. Geophys. Res. Lett. 18:2265-2268.
21       Lokke, H.; Bak, J.; Falkengren-Grerup, U.; Finlay, R. D.; Ilvesniemi, H.; Nygaard, P. H.; Starr, M. (1996)
22             Critical loads of acidic deposition for forest soils: is the current approach adequate. Ambio 25: 510-516.
23       Longstreth, J. (1999) Public health consequences of global climate change in the United States—some regions may
24             suffer disproportionately. Environ. Health Perspect. 107(suppl. 1): 169-179.
25       Lorente, J.; Redano, A.; de Cabo, X. (1994) Influence of urban aerosol on spectral solar irradiance. J. Appl.
26             Meteorol. 33:406-415.
27       Lorusso, S.; Marabelli, M.; Troili, M. (1997) Air pollution and the deterioration of historic monuments. J. Environ.
28             Pathol. Toxicol. Oncol.  16: 171-173.
29       Lovett, G. M. (1984) Atmospheric deposition to forests. In: Breece, L.; Hasbrouck, S., eds. Forest-responses to
30             acidic deposition: proceedings of the U. S.-Canadian conference; August 1983; Orono, ME. Orono, ME:
31             University of Maine, Orono, Land and Water Resources Center; pp. 7-18.
32       Lovett, G. M.; Lindberg, S. E. (1984) Dry deposition and canopy exchange in a mixed oak forest as determined by
33             analysis of throughfall. J. Appl. Ecol. 21: 1013-1027.
34       Lovett, G. M.; Lindberg, S. E. (1986) Dry deposition of nitrate to a deciduous forest. Biogeochemistry 2: 137-148.
35       Lovett, G. M.; Lindberg, S. E. (1993) Atmospheric deposition and canopy interactions of nitrogen in forests. Can. J.
36             For. Res. 23: 1603-1616.
37       Lowenthal, D. H.; Rogers, C. F.; Saxena, P.; Watson, J. G.; Chow, J. C. (1995) Sensitivity of estimated light
38             extinction coefficients to model assumptions and measurement errors. Atmos. Environ. 29: 751-766.
39       Madronich, S. (1987) Photodissociation in the Atmosphere. 1. Actinic flux and the effects of ground reflections and
40             clouds. J. Geophys. Res. [Atmos.] 92: 9740-9752.
41       Mahadev, S.; Henry, R. C. (1999) Application of a color-appearance model to vision through atmospheric haze.
42             Color Res. Appl. 24:112-120.
43       Malm, W. C.; Pitchford, M. L. (1997) Comparison of calculated sulfate scattering efficiencies as estimated from
44             size-resolved particle measurements at three national locations. Atmos. Environ. 31: 1315-1325.
45       Malm, W. C.; Sisler, J. F.; Huffman, D.; Eldred, R. A.; Cahill, T. A. (1994) Spatial and seasonal trends in particle
46             concentration and optical extinction in the United States. J. Geophys. Res. [Atmos.] 99: 1347-1370.
47       Malm, W. C.; Molenar, J. V.; Eldred, R. A.; Sisler, J. F. (1996) Examining the relationship among atmospheric
48             aerosols and light scattering and extinction in the Grand Canyon area. J. Geophys. Res. [Atmos.]
49             101:19,251-19,265.
50       Malm, W. C.; Day, D.; Kreidenweis, S. M. (1997) Comparison of measured and reconstructed scattering during an
51             intensive field study at Great Smoky Mountains National Park. Presented at: 90th annual meeting and
52             exhibition of the Air & Waste Management Association; June; Toronto, ON, CA. Pittsburgh, PA: Air &
53             Waste Management Association; 97-WA70.02.
         March 2001
4-176
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
45
46
47
48
49
 50
 51
 52
53
54
55
 Mansfeld, F. (1980) Regional air pollution study: effects of airborne sulfur pollutants on materials. Research
       Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Sciences Research Laboratory;
       EPA report no. EPA-600/4-80-007. Available from: NTIS, Springfield, VA; PB81-126351.
 Marchwinska, E.; Kucharski, R. (1987) The combined influence of SO2 and heavy metal-containing particulates on
       beans, carrots and parsley. Environ. Monit. Assess. 8: 11-25.
 Markewitz, D.; Richter, D.D.; Allen, L. H.; Urrego, J. B. (1998) Three decades of observed soil acidification in the
       Calhoun Experimental Forest: Has acid rain made a difference? Soil Sci. Soc. Amer. J. 62: 1428-1439.
 Marschner, H. (1995) Sulfur. In: Mineral nutrition of higher plants. 2nd ed. London, United Kingdom: Academic
       Press; pp. 255-265.
 Martin, M. H.; Coughtrey, P. J. (1981) Impact of metals on ecosystem function and productivity. In: Lepp, N. W.,
       ed. Effect of heavy metal pollution on plants: volume 2, metals in the environment. Barking; United
       Kingdom: Applied Science Publishers; pp. 119-158. (Mellanby, K., ed. Pollution monitoring series).
 Martin, C. E.; Gravatt, D. A.; Loeschen, V. S. (1992) Photosynthetic responses of three species to acute exposures
       of nitrate- and sulphate-containing aerosols. Atmos. Environ. Part A 26: 381-391.
 Materna, J. (1984) Impact of atmospheric pollution on natural ecosystems. In: Treshow, M., ed. Air pollution and
       plant life. New York, NY: John Wiley & Sons; pp. 397-416.
 Mathai, C.  V. (1995) The Grand Canyon visibility. EM (December): 20-31.
 Matson, P.  A.; Parton, W. J.; Power, A. G.; Swift, M. J. (1997) Agricultural intensification and ecosystem
       properties. Science (Washington, DC) 277: 504-509.
 May, P. F.; Till, A. R.; Gumming, M. J. (1972) Systems analysis of sulphur kinetics in pastures grazed by sheep.
       J. Appl. Ecol. 9: 25-49.
 McGee, E.  S.; Mossotti, V. G. (1992) Gypsum accumulation on carbonate stone. Atmos. Environ. Part B
       26: 249-253.
 McGowan, T. F.; Lipinski, G. E.; Santoleri, J. J. (1993) New rules affect the handling of waste fuels. Chem. Eng
       (March): 122-128.
 McGrattan, K. B.; Walton, W. D.; Putori, A. D., Jr.; Twilley, W. H.; McElroy, J. A.; Evans, D. D. (1995) Smoke
       plume trajectory from in situ buring of crude oil in Alaska: field experiments. Gaithersburg, MD: U.S.
       National Institute of Standards and Technology; report no. NISTIR 5764. Available from: NTIS, Springfield
       VA;PB96-131560.
 McGrattan, K. B.; Trelles, J.; Baum, H. R.; Rehm, R. G. (1996) Smoke plume trajectory over two-dimensional
       terrain. In: Beall, K., ed. Annual conference on fire research: book of abstracts; October; Gaithersburg, MD.
       Gaithersburg, MD: United States Department of Commerce, National Institute of Standards and Technology;
      report no. NISTIR 5904; pp. 65-66. Available from: NTIS, Springfield, VA; PB97-153514.
 McLachlan, M. S. (1996a) Bioaccumulation of hydrophobic chemicals in  agricultural food chains. Environ Sci   •
      Tech. 30: 252-259.
 McLachlan, M. S. (1996b) Biological uptake and transfer of polychlorinated dibenzo-p-dioxins and dibenzofurans.
      In: Hester, R. E.; Harrison, R. M., eds. Chlorinated organic micropollutants. Cambridge, United Kingdom:
      Royal Society of Chemistry; pp. 31 -52. (Issues in environmental science and technology).
 McLachlan, M. S. (1999) Framework for  the interpretation of measurements of SOCs in plants. Environ.  Sci
      Technol. 33:1799-1804.
 McMurry, P. H.; Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
      aerosols. Atmos. Environ. 23: 497-507.
 McMurry, P. H.; Zhang, X.; Lee, C.-T. (1996) Issues in aerosol measurement for optics assessments. J. Geophys
      Res.  [Atmos.] 101: 19,189-19,197.
Meiwes, K. J.; Khanna, P. K. (1981) Distribution and cycling of sulphur in the vegetation of two forest ecosystems
      in an acid rain environment. Plant Soil 60: 369-375.
Meng,  R. Z.; Karamchandani, P.; Seigneur, C. (2000) Simulation of stack plume opacity. J. Air Waste Manage
      Assoc. 50: 869-874.
Mercer, G. S. (1994) Transmissometer data reporting (IMPROVE protocol). Fort Collins, CO: Air Resource
      Specialists, Inc.; technical instruction no. 4500-5100.
Mid-Atlantic Regional Assessment (MARA) Team. (2000) Preparing for a changing climate: the potential
      conswquences of climate variability and change (Mid-Atlantic overview). Washington, DC: U.S. Global
      Change Research Program (USGCRP).
         March 2001
                                               4-177
DRAFT-DO NOT QUOTE OR CITE

-------
 1      Middleton, P. (1997) DAQM-simulated spatial and temporal differences among visibility, PM, and other air quality
 2             concerns under realistic emission change scenarios. In: Special issue on aerosols and atmospheric optics-part
 3             II, [published papers from the international specialty conference; September 1994; Snowbird, UT]. J. Air
 4             Waste Manage. Assoc. 47: 302-316.
 5      Mie, G. (1908) Beitrage zur Optik truber Medien, speziell kolloidaler Metallosungen [Optics of cloudy media,
 6             especially colloidal metal solutions]. Ann. Phys. (Leipzig) 25: 377-445.
 7      Millar, C. S. (1974) Decomposition of coniferous leaf litter. In: Dickinson, C. H.; Pugh, G. J. F., eds. Biology of
 8             plant litter decomposition: volume I. New York, NY: Academic Press; pp. 105-128.
 9      Miller, P. R.; McBride, J. R., eds. (1999) Oxidant air pollution impacts in the Montane forests  of southern
10             California: a case study of the San Bernardino Mountains. New York, NY: Springer-Verlag.
11             (Ecological Studies: v. 134.)
12      Miller, W. C.; Fornes, R. E.; Gilbert, R. D.; Speer, A.; Spence, J. (1992) Removal of CaCO3 extender in residential
13             coatings by atmospheric acidic deposition. In: Measurement of toxic and related air pollutants: proceedings
14             of the 1992 U.S. EPA/A&WMA international symposium. Pittsburgh, PA: Air & Waste Management
15             Association; pp. 129-134. (A&WMA publication VIP-25).
16      Miller, E. K.; Panek, J. A.; Friedland, A. J.; Kadlecek, J.; Mohnen, V. A. (1993) Atmospheric deposition to a
17             high-elevation forest at Whiteface Mountain, New York, USA. Tellus Ser. B 45:209-227.
18      Mims, F. M., III. (1996) Significant reduction of UVB caused by smoke from biomass burning in Brazil.
19             Photochem. Photobiol. 64: 814-816.
20      Mitchell, M. J.; Harrison, R. B.; Fitzgerald, J. W.; Johnson, D. W.; Lindberg, S. E.; Zhang, Y.; Autry, A. (1992a)
21             Sulfur distribution and cycling in forest ecosystems. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric
22             deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
23             Springer-Verlag; pp. 90-129. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
24             Ecological studies: analysis and synthesis: v. 91).
25.      Mitchell, M. J.; Foster, N. W.; Shepard, J. P.; Morrison, I. K. (1992b) Nutrient cycling in Huntington Forest and
26             Turkey Lakes deciduous stands: nitrogen and sulfur. Can. J. For. Res. 22:457-464.
27       Molenar, J. V.; Persha, G.; Malm,  W. C. (1990) Long path transmissometer for measuring ambient atmospheric
28             extinction. In: Mathai, C. V., ed. Visibility and fine particles: an A&WMA/EPA international specialty
29             conference; October 1989; Estes Park, CO. Pittsburgh PA: Air & Waste Management Association;
30             pp. 293-304. (AWMA transactions series no. TR-17).
31       Molenar, J. V.; Cismoski, D. S.; Tree, R. M.  (1992) Intercomparison of ambient optical monitoring techniques.
32             Presented at: 85th annual meeting & exhibition of the Air & Waste Management Association; June;
33             Kansas City, MO. Pittsburgh, PA: Air & Waste Management Association; paper no. 92-60.09.
34       Molenar, J. V.; Malm, W. C.; Johnson, C. E. (1994) Visual air quality simulation techniques. Atmos. Environ.
35             28:1055-1063.
36       Moran, M. D.; Pielke, R. A. (1994) Delayed  shear enhancement in mesoscale atmospheric dispersion. Presented at:
37             8th joint conference on applications of air pollution meteorology with A&WMA; January; Nashville, TN.
38             Boston, MA: Meteorological Society; pp. 96-103.
39       Moropoulou, A.; Bisbikou, K.; Torfs, K.; Van Grieken, R.; Zezza, F.; Macri, F. (1998) Origin and growth of
40             weathering crusts on ancient marbles in industrial atmosphere. Atmos. Environ. 32: 967-982.
41       Naeem, S.; Thompson, L. J.; Lawler, S. P.; Lawton, J. H.; Woodfin, R. M. (1994) Declining biodiversity can alter
42            the performance of ecosystems. Nature 368: 734-737.
43       Nash, T. H., III. (1975) Influence of effluents from a zinc factory on lichens. Ecol. Monogr. 45: 183-198.
44      National Acid Precipitation Assessment Program. (1991) Acidic deposition: state of science and technology.
45             Summary report of the National Acid  Precipitation Assessment Program. Washington, DC: U.S. National
46            Acid Precipitation Assessment Program.
47      National Oceanic and Atmospheric Administration. (1992) ASOS (Automated Surface Observing System) user's
48            guide. Washington, DC: National Oceanic and Atmospheric Administration.
49      National Park Service. (1998) IMPROVE program goals and objectives. Washington,  DC: National Park Service.
50            Available: www2.nature.nps.gov/ard/impr/index.htm [1999, November 24].
51       National Research Council. (1993) Protecting visibility in national parks and wilderness areas. Washington, DC:
52            National Academy Press. 3v.
53
         March 2001
4-178
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
National Science and Technology Council. (1998) National acid precipitation assessment program biennial report to
      Congress: an integrated assessment; executive summary. Silver Spring, MD: U.S. Department of Commerce,
      National Oceanic and Atmospheric Administration. Available:
      www.nnic.noaa.gov/CENR/NAPAP/NAPAP_96.htm [1999, November 24].
National Weather Service. (1998) Visibility. Washington, DC: National Weather Service, National Oceanic and
      Atmospheric Administration. Available: www.nws.noaa.gov/asos/vsby.htm [1999, November 24].
Nazaroff, W. W.; Cass, G. R. (1991) Protecting museum collections from soiling due to the deposition of airborne
      particles. Atmos. Environ. Part A 25: 841-852.
Neff, W. D. (1997) The Denver brown cloud studies from the perspective of model assessment needs and the role of
      meteorology. J. Air Waste Manage. Assoc.
Nielsen, K. E.; Ladekarl, U. L.; Nornberg, P. (1999) Dynamic soil processes on heathland due to changes in
      vegetation to oak and Sitka spruce. For. Ecol. Manage. 114: 107-116.
Nihlgard, B. (1985) The ammonium hypothesis-an additional explanation to the forest dieback in Europe. Ambio
      14: 2-8.
Niklihska, M.; Laskowski, R.; Maryaiislci, M. (1998) Effect of heavy metals and storage time on two types of forest
      litter: basal respiration rate and exchangeable metals. Ecotoxicol. Environ. Safety 41: 8-18.
Noble, I. R.; Dirzo, R. (1997) Forests as human-dominated ecosystems. Science (Washington, DC) 277: 522-525.
Nord, A. G.; Ericsson, T. (1993) Chemical analysis of thin black layers on building stone. Stud. Conserv. 38: 25-35.
Norton, R. A.; Childers, N. F. (1954) Experiments with urea sprays on the peach. Proc. Am. Soc. Hortic. Sci.
      63:23-31.
Novakov, T.; Penner, J. E. (1993) Large contribution of organic aerosols to cloud-condensation-nuclei
      concentrations. Nature (London) 365: 823-826.
Ockenden, W. A.; Stinnes, E.; Parker, C.; Jones, K. C. (1998) Observations on persistent organic pollutants in
      plants: implications for their use as passive air samplers and for POP cycling. Environ. Sci. Technol.
      32: 2721-2726.
Odum, E. P. (1969) The strategy of ecosystem development: an understanding of ecological succession provides a
      basis for resolving man's conflict with nature. Science (Washington, DC) 164: 262-270.
Odum, E. P. (1985) Trends expected in stressed ecosystems. BioScience 35: 419-422.
Odum, E. P. (1993) The ecosystem. In: Ecology and our endangered life-support systems. 2nd ed. Sunderland, MA:
      Sinauer Associates, Inc.; pp. 38-67.
Oehme, M.; Biseth, A.; Schlabach, M.; Wiig, O. (1995) Concentrations of polychlorinated dibenzo-p-dioxins,
      dibenzofurans and non-ortho substituted biphenyls in polar bear milk from Svalbard  (Norway). Environ.
      Pollut. 90: 401-407.
Olszyk,  D. M.; Bytnerowicz, A.; Takemoto, B. K. (1989) Photochemical oxidant pollution and vegetation: effects of
      gases, fog, and particles. Environ. Pollut. 61:11-29.
Omar, A. H.; Biegalski, S.; Larson, S. M.; Landsberger, S. (1999) Particulate contributions  to light extinction and
      local forcing at a rural Illinois site. Atmos. Environ. 33: 2637-2646.
Ormrod, D. P. (1984) Impact of trace element pollution on plants. In: Treshow, M., ed. Air  pollution and plant life.
      Chichester, United Kingdom: John Wiley & Sons; pp. 291-319.
Paerl, H. W.; Bales, J. D.; Ausley, L. W.; Buzzelli, C. P.; Crowder, L. B.; Eby, L. A.; Go, M.; Peierls, B. L.;
      Richardson, T. L.; Ramus, J. S. (2000) Recent hurricanes result in continuing ecosystem impacts on USA's
      largest lagoonal estuary: Pamlico Sound, North Carolina. EOS Trans. Am. Geophys.  U.: In Press.
Pan, W.; Tatang, M. A.; McRae, G. J.; Prinn, R. G. (1997) Uncertainty analysis of direct radiative forcing by
      anthropogenic sulfate aerosols. J. Geophys. Res. 102 (D18): 21,915-21,924.
Pan, W.; Tatang, M. A.; McRae, G. J:; Prinn, R. G. (1998) Uncertainty analysis of indirect radiative forcing by
      anthropogenic sulfate aerosols. J. Geophys. Res. 103 (D4): 3815-3823.
Parish, S. B. (1910) The effect of cement dust on citrus trees. Plant World 13: 288-291.
Parker, A. (1955) The destructive effects of air pollution on materials. Colliery Guardian 191:447-450.
Parker, G. G. (1983) Throughfall and stemflow in the forest nutrient  cycle. In: Macfadyen, A.; Ford, E. D., eds.
      Advances in  ecological research, v. 13; pp. 57-120.
Patz, J. A.; Engelberg, D.; Last, J.  (2000a) The effects of changing weather on public health. Ann. Rev. Public
      Health 21: 271-307.
         March 2001
                                                4-179
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Patz, J. A.; McGeehin, M. A.; Bernard, S. M.; Ebi, K. L.; Epstein, P. R.; Grambsch, A.; Gubler, D. J.; Reiter, P.;
 2             Romieu, I.; Rose, J. B.; Samet, J. M.; Trtanj, J. (2000b) The potential health impacts of climate variability
 3             and change for the United States: executive summary of the report of the health sector of the U.S. national
 4             assessment. Environ. Health Perspect.  108: 367-376.
 5       Pearcy, R. W.; Bjorkman, O.; Caldwell, M. M.; Keeley, J. E.; Monson, R. K.; Strain, B. R. (1987) Carbon gain by
 6             plants in natural environments. BioScience 37: 21-29.
 7       Percy, K. E. (1991) Effects of acid rain on forest vegetation: morphological and non-mensurational growth effects.
 8             In: Rennie, P. J.; Robitaille, G., eds. Effects of Acid Rain on Forest Resources: Proceedings of the
 9             Conference ; June, 1983; Sainte-Foy, Quebec. Ottawa, Canada: Canadian Forestry Service; pp. 97-111.
10       Percy, W. H.; Sutherland, J.; Christensen, J. (1991) Paradoxical relationship between substrates and agonist-induced
11             contractions of opossum esophageal body and sphincter invitro. Dig. Dis. Sci. 36: 1057-1065.
12       Perry, K. D.; Cahill, T. A.; Eldred, R. A.; Butcher, D. D.; Gill, T. E. (1997) Long-range transport of North African
13             dust to the eastern United States. J. Geophys. Res. [Atmos] 102: 11,225-11,238.
14       Peterjohn, W. T.; Adams, M. B.; Gilliam, F. S. (1996) Symptoms of nitrogen saturation in two central Appalachian
15             hardwood forest ecosystems. Biogeochemistry 35: 507-522.
16       Peterson, J. T.; Flowers, E. C.; Rudisill, J. H.  (1978) Urban-rural solar radiation and atmospheric turbidity
17     <        measurements in the Los Angeles basin. J. Appl. Meteorol. 17: 1595-1609.
18       Peterson, G.; Allen, C. R.; Rolling, C. S. (1998) Ecological resilience, biodiversity, and scale. Ecosystems 1: 6-18.
19       Petuskey, W. T.; Richardson, D. A.;  Dolske, D. A. (1995) Aspects of the deterioration of sandstone masonry in
20             Anasazi dwelling ruins at Mesa Verde  National Park, Colorado, USA. Sci. Total Environ. 167: 145-159.
21       Pimentel, D.; Wilson, C.; McCullum, C.; Huang, R.; Dwen, P.; Flack, J.; Tran, Q.;  Saltman, T.; Cliff, B. (1997)
22             Economic and environmental  benefits of biodiversity. BioScience 47: 747-757.
23       Pimm, S. L. (1984) The complexity and stability of ecosystems. Nature (London) 307: 321-326.
24       Pitchford, M. L.; Malm, W. C. (1994) Development and applications of a standard  visual index. Atmos. Environ.
25             28: 1049-1054.
26       Pitchford, M. L.; McMurry, P. H. (1994) Relationship between measured water vapor growth and chemistry of
27             atmospheric aerosol for Grand Canyon, Arizona, in winter 1990. Atmos. Environ. 28: 827-839.
28       Prescott, C. E.; Parkinson, D. (1985) Effects of sulphur pollution on rates of litter decomposition in a pine forest.
29             Can. J. Bot. 63:1436-1443.
30       Pryor, S.; Steyn, D. (1994) Visibility and ambient aerosols in southwestern British Columbia during REVEAL
31             (REgional Visibility Experimental Assessment in the Lower Fraser Valley):  application of analytical
32             techniques for source apportionment and related visibility impacts of the REVEAL fine particulate  optical
33             and scene data. Victoria,  BC,  Canada:  British Columbia Ministry of Environment, Lands and Parks.
34             Available from: NTIS, Springfield, VA; MIC-96-00314/INS.
35       Pueschel, R. F. (1993) Potential climatic effects of anthropogenic aerosols. In: Jennings, S. G., ed. Aerosol effects
36             on climate. Tucson, AZ: University of Arizona Press; pp. 110-132.
37       Rapport, D. J.; Whitford, W. G. (1999) How ecosystems respond to stress: common properties of arid and aquatic
38             systems. BioScience 49:193-203.
39       Raynal, D. J.; Joslin, J. D.; Thornton, F. C.; Schaedle, M.; Henderson, G. S. (1990) Sensitivity of tree seedlings to
40             aluminum: III. Red spruce and loblolly pine. J. Environ. Qual.  19: 180-187.
41       Realini, M.; Negrotti, R.; Appollonia, L.; Vaudan, D. (1995) Deposition of particulate matter on  stone surfaces: an
42             experimental verification of its effects  on Carrara marble. In: Saiz-Jimenez,  C., ed. The deterioration of
43             monuments: proceedings of the 2nd international symposium on biodeterioration and biodegradation;
44             January 1994; Sevilla, Spain.  Sci. Total Environ. 167: 67-72.
45       Rennie, P. J. (1955) The uptake of nutrients by mature forest growth.  Plant Soil 7:49-95.
46       Reuss, J. O. (1983) Implications of the calcium-aluminum exchange system for the effect of acid precipitation on
47             soils. J. Environ. Qual. 12: 591-595.
48       Reuss, J. O.; Johnson, D. W. (1986)  Acid deposition and the acidification of soils and waters. New York,  NY:
49             Springer-Verlag. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds. Ecological
50             studies: analysis and synthesis: v. 59).
51       Richards, L. W. (1973) Size distribution of pigment particles: measurement and characterization  by light-scattering
52             techniques. In: Patton, T. C., ed. Pigment handbook. V. III. Characterization and physical  relationships.
53             New York, NY: John Wiley & Sons; pp. 89-100.
54       Richards, L. W. (1999) Use of the deciview haze index  as an indicator for regional haze. J. Air Waste Manage.
55             Assoc. 49: 1230-1237.
         March 2001
4-180
DRAFT-DO NOT QUOTE OR CITE

-------
  1      Richards, L. W.; Blanchard, C. L.; Blumenthal, D. L., eds. (1991) Navajo generating station visibility study.
  2            Final Report. Petaluma, CA: Sonoma Technology, Inc.; report no. STI-90200-1124-FR.
  3      Riggan, P. J.; Lockwood, R. N.; Lopez, E. N. (1985) Deposition and processing of airborne nitrogen pollutants in
  4            Mediterranean-type ecosystems of southern California. Environ. Sci. Technol. 19: 781-789.
  5      Rodney, D. R. (1952) The entrance of nitrogen compounds through the epidermis of apple leaves. Proc. Am. Soc.
  6            Horde. Sci. 59: 99-102.
  7      Roelofs, J, G. M.; Kempers, A. J.; Houdijk, A. L. F. M.; Jansen, J. (1985) The effect of air-borne ammonium
  8            sulphate on Pinus nigra var. maritima in the Netherlands. Plant Soil 84: 45-56.
  9      Roelofs, J. G. M.; Boxman, A. W.; Van Dijk, H. F. G. (1987) Effects of airborne ammonium on natural vegetation
 10            and forests. In: Asman, W. A. H.; Diederen, H. S. M. A., eds. Ammonia and acidification: proceedings [of a]
 11            symposium of the European Association for the Science of Air Pollution (EURAS AP); April; Bilthoven,
 12            The Netherlands. European Association for the Science of Air Pollution; pp. 266-276.
 13      Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993) Sources of fine organic
 14            aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks.
 15            Environ. Sci. Technol. 27: 1892-1904.
 16      Rosen, H.; Hansen, A. D. A.; Gundel, L.; Novakov, T. (1978) Identification of the optically absorbing component in
 17            urban aerosols. Appl. Opt. 17: 3859-3861.
 18      Rosenberg, N. J.; Blad, B. L.; Verma, S. B. (1983) Microclimate: the biological environment. 2nd ed. New York
 19            NY:  John Wiley & Sons.
20      Ross, D. M.; Malm, W. C.; Iyer, H. K. (1997) Human visual sensitivity to plumes with a Gaussian luminance
21            distribution: experiments to develop an empirical probability of detection model. J.  Air Waste Manage.
22            Assoc. 47: 370-382.
23      Rovira, A. D.; Davey, C. B. (1974) Biology of the rhizosphere. In: Carson, E. W., ed. The plant root and its
24            environment: proceedings of an institute; July, 1971; Blacksburg, VA. Charlottesville, VA: University Press
25            of Virginia; pp. 153-204.
26      Sabbioni, C. (1995) Contribution of atmospheric deposition to  the formation of damage layers. In: Jimenez-Saiz, C.,
27            ed. The deterioration of monuments: proceedings of the  2nd international symposium on biodeterioration and
28            biodegradation; January 1994; Sevilla, Spain. Sci. Total Environ. 167: 49-56.
29      Sabbioni, C.; Zappia, G.; Gobbi, G. (1992) Carbonaceous particles on carbonate building stones in a simulated
30            system. In: Proceedings of the 1992 European aerosol conference; September; Oxford, United Kingdom.
31            J. Aerosol Sci. 23(suppl. 1): S921-S924.
32      Sabbioni, C.; Zappia, G.; Gobbi, G. (1996) Carbonaceous particles and stone damage in a laboratory exposure
33            system. J. Geophys. Res. [Atmos.] 101: 19,621-19,627.
34      Sabbioni, C.; Zappia, G.; Ghedini, N.; Gobbi, G.; Favoni, O. (1998) Black crusts on ancient mortars.
35            Atmos. Environ. 32:215-223.
36      Saiz-Jimenez, C. (1993) Deposition of airborne organic pollutants on historic buildings. Atmos. Environ  Part B
37            27:77-85.
38      Samecka-Cymerman, A.; Kempers, A. J. (1999) Bioindication of heavy metals in the town Wro/caw (Poland) with
39            evergreen plants. Atmos. Environ. 33: 419-430.
40      Sanyal, B.; Singhania, G. K.  (1956) Atmospheric corrosion of metals: part I. J. Sci. Ind. Res. Sect. B 15: 448-455.
41      Saunders, P. J. W.; Godzik, S. (1986) Terrestrial vegetation-air pollutant interactions: nongaseous air pollutants.
42            In: Legge, A. H.; Krupa, S. V., eds. Air pollutants and their effects on the terrestrial ecosystem. New York,
43            NY: John Wiley & Sons; pp. 389-394. (Advances in environmental science and technology: v. 18).
44      Saxena, P.; Hildemann, L. M.; McMurry, P. H.; Seinfeld, J. H.  (1995) Organics alter hygroscopic behavior of
45            atmospheric particles. J. Geophys. Res. [Atmos.]  100: 18,755-18,770.
46      Schiavon, N.; Chiavari, G.; Schiavon, G.; Fabbri, D. (1995) Nature and decay effects of urban soiling on  granitic
47            building stones. In: Jimenez-Saiz, C., ed. The deterioration of monuments: proceedings of the 2nd
48            international symposium on biodeterioration and biodegradation; January 1994; Sevilla, Spain. Sci. Total
49            Environ. 167:87-101.
50      Schier, G. A.; Jensen, K. (1992) Atmospheric deposition effects on foliar injury and foliar leaching in red spruce.
51            In: Eager, C.; Adams,  M. B., eds. Ecology and decline of red spruce in the eastern United States. New York,
52            NY: Springer-Verlag; pp. 271-294. [Billings, W. D.; Golley, F.; Lange, O. L.;  Olson, J. D.; Remmert, H.,
53            eds. Ecological Studies, Analysis and Synthesis Ecological Studies, v. 96].
         March 2001
4-181
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Schnug, E. (1997) Significance of sulphur for the quality of domesticated plants. In: Cram, W. J.; De Kok, L. J.;
 2             Stulen, I.; Brunold, C.; Rennenberg, H., eds. Sulphur metabolism in higher plants: molecular,
 3             ecophysiological and nutritional aspects. Leiden, The Netherlands: Backhuys Publishers; pp. 109-130.
 4       Schulze, E.-D. (1989) Air pollution and forest decline in a spruce (Picea abies) forest. Science (Washington, DC)
 5             244:776-783.
 6       Schuster, P. F.; Reddy, M. M.; Sherwood, S. I. (1994) Effects of acid rain and sulfur dioxide on marble dissolution.
 7             Mater. Perform. 33: 76-80.
 8       Schwar, M. J. R. (1998) Nuisance dust deposition and soiling rate measurements. Environ. Technol. 19: 223-229.
 9       Seckmeyer, G.; McKenzie, R. L. (1992) Increased ultraviolet radiation in New Zealand (45 degrees S) relative to
10             Germany (48 degrees N). Nature 359:135ff.
11       Seigneur, C.; Wu, X. A.; Constantinou, E.; Gillespie, P.; Bergstrom, R. W.; Sykes, I.; Venkatram, A.;
12             Karamchandani, P. (1997) Formulation of a second-generation reactive plume and visibility model. J. Air
13             Waste Manage. Assoc. 47:176-184.
14       Seinfeld, J. H. (1989) Urban air pollution: state of the science. Science 243: 745-752.
15       Sereda, P. J. (1974) Weather factors affecting corrosion of metals. In: Corrosion in natural environments: three
16             symposia presented at the seventy-sixth annual meeting [of the] American Society for Testing and Materials;
17             June 1973; Philadelphia, PA. Philadelphia, PA: American Society for Testing and Materials; pp. 7-22.
18             (ASTM special technical publication 558).
19       Shannon, J. D.; Trexler, E. C., Jr.; Sonnenblick, R. (1997) Modeling visibility for assessment. Armos. Environ.
20             31:3719-3727.
21       Sievering, H.; Rusch, D.; Marquez, L. (1996) Nitric acid, particulate nitrate and ammonium in the continental free
22             troposphere: nitrogen deposition to an alpine tundra ecosystem. Armos. Environ. 30:2527-2537.
23       Shodor Education Foundation, Inc.; Industrial Extension Service of North Carolina State University; North Carolina
24             Supercomputing Center. (1996) Air quality meteorology. Session 8: physical meteorology and visibility.
25             Part 3: visibility. Research Triangle Park, NC: U.S. Environmental Protection Agency, National Exposure
26             Research Laboratory and U.S. National Oceanic and Atmospheric Administration. Available:
27             http://www.shodor.org/metweb/outline.html [1999, November 24].
28       Shortle, W. C.; Smith, K. T. (1988) Aluminum-induced calcium deficiency syndrome in declining red spruce.
29             Science (Washington, DC) 240: 1017-1018.
30       Shortle, W. C.; Bondietti, E. A. (1992) Timing, magnitude, and impact of acidic deposition on sensitive forest sites.
31             Water Air Soil Pollut. 61:253-267.
32       Showak, W.; Dunbar, S. R. (1982) Effect of 1 percent copper addition on atmospheric corrosion of rolled zinc after
33             20 years' exposure. In: Dean, S. W.; Rhea, E. C., eds. Atmospheric corrosion of metals: a symposium;
34             May 1980; Denver, CO. Philadelphia, PA: American Society for Testing and Materials; pp. 135-162.
3 5             (ASTM special technical publication 767).
36       Shriner, D. S.; Henderson, G. S. (1978) Sulfur distribution and cycling in a deciduous forest watershed. J. Environ.
37             Qual. 7:392-397.
38       Simonich; S. L.; Kites, R. A. (1995) Organic pollutant accumulation in vegetation. Environ. Sci. Technol.
39             29:2905-2914.
40       Simpson, J. W.; Horrobin, P. J., eds. (1970) The weathering and performance of building materials. New York, NY:
41             Wiley-Interscience.
42       Sisler, J. F.; Cahill, T. A. (1993) Spacial and temporal patterns and the chemical composition of the haze and its
43             impact on visibility in Alaska. In: Proceedings [of the] 86th annual meeting & exhibition of the Air & Waste
44             Management Association. Volume 8:  general environmental topics; June; Denver, CO. Pittsburgh, PA: Air &
45             Waste Management Association; 93-MP-4.03.
46       Sisler, J. F.; Malm, W. C. (1994) The relative importance of soluble aerosols to spatial and seasonal trends of
47             impaired visibility in the United States. Atmos. Environ. 28: 851 -862.
48       Sisler, J. F.; Malm, W. C. (2000) Interpretation of trends of PM2 5 and reconstructed visibility from the IMPROVE
49             network. J. Air Waste Manage. Assoc.: submitted.
50       Sisler, J. F.; Huffman, D.; Lattimer, D. A.; Malm, W. C.; Pitchford, M. L. (1993) Spatial and temporal patterns and
51             the chemical composition of the haze  in the United States: an anlysis of data from the IMPROVE network,
52             1988-1991. Fort Collins, CO: Colorado State University, Cooperative Institute for Research in the
53             Atmosphere (CIRA).
54      Skeffington, R. A.; Wilson, E. J. (1988) Excess nitrogen deposition: issues for consideration. Environ. Pollut.
55             54:159-184.
         March 2001
4-182
DRAFT-DO NOT QUOTE OR CITE

-------
  1       Skerry, B, S.; Alavi, A.; Lindgren, K. I. (1988) Environmental and electrochemical test methods for the evaluation
  2             of protective organic coatings. J. Coat. Technol. 60: 97-106.
  3       Sloane, C. S.; Watson, J.; Chow, J.; Pritchett, L.; Richards, L. W. (1991) Size-segregated fine particle
  4             measurements by chemical species and their impact on visibility impairment in Denver. Atmos. Environ.
  5             Part A 25: 1013-1024.
  6       Smith, W. H. (1974) Air pollution - effects on the structure and function of the temperate forest ecosystem.
  7             Environ. Pollut. 6: 111-129.
  8       Smith, W. H. (1990a) Forests as sinks for air contaminants: soil compartment. In: Air pollution and forests:
  9             interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag;
 10             pp. 113-146. (Springer series on environmental management).
 11       Smith, W. H. (1990b) Forest biotic agent stress: air pollutants and disease caused by microbial pathogens. In: Air
 12             pollution and forests: interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY:
 13             Springer-Verlag; pp. 366-397.  (Springer series on environmental management).
 14       Smith, W. H. (1990c) Forests as sinks for air contaminants: vegetative compartment. In: Air pollution and forests:
 15             interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag;
 16             pp. 147-180. (Springer series on environmental management).
 17       Smith, W. H. (1990d) Forests as sources of hydrocarbons, particulates, and other contaminants. In: Air pollution
 18             and forests: interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY:
 19             Springer-Verlag; pp. 83-112. (Springer series on environmental management).
 20       Smith, W. H. (1990e) Forest nutrient cycling: toxic ions. In: Air pollution and forests: interactions between air
 21             contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag; pp. 225-268. (Springer series
 22             on environmental management).
 23       Smith, W. H. (1991) Air pollution and forest damage. Chem.  Eng. News 69(45):  30-43.
 24       Smith, W. H.; Staskawicz, B. J. (1977) Removal of atmospheric particles by leaves and twigs of urban trees:
 25             some preliminary observations and assessment of research needs. Environ. Manage. (N. Y.) 1:  317-330.
 26       Soulen, P. E.; Frederick, J.  E. (1999) Estimating biologically active UV irradiance from satellite radiance
 27            measurements: a sensitivity study. J. Geophys. Res. 104: 4117-4126.
 28       Spence, J. W.; Haynie, F. H. (1972) Paint technology and air pollution: a survey and economic assessment.
 29            Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Programs; publication no.
 30            AP-103. Available from: NTIS, Springfield, VA; PB-210736.
 31       Spence, J. W.; Haynie, F.; Upham, J. B. (1975) Effects of gaseous pollutants on paints: a chamber study. J. Paint
 32            Technol. 47: 57-63.
 33       Spence, J. W.; Haynie, F. H.; Lipfert, F. W.; Cramer, S. D.; McDonald, L. G. (1992) Atmospheric corrosion model
 34            for galvanized steel structures.  Corrosion 48:  1009-1019.
 35       Spiker, E. C.; Comer, V. J.; Hosker, R. P.; Sherwood, S. I. (1992) Dry deposition of SO2 on limestone and marble:
 36            the role of humidity. In: Rodrigues, J. D.; Henriques, F.; Jeremias, F. T., eds. Proceedings of the 7th
 37            international congress on deterioration and conservation of stone; June; Lisbon, Portugal. Lisbon, Portugal:
 38            Laboratorio Nacional de Engenharia Civil; pp. 397-406.
 39      Spinka, J. (1971) Vliv znecisteneho ovzdusi na ovocne stromy a zeleninu [Effects of polluted air on fruit trees and
 40            legumes]. Ziva 19: 13-15.
 41      Stevens, R. K.; Dzubay, T.  G.; Lewis, C. W.; Shaw, R. W., Jr. (1984) Source apportionment methods  applied to the
 42            determination of the  origin of ambient aerosols that affect visibility in forested areas. Atmos. Environ.
43             18:261-272.
44      Stoddard, J. L. (1994) Long-term changes  in watershed retention of nitrogen: its causes and aquatic consequences.
45            In: Baker, L. A., ed. Environmental chemistry of lakes and reservoirs. Washington, DC: American Chemical
46            Society; pp. 223-284. (Advances in chemistry series no. 237).
47      Stoddard, J. L.; Murdoch, P. S. (1991) Catskill Mountains. In: Charles, D. F.,  ed.  Acidic deposition and aquatic
48            ecosystems: regional case studies. New York, NY: Springer-Verlag; pp. 237-271.
49      Storm, G. L.; Fosmire, G. J.; Bellis, E. D. (1994) Persistence of metals in soil  and selected vertebrates in the vicinity
50            of the Palmerton zinc smelters. J. Environ. Qual. 23: 508-514.
51       Strandberg, H.; Johansson,  L.-G. (1997) The formation of black patina on copper in humid air containing traces  of
 52            SO2. J. Electrochem. Soc. 144:  81 -89.
53      Suzuki, D. T. (1997) Barriers to perception: from a world of interconnection to fragmentation. In: Raven, P. H., ed.
54            Nature and Human Society: the quest for a sustainable world.  Washington, DC: National Academy Press;
55            pp. 11-21.
         March 2001
4-183
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Sydberger, T.; Ericsson, R. (1977) Laboratory testing of the atmospheric corrosion of steel. Werkst. Korros.
 2             28: 154-158.
 3       Sykes, R. I.; Gabruk, R. S. (1997) A second-order closure model for the effect of averaging time on turbulent plume
 4             dispersion. J. Appl. Meteorol. 36: 1038-1045.
 5       Systems Applications International, Inc. (1998) User's guide to the regulatory modeling system for aerosols and
 6             deposition (REMSAD). San Rafael, CA: Systems Applications International, Inc.; report no.
 7             SYSAPP98-96\42r2.
 8       Tang, I. N. (1997) Thermodynamic and optical properties of mixed-salt aerosols of atmospheric importance.
 9             J. Geophys. Res. [Atmos.] 102: 1883-1893.
10       Taylor, G. E., Jr.; Johnson, D. W.; Andersen, C. P. (1994) Air pollution and forest ecosystems: a regional to global
11             perspective. Ecol. Appl. 4: 662-689.
12       Tegen, I.; Lacis, A. A. (1996) Modelling of particle size distribution and its influence on the radiative properties of
13             mineral dust aerosol. J. Geophys. Res. Atmos. 101: 19,237-19,244.
14       Tegen, I.; Lacis, A. A.; Fung, I. (1996) The influence on climate forcing of mineral aerosols from disturbed soils.
15             Nature (London) 380: 419-422.
16       Thornton, F. C.; Schaedle, M.; Raynal, D. J. (1987) Effects of aluminum on red spruce seedlings in solution culture.
17             Environ. Exp. Bot. 27: 489-498.
18       Tilman, D. (1996) Biodiversity: population versus ecosystem stability. Ecology 77: 350-363.
19       Tilman, D. (2000) Causes, consequences and ethics of biodiversity. Nature 405: 208-211.
20       Tilman, D.; Downing, J. A.  (1994) Biodiversity and stability in grasslands. Nature (London) 367: 363-365.
21       Tong, S. T. Y. (1991) The retention of copper and lead particulate matter in plant foliage and forest soil.
22             Environ. Int. 17:31-37.
23       Trettin, C.C.; Johnson, D.W.; Todd, D.E., Jr. (1999) Forest nutrient and carbon pools at Walker Branch watershed:
24             changes during a 21-year-period. Soil Sci. Soc. Arner. J. 63: 1436-1448.
25       Trijonis, J.; Shapland, D. (1979) Existing visibility levels in the United States: isopleth maps of visibility in
26             suburban/nonurban areas during 1974-76. Research Triangle Park, NC: U.S. Environmental Protection
27             Agency, Office of Air Quality Planning and Standards; report no. EPA-450/5-79-010. Available from:
28             NTIS, Springfield, VA; PB80-156177.
29       Trijonis, J. C.; Pitchford, M.; McGown, M. (1987) Preliminary extinction budget results from the RESOLVE
30             program. In: Bhardwaja, P. S., ed. Visibility protection: research and policy aspects, an APCA international
31             specialty conference; September 1986; Grand Teton National Park,  W Y. Pittsburgh, PA: Air Pollution
32             Control Association; pp. 872-883.  (APCA transactions series no. TR-10).
33       Trijonis, J. C.; Malm, W. C.; Pitchford, M.; White,  W. H. (1991) Visibility: existing and historical
34             conditions-causes and effects. In:  Irving, P.  M., ed. Acidic deposition: state of science and technology,
35             volume III: terrestrial, materials, health and visibility effects. Washington, DC: The U.S. National Acid
36             Precipitation Assessment Program. (State of science and technology report no. 24).
37       Turner, J.; Kelly, J. (1981) Relationships between soil nutrients and vegetation in a north coast forest, New South
38             Wales. Aust. For. Res. 11: 201-208
39       Turner, J.; Lambert, M. J. (1980) Sulfur nutrition of forests. In: Shriner, D. S.; Richmond, C. R.; Lingberg, S. E.,
40             eds. Atmospheric sulfur deposition: environmental impact and health effects, proceedings of the second life
41             sciences symposium; October 1979; Gatlinburg, TN. Ann Arbor, MI:  Ann Arbor Science Publishers, Inc.;
42             pp. 321-333.
43       Turner, J.; Lambert, M. J.; Gessel, S. P. (1977) Use of foliage sulphate concentrations to predict response to urea
44             application by Douglas-fir. Can. J. For. Res.  7: 476-480.
45       Twomey, S. (1974) Pollution and the planetary albedo. Atmos. Environ. 8: 1251-1256.
46       Twomey, S. (1977) Atmospheric aerosols. Amsterdam, The Netherlands: Elsevier Scientific Publishing Company;
47             p. 237.
48       Tyler, G. (1972) Heavy metals pollute nature, may reduce productivity. Ambio 1: 52-59.
49       U.S. Code. (1991) Clean Air Act, §109, national ambient air quality standards. U. S. C.  42: §7409.
50       U.S. Environmental Protection Agency. (1979) Protecting visibility: an EPA report to Congress. Research Triangle
51             Park, NC: Office of Air Quality Planning and Standards; EPA report no. EPA-450/5-79-008. Available from:
52             NTIS, Springfield, VA; PB80-220320.
53       U.S. Environmental Protection Agency. (1982) Air quality criteria for particulate matter and sulfur oxides. Research
54             Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
55             Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from:  NTIS, Springfield, VA; PB84-156777.
         March 2001
4-184
DRAFT-DO NOT QUOTE OR CITE

-------
  1       U.S. Environmental Protection Agency. (1986 ) Air quality criteria for lead. Research Triangle Park, NC: Office of
  2             Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report no.
  3             EPA-600/8-83/028aF-dF. 4v. Available from: NTIS, Springfield, VA; PB87-142378.
  4       U.S. Environmental Protection Agency. (1987) Assessing the risks of trace gases that can modify the stratosphere.
  5             Washington, DC: U.S. Environmental Protection Agency, Office of Air Radiation; report no. EPA
  6             400/1-87/001A-H.
  7       U.S. Environmental Protection Agency. (1988) Workbook for plume visual impact screening and analysis. Research
  8             Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-450/4-88-015.
  9             Available from: NTIS, Springfield, VA; PB89-151286.
10       U.S. Environmental Protection Agency. (1992) Interagency workgroup  on air quality modeling (IWAQM) work
11             plan rationale. Research Triangle Park, NC: Office of Air Quality Planning and Standards; report no.
12             EPA-454/R-92-001.
13       U.S. Environmental Protection Agency. (1993) Air quality criteria for oxides of nitrogen. Research Triangle Park,
14             NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
15             nos. EPA/600/8-9!/049aF-cF. 3v. Available from: NTIS, Springfield, VA; PB95-124533, PB95-124525, and
16             PB95-124517.
17       U.S. Environmental Protection Agency. (1995a) Inter-agency workgroup on air quality modeling (IWAQM):
18             assessment of phase I recommendations regarding the use of MESOPUFF II. Research Triangle Park, NC:
19             U.S. Environmental Protection Agency, Emissions, Monitoring and Analysis Division; report no.
20             EPA-454/R-95-006. Available from: NTIS, Springfield, VA; PB95-241931. P. 106 (1995a).
21       U.S. Environmental Protection Agency. (1995b) Testing of meteorological and dispersion models for use in
22             regional air quality modeling. Research Triangle Park, NC: Office of Air Quality Planning and Standards;
23             report no. EPA-454/R-95-005. Available from: NTIS,  Springfield, VA; PB95-215638.
24       U.S. Environmental Protection Agency. (1995c) Interim findings on the status of visibility research. Research
25             Triangle Park, NC: Office of Research and Development; report  no. EPA/600/R-95/021.
26       U.S. Environmental Protection Agency. (1996a) Air quality criteria for particulate matter. Research Triangle Park,
27             NC: National Center for Environmental Assessment-RTP Office; report nos. EPA/600/P-95/001aF-cF. 3v.
28       U.S. Environmental Protection Agency. (1996b) Air quality criteria for  ozone and related photochemical oxidants.
29             Research Triangle Park, NC: Office of Research and Development; report nos. EPA/600/AP-93/004aF-cF.
30             3v. Available from: NTIS, Springfield, VA; PB96-185582, PB96-185590, and PB96-185608. Available
31             online at: www.epa.gov/ncea/ozone.htm.
32       U.S. Environmental Protection Agency. (1997a) Nitrogen oxides: impacts on public health and the environment.
33             Washington, DC: Office of Air and Radiation; August. Available:
34             www.epa.gov/rtncaaal/tl/reports/noxrept.pdf [1999, November 24].
35       U.S. Environmental Protection Agency. (1997b) Summary of the particulate monitoring program. Washington, DC:
36             Office of Air and Radiation.
37       U.S. Environmental Protection Agency. (1999)  Protection of visibility. C. F. R. 40: "Section" 51.300-51.309.
38       U.S. Environmental Protection Agency. (2000a) Deposition of air pollutants to the great waters. Third report to
39             Congress. [Executive Summary]. Research Triangle Park, NC: U.S. Environmental Protection Agency,
40             Office of Air Quality Planning and Standards; report no. EPA-453/R-00-005.
41       U.S. Environmental Protection Agency. (2000b) Latest findings on national air quality: 1999 status and trends.
42             Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
43             Standards; report no. EPA-454/F-00-002.
44       U. S. Global Change Research Program (USGCRP). (2000) Climate Change Impacts on the United States:  the '
45             Potential Consequences of Climate Variability and Change (Overview), Report of National Assessment
46             Synthesis Team (NAST). NSTC Review Draft (September).
47       United Nations Environment Programme (UNEP). (1986) Report of the international conference on the assessment
48             of the role of carbon dioxide  and of other greenhouse gases in climate variations and associated impacts;
49             October 1985; Villach, Austria. Paris, France: International Council of Scientific Unions. [World
50             Meteorological Organization series: WMO no. 661].
51       United Nations Environment Programme (UNEP). (1998) Environmental effects of ozone depletion:  1998
52             assessment. J. Photochem. Photobiol. B 46: 1-4.
53       Unsworth, M. H.; Wilshaw, J. C. (1989) Wet, occult and dry deposition of pollutants on forests. Agric. For.
54             Meteorol. 47: 221-238.
         March 2001
4-185
DRAFT-DO NOT QUOTE OR CITE

-------
          1       Van Breemen, N.; Van Dijk, H. F. G. (1988) Ecosystem effects of atmospheric deposition of nitrogen in The
          2             Netherlands. In: Dempster, J. P.; Manning, W. J., eds. Excess nitrogen deposition. Environ. Pollut.
          3             54:249-274.
          4       Van Breemen, N.; Burrough, P. A.; Velthorst, E. J.; Van Dobben, H. F.; De Wit, T.; Ridder, T. B.; Reijnders, H. F.
          5             R. (1982) Soil acidification from atmospheric ammonium sulphate in forest canopy throughfall. Nature
          6             (London) 299: 548-550.
          7       Van Breemen, N.; Mulder, J.; Driscoll, C. T. (1983) Acidification and alkalinization of soils. Plant Soil
          8             75:283-308.
          9       Van der Leun, J. C.; Tang, X.; Tevini, M. (1995) Environmental effects of ozone depletion:  1994 assessment.
         10             Ambio24:138.
         11       Van der Leun, J. C.; Tang, X.; Tevini, M., eds. (1998) Environmental effects of ozone depletion: 1998 assessment.
         12             J. Photochem. Photobiol. B: Biology 46(1-3): 1-108.
         13       Van Dijk, H. F. G.; Roelofs, J. G. M. (1988) Effects of excessive ammonium deposition on the nutritional status
         14             and condition of pine needles. Physiol. Plant. 73: 494-501.
         15       Van Dorland, R.; Dentener, F. J.; Lelieveld, J. (1997) Radiative forcing due to tropospheric ozone and sulfate
         16             aerosols. J. Geophys. Res. [Atmos.] 102:28,079-28,100.
         17       Van Miegroet, H.; Cole, D. W. (1984) The impact of nitrification on soil acidification and cation leaching in red
         18             alder ecosystem. J. Environ. Qual. 13: 586-590.
         19       Van Miegroet, H.; Johnson, D. W.; Todd, D. E. (1993) Foliar response of red spruce saplings to fertilization with
         20             Ca and Mg in the Great Smoky Mountains National Park. Can. J. For. Res. 23: 89-95.
         21       Vasconcelos, L. A. de P.; Kahl, J. D. W.; Liu, D.; Macias, E. S.; White, W. H. (1996) Patterns of dust transport to
         22             the Grand Canyon. Geophys. Res. Lett. 23: 3187-3190.
         23       Viles, H. A. (1990) The early stages of building stone decay in an urban environment. Atmos. Environ.
         24             24A: 229-232.
         25       Vitousek, P. M.; Mooney, H. A.; Lubchenco, J.; Melillo, J. M. (1997) Human domination of Earth's ecosystems.
         26             Science (Washington, DC) 277:494-499.
         27       Vogt, K. A.; Dahlgren, R.; Ugolini, F.; Zabowski, D.; Moore, E. E.; Zasoski, R. (1987a) Aluminum, Fe, Ca, Mg, K,
         28             Mn, Cu, Zn and P in above- and belowground biomass. I. Abies amabilis and Tsuga mertensiana.
         29             Biogeochemistry 4:277-294.
         30       Vogt, K. A.; Dahlgren, R.; Ugolini, F.; Zabowski, D.; Moore, E. E.; Zasoski, R. (1987b) Aluminum, Fe, Ca, Mg, K,
         31             Mn, Cu, Zn and P in above- and belowground biomass. II. Pools, and circulation in a subalpine Abies
         32             amabilis stand. Biogeochemistry 4: 295-311.
         33       Vose, J. M.; Swank, W. T. (1990) Preliminary estimates of foliar absorption of 15N-labeled nitric acid vapor
         34            (HNO3) by mature eastern white pine (Pinus strobus). Can. J. For. Res. 20: 857-860.
         35       Waggoner, A. P.; Weiss, R. E. (1980) Comparison of fine particle mass concentration and light scattering extinction
         36             in ambient aerosol. Atmos. Environ. 14:623-626.
         37       Waggoner, A. P.; Weiss, R. E.; Ahlquist, N. C.; Covert, D. S.; Will, S.; Charlson, R. J. (1981) Optical
         38             characteristics of atmospheric aerosols. In: White, W. H.; Moore, D. J.; Lodge, J. P., Jr., eds. Plumes and
         39             visibility: measurements and model components: proceedings of the symposium; November 1980;
         40             Grand Canyon National Park, AZ. Atmos.  Environ. 15: 1891 -1909.
         41       Wagrowski, D. M.; Kites, R. A. (1997) Polycyclic aromatic hydrocarbon accumulation in urban, suburban, and rural
         42            vegetation. Environ. Sci. Technol. 31: 279-282.
         43       Wall, D. H. (1999) Biodiversity and ecosystem functioning. BioScience 49: 107-108.
         44      Wall, D. H.; Moore, J. C. (1999) Interactions underground: soil biodiversity, mutualism, and ecosystem processes.
         45             Bioscience49:109-117.
         46      Walton, J. R.; Johnson, J. B.; Wood, G. C. (1982) Atmospheric corrosion initiation by sulphur dioxide and
         47            particulate matter: II. characterisation and corrosivity of individual particulate atmospheric pollutants.
         48            Br. Corros. J. 17: 65-70.
         49      Wang, W. C.; Yung, Y. L.; Lacis, A. A.; Mo, T.; Hansen, J. E. (1976) Greenhouse effects due to man-made
         50            perturbations of trace gases. Science (Washington, DC) 194: 685-690.
         51       Waring, R. H. (1987) Nitrate pollution: a particular danger to boreal and subalpine coniferous forests. In:
         52            Fujimori, T.; Kimura, M., eds. Human impacts and management of mountain forests:  [proceedings of a
         53             symposium]. Ibaraki, Japan: Forestry and Forest Products Research Institute; pp. 93-105.
         54      Waring, R. H.; Schlesinger, W. H. (1985) The carbon balance of trees. In: Forest ecosystems: concepts and
         55            management. Orlando,  FL: Academic Press, Inc.; pp. 7-37.
                  March 2001
4-186
DRAFT-DO NOT QUOTE OR CITE
_

-------
  1      Watson, J. G.; Chow, J. C. (1994) Clear sky visibility as a challenge for society. Annu. Rev. Energy Environ.
  2             19:241-266.
  3      Watson, J. G.; Chow, J. C.; Richards, L. W.; Haase, D. L.; McDade, C.; Dietrich, D. L.; Moon, D.; Sloane, C.
  4             (1991) The 1989-90 Phoenix urban haze study, volume II: the apportionment of light extinction to sources
  5             [final report]. Reno, NV: University and Community College System of Nevada, Desert Research Institute;
  6             DRI document no. 8931.5F1.
  7      Webb, A. H.; Bawden, R. J.; Busby, A. K.; Hopkins, J. N. (1992) Studies on the effects of air pollution on
  8             limestone degradation in Great Britain. Atmos. Environ. Part B 26: 165-181.
  9      Wedin, D. A.; Tilman, D.  (1996) Influence of nitrogen loading and species composition on the carbon balance of
 10             grasslands. Science 274: 1720-1723.
 11      Weinbaum, S. A.; Neumann, P. M. (1977) Uptake and metabolism of 15N-labeled potassium nitrate by French
 12            prune (Prunus domestica L.) leaves and the effects of two surfactants. J. Am. Soc. Hortic. Sci. 102: 601-604.
 13      Wellburn, A. R. (1990) Tansley review no. 24: Why are atmospheric oxides of nitrogen usually phytotoxic and not
 14             alternative fertilizers? New Phytol. 115: 395-429.
 15      Wenny, B. N.; Schafer, J.  S.; DeLuisi, J. J.; Saxena, V. K.; Barnard, W. F.; Petropavlovskikh, I. V.; Vergamini,
 16            A. J. (1998) A study of regional aerosol radiative properties and effects on  ultraviolet-B radiation.
 17            J. Geophys. Res.  [Atmos.] 103: 17,083-17,097.
 18      Wesselink, L. G.; Meiwes, K.-J.; Matzner, E.; Stein, A. (1995) Long-term changes in water and soil chemistry in
 19            spruce and beech forests, Soiling, Germany. Environ. Sci. Technol. 29:  51-58.
 20      Westman, W. E. (1977) How much are nature's services worth? Measuring the social benefits of ecosystem
 21            functioning is both controversial and illuminating. Science (Washington, DC) 197: 960-964.
 22      White, W. H. (1976) Reduction of visibility by sulphates in photochemical smog.  Nature (London) 264: 735-736.
 23      White, W. H. (1997) Deteriorating air or improving measurements? On interpreting concatenate time series.
 24            J. Geophys. Res.  [Atmos.] 102: 6813-6821.
 25      White, W. H.; Macias, E. S. (1990) Light scattering by haze and dust at Spirit Mountain, Nevada. In: Mathai, C. V.,
 26            ed. Visibility and fine particles: an A&WMA/EPA international specialty conference; October 1989; Estes
 27            Park, CO. Pittsburgh, PA; Air & Waste Management Association; pp. 914-922. (A&WMA transactions
 28            series no. TR-17).
 29      White, W. H.; Seigneur, C.; Heinold, D. W.; Eltgroth, M. W.; Richards, L. W.; Roberts, P. T.; Bhardwaja, P. S.;
 30            Conner, W. D.; Wilson, W. E., Jr. (1985) Predicting the visibility of chimney plumes: an intercomparison of
 31            four models with observations at a well-controlled power plant. Atmos. Environ. 19: 515-528.
 32      White, W. H.; Macias, E. S.; Nininger, R. C.; Schorran, D. (1994) Size-resolved measurements of light scattering by
 33            ambient particles in the southwestern U.S.A. Atmos. Environ. 28: 909-921.
 34      White, W. H.; Macias, E. S.; Vasconcelos, L. A. de P.; Farber, R. J.; Mirabella, V. A.;  Green, M. C.; Pitchford,
 35            M. L. (1999) Tracking regional background in a haze attribution experiment. J. Air Waste Manage. Assoc.
 36            49:599-602.
 37      Williams, M. W.; Baron, J. S.; Caine, N.; Sommerfeld, R.; Sanford, R. (1996) Nitrogen saturation in the
 38            Rocky Mountains. Environ. Sci. Technol. 30: 640-646.
 39      Wilson, E. O. (1997) The creation of biodiversity. In: Raven, P. H., ed. Nature and human society: the quest for a
 40            sustainable world: proceedings of the 1997 forum on biodiversity; October; sponsored by the National
 41            Research Council. Washington, DC: National Academy Press.
 42      Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
 43            epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
 44      Winner, W. E.; Atkinson, C. J. (1986) Absorption of air pollution by plants, and consequences for growth.
 45            Trends Ecol. Evol. 1: 15-18.
 46      Winner, W. E.; Bewley,  J. D. (1978a) Terrestrial mosses as bioindicators of SO2 pollution stress: synecological
47            analysis and the index of atmospheric purity. Oecologia 35: 221-230.
 48      Winner, W. E.; Bewley,  J. D. (1978b) Contrasts between bryophyte and vascular plant  synecological responses in
49            an SO2-stressed white spruce association in central Alberta. Oecologia 33: 311-325.
 50      Wittenburg, C.; Dannecker, W. (1992) Dry deposition and deposition velocity of airborne acidic species upon
 51            different sandstones. In: Proceedings of the 1992 European aerosol conference;  September; Oxford,
 52            United Kingdom. J. Aerosol Sci. 23(suppl. 1): S869-S872.
 53      Wolff, G. T.; Collins, D. C.; Rodgers, W. R.; Verma, M. H.; Wong, C. A. (1990) Spotting of automotive finishes
 54            from the interactions between dry deposition of crustal material and wet deposition of sulfate. J. Air Waste
 55            Manage. Assoc. 40: 1638-1648.
         March 2001
4-187         DRAFT-DO NOT QUOTE OR CITE

-------
 1       World Health Organization. (1997) Nitrogen oxides. 2nd ed. Geneva, Switzerland: World Health Organization.
 2             (Environmental health criteria 188).
 3       World Meteorological Organization. (1988) Developing policies for responding to climatic change: a summary of
 4             the discussions and recommendations of workshops; September-October 1987; Villach, Austria; and
 5             November 1987; Bellagio, Austria. Geneva, Switzerland: World Meteorological Organization; report no.
 6             WMO/TD; no. 225. [World Climate Impact Programme series report no. WCIP-1].
 7       World Meteorological Organization. (1999) Scientific assessment of ozone depletion: 1998. Geneva, Switzerland:
 8             World Meteorological Organization, Global Ozone and Monitoring Project; report no. 44.
 9       World Resources Institute. (2000) World resources 2000-2001: people and ecosystems: the fraying web of life.
10             Washington, DC: World Resources Institute.
11       Wu, P.-M.; Okada, K. (1994) Nature of coarse nitrate particles in the atmosphere-a single particle approach.
12             Atmos. Environ. 28:2053-2060.
13       Wu, X. A.; Seigneur, C.; Bergstrom, R. W. (1996) Evaluation of a sectional representation of size distributions for
14             calculating aerosol optical properties. J. Geophys. Res. [Atmos.] 101: 19,277-19,283.
15       Xu, J. R.; Balik, C. M. (1989) Infrared attenuated total reflectance study of latex paint films exposed to aqueous
16             sulfur dioxide. J. Appl. Polym. Sci. 38:173-183.
17       Yanai, R. D.; Siccama, T. G.; Arthur, M.. A.; Federer, C. A.; Friedland, A. J.  (1999) Accumulation and depletion
18             of base cations in forest floors in the northeastern United States. Ecology 80: 2774-2787.
19       Yerrapragada, S. S.; Jaynes, J. H.; Chirra, S. R.; Gauri, K. L. (1994) Rate of weathering of marble due to dry
20             deposition of ambient sulfur and nitrogen dioxides. Anal. Chem. 66: 655-659.
21       Yocom, J. E.; Grappone, N. (1976) Effects of power plant emissions on materials. Palo Alto, CA: Electric Power
22             Research Institute; report no. EPRI/EC-139. Available from: NTIS, Springfield, VA; PB-257 539.
23       Yocom, J. E.; Upham, J. B. (1977) Effects of economic materials and structures. In: Stem, A. C., ed. Air pollution.
24             3rd ed. Volume II. New York, NY: Academic Press; pp. 93-94.
25       Young, P. (1996) Pollution-fueled "biodeterioration" threatens historic stone: researchers are probing the role
26             microorganisms play in the decay of buildings and artwork. Environ. Sci. Technol. 30: 206A-208A.
27       Young, J. R.; Ellis, E. C.; Hidy, G. M. (1988) Deposition of air-borne acidifiers in the western environment.
28             J. Environ. Qual. 17:1-26.
29       Youngs, W. D.; Rutzke, M.; Gutenmann, W. H.; Lisk, D. J.  (1993) Nickel and vanadium in foliage in the vicinity of
30             an oil-fired power plant. Chemosphere 27:  1269-1272.
31       Zannetti, P.; Tombach, I.; Balson, W. (1990) Calculation of visual range improvements from SO2 emission
32             controls—I. semi-empirical methodology. Atmos. Environ. Part A 24: 2361-2368.
33       Zannetti, P.; Tombach, I.; Cvencek, S.; Balson, W. (1993) Calculation of visual range improvements from SO2
34             emission controls-II. an application to the  eastern United States. Atmos. Environ. Part A 27:1479-1490.
35       Zappia, G.; Sabbioni, C.; Gobbi, G. (1993) Non-carbonate carbon content on black and white areas of damaged
3 6             stone monuments. Atmos. Environ. 27A: 1117-1121.
37       Zappia, G.; Sabbioni, C.; Pauri, M. G.; Gobbi, G. (1994) Mortar damage due to airborne sulfur compounds in a
38             simulation chamber. Mater. Struct. 27: 469-473.
39       Zappia, G.; Sabbioni, C.; Riontino, C.; Gobbi, G.; Favoni, O. (1998) Exposure tests of building materials in urban
40             atmosphere. Sci. Total Environ. 224: 235-244.
41       Zepp, R. G.; Callaghan, T. V.; Erickson, D. J. (1998) Effects of enhanced solar ultraviolet radiation on
42             biogeochemical cycles. J. Photochem. Photobiol. B 46: 69-82.
43       Zhang, X. Q.; McMurry, P. H.; Hering, S. V.; Casuccio, G. S. (1993) Mixing characteristics and water content of
44             submicron aerosols measured in Los Angeles and at the Grand Canyon. Atmos. Environ. Part A
45             27:1593-1607.
46       Zhang, X.; Turpin, B. J.; McMurry, P. H.; Hering, S. V.; Stolzenburg, M. R. (1994) Mie theory evaluation of
47             species contributions to 1990 wintertime visibility reduction in the Grand Canyon. Air Waste 44: 153-162.
48
         March 2001
4-188
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
                               APPENDIX 4A

           Excerpted Key Points from the Executive Summary of the
                    World Meteorological Organization 1998
                  Assessment of Stratospheric Ozone Depletion
                   (World Meteorological Organization, 1999)

     Among the provisions of the 1987 Montreal Protocol on Substances that Deplete the Ozone
Layer was the requirement that the Parties to the Protocol base their future decisions on available
scientific, environmental, technical, and economic information, as assessed by worldwide expert
communities. Advances in the understanding of ozone science over this decade were assessed in
1988, 1989, 1991, and 1994. This information was input to the subsequent Amendments and
Adjustments of the 1987 Protocol. The 1998 assessment summarized below is the fifth in that
series.

Recent Major Scientific Findings and Observations
     Since the Scientific Assessment of Ozone Depletion:  1994, significant advances have
continued to be made in understanding of the impact of human activities on the ozone layer, the
influence of changes in chemical composition on the radiative balance of the Earth's climate, and,
indeed, the coupling of the ozone layer and the climate system. Numerous laboratory
investigations, atmospheric observations, and theoretical and modeling studies have produced
several key ozone- and climate-related findings that are discussed below.

• The total combined abundance of ozone-depleting compounds in the lower atmosphere
 peaked in about 1994 and now is slowly declining. Total chlorine is declining, but total
 bromine is still increasing. As forecast in the 1994 Assessment, the long period of increasing
 total chlorine abundances—primarily from the chlorofluorocarbons (CFCs), carbon
 tetrachloride (CC14), and methyl chloroform (CH3CCI3)—has ended. The peak total
 tropospheric chlorine abundance was 3.7 ± 0.1 parts per billion (ppb) between mid-1992 and
 mid-1994. The declining abundance of total chlorine results principally from reduced
 emissions of methyl chloroform.  Chlorine from the major CFCs is still increasing slightly. The
       March 2001
                                       4A-1
DRAFT-DO NOT QUOTE OR CITE

-------
 1      abundances of most of the halons continue to increase (for example, Halon-1211, almost 6%
 2      per year in 1996), but the rate has slowed in recent years. These halon increases likely are
 3      caused by emissions in the 1990s from the halon "bank," largely in developed countries, and
 4      new production of halons in developing countries. The observed abundances of CFCs and
 5      chlorocarbons hi the lower atmosphere are consistent with reported emissions.
 6     • The observed abundances of the substitutes for the CFCs are increasing.  Abundances of
 7      the hydrochlorofluorocarbons (HCFCs) and hydrofluorocarbons (HFCs) are increasing as a
 8      result of continuation of earlier uses and of their use as substitutes for the CFCs. hi 1996, the
 9      HCFCs contributed about 5% to the tropospheric chlorine from the long-lived gases. This
10      addition from the substitutes offsets some of the decline in tropospheric chlorine associated
11      with methyl chloroform, but is still about 10 times less than that from the total tropospheric
12      chlorine growth rate throughout the 1980s.  The atmospheric abundances of HCFC-141b and
13      HCFC-142b calculated from reported emissions data are factors of 1.3 and 2, respectively,
14      smaller than observations. Observed and calculated abundances agree for HCFC-22 and
15      HFC-134a.
16     • The combined abundance of stratospheric chlorine and bromine is expected to peak
17      before the year 2000. The delay in this peak in the stratosphere compared with the lower
18      atmosphere reflects the average time required for surface emissions to reach the  lower
19      stratosphere.  The observations of key chlorine compounds in the stratosphere up through the
20      present show the expected slower rate of increase and show that the peak had not occurred at
21      the time of the most recent observations that were analyzed for this assessment.
22     • The role of methyl bromide as an ozone-depleting compound is now considered to be less
23      than was estimated in the 1994 Assessment, although significant uncertainties remain.
24      The current best estimate of the Ozone Depletion Potential (ODP) for methyl bromide (CH3Br)
25      is 0.4, compared with an ODP of 0.6 estimated previously.  The change is caused primarily by
26      both an increase in estimates of ocean removal processes and identification of an uptake by
27      soils, with a smaller contribution from change in our estimate of the atmospheric removal rate.
28      Recent research has shown that the science of atmospheric methyl bromide is complex and still
29      not well understood.  Current understanding of the sources and sinks of atmospheric methyl
30      bromide is incomplete.
31
       March 2001
4A-2
DRAFT-DO NOT QUOTE OR CITE

-------
  1      • The rate of decline in stratospheric ozone at mid-latitudes has slowed; hence, the
  2       projections of ozone loss made in the 1994 assessment are larger than what has actually
  3       occurred. Total column ozone decreased significantly at mid-latitudes (25 to 60°) between
  4       1979 and 1991, with estimated linear downward trends of 4.0, 1.8, and 3.8% per decade,
  5       respectively, for northern mid-latitudes in winter/spring, northern mid-latitudes in summer/fall,
  6       and southern mid-latitudes year round.  However, since 1991, the linear trend observed during
  7       the 1980s has not continued, rather, total column ozone has been almost constant at mid-
  8       latitudes in both hemispheres since the recovery from the 1991 Mt. Pinatubo eruption. The
  9       observed total column ozone losses from  1979 to the period 1994 to 1997 are about 5.4, 2.8,
10       and 5.0%, respectively, for northern mid-latitudes in winter/spring, northern mid-latitudes in
11       summer/fall, and southern mid-latitudes year round, rather than the values projected in the  1994
12       assessment assuming a linear trend: 7.6, 3.4, and 7.2%, respectively. Understanding of how
13       changes in stratospheric chlorine/bromine and aerosol loading affect ozone suggests some of
14       the reasons for the unsuitability of using a linear extrapolation of the pre-1991 ozone trend to
15       the present.
16      • The  springtime Antarctic ozone hole continues unabated. The extent of ozone depletion has
17       remained essentially unchanged since the  early 1990s. This behavior is expected given the
18       near-complete destruction of ozone within the Antarctic lower stratosphere during springtime.
19       The factors contributing to the continuing depletion are well understood.
20      • The  link between the long-term buildup of chlorine and the decline of ozone in the upper
21       stratosphere has been firmly established.  Model predictions based on the observed buildup
22       of stratospheric chlorine in the upper stratosphere indicate a depletion of ozone that is in good
23       quantitative agreement with  the altitude and latitude dependence of the measured ozone decline
24       during the past several decades, which peaks at about 7% per decade near 40 km at mid-
25       latitudes in both hemispheres.
26      • The  late-winter/spring ozone values in the Arctic were unusually low in six out of the last
27       nine years, the six being years that are characterized by unusually cold and protracted
28       stratospheric winters. The possibility of such depletions was predicted in the 1989
29       assessment. Minimum Arctic vortex temperatures are near the threshold for large chlorine
30       activation. Therefore, the year-to-year variability in temperature, which is driven by
31       meteorology, leads to particularly large variability in ozone for current chlorine loading. As a
        March 2001
4A-3
DRAFT-DO NOT QUOTE OR CITE

-------
 1       result, it is not possible to forecast the behavior of Arctic ozone for a particular year. Elevated
 2       stratospheric halogen abundances over the next decade or so imply that the Arctic will continue
 3       to be vulnerable to large ozone losses.
 4     • The understanding of the relation between increasing surface UV-B radiation and
 5       decreasing column ozone has been further strengthened by ground-based observations,
 6       and newly developed satellite methods show promise for establishing global trends in UV
 7       radiation. The inverse dependence of surface UV radiation and the overhead amount of ozone,
 8       which was demonstrated in earlier assessments, has been further demonstrated and quantified
 9       by ground-based measurements under a wide range of atmospheric conditions. In addition, the
10       influences of other variables, such as clouds, particles, and surface reflectivity, are better
11       understood. These data have assisted the development of a satellite-based method to estimate
12       global UV changes, taking into account the role of cloud cover. The satellite estimates for 1979
13       through 1992 indicate that the largest UV increases occur during spring at high latitudes in both
14       hemispheres.
15     • Stratospheric ozone losses have caused a cooling of the global lower stratosphere and
16       global-average negative radiative forcing of the climate system. The decadal temperature
17       trends in the stratosphere have now been better quantified.  Model simulations indicate that
18       much of the observed downward trend in lower stratospheric temperatures (about 0.6 °C per
19       decade from 1979 to 1994) is attributed to the ozone loss in the lower stratosphere.  A lower
20       stratosphere that is cooler results in less infrared radiation reaching the surface/troposphere
21       system. Radiative calculations, using extrapolations based on the ozone trends reported in the
22       1994 assessment for reference, indicate that stratospheric ozone losses since 1980 may have
23       offset about 30% of the positive forcing because of increases in the well-mixed greenhouse
24       gases (i.e., carbon dioxide, methane, nitrous oxide, halocarbons) over the same time period. The
25       climatic impact of the slowing of mid-latitude ozone trends and the enhanced ozone loss in the
26       Arctic has not yet been assessed.
27     • Based on past emissions of ozone-depleting substances and a projection of the maximum
28       allowances under the Montreal Protocol into the future, the maximum ozone depletion is
29       estimated to lie within the current decade or the next two decades, but its identification
30       and the evidence for the recovery of the ozone layer lie still further ahead. The falloff of
31       total chlorine and bromine abundances in the stratosphere in the next century will be much
       March 2001
4A-4
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14  .
15
16
17
  slower than the rate of increase observed in past decades, because of the slow rate at which
  natural processes remove these compounds from the stratosphere. The most vulnerable period
  for ozone depletion will be extended into the coming decades. However, extreme
  perturbations, such as natural events like volcanic eruptions, could enhance the loss from
  ozone-depleting chemicals. Detection of the beginning of the recovery of the ozone layer could
  be achievable early in the next century if decreasing chlorine and bromine abundances were the
  only factor. However, potential future increases or decreases in other gases important in ozone
  chemistry (such as nitrous oxide, methane, and water vapor) and climate change will influence
  the recovery of the ozone layer.  When combined with the natural variability of the ozone layer,
  these factors imply that unambiguous detection of the beginning of the recovery of the ozone
  layer is expected to be well after the maximum stratospheric loading of ozone-depleting gases.
REFERENCE
World Meteorological Organization (WMO). (1999) Scientific assessment of ozone depletion: 1998. Geneva,
     Switzerland: World Meteorological Organization, Global Ozone and Monitoring Project; report no. 44.
       March 2001
                                         4A-5
DRAFT-DO NOT QUOTE OR CITE

-------

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
                               APPENDIX 4B

  Excerpted Key Points from the Executive Summary of the United Nations
  Environmental Programme 1998 Assessment of Environmental Effects of
     Ozone Depletion (United Nations Environmental Programme, 1998)

     Decreased quantities of total-column ozone now are observed over large parts of the globe,
permitting increased penetration of solar UV-B radiation (280 to 315 nm) to the Earth's surface.
The present assessment deals with the possible consequences. The Atmospheric Science Panel
predicts that the ozone layer will be in its most vulnerable state during the coming two decades.
Some of the effects are expected to occur during most of the next century. Recent studies show
that the effects of ozone depletion would have been dramatically worse without protective
measures taken under the 1987 Montreal Protocol. The assessment is given in seven papers,
summarized below:

(1)  Changes in Ultraviolet Radiation
• Stratospheric ozone levels are near their lowest points since measurements began, so
 current UV-B radiation levels are thought to be close to their maximum.  Total
 stratospheric content of ozone-depleting substances is expected to reach a maximum before the
 year 2000. All other things being equal, the current ozone losses and related UV-B increases
 should be close to their maximum. Increases in surface erythemal (sunburning) UV radiation
 relative to the values in the  1970s are estimated to be
     • about 7% at Northern Hemisphere mid-latitudes in winter/spring;
     • about 4% at Northern Hemisphere mid-latitudes in summer/fall;
     • about 6% at Southern Hemisphere mid-latitudes on a year-round basis;
     • about 130% in the Antarctic in the spring; and
     • about 22% in the Arctic in the spring.
• The correlation between increases in surface UV-B radiation and decreases in overhead
 ozone has been demonstrated further and quantified by ground-based instruments under
 a wide range of conditions. Improved measurements of UV-B radiation are now providing
       March 2001
                                       4B-1
DRAFT-DO NOT QUOTE OR CITE

-------
 1     better geographical and temporal coverage. Surface UV-B radiation levels are highly variable
 2     because of sun angle, cloud cover, and, also, because of local effects, including pollutants and
 3     surface reflections. With a few exceptions, the direct detection of UV-B trends at low- and
 4     mid-latitudes remains problematic because of this high natural variability, the relatively small
 5     ozone changes, and the practical difficulties of maintaining long-term stability in networks of
 6     UV-measuring instruments.  Few reliable UV-B radiation measurements are available from
 7     pre-ozone-depletion days.
 8     • Satellite-based observations of atmospheric ozone and clouds are being used, together
 9       with models of atmospheric transmission, to provide global coverage and long-term
10       estimates of surface UV-B radiation. Estimates of long-term (1979 to 1992) trends in zonally
11       averaged UV irradiances that include cloud effects are nearly identical to those for clear-sky
12       estimates, providing evidence that clouds have not influenced the UV-B trends.  However, the
13       limitations of satellite-derived UV estimates should be recognized.  To assess uncertainties
14       inherent in this approach, additional validations involving comparisons with ground-based
15       observations are required.
16     • Direct comparisons of ground-based UV-B radiation measurements between a few
17       mid-latitude sites in the Northern and Southern Hemispheres have shown larger
18       differences than those estimated using satellite data.  Ground-based measurements show that
19       summertime erythemal UV irradiances in the Southern Hemisphere exceed those at comparable
20       latitudes of the Northern Hemisphere by up to 40%, whereas corresponding satellite-based
21       estimates yield only 10 to 15% differences.  Atmospheric pollution may be a factor in this
22       discrepancy between ground-based measurements and satellite-derived estimates.  UV-B
23       measurements at more sites are required to determine whether the larger observed differences
24       are globally representative.
25     • High levels of UV-B radiation continue to be observed in Antarctica during the recurrent
26       spring-time ozone hole. For example, during ozone-hole episodes, measured biologically
27       damaging radiation at Palmer Station, Antarctica (64 °S) has been found to approach and
28       occasionally even exceed maximum summer values at San Diego, CA (32 °N).
29     • Long-term predictions of future UV-B levels are difficult and uncertain. Nevertheless,
30       current best estimates suggest that a slow recovery to pre-ozone-depletion levels may be
31       expected during the next half-century. Although the maximum ozone depletion, and hence
        March 2001
4B-2
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
maximum UV-B increase, is likely to occur in the current decade, the ozone layer will continue
to be in its most vulnerable state into the next century.  The peak depletion and the recovery
phase could be delayed by decades because of interactions with other long-term atmospheric
changes (e.g., increasing concentrations of greenhouse gases).  Other factors that could influence
the recovery include nonratification or noncompliance with the Montreal Protocol and its
Amendments and Adjustments and future volcanic eruptions. The recovery phase for surface
UV-B irradiances probably will not be detectable until many years after the ozone minimum.

(2) Effects on Human and Animal Health
• Recent estimates suggest that the increase in the risk of cataract and skin cancer because
  of ozone depletion would not have been controlled adequately by implementation of the
  Montreal Protocol (1987) alone, but can be achieved through implementation of its later
  provisions. Risk assessments for the  United States and northwestern Europe indicate large
  increases in cataracts and skin cancers under either the "no Protocol" or the early Montreal
  Protocol scenarios. Under scenarios based on later amendments (Copenhagen, 1992) and
  Montreal (1997), increases in cataracts and skin cancer attributable  to ozone depletion return
  almost to zero  by the end of the next century.
• The increases in UV-B radiation associated with ozone depletion are likely to lead to
  increases in the incidence or severity of a variety of short- and long-term health effects, if
  current exposure practices are not modified by changes in behavior.
• Adverse effects on the eye will affect all populations irrespective of skin color.  Adverse
  impacts could include more cases of acute reactions such as "snowblindness", increases in
  cataract incidence or severity (and thus the incidence of cataract-associated blindness), and
  increases in the incidence (and mortality) from ocular melanoma and squamous cell carcinoma
  of the eye.
• Effects on the immune system also will affect all populations but may be both adverse and
  beneficial. Adverse effects include depressed resistance to certain tumors and infectious
  diseases, potential impairment of vaccination responses, and possibly increased severity of
  some autoimmune and allergic responses. Beneficial effects could include decreases in the
  severity of certain immunologic disease conditions, such as psoriasis and nickel allergy.
       March 2001
                                         4B-3
DRAFT-DO NOT QUOTE OR CITE

-------
 1      • Effects on the skin could include increases in photoaging and skin cancer with risk
 2       increasing with fairness of skin.  Increases in UV-B are likely to accelerate the rate of
 3       photoaging, as well as increase the incidence (and associated mortality) of melanoma and
 4       nonmelanoma skin cancer, basal cell carcinoma, and squamous cell carcinoma.
 5      • Research is generating much new information being used to help reduce uncertainties
 6       associated with current risk estimates. Evaluation of the impact of susceptibility genes is
 7       helping to identify highly susceptible populations so that their special risk can be assessed.
 8       Examination of the impacts of behavior changes, such as consuming diets that are high in
 9       antioxidants, avoiding sun exposure during the 4 h around solar noon, and wearing of covering
10      apparel (e.g., hats, sunglasses), is beginning to identify important exposure patterns, as well as
11       possible mitigation strategies.
12      • Quantitative risk assessments for a variety of other  effects, such as UV-B-induced
13       immunosuppression of infectious diseases, are not yet possible.  New information continues
14       to confirm the reasonableness of these concerns, but data that is adequate for quantitative risk
15       assessment are not yet available.
16
17      (3) Effects on Terrestrial Ecosystems
18      • Increased UV-B can be damaging for terrestrial organisms including plants and
19       microbes, but all these organisms also have protective and repair processes. The balance
20       between damage and protection varies among species and even varieties of crop species; many
21       species and varieties can accommodate increased UV-B. Tolerance of elevated UV-B by some
22       species and crop varieties provides opportunities for genetic engineering and breeding to deal
23       with potential crop-yield reductions because of elevated UV-B in agricultural systems.
24      • Research in the past few years indicates that increased UV-B exerts effects more often
25       through altered patterns of gene activity rather than damage. These UV-B effects on
26       regulation manifest themselves in many ways including changes in life-cycle timing, changes hi
27       plant form, and production of plant chemicals not directly involved in primary metabolism.
28       These plant chemicals play a role in protecting plants from pathogens and insect attack and
29       affect food quality for humans and grauine animals.
30      • Terrestrial ecosystem responses to increased UV-B  are evident primarily in interactions
31       among species, rather than in the performance of individual species. Much of the recent
        March 2001
4B-4
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
  experimentation indicates that increased UV-B affects the balance of competition among higher
  plants, the degree to which higher plants are consumed by insects, and susceptibility of plants to
  pathogens. These effects can be mediated, in large part, by changes in plant form and
  chemistry, but effects of UV-B on insects and microbes are also possible.  The direction of
  these UV-B mediated interactions among species is often difficult to predict based only on
  single-organism responses to increased UV-B.
• Effects of increased UV-B radiation can accumulate from year to year in long-lived
  perennial plants and from generation to generation in annual plants.  This effect has been
  shown in a few recent studies, but the generality of this accumulation among species is not
  presently known. If this phenomenon is widespread, this would amplify otherwise subtle
  responses to UV-B seen in a single growing season, for example, in forest trees.
• Effects of increased UV-B must be taken into account together with other environmental
  factors including those associated with global change. Responses of plants and other
  organisms to increased  UV-B are modified by other environmental factors (e.g., CO, water
  stress, mineral nutrient  availability, heavy metals, temperature). Many of these factors also are
  changing as the global climate is altered.

(4) Effects on Aquatic Ecosystems
• Recent studies continue to demonstrate that solar UV-B and UV-A have adverse effects on
  the growth, photosynthesis, protein and pigment content, and reproduction of
  phytoplankton, thus affecting the food web. These studies have determined biological
  weighting functions and exposure-response curves for phytoplankton and have developed new
  models for the estimation of UV-related photoinhibition. In spite of this increased
  understanding and enhanced ability to model aquatic impacts, considerable uncertainty remains
  with respect to quantifying effects of ozone-related UV-B increases at the ecosystem level.
• Macroalgae and sea grasses show a pronounced sensitivity to solar UV-B. They are
  important biomass producers in aquatic ecosystems. Most of these organisms are attached and
  so cannot avoid being exposed to solar radiation at their growth site. Effects have been found
  throughout the top 10 to 15 m of the water column.
• Zooplankton communities, as well as other aquatic organisms including sea urchins,
  corals, and amphibians, are sensitive to UV-B. There is evidence that, for some of these
       March 2001
                                         4B-5
DRAFT-DO NOT QUOTE OR CITE

-------
 1      populations, even current levels of solar UV-B radiation, acting in conjunction with other
 2      environmental stresses, may be a limiting factor, but quantitative evaluation of possible effects
 3      remains uncertain.
 4     • UV-B radiation is absorbed by and breaks down dissolved organic carbon (DOC) and
 5      participate organic carbon  (POC) and makes the products available for bacterial
 6      degradation and remineralization. The degradation products are of importance in the cycling
 7      of carbon in aquatic ecosystems. Because UV-B breaks down DOC as it is absorbed, increase
 8      in UV-B can increase the penetration of both UV-B and UV-A radiation into the water column.
 9      As a consequence, the quantity of UV-B penetrating to a given depth both influences and is
10      influenced by DOC.  Warming and acidification result in faster degradation of these substances
11      and, thus, enhance the penetration of UV radiation into the water column.
12     • Polar marine ecosystems, where ozone-related UV-B increases are the greatest, are
13      expected to be the oceanic ecosystems most influenced by ozone depletion. Oceanic
14      ecosystems are characterized by large spatial and temporal variabilities that make it difficult to
15      select out UV-B-specific effects on single species or whole phytoplankton communities.
16      Although estimates of reduction in both Arctic and Antarctic productivity are based on
17      measurable short-term effects, there remain considerable uncertainties in estimating long-term
18       consequences, including possible shifts in community structure. Reduced productivity of fish
19       and other marine crops could have an economic impact, as well as affect natural predators;
20      however, quantitative estimation of the possible effects of reduced production remain
21       controversial.
22     • Potential consequences of enhanced levels of exposure of aquatic ecosystems to UV-B
23       radiation include reduced uptake capacity for atmospheric carbon dioxide (CO2),
24       resulting in the potential augmentation of global warming. The oceans play a key role with
25       respect to the budget of greenhouse gases. Marine phytoplankton are a major sink for
26       atmospheric CO2 and they have a decisive role in the development of future trends of CO2
27       concentrations in the atmosphere. The relative importance of the net uptake of CO2 by the
28       biological pump and the possible role of increased UV-B in the ocean are still controversial.
29
30
        March 2001
4B-6
DRAFT-DO NOT QUOTE OR CITE

-------
  1      (5) Effects on Biogeochemical Cycles
  2      • Effects of increased UV-B on emissions of CO2 and carbon monoxide (CO) and on
  3       mineral nutrient cycling in the terrestrial biosphere have been confirmed by recent
  4       studies of a range of species and ecosystems. The effects, both in magnitude and direction, of
  5       UV-B on trace gas emissions and mineral nutrient cycling are species specific and operate on a
  6       number of processes. These processes include changes in the chemical composition in living
  7       plant tissue; photodegradation (breakdown by light) of dead plant matter, including litter;
  8       release of CO from vegetation previously charred by fire; changes in the communities of
  9       microbial decomposers; and effects on nitrogen-fixing micro-organisms and plants. Long-term
 10       experiments are in place to examine UV-B effects on carbon capture and storage in biomass
 11       within natural terrestrial ecosystems.
 12      • Studies in natural aquatic ecosystems have indicated that organic matter is the primary
 13       regulator of UV-B penetration. Enhanced UV-B can affect the balance between the
 14       biological processes that produce the organic matter and the chemical and microbial processes
 15       that degrade it. Changes in the balance have broad impacts on the effects of enhanced UV-B on
 16       biogeochemical cycles. These changes, which are reinforced by changes in climate and
 17       acidification, result from clarification of the water and changes in light quality.
 18      • Increased UV-B has positive and negative impacts on microbial activity in aquatic
 19       ecosystems that can affect carbon and mineral nutrient cycling, as well as the uptake and
20       release of greenhouse and chemically reactive gases. Photoinhibition of surface aquatic
21       micro-organisms by UV- B can be offset partially by photodegradation of dissolved organic
22       matter to produce substrates, such as organic acids and ammonium, that stimulate microbial
23       activity.
24      • Modeling and experimental approaches are being developed to predict and measure the
25       interactions and feedbacks between climate change in UV-B-induced changes in marine
26       and terrestrial biogeochemical cycles. These interactions include alterations in the oxidative
27       environment in the upper ocean and in the marine boundary layer and oceanic production and
28       release of CO, volatile organic compounds (VOC), and reactive oxygen species (ROS, such as
29       hydrogen peroxide and hydroxyl radicals). Climate-related changes in temperature and water
30       supply in terrestrial ecosystems interact with UV-B radiation through biogeochemical processes
31       operating on a wide range of time scales.
       March 2001
4B-7
DRAFT-DO NOT QUOTE OR CITE

-------
 1     (6)  Effects on Air Quality
 2     • Increased UV-B will increase the chemical activity in the lower atmosphere (the
 3      troposphere). Troposphere ozone levels are sensitive to local concentrations of nitrogen
 4      oxides (NOx) and hydrocarbons. Model studies suggest that additional UV-B radiation reduces
 5      tropospheric ozone in clean environments (low NOX) and increases tropospheric ozone in
 6      polluted areas (high NOJ.
 7     • Assuming other factors remain constant, additional UV-B will increase the rate at which
 8      primary pollutants are removed from the troposphere. Increased UV-B is expected to
 9      increase the concentration of hydroxyl radicals (OH) and result in faster removal of pollutants.
10      Increased concentrations of oxidants such as hydrogen peroxide and organic peroxides also are
11      expected. The effects of UV-B increases on tropospheric ozone, OH, methane, CO, and
12      possibly other tropospheric constituents, although not negligible, will be difficult to detect
13      because the concentrations of these species also are influenced by many other variable factors
14      (e.g., emissions).
15     -No significant effects on humans or the environment have been identified from
16      trifluoroacetic acid (TFA) produced by atmospheric degradation of HCFCs and MFCs.
17      Numerous studies have shown that TFA has, at most, moderate short-term toxicity. Insufficient
18      information is available to assess potential chronic, developmental, or reproductive effects.  The
19      atmospheric degradation mechanisms of most substitutes for ozone-depleting substances are
20      well established. HCFCs and HFCs are two important classes of substitutes.  Atmospheric
21      degradation of HCFC-123 (CF3CHC12), HCFC-124 (CF3CHFCI), and H FC-I34a (CF3CH2F)
22      produces TFA.  Reported measurements of TFA in rain, rivers, lakes, and oceans show it to be
23      an ubiquitous component of the hydrosphere, present at levels much higher than can be
24      explained by currently reported sources.  The levels of TFA currently produced by the
25      atmospheric degradation of HFCs and HCFCs are estimated to be orders of magnitude below
26      those of concern and make only a minor contribution to the current environmental burden of
27      TFA.
28
29      (7) Effects on Materials
30      • Physical and mechanical properties of polymers are affected negatively by increased
31      UV-B in sunlight. Increased UV-B reduces the useful lifetimes of synthetic polymer products
        March 2001
4B-8
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
  used outdoors and of biopolymer materials such as wood, paper, wool, and cotton. The
  reduction in service life of materials depends on the synergistic effect of increased UV-B and
  other factors, especially the temperature of the material during exposure to sunlight. Even
  under harsh UV exposure conditions, the higher temperatures largely determine the extent of
  increased UV-induced damage to photostabilized polyethylenes. However, accurate assessment
  of such damage to various materials is presently difficult to make because of limited availability
  of technical data, especially on the relationship between the dose of UV-B radiation and the
  resulting damage of the polymer or other material.
• Conventional photostabilizers are likely to be able to mitigate the effects of increased UV
  levels in sunlight. More effective photostabilizers for plastics have been commercialized in
  recent years. The use of these compounds allows plastic polymer products to be used in a wide
  range of different UV environments found worldwide. It is reasonable to expect existing
  photostabilizer technologies to be able to mitigate these effects of an increased UV-B on
  polymer materials. This, however, would increase the cost of the relevant polymer products,
  surface coatings, and treated biopolymer materials. However, the efficiencies of even the
  conventional photostabilizers under the unique exposure environments resulting from an
  increase in solar UV-B have not been well studied.
REFERENCE
United Nations Environment Programme (UNEP). (1998) Environmental effects of ozone depletion: 1998
     assessment. J. Photochem. Photobiol. B 46: 1-4.
       March 2001
                                         4B-9
DRAFT-DO NOT QUOTE OR CITE

-------

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
                               APPENDIX 4C
   Excerpted Key Points from the Executive Summary of the Special Report
    of the International Panel on Climate Change Working Group II on the
    Regional Impacts of Climate Change: An Assessment of Vulnerability
     Excerpts from Executive Summary materials from a Special Report of the IPCC Working
Group II, The Regional Impacts of Climate Change: An Assessment of Vulnerability (IPCC,
1998) are incorporated below in this appendix to provide an overview of key points regarding the
vulnerability of aquatic and terrestrial ecosystems, water resources, agriculture, and human
habitability in North America to climate change.

Scope of the Assessment
     The report was prepared at the request of the Conference of the Parties to the United
Nations Framework Convention on Climate Change (UNFCCC) and its subsidiary bodies
(specifically, the Subsidiary Body for Scientific and Technological Advice-SBSTA).  The special
report provides, on a regional basis, a review of state-of-the-art information on the vulnerability
to potential changes in climate of ecological systems, socioeconomic sectors (agriculture,
fisheries, water resources, and human settlements), and human health. The report reviews the
sensitivity of these systems as well as options for adaptation. Though the report draws heavily on
the sectoral impact assessments of the Second Assessment Report (SAR), it also draws on more
recent peer-reviewed literature (inter alia, country studies programs).

Nature of the Issue
     Human activities (primarily burning of fossil fuels and changes in land use and land cover)
are increasing atmospheric concentrations of greenhouse gases, which alter radiative balances
and tend to warm the atmosphere, and, in some regions, aerosols, which have an opposite effect
on radiative balances and tend to cool the atmosphere.  At present, in some locations primarily in
the Northern Hemisphere, the cooling effects of aerosols can be large enough to more than offset
the warming caused by greenhouse gases. Because aerosols do not remain in the atmosphere for
March 2001                             4C-1       DRAFT-DO NOT QUOTE OR CITE

-------
 1     long periods and global emissions of their precursors are not projected to increase substantially,
 2     aerosols will not offset the global long-term effects of greenhouse gases, which are long-lived.
 3     Aerosols can have important consequences for continental-scale patterns of climate change.
 4          These changes in greenhouse gases and aerosols, taken together, are projected to lead to
 5     regional and global changes in temperature, precipitation, and other climate variables, resulting
 6     in global changes in soil moisture; an increase in global mean sea level; and prospects for more
 7     severe extreme high-temperature events, floods, and droughts in some places. Based on the
 8     range of sensitivities of climate to changes  in the atmospheric concentrations of greenhouse gases
 9     (IPCC, 1996; WG I) and plausible changes in emissions of greenhouse gases and aerosols
10     (lS92a-f, scenarios that assume no climate policies), climate models project that the mean annual
11     global surface temperature will increase by 1 to 3.5 °C by 2100, that global mean sea level will
12     rise by 15 to 95 cm, and that changes in the spatial and temporal patterns of precipitation will
13     occur.  The average rate of warming probably would be greater than any seen in the past 10,000
14     years, although the actual annual to decadal rate would include considerable natural variability,
15     and regional changes could differ substantially from the global mean value. These long-term,
16     large-scale, human-induced changes will interact with natural variability on time  scales of days to
17     decades (e.g., the El Nino-Southern Oscillation [ENSO] phenomenon) and, thus,  influence social
18     and economic well-being.  Possible local climate effects caused by unexpected events such as a
19     climate-change-induced change of flow pattern of marine water streams (e.g., the Gulf Stream)
20     have not been considered, because such changes cannot be predicted with confidence at present.
21           Scientific studies show that human health, ecological systems, and socioeconomic sectors
22     (e.g., hydrology and water resources, food and fiber production, coastal systems,  human
23     settlements), all of which are vital to sustainable development, are sensitive to changes in
24     climate, including both the magnitude and rate of climate change, as well as to changes in
25     climate variability.  Whereas many regions are likely to experience adverse effects of climate
26     change, some of which are potentially irreversible, some effects of climate change are likely to be
27     beneficial. Climate change represents an important additional stress on those systems already
28     affected by increasing resource demands, unsustainable management practices, and pollution,
29     which in many cases may be equal to or greater than those of climate change. These stresses will
30     interact in different ways across regions but can be expected to reduce the ability of some
31     environmental systems to provide, on a sustained basis, key goods and services needed for
        March 2001
4C-2
DRAFT-DO NOT QUOTE OR CITE

-------
  1      successful economic and social development, including adequate food, clean air, and water;
  2      energy; safe shelter; low levels of disease; and employment opportunities. Climate change also
  3      will take place in the context of economic development, which may make some groups or
  4      countries less vulnerable to climate change, for example, by increasing the resources available for
  5      adaptation. Those that experience low rates of growth, rapid increases in population, and
  6      ecological degradation may become increasingly vulnerable to potential changes.
  7
  8      Approach of the Assessment
  9           The report assesses the vulnerability of natural and social systems of major regions of the
10      world to climate change. Vulnerability is defined as the extent to which a natural or social
11      system is susceptible to sustaining damage from climate change. Vulnerability is a function of
12      the sensitivity of a system to changes in climate (the degree to which a system will respond to a
13      given change in climate, including both beneficial and harmful effects) and the ability to adapt
14      the system to changes in climate (the degree to which adjustments in practices, processes, or
15      structures can moderate or offset the potential for damage or take advantage of opportunities
16      created because of a given change in climate).  Under this framework, a highly vulnerable system
17      would be one that is highly sensitive to modest changes hi climate, where the sensitivity includes
18      the potential for substantial harmful effects, and one for which the ability to adapt is severely
19      constrained.
20           Because the  available studies have not employed a common set of climate scenarios and
21      methods, and because of uncertainties regarding the sensitivities and adaptability of natural and
22      social systems, the assessment of regional vulnerabilities is necessarily qualitative.  However, the
23      report provides substantial and indispensable information  on what currently is known about
24      vulnerability to climate change.
25           In a number  of instances, quantitative estimates of impacts of climate change are cited in
26      the report. Such estimates  are dependent on the specific assumptions employed regarding future
27      changes in climate, as well as on the particular methods and models applied hi the analyses.
28      In interpreting these estimates, it is important to bear hi mind that uncertainties regarding the
29      character, magnitude, and rates of future climate change remain.  These uncertainties impose
30      limitations on the ability of scientists to project impacts of climate change, particularly at
31      regional and smaller scales.
        March 2001
4C-3
DRAFT-DO NOT QUOTE OR CITE

-------
 1           It is in part because of the uncertainties regarding how climate will change that the report
 2     takes the approach of assessing vulnerabilities rather than assessing quantitatively the expected
 3     impacts of climate change.  The estimates are interpreted best as illustrative of the potential
 4     character and approximate magnitudes of impacts that may result from specific scenarios of
 5     climate change. They serve as indicators of sensitivities and possible vulnerabilities. Most
 6     commonly, the estimates are based on changes in equilibrium climate that have been simulated to
 7     result from an equivalent doubling of carbon dioxide (CO2) in the atmosphere. Usually, the
 8     simulations have excluded the effects of aerosols. Increases in global mean temperatures
 9     corresponding to these scenarios mostly fall in the range of 2 to 5 °C.  To provide a temporal
10     context for these scenarios, the range of projected global mean warming by 2100 is 1 to 3.5 °C,
11     accompanied by a mean sea-level rise of 15 to 95 cm, according to the IPCC Second Assessment
12     Report. General circulation model (GCM) results are used in this analysis to justify the order of
13     magnitude of the changes used in the sensitivity analyses. They are not predictions that climate
14     will change by specific magnitudes in particular countries or regions. The amount of literature
15     available for assessment varies in quantity and quality among the regions.
16
17     Overview of Regional Vulnerabilities to Global Climate Change
18           The report's assessment of regional vulnerability to climate change focuses on ecosystems,
19     hydrology and water resources, food and fiber production, coastal systems, human settlements,
20     human health, and other sectors or systems (including the climate system) important to
21      10 regions that encompass the Earth's land surface.  Wide variation in the vulnerability of similar
22     sectors or systems is to be expected across regions, as a consequence of regional differences in
23     local environmental conditions; preexisting stresses to ecosystems; current resource-use patterns;
24     and the framework of factors affecting decision making, including government policies, prices,
25     preferences,  and values. Nonetheless, some general observations, based on information
26     contained in the IPCC Second Assessment Report (SAR) (IPCC, 1995) and synthesized from the
27     regional analyses in the 1998 assessment, provide a global context for assessment of each
28     region's vulnerability.  The general types of vulnerabilities are discussed first below, followed by
29     more specific discussion of projected likely regional vulnerabilities most directly applicable to
30     the United States.
31
        March 2001
4C-4
DRAFT-DO NOT QUOTE OR CITE

-------
  1      Ecosystems
  2           Ecosystems are of fundamental importance to environmental function and to sustainability,
  3      and they provide many goods and services critical to individuals and societies. These goods and
  4      services include (1) providing food, fiber, fodder, shelter, medicines, and energy; (2) processing
  5      and storing carbon and nutrients; (3) assimilating wastes; (4) purifying water, regulating water
  6      runoff, and moderating floods; (5) building soils and reducing soil degradation; (6) providing
  7      opportunities for recreation and tourism; and (7) housing the Earth's entire reservoir of genetic
  8      and species diversity.  In addition, natural ecosystems have cultural, religious, aesthetic, and
  9      intrinsic existence values.  Changes in climate have the potential to affect the geographic location
10      of ecological systems, the mix of species that they contain, and their ability to provide the wide
11      range of benefits on which societies rely for their continued existence. Ecological systems are
12      intrinsically dynamic and are constantly influenced by climate variability. The primary influence
13      of anthropogenic climate change on ecosystems is expected to be through the rate and magnitude
14      of change in climate means and extremes; climate change is expected to occur at a rapid rate
15      relative to the speed at which ecosystems can adapt and reestablish themselves; and through the
16      direct effects of increased atmospheric CO2 concentrations, which may increase the productivity
17      and efficiency of water use in some plant species.  Secondary effects of climate change involve
18      changes in soil characteristics and disturbance regimes (e.g., fires, pests, diseases), which would
19      favor some species over others and thus change the species composition of ecosystems.
20           Based on model simulations of vegetation distribution, which use GCM-based climate
21      scenarios, large shifts of vegetation boundaries into higher latitudes and elevations can be
22      expected. The mix of species within a given vegetation class likely will change. Under
23      equilibrium GCM climate scenarios, large regions show drought-induced declines in vegetation,
24      even when the direct effects of CO2 fertilization are included. By comparison, under transient
25      climate scenarios, in which trace gases increase slowly over a period of years, the full effects of
26      changes in temperature and precipitation lag the effects of a change in atmospheric composition
27      by a number of decades; hence, the positive effects of CO2, precede the full effects of changes in
28      climate.
29           Climate change is projected to occur at a rapid rate relative to the speed at which forest
30      species grow, reproduce, and reestablish themselves (past tree species' migration rates are
31      believed to be on the order of 4 to 200 km per century). For mid-latitude regions, an average
        March 2001
4C-5
DRAFT-DO NOT QUOTE OR CITE

-------
 1     warming of 1 to 3.5 °C over the next 100 years would be equivalent to a poleward shift of the
 2     present geographic bands of similar temperatures (or "isotherms") by approximately 150 to
 3     550 km, or an altitude shift of about a 150 to 550 m. Therefore, the species composition of
 4     forests is likely to change; in some regions, entire forest types may disappear, and new
 5     assemblages of species and, hence, new ecosystems may be established. As a consequence of
 6     possible changes in temperature and water availability under doubled equivalent-CO2 equilibrium
 7     conditions, a substantial fraction (a global average of one-third, varying by region from
 8     one-seventh to two-thirds) of the existing forested area of the world likely would undergo major
 9     changes hi broad vegetation types, with the greatest changes occurring in high latitudes and the
10     least in the tropics. In tropical rangelands, major alterations in productivity and species
11     composition would occur because of altered rainfall amount and seasonally and increased
12     evapotranspiration, although a mean temperature increase alone would not lead to such changes.
13           Inland aquatic ecosystems will be influenced by climate change through altered water
14     temperatures, flow regimes, water levels, and thawing of permafrost at high latitudes.  In lakes
15     and streams, warming would have the greatest biological effects at high latitudes, where
16     biological productivity would increase and lead to expansion of cool-water species' ranges, and
17     at the low-latitude boundaries of cold- and cool-water species ranges, where extinctions would be
18     greatest. Increases in flow variability, particularly the frequency and duration of large floods and
19     droughts, would tend to reduce water quality, biological productivity, and habitat in streams. The
20     geographical distribution of wetlands is likely to shift with changes in  temperature and
21     precipitation, with uncertain implications for net greenhouse gas emissions from nontidal
22     wetlands. Some coastal ecosystems (saltwater marshes, mangrove ecosystems, coastal wetlands,
23     coral reefs,  coral atolls, and river deltas) are particularly at risk from climate change and other
24     stresses. Changes in these ecosystems would have major negative effects on freshwater supplies,
25     fisheries, biodiversity, and tourism.
26           Adaptation options for ecosystems are limited, and their effectiveness is uncertain. Options
27     include establishment of corridors to assist the "migration" of ecosystems, land-use management,
28     plantings, and restoration of degraded areas. Because of the projected rapid rate of change
29     relative to the rate at which species can reestablish themselves, the isolation and fragmentation
30     of many ecosystems, the existence of multiple stresses (e.g., land-use change, pollution), and
        March 2001
4C-6
DRAFT-DO NOT QUOTE OR CITE

-------
  1      limited adaptation options, ecosystems (especially forested systems, montane systems, and coral
  2      reefs) are vulnerable to climate change.
  3
  4      Hydrology and Water Resources
  5           Water availability is an essential component of welfare and productivity. Currently,
  6      1.3 billion people do not have access to adequate supplies of safe water, and 2 billion people do
  7      not have access to adequate sanitation.  Although these people are dispersed throughout the
  8      globe, reflecting subnational variations in water availability and quality, some 19 countries
  9      (primarily in the Middle East and northern and southern Africa) face such severe shortfalls that
10      they are classified as either water-scarce or water-stressed; this number is expected to roughly
11      double by 2025, in large part because of increases in demand resulting from economic and
12      population growth. For example, most policy makers now recognize drought as a recurrent
13      feature of Africa's climate. However, climate change will further exacerbate the frequency and
14      magnitude of droughts in some places.
15           Changes in climate could exacerbate periodic and chronic shortfalls of water, particularly in
16      arid and semi-arid areas of the world.  Developing countries are highly vulnerable to climate
17      change because many are located in arid and semi-arid regions, and most derive their water
18      resources from single-point systems such as bore holes or isolated reservoirs. These systems, by
19      their nature, are vulnerable because there is no redundancy in the system to provide resources,
20      should the primary supply fail. Also, given the limited technical, financial, and management
21      resources possessed by developing countries, adjusting to shortages or implementing adaptation
22      measures will impose a heavy burden on their national economies. There is evidence that
23      flooding is likely to become a larger problem in many temperate and humid regions, requiring
24      adaptations not only to droughts and chronic water shortages but also to floods and associated
25      damages, raising concerns about dam and levee failures.
26           The impacts  of climate change will depend on the baseline condition of the water supply
27      system and the  ability of water resources managers to respond not only to climate change but also
28      to population growth and changes in demands; technology; and economic, social, and legislative
29      conditions.
30           Various approaches are available to reduce the potential vulnerability of water systems to
31      climate change. Options include pricing systems, water efficiency initiatives, engineering and
        March 2001
4C-7
DRAFT-DO NOT QUOTE OR CITE

-------
 1      structural improvements to water supply infrastructure, agriculture policies, and urban planning
 2      and management. At the national/regional level, priorities include placing greater emphasis on
 3      integrated, cross-sectoral water resources management, using river basins as resource
 4      management units, and encouraging sound pricing and management practices. Given increasing
 5      demands, the prevalence and sensitivity of many simple water management systems to
 6      fluctuations in precipitation and runoff, and the considerable time and expense required to
 7      implement many adaptation measures, the water resources sector in many regions and countries
 8      is vulnerable to potential changes in climate.
 9
10      Food and Fiber Production
11           Currently, 800 million people are malnourished; as the world's population increases and
12      incomes in some countries rise, food consumption is expected to double over the next three to
13      four decades. The most recent doubling in food production occurred over a 25-year period and
14      was based on irrigation, chemical inputs, and high-yielding crop varieties. Whether the
15      remarkable gains of the past 25 years will be repeated is uncertain. Problems associated with
16      intensifying production on land already in use (e.g., chemical and biological runoff, waterlogging
17      and salinization of soils, soil erosion and compaction) are becoming increasingly evident.
18      Expanding the amount of land under cultivation (including reducing land deliberately taken out
19      of production to reduce agricultural output) also is an option for increasing total crop production,
20      but it could lead to increases in competition for land and pressure on natural ecosystems,
21      increased agricultural emissions of greenhouse gases, a reduction in natural sinks of carbon, and
22      expansion of agriculture to marginal lands, all of which could undermine the ability to
23      sustainably support increased agricultural production.
24           Changes in climate will interact with stresses that result from actions to increase
25      agricultural production, affecting crop yields and productivity in different ways, depending on the
26      types of agricultural practices and systems  in place. The main direct effects will be through
27      changes in factors such as temperature, precipitation, length of growing season, and timing of
28      extreme or critical threshold events relative to crop development, as well as through changes in
29      atmospheric CO2 concentration (which may have a beneficial effect on the growth of many crop
30      types): Indirect effects will include potentially detrimental changes in diseases, pests, and weeds,
31      the effects of which have not yet been quantified in most available studies. Evidence continues
        March 2001
4C-8
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 to support the findings of the IPCC SAR that "global agricultural production could be maintained
 relative to baseline production" for a growing population under 2xCO2 equilibrium climate
 conditions.  In addition, the regional findings of this special report lend support to concerns over
 the "potential serious consequences" of increased risk of hunger in some regions, particularly the
 tropics and subtropics. Generally, middle to high latitudes may experience increases in
 productivity, depending on crop type, growing season, changes in temperature regimes, and the
 seasonality of precipitation. In the tropics and subtropics, where some crops are near their
 maximum temperature tolerance and where dry-land, nonirrigated agriculture predominates,
 yields are likely to decrease. The  livelihoods of subsistence farmers and pastoral peoples, who
 make up a large portion of rural populations in some regions, also could be affected negatively.
 In regions where there is a likelihood of decreased rainfall, agriculture could be significantly
 affected.
      Fisheries and fish production are sensitive to changes in climate and currently are at risk
 from overfishing, diminishing nursery areas, and extensive inshore and coastal pollution.
 Globally, marine fisheries production is expected to remain about the same in response to
 changes in climate; high-latitude freshwater and aquaculture production is likely to increase,
 assuming that natural climate variability and the structure and strength of ocean currents remain
 about the same. The principal impacts will be felt at the national and local levels, as centers of
 production shift. The positive effects of climate change, such as longer growing seasons, lower
 natural winter mortality, and faster growth rates in higher latitudes, may be offset by negative
 factors such as changes in established reproductive patterns, migration routes, and ecosystem
 relationships.
      Given the many forces bringing profound change to the agricultural sector, adaptation
 options that enhance resilience to current natural climate variability and potential changes in
 means and extremes and address other concerns (e.g., soil erosion, salinization) offer no- or
 low-regret options. For example, linking agricultural management to seasonal climate
predictions can assist in incremental adaptation, particularly in regions where climate is strongly
affected by ENSO conditions. The suitability of these options for different regions varies, in part
because of differences in the financial and institutional ability of the private sector and
governments in different regions to implement them. Adaptation options include changes in
crops and crop varieties; development of new crop varieties; changes in planting schedules and
        March 2001
                                          4C-9
DRAFT-DO NOT QUOTE OR CITE

-------
 1     tillage practices; introduction of new biotechnologies; and improved water-management and
 2     irrigation systems, which have high capital costs and are limited by availability of water
 3     resources.  Other options, such as minimum- and reduced-tillage technologies, do not require
 4     such extensive capitalization but do require high levels of agricultural training and support.
 5           In regions where agriculture is well adapted to current climate variability or where market
 6     and institutional factors are in place to redistribute agricultural surpluses to make up for
 7     shortfalls, vulnerability to changes in climate means and extremes generally is low. However, in
 8     regions where agriculture is unable to cope with existing extremes, where markets and
 9     institutions to facilitate redistribution of deficits and surpluses are not in place, or where
10     adaptation resources are limited, the vulnerability of the agricultural sector to climate change
11     should be considered high.  Other factors also will influence the vulnerability of agricultural
12     production in a particular country or region to climate change, including the extent to which
13     current temperatures or precipitation patterns are close to or exceed tolerance limits for important
14     crops, per capita income, the percentage of economic activity based on agricultural production,
15     and the preexisting condition of the agricultural land base.
16
17     Coastal Systems
1 g           Coastal zones are characterized by a rich diversity of ecosystems and a great number of
19     socioeconomic activities. Coastal human populations in many countries have been growing at
20     double the national rate of population growth. Currently, it is estimated that about half of the
21     global population lives in coastal zones, although there is large variation among countries.
22     Changes in climate will affect coastal systems through sea-level rise and an increase in
23     storm-surge hazards and possible changes in the frequency or intensity of extreme events.
24           Coasts in many countries currently face severe sea-level rise problems as a consequence of
25     tectonically and anthropogenically induced subsidence. An estimated 46 million people per year
26      currently are at risk of flooding from storm surges. Climate change will exacerbate these
27      problems, leading to potential impacts on ecosystems and human coastal infrastructure. Large
28      numbers of people also potentially are affected by sea-level rise, for example, tens of millions of
29      people in Bangladesh would be displaced by a 1 -m increase (the top of the range of IPCC
30      Working Group I estimates for 2100) in the absence  of adaptation measures.  A growing number
31      of extremely large cities are located in coastal areas, which means that large amounts of
        March 2001
4C-10       DRAFT-DO NOT QUOTE OR CITE

-------
 1      infrastructure may be affected. Although annual protection costs for many nations are relatively
 2      modest, about 0.1% of gross domestic product (GDP), the average annual costs to many small
 3      island states total several percent of GDP. For some island nations, the high cost of providing
 4      storm-surge protection would make it essentially infeasible, especially given the limited
 5      availability of capital for investment.
 6           Beaches, dunes, estuaries, and coastal wetlands adapt naturally and dynamically to changes
 7      in prevailing winds and seas, as well as sea-level changes; in areas where infrastructure
 8      development is not extensive, planned retreat and accommodation to changes may be possible.
 9      It also may be possible to rebuild or relocate capital assets at the end of their design life. In other
10      areas, however, accommodation and planned retreat are not viable options, and protection using
11      hard structures (e.g., dikes, levees, floodwalls, barriers) and soft structures (e.g., beach
12      nourishment, dune restoration, wetland creation) will be necessary. Factors that limit the
13      implementation of these options include inadequate financial resources, limited institutional and
14      technological capability, and shortages of trained personnel. In most regions, current coastal
15      management and planning frameworks do not take account of the  vulnerability of key systems to
16      changes in climate and sea level or long lead times for implementation of many adaptation
17      measures.  Inappropriate policies encourage development in impact-prone areas. Given
18      increasing population density in coastal zones; long lead times for implementation of many
19      adaptation measures; and institutional, financial, and technological limitations (particularly in
20      many developing countries), coastal systems should be considered vulnerable to changes in
21      climate.
22
23      Human Health
24           In much of the world, life expectancy is increasing; in addition, infant and child mortality
25      in most developing countries is droping. Against this positive backdrop, however, there appears
26      to be a widespread increase in new and resurgent vectorborne and infectious diseases, such as
27      dengue, malaria, hantavirus, and cholera. In addition, the percentage of the developing world's
                                                                     s:
28      population living in cities is expected to increase from 25% (in 1960) to more than 50% by 2020,
29      with percentages in some regions far exceeding these averages. These changes will bring
30      benefits only if accompanied by increased access to services such as sanitation and potable water
31      supplies; they also can lead to serious urban environmental problems, including air pollution
        March 2001
4C-11
DRAFT-DO NOT QUOTE OR CITE

-------
 1      (e-g-j participates, surface ozone, lead), poor sanitation, and associated problems in water quality
 2      and potability, if access to services is not improved.
 3           Climate change could affect human health through increases in heat-stress mortality,
 4      tropical vector-borne diseases, urban air pollution problems, and decreases in cold-related
 5      illnesses. Compared with the total burden of ill health, these problems are not likely to be large.
 6      In the aggregate, however, the direct and indirect impacts of climate change on human health do
 7      constitute a hazard to human population health, especially in developing countries in the tropics
 8      and subtropics; these impacts have considerable potential to cause significant loss of life, affect
 9      communities, and increase health-care costs and lost work days. Model projections (which entail
10      necessary simplifying assumptions) indicate that the geographical zone of potential malaria
11      transmission would expand in response to  global mean temperature increases at the upper part of
12      the IPCC-projected range (3 to 5 °C by 2100), increasing the affected proportion of the world's
13      population from approximately 45% to approximately 60% by the latter half of the next century.
14      Areas where malaria is currently endemic could experience intensified transmission (on the order
15      of 50 to 80 million additional annual cases, relative to an estimated global background total of
16      500 million cases). Some increases in non-vector-borne infectious diseases, such as
17      salmonellosis, cholera, and giardiasis, also could occur as a result of elevated temperatures and
18      increased flooding. However, quantifying the projected health impacts is difficult because the
19      extent of climate induced health disorders  depends on other factors, such as migration, provision
20      of clean urban environments, improved nutrition, increased availability of potable water,
21      improvements in sanitation, the extent of disease vector-control measures, changes in resistance
22-     of vector organisms to insecticides, and more widespread availability of health care.  Human
23      health is vulnerable to changes in climate,  particularly in urban areas, where access to space
24      conditioning may be limited, as well as in areas where exposure to vector-borne and
25      communicable diseases may increase and health-care delivery and basic services, such as
26      sanitation, are poor.
27
28      Regional Vulnerability to Global Climate Change
29           Discussions about two geographic regions (North American and Polar regions) assessed in
30      the report are included here because of their relevance to the continental United States and
31      Alaska, respectively.
        March 2001
4C-12
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 North American Region
      This region consists of Canada and the United States south of the Arctic Circle. Within the
 region, vulnerability to and the impacts of climate change vary significantly from sector to sector
 and from subregion to subregion.  This "texture" is important in understanding the potential
 effects of climate change on North America, as well as in formulating and implementing viable
 response strategies.
      Ecosystems.  Most ecosystems are moderately to highly sensitive to changes in climate.
 Effects are likely to include both beneficial and harmful changes. Potential impacts include
 northward shifts of forest and other vegetation types, which would affect biodiversity by altering
 habitats and would reduce the market and nonmarket goods and  services they provide; declines in
 forest density and forested area in some subregions, but gains in  others; more frequent and larger
 forest fires; expansion of arid land species into the great basin region; drying of prairie pothole
 wetlands that currently support over 50% of all waterfowl in North America; and changes in
 distribution of habitat for cold-, cool-, and warm-water fish. The ability to apply management
 practices to limit potential damages is likely to be low for ecosystems that are not already
 intensively managed.
      Hydrology and Water Resources. Water quantity and quality are particularly sensitive to
 climate change.  Potential impacts include increased runoff in winter and spring and decreased
 soil moisture and runoff in summer. The Great Plains and prairie regions are particularly
 vulnerable. Projected increases in the frequency of heavy rainfall events and severe flooding also
 could be accompanied by an increase in the length of dry periods between rainfall events and in
 the frequency or severity of droughts in parts of North America.  Water quality could suffer and
 would decline where minimum river flows decline. Opportunities to  adapt are extensive, but
 their costs and possible obstacles may be limiting.
      Food and Fiber Production.  The productivity of food and fiber resources of North
 America is moderately to highly sensitive to climate change. Most studies, however, have not
 fully considered the effects of potential changes in climate variability; water availability; stresses
 from pests, diseases, and fire; or interactions with other, existing  stresses. Warmer climate
 scenarios (4 to 5 °C increases in North America) have yielded estimates of negative impacts in
eastern, southeastern, and corn belt regions and positive effects in northern plains and western
regions. More moderate warming produced estimates of predominately positive effects in some
       March 2001
                                         4C-13
DRAFT-DO NOT QUOTE OR CITE

-------
 1     warm-season crops. Vulnerability of commercial forest production is uncertain, but is likely to
 2     be lower than less intensively managed systems because of changing technology and
 3     management options.  The vulnerability of food and fiber production in North America is thought
 4     to be low at the continental scale, though subregional variation in losses or gains is likely. The
 5     ability to adapt may be limited by information gaps; institutional obstacles; high economic,
 6     social, and environmental costs; and the rate of climate change.
 7          Coastal Systems. Sea level has been rising relative to the land along most of the coast of
 8     North America, and falling in a few areas, for thousands of years. During the next century, a
 9     50-cm rise in sea level from climate change alone could inundate 8500 to 19,000 square
10     kilometers of dry land, expand the 100-year flood plain by more than 23,000 square kilometers,
11     and eliminate as much as 50% of North America's coastal wetlands. The projected changes in
12     sea level because of climate change alone would underestimate the total change in sea level from
13     all causes along the eastern seaboard and Gulf Coast of North America. In many areas, wetlands
14     and estuarine beaches may be squeezed between advancing seas and dikes or seawalls built to
15     protect human settlements. Several local governments are implementing land-use regulations to
16     enable coastal ecosystems to migrate landward as sea level rises. Saltwater intrusion may
17     threaten water supplies in several areas.
18          Human Settlements. Projected changes in climate could have positive and negative impacts
19     on the operation and maintenance costs of North American land and water transportation. Such
20     changes also could increase the risks to property and human health and life as a result of possible
21     increased exposure to natural hazards (e.g., wildfires, landslides, extreme weather events) and
22     result in increased demand for cooling and decreased demand for heating energy,  with the overall
23     net effect varying across geographic regions.
24          Human Health.  Climate can have wide-ranging and potentially adverse effects on human
25     health via direct pathways (e.g., thermal stress, extreme weather and climate events) and indirect
26     pathways (e.g., disease vectors and infectious agents, environmental and occupational exposures
27     to toxic substances, food production). In high-latitude regions, some human health impacts are
28     expected because of dietary changes resulting from shifts in migratory patterns and abundance of
29     native food sources.
30           Conclusions. Taken individually, any one of the impacts of climate change  may be within
31     the response capabilities of a subregion or sector. The fact that they are projected to occur
        March 2001
4C-14
DRAFT-DO NOT QUOTE OR CITE

-------
 1      simultaneously and in concert with changes in population, technology, economics, and other
 2      environmental and social changes, however, adds to the complexity of the impact assessment and
 3      the choice of appropriate responses. The characteristics of subregions and sectors of North
 4      America suggest that neither the impacts of climate change nor the response options will be
 5      uniform.
 6           Many systems of North America are moderately to highly sensitive to climate change, and
 7      the range of estimated effects often includes the potential for substantial damages. The
 8      technological capability to adapt management of systems to lessen or avoid damaging effects
 9      exists in many instances. The ability to adapt may be diminished, however, by the attendant
10      costs, lack of private incentives to protect publicly owned natural systems, imperfect information
11      regarding future changes in climate and the available options for adaptation, and institutional
12      barriers. The most vulnerable sectors and regions include long-lived natural forest ecosystems in
13      the east and ulterior west, water resources in the southern plains, agriculture in the southeast and
14      southern plains, human health in areas currently experiencing diminished urban air quality,
15      northern ecosystems and habitats,  estuarine beaches in developed areas, and low-latitude cool
16      and cold-water fisheries. Other sectors and subregions may benefit from opportunities associated
17      with warmer temperatures or, potentially, from CO2 fertilization, including west coast coniferous
18      forests; some western rangelands;  reduced energy costs for heating in the northern latitudes;
19      reduced salting and snow-clearance costs; longer open-water seasons in northern channels and
20      ports; and agriculture in the northern latitudes, the interior west, and the west coast.
21
22      Polar (Arctic and Antarctic) Regions
23           The polar regions include some very diverse landscapes, and the Arctic and the Antarctic
24      are very different in character. The Arctic is defined here as the area within the Arctic Circle; the
25      Antarctic here includes the area within the Antarctic Convergence, including the Antarctic
26      continent, the Southern Ocean, and the sub-Antarctic islands. The Arctic can be described as a
27      frozen ocean surrounded by land, and the Antarctic as a frozen continent surrounded by ocean.
28      The projected warming in the polar regions is greater than for many other regions of the world.
29      Where temperatures are close to freezing on average, global warming will reduce land ice and
30      sea ice, the former contributing to  sea-level rise.  However, in the interiors of ice caps, increased
        March 2001
4C-15
DRAFT-DO NOT QUOTE OR CITE

-------
 1     temperature may not be sufficient to lead to melting of ice and snow, and will tend to have the
 2     effect of increasing snow accumulation.
 3           Ecosystems.  Major physical and ecological changes are expected in the Arctic. Frozen
 4     areas close to the freezing point will thaw and undergo substantial changes with warming.
 5     Substantial loss of sea ice is expected in the Arctic Ocean.  As warming occurs, there will be
 6     considerable thawing of permafrost, leading to changes in drainage, increased slumping, and
 7     altered landscapes over large areas.  Polar warming probably will increase biological production
 8     but may lead to different species composition on land and in the sea. On land, there will be a
 9     tendency for polar shifts in major biomes such as tundra and boreal forest and associated
10     animals, with significant impacts on species such as bear and caribou. However, the Arctic
11     Ocean geographically limits northward movement. Much smaller changes are likely for the
12     Antarctic, but there may be species shifts. In the sea, marine ecosystems will move poleward.
13     Animals dependent on  ice may be disadvantaged in both polar areas.
14           Hydrology and Water Resources. Increasing temperature will thaw permafrost and melt
15     more snow and ice. There will be more running and standing water. Drainage systems in the
16     Arctic are likely to change at the local scale. River and lake ice will break up earlier and  freeze
17     later.
18           Food and Fiber Production. Agriculture is severely limited by the harsh climate. Many
19     limitations will remain in the future, although some  small northern extension of farming into the
20     Arctic may be possible. In general, marine ecological productivity should rise. Warming should
21     increase growth and development rates of nonmammals; ultraviolet-B (UV-B) radiation is still
22     increasing, however, which may adversely affect primary productivity as well as fish
23     productivity.
24           Coastal Systems.  As warming occurs, the Arctic could experience a thinner and reduced
25     ice cover. Coastal and river navigation will increase, with new opportunities for water transport,
26     tourism, and trade. The Arctic Ocean could become a major global trade route.  Reductions in
27     ice will benefit offshore oil production. Increased erosion of Arctic shorelines is expected from a
28     combination of rising sea level, permafrost thaw, and increased wave action as a result of
29     increased open water.  Further breakup of ice shelves in the Antarctic peninsula is likely.
30     Elsewhere in Antarctica, little change is expected in coastlines and probably  in its large ice
31     shelves.
        March 2001
4C-16
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
      Human Settlements. Human communities in the Arctic will be substantially affected by the
projected physical and ecological changes. The effects will be particularly important for
indigenous peoples leading traditional lifestyles. There will be new opportunities for shipping,
the oil industry, fishing, mining, tourism, and migration of people. Sea ice changes projected for
the Arctic have major strategic implications for trade, especially between Asia and Europe.
      Conclusions.  The Antarctic peninsula and the Arctic are very vulnerable to projected
climate change and its impacts. Although the number of people directly affected is relatively
small, many native communities will face profound changes that impact on traditional lifestyles.
Direct effects could include ecosystem shifts, sea and river-ice loss, and permafrost thaw.
Indirect effects could include feedbacks to the climate system such as further releases of
greenhouse  gases, changes in ocean circulation drivers, and increased temperature and higher
precipitation with loss of ice, which could affect climate and sea level globally.  The interior of
Antarctica is less vulnerable to climate change, because the temperature changes envisaged over
the next century are likely to have little impact and very few people are involved. However,
there are considerable uncertainties about the mass balance of the Antarctic ice sheets and the
future behavior of the West Antarctic ice sheet (low probability of disintegration over the next
century). Changes in either could affect sea level and Southern Hemisphere climates.
REFERENCES
Intergovernmental Panel on Climate Change (IPCC). (1996) Climate change 1996: contribution of working group I
     to the second assessment of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom:
     Cambridge University Press; p. 572. [IPCC second assessment report].
Intergovernmental Panel on Climate Change (IPCC). (1998) The regional impacts of climate change: an assessment
     of vulnerability. Cambridge, United Kingdom: Cambridge University Press.
        March 2001
                                          4C-17
DRAFT-DO NOT QUOTE OR CITE

-------

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
                               APPENDIX 4D
          Excerpted Materials from the U.S. Global Change Research
      Program Assessment Overview Report on Climate Change Impacts
    on the United States (U.S. Global Change Research Program, 2000) and
                    Subsidiary Regional Assessment Reports
     The subject Assessment Overview on Climate Change Impacts on the United States
(USGCRP, 2000) was prepared by the USGCRP National Assessment Synthesis Team (NAST)
and represents an important landmark for the U.S. national process of research analyses and
dialog about coming changes in climate, their impacts, and possible adaptation measures that can
be taken. NAST consists of a committee of experts drawn from governments, academia industry,
and nongovernmental organizations (NGO's). The subject overview is based on a much more
extensive, detailed "foundation" report, written by NAST in coordination with independent
regional and sector assessment teams. The subject assessment, required by a 1990 U.S. law, was
conducted under the USGCRP in response to a request from the President's Science Advisor.
The materials presented below are excerpted from the September 2000 NSTC review draft of the
assessment overview and, if necessary, later will be appropriately corrected to reflect the final
versions of the report due out in fall 2000, after completion of all peer-review and clearance
precesses.  Selected material derived from one or another of the specific regional assessments
also are presented in this appendix. The materials selected for presentation here are meant to
provide an informative introduction to the latest available expert assessment of potential sector
and regional-scale impacts of climate change in the United States and to illustrate the difficulties
in projecting likely varying location-specific mixes of potential deleterious and beneficial effects
of climate change.
     The past record of 1000 years of global temperature and CO2 emissions change, as
depicted by the assessment  overview, is shown in Figure 4D-1. As noted in the Figure 4D-1,
there appears to be a relatively close correlation between marked parallel increases in
anthropogenic carbon emissions starting roughly in the latter part of the 18th century, increasing
atmospheric CO2 concentrations, and notable increasing global average temperature trends.
       March 2001
                                       4D-1
DRAFT-DO NOT QUOTE OR CITE

-------
                      Temperature Change
                 380-


               Q-, 360


               •§ 340

               2"
               § 320
               W 300
               e
               ra
               0- 280
               O
                 260
CO2 Concentrations
                                                  & s  s
Figure 4D-1.  Records of Northern Hemisphere surface temperatures, CO2 concentrations,
             and carbon emissions show a close correlation. Temperature Change:
             reconstruction of annual-average Northern Hemisphere surface air
             temperatures derived from historical records, tree rings, and corals (blue),
             and air temperatures directly measured (purple). CO2 Concentrations:
             record of global CO2 concentration for the last 1000 years, derived from
             measurements of CO2 concentration in air bubbles in the layered ice cores
             drilled in Antarctica (blue line) and from atmospheric measurements since
             1957.  Carbon Emissions: reconstruction of past emissions of CO2 as a result
             of land clearing and fossil fuel combustion  since about 1750 (in billions of
             metric tons of carbon per year.
 March 2001
                  4D-2
DRAFT-DO NOT QUOTE OR CITE

-------
  1          The two primary models used to project future changes in global climate in the USGCRP
  2     assessment were developed at the Canadian Climate Centre and the Hadley Centre in the United
  3     Kingdom and have been peer-reviewed extensively by other scientists. Both incorporate similar
  4     assumptions about future emissions of carbon dioxide and other major greenhouse gases (both
  5     approximate the IPCC "business as usual" scenario with a 1% per year increase in greenhouse
  6     gases and growing sulfur emissions). These models were the best fit to a list of criteria
  7     developed for the U.S. National Assessment. Climate models developed at the National Center
  8     for Atmospheric Research (NCAR), NOAA's Geophysical Fluid Dynamics Laboratory (GFDL),
  9     NASA's Goddard Institute for Space Studies (GISS), and Max Planck Institute (MPI) in
 10     Germany also were used in various aspects of the assessment. Although the physical principles
 11     driving the models are similar, they differ in how they represent the effects of some important
 12     processes, with the two primary models yielding different views of 21 st century climate. On
 13     average over the United States, the Hadley model projects a much wetter climate than does the
 14     Canadian model, although the Canadian model projects a greater increase in temperature than
 15     does the Hadley. Both projections are plausible, given current understanding.  See Figure 4D-2
 16     for plots of U.S. average temperature increases projected by the different models.  In all climate
 17     models, increases in temperature for the United States are significantly higher than global
 18     average temperature increases (see Table 4D-1), because of the fact that all models project
 19     warming to be greatest at middle to high latitudes (partly because melting snow and ice make the
20     surface less reflective of sunlight, allowing it to absorb more heat). Warming also will be greater
21      over land than over the oceans because it takes longer for the oceans to warm.
22          Uncertainties about future climate stem from a wide variety of factors (e.g., questions about
23      how to represent clouds and precipitation in climate models and uncertainties about how
24     emissions of greenhouse gases will change). These uncertainties result in differences in climate
25      model projections. Examining these differences aids in understanding the range of risk or
26      opportunity associated with a plausible range of future climate changes. These differences in
27      model projections also raise questions about how to interpret model results, especially at the
28      regional level, where projections can differ significantly.
29           One of the most important world-wide consequences of the overall global warming
30      increases projected for the 21 st century is sea level rise, and it can be expected to impact Alaska,
31      coastal areas of the continental United States, and U.S. Hawaiian and Carribean islands regions,
       March 2001
4D-3
DRAFT-DO NOT QUOTE OR CITE

-------
LL.
o
 Q)
 O>
O
Q)
•*->
£
CD
D.

I
9-

8

7

6

5



3-

2-

 1-

0-

-1-



-3

-4-
     -5
              Changes  in Temperature over the  U.S.
                     Simulated by Climate Models
                   Hadley Centre - Version 2
                   Canadian Centre
                   Max Planck Institute
                   Geophysical Fluid Dynamics Laboratory
                   Hadley Centre - Version 3
                   NCAR - Parallel Version
                   NCAR - Climate System Model
      1850
                   1900
                             I
                           1950
2000
2050
2100
Figure 4D-2. Simulation of decadal average changes in temperature from leading climate
            models on historic and projected changes in CO2 and sulfate atmospheric
            concentrations. For the 21st century, the projected global temperature
            increase is 4.9 °F for the Hadley model and 7.4 °F for the Canadian model.
            The model with the smallest projected increase in global temperature is the
            NCAR Climate System Model at 3.6 °F. By comparison, the projected
            increase in temperature for the 21st century over the contiguous United
            States is Canadian, 9.4 °F, Hadley, 5.5 °F, and NCAR Climate System
            Model, 4.0 °F.
March 2001
                                    4D-4
                                          DRAFT-DO NOT QUOTE OR CITE

-------
            TABLE 4D-1. RANGE OF PROJECTED WARMING IN THE 21ST CENTURY
                                       	Global	United States
         Hadley Model                            +4.9 °F                    +5.5 °F
         Canadian Model                          +7.4 °F                    +9.4 °F
         NCAR Climate System Model	+3.6 °F                    +4.0 °F
  1      as well. Figure 4D-3 illustrates sea level rise predicted by the Canadian and Hadley models used
  2      in the USGCRP assessment.  The sea level rise projected by either model can be expected to pose
  3      threats, not only in terms of potential inundation of low lying portions of the Hawaiian and
  4      Carribean islands, but also in terms of shoreline erosion in portions of Alaska and the continental
  5      United States. Those continental United States areas most vulnerable to future sea level rise are
  6      those low lying areas already experiencing rapid erosion rates, as depicted in Figure 4D-4.
  7      Substantial impacts can be expected, including losses of coastal wetlands important for migratory
  8      birds and degradation of estuarine sound complexes providing shallow water fishery nurseries
  9      (most immediately because of salt water incursions resulting from sea level rise and other
10      impacts resulting from more frequent and extensive algal-toxic blooms impacting coastal
11      commercial fish and shell fish harvests secondary to increased nutrient out flows caused by
12      extreme rain fall events [e.g., during hurricanes]).
13           The main climate models used all predict notable increases in the minimum and maximum
14      annual average temperatures in the United States during the next 100 years. Projected changes in
15      temperature minimum and maxima are likely more important than average temperatures, in that
16      they influence such things as human comfort, heat and cold stress in plants and animals,
17      maintenance of snow pack, and pest populations (low temperatures kill many pests and higher
18      minimum temperatures may allow increased overwinter survival of pests).  The largest increases
19      in temperature are projected over much of the southern United States in summer, dramatically
20      raising the heat index (a measure of discomfort based on temperature and humidity). Also,
21      following an average 5 to  10% increase in average U.S. precipitation over the last century, the
22      climate models project notable changes in precipitation during the 21 st century. The Canadian
       March 2001
4D-5
DRAFT-DO NOT QUOTE OR CITE

-------
                                   Canadian Model (Thermal Expansion)
                                   Hadley Model (Themial Expansion)
                                   Hadley Model (T.E. + glacial melt)
                            1850
                                        1950
                                              2000
                                                    2050
                                                           2100
Figure 4D-3. Historic and projected changes in sea level (in inches) based on the Canadian
             and Hadley model simulations. The Canadian model projection includes
             only the effects of thermal expansion of warming ocean waters. The Hadley
             projection includes both thermal expansion and the additional sea-level rise
             projected because of melting of land-based glaciers.  Neither model includes
             consideration of possible sea-level changes because of polar ice melting or
             accumulation of snow on Greenland and Antarctica.
                             Severely eroding
Figure 4D-4. This map is a preliminary classification of annual shoreline erosion
             throughout the United States, in coarse detail and resolution. The areas most
             vulnerable to future sea-level change are those with low relief that are already
             experiencing rapid erosion rates, such as the Southeast and Gulf Coast.
March 2001
4D-6
DRAFT-DO NOT QUOTE OR CITE

-------
  1     model predicts the largest percentage increases in precipitation in California and the Southwest;
  2     but east of the Rocky Mountains, the southern half of the United States is projected to have
  3     decreased precipitation (with especially large decreases in eastern Colorado, western Kansas, and
  4     in an arc stretching from Louisiana to Virginia). The Hadley model predicts largest percentage
  5     increases in southern California and the Southwest, along with lesser increases for the rest of the
  6     nation, except for small areas of the Northwest and the Gulf Coast. Both models also predict
  7     increases in frequency of heavy precipitation effects, largely because of shifts in storm activity
  8     and tracks. Soil moisture critical for both agriculture and natural ecosystems may, despite
  9     increased precipitation, actually undergo marked decreases in some areas, because of offsetting
 10     evaporation rates increased by higher temperatures during projected scenarios of likely increased
 11     periods of drought for some U.S. regions.
 12          The predicted changes in temperature and precipitation are expected to result in varying
 13     impacts of climate change on ecosystems various U.S. regions.  Such impacts will likely include
 14     the following.
 15     • Changes in productivity and carbon storage capacity of ecosystems (decreases in some places
 16       and increases in others are very likely).
 17     • Shifts in the distribution of major plant and animal species are likely.
 18     • Some ecosystems, such as alpine meadows, are likely to disappear in some places because the
 19       new local climate will not support them or there are barriers to their movement.
20     • In many places, it is very likely that ecosystem services, such as air and water purification,
21       landscape stabilization against erosion, and carbon storage capacity will be reduced.  These
22       losses likely will occur in the wake of episodic, large-scale disturbances that trigger species
23       migrations or local extinctions.
24     • In some places, it is very likely that ecosystems services will be enhanced where climate-
25       related stresses are reduced.
26          The USGCRP assessment provides extensive detailed evaluations of the above and other
27     types of impacts projected to occur as consequences of changing weather patterns (and
28     consequent shifts in temperature, precipitation, etc.). Such evaluations are summarized in the
29     overview assessment in relation to several overall sectors (water resources, agriculture, forests,
30     coastal areas and marine resources, and human health) and in relation to different regions of the
31      United States.
        March 2001
4D-7
DRAFT-DO NOT QUOTE OR CITE

-------
 1          The following concise statements highlight some of the more salient points emerging from
 2     the overall sector evaluations.
 3     • Water. Rising temperatures and greater precipitation are likely to lead to more evaporation
 4       and greater swings between wet and dry conditions. Changes in the amount and timing of rain,
 5       snow, runoff, and soil moisture are very likely. Water management, including pricing and
 6       allocation, will very likely be important in determining many impacts.
 7     • Agriculture. Overall productivity of American agriculture likely will remain high and is
 8       projected to increase throughout the 21 st century, with northern regions faring better than
 9       southern ones. Though agriculture is highly dependent on climate, it is also highly adaptive
10       Weather extremes, pests, and weeds likely will present challenges in a changing climate.
11       Falling commodity prices and competitive pressures are likely to stress farmers and rural
12       communities.
13     • Forests. Rising CO2 concentrations and modest warming are likely to increase forest
14       productively in many regions. With larger increases in temperature, increased drought is likely
15       to reduce forest productivity in some regions, notably in the Southeast and Northwest. Climate
16       change is likely to cause shifts in species ranges, as well as large changes in disturbances such
17       as fire and pests.
18     • Coastal Areas and Marine Resources. Coastal wetlands and shorelines are vulnerable to
19       sea-level rise and storm surges, especially when climate impacts are combined with the
20       growing stressed of increasing human population and development.  It is likely that coastal
21       communities will be affected increasingly by extreme events. The negative impacts on natural
22       ecosystems are very likely to increase.
23      • Human Health.  Heat-related illnesses and deaths, air pollution, injuries and deaths from
24       extreme weather events, and diseases carried by water, food, insects, ticks,  and rodents, have
25       all been raised as concerns for the United States in a warmer world.  Modern public health
26       efforts will be important in identifying and adapting to these potential impacts.
27           The USGCRP Assessment also evaluated sector impacts in relation to various U.S. regions,
28      broken out as depicted in Figure 4D-5 derived from the overview assessment (USGCRP, 2000).
29      That assessment highlighted the following important points in relation to expected major impacts
30      in each of the regions evaluated.
31
        March 2001
4D-8
DRAFT-DO NOT QUOTE OR CITE

-------
                                                                                    C
                                                                                    O
                                                                                    
                                                                                    en
                                                                                    es
                                                                                   U
                                                                                   O
                                                                                   •o
                                                                                    
-------
 1     • Alaska.  Sharp winter and springtime temperature increases are very likely to cause continued
 2       thawing of permafrost, further disrupting forest ecosystems, roads, and buildings.
 3     • Northwest. Increasing stream temperatures are very likely to further stress migrating fish,
 4       complicating restoration efforts.
 5     • Mountain West. Higher winter temperatures are very likely to reduce snowpack and peak
 6       runoff and shift the peak to earlier in the spring, reducing summer runoff and complicating
 7       water management for flood control, fish runs, cities, and irrigation.
 8     • Southwest. With an increase in precipitation, the desert ecosystems native to this region are
 9       likely to decline, whereas grasslands and shrublands likely are to expand.
10     • Midwest/Great Plains. Higher CO2 concentrations are likely to offset the effects of rising
11       temperatures on forests and agriculture for several decades, increasing productivity.
12     • Southern Great Plains.  Prairie potholes, which provide important habitat for ducks and other
13       migratory waterfowl, are likely to dry up  in a warmer climate.
14     • Great Lakes. Lake levels are likely to decline, leading to reduced water supply and more
15       costly transportation.  Shoreline damage caused by high water levels is likely to decrease.
16     • Northern and Mountain Regions.  It is very probable that warm weather recreational
17       opportunities, such as hiking, will expand, whereas cold  weather activities, such as skiing, will
18       contract.
19     • Northeast, Southeast, and Midwest. Rising temperatures are very likely to increase the heat
20       index dramatically in summer, with impacts to health and comfort.  Warmer winters are likely
21       to reduce cold-related stresses.
22     • Appalachians. Warmer and moister air very likely will  lead to more intense rainfall events,
23       increasing the potential for flash floods.
24     • Southeast.  Under warmer wetter scenarios, the range of southern tree species is likely to
25       expand.  Under hotter and drier scenarios, it is likely that far southeastern forests will  be
26       displaced by grasslands and savannas.
27     • Southeast Atlantic Coast. It is very probable that rising sea levels and storm surge will
28       threaten natural ecosystems and human coastal development and reduce buffering capacity
29       against storm impacts.
30     • Southeast Gulf Coast.  Inundation of coastal wetlands will very likely increase, threatening
31       fertile areas for marine life, migrating birds, and waterfowl.
       March 2001                               4D-10        DRAFT-DO NOT QUOTE OR CITE

-------
  1      • Islands. More intense El Nino and La Nina events are possible and are likely to create extreme
  2       fluctuations in water resources for island citizens and the tourists who sustain local economies.
  3           Other materials from the overview assessment summarize regional concerns with regard to
  4      different types of sector impacts. Tables 4D-2 and 4D-3 present two examples drawn from the
  5      overview assessment, denoting concerns or impacts regarding water resources and types of
  6      ecosystems, respectively, likely to be impacted in different U.S. regions.
          TABLE 4D-2. TYPES OF WATER CONCERNS PROJECTED TO BE IMPORTANT
               FOR U.S. REGIONS CONSEQUENT TO FUTURE CLIMATE CHANGE"
Region
Northeast
Southeast
Midwest
Great Plains
West
Northwest
Alaska
Islands
Floods
X
X
X
X
X
X

X
Droughts
X
X
X
X
X
X
X
X
Snowpack
X

X
X
X
X
X

Groundwater
X
X
X
X
X


X
Lake, River, and
Reservoir Levels


X
X
X
X


Quality
X
X
X
X
X


X
         This table identifies some of the key regional concerns about water. Many of these issues were raised and
         discussed by stakeholders during regional workshops and other Assessment meetings held between 1997 and
         2000.
 1          As seen in Table 4D-2, different types of water impacts are projected to be of important
 2     widespread concern across many different U.S. regions. It should be noted that some limited
 3     beneficial effects may occur in some regions (e.g., longer periods of open-water transportation on
 4     navigable rivers and sounds  in and around Alaska).
 5          The overview assessment notes that the information presented in Table 4D-3 represents
 6     only a partial list of potential impacts for major ecosystem types and that, although the impacts
 7     often are stated in terms of plant-community impacts, it is important to recognize that such plant-
 8     community changes also will have animal habitat effects and consequent impacts on both
 9     terrestrial and aquatic animal species. Both the plant and animal impacts can have further
10     consequent impacts on human health and welfare, which also can be expected to vary
11     considerably from region to region.
       March 2001
4D-11
DRAFT-DO NOT QUOTE OR CITE

-------
   TABLE 4D-3. PROJECTED FUTURE CLIMATE-CHANGE-INDUCED IMPACTS
    ON TYPES OF ECOSYSTEMS OF CONCERN TO DIFFERENT U.S. REGIONS
 Ecosystem
 Type
Impacts
        U.S. Regions
 Forests      Changes in tree species composition and
             alteration of animal habitat

             Displacement of forests by woodlands and
             grasslands under a warmer climate in
             which soils are drier

 Grasslands   Displacement of grasslands by woodlands
             and forests under a wetter climate

             Increase in success of nonnative invasive
             plant species

 Tundra      Loss of alpine meadows as their species
             are displaced by lower elevation species

             Loss of northern tundra as trees migrate
             poleward

             Changes in plant community composition
             and alteration of animal habitat

 Semi-arid    Increase in woody species and loss of
 and Arid     desert species under wetter climate

 Freshwater   Loss of prairie pot holes with more
             frequent drought conditions

             Habitat changes in rivers and lakes as
             amount and timing of runoff changes and
             water temperatures rise

 Coastal and  Loss of coastal wetlands as sea level rises
 Marine      and coastal development prevents
             landward migration

             Loss of barrier islands as sea-level rise
             prevents landward migration

             Changes in quantity and quality of
             freshwater delivery to estuaries and bays
             alter plant and animal habitats

             Loss of coral reefs as water temperature
             increases

             Changes in ice location and duration alter
             marine mammal habitat
                                                  NE  SE  MW   GP   WE   PNW   AK   IS
                        X
                                   X
                                              X
               X
                                                           X
                                                           X
                                                                X
March 2001
                 4D-12
DRAFT-DO NOT QUOTE OR CITE

-------
 1           The USGCRP (2000) assessment included extensive detailed evaluation of projected

 2      climate change impacts on the different U.S. regions depicted in Figure 4D-5. Overview reports

 3      on those detailed evaluations by various regional assessment teams are in various stages of

 4      preparation, with information pertaining to each being available via the internet at the following

 5      address: http://www.nacc.usgcrp.gov/regions/.

 6           The wide variation in the types of projected impacts of climate change, both deleterious and

 7      some possible beneficial effects,  can be readily illustrated by one example illustrated in

 8      Figure 4D-6. The figure depicts projected types of changes that may occur (with varying degrees

 9      of certainty indicated) as the consequence of climate change impacts on the Mid-Atlantic Region

10      (MAR) of the United States, including both potentially negative and positive impacts.

11
                    Summary of MAR impacts
                                     Positive Impact
         Most Certain

          • Agricultural production

          • Coastal zones

          • Temperature related health status
    erosion,
saltwater intrusion
                                            soybeans,
                                            possibly corn
                                            and treefruits
         Moderately Certain

          • Forestery production

          • Temperature related health status
                                                                                          less cold stress
         Uncertain

          • Biodiversity

          • Fresh water quantity

          • Fresh water quality

          • Ecological functioning

          • Vector and water-borne disease health status

          • Environmental effects from agriculture
 migration barriers,
  invasive species
forest composition,
cold water fisheries
              nutrient leaching,
                  runoff
        Figure 4D-6.  Projected climate change impacts in the Mid-Atlantic Region (MAR) of the
                      United States.


        Source:  Mid-Atlantic Regional Assessment Team (2000).
        March 2001
     4D-13
DRAFT-DO NOT QUOTE OR CITE

-------
1      REFERENCES
2      Mid-Atlantic Regional Assessment (MARA) Team. (2000) Preparing for a changing climate: the potential
3            consequences of climate variability and change (Mid-Atlantic overview). Washington, DC: U.S. Global
4            Change Research Program (USGCRP).
5 „     U. S. Global Change Research Program (USGCRP, 2000) Climate Change Impacts on the United States: the
6            Potential Consequences of Climate Variability and Change (Overview), Report of National Assessment
7            Synthesis Team (NAST).  NSTC Review Draft (September 2000).
8
       March 2001
4D-14
DRAFT-DO NOT QUOTE OR CITE

-------
                         APPENDIX 4E
    Recent Model Projections of Excess Mortality Expected in U.S. Cities
   During Summer and Winter Seasons Because of Future Climate Change,
                 Based on Kalkstein and Greene (1997)
March 2001
4E-1
DRAFT-DO NOT QUOTE OR CITE

-------
   TABLE 4E-1. MODELED PROJECTIONS OF DIRECT HUMAN HEALTH
IMPACTS OF CLIMATE CHANGE: ESTIMATED TOTAL EXCESS MORTALITY
      IN U.S. URBAN AREAS FOR AN AVERAGE SUMMER SEASON,
              ASSUMING FULL ACCLIMATIZATION3
Year 2020 Climate
SMSA
Anaheim
Atlanta
Baltimore
Birmingham
Boston
Buffalo
Chicago
Cincinnati
Cleveland
Columbus
Dallas
Denver
Detroit
Ft. Lauderdale
Greensboro
Hartford
Houston
Indianapolis
Jacksonville
Kansas City
Los Angeles
Louisville
Memphis
Miami
Minneapolis
Nassau
New Orleans
New York
March 2001
Present
climate
0
25
84
42
96
33
191
14
29
33
36
42
110
0
22
38
7
36
0
49
68
17
25
0
59
29
20
307

GFDL 89
0
,43
57
26
113
15
243
16
21
24
45
29
84
0
28
21
7
23
0
79
74
0
42
0
55
59
0
363

Year 2050 Climate
UKMO Max Planck GFDL 89
0
62
148
47
165
52
538
90
55
83
62
41
240
0
43
42
16
93
0
173
123
2
27
0
185
84
0
753

0
22
63
14
134
36
421
49
44
51
45
30
164
0
27
32
7
51
0
93
83
0
57
0
148
84
0
498
4E-2
0
60
124
40
155
34
359
54
46
51
107
35
130
0
37
38
15
55
0
121
110
0
40
0
123
110
0
460
UKMO Max Planck
0
138
164
47
194
73
583
81'
58
90
64
39
271
0
45
50
17
86
0
156
128
1
29
0
215
116
0
999
DRAFT-DO NOT QUOTE
0
33 .
131
21
160
59
550
67
53
78
44
32
256
0
29
41
6
69
0
105
116
1
49
0
186
116
0
727
OR CITE

-------
  TABLE 4E-1 (cont'd). MODELED PROJECTIONS OF DIRECT HUMAN HEALTH
  IMPACTS OF CLIMATE CHANGE: ESTIMATED TOTAL EXCESS MORTALITY
          IN U.S. URBAN AREAS FOR AN AVERAGE SUMMER SEASON,
                     ASSUMING FULL ACCLIMATIZATION8
Year 2020 Climate
SMSA
Newark
Philadelphia
Phoenix
Pittsburgh
Portland
Providence
Riverside
Salt Lake City
San Antonio
San Diego
San Francisco
San Jose
Seattle
St. Louis
Tampa
Washington, DC
Total
Present
climate
26
129
0
39
9
47
4
0
4
0
28
0
5
79
28
0
1,840
GFDL 89
83
99
0
32
13
39
6
0
0
0
24
0
1
149
68
0
1,981
UKMO
173
362
0
66
.22
80
10
0
0
0
23
0
0
173
95
0
4,128
Max Planck
111
191
0
64
11
52
6
0
0
0
23
0
2
158
28
0
2,799
Year 2050 Climate
GFLD 89
150
246
0
61
23
73
8
0
0
0
18
0
0
212
95
0
3,790
UKMO
127
477
0
83
31
96
11
0
0
0
24
0
0
155
100
0
4,748
Max Planck
161
323
0
95
14
74
7
0
0
0
23
0
1
189
47
0
3,863
 "Abbreviations: SMSA, standard metropolitan statistical area; GFDL, Geophysical Fluid Dynamics Laboratory
  Model; UKMO, United Kingdom Meteorological Office Model; Max Planck, Max Planck Institute Model.
  Values given are estimated excess deaths.

 Source: Kalkstein and Green (1997).
March 2001
4E-3
DRAFT-DO NOT QUOTE OR CITE

-------
    TABLE 4E-2. MODELED PROJECTIONS OF DIRECT HUMAN HEALTH
   IMPACTS OF GLOBAL CLIMATE CHANGE: ESTIMATED TOTAL EXCESS
   MORTALITY IN U.S. URBAN AREAS FOR AN AVERAGE WINTER SEASON,
                ASSUMING FULL ACCLIMATIZATION"
Year 2020 climate
SMSA
Anaheim
Atlanta
Baltimore
Birmingham
Boston
Buffalo
Chicago
Cincinnati
Cleveland
Columbus
Dallas
Denver
Detroit
Ft. Lauderdale
Greensboro
Hartford
Houston
Indianapolis
Jacksonville
Kansas City
Los Angeles
Louisville
Memphis
Miami
Minneapolis
Nassau
Present
Climate
2
37
0
25
0
7
2
0
2
12
32
9
34
36
0
0
24
16
0
12
100
16
23
46
0
24
GFDL 89
0
53
0
12
0
18
4
0
9
1
41
10
15
4
0
0
33
32
0
51
102
12
20
35
0
21
UKMO
0
48
0
8
0
8
3
0
10
2
33
11
20
4
0
0
29
28
0
36
78
17
17
35
0
4
Max Planck
1
52
0
11
0
17
4
0
10
1
43
10
15
5
0
0
35
33
0
46
100
12
19
37
0
20
Year 2050 climate
GFDL 89
0
50
0
11
0
5
4
0
15
3
36
11
18
3
0
0
29
34
0
42
77
19
19
32
0
5
UKMO
0
47
0
7
0
5
2
0
10
2
31
11
25
3
0
0
27
28 •
0
35
88
15
15
32
0
3
Max Planck
0
52
0
12
0
' 18
5
0
9
1
41
11
14
5
0
0
33
32
0
46
81
12
19
36
0
21
March 2001
4E-4
DRAFT-DO NOT QUOTE OR CITE

-------
  TABLE 4E-2 (cont'd). MODELED PROJECTIONS OF DIRECT HUMAN HEALTH
    IMPACTS OF GLOBAL CLIMATE CHANGE: ESTIMATED TOTAL EXCESS
    MORTALITY IN U.S. URBAN AREAS FOR AN AVERAGE WINTER SEASON,
                      ASSUMING FULL ACCLIMATIZATION3
Year 2020 climate
SMSA
New Orleans
New York
Newark
Philadelphia
Phoenix
Pittsburgh
Portland
Providence
Riverside
Salt Lake City
San Antonio
San Diego
San Francisco
San Jose
Seattle
St. Louis
Tampa
Washington, DC
Total
Present
Climate
52
102
48
85
26
19
17
27
10
5
9
17
85
3
13
50
21
19
1,067
GFDL 89
56
123
23
80
25
20
15
21
26
7
10
26
39
2
40
61
26
31
1,104
UKMO
51
150
8
14
26
29
12
34
29
9
6
16
30
4
45
68
24
38
984
Max Planck
54
120
20
73
25
21
15
33
26
8
11
24
42
2
37
60
26
30
1,098
Year 2050 climate
GFDL 89
51
152
10
36
26
24
12
35
27
8
5
16
30
3
46
53
22
20
989
UKMO
47
93
6
9
27
31
10
36
27
10
4
18
21
5
47
61
20
35
894
Max Planck
54
121
23
82
26
21
13
21
26
9
9
16
26
4
43
61
25
31
1,059
 "Abbreviations: SMSA, standard metropolitan statistical area; GFDL, Geophysical Fluid Dynamics Laboratory
 Model; UKMO, United Kingdom Meteorological Office Model; Max Planck, Max Planck Institute Model.
 Values given are estimated excess deaths.
REFERENCE
Kalkstein, L. S.; Greene, J. S. (1997) An evaluation of climate/mortality relationships in large U.S. cities and the
     possible impacts of a climate change. Environ. Health Perspect. 105: 84-93.
March 2001
4E-5
DRAFT-DO NOT QUOTE OR CITE

-------

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
  5. HUMAN EXPOSURE TO PARTICULATE MATTER
                       AND ITS CONSTITUENTS
5.1  INTRODUCTION
5.1.1 Purpose
     Exposure is defined as the contact by an individual with a pollutant for a specific duration
of time at a visible external boundary (modified from Duan 1982,1991). For airborne particulate
matter (PM), the breathing zone is considered the point of contact and the lung is the external
boundary of concern.  An individual's exposure is measured as the PM air concentration in
his/her breathing zone over time.  Understanding exposure is important, because it is individuals
who experience adverse health effects associated with elevated PM concentrations in ambient air.
     The U.S. Environmental Protection Agency's (EPA's) regulatory authority for PM applies
primarily to ambient air and those sources that contribute to ambient PM air concentrations.
»
Thus, a major emphasis must be to develop an understanding of exposure to PM from ambient
sources. However, personal exposure to total PM may result from exposure to PM from both
ambient and nonambient sources. Therefore, it will be necessary to account for both in order to
fully understand the relationship between PM and health effects. Personal exposure to PM from
nonambient sources may be a confounder in community-based epidemiological studies in which
ambient PM measures are correlated with community health parameters. In addition, an
individual's personal exposure to ambient, nonambient, and total PM would provide useful
information for studies where health outcomes are tracked individually.
     The overall purpose of this chapter is to provide current exposure information that will aid
in the understanding and interpretation of PM dosimetry, toxicology, and epidemiology studies
assessed in later chapters. The specific objectives of this chapter, which are described below, are
fourfold.
(1) To provide an overall conceptual framework of exposure science as applied to PM, including
   the identification and evaluation of factors that determine personal exposure to total PM and
   to PM from ambient and nonambient PM sources
       March 2001
                                         5-1
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15,
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
(2) To provide a concise summary and review of recent data (since 1996) and findings from
    pertinent studies of personal exposure to total PM and to PM from ambient and nonambient
    sources
(3) To characterize quantitative relationships between ambient air quality measurements (mass,
    chemical components, number, etc.), as determined by a community monitoring site, and
    total personal PM exposure as well as its ambient and nonambient components
(4) To evaluate the implications of using ambient PM concentrations as a surrogate for exposure
    in epidemiologicaL studies of PM health effects

5.1.2 Particulate Matter Mass and Constituents
     Current EPA PM regulations are based on mass as a function of aerodynamic size.
However, EPA also measures the chemical composition of PM in both monitoring and research
studies.  The composition of PM is variable and adverse health effects may be related to PM
characteristics other than mass. PM from ambient and nonambient sources also may have
differing physical and chemical characteristics and differing health effects. Ultimately, to
understand and control health impacts caused by PM, it is important to quantify and understand
exposure to those chemical constituents responsible for the adverse health effects.  The National
Research Council (NRC) recognized the distinction between measuring exposure to PM mass
and to chemical constituents when setting Research Priorities for Airborne Particulate Matter I:
Immediate Priorities and a Long-range Research Portfolio (NRC, 1998). Specifically, NRC
Research Topic 1 recommends evaluating the relationship between outdoor measures versus
actual human exposure for PM mass. The NRC Research Topic 2 recommends evaluating
exposures to biologically important constituents and specific characteristics of PM that cause
responses in potentially susceptible subpopulations and the general population. It also was
recognized by the NRC that, "a more targeted set of studies under this research topic (#2) should
await a better understanding of the physical, chemical, and biological properties of airborne
particles associated with the reported mortality and morbidity outcomes" (NRC, 1999). The
NRC also stated that the studies "should be designed to determine the extent to which members
of the population contact these biologically important constituents and size fraction of concern in
outdoor air, outdoor air that has penetrated indoors,  and air pollutants generated indoors" (NRC,
1999). Thus, when biologically important constituents are identified, exposure studies should
        March 2001
                                          5-2
DRAFT-DO NOT QUOTE OR CITE

-------
  1      include contributions from all sources. The emphasis in this chapter on PM mass reflects the
  2      current state of the science. Where available, data also have been provided on chemical
  3      constituents, although in most cases, the data are limited.  As recognized by the NRC, a better
  4      understanding of exposures to chemical constituents will be required to more fully identify,
  5      understand, and control those sources of PM with adverse health effects and to accurately define
  6      the relationship between PM exposure and health outcomes.
  7
  8      5.1.3  Relationship to Past Documents
  9           Early versions of PM criteria documents  did not emphasize total human exposure but rather
 10      focused almost exclusively on outdoor air concentrations. For instance, the 1969 Air Quality
 11      Criteria for Particulate Matter (PM AQCD) (National Air Pollution Control Administration,
 12      1969) did not discuss either exposure or indoor concentrations. The 1982 PM AQCD (U.S.
 13      Environmental Protection Agency, 1982) provided some discussion of indoor PM concentrations,'
 14      reflecting an increase in microenvironmental and personal exposure studies. The new data
 15      indicated that personal activities, along with PM generated by personal and indoor sources (e.g.,
 16      cigarette smoking), could lead to high indoor levels and high personal exposures to total PM.
 17      Some studies reported indoor concentrations that exceeded PM concentrations found in the air
 18      outside the monitored microenvironments or at nearby monitoring sites.
 19           Between 1982 and 1996, many more studies of personal and indoor PM exposure
20      demonstrated that, in most inhabited domestic environments, indoor PM concentrations and
21      personal PM exposures of the residents were greater than ambient PM concentrations measured
22      simultaneously (e.g., Sexton et al., 1984; Spengler et al., 1985; Clayton et al., 1993). As a result,
23      the NRC (1991) recognized the potential importance of indoor sources of contaminants
24      (including PM) in causing adverse health outcomes.
25           The 1996 AQCD (U.S. Environmental Protection Agency, 1996) reviewed the human PM
26      exposure literature through early 1996. Many of the studies cited showed poor correlations
27      between personal exposure or indoor measurements of PM and outdoor or ambient site
28      measurements.  Conversely, Janssen et al. (1995)  and Tamura et al. (1996a) showed that in the
29      absence of major nonambient sources, total PM exposures to individuals tracked through time
30      were highly correlated with ambient PM concentrations. Analyses of these latter two studies led
31      to consideration of ambient and nonambient exposures as separate components of total personal
        March 2001                               5-3        DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
exposure.  As a result, the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996), for
the first time, distinguished between ambient and nonambient PM exposure. This chapter builds
on the work of the 1996 PM AQCD by further evaluating the ambient and nonambient
components of PM, as well as reporting research that evaluates the relationship between ambient
concentrations and total, ambient, and nonambient personal exposure.
5.2  STRUCTURE FOR THE CHAPTER
     The chapter is organized to provide information on the principles of exposure, review the
existing literature, and summarize key findings and limitations in the information; the specific
sections are described below.
• Section 5.3 discusses the basic concepts of exposure, including definitions, methods for
  estimating exposure, and methods for estimating ambient components of exposure.
• Section 5.4 presents PM mass data, including a description of  the key available studies,
  correlations of PM exposures with ambient concentrations, and factors that effect the
  correlations.
• Section 5.5 presents data on PM constituents, including a description of the key available
  studies, correlations with ambient concentrations, and factors that effect the correlations.
• Section 5.6 discusses the implications of using ambient PM concentrations in epidemiological
  studies of PM health effects.
• Section 5.7 summarizes key findings and limitations of the information.
 5.3  BASIC CONCEPTS OF EXPOSURE
 5.3.1 Components of Exposure
      The total exposure of an individual over a discrete period of time includes exposures to
 many different particles from various sources while in different microenvironments 0/e's). Duan
 (1982) defined a microenvironment as "a [portion] of air space with homogeneous pollutant
 concentration." It also has been defined as a volume in space, for a specific time interval, during
 which the variance of concentration within the volume is significantly less than the variance
       March 2001
                                         5-4
DRAFT-DO NOT QUOTE OR CITE

-------
 1     between that //e and surrounding ^ue's (Mage, 1985). In general, people pass through a series of
 2     /^e's, including outdoor, in-vehicle, and indoor //e's, as they go through time and space.  Thus,
 3     total daily exposure for a single individual to PM can be expressed as the sum of various
 4     microenvironmental exposures that the person occupies in the day (modified from National
 5     Research Council, 1991).
 6          In a given //e, particles may originate from a wide variety of sources.  For example, in an
 7     indoor /we, PM may be generated by (1) indoor activities, (2) outdoor PM entering the indoor /^e,
 8     (3) the chemical interaction of outdoor air pollutants and indoor air or indoor sources,
 9     (4) transport from another indoor //e, or (5) personal activities. All of these disparate sources
10     have to be accounted for in a total human PM exposure assessment.
11          An analysis of personal exposure to PM mass (or constituent compounds) requires
12     definition and discussion of several classes of particles and exposure. In this chapter, PM
13     metrics may be described in terms of exposure or as an air concentration. PM also may be
14     described according to both its source (i.e., ambient, nonambient) and the microenvironment
15     where exposure occurs.  Table 5-1 provides a summary of the terms used in this chapter, the
16     notation used for these terms, and their definition. These terms will be used throughout this
17     section and will provide the terminology for evaluating personal exposure to total PM and PM
18     from ambient and nonambient sources.
19
20     5.3.2  Methods To Estimate Personal Exposure
21          Personal exposure may be estimated using either direct or indirect approaches. Direct
22     approaches measure the contact of the person with the chemical concentration in the exposure
23     media over an identified period of time. Direct measurement methods include personal exposure
24     monitors (PEMs) for PM that are worn continuously by individuals as they encounter various
25     microenvironments and perform their daily activities. Indirect approaches use available
26     information on concentrations of chemicals in microenvironments, along with information about
27     the time individuals spend in those microenvironments and personal PM generating activities.
28     The indirect approach then uses models and data on microenvironmental air concentrations and
29     time spent in microenvironments to estimate personal exposure. This section describes the
30     methods to directly measure personal exposures and microenvironmental concentrations, as well
        March 2001
5-5
DRAFT-DO NOT QUOTE OR CITE

-------
                TABLE 5-1.  CLASSES OF PARTICULATE MATTER EXPOSURE AND
               	CONCENTRATION DEFINITIONS	
         Term
                        Notation
                           Definition
Concentration

Personal Exposure


Microenvironment


Ambient PM
Ambient-Outdoor PM
Indoor PM
Ambient-Indoor PM
Indoor-Generated PM
         Personal Exposure to
         Indoor-Generated PM
         Personal Exposure to
         Ambient-Generated PM
Personal Exposure to
Personal-Activity PM
Personal Exposure to
Nonambient PM
                                C

                                E


                                //e


                                Ca
                                Cao
                                C,
                                Cai
                                Epact
                      General Definitions
Air concentration of PM in a given microenvironment, expressed in
/zg/m3
Contact at visible external boundaries of an individual with a pollutant
for a specific duration of time; quantified by the amount of PM available
in concentration units 0"g/m3) at the oral/nasal contact boundary for a
specified time period (At). General term for any exposure variable.
Volume in space, for a specific time interval, during which the variance
of concentration within the volume is significantly less than the variance
between that /ze and surrounding /zes
                    Concentration Variables
PM in the atmosphere measured at a community ambient monitoring site
either emitted into the atmosphere directly (primary PM) or formed in it
(secondary PM). Major sources of PM species are industry, motor
vehicles, commerce, domestic emissions such as wood smoke, and
natural wind-blown dust or soil.
Ambient PM in an outdoor microenvironment
All PM found indoors
Ambient PM that has infiltrated indoors (i.e., has penetrated indoors and
remains suspended)
PM generated or formed indoors
                      Exposure Variables
Sum of personal exposure resulting from indoor-generated PM
Sum of personal exposure caused by ambient-outdoor and ambient
indoor PM (does not include resuspended ambient PM previously
deposited indoors)
Small-scale PM-generating activities that primarily influence exposure of
the person performing the activity itself
Sum of personal exposure to indoor-generated and personal activity PM
     = Ei + Eact
         Personal Exposure to
         Total PM
                                 Sum of all personal exposures to ambient and nonambient PM
1      as the models used to estimate exposure. Several approaches to estimate personal exposure to
2      ambient PM also are described.
       March 2001
                                              5-6
                         DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 5.3.2.1 Direct Measurement Methods
 5.3.2.1.1 Personal Exposure Monitoring Methods
      In theory, personal exposure to total PM is measured by sampling the concentration of PM
 in inhaled air entering the nose or mouth. Practically, it is defined as that PM collected by a
 PEM worn by a person and sampling from a point near the breathing zone (but not impacted by
 exhaled breath). The inlet to a PEM normally is placed at the outer limit of the breathing zone to
 avoid a negative sampling bias resulting from dilution of the collected air by exhaled breath
 depleted of PM. However, such placement does not allow for the sampling of directly inhaled
 cigarette smoke or inhaled air that passes through a dust mask.  PEMs for PM use measurement
 techniques similar to those used for ambient PM. The PEM is a filter-based mass measurement
 of a particle size fraction (PM10 or PM2 5), usually integrated over either a 24- or 12-h period at
 flow rates of 2 to 4 L/min using battery-operated pumps. PEMs must be worn by study
 participants and, therefore, they must be quiet, compact, and battery-operated. These
 requirements limit the type of pumps and the total sample volume that can be collected.
 Generally, small sample volumes  limit personal exposure measurements to PM mass and a few
 elements detected by XRF. In most studies, PM2 5 and PMIO have not been collected
 concurrently.
     Other methods used for ambient PM also have been adapted for use as a  personal exposure
 monitor. For example, a personal nephelometer that measures particle number within a specific
 particle size range using light scattering has been used in personal exposure studies to obtain
 real-time measurements of PM.

 5.3.2.1.2 Microenvironmental Monitoring Methods
     Direct measurements of microenvironmental PM concentrations, which are used with
 models to estimate personal exposure to PM, also use methods similar to those for ambient PM.
 These methods differ from PEMs in that they are stationary with respect to the microenvironment
 (such as a stationary PEM). Microenvironmental monitoring methods include  filter-based mass
measurements of particle size fractions (PM,0, PM2 5), usually integrated over either a 24- or 12-h
period. Flow rates vary between various devices from 4 to 20 L/min. Larger sample volumes
allow more extensive chemical characterization to be conducted on microenvironmental samples.
Because more than one pumping system can be used in a microenvironment, PM2 5 and PM10 can
       March 2001
                                         5-7
DRAFT-DO NOT QUOTE OR CITE

-------
 1      be collected simultaneously. Other continuous ambient PM measurement methods that have
 2      been utilized for microenvironmental monitoring are the Tapered Element Oscillating
 3      Microbalance (TEOM) and nephelometers. Various continuous techniques for counting particles
 4      by size also have also been used (Climet, LASX, SMPS, APS). Measurement techniques are
 5      discussed in Chapter 2.
 6
 7      5 3.2.2 Indirect Methods (Modeling Methods)
 8      5.3.2.2.1 Personal Exposure Models
 9           Exposure modeling for PM2.5 mass and chemical constituents is a relatively new field
10      facing significant methodological challenges and input data limitations. Exposure models
11      typically use one of two general approaches: (1) a time-series approach that estimates
12     microenvironmental exposures sequentially as individuals go through time or (2) a time-averaged
13     approach that estimates microenvironmental exposures using average microenvironmental
14     concentrations and the total time spent hi each microenvironment. Although the time-series
15     approach to modeling personal exposures provides the appropriate structure for accurately
16     estimating personal exposures (Esmen and Hall, 2000; Mihlan et al., 2000), a time-averaged
17     approach typically is used when the input data needed to support a time-series model are not
18     available. In addition, the time-varying dose profile of an exposed individual can be modeled
19     only by using the time-series approach (McCurdy, 1997, 2000). We define the personal
20     exposure of an individual to a chemical in air to be (NRC, 1991)
21
                                            E=
22      where
23           E is the personal exposure during the time period from t, to t,, and
24           C(t) is the concentration near the nose and mouth not impacted by
25           exhaled air, at time t.
26           In general, personal exposure models combine microenvironmental concentration data with
27      human activity pattern data to estimate personal exposures. Time-averaged models also can be
28      used to estimate personal exposure for an individual or for a defined population. Total personal
        March 2001
5-8
DRAFT-DO NOT QUOTE OR CITE

-------
 1      exposure models estimate exposures for all of the different microenvironments in which a person
 2      spends time, and total average personal exposure is calculated from the sum of these
 3      microenvironmental exposures:
                                                                                         (5-1)
 4
 5      where Ey is the personal exposure in each microenvironment,/ (Duan, 1982). Example
 6      microenvironments include outdoors, indoors at home, indoors at work, and in transit. Each
 7      microenvironmental exposure, E;, is calculated from the average concentration in
 8      microenvironment/,  Cj , weighted by the time spent in microenvironment/, t;. T is the sum of t,
 9      over all/. It is important to note that, although measurement data may be an average
10      concentration over some time period (i.e., 24 h), significant variations in PM concentrations can
11      occur during that time period. Thus, an error may be introduced if real-time concentrations are
12      highly variable, and an average concentration for a microenvironment is used to estimate
13      exposure when the individual is in that microenvironment for only a fraction of the total time.
14      This exposure formulation has been applied to concentration data in a number of studies (Ott,
15      1984; Ott et al., 1988, 1992; Miller et al., 1998; Klepeis et al., 1994; Lachenmyer and Hidy,
16      2000).
17          Microenvironmental concentrations used in the exposure models can be measured directly
18      or estimated from one or more microenvironmental models. Microenvironmental models vary in
19      complexity, from a simple indoor/outdoor ratio to a multi-compartmental mass-balance model.
20      A discussion of microenvironmental models is presented below in Section 5.3.2.2.2.
21          On the individual level, the time spent in the various microenvironments is obtained from
22      time/activity diaries that are completed by the individual.  For population-based estimates, the
23      time spent in various microenvironments is obtained from human activity databases.  Many of
24      the largest human activity databases have been consolidated by EPA's National Exposure
25      Research Laboratory (NERL) into one comprehensive database called the Consolidated Human
26      Activity Database (CHAD). CHAD contains over 22,000 person-days of 24-h activity data from
27      11 different human activity pattern studies. Population cohorts with diverse characteristics can
       March 2001
5-9
DRAFT-DO NOT QUOTE OR CITE

-------
 1     be constructed from the activity data in CHAD and used for exposure analysis and modeling
 2     (McCurdy, 2000). Table 5-2 is a summary listing of the human activity studies in CHAD.
 3          Methodologically, personal exposure, models can be divided into three general types:
 4     (1) statistical models based on empirical data obtained from one or more personal monitoring
 5     study, (2) simulation models based upon known or assumed physical relationships, and
 6     (3) physical-stochastic models that include Monte Carlo or other techniques to explicitly address
 7     variability and uncertainty in model structure and input data (Ryan,  1991; Macintosh et al.,
 8     1995). The attributes, strengths, and weaknesses of these model types are discussed by Ryan
 9     (1991), National Research Council (1991), Frey and Rhodes (1996), and Ramachandran and
10     Vincent (1999).  GIS-based approaches to estimate health risks of environmental concentrations
11     also have been developed (e.g., Beyea and Hatch, 1999; Jensen, 1999). A recent summary
12     review of the logic of exposure modeling is found in Klepeis (1999).
13          Personal exposure models that have been developed for PM are summarized in Table 5-3.
14     The regression-based models (Johnson et al., 2000; Janssen et al., 1997; Janssen et al., 1998a)
15     were developed for a specific purpose (i.e., to account for the observed difference between
16     personal exposure and microenvironmental measurements) and are based on data from a single
17     study, which limits their utility for broader purposes. Other types of models in Table 5-3  were
18     limited by a lack of data for the various model inputs. For example, ambient PM monitoring data
19     is not generally of adequate spatial and temporal resolution for these models.  Lurmann and Korc
20     (1994) used site-specific coefficient of haze (COH) information to stochastically develop a time
21     series of 1-h PM10 data from every sixth day 24-h PM10 measurements. A mass-balance model
22     typically was used for indoor microenvironments when sufficient data was available, such as for
23     a residence. For most other microenvironments, indoor/outdoor ratios were used because of the
24     lack of data for a mass-balance model. In addition, only the deterministic model PMEX included
25     estimation of inhaled dose from activity-specific breathing rate information. Data from recent
26     PM personal exposure and microenvironmental measurement studies will help facilitate the
27     development of improved personal exposure models for PM.
28          An integrated human exposure source-to-dose modeling system that will include exposure
29     models to predict population exposures to environmental pollutants such as PM currently is
30     being developed by NERL. A first-generation population exposure model for PM, called the
31     Stochastic Human Exposure and Dose Simulation (SHEDS-PM) model, recently has been
        March 2001
5-10
DRAFT-DO NOT QUOTE OR CITE

-------
ati

ce
           Is,
Calen
Time
          II
                i.
                              D
                              a.
                il
                •o „
                u~
                tl
                |£-
                ,2 o
               •3
               •s
                -
               ON
               ON
rt of COP
ys; 55 P-D
P-D
OO ^—.  ,-x
ON J3  «

^ON  ON
_; ON  ON

rt ^-  ^^

| "3  73

§ S  o

1^  S>


5i  '^
                    z1
—    «2
¥    5¥
Akl
Joh
4h St
h Sta
                              73 S  73 E
                                   § S
Standard
                              5
                         —    \o
                         _    oo
               vo    —•
r
85
               e.^  -Z
               rH _L  *r? Q.
                    22  22  <:
           2  2*
                              If II  I

                   s-s

                   IS

                             s

                             !
                             o.
ati
days
                                     a
                                     c-
                                     *0

                                     •&

                                     
-------
1
z
1
     1
            4.
            X
            tu
                          I
                          GO
                     X
                     u

                     S
                     oi
                                 UJ
            *C3

            §
                 I
-day
                                                   §
                                                   M
                                                   *
                                                 ;S.8 §
                                         S
                                         00

                                                             11111:1
                                                        i
Distributioi

population
                                                              !l
                                                          ™J3' § " S
                                                                 ^ S o
                                                                .fill
S
I
00
                                         o
March 2001
                             5-12
                                     DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
developed. The SHEDS-PM model uses a 2-stage Monte Carlo sampling technique previously
applied by Macintosh et al. (1995) for benzene exposures. This technique allows for separate
characterization of variability and uncertainty in the model predictions (to predict the distribution
of total exposure to PM for the population of an urban/metropolitan area and to estimate the
contribution of ambient PM to total PM exposure).  This model is yet to be evaluated and is
discussed for information purposes only because results from the case study have been only
recently reported in a journal article submitted for peer review (Burke et al., 2001).

5.3.2.2.2 Microenvironmental Models
     The mass balance model has been used extensively in exposure analysis to estimate PM
concentrations in indoor microenvironments (Calder, 1957; Sexton and Ryan, 1988; Duan, 1982,
1991; McCurdy, 1995; Johnson, 1995; Klepeis et al., 1995; Dockery and Spengler, 1981; Ott,
1984; Ott et al., 1988, 1992, 2000; Miller et al., 1998; Mage et al., 1999; Wilson et al., 2000).
The mass balance model describes the infiltration of particles from outdoors into the indoor
microenvironment and the generation of particles from indoor sources:
V
                                = vPCa-vQ-kVCi+Qi,
                                                                    (5-2)
where
V
           Ca
           k
  volume of the well-mixed indoor air (cubic meters),
  concentration of indoor PM;
  volumetric air exchange rate between indoors and outdoors (cubic
  meters per hour);
  penetration ratio, the fraction of ambient (outdoor) PM that is not
  removed from ambient air during its entry into the indoor volume;
  concentration of PM in the ambient air (micrograms per cubic meter);
  removal rate (per hour); and
  indoor sources of particles (micrograms per hour).
     Qi contains a variety of indoor, particle-generating sources, including combustion or
mechanical processes, condensation of vapors formed by combustion or chemical reaction,
       March 2001
                                         5-13
                                 DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
suspension from bulk material, and resuspension of previously deposited PM.  The removal rate,
k, includes dry deposition to interior surfaces by diffusion, impaction, electrostatic forces, and
gravitational fallout. It may include other removal processes such as filtration by forced air
heating, ventilation, or air-conditioning (HVAC) or by independent air cleaners. All parameters
except V are functions of time.  P and k also are functions of particle aerodynamic diameter
andv.
     In addition to the mass balance model, a number of single-source or single-
microenvironment models exist. However, most are used to estimate personal exposures to
environmental tobacco smoke (ETS). These models include both empirically based statistical
models and physical models based on first principles; some are time-averaged, whereas others
are time-series.  These models evaluate the contribution of ETS to total PM exposure in an
enclosed microenvironment and can be applied as activity-specific components of total personal
exposure models. Examples of ETS-oriented personal exposure models are Klepeis (1999),
Klepeis et al. (1996,2000), Mage and Ott (1996), Ott (1999), Ott et al. (1992,  1995),  and
Robinson etal. (1994).

5.3.2.3  Methods of Estimating Personal Exposure to Ambient Particulate Matter
     In keeping with the various components of PM exposure described above in Section 5.3.1,
personal exposure to PM can be expressed as the sum of exposure to particles from different
sources summed over all microenvironments in which exposure occurs. Total personal exposure
may be expressed as
                                     Et — Eag  + Eig + Epact
                                     Et = Eag + Ei
                                                                                         (5-3)
                                           .nonag,
where Et is the total personal exposure to ambient and nonambient PM, Eag is personal exposure
to ambient PM (the sum of ambient PM while outdoors and ambient PM that has infiltrated
indoors, while indoors), Eig is personal exposure to indoor-generated PM, Epact is personal
exposure to PM from personal activity, and Enonag is personal exposure to nonambient PM.
Although personal exposure to ambient and nonambient PM cannot be measured directly, they
       March 2001
                                         5-14
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13

 14
 can be calculated or estimated from other measurement data. Approaches for estimating these
 components of PM exposure are described in the following section.

 5.3.2.3.1 Mass Balance Approach
 Ambient-Indoor Concentrations of Participate Matter
      The mass balance model described above (Equation 5-2) has been used to estimate PM
 concentrations in indoor microenvironments. This model also may be used to estimate ambient-
 indoor (Cai) and indoor-generated (Cig) PM concentrations. The mass balance model can be
 solved for Cai and Cig assuming equilibrium conditions, and assuming that all variables remain
 constant (Ott et al., 2000; Dockery and Spengler, 1981; Koutrakis et al., 1992). By substituting
 dCai + dCig for dQ in equation 5-2 and assuming dCai and dCig = 0, ambient-indoor PM (Cai) and
 indoor-generated PM (Cig), at equilibrium, are given by
                                                                                  (5-4)
15
16
17
18
19
20
21
22
23
24
25
                                                                                         (5-5)
where a = v/V, the number of air exchanges per hour. Equations 5-4 and 5-5 assume equilibrium
conditions and, therefore, are valid only when the parameters k, a, Cao, and Qi are not changing
rapidly and when the Cs are averaged over several hours. Under certain conditions (e.g.,
air-conditioned homes, homes with HVAC or air cleaners that cycle on and off, ambient
pollutants with rapidly varying concentrations), nonequilibrium versions  of the mass balance
model (Ott et al., 2000; Freijer and Bloeman,  2000; Isukapalli and Georgopoulos, 2000) are
likely to provide a more accurate estimate of Cai and Cig.  However, the equilibrium model
provides a useful, if simplified, example of the basic relationships (Ott et al., 2000).
     Equation 5-4 may be rearranged further  to give Cai/Cao, the equilibrium fraction of ambient
PM that is found indoors, defined as the infiltration factor (F^) (Dockery and Spengler, 1981).
       March 2001
                                         5-15
DRAFT-DO NOT QUOTE OR CITE

-------
                                             Fai
                                       INF = „
                                                  Pa
                                                                                 (5-6)
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22

23
The penetration ratio (P) and the decay rate (&) can be estimated using a variety techniques.
A discussion of these variables and estimation techniques is given in Section 5.4.3.2.2.  Because
both P and k are a function of particle aerodynamic diameter, F^p also will be a function of
particle aerodynamic diameter.

Personal Exposure to Ambient-Generated Participate Matter
     Personal exposure to ambient-generated PM (Eag) may be estimated using ambient-indoor
PM concentration (Cai) from the mass balance model, ambient outdoor PM concentrations (Cao)
and information on the time an individual spent hi the various microenvironments.
Mathematically, this may be expressed as
                                       ag
                                                                                         (5-7)
        is the fraction of time that an individual spent outdoors, and (1 -y) is the fraction of time
spent indoors.
      It is convenient to express personal exposure to ambient generated PM (Eag) as the product
of the ambient PM concentration (Cao or CJ and a personal exposure or attenuation factor.
Following the usage in several recent papers (Zeger et al., 2000; Dominici et al., 2000; Ott et al.,
2000), the symbol a will be used for this attenuation factor.  Equation 5-7 can be rearranged to
obtain an expression for a:
                                                ' Pa 1
                                                a+k}
                            (5-8)
        March 2001
                                          5-16
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
Substituting equation 5-6 in equation 5-8 gives a relationship for a in terms of the infiltration
factor FJNF and the fraction of time spent in the various microenvironments:
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
                                                 -y)
                                                                                  (5-9)
Thus, personal exposures to ambient PM (Eag) may be calculated from measurable quantities:
                                      — OC Cao.
                            (5-10)
The factor a can be measured directly or calculated from measured or estimated values of the
parameters a, k, and P and the time spent in various microenvironments from activity pattern
diaries (Wilson et al., 2000).
     The use of a mass balance model to separate personal exposure into two components
because of exposure to ambient and nonambient concentrations is not novel. This approach,
based on Equation 5-3 as given in Duan (1982) and called superposition of component
concentrations, has been applied using multiple microenvironments to carbon monoxide (Ott,
1984; Ott et al., 1988, 1992), volatile organic compounds (Miller et al., 1998), and particles
(Koutrakis et al., 1992; Klepeis et al., 1994).  However, in these studies, and in most of the
exposure literature, the ambient and nonambient components are added to yield a personal
exposure from all sources of the pollutant.  The use of the mass balance model, ambient
concentrations, and exposure parameters to estimate exposure to ambient-generated PM and
exposure to indoor-generated PM separately as different classes of exposure has been discussed
in Wilson  and Suh (1997) and in Wilson et al. (2000).

5.3.2.3.2 Tracer Species as Surrogates of Ambient-Generated Particulate Matter
     The ratio of personal exposure to ambient concentration for a PM component that has no
indoor sources may be used as a measure of the ratio of personal exposure to ambient PM to the
ambient concentration of PM for PM of similar aerodynamic diameter (Wilson et al., 2000).
Sulfate, in particular,  often is used as a marker of outdoor air in indoor microenvironments
        March 2001
                                         5-17
DRAFT-DO NOT QUOTE OR CITE

-------
 1      (Jones et al., 2000). It is found primarily in the PM2 5 fraction of the aerosol (Cohen et al, 2000).
 2      Ozkaynak et al. (1996a, b) and Janssen et al. (1999a) report, in the PTEAM and Netherlands
 3      studies respectively, that XRF analyses of indoor PM and the immediate outdoor PM show that
 4      sulfur is the only element reported with virtually identical mass concentrations in both indoor and
 5      outdoor air. Therefore, where there are no indoor sources of fine-mode sulfates, one may deduce
 6      that the ambient-to-personal relationship found for sulfates probably would be the same as that
 7      for unspeciated particulate matter of the same aerodynamic size range.  This assumption has not
 8      been validated, however, and ambient PM with different physical or chemical characteristics may
 9      not behave similarly to sulfate.
10           Particulate sulfate is formed in the ambient air via photochemical oxidation of gaseous
11      sulfur dioxide arising from the primary emissions from the combustion of fossil fuels  containing
12      sulfur.  They also arise from the direct emissions of sulfur-containing particles from
13      nonanthropogenic sources (e.g., volcanic activity, wind-blown soil). In the indoor environment,
14      the only common sources of sulfate may be resuspension by human activity of deposited PM
15      containing ammonium sulfates or soil sulfates that were tracked into the home. In some homes
16      an unvented kerosene heater using a high-sulfur fuel may be a major contributor during winter
17      (Leaderer et al., 1999). Use of matches to light cigarettes or gas stoves are also a source of
18      sulfates. Studies that have used sulfate as a surrogate for ambient PM are discussed in
19      Section 5.4.3.1 (i.e., Oglesby et al., 2000a; Sarnat et al., 2000; Ebelt, 2000).
20
21      5.3.2.5.3 Source-Apportionment Techniques
22           Source apportionment techniques provide a method for determining personal exposure to
23      PM from specific sources. If a sufficient number of samples are analyzed with sufficient
24      compositional detail, it is possible to use statistical techniques to derive source category
25      signatures, identify indoor and outdoor source categories, and estimate their contribution to
26      indoor and personal PM. Daily contributions from sources that have no indoor component can
27      be used as tracers to generate exposure to ambient PM of similar aerodynamic size or  directly as
28      exposure surrogates in epidemiologic analyses.  Studies that have used source-apportionment are
29      discussed in Section 5.4.3.3 (i.e., Ozkaynak and Thurston, 1987; Yakovleva et al., 1999; Mar
30      et al. 2000; Laden et al., 2000).
31
        March 2001
5-18
DRAFT-DO NOT QUOTE OR CITE

-------
  1      5.4  SUMMARY OF PARTICULATE MATTER MASS DATA
  2      5.4.1 Types of Particulate Matter Measurement Studies
  3           A variety of field measurement studies have been conducted to quantify personal exposure
  4      to PM mass, measure microenvironmental concentrations of PM, and evaluate the relationship
  5      between personal exposure to PM and PM air concentrations measured at ambient sites.
  6      In general, exposure measurement studies are of two types depending on how the participants are
  7      selected for the study. In a probability study, participants are selected using a probability
  8      sampling design where every member of the defined population has a known, positive probability
  9      of being included into the sample. Probability study results can be used to make statistical
10      inferences about the target population. In & purposeful or nonprobability design, any convenient
11      method may be used to enlist participants and the probability of any individual in the population
12      being included in the sample is unknown.  Participants in purposeful samples (also referred to as
13      a "convenience" samples) may not have same the characteristics that would lead to exposure as
14      the general population.  Thus, results of purposeful studies apply only to the subjects sampled on
15      the days that they were  sampled. In a purposeful study, statistically valid inferences cannot be
16      made to any other population or period of time. Although such studies may report significant
17      differences, confidence intervals, andp values, they have no inferential validity (Lessler and
18      Kalsbeek, 1992). However, most purposeful studies of PM personal exposure can provide data to
19      develop relationships on important exposure factors and useful information for developing and
20      evaluating either statistical or physical/chemical human exposure models.
21           Regardless of the  sampling design  (probability or purposeful) there are three general
22      categories of study design that can be used to measure personal exposure to PM and evaluate the
23      relationship between personal PM exposure levels and ambient PM concentrations measured
24      simultaneously: (1) longitudinal, (2) daily-average, and (3) pooled.  These are discussed in
25      Section 5.4.3.1.1.
26
27      5.4.2 Available Data
28      5.4.2.1  Personal Exposure Data
29           Table 5-4 gives an overview of the personal exposure studies that have been conducted and
30      are reviewed in this section.  This includes studies that have been reported since the 1996 AQCD.
       March 2001
5-19
DRAFT-DO NOT QUOTE OR CITE

-------











_
H
02
P
O

s>
w
o
C/3
05
6-J
PH
1
S
§
g
o
1
en
"T

W
S
H






8

0
^

^
1 *fl
^ *M
i1
'i
s
S
55
S
cu
•i
l|
S
DU


It

f
£
5
t
t/5

o |
o S1

•0
Study Location an
Population
&
Study Des



















en
 a ^s S
^5 -a •« g g -s g
OO O •? O cu  Z en

2 ^
= a 3"= 2
S S SS 2
cu cu cu cu Cu

0" cu 0" 0
•J* H-T ^T ^-T
cu cC cu oT


 Q>
"5*3 *o *o
So o S
cu a. cu a,



















CA

-3
S

1
1
(S












• _

js. §
£" ^, g "~T o


^ ^ T3 . Ct 2 J T3 g
"S3 "a> w4_.'rt2'crt
g | i s « 1 | s
II 1 1 1 I 1 1
A
O
z
os o
8S 8

g2 1 igs g g i
5 ^ 555^^55
S S S)SSS*S2S
ft. ftt O-, O-r G« D^ O^ O-t D-

« o o" o"cf o o o
< «r _r _r ,-r i-r »-r *-? <3
cu" &T cu" aT cu" cu" cu* cu* cu*


II g SS •* 2 • « -
0-1 Tf m J2 OO ^O
il S s " 33
g- - | il i
2°* oo o\'^1^ 2T!~ §•
1 If 1 11 Ji II 1 !l

co <^ 3C ^^ v^ ^o o oo o ^o

5 §
•3 t Q £ 1 -
' ill Hti ll II js It if ll
c c G c c c c cj
I a alalal
'ob *ob "So 'ob oo CD 5o DO
33 333333
March 2001
5-20
DRAFT-DO NOT QUOTE OR CITE

-------








G
8
_c
co
1
H
t/5
fjj
s
£3
00
O
td
SONAL]
K
g
PH
H
U


O
fe

53
1^
1
**.
^

Is
0
u
TABLE 5-4








8
P
£
2 «a
— is
o
1
TO
U
2
8
OT
s
a.
"S
|S
Q w

0>
OX)


T3
O
1
>»
1
t/3
°f
6 ^o*



•a
Study Location an
Population
§>
%
Q
t
OT




















5
is
ieful Studies (
§.
1

















r-
o\
ON
O
"3

u
1







0
s"
Cu

P, A, School
oo

2
o


to
2
f
1

^




Amsterdam and
Wageningen, Neth.,
school children
I
•a
i?
oo ^-^
OS O

c. §,
*t5 »-:
t3 g
4J •*-*
1 1
C?
OT
li

0
Jg"
(X
o "S
s s"
0. 0.

oT cu
S 2
oo

— -H

oo"

*" Os
•* i c"
OS S O.

en — '




Amsterdam, adults
Baltimore, elderly
subjects
c c
I I





^
"3 '

S
1


o1




_
Cu
OT

O
— "
oT
rn








s
o\

oo




Tokyo, Japan,
elderly housewives
c
•3
1
s

ON OO
O\ Os

• "3 —
w «
S "

1 1

Q-
aO
o
OH
0
S
°i 2

OH CM

01
.£
o 1 g" 1
E!E 1
Cw, cu O O
f, ~
It




o! fc

o 3 |
O\ "^"S
S C *o
ca o. c
u- &o 5

*«o o
(N O




8
c 1
1 f.
13
-S
3
I 1
































1
ja
TO

HjJ
vimetric measurements.
: indoors, O = outdoors,
Si
8 I
S 8.
March 2001
5-21
DRAFT-DO NOT QUOTE OR CITE

-------
 1     Major studies that were reported before that time also have been included to provide a
 2     comprehensive evaluation of data in this area. Table 5-4 gives information on the sampling and
 3     study designs, the study population, the season, number of participants, PM exposure metric, and
 4     the PM size fraction measured.
 5          Although there are a number of studies listed in the table, the data available to answer the
 6     important questions related to exposure are limited. Few are based on probability sampling
 7     designs that allow study results to be inferred to the general population. Unfortunately, none of
 8     these probability studies uses a longitudinal study design.  This limits our ability to provide
 9     population estimates on the relationship between personal  PM exposures and ambient site
10     measurements. In addition, most of the probability studies of PM exposure were conducted
11     during a single season, thus variations in ambient concentrations, air exchange rates, and
12     personal activities are not accounted for across seasons. In these cases, study results are only
13     applicable to a specific time period. Longitudinal studies,  on the other hand, generally have
14     small sample sizes and use a purposeful sampling design.  Many of these studies did not include
15     ambient site measurements to allow comparisons with the  exposure data, and approximately half
16     of these studies monitored PM25.
17          Four large-scale probability studies that quantify personal exposure to PM under normal
18     ambient source conditions have been reported in the literature. These include the EPA's Particle
19     Total Exposure Assessment Methodology (PTEAM) study (Clayton et al., 1993; Ozkaynak et al.,
20     1996a,b); the Toronto, Ontario, study (Clayton et al., 1999a and Pellizzari et al.,  1999); the Air
21     Pollution Exposure Distribution within Adult Urban Populations in Europe (EXPOLIS) exposure
22     study (Jantunen et al., 1998, 2000; Oglesby, et al., 2000); and a study of a small, highly polluted,
23     area in Mexico City (Santos-Burgoa et al., 1998). Only preliminary results have been reported
24     for the EXPOLIS study.  A fifth study conducted in Kuwait during the last days of the oil-well
25     fires (Al-Raheem et al., 2000) is not reported here because the ambient PM levels were not
26     representative of normal  ambient source conditions.
27          Recent longitudinal exposure studies have focused on potentially susceptible
28     subpopulations such as the elderly with preexisting respiratory and heart diseases (hypertension,
29     chronic obstructive pulmonary disease, and congestive heart disease). This is in keeping with air
30     pollution analyses that indicate mortality associated with high levels of ambient PM2 5 is greatest
31     for elderly people with cardiopulmonary disease (U.S. Environmental Protection Agency, 1996).
        March 2001
5-22
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Longitudinal studies were conducted in the Netherlands by Janssen (1998) and Jassen et al.
(1997, 1998a,b, 1999a,b) on purposefully selected samples of adults (50 to 70 years old) and
children (10 to 12 years old).  Several additional studies have focused on nonsmoking elderly
populations in Amsterdam and Helsinki (Janssen et al., 2000), Tokyo (Tamura et al.,  1996a),
Baltimore (Liao et al., 1999; Williams et-al., 2000a,b,c), and Fresno, CA (Evans et al. 2000).
These cohorts were selected because of the low incidence of indoor sources of PM (such as
combustion or cooking).  This should allow an examination of the relationship between personal
and ambient PM concentrations without the large influences caused by smoking, cooking, and
other indoor particle-generating activities. The EPA has a research program focused on
understanding PM exposure characteristics and relationships. Within the program, longitudinal
studies are being conducted on elderly participants with underlying heart and lung disease
(COPD, patients with cardiac  defibrillator, and myocardial infarction), an elderly environmental
justice cohort, and asthmatics. These studies are being conducted in several cities throughout the
United States and over several seasons.  Only preliminary data are currently available, and results
are not reported in this document.
      A series of studies by Phillips et al. (1994, 1996, 1997a,b, 1998a,b, 1999) examined
personal ETS exposure in several European cities. Participants varied by age and occupation.
Respirable Particulate Matter (RSP) concentrations were reported. These studies are not
included in Table 5-4 because of their focus on ETS exposure, which is not the focus of this
chapter. A small personal exposure study in Zurich, Switzerland, was reported by Monn et al.,
(1997) for PM10. This study also is not listed in Table 5-4 because indoor and outdoor
measurements were not taken  simultaneously with the personal measurements, and other details
of the study were not published.

5.4.2.2 Microenvironmental Data
      Usually, personal PM monitoring is conducted using integrated measurements over a 12- or
24-h period. As such, total PM exposure estimates based on PEM measurements do not capture
data from individual microenvironments. Recent studies have examined PM concentrations in
various microenvironments using a number of different types of instruments ranging from filter-
based to continuous particle monitors. Details on the  instruments used, measurements collected,
and findings of these studies according to microenvironment (residential indoor, nonresidential
       March 2001
                                          5-23
DRAFT-DO NOT QUOTE OR CITE

-------
 1     indoor, and traffic-related) are summarized in Table 5-5. Studies that collected
 2     microenvironmental data as part of a personal exposure monitoring study are summarized in
 3     Table 5-4.  In general, the studies listed in Table 5-5 are relatively small, purposeful studies
 4     designed to provide specific data on the factors that effect microenvironmental concentration of
 5     PM from both ambient and nonambient sources.
 6          Recently published studies have used various types of continuous monitors to examine
 7     particle concentrations in specific microenvironments and resulting from specific activities.
 8     Continuous particle monitors such as the SMPS, APS, and Climet have been used to measure
 9     particle size distributions in residential microenvironments (Abt et al., 2000a; Long et al., 2000a;
10     Wallace et al., 1997; Wallace, 2000a; McBride et al.,  1999; Vette et al., 2001). These studies
11     have been able to assess penetration efficiency for ambient particles and microenvironments
12     indoors as well as penetration factors and deposition rates.  Continuous instruments are also a
13     valuable tool for assessing the impact of particle resuspension caused by human activity.
14     A semi-quantitative estimate of PM exposure can be obtained using personal nephelometers that
15     measure PM using light-scattering techniques. Recent PM exposure studies have used personal
16     nephelometers (1 min avg time) to measure PM continuously (Howard-Reed et al., 2000;
17     Quintana et al., 2000) in various microenvironments.  These data have been used to identify the
18     most important ambient and nonambient sources of PM, to provide an estimate of source
19     strength, and to compare modeled time activity data and PEM 24-h mass data to nephelometer
20     measurements (Rea et al., 2001). Several studies also have examined PM exposure in vehicles
21     using both continuous and filter-based techniques.
22
23     5.4.2.3 Interpretation of Participate Matter Exposure Data
24           Papers that have reanalyzed and interpreted the data collected in previous PM exposure
25     studies are summarized  in Table 5-6. These analyses are directed towards understanding the
26     personal cloud, the variability in total PM exposure, and the personal exposure-to-ambient
27     concentration relationships for PM.  Results are highlighted here and given in more detail in
28     Section 5.4.3. Brown and Paxton (1998) determined that the high variability in personal
29     exposure to PM makes the personal-to-ambient PM relationship difficult to predict. Wallace
30     (2000b) used data from  a number of studies to test two hypotheses: elderly COPD patients have
31     (1) smaller personal clouds and (2) higher correlations between personal exposure and ambient
        March 2001
5-24
DRAFT-DO NOT QUOTE OR CITE

-------



c»
fi
5ASUREMENT STUD
ROENVIRONMENTAL MI
Q.
^
H
r^
(s
L
Pi
Pi
5
&
^
g

v!
i
m

|
H








a
Notes/Findii
i
•s
fr
C3
00

0
l|
UH


'tn'
i
g
J

g
'a* H.
l|
Q








1

1

















 2 h'')







GO •<



1°"»

o § J?

CN CN VO



1?
g
§1
*e3 ^
II
s
|o S
•H '5 w
1 M § J
Sources of fine particles: cookinj
Sources of coarse particles: cook
activities. 50% of particles by vol
indoor events were ultrafine partii
Continuous PM distributions and size
distributions obtained for indoor and
outdoor air


o



a?
o\
o%
^<
•a >
t; ~-
§J
JQ
W
•s
.™
•£
e ®
5|
*l
il
hidoor PM concentrations were k
conditioning (AC) than non-air-ci
24 h mean: Regional air 20.2 ±9.9 n^rtf
;n = 50); Outdoor homes 21. 8 ± 14.8 fig/m3
;n = 43); Indoor w/ AC 18.7 ± 13.2 /ig/m3
;n = 49); Indoor w/o AC 2 1 . 1 ±7.5 ^g/m3
;n = 9)



o




















$ i
1 ~^
Is*
si^
g^-g
u 73 t>
ll|
§ l •,?
0.3- to 0.5-/im particles linked to
frying, broiling; 0.5- to 2.5-yum pj
cooking events; >2.5-i*m particle!
movement.
Fime activity data, whole-house air
;xchange rates
Continuous carbon monoxide: descriptive
lata for monitored pollutants; size profiles
"or six indoor particle sources

«•
1 1 1 _,_ |
*tq § S O Tf
vo _• o O O

g
•S
cS
•*-* u
§
u «
M§
II
•o 8 J1
|'5-g
•S£ 1
a o «
« -o "2
•o « §
111
Time-series plots of personal nepl
that each participant's PM exposi
of short-term peaks, imposed on a
ambient PM concentrations.
Continuous (15-min avg) PM and time
ictivity data; 24-h PM mass; participants
rom Baltimore and Fresno PM panel
itudies.
Descriptive statistics from each study for
we microenvironments


o
— «n
O CN

S
"^ § x-v
•I*!!! w
£-• -ZZs *Z CU

OT
"c
>M
"o
ra
O.
•/^


|
§.
•a
s a
**? S
^02
1 = 1
111
"8
^i ^
i^ l-s
O CO o ^3
8~ g-C
(D C S °
2 S'^s'Sl
as IT
K S £ •*
54 ± 3 1 % of average daily PM2.5 e
indoor residences, where participz
their time. A significant portion o
occurred where participants spent
Continuous (15-min avg) PM and time
ictivity data; 24-h PM mass;
vlodeled PM mass and time activity data to
ipportion time spent in a location. Good
:omparison with nephelometer mass (6-
!0%)

o
s i
—* *n
O 
-------
H
CO
 §
^

«?
«n

S

I
    .Si
     '&
D-. t,

11
            •i-81
            •a* §
            2S2SS
O OO
'
                                      I

                                      .5
                       •a!.
II
^ o
            22


            11
               •s
           — o. o. «
            •» o
ea oo

if
o-^
            1°
                       S.S
                          a
'3;
                                                0-.

                                                I
                                              3 2 0
                                              M flj r-~.
                                              "
        112
                                                 -4
                                  ^ r. ••
                                  "e 713

                                  3. "5*00
                                                             o «> *o
                                                            111
                                                            «§<§•§
                                          •S « !

                                          cu a, i
                                              ^3
I:
                                                            09
                                                         Q e i
                                                            -a
                                                          i-liil
                                                          i 111 E
                                                                    o
                                                                    o
                                                         ci

                                                         15 O
                                                         * o
                                                         u c
                                                         33 s
                                                         > £
                                            8>1
                                            § O

                                            5 I

                                            11

                                            •§"=

                                            ^ J
                                            (U ft,

                                            •z:
                                            cd w
                                            I
                                           •§
                                           CQ
 March 2001
                           5-26
           DRAFT-DO NOT QUOTE OR CITE

-------
fc
Pi
tt
pt

B

§
«3
I
            3
            •o
            1
            !
            M
            1
            t»
            f
            Di £
           a-g S-a
           lill
           Z ^ c o.
           o ^

           §,0
                     S|
                     55 .§
                     .tg2
                     < S
                     _• 8
                     M .g
                     •S g
                     (GQ C3
                     '53 [i,
                       tcj
                      &o
                                                        73
                                             g i
                                             i§
                                      II
                                      M
                                      §i
                                      t3 3
                                      £ O
                                                                         "e4
                                                                         4 ?
                                                                         V
                                                                         E1?
                                                                         2
                                                                         lg
                         II
                          '}•>
    < i
V ^  "S»

1 O  e
                                            ts*
                                            S3
                                                     a oa
                                                              i c
                                                              II
                                                              11
                                                              5 s.s
                                                              §.g .a
                                                             •g i •s
F


I  O
-  CQ

•O  b"
                                        2<^ §
                                        co c->
           21
           § 00
           Jl
March 2001
                5-27
                                               DRAFT-DO NOT QUOTE OR CITE

-------
CO

W
CO

H
Z
ps
Z
O


P
 V)


 5

 H

           °!S
           gs
           II
           iSS
3  1
IS-§
                  h
                  SB
                  sS
                  > m
                   • QJ

                  I 2
                  73 S)
                  o o
                  •s s
                  •^•n
                   •3
            •£%• S  .1
            1 E.E  6
                   11

                   ill
                   II
             oo g .2

-Ni
                                 IS
                                •- &
                                JS 5
                                        E

                                       S
                                       o S
                                       U1 .g
                            ^
                            ^.-^
                                       8
                            (8 ^

                            M . .
S g
a. > «
                                                                  ed vj

                                                                  4 81

                                                                  2.1^


                                                                  II"
                                                                  > > (N
                                                                  -5-5^
* U *—' ^ *"*

 |3l^
5 09 — w oo
                                                           §
                                                           o
                                                           00 I
                                                           •g
                                               §11
                                               •g -S s?
                                                           §1
                                                           I §
 March 2001
                                    5-28
                                  DRAFT-DO NOT QUOTE OR CITE

-------
h-
O
w
g
5
A
>
ta
PH
O
g
£
e
ci
c£
vo
•A
S
   £ §

     4s  |.
     -c S  e
     S g
     % M S -O -2
Jsl?*8.!
l^llpl
111^1
Pltltfe"
ill=a?g:

1§^11211
±: »vo «,£:—• cxu
k et al,
r 1988
   o 2  .
   ":•_-  -3
   -3f  «
ioy etal.,
COPD (B
m COPD (
                __ o\  ^-*
                •S
AM (
., 1989
., 1990
ES(Li
ville
sterd
1
Bos
              1
              1
              I
                            •§
                      .s«r
                      "S o
                    "^-l<^
                    o-SZ^-
                    •-^ IK ta «
                    ^l§l
                    I|S| ffi
                    Sg s s =~
                    f- O. M CO
     e
     I
                          la
                          CQ ^,
                         ^B
March 2001
                     5-29
                             DRAFT-DO NOT QUOTE OR CITE

-------
 1     concentrations, compared to healthy elderly, children, and the general population. The analysis
 2     by Wallace (2000a) and three subsequent longitudinal studies (Williams 2000a,b,c; Ebelt et al.,
 3     2000; Samat et al., 2000) support hypothesis 1 but not hypothesis 2. Ozkaynak and Sperigler
 4     (1996) show that at least 50% of personal PM10 exposure for the general population is because of
 5     ambient particles that significantly contribute to inhaled particles. Wilson and Suh (1997)
 6     conclude that fine and coarse particles should be treated as separate classes of pollutants because
 7     of differences in characteristics and potential health effects. Wilson et al. (2000) give a review of
 8     what they call the "exposure paradox" and determine that personal PM needs to be divided into
 9     different classes according to source type, and that correlations between personal and ambient
10     PM will be higher when nonambient sources of PM are removed from the personal PM
11     concentration. Mage (1998) conducted analysis using the PTEAM data and showed that on
12     average a person is exposed to >75% of ambient PM2 5 and >64% of ambient PM10.  Mage et al.
13     (1999) use an algorithm to fill in missing data and outliers to analyzed data sets and show that
14     variation in  daily personal exposures for subjects with similar activity patterns and no ETS
15     exposure are driven by variation in ambient PM concentrations.
16
17     5.4.3  Factors Influencing and Key Findings on Particulate Matter Exposures
18     5.4.3.1  Correlations of Personal/Microenvironmental Particulate Matter with Ambient
19             Particulate Matter
20          The relationship between measured personal PM exposure  and PM concentrations
21     measured at ambient sites has been of interest to exposure analysts. Many of the studies,
22     summarized above in Table 5-4, have analyzed this relationship using measurements of personal
23     PM exposures and ambient PM concentrations.  The statistical correlation between these
24     measurements for the various personal exposure studies is discussed in this section.  .
25
26     5.4.3.1.1  Types of Correlations
27          The three types of correlation data that will be discussed in this section are longitudinal,
28     "pooled", and daily-average correlations. Longitudinal correlations are calculated when data
29     from a  study includes measurements over multiple days for each subject (longitudinal study
30     design). Longitudinal correlations describe the temporal relationship between daily personal PM
31     exposure  and daily ambient PM concentration for each individual subject. The longitudinal
        March 2001
5-30
DRAFT-DO NOT QUOTE OR CITE

-------
 1      correlation coefficient, r, may differ for each subject, and an analysis of the variability in r across
 2      subjects can be performed with this type of data. Typically, the median r is reported along with
 3      the range across subjects in the study. Pooled correlations are calculated when a study involves
 4      one or only a few measurements per subject and different subjects are studied on subsequent days
 5      (sometimes called a "cross-sectional" study design).  The different subject/different day data are
 6      combined, pooled, for the correlation calculation. Pooled correlations describe the relationship
 7      between daily personal PM exposure and daily ambient PM concentration across all subjects in
 8      the study.  This type of correlation is sometimes called cross-sectional, but will be called pooled
 9      in this chapter because only a limited number of participants are monitored on any given day.
10      Daily-average correlations are calculated using the average exposure across subjects for each
11      day.  Daily-average correlations describe the relationship between daily community-averaged
12      personal PM exposure and daily ambient PM concentration. This type of correlation could be
13      called cross-sectional, but given that the pooled correlation also is referred to as cross-sectional,
14      the term daily average is used here.
15           Studies that have reported longitudinal correlations also typically have reported pooled
16      correlations. However, pooling of the data for the correlation has been handled differently across
17      the various studies. For some studies, the multiple days of measurements for each subject were
18      assumed to be independent (after autocorrelation and sensitivity analysis) and combined together
19      in the correlation calculation (Ebelt et al., 2000). In other studies, daily averages across subjects
20      were calculated and the correlation determined from the daily averages (Williams et al., 2000b).
21      A third approach also was used in other studies  to simulate a cross-sectional study design
22      (Janssen et al., 1997, 1998a,  1999c). In this  approach, a random-sampling procedure was used to
23      select a random day from each subject's measurements to use for the correlation. This procedure
24      was repeated many times, and statistics such as the mean and standard deviation of the pooled
25      correlation coefficient were reported.
26           The type  of correlation  analysis can have a substantial effect on the resulting correlation
27      coefficient. Mage et al. (1999) mathematically demonstrated that very low correlations between
28      personal exposure and ambient concentrations could be obtained when people with very different
29      nonambient exposures are pooled, even though their individual longitudinal correlations are high.
30      The longitudinal studies conducted by Tamura et al. (1996a) and Janssen et al. (1997, 1998a,
31      1999c) determined that the longitudinal correlations between personal exposure and ambient PM
        March 2001
5-31
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
concentrations were much higher than the correlations obtained from a pooled data set. Wallace
(2000a) reviewed a number of longitudinal studies found that the median longitudinal correlation
coefficient was much higher than the pooled correlation coefficient for the same data (see
Tables 1 and 2 Wallace, 2000a). Williams et al. (2000a,b) and Evans et al. (2000) reported
higher correlation coefficients for daily-average correlations compared to longitudinal
correlations.

5.4.3.1.2 Correlation Data from Personal Exposure Studies
     Measurement data and correlation coefficients for the personal exposure studies described
hi Section 5.4.2.1 are summarized in Table 5-7. All data are based on mass measurements. The
studies are grouped by the type of study design, longitudinal or pooled. For each study in
Table 5-7, summary statistics for the total personal PM exposure measurements are presented,
as well as statistics for residential indoor, residential outdoor, and ambient PM concentrations,
where available. The correlation coefficient (r) between total personal PM exposures and
ambient PM concentrations also are presented and classified as longitudinal or pooled
correlations. When reported,/?-values for the correlation coefficients are included. Correlation
coefficients between personal, indoor, outdoor, and ambient also are reported, when available.

5.4.3.1.3 Correlations Between Personal Exposures, Indoor, Outdoor, and Ambient
         Measurements
     Longitudinal and pooled correlations between personal exposure and ambient or outdoor
PM concentrations varied considerably between study and study subjects.  Most studies report
longitudinal correlation coefficients that range from <0 to ~ 1, indicating that an individual's
activities and residence type may have a significant effect on total personal exposure to PM.
General population studies tend to show lower correlations because of the higher variation in the
levels of PM generating activities.  In contrast, the absence of indoor sources for the populations
in several of the longitudinal studies resulted in high correlations between personal exposure and
ambient PM within subjects over time for these populations. But even for  these studies,
correlations varied by individual, depending on then: activities and the microenvironments that
they occupied.
        March 2001
                                          5-32
DRAFT-DO NOT QUOTE OR CITE

-------

£?H
0
HH
3
H
1
0
U
Q
9
1

..

&H


H H
^ij fr.
g w
w o

i-l H
^ o
PARTIC
TION C<
g|
PY- {•••
co 5
S5
P Q
co 5
o
(2j
S
o
H
g
0
§
1
CO
PS
P*
t-"
uo
J
PH
H
^
CO
.1
'3

0
1
a




£
... r*
11
fi
< o
.2 U
rt -.
c §

OH 2^
U








1="
•a
1
J3
1
i
§

•a
i
S
S








"a
§


I

1






z

x-v
0>
CO
c
s»- •*
1
>

^
ft
1
1

*M ^
-a "C3
1°
"3
18
*a *a
"S -^
c^

_ ,
1
(§

«
1^
o
"^
1

fl
<£=

g
"1












































.2
1
"«
_g
1
I
0?C?
99
O oo
— •" OO

•**• O Tl O fvj rO
O O O O O O

JJ
X t I I * *
cL-or cuSc S S


a a a
8 g 1 a 1 a S 1 2
I ft 10 P a! 1 f I
8 ,| | 111 ll 1| 111

I II """!! < «> «"iao Ido louQ Q
Sft/l^Q 2P S'^CflS1 t2o5°° I^OOOOOT^W^1
J J: 'f ? "f
§ S f 'S S f S Is 1 S • S S
1 §• 5 J r
"3 2 u o S
1 # 1 # i i 1 i I i^ s3 ss
^ CU > CU CUcCU gCU R P-i P-i fU
March 2001
5-33
DRAFT-DO NOT QUOTE OR CITE

-------





Q
g
5
NATIONS
1
Z
&M
o
z
o
u
o
P-")
in
sl
il
co O
S U
fi 5

S^
Sr H
M S
MONITOK
CORI
M

^j*
0
CO

'"•^

AH
o
a
0
~—^
f^
V)
W
S

H




a
.1

& *5
j3 -P
O c
1




S
„ 13
11
lo ts
Personal-Am
orrelation Coef
o







ion Levels O'g/m3)
S
c
o
1
*o

s

s












•eT
U)

&_
0
[ra
^

1





S
z
1
¥

1

1
^D
1
1 h

Ii
"rt
11
1


"3
S

cu


"|7§
n CO

.£2
1
s
U)
oi|

r- CM * r^ ' en Os — CM 00 SO rf en -H u-j ~H CM OO OO so — ^ t ^1 en m CN — Si CM § O *> ^** Q u -H J * -H J ^ ^" "3 CM CM ™ — a 293 S S S 1* fc pfe o o JJ •a -d SS (0.0-0.72)' 3 Median L r- . oo CM r* CM •n en CM oo rf -H « CM CM r^ ' en OS s Q «* co S1 -H § ^ CM o ^ S O O JJ •a -d U U b S3 - u i | E S o3 § s ^ E .S f. ~. « 77-S^ *i c c 12 S .3 > C- ON OO <*. 99 0 — 00 S Tl s. 2.S O VO u^ OS wi O *f fn f- CN oo r-- «n TT o o o o o o o _] -J -1 .9 .1 .i S ' S 2 a. a. a. a. 2 § jch oo ^ ^r ^ 3S 25 -H — -H -H NO ' 1^ to •X « NO* 06 — T (S — IN " « a «> ,. Q Q tn "" £ ») en -H c*j ° 41 -H is S rsi 0^3- 2 J i ~f .£ J 0- ^ erf ^ od Sz Sz CO ^ CO ^ ff ft 9' ° ^ ^ i-E 11 SK =S o o" o o J _] -J-J ii ii CM Ji ^ 3 en t"* 06 T " i2 o os -H -H "* ^ en oo ^ ^ en CM r~- os KS £S a a a a -H -H -H -H IX IX IX IX J. J.


-------
    co
 §
 »
 s
,—,
CA
g
1
| 8
.0
1
3
'^^
~ 2
|j
2 3
in '-C
""3


£~-
S
^
•3

«j

S
1
I
1
S




e

a
£
:>

"l


1
"g

S
"u
^s w
ii
.2 "3
(2°

*rt
1 0

-------




Q
Z
^t
1
CH^

S
H
W
U
Z
o
u
Q
w
5
^J<
W l

8*5ra o, v *2 S .2 II II II II II II S I o ~ "3 § g > a o s ra o •" o« a.ti gwUwneuOw ^>;O> h- ^i«r4rn^*tovovo g 2 f~ . "2 _g cd B! fe ^ March 2001 5-36 DRAFT-DO NOT QUOTE OR CITE


-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
Probability Studies
      In the Toronto study (Pellizzari et al., 1999), pooled correlations were derived for personal,
indoor, outdoor, and fixed site ambient measurements. This study was conducted in Toronto on
a probability sample of 732 participants who represented the general population, 16 years and
older. The study included between 185 and 203 monitoring periods with usable PM data for
personal, residential indoor, and outdoor measurements.  For PM10, measurements, the mean
concentrations were 67.9 /ug/m3 for personal, 29.8 /J-g/m? for indoor air, and 24.3 ,ug/m3 for
outdoor air samples. For PM2 5, the mean concentrations were 28.4 //g/m3 for personal,
21.1 yUg/m3 for indoor air, and 15.1 yug/m3 for outdoor air samples. A low but significant
correlation (r = 0.23, p < 0.01) was reported between personal exposure and ambient
measurements.  The correlations between indoor concentrations and the various outdoor
measurements of PM25 ranged from 0.21 to 0.33. The highest correlations were for outdoor
measurements at the residences with the ambient measurements made at the roof site (0.88) and
the other fixed site (0.82). Pellizzari et al. (1999) state that much of the difference among the
data for personal/indoor/outdoor PM

      ...  can be attributed to tobacco smoking, since all variables reflecting smoking... were found to be
      highly correlated with the personal (and indoor) particulate matter levels, relative to other variables that
      were measured... none of the outdoor concentration data types (residential or otherwise) can
      adequately predict personal exposures to particulate matter, (p. 729)

      Santos-Burgoa et al. (1998) describe a 1992 study of personal exposures and indoor
concentrations to a randomly sampled population near Mexico City. The sample of 66 monitored
subjects included children, students, office and industrial workers, and housewives. None of the
people monitored were more than 65 years old. The mean 24-h personal exposure and indoor
concentrations were 97 ± 44 (SD) and 99 ± 50 ^g '3,  respectively, with an rPersonal/Ambient = 0.26
(p = 0.099).  Other correlations of interest were rPeisonal/Indoor = 0.47 (p = 0.002) and rIndoor/Ambient =
0.23 (p = 0.158). A strong statistical association was found between personal exposure and
socioeconomic class (p = 0.047) and a composite index of indoor sources at the home
(p = 0.039).
        March 2001
                                          5-37
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Correlation analysis for personal exposure has not yet been reported for EXPOLIS. Some
 2     preliminary results (Jantunen et al., 2000) show that in Basel and Helsinki, a single ambient
 3     monitoring station was sufficient to characterize the ambient PM2 5 concentration in each city.
 4     Using microenvironmental concentration data collected while the subjects were at home, at work,
 5     and outdoors, they calculated the sum of the time-weighted-averages of these data and found the
 6     results closely match the personal PM2 5 exposure data collected by the monitors carried by most
 7     of the subjects, with a few subjects, mostly smokers, being noticeable exceptions.
 8
 9     Longitudinal Studies
10           A number of longitudinal studies using a purposeful sampling design have been conducted
11     and reported in the literature since 1996. A number of these studies (Janssen et al., 1998a,
12      1999b, 2000; Williams et al., 2000b; Evans et al., 2000) support the previous work by Janssen
13     et al. (1995) and Tamura et al.  (1996a) and demonstrate that, for individuals with little exposure
14     to nonambient sources of PM, correlations between total PM exposure and ambient PM
15     measurements are high. Other studies (Ebelt et al., 2000; Samat et al., 2000) show strong
16     correlations for the SO4"2 component of PM2 5 but poorer correlations for PM2 5 mass.  Still other
17     studies show only weak correlations (Rojas-Bracho et al., 2000; Linn et al., 1999; Bahadori et al.,
18     2001). Even when strong longitudinal correlations are demonstrated for individuals in a study,
19     the variety of living conditions may lead to variations in the fraction of ambient PM contributing
20     to personal exposure. Groups with similar living conditions, especially if measurements are
21     conducted during one season, may have similar a and, therefore, very high correlations between
22     personal exposure and ambient concentrations. However, when a panel contains subjects with
23     homes of very different ventilation characteristics or covers more than one season, variations in a
24     can be high across subjects.
25           Elderly Subjects.  Janssen et al. (2000)  continued their longitudinal studies with
26     measurements of personal, indoor, and outdoor concentrations of PM25 for elderly subjects with
27     doctor-diagnosed angina pectoris or coronary heart disease. Studies were conducted in
28     Amsterdam and Helsinki, Finland, in the winter and spring of 1998 and 1999.  In the Amsterdam
29     study, with 338 to 417 observations, the mean concentrations were 24.3, 28.6, and 20.6 //g/m3 for
30     personal, indoor, and outdoor samples, respectively. If the measurements with ETS in the home
31     were excluded, the mean indoor concentration dropped to 16 Aig/m3, which was lower than
        March 2001
5-38
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
27
28
29
30
 outdoors. In the Helsinki study, the mean PM2 5 concentrations were 10.8 yug/m3 for personal,
 11.0 //g/m3 for indoor air, and 12.6 yug/m3 outdoor air samples.  The authors note that for this
 group of subjects, personal exposure, indoor concentrations, and ambient concentrations of PM25
 were highly correlated within subjects over time. Median Pearson's correlation coefficients
 between personal exposure and outdoor concentrations were 0.79 in Amsterdam and 0.76 in
 Helsinki. The median Pearson's r for the indoor/outdoor relationship was 0.85 for the
 Amsterdam study, excluding homes with ETS. The correlation for indoors versus outdoors was
 0.70 for all homes.
      A series of PM personal monitoring studies involving elderly subjects was conducted in
 Baltimore County, MD, and Fresno, CA. The first study was a 17-day pilot (January-February
 1997) to investigate daily personal and indoor PM15 concentrations, and outdoor PM2 5 and
 PM2.s-io concentrations experienced by nonsmoking elderly residents of a retirement community
 located near Baltimore (Liao et al, 1999; Williams et al., 2000c).  The 26 residents were aged
 65 to 89 (mean = 81), and 69% of them reported a medical condition, such as hypertension or
 coronary heart disease. In addition, they were quite sedentary; less than 5 h day"1, on average,
 was spent on ambulatory activities.  Because most of the residents ate meals in a communal
 dining area, the average daily cooking time in the individual apartments was only 0.5 h (range 0
 to 4.5 h). About 96% of the residents' time was spent indoors (Williams et al., 2000c).  Personal
 monitoring, conducted for five subjects, yielded longitudinal correlation coefficients between
 ambient concentrations and personal exposure ranging from 0.00 to 0.90.
     Subjects with COPD.  Linn et al. (1999) describe a 4-day longitudinal assessment of
 personal PM2 5 and PMIO exposures (on alternate days) in 30 COPD subjects aged 56 to 83;
 concurrent indoor and outdoor monitoring were conducted at their residences. This study
 occurred in the summer and autumn of 1996 in the Los Angeles area.  PM10 data from the nearest
 fixed-site monitoring station to each residence also was obtained. Pooled correlations for
personal exposure to outdoor measurements were 0.26 and 0.22 for PM2 5 and PM10, respectively.
Day-to-day changes in PM2 5 and PM10 measured outside the homes tracked concurrent PM10
measurements at the nearest ambient monitoring location, with R2 values of 0.22 and 0.44,
respectively.  Day to day changes in PM mass measured indoors  also tracked outdoors at the
homes with R2 values of 0.27 and 0.19 for PM10 and PM25  respectively.
       March 2001
                                         5-39
DRAFT-DO NOT QUOTE OR CITE

-------
 1           Personal, indoor, and outdoor PM2 5? PM,0i and PM2.5.10 correlations were reported by
 2      Rojas-Bracho et al. (2000) for a study conducted in Boston, MA, on 18 individuals with COPD.
 3      Both the mean and median personal exposure concentrations were higher than the indoor
 4      concentrations, which were higher than outdoor concentrations for all three PM measurement
 5      parameters.  Geometric mean indoor/outdoor ratios were 1.4 ± 1.9 for PM10, 1.3 ± 1.8 for PM2 5,
 6      and 1.5 ± 2.7 for PM2.s.10. Median longitudinal R2s between personal exposure and ambient PM
 7      measurements were 0.12 for PM10, 0.37 for PM25 and 0.07 for PM2.5.I0. The relationship between
 8      the indoor and outdoor concentrations was strongest for PM25, with a median R2 of 0.55 and
 9      11 homes having significant R2 values.  For PM10 the median R2 value was 0.25, with significant
10      values for eight homes. Only five homes had significant indoor/outdoor associations for PM2 5.10,
11      with an insignificant median R2 value of 0.04.
12          Bahadori et al. (2001) report a pilot study of the PM exposure of 10 nonrandomly chosen
13      chronic obstructive pulmonary disease (COPD) patients in Nashville, TN, during the summer of
14     1995.  Each subject alternately carried a personal PM25 or PM10 monitor for a 12-h daytime
15     period (8 a.m. to 8 p.m.) for 6 consecutive days. These same pollutants were monitored
16     simultaneously indoors and outdoors at their homes. All of the homes were air-conditioned and
17     had low air exchange rates (mean = 0.57 hf1), which may have contributed to the finding that
18     mean indoor PMZ5 was 66% of the mean ambient PM2 5. This can be contrasted with the
19     PTEAM study in Riverside, CA, where no air conditioners were in use and the mean indoor
20     PM2-S was 98% of the mean ambient PM2.5 (Clayton et al., 1993). Data sets were pooled for
21     correlation analysis.  Resulting pooled correlations between personal and outdoor concentrations
22     were r= 0.09 for PM2.5 and r=-0.08 for PM10.
23
24     5.4.3.1.4 A Correlation Between a Daily-Average Exposure and Ambient Concentrations
25           A recent biostatistical analysis (Zeger et al., 2000) suggests that the community mean
26     exposure is the appropriate parameter for analyzing exposure error in community time-series
27     epidemiology. Ott et al. (2000) suggest that the correlation of the community mean exposure
28     with ambient concentrations will approach 1.0 for a large community and demonstrated this
29     using data from the PTEAM study. Mage et al. (1999) calculated the daily-average exposure for
30     three earlier studies with sufficient data and found that the coefficients for the correlation of daily
        March 2001
5-40
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
32
averages with the ambient concentrations were high (R2 = 0.9).  Two new studies have obtained
sufficient data to calculate a daily-average correlation.
      The 1997 Baltimore pilot study was followed up in July and August 1998 by a more
extensive study in which identical samplers were used for personal, indoor, and outdoor
measurements. The participants, aged 72 to 93 (mean 81) years, included healthy members, as
well as subjects with COPD and cardiovascular disease. The participants lived in an 18-story
retirement facility that provided a self-contained living environment. There was a central HVAC
system for common areas but each apartment had an individually controlled HVAC system. The
subjects had limited exposures to indoor-generated sources of PM because of their low
frequency/duration of activities like cooking, cleaning, or interacting with tobacco smokers
(Williams et al., 2000a,b). As a result, the daily-average correlation coefficient was very high
(r = 0.89) between personal exposure and ambient concentrations of PM2 5. Median longitudinal
correlations were also high (r = 0.81; range = 0.38 to 0.98).
      Evans et al. (2000) report two panel studies in Fresno, CA, with daily-average correlation
coefficients of 0.41  and 0.84.

5.4.3.1.5 Correlations Using Sulfate as a Surrogate for Personal Exposure to Ambient
         Particulate Matter
      A study, conducted in Vancouver, involving sixteen COPD patients aged 54 to 86, reported
low median longitudinal (r = 0.48) and pooled (r = 0.15) correlation coefficients between
personal exposure and ambient  concentrations of PM25 (Ebelt et al., 2000). However, the
correlation between personal exposure and ambient concentrations of SO42" was much higher.
The results for PM2 5 and sulfate are compared in Figure 5-1. Ebelt et al. (2000) conclude the
following.

         We found SO42' to be a good measure of exposure to accumulation mode PM of ambient
      origin.  Personal and ambient measures of SO42" were highly correlated over time, unlike the
      moderate correlation found for PM2 5. The individual correlations demonstrated that ambient
      SO42' was a consistently strong predictor across all individuals and all levels of exposure, whereas
      for PM2.5 correlations varied by individual and were dependent upon the level of personal
      exposure. Although indoor sources likely contribute to personal exposures of PM2 5, accounting
      for such variables did not lead to models with the same predictive power as found for SO42".
        March 2001
                                          5-41
DRAFT-DO NOT QUOTE OR CITE

-------
                                Correlation Coefficient for Individuals

1.00 -

0.75 -
0.50 -

0.25 -

f 0.00 -
-0.25 -
-0.50 -
-0.75 -
_-i nn -
Ebelt et al., 2000
Pearson's r
P=.
PM2.S 5





I







Sulfate









Percentile

90th Percentile
75th Percentile
Median


25th Percentile
1 0th Percentile




Sarnatetal., 2000
Spearman's r
PM25
&

^
I



T_


~

I




NSSSSN Surr
i i Win


gL

T
Sulfate





imer
ter

                     PM25     Sulfate
                                                            PM25     Sulfate
      Figure 5-1.  Comparison of correlation coefficients for longitudinal analysis of personal
                  exposure versus ambient concentrations for individual subjects for PM2 5 and
                  sulfate.
1
2
3
4
5
6
7
     Similarly, we found that accounting for spatial variability in ambient levels did not improve the
     relationship between ambient concentrations and measured personal exposures. Overall, we have
     shown that a personal measure of exposure to outdoor source PM is highly related to variation in
     ambient levels of PM.

     Another study conducted in Baltimore, MD, involved 15 nonsmoking adult subjects
(>64 years old) who were monitored for 12 days during summer 1998 and winter 1999 (Sarnat
      March 2001
                                         5-42
DRAFT-DO NOT QUOTE OR CITE

-------
  1      et al., 2000).  All subjects (nonrandom selection) were retired, physically healthy, and lived in
  2      nonsmoking private residences.  Each residence, except one, was equipped with central
  3      air-conditioning; however, not all residences used air-conditioning throughout the summer. The
  4      average age of the subjects was 75 years (±6.8 years).  Sarnat et al. (2000) reported higher
  5      longitudinal and pooled correlations for PM2 5 during summer than winter.  Similar to Ebelt et al.
  6      (2000), Sarnat et al. (2000) reported stronger associations between personal exposure to SO42"
  7      and ambient concentrations of SO42". The ranges of correlations are shown in Figure 5-1 along
  8      with similar data from Ebelt et al. (2000).
  9           The study conducted by Sarnat et al. (2000) also illustrates the importance of ventilation on
10      personal exposure to PM. During the summer, subjects recorded the ventilation status of every
11      visited indoor location (e.g., windows open, air-conditioning use).  As a surrogate for the
12      air-exchange rate, personal exposures were classified by the fraction of time the windows were
13      open while a subject was in an indoor environment (Fv).  Sarnat et al. (2000) report regression
14      analyses for personal exposure on ambient concentration for total PM2 s and for sulfate for each
15      of the three ventilation conditions.  Personal exposure to sulfate may be taken as a surrogate for
16      personal exposure to ambient accumulation-mode PM in the absence of indoor sulfate sources.
17      Figure 5-2 shows a comparison of the regressions and indicates how the use of a sulfate tracer as
18      a surrogate for PM of ambient origin improves the correlation coefficient. The improvement is
19      especially pronounced for the lowest ventilation conditions.  For the lowest ventilation condition,
20      R2 improves from 0.25 to 0.72.
21           The Ebelt et al. (2000) and Sarnat et al. (2000) studies did not use their sulfate data to
22      develop relationships between personal exposure to ambient PM and ambient PM concentrations
23      for individual subjects, as suggested by Wilson et al. (2000). However, the higher correlation
24      coefficients and the narrower range of the correlation coefficient for sulfate suggest that
25      removing indoor-generated and personal activity PM from total personal PM would result in a
26      higher correlation with ambient concentrations.  However, the variation in ventilation status (and
27      thus in the attenuation coefficient a) still would cause variations between ambient concentrations
28      of PM and personal exposure to ambient PM, especially if the study continued long enough to
29      extend through more than one season.
30
        March 2001
5-43
DRAFT-DO NOT QUOTE OR CITE

-------
   60'
g  40-


Q.  30-


"g  20-


IX  10-
                     Well Ventilated Indoor Environment
                                        35-
       R2=0.80
                                        30

                                        25

                                        20

                                        15

                                        10

                                         5
                                            R2 = 0.88
   60'
^  50-

1
g-  40-


8.  30-


"S  20-


(?  10-
                   Moderately Vented Indoor Environment
                                        35
       R2 = 0.57
   /•
        /
<*'•'• '
                                        30'

                                        25

                                        20

                                        15

                                        10

                                         5
                                            R2 = 0.73
   60
   50-
 g- 40-


 §. 30-


 1 20-
                     Poorly Ventilated Indoor Environment
                                        35
       R2 = 0.25
           10    20    30    40    50

                    PM2.5
                                        30-

                                        25-

                                        20-

                                        15-

                                        10-

                                         5-

                                         0
                                            R2 = 0.72
                                     60   0
                                          10   15   20   25   30   40
                                                   SO4
                         Ambient Concentration (M9/m3)

Figure 5-2.  Personal exposure versus ambient concentrations for PM2 s and sulfate. (Slope
           estimated from mixed models).

Source: Sarnat et al. (2000).
March 2001
                                    5-44
                                      DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
5.4.3.1.6 Correlations Between Personal Exposure to Ambient and Nonambient
         Particulate Matter
     The utility of treating personal exposure to ambient PM, Eag, and personal exposure to
nonambient PM, Enonag, as separate and distinct components of total personal exposure to PM, Et,
was pointed out by Wilson and Suh (1997). The PTEAM study measured, in addition to indoor,
outdoor, and personal PM, the air exchange rate for each home and collected information on the
time spent in various indoor and outdoor #e. This information is available for 147, 12-h daytime
periods. With this information, it is possible to estimate the daytime Eag and Enonag as described
in Section 5.3.2.3.1. Various examples of this information have been reported (Mage et al.,
1999; Wilson et al., 2000). Graphs showing the relationships between ambient concentration and
the various components of personal exposure (Et, Eag, and  Enonag) are shown in Figure 5-3. The
correlation coefficient for the pooled data set improves from r = 0.377 for Et versus Ca
(Figure 5-3a) to r = 0.856 for Eag versus Ca (Figure 5-3b) because of the removal of the Enonag ,
which, as shown in Figure 5-3c, is highly variable and independent of Ca.  The correlation
between Eag and Ca is less than  1 because of the day-to-day variation in ait. The regression
analysis with E, total PM gives  O~= 0.711 and N = 81.6 ^g/m3. The regression analysis with Eag
gives a = 0.625. The regression with Enonag gives N = 79.2 Aig/m3.  The finite intercept in the
regression with Eag must be attributed to bias or error in some of the measurements. No studies,
other than PTEAM, have provided the quantity of data on Et, Ca, Ci5 and a required to conduct
an analysis comparable to that shown in Figure 5-3.
     The higher correlations found between daily-average personal exposures and ambient PM
concentrations, as opposed to lower correlations found between individual exposures  and
ambient PM levels, recently have been attributed to statistical rather  than physical causes. Ott
et al. (2000), using their Random Component Superposition (RCS) model, solely attribute this to
the averaging process.  Because personal exposures also include contributions from ambient
concentrations, the correlation between personal exposure and ambient concentrations increases
as the number of subjects measured daily increases. Based on theory, Ott et al. (2000) predict
expected correlations above 0.9 if 25 subjects had been studied during the PTEAM study and
above 0.70 in the New Jersey study reported by Lioy et al. (1990).
        March 2001
                                         5-45
DRAFT-DO NOT QUOTE OR CITE

-------
                  250
                 5150-1
               S ^
               0) en
               |u!10°-
               fig
               I! 5°-
               «|
                 i  0'
                   -SO'
                       r» 0.051
                       R'» 0.0026
                       N=39.6+0.086C
                       *=147	
            *

           * *
^M^T^
       ;     *'  V    T~»
                              50       100       150      200
                                  Ambient Concentration, \ig/m3
                                                                 250
Figure 5-3. Regression analyses of aspects of daytime personal exposure to PM10 estimated
           using data from the PTEAM study, (a) Total personal exposure to PM, E,,
           regressed on ambient concentration, Ca.  (b) Personal exposure to ambient PM,
           Eag regressed on Ca. (c) Personal exposure to  nonambient PM, Enonag regressed
           onCa.
Source: Data taken from Clayton et al. (1993).

March 2001
               5-46
DRAFT-DO NOT QUOTE OR CITE

-------
  1          The RCS model introduced by Ott et al.(2000) presents a modeling framework to determine
  2     the contribution of ambient PM10 and indoor-generated PM10 on personal exposures in large
  3     urban metropolitan areas. The model has been tested using personal, indoor and outdoor PM10
  4     data from three urban areas (Riverside, CA; Toronto; and Phillipsburg, NJ). Results suggest that
  5     it is possible to separate the ambient and nonambient PM contributions to personal exposures on
  6     a community-wide basis.  However, as discussed in the paper, the authors make some
  7     assumptions that require individual consideration in each-city specific application of the model
  8     for exposure or health effects investigations. Primarily, housing factors, air-conditioning,
  9     seasonal differences, and complexities in time-activity profiles specific to the cohort being
 10     studied have to be taken into account prior to adopting the model to a given situation.
 11
 12     5.4.3.2 Factors That Affect Correlations
 13          A number of factors will affect the relationship between personal exposure and PM
 14     measured at ambient-site community monitors. Spatial variability in outdoor microenvironments
 15     and penetration into indoor microenvironments will influence the relationship for ambient-
 16     generated PM, air-exchange rates and decay rates in indoor microenvironments will influence the
 17     relationship for both ambient-generated and total PM, whereas personal activities will influence
 18     the relationship for total PM but not ambient-generated PM. Information on these effects is
 19     presented in detail in the following section.
20
21      5.4.3.2.1 Spatial Variability and Correlations Over Time
22          Chapter 3 (Section 3.2.3) presents information on the spatial variability of PM mass and
23      chemical components at fixed-site ambient monitors; for purposes of this chapter, this spatial
24      variability is called an "ambient gradient". The data presented in Section 3.2.3 indicate that
25      ambient gradients of PM and its constituents exist in urban areas to a greater or lesser degree.
26      This gradient, and any that may exist between a fixed-site monitor and the outdoor //e near where
27      people live, work, and play, obviously affects the exposure. The purpose of this section is to
28      review the available  data on ambient monitor-to-outdoor microenvironmental concentration
29      gradients, or relationships, that have been measured by researchers since 1996. A few outdoor-
30      to-outdoor monitoring studies also are included to highlight relationships among important yue
31      categories.  To assess spatial variability or gradients, the spatial correlations in the data are
        March 2001
5-47
DRAFT-DO NOT QUOTE OR CITE

-------
 1     usually analyzed. However, it should be noted that high temporal correlation between two
 2     monitoring locations does not imply low spatial variability or low ambient gradients. High
 3     temporal correlation between two sites indicates that changes in concentrations at one site can be
 4     estimated from data at another site.
 5          Oglesby et al. (2000), in a paper on the EXPOLIS-EAS study, conclude that very little
 6     spatial variability exists in Basel, Switzerland, between PM levels measured at fixed site
 7     monitors and the participant's outdoor //e. The authors report a high correlation between home
 8     outdoor PM2.5 levels (48-h measurements beginning and ending at 8:00 a.m.) and the
 9     corresponding 24-h average PM4 (time-weighted values calculated from midnight to midnight)
10     measured at a fixed monitoring station ^ = 38,^ = 0.96, p< 0.001). They considered each
11     home outdoor monitor as a temporary fixed monitor and concluded that "the PM2 5 level
12     measured at home outdoors ... represents the fine particle level prevailing in the city of Basel
13     during the 48-h measuring period...."
14          In a study conducted in Helsinki, Finland, Buzorius et al. (1999) conclude that a single
15     monitor may be used to adequately describe the ambient gradient across the metropolitan area.
16     Particle size distributions were measured using a differential mobility particle sizer (DMPS;
17     Wintlmayer) coupled with a condensation particle counter (CPC TSI3010, 3022) at four
18     locations including the official air monitoring station, which represented a "background" site.
19     The monitoring period varied between 2 weeks and 6 mo for the sites and data were reported for
20     10-min and 1-, 8-, and 24-h averages. As expected, temporal variation decreased as the
21     averaging time increased.  The authors report that particle number concentration varied in
22     magnitude with local traffic intensity. Linear correlation coefficients computed for all possible
23     site-pairs and averaging times showed that the correlation coefficient improved with increasing
24     averaging time.  Using wind speed and direction vectors, lagged correlations were calculated and
25     were generally higher than the  "raw" data correlations.  Weekday correlations were higher than
26     weekend correlations as "traffic provides relatively uniform spatial distribution of particulate
27     matter" (p. 565). The authors conclude that, even for time periods of 10 min and 1 h,  sampling at
28     one station can describe changes across relatively large areas of the city with a correlation
29     coefficient >0.7.
30          Dubowsky et al. (1999) point out that, although the variation of PM2 5 mass concentration
31     across a community may be small, there may be significant spatial variations of specific
        March 2001
5-48
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
 components of the total mass on a local scale. An example is given of a study of concentrations
 of polycyclic aromatic hydrocarbons (PAH) at three indoor locations in a community; an urban
 and a semi-urban site separated by 1.6 km, and a suburban site located further away.  The authors
 found the geometric mean PAH concentrations at these three locations varied respectively as
 31:19:8 ng/m3, and suggest that the local variations hi traffic density were responsible for this
 gradient. Note that these concentrations are 1,000 times lower than the total PM mass
 concentration, so that such a small gradient would not be detectable for total PM2 5 mass
 measurements on the order of 25 //g m"3.
      Leaderer et al. (1999) monitored 24-h PM10, PM2 5, and sulfates during the summers of
 1995  and 1996 at a regional site in Vinton, VA (6 km from Roanoke, VA). One similar 24-h
 measurement was made outdoors at residences in the surrounding area, at distances ranging from
 1 km  to > 175 km from the Vinton site,  at an average separation distance of 96 km. The authors
 reported significant correlations for PM2 5 and sulfates between the residential outdoor values and
 those  measured at Vinton on the same day. In addition, the mean values of the regional site and
 residential site PM2 5 and sulfates  showed no significant differences in spite of the large distance
 separations and mountainous terrain intervening in most directions. However, for the
 concentrations of PM2 5_10, estimated as  PM,0-PM25, no significant correlation among these sites
 was found (n = 30, r = -0.20).
     Lillquist et al. (1998) found no significant gradient in PMIO concentrations in Salt Lake
 City, UT, when levels were low, but a gradient existed when levels were high. PMIO
 concentrations were measured outdoor at three hospitals using a Minivol 4.01 sampler
 (Airmetrics, Inc.) operating at 5 L min'1 and at the Utah Department of Air Quality (DAQ)
 ambient monitoring station located between 3 and  13 km from the hospitals for a period of about
 5 mo.
     Pope et al. (1999) monitored ambient PMIO concentrations in Provo, UT (Utah Valley),
 during the same time frame the following year and reported nearly identical concentrations at
three sites separated by 4 to 12 km. Pearson correlation coefficients for the data were between
0.92 and 0.96. The greater degree of variability in the Salt Lake City PM10 data relative to the
Provo data maybe related to the higher incidence of wind-blown crustal material in Salt Lake
City.  Pope et al. (1999) reported that increased health effects in the Utah Valley were associated
with stagnation and thermal inversions trapping anthropogenically derived PM10, whereas, no
        March 2001
                                         5-49
DRAFT-DO NOT QUOTE OR CITE

-------
 1      increases in health effects were observed when PMIO levels were increased during events of wind
 2      blown crustal material.
 3           Vakeva et al. (1999) found significant vertical gradients in submicron particles existed in
 4      an urban street canyon of Lahti, Finland. Particle number concentrations were measured using a
 5      TSI screen diffusion battery and a condensation particle counter at 1.5 and 25 m above the street
 6      at rooftop level. The authors found a fivefold decrease in concentration between the two
 7      sampling heights and attributed the vertical gradient to dilution and dispersion of pollutants
 8      emitted at street level.
 9           White (1998) suggests that the higher random measurement error for the coarse PM
10      fraction  compared to the error for the fine PM fraction may be responsible for a major portion of
11      the apparent greater spatial variability of coarse ambient PM concentration compared to fine
12      ambient PM concentration in a community (e.g., Burton et al., 1996; Leaderer et al., 1999).
13      When PM2.5 and PM10 are collected independently, and the coarse fraction is obtained by
14      difference (PM2 S_IO = PMi0-PM2 5), then the expected variance in the coarse fraction is the sum of
15      the variances of the PM10 and PM25 measurements.  When a dichotomous sampler collects PM25
16      and PM2 5_10 on two separate filters, the coarse fraction also is expected to have a larger error than
17      the fine fraction.  There is a possible error caused by loss of mass below the cut-point size and a
18      gain of mass above the cut-point size that is created by the asymmetry of the product of the
19      penetration tunes PM concentration about the cut-point size.  Because a dichotomous PM
20      sampler collects coarse mass using an upper and lower cut-point, it is expected to have a larger
21      variance than for the fine mass collected using the same lower cut-point.
22           Wilson and Suh (1997) conclude that PM25 and PMi0 concentrations are correlated more
23      highly across Philadelphia than are PM2^.10 concentrations. Ambient monitoring data from 1992
24      to 1993 was reviewed for PM2 5, PM2 5.10, and PM10, as well as for PM2 5 and PM2 5_10 dichotomous
25      data for  212 site-years of information contained in the AIRS database. The authors also observed
26      that PM10 frequently was correlated more highly with PM2 5 than with PM2 5.10. The authors note
27      that PM2iS constitutes a large fraction of PM10, and that this is the likely reason for the strong
28      agreement between PM2 5 and PM10.  Similar observations were made by Keywood et al. (1999)
29      in six Australian cities.  The authors reported that PM10 was more highly correlated with PM2 5
30      than with coarse PM (PM2 5_10), suggesting that "variability in PM10 is dominated by variability in
31      PM2.5."
        March 2001
5-50
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
      Lippmann et al. (2000) examined the site-to-site temporal correlations in Philadelphia
 (1981 to 1994) and found the ranking of median site-to-site correlation was O3 (0.83), PM10
 (0.78), TSP (0.71), N02 (0.70), CO (0.50), and SO2 (0.49). The authors explain that O3 and a
 fraction of TSP and PM10 (e.g., sulfate) are secondary pollutants that would tend to be distributed
 spatially more uniformly within the city than primary pollutants such as CO and SO2, which are
 more likely to be influenced by local emission sources.  Lippman et al. (2000) conclude:  "Thus,
 spatial uniformity of pollutants may be due to area-wide sources, or to transport (e.g., advection)
 of fairly stable pollutants into the urban area from upwind sources. Relative spatial uniformity of
 pollutants would therefore vary from city to city or region to region."

 5.4.3.2.2 Physical Factors Affecting Indoor Microenvironmental Particulate Matter
         Concentrations
      Several physical factors affect ambient particle concentrations in the indoor /j.e, including
 air exchange, penetration, and particle deposition. Combined, these factors are critical variables
 that describe ambient particle dynamics in the indoor #e and, to a large degree, significantly
 affect an individual's personal exposure  to ambient-generated particles while indoors. The
 relationship between ambient outdoor particles and ambient particles that have infiltrated indoors
 is given by
                                                                                   (5-10)
where Cai and Cao are the concentration of ambient indoor and outdoor particles, respectively;
P is the penetration factor; a is the air exchange rate; and k is the particle deposition rate (as
discussed in Section 5.3.2.3.1, use of this model assumes equilibrium conditions and assumes
that all variables remain constant).  Particle penetration is a dimensionless quantity that describes
the fraction of ambient particles that effectively penetrates the building shell. "Air exchange" is
a term used to describe the rate at which the indoor air in a building or residence is replaced by
outdoor air. The dominant processes governing particle penetration are air exchange and
deposition of particles as they traverse through cracks and crevices and other routes of entry into
the building.  Although air-exchange rates have been measured in numerous studies, very few
field data existed prior to 1996 to determine size-dependent penetration factors and particle
        March 2001
                                          5-51
DRAFT-DO NOT QUOTE OR CITE

-------
 1      deposition rates. All three parameters (P, a, and K) may vary substantially depending on building
 2      type, region of the country, and season.  In the past several years, researchers have made
 3      significant advancements in understanding the relationship between particle size and penetration
 4      factors and particle deposition rates. This section will highlight the studies that have been
 5      conducted to better understand physical factors affecting indoor particle dynamics.
 6
 7      Air-Exchange Rates
 8           The air-exchange rate, a, in a residence varies depending on a variety of factors, including
 9     geographical location, age of the building, the extent to which window and doors are open, and
10     season. Murray and Burmaster( 1995) used measured values of a from households throughout
11      the United States to describe empirical distributions and to estimate univariate parametric
12     probability distributions of air-exchange rates. Figure 5-4 shows the results classified by season
13     and region. In general, a is highest in the warmest region and increases from the coldest to the
14     warmest region during all seasons. Air-exchange rates also  are quite variable within and between
15     seasons, as well as between regions (Figure 5-4). Data from the warmest region in summer
16     should be viewed cautiously as many of the measurements were made in Southern California in
17     July when windows were more likely to be open than in other areas of the country where
18     air-conditioning is used.  Use of air-conditioning generally results in lowering air-exchange rates.
19     In a separate analyses of these data, Koontz and Rector (1995) suggested that a conservative
20     estimate for air exchange in residential settings would be 0.18 h'1 (1 Oth percentile) and a typical
21     air exchange would be 0.45 h"1 (50th percentile).
22           These data provide reasonable experimental evidence that a varies by season in locations
23     with distinct seasons. As a result, infiltration of ambient particles may be more efficient during
24     warmer seasons when windows are likely to be opened more frequently and air-exchange rates
25     are higher.  This suggests that the fraction of ambient particles present in the indoor //e would be
26     greater during warmer seasons than colder seasons.  For example, in a study conducted in
27     Boston, MA, participants living in non-air-conditioned homes kept the windows closed except
28     during the summer (Long et al., 2000a). This resulted in higher and more variable air-exchange
29     rates in summer than during any other season (Figure 5-5).  During nighttime periods, when
30     indoor sources are negligible, the indoor/outdoor concentration ratio or infiltration factor may be
31     used to determine the relative contribution of ambient particles in the indoor y.e.  Particle data
        March 2001
5-52
DRAFT-DO NOT QUOTE OR CITE

-------
        O»
        WD
        S
           3 -
           _
           2H
       •*  1 H
           o
                   Coldest Region
                   Colder Region
                   Warmer Region
                   Warmest Region
                         Winter        Spring      Summer
                                               Season
                                                                Fall
      Figure 5-4. Air-exchange rates measured in homes throughout the United States. Climatic
                 regions are based on heating-degree days: Coldest region £ 7000, Colder
                 region = 5500 to 6999, Warmer region = 2500 to 4999, and Warmest region
                 <. 2500 heating-degree days.
      Based on data from Murray and Burmaster (1995).
1
2
3
4
5
collected during this study (Figure 5-6) shows the indoor/outdoor concentration ratios by particle
size. Data show that for these nine homes in Boston, the fraction of ambient particles penetrating
indoors is higher during summer when air exchange rates were higher than fall (Long et al.,
2000b).
      March 2001
                                       5-53
DRAFT-DO NOT QUOTE OR CITE

-------
             8
             7 -
        =S    6H
        U)
        03
        O)
        §    6

        I   4
        OJ
        CO
        01
        £
co
i
        £
             0 -
                                                                             95%
                                       .90%
                                                                                  Median
                       Fall
Winter
     Spring
Season
Summer
      Figure 5-5. Box plots of hourly air-exchange rates stratified by season in Boston, MA,
                  during 1998.
      Source: Long et al. (2000a).
1      Particle Deposition Rates and Penetration Factors
2           Physical factors affecting indoor particle concentrations, including particle deposition rates,
3      k, and penetration factors, P, are possibly the most uncertain and variable quantities. Although k
4      can be modeled with some success, direct measurements are difficult and results often vary from
5      study to study.  Particle deposition rates vary considerably depending on particle size because of
6      the viscous drag of air on the particles hindering their movement to varying degrees. The nature
7      and composition of particles also affect deposition rates. Surface properties of particles, such as
8      their electrostatic properties, can have a significant influence on deposition rates. In addition,
9      thennophoresis can also affect k, but probably to a lesser degree in the indoor /we because
       March 2001
          5-54
          DRAFT-DO NOT QUOTE OR CITE

-------
                         a) Home NEW2
            CO
            LL
            c
            .o
            13
           <*=
 1.1
 1.0 -
 0.9 -
 0.8 -
 0.7 -
 0.6 -
0.5 -
0.4 -
0.3 -
0.2 -
 0.1 -
                0.0
                                   0.1
                                       Summer   Fall
                      eo
                      o
                      I
                      o
          P   8
          9   9
          8   S
   S*-   in
 .  o   T
S  S   9
Poo
8   °   5
        o
           —I
            co
            o
       in
       o
CO
c\i
•^  m
CO  4
o
S
o   o o
 Particle Size (|jm)
        Figure 5-6.  Geometric mean infiltration factor (indoor/outdoor ratio) for hourly nighttime,
                    nonsource data for two seasons.  Box plots of air exchange rates are shown as
                    inserts for each plot.  (Boston, 1998)
        Source:  Long et al. (2000b).
 1      temperatures generally vary over a small range. Combined, these effects can produce order of
 2      magnitude variations in k between particles of different size and, in the case of electrophoresis
 3      and thermophoresis, particles of the same size.
 4           Particle penetration efficiency into the indoor /we depends on particle size and air exchange
 5      rates.  Penetration varies with particle size because of the size-dependent deposition of particles
 6      caused by impaction, interception, and diffusion of particles onto surfaces as they traverse
 7      through cracks and crevices. Penetration also is affected by air exchange rates.  When air
 8      exchange rates are high, P approaches unity because the majority of ambient particles have less
 9      interaction with the building shell. In contrast, when air exchange rates are low, P is governed by
10      particle deposition as particles travel through cracks and crevices.
       March 2001
                                  5-55
              DRAFT-DO NOT QUOTE OR CITE

-------
 1          Significant advancements have been made in the past few years to better characterize
 2     particle deposition rates and penetration factors. Several new studies, including two in which
 3     semi-continuous measurements of size distributions were measured indoors and outdoors, have
 4     produced new information on these quantities, which are key to understanding the contributions
 5     of ambient PM to indoor PM concentrations (Equation 5-10).
 6          Studies involving semi-continuous measurements of indoor and outdoor particle size
 7     distributions have been used to estimate k and P as a function of particle size (Vette et al., 2001;
 8     Long et al., 2000b; Abt et al., 2000b).  These studies each demonstrated that the indoor/outdoor
 9     concentration ratios (Cao/C in Equation 5-11) were highest for accumulation mode particles and
10     lowest for ultrafine and coarse-mode particles.  Various approaches were used to estimate size-
11     specific values for k and P. Vette et al. (2001) and Abt et al. (2000b) estimated k by measuring
12     the decay of particles at times when indoor levels were significantly elevated. Vette et al. (2001)
13     estimated P using measured values of k and indoor/outdoor particle measurements during
14     nonsource nighttime periods. Long et al. (2000b) used a physical-statistical model, based on
15     Equation 5-10, to estimate k and P during nonsource nighttime periods. The results for k
16     reported by Long et al. (2000b) and Abt et al. (2000b) are compared with other studies in
17     Figure 5-7. Although not shown in Figure 5-7, the results for k obtained by Vette et al. (2001)
18     were similar to the values of £ reported by Abt et al. (2000b) for particle sizes up to 1 /u.m.
19     Results for P by Long et al. (2000b) show that penetration was highest for accumulation-mode
20     particles and decreased substantially for coarse-mode particles (Figure 5-8).  The results for
21     P reported by Vette et al. (2001) show similar trends, but are lower than those reported by Long
22     et al. (2000b). This likely is because of lower air-exchange rates in the Fresno, CA, residence
23     (a ~ 0.5 h"1; Vette et al., 2001) than the Boston, MA, residences (a > 1 h'1; Long et al., 2000b).
24     These data for P and k illustrate the role that the building shell may provide in increasing the
25     concentration of particles because of indoor sources and reducing the concentration of indoor
26     particles from ambient sources, especially for homes with low air-exchange rates.
27
28     Compositional Differences Between Indoor-Generated and Ambient-Generated
29     Particulate Matter
30          Wilson et al. (2000) discuss the differences in composition between particles from indoor
31     and outdoor sources. They note that, because of the difficulty in separating indoor PM into
        March 2001
5-56
DRAFT-DO NOT QUOTE OR CITE

-------





10-


^_
J
S 1-
0.1-
-

0.









•



31
o Byrne (1992)
0 Foghefa/. (1997)a
v Ligockiefa/. (1990)"
v Ozkaynak ef a/. (1 996a)c
n Sinclair ef al. (1 988, 1 990)b
0 Thatcher and Layton (1 995)"
O Wallace ef al. (1997)d
T Abtefa/. (2000b)d
• Longefa/. (2000b)d'e
• • • Lai and Nazaroff (2000)'
f -
• """"•--.. * *


0.1









f-
* * i












' ,
' i












«













'f \s$]
r
i
error bar
hdudes 0
2,








^»
ifj

error bar
IndudesO
5




.-'
.*
.-•*

f
IT 1
."f ,







_.'*




f
J.
r
i



10
                                                       Particle size (|jm)
                    'Decay rates represent Summary Estimates from the four houses examined.
                    "Decay rates are based on sulfate and are presented as <2.5 um.
                    Estimates were computed using a surface-to-volume ratio of 2 m-1 (Koutrakis ef at, 1992).
                    'Data represent PM^j
                    "Particle sizes are the midpoint of the ranges examined.
                    •Decay rates presented are estimates of k for nightiy average data from all nine study homes.
                    Decay rates are theorectically modeled deposition values for smooth indoor surfaces and homogeneous and isotropically turbulent air flow.
                    Presented curves assume typical room dimensions (3 m x 4 m x S m) and a friction velocity of 1.0 cm/s.

        Figure 5-7.  Comparison of deposition rates from this study with literature values (adapted
                      from Abt et al., 2000b). Error bars represent standard deviations for same-
                      study estimates.

        Source:  Long et al. (2000b).
1
2
3
4
5
6
1
8
ambient and nonambient PM, there is little direct experimental information on the composition
differences between the two.  Although experimental data are limited, Wilson et al. (2000)
suggest the following.

      Photochemistry is significantly reduced indoors; therefore, most secondary sulfate [H2SO4,
      NH4HSO4, and (NH4)2SO4] and nitrate (NH4NO3) found indoors come from ambient sources.
      Primary organic emissions from incomplete combustion may be similar, regardless of the source.
      However, atmospheric reactions of polyaromatic hydrocarbons and other organic compounds
       March 2001
                                                5-57
DRAFT-DO NOT QUOTE OR CITE

-------
£

•o

UJ
c
1.2


1.1


1.0



0.9


0.8






0.6


0.5


0.4


0.3



0.2


0.1


0.0
                                                                                             1.2


                                                                                             1.1


                                                                                             1.0


                                                                                             0.9


                                                                                             0.8 ^



                                                                                             0.7  jiT
                                                                                                 £
                  8.

                  1
                  o"
                         o
                         o
                                  9

                                  o
9  ^

o  a
                                                                          s
                                                                                             0.6
                                                                                             0.5  g

                                                                                                 s-
                                                                                             0.4  Q
                                                                                             0.3


                                                                                             0.2


                                                                                             0.1


                                                                                             0.0
                                              Size Interval (pm)
       Figure 5-8. Penetration efficiencies and deposition rates from models of nightly average

                    data. Error bars represent standard errors.  (Boston, 1998, winter and

                    summer)



       Source: Long et al. (2000b).
 1           produce highly oxygenated and nitrated products, so these species are also of ambient origin.


 2           Gasoline, diesel fuel, and vehicle lubricating oil all contain naturally present metals or metal


 3           additives.  Coal and heavy fuel oil also contain more metals and nonmetals, such as selenium and


 4           arsenic, than do materials such as wood or kerosene burned inside homes. Environmental


 5           tobacco smoke (ETS), however, with its many toxic components, is primarily an indoor-generated


 6           pollutant


 7


 8           Particles  generated indoors may have different chemical and physical properties than those


 9      generated by anthropogenic ambient sources.  Siegmann et al. (1999) have demonstrated that


10      elemental carbon in soot particles generated indoors have different properties than in those


11      generated outdoors by automotive or diesel engines. In the United States, combustion-product
        March 2001
                                           5-58
                                                    DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
 PM in the ambient/outdoor air generally is produced by burning fossil fuels (e.g., coal, gasoline,
 fuel oil) and wood, whereas combustion-product PM from indoor sources is produced by
 biomass burning (e.g., tobacco, wood, foods, etc.). However, some indoor sources of PM (such
 as cigarette smoking, meat cooking, and coal burning) occur both indoors and outdoors and may
 constitute an identifiable portion of measured ambient PM (Cha et al., 1996; Kleeman and Cass,
 1998).

 Indoor Air Chemistry
      Gas- and aerosol-phase chemical reactions in the indoor jj,e are responsible for secondary
 particle formation and modification of existing particles. Homogeneous gas phase reactions
 involving ozone and terpenes (specifically d-limonene, a-terpinene, and a-pinene) have been
 identified as an important source of submicron particles (Weschler and Shields, 1999). Terpenes
 are present in several commonly available household cleaning products and d-limonene has been
 identified in more than 50% of the buildings monitored in the BASE study (Hadwen et al.,  1997).
 Long et al. (2000a) found that when PineSol (primary ingredient is a-pinene) was used indoors,
 indoor PM2 5 mass concentrations increased by 3 to 32 //g m'3 (indoor ozone concentrations
 unknown, but ambient ozone concentrations were 44 to 48 ppb). Similarly, a 10-fold increase in
 number counts of 0.1 to 0.2 //m particles was observed in an experimental office containing
 supplemented d-limonene and normally encountered indoor ozone concentrations (< 5 to
 45 ppb), resulting in an average increase in particle mass concentration of 2.5 to 5.5 //g m"3
 (Weschler and Shields, 1999). Ozone appears to be the limiting reagent as particle number
 concentration varied proportionally to ozone concentrations (Weschler and Shields, 1999).  Other
 studies showed similar findings (e.g., Jang and Kamens, 1999; Wainman et al., 2000).

 Indoor Sources of Particles
     The major sources of indoor PM in nonsmoking residences and buildings include
 suspension of PM from bulk material, cooking, cleaning, and the use of combustion devices,
 such as stoves and kerosene heaters. Human and pet activities also lead to PM detritus
production (from tracked-in soil, fabrics, skin and hair, home furnishings, etc.), which is found
ubiquitously in house dust deposited on floors and other ulterior surfaces. House dust and lint
particles may be resuspended indoors by agitation (cleaning) and turbulence (HVAC systems,
       March 2001
                                         5-59
DRAFT-DO NOT QUOTE OR CITE

-------
 1     human activities, etc.). Ambient particles that have infiltrated into the indoor /we also may be
 2     resuspended after deposition to indoor surfaces. Typically, resuspension of particles from any
 3     source involves coarse-mode particles (>1 Aim); particles of smaller diameter are not resuspended
 4     efficiently.  On the other hand, cooking produces both fine- and coarse-mode particles, whereas
 5     combustion sources typically produce fine-mode particles.
 6           Environmental tobacco smoke (ETS) is also a major indoor source of PM.  It is, however,
 7     beyond the scope of this chapter to review the extensive literature on ETS. A number of articles
 8     provide source strength information for cigarette or cigar smoking (e.g., Daisey et al. [1998] and
 9     Nelson etal. [1998]).
10           A study conducted on two homes in the Boston metropolitan area (Abt et al., 2000a)
11     showed that indoor PM sources predominate when air exchange rates were <1 h"1, and outdoor
12     sources predominate when air exchange rates were >2 h"1.  The authors attributed this to the fact
13     that when air-exchange rates were low (<1 h'1), particles released from indoor sources tend to
14     accumulate because particle deposition is the mechanism governing particle  decay and not air
15     exchange. Particle deposition rates are generally <1 h'1, especially for accumulation-mode
16     particles. When air-exchange rates were higher (>2 h'1), infiltration of ambient aerosols and
17     exfiltration of indoor-generated aerosols occur more rapidly, reducing the impact of indoor
18     sources on indoor particle levels. The study also confirmed previous findings that the major
19     indoor sources of PM are cooking, cleaning, and human activity.  They discuss the size
20     characteristics of these ubiquitous sources and report the following.
21
22           The size of the particles generated by these activities reflected their formation processes.
23           Combustion processes (oven cooking, toasting, and barbecuing) produced fine particles and
24           mechanical processes (sauteing, firing, cleaning, and movement of people) generated coarse
25           particles. These activities increased particle concentrations by many orders of magnitude higher
26           than outdoor levels and altered indoor size distributions. (Abt et al., 2000a; p. 43)
27
28      They also note that variability in indoor PM for all size fractions  was greater than for outdoor
29      PM, especially for short averaging times (2 to 33 times higher).
30           In a separate study conducted in nine nonsmoking homes in the Boston area, Long et al.
31      (2000a) concluded that the predominant source of indoor fine particles was infiltration of outdoor
        March 2001
5-60
DRAFT-DO NOT QUOTE OR CITE

-------
  1      particles, and that cooking activities were the only other significant source of fine particles.
  2      Coarse particles, however, had several indoor sources, such as cooking, cleaning, and various
  3      indoor activities. This study also concluded that more than 50% of the particles (by volume)
  4      generated during indoor events were ultrafine particles.  Events that elevated indoor particle
  5      levels were found to be brief, intermittent, and highly variable, thus requiring the use of
  6      continuous instrumentation for their characterization. Table 5-8 provides information on the
  7      mean volume mean diameter (VMD) for various types of indoor particle sources. The
  8      differences in mean VMD confirm the clear separation of source types and suggest that there is
  9      very little resuspension of accumulation-mode PM.  In addition, measurements of organic and
10      elemental carbon indicated that organic carbon had significant indoor sources, whereas elemental
11      carbon was primarily of ambient origin.
12           Vette et al. (2001) found that resuspension was a significant indoor source of particles
13      >1 /zm, whereas fine- and accumulation-mode particles were not affected by resuspension.
14      Figure 5-9 shows the diurnal variability in the indoor/outdoor  aerosol concentration ratio from an
15      unoccupied residence in Fresno.  The study was conducted in the absence of common indoor
16      particle sources such as cooking and cleaning. The data  in Figure 5-9 show the mean
17      indoor/outdoor concentration ratio for particles >1 /zm increased dramatically during daytime
18      hours. This pattern was consistent with indoor human activity levels. In contrast, the mean
19      indoor/outdoor concentration ratio for particles <1 //m (fine- and accumulation-mode particles)
20      remain fairly constant during both day and night.
21
22      5.4.3.2.3  Time/Activity Patterns
23           Total exposure to PM is the sum of various microenvironmental exposures that an
24      individual encounters during the day and will depend on  the microenvironments occupied.
25      As discussed previously, PM exposure in each microenvironment is the sum of exposures from
26      ambient sources (Eag), indoor sources (Eig), and personal  activities (Epact). Eag and Eig are
27      determined by the microenvironments in which an individual spends time; whereas Epact is
28      determined by the personal activities that he/she conducts while in those microenvironments.
29           Determining microenvironments and activities that contribute significantly to human
30      exposure begins with establishing human activity pattern information for the general population,
31      as well as subpopulations. Personal exposure and time activity pattern studies have shown that
        March 2001
5-61
DRAFT-DO NOT QUOTE OR CITE

-------
             TABLE 5-8. VOLUME MEAN DIAMETER (VMD) AND MAXIMUM PM2 5
                     CONCENTRATIONS OF INDOOR PARTICLE SOURCES a'b
Size Statistics
Particle Source
Cooking
Baking (Electric)
Baking (Gas)
Toasting
Broiling
Sauteing
Stir-Frying
Frying
Barbecuing
Cleaning
Dusting
Vacuuming
Cleaning with Pine Sol
General Activities
Walking Vigorously (w/Carpet)
Sampling w/Carpet
Sampling w/o Carpet
Burning Candles
N

8
24
23
4
13
3
20
2

11
10
5

15
52
26
7
Indoor Activity
Mean VMD
Cum)

0.1 89f
0.1 07f
0.138f
0.1 14f
0.184f,3.48g
0.135f
0.173f
0.159f

5.388
3.86g
0.097f

3.96g
4.25g
4.28B
0.311f
Background3'11
Mean VMD
Cum)

0.22 lf
0.224f
0.222f
0.236f
0.223f, 2.93s
0.277f
0.223f
0.205f

3.53E
2.79B
0.238f

3.18g
2.63B
2.93g
0.224f
Maximum
Mean

14.8
101.2
54.9
29.3
65.6
37.2
40.5
14.8

22.6
6.5
11.0

12.0
8.0
4.8
28.0
PM2.5
Concentration0'11
SD

7.4
184.9
119.7
43.4
95.4
31.4
43.2
5.2

22.6
3.9
10.2

9.1
6.6
3.0
18.0
        Notes:
        "All concentration data corrected for background particle levels.
        Includes only individual particle events that were unique for a given time period and could be detected above
        background particle levels.
        CPM concentrations in yug/m3.
        dMaximum concentrations computed from 5-min data for each activity.
        'Background data are for time periods immediately prior to the indoor event.
        fSize statistics calculated for PV0 02.0 5 using SMPS data.
        8Size statistics calculated for PV0 7.10 using APS data.

        Source: Long et al. (2000a).
1      different populations have varying time activity patterns and, accordingly, different personal PM

2      exposures. Both characteristics will vary greatly as a function of age, health status, ethnic group,

3      socioeconomic status, season, and region of the country.  Collecting detailed time activity data

4      can be very burdensome on participants but is clearly valuable in assessing human exposure and
       March 2001
5-62
DRAFT-DO NOT QUOTE OR CITE

-------
                     o
                     1
                      o
                      o
                     1
                     O
                      o
                     I
                         2.0'
                          0.0
                                            Time
       Figure 5-9. Mean hourly indoor/outdoor particle concentration ratio from an unoccupied
                   residence in Fresno, CA, during spring 1999.
       Source: Vette et al. (2001).
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
microenvironments. For modeling purposes, human activity data frequently come from general
databases that are discussed below.
     The gathering of human activity information, often called "time-budget" data, started in the
1920s; however, their use for exposure assessment purposes only began to be emphasized in the
1980s. Many of the largest U.S. human activity databases have been consolidated by EPA's
National Exposure Research Laboratory's (NERL) into one comprehensive database containing
over 22,000 person-days of 24-h activity known as the Consolidated Human Activity Database,
or CHAD (Glen et al., 1997). The information in CHAD will be accessible for constructing
population cohorts of people with diverse characteristics that are useful for analysis and
modeling (McCurdy, 2000).  See Table 5-2 for a summary listing of human activity studies in
CHAD.  Most of the databases in CHAD are available elsewhere, including the National Human
       March 2001
                                         5-63
DRAFT-DO NOT QUOTE OR CITE

-------
 1      Activity Pattern Survey (NHAPS), California's Air Resources Board (CARB), and the University
 2      of Michigan's Institute for Survey Research data sets.
 3           Although CHAD provides a very valuable resource for time and location data, there is less
 4      information on PM generating personal activities.  In addition, very few of the time-activity
 5      studies have collected longitudinal data within a season or over multiple seasons. Such
 6      longitudinal data are important in understanding potential variability in activities and how they
 7      impact correlations between PM exposure and ambient site measurements for both total PM and
 8      PM of ambient origin.
 9
10      5.4.3.3 Impact of Ambient Sources on Exposures to Particulate Matter
11           Different sources may generate ambient PM with different aerodynamic and chemical
12     characteristics, which may, in turn, result in different health responses. Thus, to fully understand
13      the relationship between PM exposure and health outcome, exposure from difference sources
14     should be identified and quantified. Source apportionment techniques provide a method for
15     determining personal exposure to PM from specific sources. Daily contributions from sources
16     that have no indoor component can be used as tracers to generate exposure to ambient PM of
17     similar aerodynamic size or directly as exposure surrogates in epidemiologic analyses. The
18     recent EPA PM Research Needs Document (U. S.  Environmental Protection Agency, 1998)
19     recommended use of source apportionment techniques to determine daily time-series of source
20     categories for use in community, time-series epidemiology.
21           A number of epidemiological studies (discussed more fully in Chapter 6) have evaluated
22     the relationship between health outcomes and sources of particulate matter determined from
23     measurements at a community monitor. These studies suggest the importance of examining
24.    sources and constituents of indoor, outdoor, and personal PM. Ozkaynak and Thurston (1987)
25     evaluated the relationship between particulate matter sources and mortality in 36 Standard
26     Metropolitan Statistical Areas (SMSAs). Particulate matter samples from EPA's Inhalable
27     Particle (IP) Network were analyzed for SO42" and NCy by automated colorimetry, and elemental
28     composition was determined with X-ray fluorescence (XRF). Mass concentrations from five
29     particulate matter source categories were determined from multiple regression of absolute factor
30      scores on the mass concentration: (1) resuspended soil, (2) auto exhaust, (3) oil combustion,
31      (4) metals, and (5) coal combustion.
        March 2001
5-64
DRAFT-DO NOT QUOTE OR CITE

-------
  1           Mar et al. (2000) applied factor analysis to evaluate the relationship between PM
  2      composition (and gaseous pollutants) in Phoenix. In addition to daily averages of PM2 5 elements
  3      from XRF analysis, they included in their analyses organic and elemental carbon in PM2 s and
  4      gaseous species emitted by combustion sources (CO, NO2, and SO2).  They identified five factors
  5      classified as (1) motor vehicles, (2) resuspended soil, (3) vegetative burning, (4) local SO2, and
  6      (5) regional sulfate.
  7           Laden et al. (2000) applied specific rotation factor analysis to particulate matter
  8      composition (XRF) data from six eastern cities (Ferris et al., 1979). Fine particulate matter was
  9      regressed on the recentered scores to determine the daily source contributions. Three main
10      sources were identified: (1) resuspended soil (Si), (2) motor vehicle (Pb), and (3) coal
11      combustion (Se).
12           Source apportionment or receptor modeling has been applied to the personal exposure data
13      to understand the relationship between personal and ambient sources of particulate matter.
14      Application of source apportionment to ambient, indoor, and personal PM composition data is
15      especially useful in sorting out the effects of particle size and composition.  If a sufficient
16      number of samples are analyzed with sufficient compositional detail, it is possible to use
17      statistical techniques to derive source category signatures, identify indoor and outdoor source
18      categories and estimate their contribution to indoor and personal PM.
19       .    Positive Matrix Factorization (PMF) has been applied to the PTEAM database by
20      Yakovleva et al. (1999). The authors utilize mass and XRF elemental composition data from
21      indoor and outdoor PM25 and personal, indoor, and outdoor PM10 samples.  PMF is an advance
22      over ordinary factor analysis because it allows measurements below the quantifiable limit to be
23      used by weighting them by their uncertainty. This effectively increases the number of species
24      that can be used in the model. The factors used by the authors correspond to general source
25      categories of PM, such as outdoor soil, resuspended indoor soil, indoor soil, personal activities,
26      sea-salt, motor vehicles, nonferrous metal smelters, and secondary sulfates.  PMF, by identifying
27      not only the various source factors but also apportioning them  among the different monitor
28      locations (personal, indoor, and outdoor), was able to quantify an estimate of the contribution of
29      resuspended indoor dust to the personal cloud (15% from indoor soil and 30% from resuspended
30      indoor soil).  Factor scores for these items then were used in a regression analysis to estimate
31      personal exposures (Yakovleva etal., 1999).
        March 2001
5-65
DRAFT-DO NOT QUOTE OR CITE

-------
 1          The most important contributors to PMIO personal exposure were indoor soil, resuspended
 2     indoor soil, and personal activities; these accounted for approximately 60% of the mass
 3     (Yakovleva et al., 1999). Collectively, they include personal cloud PM, smoking, cooking, and
 4     vacuuming.  For both PM2.5 and PM10, secondary sulfate and nonferrous metal operations
 5     accounted for another 25% of PM mass. Motor vehicle exhausts, especially starting a vehicle
 6     inside of an attached garage, accounted for another 10% of PM mass. The authors caution that
 7     these results may not apply to other geographic areas, seasons of the year, or weather conditions.
 8          Simultaneous measurement of personal (PM10) and outdoor measurements (PM25 and
 9     PMIO) were evaluated as a three-way problem with PMF, which allowed for differentiation of
10     source categories based on their variation in time and type of sample, as well as their variation in
11     composition. By use of this technique,  it was possible to identify three sources of coarse-mode,
12     soil-type PM. One was associated with ambient soil, one was associated with indoor soil
13     dispersed throughout the house, and one was  associated with soil resulting from the personal
14     activity of the subject.
15          Two other source apportionment models have been applied to ambient measurement data
16     and can be used for the personal exposure studies.  The effective variance weighted Chemical
17     Mass Balance (CMB) receptor model (Watson et al., 1984, 1990, 1991) solves a set of linear
18     equations that incorporate the uncertainty in the sample and source composition. CMB requires
19     the composition of each potential source of PM and the uncertainty for the sources and ambient
20     measurements. Source apportionment with CMB can be conducted on individual samples,
21     however, composition of each of the sources  of PM must be known. An additional source
22     apportionment model, UNMIX (Henry  et al., 1994) is a multivariate source apportionment
23     model. UNMIX is similar to PMF, but does not use explicitly the measurement uncertainties.
24     Because measurement uncertainties are not used, only species above the detection limit are
25     evaluated in the model. UNMIX provides the number of sources and source contributions and
26     requires a similar number of observations as PMF.
27          The Yakovleva et al. (1999) study demonstrates that source apportionment techniques also
28     could be very useful in determining parameters needed for exposure models and for determining
29     exposure to ambient-generated PM. Exposure information,  similar to that obtained hi the
30     PTEAM study, but including other PM components useful for definition of other source
31     categories (e.g., elemental [EC] and organic carbon [OC]; organic tracers for elemental carbon
       March 2001
5-66
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 from diesel vehicle exhaust, gasoline vehicle exhaust, and wood combustion; nitrate; Na; Mg and
 other heavy metal tracers; and, also, gas-phase pollutants) would be useful as demonstrated in the
 use of EC/OC and gas-phase pollutants by Mar et al. (2000).

 5.4.3.4  Correlations of Particulate Matter with Other Pollutants
      Several epidemiological studies have included the gaseous pollutants CO, NO2, SO2, and
 O3 along with PM10 or PM2 5 in the analysis of the statistical association of health responses with
 pollutants. In a recent study, the personal exposure to O3 and NO2 were determined, as well as
 that to PM25 and PM25.10 for a cohort 15 elderly subjects in Baltimore, MD. Sarnat et al. (2000)
 conclude that the potential for confounding of PM2 5 by O3, NO2, or PM 2 5.10 appears to be
 limited, because, despite significant correlations observed among ambient pollutant
 concentrations, the correlations among personal exposures were low. Spearman correlations for
 14 subjects in summer and 14 subjects  in winter are given in Table 5-9 for relationships between
 personal PM2 5 and ambient concentrations of PM25, PM2 5.10, O3, and NO2.  In contrast to ambient
 concentrations, neither personal exposure to total PM2 5 nor PM2 5 ambient origin was correlated
 significantly with personal exposures to the co-pollutants, PM25.10, nonambient PM25, O3, NO2,
 and SO2. Personal-ambient associations for PM2 5.10, O3, NO2, and SO2 were similarly weak and
 insignificant. It should be noted that measured personal exposures to O3, NO2, and SO2 were
 below their respective LOD for 70% of the samples.
     A newly developed Roll-Around  System (RAS) was used to evaluate the hourly
 relationship between gaseous pollutants (CO, O3, NO2, SO2, and VOCs) and PM (Chang et al.,
 2000). Exposures were characterized over a 15-day period for the summer and winter in
 Baltimore, based on scripted activities to simulate activities performed by older adults (65+ years
 of age).  Spearman rank correlations were reported for PM2 5, O3,  CO, and toluene for both the
 summer and winter and the correlations are given for each microenvironment in Table 5-10:
 indoor residence, indoor other, outdoor near roadway, outdoor away from road,  and in vehicle.
No significant relationships (p < 0.05) were found between hourly PM2 5 and O3. Significant
relationships were found between hourly PM2 5 and CO: indoor residence, winter; indoor other,
 summer and winter; and outdoor away from roadway,  summer. Significant relationships also
were found between hourly PM2 5 and toluene: indoor residence, winter; indoor other, winter;
and in vehicle, winter. The significant relationships between CO  and PM2 5 in the winter may be
       March 2001
                                         5-67
DRAFT-DO NOT QUOTE OR CITE

-------
TABLE 5-9. CORRELATIONS BETWEEN PERSONAL PM2.5 AND AMBIENT
               POLLUTANT CONCENTRATIONS1
Personal PM2.5
vs. Ambient:
SUMMER














WINTER













Median
Median
Subject
SA1
SA2
SA5
SB1
SB2
SB3
SB4
SB5
SB6
SCI
SC2
SC3
SC4
SC5
WAI
WA2
WA4
WAS
WB1
WB2
WB3
WB4
WC1
WC2
WC3
WC4
WC5
WC6
Summer
Winter
PM2.5 03
0.55
0.55
0.59
0.65
-0.21
0.52
0.75
0.73
0.53
• 0.95
0.78
0.85
0.78
0.55
0.22
-0.38
-0.18
0.22
0.50
0.62
0.55
-0.12
0.74
0.79
0.28
0.19
0.57
0.01
0.76
0.25
'Correlations represent Spearman'
0.15
0.31
0.18
0.40
-0.62
0.55
0.62
0.45
0.15
0.75
0.65
0.75
0.66
0.51
-0.18
-0.07
0.67
-0.43
-0.84
-0.32
-0.45
-0.01
-0.62
-0.88
-0.42
-0.84
-0.62
-0.03
0.48
-0.43
NO2
0.38
0.66
0.52
-0.15
0.57
-0.14
-0.34
-0.42
-0.38
0.66
0.36
0.73
0.59
0.32
-0.26
-0.36
-0.22
0.67
0.77
0.59
0.62
0.34
-0.15
0.17
0.03
0.50
0.08
0.65
0.37
0.26
PM2.5.10
-0.12
0.57
0.64
0.38
0.15
-0.04
-0.12
0.23
0.12
0.65
0.51
0.65
0.70
0.43
-0.05
-0.70 .
-0.29
0.50
0.41
0.09
0.04
-0.10
0.44
0.77
0.57
0.45
0.57
0.37
0.41
0.39
Personal PM2 5
of Ambient Origin vs. Ambient:
03
0.27
0.21
0.33
0.59
0.26
0.52
0.45
0.36
-0.03
0.55
0.66
0.69
0.50
0.34
-0.75
-0.15
-0.33
-0.72
-0.57
-0.76
-0.77
-0.50
-0.64
-0.57
-0.77
-0.72
-0.76
-0.75
0.41
-0.76
s r values; italicized values indicate significance at the
NO2
0.77
0.64
0.57
-0.74
0.08
-0.20
-0.29
-0.48
-0.57
0.65
0.65
0.77
0.50
0.33
-0.04
-0.15
0.20
-0.09
0.53
0.59
0.56
0.65
0.02
0.25
0.30
0.22
0.05
0.19
0.42
0.21
a = 0.05 level.
PM2.5.10
0.15
0.68
0.79
-0.03
0.33
0.00
-0.14
0.33
0.32
0.57
0.76
0.50
0.51
0.27
-0.24
0.02
0.00
0.40
0.66
0.59
0.60
0.48
0.69
0.77
-0.45
0.67
0.42
-0.45
0.33
0.45

Source: Samat et al. (2000).
March 2001



5-68
DRAFT-DO NOT QUOTE OR CITE

-------
            TABLE 5-10. CORRELATIONS BETWEEN HOURLY PERSONAL PM25 AND
                                      GASEOUS POLLUTANTS
Indoor
Residence

PM^vs.
Summer
Winter
PM^vs.
Summer
Winter
PM25vs.
Summer
Winter
N
03
35
56
CO
41
59
Toluene
46
66
rs

0.29
0.05

0.25
0.43a

0.23
0.38a
Indoor Other
N

16
37

19
39

21
47
rs

-0.14
-0.06

0.59a
0.62"

-0.14
0.44a
Outdoor Near
Roadway
N

10
11

13
13

14
17
rs

0.05
-0.28

0.14
0.37

0.26
0.40
Outdoor Away
from road
N

12
7

12
8

14
8
rs

0.45
0.04

0.62
0.41

0.02
0.48
In Vehicle
N

37
34

46
37

48
42
rs

0.21
-0.10

0.23
0.10

0.12
0.43a
         "Correlations represent Spearman's r values; italicized values indicate significance at the a = 0.05 level.
         Source: Chang et al. (2000).
 1      caused by reduced air-exchange rates that could allow them to accumulate (Chang et al., 2000).
 2      Although no significant correlation was found between in vehicle PM2 5 and CO, toluene, which
 3      is a significant component of vehicle exhaust (Conner et al., 1995), was correlated significantly
 4      to PM2 5 in the winter.
 5          Carrer et al. (1998) present data on the correlations among personal and
 6      microenvironmental PM10 exposures and concentrations and selected environmental chemicals
 7      that were monitored simultaneously (using a method that was not described). These chemicals
 8      were nitrogen oxides (NOJ, carbon monoxide (CO), and total volatile organic compounds
 9      (TVOC), benzene, toluene, xylene, and formaldehyde. The Kendall T correlation coefficient was
10      used; only results significant at p < 0.05 are mentioned here. Significant associations were found
11      only between the following pairs of substances (T shown in parentheses):  personal PM10 (24 h)
12      and NOX (0.34), CO (0.34), TVOC (0.18), toluene (0.19), and xylene (0.26); office PM10 and NOX
13      (0.31); home PM10 and NOX (0.24), CO (0.24), toulene (0.17), and xylene (0.25). Surprisingly,
       March 2001
5-69
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
because most of the chemical substances are associated with motor vehicular emissions, there
was no significant correlation between "commuting PM10" and any of the substances (Carrer
etal., 1998).
5.5. SUMMARY OF PARTICULATE MATTER CONSTITUENT DATA
5.5.1  Introduction
     Atmospheric PM contains a number of chemical constituents that may be of significance
with respect to the human exposure and health effects. These constituents may be either
components of the ambient particles or bound to the surface of particles.  They may be elements,
inorganic species, or organic compounds. A limited number of studies have collected data on
concentrations of elements, acidic aerosols, and polycyclic aromatic hydrocarbons (PAHs) in
ambient, personal, and microenvironmental PM samples.  But, there have not been extensive
analyses of the constituents of PM in personal or microenvironmental samples. Data from
relevant studies are summarized in this section. The summary does not address bacteria,
bioaerosols, viruses, or fungi (e.g., Owen et al., 1992; Ren et al., 1999).

5.5.2  Monitoring Studies That Address Particulate Matter Constituents
     A limited number of studies have measured the constituents of PM in personal or
microenvironmental samples. Relevant studies published in recent years are summarized in
Tables 5-11 and 5-12 for personal exposure measurements of PM and microenvironmental
samples, respectively. Studies that measured both personal and microenvironmental samples are
included in Table 5-11.
     The largest database on personal, microenvironmental, and outdoor measurements of PM
elemental concentrations is the PTEAM study (Ozkaynak et al., 1996b).  The results are
highlighted in the table and discussed below. The table shows that a number of studies have
measured aerosol acidity, sulfate, ammonia, and nitrate concentrations. Also, a number of
studies have measured PAHs, both indoors and outdoors.  Other than the PAHs, there is little
data on organic constituents of PM.
       March 2001
                                         5-70
DRAFT-DO NOT QUOTE OR CITE

-------






W2
CE
•
P.
|

£E

5
x




















Summary of Results

8
"3
CO
'o
o

CO
e
o

JS
S
Q.
g

U
O
S
CO

0
1
tju
o
S:
ea
2
CO
§
c
6
s
a.
«•§
llllj
S •= T3 3 °i
Illia
a s ° ^ v, •
lir was the major source for mo
iroviding 70 to 100% of the obi
tions for 12 of the 15 elements.
ts for central monitoring site v<
nces were 0.98 for sulfur and 0
[except copper).
g 01- S .2 •§ -2
| s s ;! s s
111 §11






a
"3
•5

oo
•<
0
u"
T3
'e
.S

OS
"«
•s


•i
0
1
s

„
§c
_
w
o
door or outdo
omparison.
\s and Pb levels higher than in<
3 community ambient site for ci
1--
W *fl>
£ S






8

S

r-
»n
§
•a
S
1

1
^
S
o
?.
J3
OT
^
1

£
e
«
"^
S^
loors. ButPti
cations than s
2.5 Mn higher outdoors than ind
ntrations higher at two fixed lo<
ts' homes.

-------
  C/3
   Ct
ged from 0.7 to 0.9 for three PM

                4J (4-1
                i 2
                g a
              t
              •g  2
              ill
                i




              S
              o
                                          5
i
                                          3
         a   •§
                       
       P
       U
           2
           1
                                            o
                                            o
                                                          <
                                                          u
                                                          o-
                §
                J2o
                                 .g
                                 "a
                                 Q
                                                          (X 0.
March 2001
5-72
                               DRAFT-DO NOT QUOTE OR CITE

-------
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
 5.5.3 Key Findings
 5.5.3.1 Correlations of Personal and Indoor Concentrations with Ambient Concentrations
        of Participate Matter Constituents
      The elemental composition of PM in personal samples was measured in the PTEAM study,
 the first probability-based study of personal exposure to particles. A number of important
 observations, made from the PTEAM data collected in Riverside, CA, are summarized by
 Ozkaynak et al. (1996b). Population-weighted daytime personal exposures averaged
 150 ± 9 Aig/m3, compared to concurrent indoor and outdoor concentrations of 95 ± 6 yUg/m3. The
 personal exposure measurements suggested that there was a "personal cloud" of particles
 associated with personal activities. Daytime personal exposures to 14 of the 15 elements
 measured in the samples were considerably greater than concurrent indoor or outdoor
 concentrations, with sulfur being the only exception.
      The PTEAM data also showed good agreement between the concentrations  of the elements
 measured outdoors at the backyard of the residences with the concentrations measured at the
 central site in the community. The agreement was excellent for sulfur. Although the particle and
 element mass concentrations were higher in personal samples than for indoor or outdoor samples,
 a nonlinear mass-balance method showed that the penetration factor was nearly 1  for all particles
 and elements.
      Similarly to the PTEAM results, recent measurements of element concentrations in
 NHEXAS showed elevated concentrations of As and Pb in personal samples relative to indoor
 and outdoor samples (Clayton et al., 1999b). The elevated concentrations of As and Pb were
 consistent with elevated levels of PM in personal samples (median particle exposure of
 101 //g/m3), compared to indoor concentrations (34.4 ^g/m3).  There was a strong association
 between personal and indoor concentrations and indoor and outdoor concentrations for both As
 and Pb. However, there were no central site ambient measurements for comparison to the
 outdoor or indoor measurements at the residences.
     Manganese (Mn) concentrations were measured in PM2 5 samples collected in Toronto
 (Crump, 2000). The mean PM2 5 Mn concentrations were higher outdoors than indoors. But the
outdoor concentrations measured at the participant's homes were lower than those measured at
two fixed locations. Crump (2000) suggested that the difference in the concentrations may have
       March 2001
                                        5-73
                                                           DRAFT-DO NOT QUOTE OR CITE

-------
 1     been because the fixed locations were likely closer to high-traffic areas than were the
 2     participant's homes.
 3          Studies of acidic aerosols and gases typically measure strong acidity (H+), SO42', NH4+, and
 4     NO3". The relationship between the concentrations of these ions and the relationship between
 5     indoor and outdoor concentrations have been addressed in a number of studies during which
 6     personal samples, microenvironmental, and outdoor samples have been collected, as shown in
 7     Tables 5-11 and 5-12. Key findings from these studies include those shown below.
 8          • Acid aerosol concentrations measured at the residences in the Uniontown, PA, study were
 9            significantly different from those measured at a fixed ambient site located 16 km from the
10            community.  But, Leaderer et al. (1999) reported that the regional ambient air monitoring
11            site in Vinton, VA, provided a reasonable estimate of indoor and outdoor sulfate
12            measurements during the summer at homes  without tobacco combustion.
13          • Approximately 75% of the fine aerosol indoors during the summer was associated with
14            outdoor sources based on I/O sulfate ratios measured in the  Leaderer et al. (1999) study.
15          • Personal exposures to strong acidity (ET) were lower than corresponding outdoor levels
16            measured in studies by Brauer et al. (1989,1990) and Suh et al. (1992). But the personal
17            exposure levels measured by Suh et al. (1992) were higher than the indoor
18            microenvironmental levels.
19          • Personal exposures to NH4+, and NO3' were reported by Suh et al. (1992) to be lower than
20            either indoor or outdoor levels.
21          • Personal exposures to SO42" were also lower than corresponding outdoor levels, but
22            higher than the indoor microenvironmental levels (Suh et al., 1992; 1993a,b), as shown in
23            Table 5-13.
24          The fact that the personal and indoor H* concentrations were  substantially lower than
25      outdoor concentrations suggests that a large fraction of aerosol strong acidity is neutralized by
26      ammonia. Ammonia is emitted in relatively high concentrations in exhaled breath and sweat.
27      The difference between indoor and outdoor H+ concentrations in the Suh et al. (1992, 1993a,b)
28      studies was also much higher than the difference for indoor and outdoor SO42", indicative of
29      neutralization of the H+. Results of the Suh et al. (1992, 1993a,b) studies also showed substantial
30      interpersonal variability of H+ concentrations that could not be explained by variation in outdoor
31      concentrations.
        March 2001
5-74
DRAFT-DO NOT QUOTE OR CITE

-------
              TABLE 5-13.  SUMMARY STATISTICS FOR PERSONAL, INDOOR, AND
          OUTDOOR CONCENTRATIONS OF SELECTED AEROSOL COMPONENTS IN
                              TWO PENNSYLVANIA COMMUNITIES
Aerosol
State College
NO3-
SO42'
NH/
H+
Uniontown
SO42'
NH/
H+
Home Type

A/C Homes0
Non-A/C
A/C Homes
Non-A/C
All Homes'*
All Homes
A/C Homes
Non-A/C
All Homes0

All Homes'
All Homes0
All Homes0
Sample Site
(In/Out)a

53/71
254/71
56/75
259/75
214/76
314/155
28/74
230/74
163/75

91/46
91/44
91/46

Indoor (12 h)
GM ± GSDb

2.1 ±2.7
3.2 ±2.3
61.8 ±2.5
96.7 ± 2.5
69.1 ±2.6
154.7 ±2.8
4.2 ± 4.3
11.2±3.1
9.1 ±3.5

87.8 ±2.1
157.2 ±2.8
13.7 ±2.5
Concentration
(nmol m"3)
Outdoor (24 h)
GM ± GSDb

1.4 ±2.1
1.4±2.1
109.4 ± 2.4
109.4 ± 2.4
91.0 ±2.5
104.4 ±2.3
82.5 ±2.6
82.5 ± 2.6
72.4 ± 2.9

124.9 ±1.9
139.4 ±2.1
76 6 ± 2 7

Personal (12 h)
GM ± GSDb

— •
71. 5 ±2.4
—
18.4 ±3.0

110.3 ±1.8
167.0 ±2.0
42 8 ± 2 2
        "In/Out = Indoor sample site/outdoor sample site.
        bGM ± GSD = Geometric mean ± geometric standard deviation.
        °A/C Homes = Homes that had air-conditioning (A/C); this does not imply that it was on during the entire
        sampling period.
        Non-A/C = Homes without air conditioning.
        dThe sample size (n) for the personal monitoring = 209.
        °n = 174 for personal monitoring.

        Source:  Suhetal. (1992,1993a,b).
1

2

3

4

5

6
     Similar results for ammonia were reported by Waldman and Liang (1993). They reported
that levels of ammonia in institutional settings that they monitored were 10- to 50- times higher
than outdoors, and that acid aerosols were largely neutralized.  Leaderer et al. (1999) reported
that ammonia concentrations during both winter and summer in residences were an order of

magnitude higher indoors than outdoors, consistent with results of other studies and the presence
of sources of ammonia indoors.
      March 2001
                                         5-75
DRAFT-DO NOT QUOTE OR CITE

-------
1          Sulfate aerosols appear to penetrate indoors effectively. Waldman et al. (1990) reported
2     I/O ratios of 0.7 to 0.9 in two nursing care facilities and a day-care center. Sulfate I/O ratios were
3     measured for three particle size fractions in 12 residences in Birmingham, England, by Jones
4     et al. (2000). The sulfate I/O ratios were 0.7 to 0.9 for PM < 1.1 ywm, 0.6 to 0.8 for PM 1.1 to
5     2.1 /zm, and 0.7 to 0.8 for PM 2.1 to 10 //m. Suh et al. (1993b) reported that personal and
6     outdoor sulfate concentrations were highly correlated, as depicted in Figure 5-10.
              600
         Q.
                   0
100       200       300       400       500
    Outdoor Sulfate (nmoles/m3)
                         600
       Figure 5-10. Personal versus outdoor SO4= in State College, PA. Open circles represent
                   children living in air conditioned homes; the solid line is the 1:1 line.
       Source: Suh et al. (1993b).
       March 2001
                  5-76
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
     Indoor/outdoor relationships were measured for a number of PM25 components and related
species in Lindon, UT, during January and February of 1997 by Patterson and Eatough (2000).
Outdoor samples were collected at the Utah State Air Quality monitoring site.  Indoor samples
were collected in the adjacent Lindon Elementary School. The infiltration factors, Cai/Cao, given
by the slope of the regression lines (Table 5-14), were low (0.27 for sulfate and 0.12 for PM2 5),
possibly because of removal of particles in the air heating and ventilation system. The authors
concluded that the data indicate that indoor PM2 5 mass may not always be a good indicator of
exposure to ambient combustion material caused by the influence of indoor sources of particles.
However, ambient sulfate, SO2, nitrate, soot, and total particulate number displayed strong
correlations with indoor exposure. Ambient PM2 5 mass was not a good indicator of indoor PM2 5
mass exposure.
        TABLE 5-14. STATISTICAL CORRELATION OF OUTDOOR (x) VERSUS INDOOR
                        (y) CONCENTRATION FOR MEASURED SPECIES
                    (Units are nmol m'3, except for soot and metals, which are /
                               and absorption units m"3, respectively.)3
Species
SO2 All Samples
SO2 Day Samples
SO2 Night Samples
Sulfate All Samples
Sulfate Day Samples
Sulfate Night Samples
Nitrate All Samples
Nitrate Day Samples
Nitrate Night Samples
Soot Day Samples
Soot Night Samples
Total Acidity All Samples
Metals All Samples
Slope
0.0272 ± 0.0023
0.0233 ± 0.0037
0.0297 ± 0.0029
0.267 ±0.024
0.261 ± 0.034
0.282 ± 0.035
0.0639 ± 0.0096
0.097 ± 0.0096
0.047 ±0.0 11
0.43 ± 0.25
0.33 ±0.13
0.04 ± 0.73
0.10 ±0.30
Intercept
0.34 ±0.1 3
0.75 ± 0.26
0.099 ± 0.075
-0.14 ±0.48
0.40 ± 0.66
-0.84 ±0.68
0.9 ±1.5
-0.4 ± 1.4
1.5 ±1.8
3.5 ±1.7
0.00 ± 0.55
0.42 ± 0.23
0.0014 ± 0.0042
r2
0.73
0.62
0.82
0.70
0.71
0.70
0.54
0.88
0.44
0.43
0.69
0.00
0.01
Average
Outdoors
38
56
20
16
16
16
134
126
139
6
4
0.2
0.0042
        "Linden Elementary School, Lindon, UT, January and February 1997.
        Source: Patterson and Eatough (2000).
       March 2001
                                        5-77
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
     Oglesby et al. (2000) conducted a study to evaluate the validity of fixed-site fine particle
concentration measurements as exposure surrogates for air pollution epidemiology. Using 48-h
EXPOLIS data from Basel, Switzerland, they investigated the personal exposure/outdoor
concentration relationships for four indicator groups: (1) PM2 5 mass, (2) sulfur and potassium
for regional air pollution, (3) lead and bromine for traffic-related particles, and (4) calcium for
crustal particles. The authors reported that personal exposures to PM2 5 mass were not correlated
to corresponding home outdoor levels (n = 44, r = 0.07). In the study group reporting neither
relevant indoor sources nor relevant activities, personal exposures and home outdoor levels of
sulfur were highly correlated (n = 40, r = 0.85).  Oglesby et al. (2000) concluded that

     "for regional air pollution, fixed-site fine particle levels are valid exposure surrogates. For source-
     specific exposures, however, fixed-site data are probably not the optimal measure.  Still, in air pollution
     epidemiology, ambient PM2^ levels may be more appropriate exposure estimates than total personal
           exposure, since the latter reflects a mixture of indoor and outdoor sources."
     PAHs have been measured in studies by EPA and the California Air Resources Board.
PAH results from a probability sample of 125 homes in Riverside are discussed in reports by
Sheldon et al. (1992a,b) and Ozkaynak et al. (1996b). Data for two sequential 12-h samples were
reported for PAHs by ring size (3 to 7) and for individual phthalates. The results are summarized
below.
     • The particulate-phase 5- to 7-ring species had lower relative concentrations than the more
       volatile 3- to 4-ring species.
     • The 12-h indoor/outdoor ratios for the 5- to 7-ring species ranged from 1 . 1 to 1 .4 during
       the day and from 0.64 to 0.85 during the night (Sheldon et al., 1993a).
     • An indoor air model used to calculate indoor "source strengths" for the PAHs showed
       that smoking had the strongest effect  on indoor concentrations.
     Results from a larger PAH probability study in 280 homes in Placerville and Roseville
(Sheldon et al., 1993b,c) were similar to the 125-home study. The higher-ring, particle-bound
PAH's had lower indoor and outdoor concentrations than the lower-ring species.  For most
PAHs, the I/O ratio was greater than 1 for smoking and smoking/fireplace homes  and less than
1 for fireplace-only, wood stove, wood stove/gas heat, gas heat, and "no source" homes.
March 2001
                                                  5-78
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
28
29
30
31
      A study of PAHs in indoor and outdoor air was conducted in 14 inner-city and 10 rural
 low-income homes near Durham, NC, in two seasons (winter and summer) in 1995 (Chuang
 et al., 1999). Fine-particle-bound PAH concentrations measured with a real-time monitor were
 usually higher indoors than outdoors (2.47 ± 1.90 versus 0.53 ± 0.58 yug/m3). Higher indoor
 levels were seen in smoker's homes compared with nonsmoker's homes, and higher outdoor and
 indoor PAH levels were seen in urban areas compared with rural areas.
      In a study reported by Dubowsky et al. (1999), the weekday indoor PAH concentrations
 attributable to traffic (indoor source contributions were removed) were 39 ± 25 ng/m3 in a
 dormitory that had a high air exchange rate because of open windows and doors, 26 ± 25 ng/m3
 in an apartment, and 9 ± 6 ng/m3 in a suburban home. The study showed that both
 outdoor—especially motor vehicular traffic—and indoor sources contributed to indoor PAH
 concentrations.

 5.5.4 Factors Affecting Correlations Between Ambient Measurements and
       Personal or Microenvironmental Measurements of Particulate Matter
       Constituents
      The primary factors affecting correlations between personal exposure and ambient air PM
 measurements have been discussed in Section 4.3.2.  These include air-exchange rates, particle
 penetration factors, decay rates and removal mechanisms, indoor air chemistry, and indoor
 sources. The importance of these factors varies for different PM constituents. For acid aerosols,
 indoor air chemistry is particularly important as indicated by the discussion of the neutralization
 of the acidity by ammonia, which is present at higher concentrations indoors because of the
presence of indoor sources. For SVOCs, including PAHs and phthalates, the presence of indoor
sources will impact substantially the correlation between indoor and ambient concentrations
(Ozkaynak et al., 1996b). Penetration factors for PM will impact correlations between indoors
and outdoors for most elements, except Pb, which may have significant indoor sources in older
homes. Indoor air chemistry, decay rates, and removal mechanisms may affect soot and organic
carbon. These factors  must be fully evaluated when attempting to correlate ambient, personal,
and indoor PM concentrations.
       March 2001
                                        5-79
                                                           DRAFT-DO NOT QUOTE OR CITE

-------
 1      5.5.5 Limitations of Available Data
 2           The previous discussion demonstrates that there is very limited data available that can be
 3      used to compare personal, microenvironmental, and ambient air concentrations of PM
 4      constituents. Because of resource limitations, PM constituents have not been measured in many
 5      studies of PM exposure. Although there is some data on acid aerosols, the comparisons between
 6      the personal and indoor data generally have been with outdoor measurements at the participant's
 7      residences, not with community ambient air measurement sites. The relationship between
 8      personal exposure and indoor levels of acid aerosols is not clear because of the limited database.
 9      The exception is sulfate, for which there appears to be a strong correlation between indoor and
10      ambient concentrations.
11           With the exception of PAHs, there are practically no data available to relate personal or
12     indoor concentrations with outdoor or ambient site concentrations of SVOCs, which may be
13     generated from a variety of combustion and industrial sources. The relationship between
14     exposure and ambient concentrations of particles from specific sources, such as diesel engines,
15     has not been determined.
I g          Although there is an increasing amount of research being performed to measure PM
17     constituents in different PM size fractions, the current data are inadequate to adequately assess
18     the relationship between indoor and ambient concentrations of most PM constituents.
19
20
21     5.6. IMPLICATIONS OF USING AMBIENT PARTICULATE MATTER
22          CONCENTRATIONS IN EPIDEMIOLOGIC  STUDIES OF
23          PARTICULATE MATTER HEALTH EFFECTS
24          m this section, the exposure issues that relate to the interpretation of the findings from
25     epidemiologic studies of PM health effects are examined.  This section examines the  errors that
26     may be associated with using ambient PM concentrations in epidemiologic analyses of PM health
27     effects. Fkst, implications of associations found between personal exposure and ambient PM
 28     concentrations are reviewed. This is discussed separately in the context of either community
 29     time-series studies or long-term, cross-sectional studies of chronic effects. Next, the  role of
 30      compositional and spatial differences hi PM concentrations are discussed and how these may
 31      influence the interpretation of findings from PM epidemiology.  Finally, using statistical
        March 2001
5-80
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
methods, an evaluation of the influence of exposure measurement errors on PM epidemiology
studies is presented.

5.6.1  Potential Sources of Error Resulting from Using Ambient Particulate
       Matter Concentrations in Epidemiologic Analyses
     Measurement studies of personal exposures to PM are still few and limited in spatial,
temporal, and demographic coverage. Consequently, with the exception of a few longitudinal
panel studies, most epidemiologic studies of PM health effects rely on ambient community
monitoring data giving 24-h average PM concentration measurements.  Moreover, because of
limited sampling for PM2 5, many of these epidemiologic studies had to use available PM10or in
some instances had to rely on historic data on other PM measures or indicators, such as TSP,
SO4=, IP15, RSP, COH, KM, etc. A critical question often raised in the interpretation of results
from acute or chronic epidemiologic community-based studies of PM is whether the use of
ambient stationary site PM concentration data influences or biases the findings from these
studies. Because the  health outcomes are measured on individuals,  the epidemiologists might
prefer to use personal exposure measurements (total, ambient, or nonambient) instead of
surrogates, such as ambient PM concentration measurements collected at one or more ambient
monitoring sites in the community. Use of ambient concentrations could lead to
misclassification of individual exposures and to errors in the epidemiologic analysis of pollution
and health data depending on the pollutant and on the mobility and lifestyles of the population
studied. Ambient monitoring stations can be some distance away from the individuals and can
represent only a fraction of all likely outdoor microenvironments that individuals come in contact
with during the course of their daily lives. Furthermore,  most individuals are quite mobile and
move through multiple microenvironments (e.g., home, school, office, commuting, shopping,
etc.) and engage in diverse personal activities at home (e.g, cooking, gardening, cleaning,
smoking).  Some of these microenvironments and activities may have different sources of PM
and result in distinctly different concentrations of PM than that monitored by the fixed-site
ambient monitors.  Consequently, exposures of some individuals will be classified incorrectly if
only ambient monitoring data are used to estimate individual level exposures to PM. Thus, bias
or loss  of precision in the epidemiologic analysis may result from improper assessment of
exposures using data routinely collected by the neighborhood monitoring stations.
       March 2001
                                         5-81
DRAFT-DO NOT QUOTE OR CITE

-------
 1          Because individuals are exposed to particles in a multitude of indoor and outdoor
 2     microenvironments during the course of a day, concern over error introduced in the estimation of
 3     PM risk coefficients using ambient, as opposed to personal, PM measurements has received
 4     considerable attention recently from exposure analysts, epidemiologists, and biostatisticians.
 5     Some exposure analysts contend that, for community time-series epidemiology to yield
 6     information on the statistical association of a pollutant with a health response, there must be an
 7     association between personal exposure to a pollutant and the ambient concentration of that
 8     pollutant because people tend to spend around 90% time indoors and are exposed to both indoor
 9     and outdoor-generated PM (cf. Wallace, 2000b; Brown and Paxton, 1998; Ebelt et al., 2000).
10     Consequently, numerous findings reported in the epidemiologic literature on significant
11     associations between  ambient PM concentrations and various morbidity and mortality health
12     indices, in spite of the low correlations between ambient PM and concentrations and measures of
13     personal exposure, has been described by some exposure analysts as an exposure paradox
14     (Lachenmyer and Hidy, 2000, Wilson et al., 2000).
15          To resolve the so-called exposure paradox several types of analyses need to be considered.
16     The first type  of analysis has to examine the correlations between ambient PM concentrations
17     and personal exposures that are relevant to most of the existing PM epidemiology studies using
18     either pooled, daily-average, or longitudinal exposure data. The second approach has to  study the
19     degree of correlations between the two key components of personal PM exposures (i.e.,
20     exposures caused by ambient-generated PM and exposures caused by nonambient PM) with
21     ambient or outdoor PM concentrations, for each of the three types of exposure study design.
22     In addition, several factors influencing either the exposure or health response characterization of
23     the subjects have to be addressed. These include such factors as
24           »spatial variability of PM components,
25           • health  or sensitivity status of subjects,
26           • variations of PM with other co-pollutants,
27           • formal evaluation of exposure errors in the analysis of health data, and
28           • how the results may depend on the variations in the design of the epidemiologic study.
29           To facilitate the discussion of these topics, a brief review of concepts pertinent to exposure
30     analysis issues in epidemiology is presented.
31
        March 2001
5-82
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 5.6.2 Associations Between Personal Exposures and Ambient Particulate
       Matter Concentrations
      As defined earlier in Sections 5.3 and 5.4, personal exposures to PM result from an
 individual's exposures to PM in many different types of microenvironments (e.g., outdoors near
 home, outdoors away from home, indoors at home, indoors at office or school, commuting,
 restaurants, malls, other public places, etc.). Total personal exposures (Et) that occur in these
 indoor and outdoor microenvironments can be classified as those resulting from PM of outdoor
 origin (Eag)  and those primarily generated by indoor sources and personal activities (Enonag=
 Eig+Epact). The associations between personal exposures and ambient PM concentrations that
 have been reported from various personal exposure monitoring studies under three broad
 categories of study design: (1) longitudinal, (2) daily-average, or (3) pooled exposure studies are
 summarized below.
      In the previous Sections 5.4.3.1.2 and 5.4.3.1.3, some of the recent studies conducted
 primarily in the United States, involving children, elderly, and subjects with COPD were
 reviewed, and they indicated that both intra- and interindividual variability in the relationships
 between personal exposures and ambient PM concentrations were observed. A variety of
 different physical, chemical, and personal or behavioral factors were identified by the original
 investigators that seem to influence the magnitude and the strength of the associations reported.
      Clearly, for cohort studies in which individual daily health response are obtained,
 individual longitudinal PM personal exposure data (including ambient-generated and nonambient
 components) provide the appropriate indicators. In this case, health responses of each individual
 can be associated with the total personal exposure, the ambient-generated exposure, or the
 nonambient exposure of each individual. Also, the relationships of personal exposure indicators
 with ambient concentration can be investigated. In the case of community time-series
 epidemiology, however, it is not feasible to obtain experimental measurements of personal
 exposure for the millions of people over time periods of years that are needed to investigate the
relationship between air pollution and infrequent health responses such as deaths or even hospital
 admissions.  The epidemiologist must work with the aggregate number of health responses
occurring each day and a measure of the ambient concentration that is presumed to be
representative of the entire community. The relationship of PM exposures of the potentially
susceptible groups to monitored ambient PM concentrations depends on their activity pattern and
       March 2001
                                         5-83
DRAFT-DO NOT QUOTE OR CITE

-------
 1     level, residential building and HVAC factors (which influence the infiltration factor), status of
 2     exposure to ETS, amount of cooking or cleaning indoors, and seasonal factors, among others.
 3     Average personal exposures of these special subgroups to ambient-generated PM are correlated
 4     well with ambient PM concentrations regardless of individual variation in the absence of major
 5     microenvironmental sources.
 6           There seem to be clear differences in the relationships of ambient (Eag) and nonambient
 7     (Enonjg) exposure with ambient concentration (CJ. Various researchers have shown that Enonag is
 8     independent of Ca, but that Eag is a function of Ca. Wilson et al. (2000) explains the difference
 9     based on different temporal patterns that effect PM concentrations.  "Concentrations of ambient
10     PM are driven by meteorology and by changes in the emission rates and locations of emission
11     sources, while concentrations of nonambient PM are driven by the daily activities of people."
12           Ott et al. (2000) also discuss the reasons for assuming that Enonag is independent of Eag and
13     Ca. They show that the nonambient component of total personal exposure is uncorrelated with
14     the outdoor concentration data.  Ott et al. (2000) show the Enonag is similar for three population-
15     based exposure studies, including two large probability-based studies, the PTEAM study
16     conducted in Riverside (Clayton et al., 1993; Thomas et al., 1993; Ozkaynak et al., 1996a,b) and
17     a study in Toronto (Pelizzarri et al., 1999; Clayton et al., 1999a), as well as a nonprobability-
18     based study, conducted in Phillipsburg (Lioy et al., 1990).  Based on these three studies, they
19     conclude that Enonag and the distribution of(Enonag)it can be treated as constant from city to city.
20           Dominici et al. (2000) examined a larger database consisting of five different PM exposure
21     studies and concluded that Enonag  can be treated as relatively constant from city to city.
22     If (Enonag), were constant, this would imply that it would have a zero correlation with (Ca)t.
23     However, this hypothesis of constant (Enonag)it has not been established fully because only a few
24     studies have obtained the data needed to estimate (Enonag)it.  Although Enonag is independent of
25     Ca, it may not be independent of a. Sarnat et al. (2000) show that Enonag goes up as the
26     ventilation rate (and a) goes down. Lachenmeyer and Hidy (2000) also  show, by comparing
27     winter and summer regression equations, that as the slope (a) goes down, the intercept (Enonag)
28     goes up.
29           Mage et al. (1999) assume that the PM!0 concentration component from indoor sources,
30     such as smoking, cooking, cleaning, burning candles, and so on, is not correlated with the
        March 2001
5-84
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
outdoor concentration. They indicate that this lack of correlation is expected, because people are
unaware of ambient concentrations and do not necessarily change their smoking or cooking
activities as outdoor PM10 concentrations vary, an assumption supported by other empirical
analyses of personal exposure data.  For the PTEAM data set, Mage et al. (1999) have shown that
Eig and Ca have r near zero (R2 = 0.005). Wilson et al. (2000) have shown the Cai and Cig also
have r near zero (R2 = 0.03).  Figure 5-11 shows the relationship of estimated (Enonag)it and Enonag
with Ca (calculated by EPA from PTEAM and THEES data).
     Based on these results it is reasonable to assume that ordinarily Enonag has no relationship
with Ca. Therefore, in linear nonthreshold models of PM health effects, Enonag is not expected to
contribute to the relative risk determined hi a regression of health responses on Ca. Furthermore,
in time-series analysis of pooled or daily heath data, it is expected that Eag rather than E, will have
the stronger association with Ca.

5.6.3  Role of Compositional Differences in Exposure Characterization for
       Epidemiology
     The majority of the available data on PM exposures and relationships with ambient PM
have come from a few large-scale studies, such as PTEAM, or longitudinal studies on selected
populations, mostly the elderly.  Consequently, for most analyses, exposure scientists and
statisticians had to rely on PM10 or PM25 mass data, instead of elemental or chemical
compositional information on individual or microenvironmental samples. In a few cases,
researchers have examined the factors influencing indoor outdoor ratios or penetration and
deposition coefficients using elemental mass data on personal, indoor, and outdoor PM data (e.g.,
Ozkaynak et al. 1996a,b; Yakovleva et al. 1999). These results have been informative in terms
of understanding relative infiltration of different classes of particle sizes and sources into
residences (e.g., fossil fuel combustion, mobile source emissions, soil-derived, etc.).  Clearly, in
the accumulation-mode, particles associated with stationary or mobile combustion sources have
greater potential for penetration into homes and other microenvironments than do crustal
material. The chemical composition of even these broad categories of source classes may have
distinct composition and relative toxicity. Moreover, when particles and reactive gases are
present indoors in the presence of other pollutants or household chemicals, they may react to
form additional or different compounds and particles with yet unknown physical, chemical, and
        March 2001
                                         5-85
DRAFT-DO NOT QUOTE OR CITE

-------
                        200
                                      50
                             100
150
200
250
  «£   60'
gl
                   0)  O)
                   Q.S
                  izf
                    >
                          50-
                          40-
                   il   ^
                          20'
                                                                           (b)
                                        50           100          150
                                       Ambient Concentration, pg/m3
                                                           200
      Figure 5-11. Plots of nonambient exposure to PM10, (a) daytime individual values from
                   PTEAM data and (b) daily-average values from TREES data.
      Source: Data taken from (a) Clayton et al. (1993) and (b) Lioy et al. (1990).
1     toxic composition (Isukapalli et al. 2000).  Thus, if indoor-generated and outdoor-generated PM
2     were responsible for different types of health effects, or had significantly different toxicities on a
3     per unit mass basis, it would be then be important that Eag and Enonag should be separated and
4     treated as different species, much like the current separation of PM10 into PM25 and PM10.2 5.
5     These complexities in personal exposure profiles may introduce nonlinearities and other
      March 2001
                             5-86
  DRAFT-DO NOT QUOTE OR CITE

-------
 1      statistical challenges in the selection and fitting of concentration-response models.
 2      Unfortunately, PM health effects models have not yet been able to meaningfully consider such
 3      complexities. The relationships of toxicity to the chemical and physical properties of PM are
 4      discussed in Chapter 7.
 5           It is important also to note that individuals spend time in places other than their homes and
 6      outdoors. Many of the interpretations reported in the published literature on factors influencing
 7      personal PM]0 exposures, as well as in this chapter, come from the PTEAM study. The PTEAM
 8      study was conducted 10 years ago in one geographic location in California, during one season,
 9      and most residences had very high and relatively uniform air-exchange rates. Nonhome indoor
10      microenvironments were not monitored directly during the PTEAM study.  Commuting
11      exposures from traffic or exposures in a variety of different public places or office buildings
12      could not be assessed directly.  Nonresidential buildings may have lower or higher ambient
13      infiltration rates depending on the use and type of the mechanical ventilation systems employed.
14      Because the source and chemical composition of particulate matter effecting personal exposures
15      in different microenvironments vary by season, day-of-the-week, and time of day, it is likely that
16      some degree of misclassification of exposures to PM toxic agents of concern will be introduced
17      when health effects models use only daily-average mass measures such as PM10 or PM2 5.
18      Because of the paucity of currently available data on many of these factors, it is impossible to
19      ascertain at this point the magnitude and severity of these more complex exposure
20      missclassification problems in the interpretation of results from PM epidemiology.
21
22      5.6.4 Role of Spatial Variability in Exposure Characterization for
23            Epidemiology
24           Chapter 3 (Section 3.2.3) and Chapter 5 (Section 5.3) present information on the spatial
25      variability of PM mass and chemical components at fixed-site ambient monitors; for purposes  of
26      this chapter, this spatial variability is called an "ambient gradient." Any gradient that may exist
27      between a fixed-site monitor and the outdoor /we near where people live, work, and play,
28      obviously affects the concentration profile actually experienced by people as they go about their
29      daily lives.
30           However, the evidence so far indicates that PM concentrations, especially fine PM (mass
31      and sulfate), generally are distributed uniformly in most metropolitan areas. This reduces the
        March 2001
5-87
DRAFT-DO NOT QUOTE OR CITE

-------
 1     potential for exposure misclassification because of outdoor spatial gradients when a limited
 2     number of ambient PM monitors are used to represent population average ambient exposures in
 3     time-series or cross-sectional epidemiologic studies of PM. This topic is further discussed below
 4     in Section 5.6.5. However, as discussed earlier, the same assumption is not necessarily true for
 5     different components of PM, because source-specific and other spatially nommiform pollutant
 6     emissions could alter the spatial profile of individual PM components in a community.
 7     For example, particulate and gaseous pollutants emitted from motor vehicles tend to be higher
 8     near roadways and inside cars. Likewise, acidic and organic PM species may be location- and
 9     time-dependent. Furthermore, human activities are complex, and if outdoor PM constituent
10     concentration profiles are either spatially or temporally variable, it is likely that exposure
11     misclassification errors could be introduced in the analysis of PM air pollution and health data.
12
13     5.6.5 Analysis of Exposure Measurement Error Issues in Particulate Matter
14            Epidemiology
15           The effects of exposure misclassification on relative risk estimates of disease using
16     classical 2x2 contingency design (i.e., exposed/nonexposed versus diseased/nondiseased) have
17     been studied extensively in the epidemiologic literature. It has been shown that the magnitude of
18     the exposure-disease association (e.g., relative risk) because of either misclassification of
19     exposure or disease alone (i.e., nondifferential misclassification) biases the effect results toward
20     the null, and differential misclassification (i.e, different magnitudes of disease misclassification
21     in exposed and nonexposed populations) can bias the effect measure toward or away from the
22     null value relative to the true measure of association (Shy et al.,  1978; Gladen and Rogan, 1979;
23     Copeland et al., 1977; Ozkaynak et al.,  1986). However, the extension of these results from
24     contingency analysis design to multivariate (e.g., log-linear regression, Poissson, logit) models
25     typically used in recent PM epidemiology has been more complicated.  Recently, researchers
26     have developed a framework for analyzing measurement errors typically encountered in the
27     analysis of time-series mortality arid morbidity effects from exposures to ambient PM (cf. Zeger
28     et al., 2000; Dominici et al., 2000; Samet et al., 2000). Some analysis in the context of cross-
29     sectional epidemiology have also been conducted (e.g. Navidi et al., 1999).
30           The appropriateness of using ambient PM concentration as an exposure  metric in the
31     context of epidemiologic analysis of health effects associated with exposure to PM recently has
        March 2001
5-88
DRAFT-DO NOT QUOTE OR CITE

-------
 1     been examined by a number of investigators (cf. Zeger et al., 2000; Dominici et al., 2000; Navidi
 2     et al., 1999; Ozkaynak and Spengler, 1996). hi the following section, the error analysis model
 3     framework developed in Zeger et al. (2000) will be discussed in the context of time-series
 4     epidemiology. After which, issues and implications of exposure errors to findings from long-
 5     term/chronic or cross-sectional epidemiology will be discussed briefly.
 6
 7     5.6.5.1  Analysis of Exposure Measurement Errors in Time-Series Studies
 8          Zeger et al. (2000) provide a useful framework for analyzing exposure error in community
 9     time-series epidemiology.  This framework, coupled with results from recent exposure studies,
10     makes it possible to clarify some important questions regarding relationships among the three
11     aspects of personal exposure (1) total personal, (2) personal caused by ambient PM, and
12     (3) personal resulting from nonambient PM and ambient concentration.  Consider the regression
13     of a health response (i.e., mortality rate on day t, Yt, against the ambient concentration of PM on
14     day t, C,). In analyzing pollution-level data on mortality and air pollution, log-linear regressions
15     of the form:
16
17
18
19
20
21
22
23
24
25
26
27
                                                                                        (5-11)
are fit, where Yj is the expected mortality rate; s(t) is an arbitrary but smooth function of time,
introduced to control for the confounding of longer trends and seasonality; Ct, is the average of
multiple monitor measurements of ambient pollution measurement for day t; and u, are other
possible confounders such as temperature and dew point on the same or previous day. Each
coefficient, P, in Equation 5-11 gives the expected change in the health response, Y, because of a
unit change in its corresponding variable.
     However, instead of Equation 5-11, Zeger et al. (2000) suggest that the analyst would like
to know the corresponding relationship for personal exposure rather than ambient concentration,

                        Yt  - exp|>0) + EtPE + utfiu ].                        (5-12)

Zeger et al. (2000) do not differentiate among the three aspects of personal or community
exposure.  To understand the error in P caused by using ambient concentrations instead of
        March 2001
                                         5-89
DRAFT-DO NOT QUOTE OR CITE

-------
 1      personal exposure in the regression analysis, it is necessary to examine the relationship between
 2      Pc, based on a unit change in the ambient concentration, C, and PE, based on a unit change in one
 3      of the three aspects of personal exposure, E. In considering the consequences for Pc, as an
 4      estimate of PE, of having a measure of ambient pollution C,, rather than actual personal exposure
 5      Eit, it is convenient to express the desired pollution measurement, Eit, as Ct plus three error terms:
 6


 8
 9           Here Et represents the daily, community-average personal exposure. The first term,
10      (Eit — Et), is  the error resulting from having only aggregated or community-averaged exposure
11      rather than individual-level exposure data.  The second term, (Et - Ct), is the difference
12      between the average personal exposure and the true ambient pollutant level, and the third term,
13      (C*, -C,), represents the difference between the true and the measured ambient concentration.
14           In the evaluation of these error terms, two types of measurement error often are considered
15      in the context of epidemiology.  The classical error model assumes that measurement error,
16      (Q-Et), depends on ambient measurements [simply referred to as Ct here instead of (Ca)J. The
17      Berkson error model assumes that the measurement error is dependent on the true value or the
18      personal exposure (E,). The regression coefficient (Pc), estimated from the health effects model
19      in the Berkson error case, gives  an unbiased estimate of PE. In the classical error case, Pc is a
20      biased estimate of pE, and the degree of bias depends on the correlation between the
21      measurement error and Ct. The measurement error analysis of Zeger et al. (2000) includes three
22      components: (1) an individual's deviation from the risk-weighted average personal exposure;
23      (2) the difference between the average personal exposure and the true ambient level; and (3) the
24      difference between the measured and the true ambient levels, which include the spatial variation
25      of outdoor PM and instrument sampling error.  Zeger et al. (2000) conclude that the first and
26      third components are of the Berkson type and, therefore, are likely to have smaller effects on the
27      relative risk estimates for PM. However, the second component can be a source of substantial
28      bias if, for example, there are short-term associations of the contributions of indoor sources  with
29      ambient concentrations. However, recent analysis of PTEAM data (Mage et al., 2000) and
30      theoretical considerations (Ott et al., 2000) indicate that it is unlikely that nonambient exposures
        March 2001
5-90
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 will be correlated with the ambient concentration.  Therefore, this type of bias is unlikely.
 However, if the community average exposure to ambient PM is less than the ambient
 concentration, the risk regression coefficient, pc, will be biased low. According to Carrol (1995),
 Pc = cc PE, where Pc is the percentage increase in risk because of a unit increase in ambient
 concentration, and PE is the estimated percentage increase in risk because of a unit increase in the
 community-average personal exposure to ambient PM. Both Zeger et al. (2000) and Dominici
 et al. (2000) examine the nature of error with this second component.  Both of these analyses
 conclude that the error introduced because of measured differences between the average personal
 exposure and ambient levels can bias the regression coefficients.  In both cases they find the Pc is
 close to a PE.
     This framework analysis demonstrates the importance of the daily community-average
 exposure, Et, in community time-series epidemiology. It is Et, not the random, pooled values of
 Ej t, that need to have a statistically significant correlation with Ct for proper interpretation of
 community time-series epidemiology studies based on ambient  monitoring data, as discussed
 further in Wilson et al. (2000) and Mage et al. (1999).
     A critical assumption in the above analysis is that the risk varies linearly with C or E (i.e.,
 Pc and PE are constant). This assumption does not permit a threshold (a concentration below
 which there  is no effect).  It also includes the assumption that the appropriate metric for
 determination of a health response is the 24-h average PM mass concentration. Zeger et al.
 (2000) show that the likely consequence of using ambient concentrations instead of the risk-
 weighted average personal exposure measures is to underestimate the pollution effects.
 According to Zeger et al. (2000) the largest biases in inferences about the mortality-personal
 exposure relative risk will occur because of more complex errors between ambient concentration
 and daily-average personal exposure measures. It is important to note that both the Zeger et al.
 (2000) and the Dominici et al. (2000)  error analyses used personal PM10data from the PTEAM
 study data. However, effects of measurement error estimates may differ by particle size and
 composition. It is possible that PM25  , ultrafine particle measures, or another component of PM,
may better reflect personal exposures to PM of outdoor origin. Finally, the seasonal or temporal
variations in the measurement errors and correlations between different PM concentration
measures and co-pollutants (e.g. SO2, CO, NO2, O3) could influence the error analysis results
reported by the investigators cited above.
        March 2001
                                          5-91
DRAFT-DO NOT QUOTE OR CITE

-------
 1     5.6.5.2 Analysis of Exposure Measurement Errors in Long-Term Epidemiology Studies
 2          The Six Cities (Dockery et al., 1993) and ACS (Pope et al., 1995) studies have played an
 3     important role in assessing the health effects from long-term exposures to particulate pollution.
 4     Even though these studies often have been considered as chronic epidemiologic studies, it is not
 5     easy to differentiate the role of historic exposures from those of recent exposures on chronic
 6     disease mortality.  In the Six Cities study, fine particles and sulfates were measured at the
 7     community level, and the final analysis of the database used six city-wide average ambient
 8     concentration measurements. This limitation also applies to the ACS study but has less impact
 9     because of the larger number of cities considered in that study. In a HEI-sponsored reanalysis of
10     the Six Cities and the ACS data sets, Krewski et al. (2000) attempted to examine some  of the
11     exposure misclassification issues either analytically or through sensitivity analysis of the
12     aerometric and health data. The HEI reanalysis project also addressed exposure measurement
13     error issues related to the Six Cities study. For example, the inability to account for exposures
14     prior to the enrollment of the cohort, hampered accurate interpretation of the relative risk
15     estimates in terms of acute versus chronic causes. Although the results seem to suggest past
16     exposures are more strongly associated with mortality than recent exposures, the measurement
17     error for long-term averages could be higher, thus influencing these interpretations. For example,
18     Krewski et al. (2000), using the individual mobility data available for the Six Cities cohort,
19     analyzed the mover and nonmover groups separately. The relative risk of fine particle effects on
20     all-cause mortality was shown to be higher for the nonmover group than for the mover  group,
21     suggesting the possibility of higher exposure misclassification biases for the movers. The issue
22     of using selected ambient monitors in the epidemiologic analyses also was  investigated by the
23     ACS and Six Cities studies reanalysis team. Krewski et al.(2000) presented the sensitivity of
24     results to choices made in selecting stationary or mobile-source-oriented monitors. For the ACS
25      study, reanalysis of the sulfate data using only those monitors designated as residential  or urban,
26      and excluding sites designated as industrial, agricultural, or mobile did not change the risk
27      estimates appreciably. On the other hand, application of spatial analytic methods designed to
28      control confounding at larger geographic scales (i.e., between cities) caused changes in the
29      particle and sulfate risk coefficients. Spatial adjustment may account for differences in pollution
30     mix or PM composition, but many other cohort-dependent risk factors will vary across  regions  or
31      cities  in the United States. Therefore, it is difficult to interpret these findings solely in  terms of
        March 2001
5-92
DRAFT-DO NOT QUOTE OR CITE

-------
  1      spatial differences in pollution composition or relative PM toxicity until further research is
  2      concluded.
  3           Another study that has examined the influence of measurement errors in air pollution
  4      exposure and health effects assessments is the one reported by Navidi et al. (1999). This study
  5      developed techniques to incorporate exposure measurement errors encountered in long-term air
  6      pollution health effects studies and tested them on the data from the University of Southern
  7      California Children's Health Study conducted in 12 communities in California. These
  8      investigators developed separate error analysis models for direct (i.e., personal sampling) and
  9      indirect (i.e., microenvironmental) personal exposure assessment methods.  These models were
 10      generic to most air pollutants, but a specific application was performed using a simulated data set
 11      for studying ozone health effects on lung function decline in children. Because the assumptions
 12      made in their microenvironmental simulation modeling framework were similar to those made in
 13      estimating personal PM exposures, it is useful to consider the conclusions from Navidi et. al.
 14      (1999). According to Navidi et al. (1999), neither the microenvironmental nor the personal
 15      sampler method produces reliable estimates of the exposure-response slope (for O3) when
 16      measurement error is uncorrected. Because of nondifferential measurement error, the bias was
 17      toward zero under the assumptions made in Navidi et al. (1999) but could be away from zero if
 18      the measurement error was correlated with the health response.  A simulation analysis indicated
 19      that the standard error of the estimate of a health effect increases as the  errors in exposure
20      assessment increase (Navidi et al., 1999).  According to Navidi et al.  (1999), when a fraction of
21      the ambient level in a microenvironment is estimated with a standard error of 30%, the standard
22      error of the estimate is 50% higher than it would be if the true exposures were known.  It appears
23      that errors in estimating ambient PM indoor/ambient PM outdoor ratios have much more
24      influence on the accuracy of the microenvironmental approach than do errors in estimating time
25      spent in these microenvironments.
26
27      5.6.5.3 Conclusions from Analysis of Exposure Measurement Errors on Particulate Matter
28            Epidemiology
29           Personal exposures to PM are influenced by a number of factors and sources of PM located
30      in both indoor and outdoor microenvironments. However, PM resulting from ambient sources
31      does penetrate into indoor environments, such as residences, offices, public buildings, etc., in
       March 2001
5-93
DRAFT-DO NOT QUOTE OR CITE

-------
 1     which individuals spend a large portion of their daily lives.  The correlations between total
 2     personal exposures and ambient or outdoor PM concentrations can vary depending on the relative
 3     contributions of indoor PM sources to total personal exposures.  Panel studies of both adult and
 4     young subjects have shown that, in fact, individual correlations of personal exposures with
 5     ambient PM concentrations could vary person to person, and even day to day, depending on the
 6     specific activities of each person.  Separation of PM exposures into two components,
 7     ambient-generated PM and nonambient PM, would reduce uncertainties in the analysis and
 8     interpretation of PM health effects data. Nevertheless, because ambient-generated PM is an
 9     integral component of total personal exposures to PM, statistical analysis of cohort-average
10     exposures are strongly correlated with ambient PM concentrations when the size of the
11     underlying population studied is large. Using the PTEAM study data, analysis of exposure
12     measurement errors, in the context of tune-series epidemiology, also has shown that  errors or
13     uncertainties introduced by using surrogate exposure variables, such as ambient PM
14     concentrations, could lead to biases in the estimation of health risk coefficients. These then
15     would need to be corrected by suitable calibration of the PM health risk coefficients.
16     Correlations between the PM exposure variables and other covariates (e.g., gaseous
17     co-pollutants, weather variables, etc.) also could influence the degree of bias in the estimated PM
18     regression coefficients. However, most time-series regression models employ seasonal or
19     temporal detrending of the variables, thus reducing the magnitude of this cross-correlation
20     problem (Ozkaynak and Spengler 1996).
21           Ordinarily, exposure measurement errors are not expected to influence the interpretation of
22     findings from either the cross-sectional or time-series epidemiologic studies that have used
23     ambient concentration data if they include sufficient adjustments for seasonality and key
24     confounders. Clearly, there is no question that better estimates of exposures to components of
25     PM of health concern are beneficial.  Composition of PM may vary in different geographic
26     locations and different exposure microenvironments. Compositional and spatial variations could
27     lead to further errors in using ambient PM measures as surrogates for exposures to PM. Even
28     though the spatial variability of PM (PM2 5 in particular) mass concentrations in urban
29     environments seems to be small, the same conclusions drawn above regarding the influence of
30     measurement errors may not necessarily hold for all of the PM toxic components.  Again, the
31     expectation based on statistical modeling  considerations is that these exposure measurement
        March 2001
5-94
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
 errors or uncertainties will most likely reduce the statistical power of the PM health effects
 analysis, making it difficult to detect a true underlying association between the correct exposure
 metric and the health outcome studied. However, until more data on exposures to toxic agents of
 PM become available, existing studies on PM exposure measurement errors must be relied on;
 these indicate that use of ambient PM concentrations as a surrogate for exposures is not expected
 to change the principal conclusions from PM epidemiologic studies, utilizing community average
 health and pollution data.
5.7  SUMMARY OF KEY FINDINGS AND LIMITATIONS
Exposure Definitions and Components
• Personal exposure (E) to PM mass or its constituents results when individuals come in contact
  with particulate pollutant concentrations (C) in locations or microenvironments (jj,e) that they
  frequent during a specific period of time.  Various PM exposure metrics can be defined
  according to its source (i.e., ambient, nonambient) and the microenvironment where exposure
  occurs.
• Personal exposure to PM results from an individual's exposure to PM in many different types
  of microenvironments (e.g., outdoors near home, outdoors away from home, indoors at home,
  indoors at office or school,  commuting, restaurants, malls, other public places,  etc.). Thus, total
  daily exposure to PM for a  single individual (E,) can be expressed as the sum of various
  microenvironmental exposures that the person encounters during the course of a day.
• In a given fte, particles may originate from a wide variety of sources. In an indoor
  microenvironment, PM may be generated  from within as a result of PM generating activities
  (e.g., cooking, cleaning, smoking, resuspending PM from PM resulting from both indoor and
  outdoor sources that had settled out), from outside (outdoor PM entering through cracks and
  openings in the structure), and from the chemical interaction of pollutants from outdoor air with
  indoor-generated pollutants.
• The total daily exposure to  PM for a single individual (Et) also can be expressed as the sum of
  contributions of ambient-generated (Eag) and nonambient-generated (Enonag) PM (i.e.,
  E = Eag + Enonag). Enonag, in turn, is composed of PM generated by indoor sources (Eig ) and PM
  generated by personal activities  (Epact) (i.e., Enonag = Eig + Epact). Eag is composed of exposures to
March 2001                              5-95         DRAFT-DO NOT QUOTE OR CITE

-------
 1       ambient PM concentrations while outdoors,   Ca Afa, and ambient PM that has infiltrated
                                                 /
 2       indoors, JJ CflI.Af, while indoors (i.e., Eag = JJ CaAffl + £ q,,Af, ).
                  »                              /           /
 3     * Exposure models are useful tools for examining the importance of sources, microenvironments,
 4       and physical and behavioral factors that influence personal exposures to PM. However,
 5       development and evaluation of population exposure models for PM and its components has
 6       been limited. Improved modeling methodologies and new model input data are needed.
 7
 8     Factors Affecting Concentrations and Exposures to Participate Matter
 9     • Concentrations of PM indoors are affected by several factors and mechanisms: ambient
10       concentrations outdoors; air exchange rates; particle penetration factors; particle production
11       from indoor sources and indoor air chemistry; and indoor particle decay rates and removal
12       mechanisms caused by physical processes or resulting from mechanical filtration, ventilation or
13       air-conditioning devices.
14     • Average personal exposures to PM mass and its constituents are influenced by
15       microenvironmental PM concentrations and by how much tune is spent by each individual in
16       these various indoor and outdoor microenvironments. Nationwide,  individuals, on average,
17       spend nearly 90% of their time indoors (at home and in other indoor locations) and about 6% of
18       their tune outdoors.
19     • The relative size of personal exposure to ambient-generated PM relative to nonambient-
20       generated PM depends on the ambient concentration, the infiltration rate of outdoor PM into
21       indoor microenvironments, the amount of PM generated indoors (e.g., ETS, cooking and
22       cleaning emissions), and the amount of PM generated by personal activity sources. Infiltration
23       rates primarily depend on air-exchange rate, size-dependent particle penetration across the
24       building membrane, and size-dependent removal rates. All of these factors vary over time and
25       across subjects and building types.
26     • The relationship  between PM exposure and health outcome could depend on the concentration,
27       composition, and toxicity of the PM originating from different sources. Application of source
28       apportionment techniques to ambient, indoor, and personal PM composition data have
29       identified the following general source categories of importance:  outside soil, resuspended
30       indoor soil, indoor soil, personal activities, sea-salt, motor vehicles, nonferrous metal smelters,
31       and secondary sulfates.
       March 2001                              5-96        DRAFT-DO NOT QUOTE OR CITE

-------
  1      • There have been only a limited number of studies that have measured the physical and
  2        chemical constituents of PM in personal or microenvironmental samples.  Available data on
  3        PM constituents indicate that
  4             - personal and indoor sulfate measurements often are correlated highly with outdoor and
  5              ambient sulfate concentration measurements;
  6             - for acid aerosols, indoor air chemistry is particularly important because of the
  7              neutralization of the acidity by ammonia, which is present at higher concentrations
  8              indoors because of the presence of indoor sources of ammonia;
  9             - for S VOCs, including PAHs and phthalates, the presence of indoor sources will
10              substantially impact the relation betwebn indoor and ambient concentrations;
11             - penetration and decay rates are a functions of size and will cause variations in the
12              attenuation factors as a function of particle size; infiltration rates will be higher for PM,
13              and PM2 5 than for PM10, PM10.2 5 or ultrafine particles; and
14             - Indoor air chemistry may increase indoor concentrations of organic PM.
15      • Even though there is an increasing amount of research being performed to measure PM
16       constituents in different PM size fractions, with few exceptions (i.e., sulfur or sulfates), the
17       current data are inadequate to adequately assess the relationship between personal, indoor, and
18       ambient concentrations of most PM constituents.
19
20      Correlations Between Personal Exposures, Indoor, Outdoor, and Ambient Measurements
21      • Most of the available personal data on PM measurements and information on the relationships
22       between personal and ambient PM come from a few large-scale studies, such as the PTEAM
23       study, or the longitudinal panel studies, which have been conducted on selected populations,
24       such as the elderly.
25      • Panel and cohort studies that have measured PM exposures and concentrations typically have
26       reported their results in terms of three types of correlations:  (1) longitudinal, (2) pooled, and
27       (3) daily-average correlations between personal and ambient or outdoor PM.
28      • The type of correlation analysis performed can have a substantial effect on the resulting
29       correlation coefficient.  Low correlations with ambient concentrations could result when people
30       with very different nonambient exposures are pooled, even though temporally, their individual
31       personal exposures may be correlated highly with ambient concentrations.
        March 2001
5-97
DRAFT-DO NOT QUOTE OR CITE

-------
 1      • Recent studies conducted by EPA of the elderly subjects living in a retirement facility in
 2       Baltimore and a group of elderly living in Fresno produced higher correlation coefficients
 3       between personal and ambient PM for daily-average correlations compared to longitudinal
 4       correlations. This supports earlier analyses showing the daily-average correlations are higher
 5       than pooled correlations.
 6      • Longitudinal and pooled correlations between personal exposure and ambient or outdoor PM
 7       concentrations reported by various investigators varied considerably among the different
 8       studies and in each study between the study subjects. Most studies report longitudinal
 9       correlation coefficients that range from close to zero to near one, indicating that individual's
10       activities and residence type may have a significant effect on total personal exposures to PM.
11      • Longitudinal studies that measured sulfate found high correlations between personal and
12       ambient sulfate.
13      • In general, probability-based population studies tend to show low pooled correlations because
14       of the high differences in levels of nonambient PM generating activities from one subject to
15       another. In contrast, the absence of indoor sources for the populations in several of the
16       longitudinal panel studies resulted in high correlations between personal exposure and ambient
17       PM within subjects over time for these populations. But even for these studies, correlations
18       varied by individual depending on their  activities and on the microenvironments that they
19       occupied.
20
21      Potential Sources of Error Resulting from Using Ambient Particulate Matter
22      Concentrations in Epidemiologic Analyses
23      • There is, as yet, no clear consensus among exposure analysts as to how well ambiently
24       measured PM concentrations represent a surrogate for personal exposure to total PM or to
25       ambient-generated PM.
26      • Measurement studies of personal exposures to PM are still few and limited in spatial, temporal,
27       and demographic coverage. Consequently, with the exception of a few longitudinal panel
28       studies, most epidemiologic studies on PM health effects have relied on daily-average PM
29       concentration measurements obtained from ambient community monitoring data as a surrogate
30       for the exposure variable.
        March 2001
5-98
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
  • Because individuals are exposed to particles in a multitude of indoor and outdoor
   microenvironments during the course of a day, concerns over error introduced in the estimation
   of PM risk coefficients using ambient, as opposed to personal PM measurements, have been
   raised.
  • Total personal exposures to PM could vary from person to person, and even day to day,
   depending on the specific activities of each person. Separation of PM exposures into two
   components, ambient-generated PM and nonambient-generated PM, would reduce potential
   uncertainties in the analysis and interpretation of PM health effects data.
 • Available data indicate that PM mass concentrations, especially fine PM, typically are
  distributed uniformly in most metropolitan areas, thus reducing the potential for exposure
  ^classification because of spatial variability when a limited number of ambient PM monitors
  are used to represent population average ambient exposures in community time-series or
  long-term, cross-sectional epidemiologic studies of PM.
 • Even though the spatial variability of PM (in particular, PM2.5) mass concentrations in urban
  environments seems to be small, the same conclusions drawn above regarding the influence of
  measurement errors may not necessarily hold for all of the PM components.
 • There are important differences in the relationship of ambient PM concentrations (CJ with
  exposures to ambient PM (Eag), and with exposures to nonambient PM (Enonag).  Various
  researchers have shown that Eag is a function of Ca, and that concentrations of ambient PM are
  driven by meteorology, by changes in source emission rates, and in locations of emission
  sources relative to the measurement site. However, Enonag is independent of Ca, because
  concentrations of nonambient PM are driven by the daily activities of people.
• Because personal exposures also include a contribution from ambient concentrations, the
  correlation between daily-average personal exposure and the daily-average ambient
  concentration increases as the number of subjects measured daily increases. An application of
  a Random Component Superposition (RCS) model has shown that the contributions of ambient
 PM10 and indoor-generated PM10 to community mean exposure can be decoupled in modeling
 urban population exposure distributions.
• If linear nonthreshold models are assumed in time-series analysis of daily-average ambient PM
 concentrations and community health data, Enonag is not expected to contribute to the relative
 risk estimates determined by regression of health responses on Ca.
       March 2001
                                                 5-99        DRAFT-DO NOT QUOTE OR CITE

-------
1      • Using the PTEAM study data, analysis of exposure measurement errors in the context of
2       time-series epidemiology has shown that errors or uncertainties introduced by using surrogate
3       exposure variables, such as ambient PM concentrations, could lead to biases in the estimation
4       of health risk coefficients.
5      - Because sources and chemical composition of particulate matter affecting personal exposures in
6       different microenvironments vary, by season, day-of-the-week, and time of day, it is likely that
7       some degree of misclassification of exposures to PM toxic agents of concern will be introduced
8       when health effects models use only daily-average  mass measures such as PM10 or PM2 5.
9       Because of the paucity of currently available data on many of these factors, it is impossible to
10      ascertain at this point the significance of these more complex exposure misclassification
11       problems in the interpretation of results from PM epidemiology.
12     - Exposure measurement errors may depend on particle size and composition. PM2.5 better
13       reflects personal exposure to PM of outdoor origin than PM10. It is possible that various
14       ultrafme particle measures, or other components of PM may be better exposure indicators for
15       epidemiologic studies.
16      • Seasonal or temporal variations in the measurement errors and their correlations between
17       different PM concentration measures and co-pollutants (e.g., SO2, CO, NO2, O3) could
18       influence the error analysis results but not likely the interpretation of current findings.
19      • Ordinarily, PM exposure measurement errors are not expected to influence the interpretation of
20       findings from either the community time-series or long-term epidemiologic studies that have
21       used ambient concentration data if they include sufficient adjustments for seasonality and key
22       personal and geographic confounders.
23      • To reduce exposure misclassification errors hi PM epidemiology, conducting new cohort
24      studies of sensitive populations with better real-time techniques for exposure monitoring and
25       further speciation of indoor-generated, ambient, and personal PM mass are essential.
26     • Based on statistical modeling considerations, it is expected that existing PM exposure
27      measurement errors or uncertainties  most likely will reduce the statistical power of the PM
28      health effects analysis, thus making  it difficult to detect a true underlying association between
 29       the correct exposure metric and the health outcome studied.
         March 2001
                                                  5-100
DRAFT-DO NOT QUOTE OR CITE

-------
1
2
3
4
Currently available studies on PM exposure measurement errors indicate that use of ambient
PM concentrations as a surrogate for personal exposures is not expected to change the key
conclusions derived from most of the recent epidemiologic studies on PM health effects.
      March 2001
                                      5-101
                                                          DRAFT-DO NOT QUOTE OR CITE

-------
 1
REFERENCES
 2       Abt, E.; Suh, H. H.; Allen, G.; Koutrakis, P. (2000a) Characterization of indoor particle sources: a study conducted
 3             in the metropolitan Boston area. Environ. Health Perspect. 108: 35-44.
 4       Abt, E.; Suh, H. H.; Catalano, P.; Koutrakis, P. (2000b) The relative contribution of outdoor and indoor particle
 5             sources to indoor concentrations. Environ. Sci. Technol. 34: 3579-3587.
 6       Akland, G. G.; Hartwell, T. D.; Johnson, T. R.; Whitmore, R. W. (1985) Measuring human exposure to carbon
 7             monoxide in Washington, D.C., and Denver, Colorado, during the winter of 1982-1983. Environ. Sci.
 8             Technol. 19:911-918.
 9       Al-Raheem, M. Y. A.; El-Desouky, M.; Mage, D. T.; Kollander, M.; Wallace, L. A.; Spengler, J. D. (2000)
10             Personal exposures to PM and gaseous air pollutants at the end of the 1991 Kuwait oil fires. Presented at:
11             PM2000: particulate matter and health—the scientific basis for regulatory decision-making, specialty
12             conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management Association.
13       Aim, S.; Jantunen, M. J.; Vartiainen, M. (1999) Urban commuter exposure to particle matter and carbon monoxide
14             inside an automobile. J. Exposure Anal. Environ. Epidemiol. 9: 237-244.
15       Anuszewski, J.; Larson, T. V.; Koenig, J. Q. (1998) Simultaneous indoor and outdoor particle light-scattering
16             measurements at nine homes using a portable nephelometer. J. Exposure Anal. Environ. Epidemiol.
17             8:483-493.
18       Back, S.-O.; Kim, Y.-S.; Perry, R. (1997) Indoor air quality in homes, offices and restaurants in Korean urban
19             areas-indoor/outdoor relationships. Atmos. Environ. 31: 529-544.
20       Bahadori, T.; Suh, H. H.; Rojas-Bracho, L.; Koutrakis, P. (2001) Personal exposure to particulate matter of
21             individuals with chronic obstructive pulmonary disease (COPD). J. Expos. Anal. Environ. Epidemiol.:
22             Accepted.
23       Beyea, J.; Hatch, M. (1999) Geographic exposure modeling:  a valuable extension of geographic  information
24             systems for use in environmental epidemiology. Environ. Health Perspect. Suppl. 107(1):  181-190.
25       Bohadana, A. B.; Massin, N.; Wild, P.; Toamain, J.-P.; Engel, S.; Goutet, P. (2000) Symptoms, airway
26             responsiveness, and exposure to dust in beech and oak wood workers. Occup. Environ. Med. 57: 268-273.
27       Brauer, M.; 't Mannetje, A. (1998) Restaurant smoking restrictions and environmental tobacco smoke exposure.
28             Am. J. Public Health 88:1834-1836.
29       Brauer, M.; Koutrakis, P.; Spengler, J. D. (1989) Personal exposures to acidic aerosols and gases. Environ. Sci.
30             Technol. 23:1408-1412.
31       Brauer, M.; Koutrakis, P.; Keeler, G. J.; Spengler, J. D. (1990) Indoor and outdoor concentrations of acidic aerosols
32             and gases. In: Indoor air '90: precedings of the 5th international conference on indoor air quality and climate,
33             volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada:
34             International Conference on Indoor Air Quality and Climate, Inc.; pp. 447-452.
35       Brauer, M.; Bartlett, K.; Regalado-Pineda, J.; Perez-Padilla, R. (1996) Assessment of particulate concentrations
36             from domestic biomass combustion in rural Mexico. Environ. Sci. Technol. 30: 104-109.
37       Brauer, M.; Hirtle, R. D.; Hall, A. C.; Yip, T. R. (1999) Monitoring personal fine particle exposure with a particle
38             counter. J. Exposure Anal. Environ. Epidemiol. 9:228-236.
39       Brown, K. G.; Paxton, M. B. (1998) Predictability of personal exposure to airborne particulate matter from ambient
40             concentrations in four communities. In: Vostal, J. J., ed. Health effects of particulate matter in ambient air.
41             Proceedings of an international conference; 1997; Prague, Czech Republic. Pittsburgh, PA: Air & Waste
42             Management Association; pp. 333-350. (A&WMA publication VIP-80).
43       Burke, J. M.; Zufall, M. J.; Ozkaynak, H. A.; Xue, J.; Zidek, J. V. (2000) Predicting population exposures to PM:
44             the importance of microenvironmental concentrations and human activities. Presented at:  PM2000:
45             particulate matter and health~the scientific basis for regulatory decision-making, specialty conference &
46             exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management Association.
47       Burke, J.; Zufall, M.; Ozkaynak, H. (2001) A population exposure model for particulate matter: case study results
48             for PM2.5 in Philadelphia, PA. J. Expos. Anal. Environ. Epidemiol.: submitted.
49       Burton, R. M.; Suh, H. H.; Koutrakis, P. (1996) Spatial variation in particulate concentrations within metropolitan
50             Philadelphia. Environ. Sci. Technol. 30: 400-407.
51       Buzorius, G.; Hameri, K.; Pekkanen, J.; Kulmala, M. (1999) Spatial variation of aerosol number concentration in
52             Helsinki city. Atmos. Environ. 33: 553-565.
53       Byrne, M. A.; Lange, C.; Goddard, A. J. H.; Roed, J. (1992) Indoor aerosol deposition studies using
54             neutron-activatable tracers. J. Aerosol Sci. 23(suppl. 1): S543-S546.
         March 2001
                                                 5-102
DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
45
46
47
48
49
 50
 51
52
53
 Calder, K. L. (1957) A numerical analysis of the protection afforded by buildings against B W aerosol attack.
       Fort Detrick, MD: Office of the Deputy Commander for Scientific Activities; BWL tech. study no. 2.
 Carrer, P.; Alcini, D.; Cavallo, D.; Bollini, D.; Ghalandar, R.; Lovato, L.; Ploia, E.; Vercelli, F.; Visigalli, F.;
       Maroni, M. (1998) Personal exposure to PM10 of subjects living in Milan. In: Vostal, J. J., ed. Health effects
       of paniculate matter in ambient air. Proceedings of an international conference; 1997; Prague, Czech
       Republic. Pittsburgh, PA: Air & Waste Management Association; pp. 359-365. (A&WMA publication •
       VIP-80).
 Carroll, R. J.; Ruppert, D.; Stefanski, L. A. (1995) Measurement error in nonlinear models. London, United
       Kingdom: Chapman & Hall. (Cox, D. R.; Hinkley, D. V.; Keiding, N.; Reid, N.; Rubin, D. B.; Silverman,
       B. W., eds. Monographs on statistics and applied probability: v. 63).
 Cha, Q.; Chen, Y.-Z.; Du, Y.-X. (1996) Influence of indoor air pollution on quality of outdoor air in city of
       Guangzhou, China. In: Indoor air '96: the 7th  international conference on indoor air quality, volume I; July;
       Nagoya, Japan. Tokyo, Japan: Institute of Public Health; pp. 1131 -1135.
 Chang, L.-T.; Koutrakis, P.; Catalano, P. J.; Suh, H.  H. (2000) Hourly personal exposures to fine particles and
       gaseous pollutants - results from Baltimore, Maryland.  J. Air Waste Manage. Assoc. 50:  1223-1.235.
 Chuang, J. C.; Callahan, P. J.; Lyu, C. W.; Wilson, N. K. (1999) Polycyclic aromatic hydrocarbon exposures of
       children in low-income families. J. Exposure Anal. Environ. Epidemiol. 2: 85-98.
 Clayton, C. A.; Perritt, R. L.; Pellizzari, E. D.; Thomas, K. W.; Whitmore, R. W.; Wallace, L. A.; Ozkaynak, H.;
       Spengler, J. D. (1993) Particle total exposure  assessment methodology (PTEAM) study: distributions of
       aerosol  and elemental concentrations in personal, indoor, and outdoor air samples in a southern California
       community. J. Exposure Anal. Environ. Epidemiol.  3: 227-250.
 Clayton, C. A.; Pellizzari, E. D.; Rodes, C. E.; Mason, R. E.; Piper, L. L. (1999a) Estimating distributions of
       long-term particulate matter and manganese exposures for residents of Toronto, Canada. Atmos Environ
       33:2515-2526.
 Clayton, C. A.; Pellizzari, E. D.; Whitmore, R. W.; Perritt, R. L., Quackenboss, J. J. (1999b) National human
       exposure assessment survey (NHEXAS): distributions and associations of lead, arsenic, and volatile organic
       compounds in EPA Region 5. J. Exposure Anal. Environ. Epidemiol. 9: 381-392.
 Cohen, B. S.; Li, W.; Xiong, J. Q.; Lippmann, M. (2000) Detecting H+ in ultrafme ambient aerosol using iron
       nano-film detectors and scanning probe microscopy. Appl. Occup. Environ. Hyg. 15: 80-89.
 Conner, T. L.;  Lonneman, W. A.; Seila, R. L. (1995) Transportation-related volatile hydrocarbon source profiles
       measured in Atlanta. J. Air Waste Manage. Assoc. 45: 383-394.
 Copeland, K. T.; Checkoway, H.; McMichael, A. J.;  Holbrook, R. H. (1977) Bias due to misclassification in the
       estimation of relative risk. Am. J. Epidemiol. 105: 488-495.
 Crump, K. S. (2000) Manganese exposures in Toronto during use of the gasoline additive, methylcyclopentadienyl
       manganese tricarbonyl. J. Exposure Anal. Environ. Epidemiol.  10: 227-239.
 Daisey, J. M.; Mahanama, K. R. R.; Hodgson, A. T. (1998) Toxic volatile organic compounds in simulated
       enviromental tobacco smoke: emission factors for exposure assessment. J. Exposure Anal. Environ
       Epidemiol. 8: 313-334.
 Dockery, D. W.; Spengler, J. D. (1981) Indoor-outdoor relationships of respirable sulfates and particles.
       Atmps. Environ. 15: 335-343.
 Dockery, D. W.; Pope, C. A., Ill; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G., Jr.; Speizer, F. E.
       (1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J Med
       329: 1753-1759.
 Dominici, F.; Zeger, S. L.; Samet, J. (2000) A measurement error model for time-series studies of air pollution and
      mortality. Biostatistics 1: 157-175.
 Donham, K. J.; Cumro, D.; Reynolds, S. J.; Merchant, J. A. (2000) Dose-response relationships between
      occupational aerosol exposures and cross-shift declines of lung function in poultry workers:
      recommendations for exposure limits. J. Occup. Environ. Med. 42: 260-269.
Duan, N. (1982) Models for human exposure to air pollution. Environ. Int. 8: 305-309.
Duan, N. (1991) Stochastic microenvironment models for air pollution exposure. J. Exp. Anal. Environ. Epidemiol
       1:235-257.
Dubowsky, S. D.; Wallace, L. A.; Buckley, T. J. (1999) The contribution of traffic to indoor concentrations of
      polycyclic aromatic hydrocarbons. J. Exposure Anal. Environ. Epidemiol.  9: 312-321.
         March 2001
                                                5-103
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Ebelt, S. T.; Petkau, A. J.; Vedal, S.; Fisher, T. V.; Brauer, M. (2000) Exposure of chronic obstructive pulmonary
 2            disease patients to particulate matter: relationships between personal and ambient air concentrations. J. Air
 3            Waste Manage. Assoc. 50:1081-1094.
 4       Esmen, N. A.; Hall, T. A. (2000) Theoretical investigation of the interrelationship between stationary and personal
 5            sampling in exposure estimation. Appl. Occup. Environ. Hyg. 15: 114-119.
 6       Evans, G. F.; Highsmith, R. V.; Sheldon, L. S.; Suggs, J. C.; Williams, R. W.; Zweidinger, R. B.; Creason, J. P.;
 7            Walsh, D.; Rodes, C. E.; Lawless, P. A. (2000) The 1999 Fresno particulate matter exposure studies:
 8            comparison of community, outdoor, and residential PM mass measurements. J. Air Waste Manage. Assoc.
 9            50: 1887-1896.
10       Ferris, B. G., Jr.; Speizer, F. E.; Spengler, J. D.; Dockery, D.; Bishop, Y. M. M.; Wolfson, M.; Humble, C. (1979)
11            Effects of sulfur oxides and respirable particles on human health: methodology and demography of
12            populations in study. Am. Rev. Respir. Dis. 120: 767-779.
13       Fogh, C. L.; Byrne, M. A.; Roed, J.; Goddard, A. J. H. (1997) Size specific indoor aerosol deposition measurements
14            and derived I/O concentrations ratios. Atmos. Environ. 31: 2193-2203.
15       Freijer, J. I.; Bloemen, H. J. T. (2000) Modeling relationships between indoor and outdoor air quality. J. Air Waste
16            Manage. Assoc. 50:292-300.
17       Frey, H. C.; Rhodes, D. S.  (1996) Characterizing, simulating and analyzing variability and uncertainty:
18            an illustration of methods using an air toxic emissions example. Hum. Ecol. Risk Assess. 2: 762-797.
19       Gladen, B.; Rogan, W. J. (1979) Misclassification and the design of environmental studies. Am. J. Epidemiol.
20             109:607-616.
21       Glen, G.; Lakkadi, Y.; Tippett, J. A.;  del Valle-Torres, M. (1997) Development of NERL/CHAD: the National
22            Exposure Research Laboratory consolidated human activity database. Research Triangle Park, NC: U.S.
23            Environmental Protection Agency, Office of Research and Development; contract no. 68-D5-0049.
24       Goldstein, B. R.; Tardiff, G.; Hoffnagle, G.; Kester, R. (1992) Valdez health study: summary report. Ankorage, AK:
25            Alyeska Pipeline Service Company.
26       Hadwen, G. E.; McCarthy, J. F.; Womble, S. E.; Girman, J. R.; Brightman, H.  S. (1997) Volatile organic compound
27            concentrations in 41 office buildings in the continental United States. In: Woods, J. E.; Grimsrud, D. T.;
28            Boschi, N., eds. Proceedings of healthy buildings/indoor air quality international conference; September;
29            Bethesda, MD. Washington, DC: Healthy Buildings/IAQ. [Healthy Buildings - International Conference;
30            v. 2, pp. 465-470.]
31       Hartwell, T. D.; Clayton, C. A.; Michie, R. M., Jr.; Whitmore, R. W.; Zelon, H. S.; Whitehurst, D. A.; Akland,
32            G. G. (1984) Study of carbon monoxide exposures of residents of Washington, D.C. Research Triangle Park,
33            NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report no.
34             EPA-600/D-84-177. Available from: NTIS, Springfield, VA; PB84-230051.
35       Hayes,  S. R.; Marshall, J. R. (1999) Designing optimal strategies to attain the new U.S. particulate matter standards:
36             some initial concepts. J. Air Waste Manage. Assoc. 49: PM192-PM198.
37       Henry, R. C.; Lewis, C. W.; Collins, J. F. (1994) Vehicle-related source compositions from ambient data:
38             the GRACE/SAFER method. Environ. Sci. Technol. 28: 823-832.
39       Houseman, E. A.; Ryan, L.; Levy; Richardson, D.; Spenger, J. J. (2001) Agri. Biol. Environ. Stat.: In Press.
40       Howard-Reed, C.; Rea, A. W.; Zufall, M. J.; Burke, J. M.; Williams, R. W.; Suggs, J. C.; Sheldon,  L. S.; Walsh, D.;
41             Kwock, R. (2000) Use of a continuous nephelometer to measure personal exposure to particles during the
42             U.S. Environmental Protection Agency Baltimore and Fresno panel studies. J. Air Waste Manage. Assoc.
43             50:1125-1132.
44       Institute for Social Research. (1997) Child development supplement to the panel study of income dynamics.
45             Ann Arbor, MI: University of Michigan. Available: www.isr.umich.edu/src/child-development/home.html
46             [2001, January 31].
47       Isukapalli, S. S.; Georgopoulos, P. G. (2000) Modeling of outdoor/indoor relationships of gas phase pollutants and
48             particulate matter. Presented at: PM2000: particulate matter and health—the scientific basis for regulatory
49             decision-making, specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste
50             Management Association.
51       Jang, M.; Kamens, R. M. (1999) Newly characterized products and composition of secondary aerosols from the
52             reaction of ec-pinene with ozone. Atmos. Environ. 33: 459-474.
53       Janssen, N. (1998) Personal exposure to airborne particles: validity of outdoor concentrations as a measure of
54             exposure in time series studies [dissertation]. Wageningen, Netherlands: Landbouwuniversiteit Wageningen.
         March 2001
5-104
DRAFT-DO NOT QUOTE OR CITE

-------
   1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
45
46
47
48
49
50
51
52
53
 Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1995) A relation between personal and ambient.
       Epidemiology 6(suppl.): S45.
 Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1997) Childhood exposure to PM10: relation between
       personal, classroom, and outdoor concentrations. Occup. Environ. Med. 54: 888-894.
 Janssen, N. A. H.; Hoek, G.; Brunekreef, B.; Harssema, H.; Mensink, I.; Zuidhof, A. (1998a) Personal sampling of
       particles in adults: relation among personal, indoor, and outdoor air concentrations. Am. J. Epidemiol
       147: 537-547.
 Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1998b) Personal sampling of airborne particles:
       method performance and data quality. J. Exposure Anal. Environ. Epidemiol. 8: 37-49.
 Janssen, N. A. H.; Hoek, G.; Brunekreef, B.; Harssema, H. (1999a) Mass concentration and elemental composition
       of PM10 in classrooms. Occup. Environ. Med. 56: 482-487.
 Janssen, L. H. J. M.; Buringh, E.; van der Meulen, A.; van den Hout, K. D. (1999b) A method to estimate the
       distribution of various fractions of PMIO in ambient air in the Netherlands. Atmos. Environ. 33: 3325-3334.
 Janssen, N. A. H.; De Hartog, J. J.; Hoek, G.; Brunekreef, B.; Lanki, T.; Timonen, K. L.; Pekkanen, J. (2000)
       Personal exposure to fine paniculate matter in elderly subjects: relation between personal, indoor, and
       outdoor concentrations. J. Air Waste Manage. Assoc. 50: 1133-1143.
 Jantunen, M. J.; Hanninen, O.; Katsouyanni, K.; Knoppel, H.; Kuenzli, N.; Lebret, E.; Maroni, M.; Saarela, K.;
       Sram, R.; Zmirou, D. (1998) Air pollution exposure in European cities: the EXPOLIS study. J. Exposure
       Anal. Environ. Epidemiol. 8: 495-518.
 Jantunen, M. J.; Kousa, A.; Hanninen, O.; Koistinen, K.; Sram, R.; Lanki, T.; Lioy, P. J. (2000) Long term
       population PM exposures correlate well with outdoor air levels - short term individual exposures do not.
       Presented at: PM2000: particulate matter and health—the scientific basis for regulatory decision-making,
       specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management
       Association.
 Jenkins, R. A.; Palausky, M. A.; Counts, R. W.; Guerin, M. R.; Dindal, A. B.; Bayne, C. K. (1996a) Determination
       of personal exposure of non-smokers to environmental tobacco smoke in the United States  Lung Cancer
       14(suppl. 1): S195-S213.
 Jenkins, R. A.; Palausky, A.; Counts, R. W.; Bayne, C. K.; Dindal,  A. B.; Guerin, M.  R. (1996) Exposure to
       environmental tobacco smoke in sixteen cities in the United States as determined by personal breathing zone
       air sampling. J. Exposure Anal. Environ. Epidemiol. 6: 473-502.
 Jensen, S. S. (1999) A geographic approach to modeling human exposure to traffic air pollution using GIS.
       Roskilde, Denmark: National Environmental Research Institute.
 Johnson, T. (1984) A study of personal exposure to carbon monoxide in Denver, Colorado. Research  Triangle Park,
       NC:  U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report no
       EPA-600/4-84-014.
 Johnson, T. (1989) Human activity patterns in Cincinnati, Ohio [final report]. Palo Alto, CA: Electric Power
      Research Institute; project no. RP940-06; report no. EN-6204.
 Johnson, T. R. (1995) Recent advances in the estimation of population exposure to mobile source pollutants.
      J. Exposure Anal. Environ. Epidemiol. 5: 551-571.
Johnson, T.; Long, T.; Ollison, W. (2000) Prediction of hourly microenvironmental concentrations of fine particles
      based on measurements obtained from the Baltimore scripted activity study. J. Exposure Anal Environ
      Epidemiol.  10:403-411.
Jones, N. C.; Thornton, C. A.; Mark, D.; Harrison, R. M. (2000) Indoor/outdoor relationships of particulate matter
      in domestic homes with roadside, urban and rural locations. Atmos. Environ. 34: 2603-2612.
Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
      segregated mass concentration data and ultrafine particle number concentrations in urban areas.
      Atmos. Environ. 33: 2907-2913.
Kleeman, M. J.;  Cass, G. R. (1998) Source contributions to the size and composition distribution of urban
      particulate air pollution. Atmos. Environ. 32: 2803-2816.
Klepeis, N. E. (1999) An introduction to the indirect exposure assessment approach: modeling human  exposure
      using microenvironmental measurements and the recent national human activity pattern survey.  Environ
      Health Perspect. Suppl. 107(2): 365-374.
         March 2001
                                                5-105
                                                                       DRAFT-DO NOT QUOTE OR CITE

-------
 1       Klepeis, N. E.; Ott, W. R.; Switzer, P. (1994) A Total Human Exposure Model (THEM) for Respirable Suspended
 2             Particles (RSP). Presented at: 87th annual meeting & exhibition of the Air & Waste Management
 3             Association; June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association; paper no. 94-
 4             WA75A.03. Available from: NTIS, Springfield, VA; PB94-197415.
 5       Klepeis, N. E.; Ott, W. R.; Switzer, P. (1995) Modeling the time series of respirable suspended particles and carbon
 6             monoxide from multiple smokers: validation in two public smoking lounges. Presented at: 88th annual
 7             meeting and exhibition of the Air & Waste Management Association; June; San Antonio, TX. Pittsburgh,
 8             PA: Air & Waste Management Association; paper no. 95-MPH.03.
 9       Klepeis, N. E.; Ott, W. R.; Switzer, P. (1996) A multiple-smoker model for predicting indoor air quality in public
10             lounges. Environ. Sci. Technol. 30: 2813-2820.
11       Klepeis, N. E.; Switzer, P.; Ott, W. R. (2000) Estimating exposure to environmental tobacco smoke particles using
12             24-hour time-activity recall diaries. Presented at: PM2000: particulate matter and health—the scientific basis
13             for regulatory decision-making, specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA:
14             Air & Waste Management Association.
15       Koontz, M. D.; Rector, H. E. (1995) Estimation of distributions for residential air exchange rates: final report.
16             Washington, DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics;
17             contractnos. 68-D9-0166 and 68-D3-0013.
18       Koontz, M. D.; Niang, L.  L. (1998) California population indoor exposure model (CPIEM): version 1.4F. Final
19             report A933-157: user's guide. Sacramento, CA: California Air Resources Board; GEOMET report no.
20             IE-2632.
21       Koutrakis, P.; Briggs, S. L. K.; Leaderer, B. P. (1992) Source apportionment of indoor aerosols in Suffolk and
22             Onondaga Counties, New York. Environ. Sci. Technol. 26: 521 -527.
23       Krewski, D.; Burnett, R. T.; Goldberg, M. S.; Hoover, K.; Siemiatycki, J.; Jerrett, M.; Abrahamowicz, M.;
24             White, W. H. (2000) Reanalysis of the Harvard Six Cities study and the American Cancer Society study of
25             particulate air pollution and mortality: a special report of the Institute's Particle Epidemiology Reanalysis
26             Project. Cambridge, MA: Health Effects Institute.
27       Lachenmyer, C.; Hidy, G. M. (2000) Urban measurements of outdoor-indoor PM2 5 concentrations and personal
28             exposure in the deep south. Part I. Pilot study of mass concentrations for nonsmoking subjects. Aerosol Sci.
29             Technol. 32:34-51.
30       Laden, F.; Neas, L. M.; Dockery, D. W.; Schwartz, J. (2000) Association of fine particulate matter from different
31             sources with daily mortality in six U.S. cities. Environ. Health Perspect. 108: 941-947.
32  '     Lai, A. C. K.; Nazaroff, W. W. (2000) Modeling indoor particle deposition from turbulent flow onto smooth
33             surfaces. J. Aerosol Sci. 31: 463-476.
34       Leaderer, B. P.; Naeher, L.; Jankun, T.; Balenger, K.; Holford, T. R.; Toth, C.; Sullivan, J.; Wolfson, J. M.;
35             Koutrakis, P. (1999) Indoor, outdoor, and regional summer and winter concentrations of PM10, PM25, SO42",
36             H+, NH/, NO3", NH3, and nitrous acid in homes with and without kerosene space heaters. Environ. Health
37             Perspect. 107:223-231.
38       Lee, S. C.; Chang, M. (1999) Indoor air quality investigations at five classrooms. Indoor Air 9: 134-138.
39       Lessler, J. T.; Kalsbeek, W. D. (1992) Nonsampling error in surveys. New York, NY: John Wiley & Sons, Inc.;
40             p. 7.
41       Liao, D.; Creason, J.; Shy, C.; Williams, R.; Watts, R.; Zweidinger, R. (1999) Daily variation of particulate air
42             pollution and poor cardiac autonomic control in the elderly. Environ. Health Perspect. 107: 521-525.
43       Ligocki, M. P.; Liu, H. I.  H.; Cass, G. R.; John, W. (1990) Measurements of particle deposition rates inside
44             Southern California museums. Aerosol Sci. Technol. 13: 85-101.
45       Lillquist, D. R.; Lee, J. S.; Ramsay, J. R.; Boucher, K. M.; Walton, Z. L.; Lyon, J. L. (1998) A comparison of
46             indoor/outdoor PM10 concentrations measured at three hospitals and a centrally located monitor in Utah.
47             Appl. Occup. Environ. Hyg. 13:409-415.
48       Linn, W. S.; Gong, H., Jr.; Clark, K. W.; Anderson, K. R. (1999) Day-to-day particulate exposures and health
49             changes in Los Angeles area residents with severe lung disease. J. Air Waste Manage. Assoc.
50             49:PM108-PM115.
51       Lioy, P. J.; Waldman, J. M.; Buckley, T.; Butler, J.; Pietarinen, C. (1990) The personal, indoor and outdoor
52             concentrations of PM-10 measured in an industrial community during the winter. Atmos. Environ. Part B
53             24:57-66.
54       Lioy, P. J.; Wainman, T.; Zhang, J.; Goldsmith, S. (1999) Typical household vacuum cleaners: the collection
55             efficiency and emissions characteristics for fine particles. J. Air Waste Manage. Assoc. 49: 200-206.
         March 2001
5-106
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
40
 41
 42
 43
 44
 45
46
47
48
49
 50
 51
52
53
54
 Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of paniculate matter components with daily
       mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report no.

 Long, C. M.; Suh, H. H.; Koutrakis, P. (2000a) Characterization of indoor particle sources using continuous mass
       and size monitors. J. Air Waste Manage. Assoc. 50: 1236-1250.
 Long, C. M.; Suh, H. H.; Koutrakis, P. (2000b) Using time- and size-resolved particulate data to explore infiltration
       and deposition behavior. Environ. Sci. Technol.: accepted.
 Lurmann,  F. W.; Korc, M. E. (1994) User's guide to the regional human exposure (REHEX) model. San Francisco,
       CA: Bay Area Air Quality Management District; draft report no. STI-93150-1414-DR.
 Macintosh, D. L.; Xue, J.; Ozkaynak, H.; Spengler, J. D.; Ryan, P. B. (1995) A population-based exposure model
       for  benzene. J. Expos. Anal. Environ. Epidemiol. 5: 375-403.
 Mage, D. T. (1985) Concepts of human exposure assessment for airborne particulate matter Environ Int
       11:407-412.                                                                       '   '
 Mage, D. T. (1998) How much PM of ambient origin are people exposed to? In: Measurement of toxic and related
       air pollutants. Volume II: proceedings of a specialty conference; September; Gary, NC. Pittsburgh, PA: Air &
       Waste Management Association; pp. 666-681. (A&WMA publication VIP-85).
 Mage, D. T.; Ott, W. R. (1996) Accounting for nonuniform mixing and human exposure in indoor environments.
       In: Tichenor, B. A., ed. Characterizing sources of indoor air pollution and related sink effects. Philadelphia,
       PA: American Society for Testing and Materials; pp. 263-278.
 Mage, D.;  Wilson, W.; Hasselblad, V.; Grant, L. (1999) Assessment of human exposure to ambient particulate
       matter. J. Air Waste Manage. Assoc. 49: 174-185.
 Mar, T. F.; Norris, G. A.; Koenig, J. Q.; Larson, T. V. (2000) Associations between air pollution and mortality in
       Phoenix, 1995-1997. Environ. Health Perspect. 108: 347-353.
 McBride, S. J.; Ferro, A. R.; Ott, W. R; Switzer, P.; Hildemann, L. M. (1999) Investigations of the proximity effect
       for pollutants in the indoor environment. J. Exposure Anal. Environ. Epidemiol. 9: 602-621.
 McCurdy, T. (1995) Estimating human exposure to selected motor vehicle pollutants using the NEM series of
      models: lessons to be learned. J. Exposure Anal. Environ. Epidemiol. 5: 533-550.
 McCurdy, T. (1997) Modeling the dose profile in human exposure assessments: ozone as an example Rev Toxicol
       1:3-23.
 McCurdy, T. (2000) Conceptual basis  for multi-route intake dose modeling using an energy expenditure approach
      J. Exposure Anal. Environ. Epidemiol. 10: 86-97.
 Mihlan, G. J.; Todd, L. A.; Truong, K. N. (2000) Assessment of occupational exposure patterns by
      frequency-domain analysis of time series data. Appl. Occup. Environ. Hyg. 15: 120-130.
 Miller, S. L.; Branoff, S.; Nazaroff, W. W. (1998) Exposure to toxic air contaminants in environmental tobacco
      smoke: an assessment for California based on personal monitoring data. J. Exposure Anal Environ
      Epidemiol. 8:287-311.
 Monn, Ch.; Carabias, V.; Junker, M.; Waeber, R.; Karrer, M.; Wanner, H. U. (1997) Small-scale spatial variability
      of particulate matter <10 urn (PM10) and nitrogen dioxide. Atmos. Environ. 31: 2243-2247.
 Murray, D.  M.; Burmaster, D. E. (1995) Residential air exchange rates in the United States: empirical and estimated
      parametric distributions by season and climatic region. Risk Anal. 15: 459-465.
 National Air Pollution Control Administration. (1969) Air quality criteria for particulate matter. Washington, DC:
      U.S. Department of Health, Education, and Welfare, Public Health Service; NAPCA publication no AP-49
      Available from: NTIS, Springfield, VA; PB-190251 IE A.
 National Research Council. (1991) Human exposure assessment for airborne pollutants: advances and opportunities
      Washington, DC: National Academy of Sciences.
 National Research Council. (1998) Research priorities for airborne particulate matter. I. Immediate priorities and a
      long-range research portfolio. Washington, DC: National Academy Press.
 National Research Council. (1999) Research priorities for airborne particulate matter. II. Evaluating research
      progress and updating the portfolio. Washington, DC: National Academy Press.
Navidi, W.; Thomas, D.; Langholz, B.; Stram, D. (1999) Statistical methods for epidemic-logic studies of the health
      effects of air pollution. Cambridge, MA: Health Effects Institute;  research report no. 86.
Nelson, P. R.; Conrad, F. W.; Kelly, S. P.; Maiolo, K. C.; Richardson, J. D.; Ogden, M. W. (1998) Composition of
      environmental tobacco smoke (ETS) from international cigarettes part II: nine country follow-up Environ
      Int. 24: 251-257.
         March 2001
                                                5-107
                                                                       DRAFT-DO NOT QUOTE OR CITE

-------
 1       Nieuwenhuijsen, M. J.; Noderer, K. S.; Schenker, M. B.; Vallyathan, V.; Olenchock, S. (1999) Personal exposure to
 2             dust, endotoxin and crystalline silica in California agriculture. Ann. Occup. Hyg. 43: 35-42.
 3       Oglesby, L.; Kiinzli, N.; Roosli, M.; Braun-Fahrlander, C.; Mathys, P.; Stem, W.; Jantunen, M.; Kousa, A. (2000)
 4             Validity of ambient levels of fine particles as surrogate for personal exposure to outdoor air
 5             pollution—results of the European EXPOLIS-EAS study (Swiss Center Basel). J. Air Waste Manage. Assoc.
 6             50:  1251-1261.
 7       Ott, W. R. (1984) Exposure estimates based on computer generated activity patterns. J. Toxicol. Clin. Toxicol.
 8             21:97-128.
 9       Ott, W. R. (1999) Mathematical models for predicting indoor air quality from smoking activity. Environ. Health
10          '   Perspect. 107(suppl. 2): 375-381.
11       Ott, W.; Thomas, J.; Mage, D.; Wallace, L. (1988) Validation of the simulation of human activity and pollutant
12          '   exposure (SHAPE) model using paired days from the Denver, CO, carbon monoxide field study. Atmos.
13             Environ. 22:2101-2113.
14       Ott, W. R.; Mage, D. T.; Thomas, J. (1992) Comparison of microenvironmental CO concentrations in two cities for
15             human exposure modeling. J. Exposure Anal. Environ. Epidemiol. 2: 249-267.
16       Ott, W. R.; Klepeis, N. E.; Switzer, P. (1995) Modeling environmental tobacco smoke in the home using transfer
17          '   functions. Presented at: 88th annual meeting and exhibition of the Air & Waste Management Association;
1 g             June; San Antonio, TX; paper no. 95-WP84B.03. Research Triangle Park, NC: U.S. Environmental
19             Protection Agency, Atmospheric Research And Exposure Assessment Laboratory; report no.
20             EPA/600/A-95/079. Available from: NTIS, Springfield, VA; PB95-225488.
21       Ott, W.; Switzer, P.; Robinson, J. (1996) Particle concentrations inside a tavern before and after prohibition of
22             smoking: evaluating the performance of an indoor air quality model. J. Air Waste Manage. Assoc.
23             46:1120-1134.
24       Ott, W.; Wallace, L.; Mage, D. (2000) Predicting particulate  (PM,0) personal exposure distributions using a random
25             component superposition (RCS) statistical model. J. Air Waste Manage. Assoc. 50: 1390-1406.
26       Owen, M. K.; Ensor, D. S.; Sparks, L. E. (1992) Airborne particle sizes and sources found in indoor air.
27             Atmos. Environ. Part A 26: 2149-2162.
28       Ozkaynak, H.; Spengler, J. (1996) The role of outdoor particulate matter in assessing total human exposure.
29             In: Wilson, R.; Spengler, J., eds. Particles in our air: concentrations and health effects. Cambridge, MA:
30             Harvard University Press.
31       Ozkaynak, H.; Thurston, G. D. (1987) Associations between 1980 U.S. mortality rates and alternative measures of
32             airborne particle concentration. Risk Anal. 7: 449-461.
33       Ozkaynak, H.; Ryan, P. B.; Spengler, J. D.; Laird, N. M. (1986) Bias due to misclassification of personal exposures
34            in epidemiologic studies of indoor and outdoor air pollution. In: Berglund, B.; Berglund, U.; Lindvall, T.;
35             Spengler, J.; Sundell, J., eds. Indoor air  quality: papers from the third international conference on indoor air
36            quality and climate; August 1984; Stockholm, Sweden. Environ. Int. 12: 389-393.
37      Ozkaynak, H.; Spengler, J. D.; Ludwig, J. F.; Butler, D. A.; Clayton, C. A.; Pellizzari, E.; Wiener, R. W. (1990)
38            Personal exposure to particulate matter: findings from the Particle Total Exposure Assessment Methodology
39            (PTEAM) prepilot study. In: Indoor air '90: precedings of the 5th international conference on indoor air
40            quality and climate, volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON,
41             Canada: International Conference on Indoor Air Quality and Climate, Inc.; pp. 571-576.
42      Ozkaynak, H.; Spengler, J. D.; Xue, J.; Koutrakis, P.; Pellizzari, E. D.; Wallace, L. (1993) Sources and factors
43            influencing personal and indoor exposures to particles, elements and nicotine: findings from the particle
44            TEAM pilot study. In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma,
45            H., eds. Indoor air '93: proceedings of the 6th international conference on indoor air quality and climate, v. 3,
46            combustion products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor A.ir '93;
47            pp. 457-462.
48      Ozkaynak, H.; Xue, J.; Spengler, J.; Wallace, L.; Pellizzari, E.; Jenkins, P. (1996) Personal exposures to airborne
49            particles and metals: results from the particle TEAM study in Riverside, California. J. Exp. Anal. Environ.
50            Epidemiol. 6: 57-78.
51       Ozkaynak, H.; Xue, J.; Weker, R.; Butler, D.; Koutrakis, P.; Spengler, J. (1996b) The particle TEAM (PTEAM)
52            study: analysis of the data:  final report,  volume III. Research Triangle Park, NC: U.S. Environmental
53            Protection Agency, Atmospheric Research and Exposure Assessment Laboratory; report no.
54            EPA/600/R-95/098. Available from: NTIS, Springfield, VA;  PB97-102495.
          March 2001
5-108
DRAFT-DO NOT QUOTE OR CITE

-------
  1      Patterson, E.; Eatough, D. J. (2000) Indoor/outdoor relationships for ambient PM2 5 and associated pollutants:
  2            epidemiological implications in Lindon, Utah. J. Air Waste Manage. Assoc. 50: 103-110.
  3      Pellizzari, E. D., Mason, R. E.; Clayton, C. A.; Thomas, K. W.; Cooper, S.; Piper. L.; Rodes, C.; Goldberg, M.;
  4            Roberds, J.; Michael, L. (1998) Manganese exposure study (Toronto). Final report. Research Triangle Park,
  5            NC: Research Triangle Institute, Analytical and Chemical Sciences; report no. RTI/6312/02-01 F.
  6      Pellizzari, E. D.; Clayton, C. A.; Rodes, C. E.; Mason, R. E.; Piper, L. L.; Fort, B.; Pfeifer, G.; Lynam, D. (1999)
  7            Participate matter and manganese exposures in Toronto, Canada. Atmos. Environ. 33: 721-734.
  8      Phillips, K.; Howard, D. A.; Browne, D.; Lewsley, J. M. (1994) Assessment of personal exposures to environmental
  9            tobacco smoke in British nonsmokers. Environ. Int. 20: 693-712.
 10      Phillips, K.; Bentley, M. C.; Howard, D. A.;.Alvan, G. (1996) Assessment of air quality in Stockholm by personal
 11            monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Scand. J.
 12            Work Environ. Health 22(suppl. 1):  1 -24.
 13      Phillips, K.; Bentley, M. C.; Howard, D. A.; Alvan, G.; Huici, A. (1997a) Assessment of air quality in Barcelona by
 14            personal monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke.
 15            Environ. Int. 23:  173-196.
 16      Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1997b) Assessment of air quality in Turin by personal
 17            monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Environ.  Int.
 18            23:851-871.
 19      Phillips, K.; Bentley, M. C.; Howard, D. A.; Alvan, G. (1998a) Assessment of air quality in Paris by personal
20            monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Environ.  Int.
21            24:405-425.
22      Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1998b) Assessment of environmental tobacco smoke and
23            respirable suspended particle exposures for nonsmokers in Lisbon by personal monitoring. Environ. Int.
24            24:301-324.
25      Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1999)  Assessment of environmental tobacco smoke and
26            respirable suspended particle exposures for nonsmokers in Basel by personal monitoring. Atmos. Environ.
27            33: 1889-1904.
28      Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.;  Speizer, F. E.;  Heath, C. W., Jr.
29            (1995) Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J.
30            Respir. Crit. Care Med. 151: 669-674.
31      Pope, C. A., Ill; Hill, R. W.; Villegas, G. M. (1999) Particulate air pollution and daily mortality on Utah's Wasatch
32            Front. Environ. Health Perspect.: 107: 567-573.
33      Praml, G.; Schierl, R. (2000) Dust exposure in Munich public transportation: a comprehensive 4-year survey in
34            buses and trams. Int. Arch. Occup. Environ. Health 73: 209-214.
35      Quintana, P. J. E.; Samimi, B. S.; Kleinman, M. T.; Liu, L.-J.; Soto, K.; Warner, G. Y.; Bufalino, C.; Valencia, J.;
36            Francis, D.; Hovell, M. H.; Delfino, R. J. (2000) Evaluation of a real-time passive personal particle monitor
37            in fixed site residential indoor and ambient measurements. J. Exposure Anal. Environ. Epidemiol.
38            10:437-445.
39      Ramachandran, G.; Vincent, J. H. (1999) A Bayesian approach to retrospective exposure assessment. Appl. Occup.
40            Environ. Hyg. 14: 547-557.
41      Rea, A. W.; Zufall, M. J.; Williams, R. W.; Howard-Reed, C.;  Sheldon, L. S. (2001) The influence of human
42            activity patterns on personal PM exposure: a comparative analysis of filter-based and continuous particle
43            measurements. J. Air Waste Manage. Assoc.: In Press.
44      Ren, P.; Jankun, T. M.; Leaderer, B. P. (1999) Comparisons of seasonal fungal prevalence in indoor and outdoor air
45            and in house dusts of dwellings in one northeast American county. J. Exposure Anal. Environ. Epidemiol.
46            9:560-568.
47      Robinson, J. P.; Wiley, J. A.; Piazza, T.; Garrett,  K.; Cirksena, K. (1989) Activity patterns of California residents
48            and their implications for potential exposure to pollution. Sacramento, CA: California Air Resources Board
49            report no. CARB-A6-177-33.
50      Robinson, J. P.; Switzer, P.; Ott, W. (1994) Microenvironmental factors related to Californians'  potential exposure
51            to environmental tobacco smoke (ETS). Research Triangle Park, NC: U.S. Environmental Protection Agency,
52            Atmospheric Research and Exposure Assessment Laboratory;  report no. EPA/600/R-94/116. Available from:
53            NTIS, Springfield, VA; PB94-214814.
54
         March 2001
5-109
DRAFT-DO NOT QUOTE OR CITE

-------
 1       Rodes, C.; Sheldon, L.; Whitaker, D.; Clayton, A.; Fitzgerald, K.; Flanagan, J.; DiGenova, F.; Hering, S.;
 2             Frazier, C. (1998) Measuring concentrations of selected air pollutants inside California vehicles [final
 3             report]. Sacramento, CA: California Environmental Protection Agency, Air Resources Board; contract no.
 4             95-339.
 5       Rojas-Bracho, L.; Suh, H. H.; Koutrakis, P. (2000) Relationships among personal, indoor, and outdoor fine and
 6             coarse particle concentrations for individuals with COPD. J. Exposure Anal. Environ. Epidemiol.
 7             10:294-306.
 8       Roorda-Knape, M. C.; Janssen, N. A. H.; De Hartog, J. J.; Van Vliet, P. H. N.; Harssema, H.; Brunekreef, B. (1998)
 9             Air pollution from traffic in city districts near major motorways. Atmos. Environ. 32: 1921-1930.
10       Ryan, P. B. (1991) An overview of human exposure modeling. J. Exposure Anal. Environ. Epidemiol. 1: 453-474.
11       Santos-Burgoa, C.;  Rojas-Bracho, L.; Rosas-Perez, I.; Ramirez-Sanchez, A.; Sanchez-Rico, G.;
12             Mejia-Hernandez, S. (1998) Modelaje de exposition a particulas en poblacion general y riesgo de
13             enfermedad respiratoria [Particle exposure modeling in the general population and risk of respiratory
14             disease]. Gac. Med. Mex. 134:407-417.
15       Sarnat, J. A.; Koutrakis, P.; Suh, H. H. (2000) Assessing the relationship between personal particulate and gaseous
16             exposures of senior citizens living in Baltimore, MD. J. Air Waste Manage. Assoc. 50: 1184-1198.
17       Schwartz, J.; Levin, R. (1999) Drinking water turbidity and health. Epidemiology 10: 86-90.
18       Sexton, K.; Ryan, P. B. (1988) Assessment of human exposure to air pollution: methods, measurements, and
19             models. In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the automobile, and public health.
20             Washington, DC: National Academy Press; pp. 207-238.
21       Sexton, K.; Spengler, J. D.; Treitman, R. D. (1984) Personal exposure to respirable particles: a case study in
22             Waterbury, Vermont. Atmos. Environ. 18: 1385-1398.
23       Sheldon, L.; Clayton, A.; Jones, B.; Keever, J.; Perritt, R.; Smith, D.; Whitaker, D.; Whitmore, R. (1992a) Indoor
24             pollutant concentrations and exposures. Sacramento, CA: California Air Resources Board; report no.
25             ARB/R-92/486.
26       Sheldon, L.; Clayton, A.; Keever, J. Perritt, R.; Whitaker, D. (1992b) PTEAM: monitoring of phthalates and PAHs
27             in indoor and outdoor air samples in Riverside, California, volume II [final report]. Sacramento, CA:
28             California Environmental Protection Agency, State Air Resources Board, Research Division; contract no.
29             A933-144. Available from: NTIS, Springfield, VA; PB93-205649/XAB.
30       Sheldon, L.; Clayton, A.; Perritt, R.; Whitaker, D. A.; Keever, J. (1993a) Indoor concentrations of polycyclic
31             aromatic hydrocarbons in California residences and their relationship to combustion source use. In: Jantunen,
32             M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air '93: proceedings
33             of the 6th international conference on indoor air quality and climate. Volume 3: combustion products, risk
34             assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 29-34.
35       Sheldon, L.; Whitaker, D.; Keever, J.; Clayton, A.;  Perritt, R. (1993b) Phthalates and PAHs in indoor and outdoor
36             air in a southern California community. In: Jantunen, M.; Kalliokoski,  P.; Kukkonen, E.; Saarela, K.;
37             Seppanen, O.; Vuorelma, H., eds. Indoor air '93: proceedings of the 6th international conference on indoor
38             air quality and climate. Volume 3: combustion products, risk assessment, policies; July; Helsinki, Finland.
39             Helsinki, Finland: Indoor Air '93; pp. 109-114.
40       Sheldon, L.; Clayton, A.; Keever, J.; Perritt, R.; Whitaker, D. (1993c) Indoor concentrations of polycyclic aromatic
41             hydrocarbons in California residences. Sacramento, CA: Air Resources Board; ARB contract no. A033-132.
42       Shy, C. M.; Kleinbaum, D. G.; Morgenstern, H. (1978) The effect of misclassification of exposure status in
43             epidemiological studies of air pollution health effects. Bull. N. Y. Acad. Med. 54: 1155-1165.
44       Siegmann, K.; Scherrer, L.; Siegmann, H. C. (1999) Physical and chemical properties of airborne nanoscale
45             particles and how to measure the impact on human health. J. Mol. Struct. (Theochem) 458: 191-201.
46       Sinclair, J. D.; Psota-Kelty, L. A.; Weschler, C. J. (1988) Indoor/outdoor ratios  and indoor surface accumulations of
47             ionic substances at Newark, New Jersey. Atmos. Environ. 22:461 -469.
48       Sinclair, J. D.; Psota-Kelty, L. A.; Weschler, C. J.; Shields, H. C. (1990) Measurement and modeling of airborne
49             concentrations and indoor surface accumulation rates of ionic substances at Neenah, Wisconsin. Atmos.
50             Environ. Part A 24:627-638.
51       Spengler, J. D.; Treitman, R. D.; Tosteson, T. D.; Mage,  D. T.; Soczek, M. L. (1985) Personal exposures to
52             respirable particulates and implications for air pollution epidemiology. Environ. Sci. Technol. 19: 700-707.
53
54
         March 2001
5-110
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
47
48
49
 50
51
52
53
54
 Spengler, J. D.; Ozkaynak, H.; Ludwig, J.; Allen, G.; Pellizzari, E. D.; Wiener, R. (1989) Personal exposures to
       participate matter: instruments and methodologies P-TEAM. In: Measurement of toxic and related air
       pollutants: proceedings of the 1989 U.S. EPA/A&WMA international symposium; May; Raleigh, NC.
       Pittsburgh, PA: Air & Waste Management Association; pp. 449-463. (A&WMA publication VIP-13).
 Spier, C. E.; Little, D. E.; Trim, S. C.; Johnson, T. R.; Linn, W. S.; Hackney, J. D. (1992) Activity patterns in
       elementary and high school students exposed to oxidant pollution. J. Exposure Anal. Environ. Epidemiol
       2:277-293.  .
 Suggs, J. C.; Burton, R. M. (1983) Spatial characteristics of inhalable particles in the Philadelphia metropolitan
       area. J. Air Pollut. Control Assoc. 33: 688-691.
 Suh, H. H.; Spengler, J. D.; Koutrakis, P. (1992) Personal exposures to acid aerosols and ammonia  Environ  Sci
       Technol. 26: 2507-2517.
 Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1993a) Indoor and outdoor acid aerosol and gas concentrations.
       In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air
       '93: proceedings of the 6th international conference on indoor air quality and climate. Volume 3: combustion
       products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 257-262.
 Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1993b) Validation of personal exposure models for sulfate and aerosol
       strong acidity. J. Air Waste Manage. Assoc. 43: 845-850.
 Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1994) The relationship between airborne acidity and ammonia in indoor
       environments. J. Exposure Anal. Environ. Epidemiol. 4: 1-23.
 Tamura, K.; Ando, M.; Sagai, M.; Matsumoto, Y. (1996a) Estimation of levels of personal exposure to suspended
       particulate matter and nitrogen dioxide in Tokyo. Environ. Sci. (Tokyo) 4: 37-51.
 Tamura, K.; Ando, M.; Miyazaki, T.; Funasaka, K.; Kuroda, K. (1996b) Indoor concentrations and personal
       exposure to particulate matter at roadside houses. In: Indoor air '96: the 7th international conference on
       indoor air quality, volume I; July; Nagoya, Japan. Tokyo, Japan: Institute of Public Health; pp. 599-602
 Teschke, K.; Marion, S. A.; Vaughan, T. L.; Morgan, M. S.; Camp, J. (1999) Exposures to wood dust in U.S.
       industries and occupations, 1979 to 1997. Am. J. Ind. Med. 35:  581-589.
 Thatcher, T. L.; Layton, D. W. (1995) Deposition, resuspension, and penetration of particles within a residence
       Atmos. Environ. 29: 1487-1497.
 Thomas, K. W.; Pellizzari, E. D.; Clayton, C. A.; Whitaker, D. A.; Shores, R. C.; Spengler, J.; Ozkaynak, H.;
       Froehlich, S. E.; Wallace, L. A. (1993) Particle total exposure assessment methodology (PTEAM) 1990
       study: method performance and data quality for personal, indoor, and outdoor monitoring. J. Exposure Anal.
       Environ. Epidemiol. 3:203-226.
 Tsang, A. M.; Klepeis, N. E. (1996) Descriptive statistics tables from a detailed analysis of the national human
       activity pattern survey (NHAPS) data. Las Vegas, NV: U.S. Environmental Protection Agency, Office of
       Research and Development; report no. EPA/600/R-96/148.
 U.S. Environmental Protection Agency. (1982) Air quality criteria for particulate matter and sulfur oxides. Research
      Triangle Park, NC: Office of Health and Environmental  Assessment, Environmental Criteria and Assessment
      Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
 U.S. Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park,
      NC: National Center for Environmental Assessment-RTP Office; report nos. EPA/600/P-95/001aF-cF.  3v
 U.S. Environmental Protection Agency. (1998) Particulate matter research needs for human health risk assessment
      to support future reviews of the national ambient air quality standards for particulate matter. Research
      Triangle Park, NC: National Center for Environmental Assessment; report no. EPA/600/R-97/132F.
 U.S. Environmental Protection Agency. (2000) Aerometric Information Retrieval System (AIRS). Research
      Triangle Park, NC: Office of Air Quality Planning and Standards. Available: www.epa.gov/ttn/airs/
      [2000, October 6].
 Vakeva, M.; Hameri, K.; Kulmala, M.; Lahdes, R.; Ruuskanen, J.; Laitinen, T. (1999) Street level versus rooftop
      concentrations of submicron aerosol particles and gaseous pollutants in an urban street canyon
      Atmos. Environ. 33: 1385-1397.
Van Vliet, P.; Knape, M.; De Hartog, J.; Janssen, N.; Harssema, H.; Brunekreef, B. (1997) Motor vehicle exhaust
      and chronic respiratory symptoms in children living near freeways. Environ. Res. 74: 122-132.
Vette, A. F.; Rea, A. W.; Lawless, P. A.; Rodes, C. E.; Evans, G.; Highsmith, V. R.; Sheldon, L. (2001)
      Characterization of indoor-outdoor aerosol concentration relationships during the Fresno PM exposure
      studies. Aerosol Sci. Technol.: accepted.
         March 2001
                                                5-111
                                                                       DRAFT-DO NOT QUOTE OR CITE

-------
 1      Wainman, T.; Zhang, J.; Weschler, C. J.; Lioy, P. J. (2000) Ozone and limonene in indoor air: a source of
 2             submicron particle exposure. Environ. Health Perspect. 108: 1139-1145.
 3      Waldman, J. M.; Liang, C. S. K. (1993) Penetration and neutralization of acid aerosols at institutional settings.
 4             ImJantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air
 5             '93: proceedings of the 6th international conference on indoor air quality and climate. Volume 3: combustion
 6             products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 263-268.
 7      Waldman, J. M.; Liang, C. S. K.; Menon, P. (1990) Indoor exposures to acidic aerosols at child and elderly care
 8             facilities. In: Indoor air '90: precedings of the 5th international conference on indoor air quality and climate,
 9             volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada:
10             International Conference on Indoor Air Quality and Climate, Inc.; pp. 607-612.
11      Wallace, L. (1996) Indoor particles: a review. J. Air Waste Manage. Assoc. 46: 98-126.
12      Wallace, L. (2000a) Real-time monitoring of particles, PAH, and CO in an occupied townhouse. Appl. Occup.
13             Environ. Hyg. 15: 39-47.
14      Wallace, L. (2000b) Correlations of personal exposure to particles with outdoor air measurements: a review of
15             recent studies. Aerosol Sci. Technol. 32: 15-25.
16      Wallace, L.; Quackenboss, J.; Rodes, C. (1997) Continuous measurements of particles, PAH, and CO in an
17             occupied townhouse in Reston, VA. In: Measurement of toxic and related air pollutants. Volume II:
1 g             proceedings of a specialty conference; April-May; Research Triangle Park, NC. Pittsburgh, PA: Air & Waste
19             Management Association, pp. 860-871. (A&WMA publication VIP-74).
20       Watson, J. G.; Cooper, J. A.; Huntzicker, J. J. (1984) The effective variance weighting for least squares calculations
21             applied to the mass balance receptor model. Atmos. Environ. 18: 1347-1355.
22       Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
23             (1990) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
24       Watson, J. G.; Chow, J. C.; Pace, T. G. (1991) Chemical mass balance. In: Hopke, P. K., ed. Receptor modeling for
25             a'ir quality management. New York, NY: Elsevier; pp. 83-116. (Data handling in science and technology—
26             volume 7).
27       Weschler, C. J.; Shields, H. C. (1999) Indoor ozone/terpene reactions as a source of indoor particles. Atmos.
28             Environ. 33:2301-2312.
29       White, W. H. (1998) Statistical considerations in the interpretation of size-resolved particulate mass data. J. Air
30             'waste Manage. Assoc. 48:454-458.
31       Wiener, R.  W. (1988) Measurement and evaluation of personal exposure to aerosols. In: Measurement of toxic and
32             related air pollutants: proceedings of the 1988 EPA/APCA international symposium; May; Research Triangle
33             Park, NC. Pittsburgh, PA: Air Pollution Control Association; pp. 84-88. (APCA publication VIP-10).
34       Wiener, R.  W. (1989) Particle total exposure assessment methodology-an overview of planning and
35             accomplishments. In: Measurement of toxic and related air pollutants: proceedings of the 1989 U.S.
36            EPA/A&WMA international symposium; May; Raleigh, NC. Pittsburgh, PA: Air & Waste Management
37            Association; pp. 442-448. (A&WMA publication VIP-13).
38      Wiener, R. W.; Wallace, L.; Pahl, D.; Pellizzari, E.; Whittaker, D.; Spengler, J.; Ozkaynak, H. (1990) Review of the
39            particle TEAM 9 home field study. In: Measurement of toxic and related air pollutants: proceedings of the
40             1990 EPA/A&WMA international symposium, v. 2; May; Raleigh, NC. Pittsburgh, PA: Air & Waste
41             Management Association; pp. 452-460. (A&WMA publication VIP-17).
42      Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991a) Study of children's activity
43            patterns: final report. Sacramento, CA: California Air Resources Board; report no. ARB-R-93/489.
44      Williams, R.; Suggs, J.; Zweidinger, R.; Evans, G.; Creason, J.; Kwok, R.; Rodes, C.; Lawless, P.; Sheldon, L.
45            (2000a) The 1998 Baltimore Particulate Matter Epidemiology-Exposure Study: part 1. Comparison of
46            ambient, residential outdoor, indoor and apartment particulate matter monitoring. J. Exposure Anal. Environ.
47            Epidemiol. 10:518-532.
48      Williams, R.; Suggs, J.; Creason, J.; Rodes, C.; Lawless, P.; Kwok, R.; Zweidinger, R.; Sheldon, L. (2000b) The
49             1998 Baltimore particulate matter epidemiology-exposure study: part 2. Personal exposure assessment
 50             associated with an elderly study population. J. Exposure Anal. Environ. Epidemiol.: 10: 533-543.
 51      Williams, R.; Creason, J.; Zweidinger, R.; Watts,  R.; Sheldon, L.; Shy, C. (2000c) Indoor, outdoor, and personal
 52             exposure monitoring of particulate air pollution: the Baltimore elderly epidemiology-exposure pilot study.
 53             Atmos. Environ. 34:4193-4204.
 54      Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
 55             epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
          March 2001
5-112
DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
Wilson, W. E.; Mage, D. T.; Grant, L. D. (2000) Estimating separately personal exposure to ambient and
      nonambient particulate matter for epidemiology and risk assessment: why and how. J. Air Waste Manage.
      Assoc. 50: 1167-1183.
Yakovleva, E.; Hopke, P. K.; Wallace, L. (1999) Receptor modeling assessment of particle total exposure
      assessment methodology data. Environ. Sci. Technol. 33: 3645-3652.
Zeger, S. L.; Thomas, D.; Dominici, F.; Samet, J. M.; Schwartz, J.; Dockery, D.; Cohen, A. (2000) Exposure
      measurement error in time-series studies of air pollution: concepts and consequences. Environ. Health
      Perspect. 108: 419-426.
Zmirou, D.; Masclet, P.; Boudet, C.; Dor, F.; Dechenaux, J. (2000) Personal exposure to atmospheric polycyclic
      aromatic hydrocarbons in a general adult population and lung cancer risk assessment. J. Occup. Environ.
      Med. 42: 121-126.
        March 2001
                                               5-113
DRAFT-DO NOT QUOTE OR CITE

-------

-------

-------
United States
Environmental Protection Agency/ORD
National Center for
   Environmental Assessment
Washington, D.C. 20460
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.

If you do not wish to receive these reports CHECK HEREtH ;
detach, or copy this cover, and return to the address in the
upper left-hand corner.
Official Business
Penally for Private Use
S300
EP/V600/P-99/002aB

-------