United States
Environmental Protection
Agency
Environmental Criteria and
Assessment Office
Research Triangle Park NC 27711
EPA-600/8-82-029b
December 1982
                             FINAL
Research and Development
Air Quality
Criteria for
Particulate Matter
and Sulfur Oxides
Volume II

-------
                         EPA-600/8-82-029b

                             December 1982
  Air Quality Criteria
for P^tieulate Matter
  and Sulfur Oxides

        Volume  II
  U.S. ENVIRONMENTAL PROTECTION AGENCY
    Office of Research and Development
  Environmental Criteria and Assessment Office
     Research Triangle Park, NC 27711

-------
                     NOTICE

Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
                        ii

-------
                                   Preface

     This  document  is  Volume II  of a three-volume  revision of  Air  Qual1ty
Criteria for  Particulate Matter  and Air Quality Criteria  for  Sulfur  Oxides,
first published in  1969 and 1970, respectively.  By law, air quality criteria
documents are the basis for establishment of the National Ambient Air Quality
Standards (NAAQS).  The Air Quality Criteria document of which this volume is
a part  has  been  prepared in response to  specific  requirements  of Section 108
of the Clean Air Act,  as amended in 1977.   The Clean Air Act requires that the
Administrator  periodically  review,  and  as  appropriate,  update  and  reissue
criteria for NAAQS.                                              "'  •-'
     As  the legally  prescribed basis  for deciding  on National  Ambient Air
Quality Standards,  the  present document,  Air Quality Criteria for Particulate
Matter  and  Sulfur Oxides,  focuses on characterization  of  health  and  welfare
effects associated  with exposure  to particulate matter  and  sulfur oxides and
pollutant  concentrations  which cause  such  effects.   The  major  health and
welfare  effects  of  particulate matter  and  sulfur oxides  are discussed  in
Chapters 8 through 14 in Volume III of this document.   To assist the reader in
putting the effects into perspective with the real-world environment, Chapters
2  through  7  in  the  present  volume  (Volume  II)  have  been prepared.   The
chapters of Volume II discuss essential points regarding:  physical and chemi-
cal properties; air monitoring and analytical measurement techniques;  sources
and  emissions;   transport,  transformation,  and  fate;  and  observed  ambient
concentrations of the  pollutants.   Also,  Chapter 7 in  this volume introduces
the  reader to  the contemporary  problem  of  acidic deposition  and potential
contributions of sulfur oxides to acidic deposition phenomena.
     Volume I introduces the criteria document, explains the rationale behind
combining the evaluation of criteria for particulate matter and sulfur oxides
in a single document and briefly summarizes the content of the entire criteria
document.   However, for a  fuller understanding  of  the  health  and  welfare
effects  of particulate  matter  and sulfur oxides, both Volumes  II and III of
the document should be  consulted.
     The  Agency  is pleased  to acknowledge  the efforts  of all  persons and
groups  who  have  contributed to the preparation of this document.  In the last
analysis,  however,  the  Environmental  Protection Agency  accepts full  respon-
sibility for its content.
                                    n i

-------
Page Intentionally Blank

-------
                                   VOLUME II
                                   CONTENTS
2.   PHYSICS AND CHEMISTRY OF SULFUR OXIDES AND PARTICULATE MATTER.....      2-1
    2.1  INTRODUCTION	.•	      2-1
    2.2  ATMOSPHERIC DOMAIN AND PROCESSES.	      2-3
    2.3  PHYSICS AND CHEMISTRY OF SULFUR OXIDES	      2-8
         2.3.1  Physical Properties of Sulfur Oxides in the Gas
                Phase	      2-8
         2.3.2  Solution Physical Properties	      2-8
                2.3.2.1  Sulfur Dioxide	      2-8
                2.3.2.2  Sulfur Trioxide and Sulfuric Acid	     2-12
         2.3.3  Gas-Phase Chemical Reactions of Sulfur Dioxide	     2-12
                2.3.3.1  Elementary Reactions.......		     2-14
                2.3.3.2  Tropospheric Chemistry of Sulfur Dioxide
                         Oxidation	     2-15
         2.3.4  Solution-Phase Chemical Reactions	     2-22
                2.3.4.1  S(IV)-Q2 - H20 System	.	     2-23
                2.3.4.2  S(IV) - Catalyst - 02 - H20 System	     2-27
                2.3.4.3  S(IV) - Carbon Black - 02 - H20.......	     2-35
                2.3.4.4  S(IV) - Dissolved Oxidants - H20	     2-35
                2.3.4.5  The Influence of Ammonia.	     2-37
         2.3.5  Surface Chemical Reactions	     2-38
         2.3.6  Estimates of S02 Oxidation		     2-40
    2.4  PHYSICS AND CHEMISTRY OF PARTICULATE MATTER	     2-41
         2.4.1  Definitions	     2-42
         2.4.2  Physical Properties of Gases and Particles	     2-45
                2.4.2.1  Physical Properties of Gases	     2-4S
                2.4.21.2  Physical Properties of Particles	     2-46
         2.4.3  Dynamics of Single Particles	     2-60
         2.4.4  Formation and Growth of Particles	     2-62
                2.4.4.1  Growth Dynamics	     2-65
                2.4.4.2  Sulfuric Acid - Water Growth Dynamics..	     2-67
                2.4.4.3  Dynamics of Growth by Chemical Reaction	     2-67
                2.4.4.4  Dynamics of Desorption	     2-68
         2.4.5  Characterization of Atmospheric Aerosol	     2-69
                2.4.5.1  Distribution...	     2-69
                2.4.5.2  Composition of Particles	     2-75
         2.4.6  Particle-Size Spectra Evolution.	     2-80
                2.4.6.1  General Dynamics Equation (GDE)	     2-80
                2.4.6.2  Application of the GDE.	     2-81
    2.5  REFERENCES	     2-86

3.  TECHNIQUES FOR THE COLLECTION AND ANALYSIS OF SULFUR OXIDES,
    PARTICULATE MATTER, AND ACID PRECIPITATION	      3-1
    3.1  INTRODUCTION	      3-1
    3.2  MEASUREMENT TECHNIQUES FOR SULFUR DIOXIDE	      3-2
         3.2.1  Introduction	       3-2
         3.2.2  Manual Methods.	       3-2

-------
                             CONTENTS (continued)
                                                                          Page
                3.2.2.1  Sample Collection	        3-2
                3.2.2.2  Calibration	        3-3
                3.2.2.3  Measurement Methods	        3-4
         3.2.3  Automated Methods.	       3-12
                3.2.3.1  Sample Collection	       3-12
                3.2.3.2  Calibration	       3-12
                3.2.3.3  Measurement Methods	       3-13
                3.2.3.4  EPA Designated Equivalent Methods	       3-17
         3.2.4  Summary	       3-21
    3.3  PARTICULATE MATTER (PM).	       3-24
         3.3.1  Introduction	       3-24
         3.3.2  Gravimetric PM Mass Measurements	       3-30
                3.3.2.1  Filtration Samplers	       3-32
                3.3.2.2  Impactor 'Samplers  	       3-49
                3.3.2.3  Dustfall Sampling	       3-54
         3.3.3  Nongravimetn'c Mass Measurements	       3-54
                3.3.3*. 1  Filtration and Impaction Samplers	       3-54
                3.3.3.2  In Situ Analyzers.	       3-60
         3.3,4  Particle Composition..		      3-62
                3.3.4.1  Analysis of Sulfates	      3-63
                3.3.4.2  Ammonium and Gaseous Ammonia Determination..      3-70
                3.3.4.3  Analysis of Nitrates	      3-71
                3.3.4.4  Analysis of Trace Elements	      3-75
                3.3.4.5,  Analysis of Organic Compounds	      3-79
                3.3.4.0  Analysis of Total Carbon and Elemental
                         Carbon	      3-80
         3.3.5  Particle Morphology Measurements	      3-81
         3.3.6  Intercomparison of Particulate Matter Measurements...      3-81
         3.3.7  Summary	      3-83
    3.4  MEASUREMENT TECHNIQUES FOR ACIDIC DEPOSITION.	»	....      3-85
         3.4.1  Introduction	      3-85
         3.4.2  U.S. Precipitation Studies	      3-86
         3.4.3  Analytical Techniques	      3-8i
                3.4.3.1  Introduction	      3-89
                3.4.3.2  Analysis of Acidic Deposition Samples	      3-89
         3.4.4  Interlaboratory Comparisons	      3-93
    3.5  REFERENCES...,	      3-96

    APPENDIX 3-A.....	.........		     3-120

4.  SOURCES AND EMISSIONS 	     4-1
    4.1  INTRODUCTION	,	     4-1
    4.2  OATA SOURCES AND ACCURACY	     4-2
    4.3  NATURAL SOURCES AND EMISSIONS	     4-3
         4.3.1  Terrestrial Dust.		.........     4-4
         4.3.2  Sea Spray		....     4-7
         4.3.3  Biogenic Emanations	     4-7
         4.3.4  Volcanic Emissions......			     4-9
         4.3.5  Wildfires	    4-10
    4.4  MANMADE SOURCES AND EMISSIONS	    4-11
         4.4.1  Historical Emission Trends	    4-11
         4.4.2  Stationary Point Source Emissions.	    4-13
                                     VI

-------
                         CONTENTS (continued)
                                                                      Page
            4.4.2.1  Fuel Combustion	     4-24
            4.4.2.2  Industrial Processes	     4-27
     4.4.3  Industrial Process Fugitive Particulate Emissions	     4-30
     4.4.4  Nonindustrial Fugitive Particulate Emissions	      4-33
     4.4.5  Transportation Source Emissions	     4-35
4.5  SUMMARY	     4-36
4,6  REFERENCES	     4-38

ENVIRONMENTAL CONCENTRATIONS AND EXPOSURE	      5-1
5.1  INTRODUCTION	      5-1
5.2  AMBIENT MEASUREMENTS OF SULFUR DIOXIDE.	      5-2
     5.2.1  Monitoring Factors.	      5-4
     5.2.2  Sulfur Dioxide Concentrations	      5-5
     5.2.3  Sulfur Dioxide Concentration by Site and Region	      5-7
            5.2.3.1  Analyses by Various Site Classifications	      5-7
          '  5.2.3.2  Regional Comparisons	      5-7
     5.2.4  Peak Localized Sulfur Dioxide Concentrations	     5-12
            5.2.4.1  1978 Highest Annual Average Concentrations...     5-12
            5.2.4.2  1978 Highest Daily Average Concentrations	     5-12
          *  5.2.4.3  Highest 1-Hour Sulfur Dioxide Concentra-
                     tions-1978 National Aerometric Data Bank
                     (NADB) Data	     5-12
     5.2.5  Temporal Patterns in Sulfur Dioxide Concentrations....     5-13
            5.2.5.1  Diurnal Patterns	     5-13
            5.2.5.2  Seasonal Patterns	     5-16
            5.2.5.3  Yearly Trends	     5-16
5.3  AMBIENT MEASUREMENTS OF SUSPENDED PARTICULATE MASS	     5-22
     5.3.1  Monitoring Factors	     5-23
            5.3.1.1  Sampling Frequency.,	     5-23
            5.3.1.2  Monitor Location	     5-27
     5.3.2  Ambient Air TSP Values	     5-27
     5.3.3  TSP Concentrations by Site and Region	     5-30
            5.3.3.1  TSP by Site Classifications	     5-31
            5.3.3.2  Intracity Comparisons	     5-31
            5.3.3.3  Regional Differences in Background
                     Concentrations	 — .     5-33
            5.3.3.4  Peak TSP Concentrations	     5-33
     5.3.4  Temporal Patterns in TSP Concentrations	     5-35
            5.3.4.1  Diurnal Patterns	     5-35
            5.3.4.2  Weekly Patterns	     5-35
            5.3.4.3  Seasonal Patterns	     5-37
            5.3.4.4  Yearly Trends	     5-37
5,4  SIZE OF ATMOSPHERIC PARTICLES	     5-46
     5.4.1  Introduction	     5-46
     5.4.2  Size Distribution of Particle Mass	     5-47
5.5  FINE PARTICLES IN AIR	     5-57
     5.5.1  Sulfates	,	     5-58
            5.5.1.1  Spatial and Temporal Variations.	     5-58
            5.5.1.2  Urban Variations	     5-64
     5.5.2  Nitrates	     5-73
     5.5.3  Carbon and Organics	     5-77
                                 vi i

-------
                             CONTENTS (continued)
                                                                          Pa
                5.5,3.1  Physical Properties of Participate Qrganies..
                5.5.3.2  Carbon and Total Organic Mass....	     5-79
                5.5.3.3  Chemical Composition of Participate Organic
                         Matter	     5-85
         5.5.4  Metallic Components of Fine Particles.	     5-87
                5.5.4.1  Lead......	     5-92
                5.5.4.2  Vanadium, Nickel, and Other Metals	     5-92
         5.5.5  Acidity of Atmospheric Aerosols	     5-96
    5.6  COARSE PARTICLES IN AIR.	     5-99
         5.6.1  Introduction	     5-99
         5.6.2  Elemental Analysis of Coarse Particles	,..    5-100
         5.6.3  Evidence from Microscopical Evaluation of Coarse
                Particles			    5-103
         5.6.4  Fugitive Dust	    5-106
         5.6.5  Summary	    5-109
    5.7  SOURCE-APPORTIONMENT OR SOURCE-RECEPTOR MODELS	    5-109
    5.8  FACTORS INFLUENCING EXPOSURE.	    5-115
         5.8.1  Introduction...	    5-115
         5.8.2  Indoor Concentrations of Sulfur Dioxide	    5-117
         5.8.3  Particle Exposures Indoors	    5-118
                5.8";3.1  Introduction.	     5-118
                5.8.3.2  Coarse-Particle Concentrations Indoors	     5-122
                5.8j3.3  Fine Particles Indoors	     5-127
         5.8,4  Monitoring and Estimation of Personal Exposures	     5-131
    5.9  SUMMARY OF- ENVIRONMENTAL CONCENTRATIONS AND EXPOSURE	    5-136
    5.10 REFERENCES	    5-139

6.  ATMOSPHERIC TRANSPORT, TRANSFORMATION, AND DEPOSITION	      6-1
    6.1  INTRODUCTION	      6-1
    6.2  CHEMICAL TRANSFORMATION PROCESSES	      6-1
         6.2.1  Chemical Transformation of Sulfur Dioxide and
                Particulate Matter	      6-3
         6.2.2  Field Measurements on the Rate of Sulfur Dioxide
                Oxidation	      6-3
    6.3  PHYSICAL REMOVAL PROCESSES	      6-6
         6.3.1  Dry- Deposition.	      6-7
                6.3.1.1  Sulfur Dioxide Dry Deposition	      6-8
                6.3.1.2  Particle Dry Deposition	     6-10
         6.3.2  Precipitation Scavenging	     6-17
                6.3.2.1  Sulfur Dioxide Wet Removal	     6-19
                6.3.2.2  Particle Wet Removal	     6-20
    6.4  TRANSPORT AND DIFFUSION	     6-23
         6.4.1  The Planetary Boundary Layer	     6-23
         6.4.2  Horizontal Transport and Pollutant Residence Times	     6-27
    6.5  AIR QUALITY SIMULATION MODELING	     6-30
         6.5.1  Gaussian Plume Modeling Techniques	     6-31
         6.5.2  Long-Range Air Pollution Modeling		     6-32
         6.5.3  Model Evaluation and Data Bases		     6-36
         6.5.4  Atmospheric Budgets.......	     6-37
    6.6  SUMMARY			     6-38
    6.7  REFERENCES	     6-39
                                    viii

-------
                             CONTENTS (continued)
7.   ACIDIC DEPOSITION	
    7.1  INTRODUCTION	
         7.1.1  Overview of the Problem	
         7.1.2  Ecosystem Dynamics	
    7.2  CAUSES OF ACIDIC PRECIPITATION.	
         7.2.1  Emissions of Sulfur and Nitrogen Oxides....	
         7.2.2  Transport of Nitrogen and Sulfur Oxides	
         7.2,3  Formation		
                7.2.3.1  Composition and pH of Precipitation	
                7.2.3.2  Geographic Extent of Acidic Precipitation	
         7.2.4  Acidic Deposition	
    7.3  EFFECTS OF ACIDIC DEPOSITION	,	
         7.3.1  Aquatic Ecosystems....	
                7.3.1.1  Acidification of Lakes and Streams..	
                7.3.1.2  Effects on Decomposition		
                7.3.1.3  Effect on Primary Producers and Primary
                         Productivity	
                7.3.1.4  Effects on Invertebrates	
                7.3.1.5  Effects on Fish	
                7.3.1.6  Effects on Vertebrates other than Fish..	
         7.3.2  Terrestrial Ecosystems	
                7.3.2.1  Effects on Soils	
                7.3.2.2  Effects on Vegetation,	
                7.3.2.3  Effects on Human Health	
               ' 7.3.2.4  Effects of Acidic Precipitation on Materials,
    7.4  ASSESSMENT OF SENSITIVE AREAS	
         7.4.1  Aquatic Ecosystems	
         7.4.2  Terrestrial Ecosystems	
    7.5  SUMMARY	
    7.6  REFERENCES	

-------
                             CONTENTS (continued)



                                    FIGURES

FIGURE

2-1  • The global sulfur cycle, showing the major reservoirs, pathways,
      and forms of occurrence of sul fur ................. , ..............     2-4
2-2   Interrelations of pathways, processes, and properties of
      sulfur oxides and particulate matter and effects... ---- ..... .......     2-7  !
2-3   The distribution of species for the S02 • H20-HS03-SO|~ system
      as a function of pH.  Also, the ratio of the concentrations of
             to the total quantity dissolved in water is shown.,...' .....  2-11
2-4   Schematic of the polluted atmospheric photooxidation cycle ........   2-18
2-5   The theoretical rate of reaction (percent per hour) of various
      free-radical species on S02 is shown for a simulated sunlignt-
      irradiated (solar zenith angle of 40°) polluted atmosphere..... ___   2-20
2-6   Percentage conversion at midday of sulfur dioxide to sul fate
      by HO and by HO, H02, and CH302 radicals as a function of °N
      1 ati tude i n summer and wi nter ....................... . ......... ....   2-21

2-7   Solubility diagram for the H+-NHt-SO|~-H20 system at
      equi 1 ibrium (30°C) ..... .... ..... . ................ . ...............    2-49

2-8   Growth of H+-NHt~SQi~ particles as a function of RH ....... .......    2-50
2-9   Condensational growth and evaporation of (NH4)2S04 particles as
      a function of relative humidity at 25°C ..........................    2-52
2-10  The equilibrium size of sul f uric acid solution droplets as a
      function of relative humidity ....................................    2-54
2-11  NHs and HN03 partial pressures as a function of droplet' s nitrate
      (C..Q-) and sulfate (Ccg2~) concentrations at 85 percent relative

      humidity, 25°C .................. .... ....... . ____ . ......... . ......    2-58
2-12  Frequency plots of number, surface, and volume distributions for
      1969 Pasadena smog aerosol ............................ . ..........    2-71
2-13a Idealized size distribution for particles found in typical urban
      aerosols (mainly from anthropogenic sources) under varying
      weather conditions ......................... . . ....................    2-73
2~13b Idealized size distribution for atmospheric particles from
      anthropogenic sources ........... ........... ......................    2-73
2-13c Idealized size distribution for atmospheric particles from
      natural sources in a marine setting ..............................    2-74
2-13d Idealized size distribution for atmospheric particles from
      natural sources in a continental setting ......... ................    2-74
2-14  Idealized representation of typical fine- and coanse-particle
      mass and chemical composition distribution in an urban aerosol —    2-76

3-1   Respiratory deposition models used as patterns for sampler •
      cutpoints ............................................... : T.~ ."..'...    3-25
3-2   Plots illustrating the relationship of particle number,
      surface area, and volume distribution as a function of
      particle size ...................... . ................... -. .........    3-27
3-3   Typical ambient mass distribution data for particles
      up to 200 pm ..................... . ...............................    3-28
3-4   Sampling effectiveness of a Hi-Vol sampler as a function of

-------
                              CONTENTS (continued)
 Figure                                                                 ,   Page
 3-5   Sampling effectiveness  of the dichotomous  sampler inlet as  a
       function of windspeed	,	    3-33
 3-6   Sampling effectiveness  of the Wedding IP inlet.	    3-34
 3-7   Sampling effectiveness  of UM-LBL IP in-let	    3-35
 3-8   Effect of sampler flowrate on the performance of a Hi-Vol for
       30 pm particles  at a windspeed of B km/hr		    3-38
 3-9   Separator efficiency and wall losses of the dichotomous
       sampler at 2.5 jjm	    3-41
 3-10  Sampling effectiveness  for the 3.5 jam cutpoint CHESS
       cyclone sampler	    3-43
 3-11  Fraction of methylene blue particles deposited in a cyclone
       sampler as a function of the aerodynamic particle diameter,......    3-45
 3-12  Sampling effectiveness  for the size-selective inlet Hi-Vol
       sampl er		    3-46
 3-13  Effect of windspeed upon cutpoint size of  the size-selective
       inlet	    3-47
 3-14  Effect of sampler flowrate on the sampling effectiveness of
       the size-selective inlet Hi-Vol  for a particle size of
       15.2 pm and windspeed of 2 km/hr	    3-48
 3-15  An example of mass size distribution obtained using a cascade
       impactor	    3-50
'3-16  Fractional particle collection of the CHAMP fractionator inlet
       at a sampler flowrate of 1133 liters/min under static windspeed
       conditions		    3-52
 3-17  Efficiency of the single impaction stage of the CHAMP Hi-Vol
       sampl er	    3-53
 3-18  Sampling effectiveness  of the inlet alone  and through the
       entire flow system of the British Smoke Shade sampler	    3-56
 3-19  Response of a Piezoelectric Microbalance to relative humidity
       for various particle types	    3-60
 3-20  Light scattering and absorption expressed  per unit volume of
       aerosol	.	,		    3-61

 5-1   Distribution of  annual  mean sulfur dioxide concentrations across
       an urban complex, as a  function of various spatial scales	     5-3
 5-2   Histogram delineating annual average sulfur dioxide concentrations
       for valid continuous sampling sites in the United States in 1978.     5-6
 5-3   Characterization of 1974-76 national SQ^ status is shown by
       second highest 24-hour  average concentration	    5-10
 5-4   Composite diurnal pattern of hourly sulfur dioxide concentrations
       are shown for Watertown, Massachusetts, for December 1978........    5-14
 5-5   Monthly means of hourly sulfur dioxide concentrations are shown
       for St. Louis (city site 26-4280-007, "Broadway & Hurck") for
       February 1977 and 1978	     5-15
 5-6   Monthly means of hourly sulfur dioxide concentrations are
       shown for Steubenville, Ohio (NOVAA site 36-6420-012) for
       June 1976 and July 1977	     5-17
 5-7   Seasonal variations in  S02 levels are shown for Steubenville,
       St. Louis, and Watertown.	     5-18

-------
                             CONTENTS (continued)
Figure                                                                    Page
5-8   Annual average sulfur dioxide concentrations are shown for 32
       urban NASN stations	    5-19
5-9   Nationwide trends in annual average sulfur dioxide concentrations
      from 1972 to 1977 are shown for 1233 sampling sites	    5-20
5-10  Distribution shows the number of TSP observations per valid site
      in 1978; total of 2882 sites	    5-24
5-11  The 95 percent confidence intervals about an annual mean TSP
      concentration of 75 ug/m3 is shown for various sampling
      frequencies	    5-26
5-12  Distribution of mean and 90th percentile TSP concentrations is
      shown for valid 1978 sites	    5-28
5-13  Histogram of number of sites against concentration shows that
      over one-third of the sites had annual mean concentrations
      between 40 and 60 ug/m3	    5-29
5-14  Histogram of mean TSP levels by neighborhood shows lowest levels
      in residential areas, higher levels in commercial areas, and
      highest levels in industrial areas	,	 —	    5-32
5-15  Average estimated contributions to nonurban levels in the East,
      Midwest, and West are most variable for transported secondary
      and continental sources	    5-34
5-16  Severity of TSP peak exposures is shown on the basis of the
      90th percentile concentration.  Four AQCR's did not report	    5-36
5-17  Seasonal variations in urban, suburban, and rural areas
      for four size ranges of particles	    5-38
5-18  Monthly mean TSP concentrations are shown for the Northern Ohio
      Valley Air Monitoring Headquarters, Steubenville, Ohio.  No
      cl ear seasonal pattern 1 s apparent.	    5-39
5-19  Annual geometric mean TSP trends are shown for selected NASN
      sites	    5-40
5-20  (Top) Nationwide trends in annual mean total suspended
      particulate concentrations from 1972 to 1977 are shown for
      2707 sampling sites.  (Bottom) Conventions for box plots	    5-42
5-21  Regional trends of annual mean total suspended particulate
      concentrations, 1972-1977, Eastern states	    5-44
5-22  Regional trends of annual mean total suspended particulate
      concentrations, 1972-1977, Western states	    5-45
5-23  Linear-log plot of the volume distributions for the four
      background distributions	    5-49
5-24  Linear-log plot of the volume distributions for two urban
      aerosols and a typical distribution measured in the Labadie
      coal-fired power plant plume near St. Louis.  Size distri-
      butions measured above a few hundred meters above the
      ground generally have a rather small coarse particle mode	    5-50
5-25  Incursion of aged smog from Los Angeles at the Goldstone
      tracking station in the Mojave Desert in California	    5-51
5-26  Sudden growth of the coarse particle mode due to local dust
      sources measured at the Hunter-Liggett Military Reservation
      in California.  This shows the independence of the
      accumulation and coarse particle mode.	    5-52
5-27  Inhalable particle network sites established as of
      March 19, 1980	    5-54
5-28  Contour maps of sulfate concentrations for 1974 are shown for:
      (a) annual average; (b) winter average; (c) summer average	    5-59

-------
                             CONTENTS (continued;
Figure                                                                    Page
5-29  Intensive Sulfate Study area in eastern Canada shows the
      geometric mean of the concentration of particulate soluble
      sulfate during the study period.  Units are micrograms of
      sulfate per cubic meter	,	     5-61
5-30  Map of SURE regions shows locations of ground measurement
      stations	,	     5-62
5-31  Cumulative plots show the frequency of sulfate concentrations
      in the SURE region on the basis of the 1974-75 historical data...     5-63
5-32  Map shows the spatial distribution of number of days per month
      that the sulfate concentration equaled or exceeded 10 MS/1"3	     5-65
5-33  1977 seasonal patterns of S02 emissions and 24-hr average S02
      and S04 ambient levels in the New York area are normalized to
      the annual average values	      5-66
5-34  Monthly variation in monthly mean of 24-hour average sulfate
      concentration at downtown Los Angeles is compared with monthly
      mean 1973 Los Angeles County power plant S02 emissions	     5-67
5-35  Map shows annual mean 24-hr average sulfate levels in micrograms
      per cubic meter in the New York area, based on 1972 data from
      Lynn et al. (1975).  Squares are locations of three CHAMP site
      stations.  The fourth station is at the tip of Long Island
      about 160 km from Manhattan	     5-70
5-36  Distribution of annual average sulfate concentration in
      micrograms per cubic meter in the greater Los Angeles area
      based on 1972-1974 data....				     5-71
5-37  Map shows U.S. mean annual ambient nitrate levels in micrograms
      per cubic meter. ,...*....	     5-74
5-38  Mean nitrate concentrations in micrograms per cubic meter at
      nonurban sites in the U.S. based on valid annual average from
      1971 through 1974......		     5-75
5-39  Calculated distributions of aerosol constituents for two aerosol
      samples taken in the Los Angeles Basin		     5-82
5-40  Benzo(a)pyrene seasonality and trends (1966-75) in the
      50th and 90th percentiles for 34 NASN urban sites	     5-84
5-41  Seasonal patterns and trends in quarterly average urban lead
      concentrations	     5-94
5-42  Regional trends in the 90th percentile of the annual averages
      for vanadium	     5-95
5-43  Seasonal variation in quarterly averages for nickel and
      vanadium at urban sites in the northeast	     5-97
5-44  Elemental compositions of some coarse particle components	    5-102
5-45  Diurnal variation- of particle concentrations and Plymouth
      Avenue traffic volume at Falls River, Mass., during March
      through June 1979 (weekdays only), shows contribution from
      reentrained particles	    5-108
5-46  Types of receptor source apportionment models	    5-110
5-47  Source contributions at RAPS sites estimated by chemical element
      balance are illustrated	    5-112
5-48  Monthly averages of size fractionated Denver aerosol mass and
      composition for January and May, 1979	    5-113
5-49  Aerosol source in downtown Portland, annual stratified
      arithmetic average.  Does not include the 17%, on the average,
      of material collected with the standard Hi-Vol sampler which
      was not collected and characterized with the ERT-TSP sampler	    5-114
                                     xm

-------
                             CONTENTS (continued)
                                                                          Page
     "Retention of Au198 labeled FegOa particles from human lungs—
      comparison of 9 non-smoking subjects with three smoker
      subjects	    5-116
5-51  Annual S02 concentrations averaged across each community's
      indoor and outdoor network (May 1977-April 1978).	    5-119
5-52  Monthly mean S02 concentrations averaged across Watertown's
      indoor and outdoor network (November 1976-April 1978)	    5-120
5-53  Monthly mean S02 concentrations averaged across Steubenville's
      indoor and outdoor network (November 1976-April 1978)	    5-121
5-54  Annual respirable particulate concentrations averaged across
      each community's indoor and outdoor network (May 1977-
      Apri 1 1978)	    5-130
5-55  An example of personal exposure to respirable particles	    5-132
5-56  Normalized distribution of personal (12-hour) exposure samples
      ((jm/m3) for non-smoke exposed and smoke exposed samples	    5-134
5-57  Daily mean indoor/outdoor and personal concentrations (ng/m3)
      of respirable particles.  Daily means averaged over 24 homes
      and outdoor locations and up to 46 personal samples.   Samples
      col 1 ected during May and June 1979	    5-135

6-1   Pathway processes of airborne pollutants.	      6-2
6-2   Predicted deposition velocities at 1 |j for 1^=30 cm s   and
                                              -3
      particle densities of 1,4, and 11.5 g cm	     6-16
6-3   Basic factors influencing precipitation scavenging	     6-18
6-4   Relationship between rain scavenging rates and particle
      size.	     6-24
6-5   Percentages of aerosol particles of various sizes removed by
      precipitation scavenging	     6-25
6-6   Estimated residence times for select pollutant species and
      their associated horizontal transport scale	     6-29
6-7   Trajectory modeling approaches are shown	     6-34

7-1   Schematic representation of the nitrogen cycle, emphasizes human
      activities that affect fluxes of nitrogen	     7-10
7-2   Law of tolerance	     7-12
7-3   Historical patterns of fossil fuel consumption in the
      United States	     7-15
7-4   Forms of coal usage in the United States	     7-16
7-5a  Trends in emissions of sulfur dioxide	     7-17
7-5b  Trends in emissions of nitrogen oxides	     7-17
7-6   Characterization of U.S. SO  emissions density by state	     7-18
7-7   Characterization of U.S. NO  emissions density by state	     7-19
7-8   Trends in mean annual concentrations of sulfate, ammonium,
      and nitrate in precipitation	     7-24
7-9   Comparison of weighted mean monthly concentrations of sulfate in
      incident precipitation collected in Walker Branch Watershed,
      Tennessee (WBW) and four MAP3S precipitation chemistry monitoring
      stations in New York, Pennsylvania, and Virginia	     7-27
7-10  Seasonal variations in pH (A) and ammonium and nitrate
      concentrations (B) in wet-only precipitation at Gainesville,
      Florida	     7-28
                                   xiv

-------
                             CONTENTS (continued)
      Seasonal variations of precipitation pH in the New York
      Metropol Han Area	     7-31
7-12  History of acidic precipitation at various sites in and
      adjacent to State of New York	     7-32
7-13  pH of rain sample, as measured in the laboratory,  used in
      combination with the reported amount of precipitation	     7-35
7-14  Annual mass transfer rates of sulfate expressed as a percentage
      of the estimated total annual flux of the element to the forest
      floor beneath a representative chestnut oak stand	     7-37
7-15  Schematic representation of the hydrogen ion cycle	     7-39
7-16  pH and calcium concentrations in lakes in northern and northwestern
      Norway sampled as part of the regional survey of 1975, in lakes  in
      northwestern Norway sampled in 1977 (o) and in lakes in southernmost
      and southeastern Norway sampled in 1974 (o)	    7-43
7-17  The pH value and sulfur loads in lake waters with extremely
      sensitive surroundings (curve 1) and with slightly less
      sensitive surroundings (curve 2)	    7-44
7-18  Total dissolved Al as a function of pH level in lakes in
      acidified areas in Europe and North America.,	•    7-45
7-19  pH levels in Little Moose Lake, Adirondack region of New York
      State, at a depth of 3 meters and at the lake outlet	    7-47
7-20  Numbers of phytoplankton species in 60 lakes having different
      pH values on the Swedish west coast, August 1976 are compared	    7-51
7-21  Percentage distribution of phytoplankton species and their
      biomasses. September 1972, West Coast of Sweden.		...    7-52
7-22  The number of species of crustacean zooplankton observed in
      57 lakes during a synoptic survey of lakes in southern Norway	    7-56
7-23  Frequency distribution of pH and fish population status in
      Adirondack Mountain lakes greater than 610 meters elevation	    7-60
7-24  Frequency distribution of pH and fish population status in 40
      Adirondack lakes greater than 610 meters elevation,
      surveyed during the period 1929-1937 and again in 1975	     7-61
7-25  Norwegian salmon fishery statistics for 68 unacidified and 7
      acidified rivers	     7-62
7-26  Showing the exchangeable ions of a soil with pH 7, the soil

      solution composition, and the replacement of Na  by H  from
      acid rain	     7-71
7-27  Regions in North America with lakes that are sensitive to acid-
      ification by acid precipitation by virtue of their underlying
      bedrock characteristics	     7-96
7-28  Soils of the eastern United States sensitive to acid rainfall
      are mapped.	    7-100
                                  xv

-------
                             CONTENTS (continued)
                                    TABLES
Table                                                                   Page

 2-1  Estimates of environmental sulfur annual fluxes (tg/year)	    2-5
 2-2  Characteristic times and lengths for observation of effects....    2-6
 2-3  Dilute sulfur dioxide-water system.	   2-10
 2-4  Relative strengths of acids in water solution (25°C)	   2-13
 2-5  Rate constants for hydroxyl,  peroxyl, and methoxyl radicals	   2-15
 2-6  Investigations of S02 - 02 aqueous systems	   2-24
 2-7  Values of k  and kh for reaction type 1	   2-26
 2-8  Values of 1C for reaction types 2	   2-27
 2-9  Investigations of S02 - manganese -  02 aqueous system	   2-29
2-10  Rate expression for the manganese-catalyzed oxidation	   2-30
2-11  Investigations of SQ2 ~ iron - 02 aqueous system.	   2-31
2-12  Rate expression for the iron-catalyzed oxidation	   2-32
2-13  Investigations of S02 - copper - 02  aqueous systems	....	   2-34
2-14  Estimates of S02 oxidation rates in  well-mixed troposphere	   2-40
2-15  Estimate of global tropospheric particulate matter production
      rates	   2-43
2-16  Particle shapes and source types	   2-46
2-17  Deliquescence and efflorescence points of salt particles...	   2-51
2-18  Sulfuric acid solution values (25°C)	   2-55
2-19  Conditions for the single-particle regime		   2-60
2-20  Mass transport parameters for air	   2-63
2-21  Dependence of particle behavior on air temperature, pressure,
      and viscosity	   2-64
2-22  Classification of major chemical species associated with
      atmospheric particles	   2-77
2-23  Application of GDE to describe particle size evaluation	   2-82

 3-1  Temperature effect on collected S02-TCM samples (EPA reference
      method)	    3-6
 3-2  Performance specifications for EPA equivalent methods for S02
      (continuous analyzers)	   3-18
 3-3  List of EPA designated equivalent methods for S02 (continuous
      analyzers)	   3-19
 3-4  Interferent test concentrations (parts per million) used in the
      testing of EPA equivalent methods for S02	   3-20
 3-5  Comparison of EPA designated equivalent methods for S02
      (continuous analyzers)	   3-22
 3-6  Recommended physical /chemical parameters for analysis	   3-90
 3-7  Results of WMO intercomparisons on synthetic precipitation
      samples.	   3-94
 3-8  Coefficients of variation of WMO intercomparisons on
      synthetic precipitation samples	   3-95

 4-1  Two EPA estimates of 1977 emissions  of particulate matter,
      and sulfur oxides (106 metric tons per year)	    4-2
 4-2  Summary of natural source particulate and sulfur emissions	    4-5
 4-3  Aerosol enrichment factors relative  to Al	,	    4-6
 4-4  Summary of estimated annual manmade  emissions (1978)	   4-11
 4-5  (a) National estimates of particulate emissions (106 metric
      tons per year)	   4-13
      (b) National estimates of SO  emissions (106 metric tons per
      year)	   4-13
                                   xvi

-------
                              CONTENTS (continued)
 Tab! e                                                                  Page
-"4-6  1978 estimates of participate and sulfur oxide emissions
       from stationary point sources	   4-14
  4-7  State-by-State listing of total  particulate and sulfur oxide
       emissions from stationary point  sources (1977), population,
       and density factors	   4-16
  4~8  Examples of uncontrolled particulate emission characteristics..   4-20
  4-9  Size-specific particulate emissions from coal-fired boilers.  .,   4-23
 4-10  Trace element air emissions vs.  solid waste:   percent from
       conventional  stationary fuel  combustion sources, and total
       (metric tons  per year)	   4-25
 4-11  Uncontrolled  industrial process  fugitive particulate emissions.   4-31
"4-12  Estimated annual particulate emissions from nonindustrial
       fugitive sources....	   4-34
 4-13  Estimated particle size distributions for several
       nonindustrial fugitive source categories in California's
       south coast air basin	   4-34

 5-1   Crosstabulation of annual mean SQ% concentration by method
       (bubbler or continuous) for population-oriented and for
       source-oriented center-city sites.	,....     5-8
 5-2   Continuous S02 monitor results by region, ug/m3	     5-9
 5-3   Eleven S02 monitoring sites with the highest annual mean
       concentrations in 1978 (valid continuous sites only)	    5-11
 5-4   Comparison of frequency distribution of S02 concentration (ppm)
       during 1962-67 and during 1977.	    j-',\.
 5-5   Range of annual geometric mean concentrations in areas with
       high TSP concentrations in 1977	   5-33
 5-6   Regional summaries of TSP values from valid monitors	   5-43
 5-7   Fine and coarse aerosol concentrations from some urban
       measurements  compared to clean areas	   5-53
 5-8   Fine fraction and coarse fraction dichotomous sampling by
       Environmental Science Research Lab, US EPA in four locations...   5-55
 5-9   Recent dichotomous sampler and TSP data from selected sites—
       arithmetic averages	   5-56
 5-10  Some characteristics of pollution in the New York and
       Los Angeles areas	^ .   .   5-M
 5-11  Primary ranking of variables for correlating airborne S0|
       in two cities based on a stepwise linear reg--p . , ..A *v;
       15 variables  from CHAMP and related monitor ing b ,«.,.,
 5-12  Typical values of aerosol concentration for different
       geographic areas (annual averages)	   5-81
 5-13  Annual averages of organic fractions in TSP, New York City,
       dispersion normalized	   5-83
 5-14  Composition of the organic fraction of airborne PM
       collected in Detroit	   5-85
 5-15  Comparison of urban and nonurban annual average concentrations
       for selected metals, 1970-74 (ug/m3)	   5-88
 5-16  Ratios of urban (U) to suburban (S) concentrations in air,
       Cleveland, Ohio, area	   5-09
 5-17  Correlations of chemical content with particle sive   ,   ...  ..   S-OO
 5-18  Particulate analyses from selected urban locations	
 5-19  Trends in reported urban metal concentrations" and their possible
       causes	   5-93
                                   xvi i

-------
                             CONTENTS (continued)
Table                                                                 Page
5-20  Coarse particle silicon, aluminum, calcium, and iron	  5-101
5-21  Relative amounts of fine, coarse, and super-coarse particles at
      selected sites	  5-104
5-22  Fourteen-city study - microscopical identification of coarse
      particles collected in urban atmospheres	  5-105
5-23  Summary of indoor/outdoor (I/O) PM monitoring studies by
      method	  5-123
5-24  Measurements in principal room of study	  5-128
5-25  Measurements in various closed rooms..	  5-128
5-26  Respirable particulate concentrations outdoors and indoors by
      amount of smoking	,	  5-129

 6-1  Field measurements on the rates of S02 oxidation in plumes.....    6-4
 6-2  Average dry deposition velocity of S02 by surface type	    6-9
 6-3  Laboratory measurements of deposition velocities of particles..   6-11
 6-4  Field measurements of deposition velocities of particles	...   6-13
 6-5  Predicted particle deposition velocities		.	   6-17
 6-6  Field measurements of scavenging coefficients of particles	   6-21
 6-7  Summary of long range transport air pollution models	   6-35

 7-1  Composition of ecosystems.	   7-8
 7-2  Mean pH values in the New York metropolitan area		  7-30
 7-3  Storm type classification	  7-30
 7-4  Chemical composition (Mean ± standard deviation) of acid lakes
      (pH <5) acidic precipitation (pH <4.5), and of soft-water lakes
      in areas not subject to'highly acidic precipitation
      (pH >4.8)	  7-41
 7-5  pH levels identified in field surveys as critical to
      long-term survival of fish populations	  7-63
 7-6  Changes in aquatic biota likely to occur with increasing
      acidity	  7-67
 7-7  Summary of effects on aquatic organisms associated with a
      range in pH	-.	  7-68
 7-8  Potential effects of acid precipitation on soils	  7-72
 7-9  Types of direct, visible injury reported in response to acidic
      wet deposition	  7-81
7-10  Thresholds for visible injury and growth effects associated with
      experimental studies of wet deposition of acidic substances	  7-84
7-11  Lead and copper concentration and pH of water from pipes
      carrying outflow from Hinckley Basin and Hanns and Steele
      Creek Basi n, near Amsterdam, New York	  7-91
7-12  Composition of rain and hoarfrost at Headingley, Leeds	  7-93
7-13  The sensitivity to acid precipitation based, on:   buffering
      capacity against pH-change, retention of H , and adverse
      effects on soils	  7-98
                                      xvm

-------
               2.   PHYSICS AND CHEMISTRY OF SULFUR OXIDES AND PARTICULATE MATTER

2.1  INTRODUCTION
     This chapter  presents the  current  state of  knowledge of  the  chemistry and  physics  of
sulfur oxides  (SO  )  and  particulate matter (PM) that  is  pertinent to tropospheric phenomena,
effects, and sampling  methodology.   The  1970 Air Quality Criteria for Sulfur Oxides (National
Air Pollution  Control  Administration,  1970)  and the 1969 Air Quality Criteria for Particulate
Matter (National Air  Pollution  Control  Administration, 1969) at the time of their publication
adequately described  the existing  knowledge  of the  ambient chemistry and  physics  of sulfur
oxides and particulate matter.  However,  significant progress has been made since that time  in
understanding  tropospheric  properties  and processes.   While this  chapter focuses  mainly  on
advances made in the past decade, earlier work is mentioned for the sake of comprehensiveness.
     This chapter is organized into three principal parts, as follows:
     A.  Atmospheric domain and processes
         --global  sulfur cycle
         --atmospheric sulfur cycle
         —pathways and processes
     B.  Physics and chemistry of sulfur oxides
         --gaseous physical properties
         —solution physical properties
         —gas-phase chemical reactions (elementary rate consists for S0« oxidation; influence
           of volatile organics and nitrogen oxides)
         —solution-phase  chemical  reactions (reactions  kinetics  for oxidation, by  0,,  H20P,
           and .catalysts; limitations of reported studies)
         —surface chemical reactions (metal  oxides and carbon)
     C.  Physics and chemistry of particulate matter
         —definitions of aerosol science terms
         —physical -properties of gases and particles (size, shape, density, morphology,  charg-
           ing, adhesion, vapor pressure; optics are discussed in Chapter 9)
         —dynamics of single particles (sedimentation, impaction, diffusion, electrodynamics)
         --formation  and  growth  of particles  (nucleation,  coagulation,   condensation,  gas-
           particle chemical reaction)
         --characterization of atmospheric aerosol (size, distribution by number, area,
           volume, and mass distributions; composition of fine and coarse particle mass
           fractions)
     Available evidence leads to the following conclusions:
     A.   The physics and chemistry of S0?
                                             2-1

-------
          i.   me  tnermoaynamtc properties,  molecular  structure  and bonding,  and electro-
magnetic  absorption spectra  in  air and  dissolved  in water are well  established,  except for
hygroscopic/deliquescent properties of internally mixed salts.
          2.   Of  the homogeneous  gas-phase  SOp  oxidation reactions,  only three  have  been
identified as likely being significant in the troposphere:
               a.   HO radical attack on S02
               b.   HOp radical attack on SOp
               c,   CHgO,, radical attack on SOp
          3.   The  auto-oxidation (uncatalyzed)  reaction  of SOp dissolved  in  liquid water is
too slow to be an important reaction in the troposphere.
          4.   The  Mn(II)-  and Fe(III)-catalyzed  oxidation of SOp  dissolved  in liquid water
may be  an important reaction  in  the troposphere.   However, there  is  serious  doubt regarding
the rate expression for th Mn(II)-catalyzed oxidation.
          5.   The  effectiveness  of Cu(II),  V(V),  V(IV), Ni(II), Zn(II),  and  Pb(II) as cata-
lysts for the oxidation of SOp dissolved in liquid water are unknown.
          6.   Lack of data  on the effectiveness of dissolved organics and of bicarbonate ion
(HCO») as  inhibitors  prevents the confident use of aqueous-phase SOp catalyzed oxidation rate
expressions for tropospheric model prediction.
          7.   Elemental carbon  (soot)  particles coated with an aqueous film may be important
catalysts for S0? oxidation in the troposphere.
          8.   The  reaction  rate  expressions of dissolved 03 and dissolved N0« with dissolved
SOp species are  known,  but these reactions  appear  to be ineffective for sulfate formation in
the troposphere.
          9.   The  rate expression  for dissolved HpOp and  dissolved SO,  species is known and
appears to be a highly effective reaction for sulfate formation in the troposphere.
         10.   SO,  reactions with  solid-particle  surfaces are  not effective  for sustained
  2-
SO* formation in the troposphere.
     B.    The physics and chemistry of particulate matter
          1.   The  physical  properties  of gases that affect aerosol  behavior  are well known.
          2.   The physical characteristics of tropospheric particles are highly variable, but
the physics through which they influence aerosol behavior is well known.
          3.   The  particle  mass  distribution  function  (AM/A  log- Diam.  v.  log  Diam.) for
tropospheric aerosols over  land  is  often multimodal.   The  fine particles (diameter less than
about 2 |jm) may  have two (or  more)  modes,  usually at about 0.02 urn and at about 0.2 pm.   The
coarse particles (greater than 2.5 pm) generally have one mode in the range 5 to 50 |jm.
          4.   The  mass  composition of  the  coarse  particles  is dominated  by minerals whose
direct source types are well-known.
                                             2-2

-------
          5.   The  mass composition  of the  fine particles  Is  dominated by  so|~,  organfcs,
elemental carbon (soot), NO- and NH, whose direct and indirect source types are quantitatively
well known.
                                                                2~                  -
          6.   The detailed chemical pathways for forming the SO, , organics, and NO- found in
the fine particles have not been established.
          7.   Strong acids are  often found in the  fine  mass fraction, while bases are found
in the coarse mass fraction.
          8.   The  molecular  composition of  the  organic components  (generally found  in  the
fine mass fraction) is not well characterized.
          9.   Water is the major constituent of the particle mass, but the deliquescence and
hygroscopic properties of mixed salts cannot be predicted reliably.
         10.   The  dynamics  of motion  of a particle with  a diameter of  less  than  10  \im can
best be  described  in  terms of the  physical  characteristics of the particle, the force fields
present, and  the motion  of the  suspending  gas.   However,  this ability  does  not adequately
extend to  larger particles,  especially in the presence  of  nonsteady force fields and motions
of  the  suspending  gas.   This limitation requires  empirical design  of inlets or  the  use of
non-aspirating methods for particles with diameters greater than about 10 jjm.
         11.   Fundamental  questions  remain about  the  aerosol  processes  of nucleation,  con-
densation,  and  coagulation,  but  these processes are  understood sufficiently  to  explain and
predict the qualitative behavior of aerosols in the troposphere.
         12.   Atmospheric  observations  confirm  theoretical  predictions that condensation,
gas-particle  reactions,  and  coagulation  are  important  processes only  for the  growth of fine
particles as opposed to coarse particles.
2.2  ATMOSPHERIC DOMAIN AND PROCESSES
     Scientists  are interested in the physical properties and chemistry of SO  and PM because
sulfur  has an  important  natural  cycle  (see  Figure 2-1)  in the environment.   It undergoes
various  oxidation  and  reduction  reactions and trans!ocations among the atmosphere, biosphere,
hydrosphere,  pedosphere,  and  lithosphere.   Human activity (especially fossil-fuel combustion)
has added  a major perturbation to  the  natural  cycle,  and possibly modified natural rates and
reservoirs.  The fluxes of sulfur translocation between reservoirs have been estimated for the
paths numbered  in  Figure  2-1.  Several estimates of annual fluxes are presented in Table 2-1.
The agreement among the  reported values is  not good;  for  example,  the  estimates of annual
anthropogenic sulfur  fluxes  to the atmosphere range from 11  to 45 percent of the total sulfur
involved  in  the atmospheric  balance.   The global  cycles  of carbon and nitrogen  and their
mutual inter-actions with the  sulfur cycle are important, but are too complex to present here.
     While  the  global  sulfur cycle (Figure 2-1)  and the annual  fluxes of sulfur between com-
partments  (Table 2-1)  provide a broad  view  of the  processes that may lead to adverse impacts
upon  mankind  and  ecological  systems, the  global scale  is  clearly beyond  the scope of this
                                             2-3

-------
                                      ATMOSPHERIC :
Figure 2-1. The global sulfur cycle, showing the major reservoirs, pathways, and forms of
occurrence of sulfur. Figures enclosed in circles (e.g., 1) refer to the individual fluxes and
correspond to figures in column 1, Table 2-1.

Source:  Moss (1978).
                                        2-4

-------
                              TABLE 2-1.   ESTIMATES3 OF ENVIRONMENTAL SULFUR ANNUAL FLUXES (TG/YEAR)
rsj
tn

Source of sulfur
Biological decay (land)
Biological decay (ocean)
Volcanic activity
Sea spray (total)
To ocean
To land
Anthropogenic
Precipitation (land)
Dry deposition
Absorption (vegetation)
Precipitation and dry
deposition (ocean)
Absorption (ocean)
Total sulfur involved in
atmospheric balance
Atmospheric balance
Land -> sea
Sea -» land
Fertilizer
Rock weathering
Pedosphere •* river runoff
Total river runoff
Flux number
in Figure 2-1
1
2
3
4
4i
42
5
6
7
8 ,

9
10 '


'


11
12
13
14
Eriksson
(1960,1963)
110
170

45
(40)
(5)
40
65
100
75

100
100

365

-10
5
10
15
55
80
Robinson
and Robbins
(1968,1970)
68
30
--
44
—
—
70
70
20
26

71
25

212

+26
4
: 11
14
J48
73
Kellogg
et al., (1972)

90
1.5
47
(43)
(4)
50
86
10
15

72
—

183
* \
• ' +5
- 4
—
.
—
"*™*
Friend
(1973)
58
48
2
44
(40)
(4)
65
86
20
15

71
25

217

+8
4
26
42
89
136
Granat
et al. (1976)
5
27
3
44
(40)
(4)
65
43
yo
i.O

70
/ -3

144

+18
17
—
—
__
122

     Note;  The numbers in parentheses for sub-pathways 4-.  and 4- are estimates of the decomposition of the total pathway 4.
      Sources:  As cited in each column and,  in part,  Friend (1973, Table 4).

-------
document.   Also,  the sulfur  and particulate matter  emissions are  not uniformly distributed

over the  land  mass of the United States,  nor is the time scale of one year adequately sensi-
tive to relate emissions to effects.

     Cause-effect  relationships are  described  in  terms of  length  and time  scales.   Conse-

quently,  any  discussion of  the atmospheric  physics  and chemistry of  sulfur  oxides  and par-

ticulate matter must be in terms of those length and time scales.   The characteristic time and

length scales  for  typical  effects are shown  in  Table 2-2;  also given are the parameters that

control the functions  relating the  effects to pollutants.   The relationships  of emissions to

effects, as shown in Table 2-2, require that we understand the physics and chemistry of sulfur
oxides and  particulate  matter on time scales of one hour to decades and length scales of 1 cm

to thousands of  kilometers.   With these constraints,  our attention is focused on that portion
of the global  sulfur cycle that consists of the perturbed atmosphere over land surfaces.   Most

of the  natural  and anthropogenic emissions of sulfur and PM are contained within  the tropo-

sphere, which  is the layer of air contained in the zone from ground level to a height of 12 ±

5 km.  This zone contains most of the pollutants  emitted into the atmosphere.

     Thus,  in  order to  understand  the .relationships  between  sources  and  effects,  it  is

necessary to have detailed knowledge of the pathways,  properties,  and processes that are shown
in Figure 2-2.
           TABLE 2-2.  CHARACTERISTIC TIMES AND LENGTHS FOR OBSERVATION OF EFFECTS
       Types of Effects
    Function
Time
Length
     Damage to ecosystems and
      materials due to S02 and
      particulate mass deposition
      (see Chapters 7, 8 and 10)

     Loss of visual quality
      (see Chapter 9)
     Climate modification
      (see Chapter 9)
     Damage to human lungs
      due to S02 and particulate
      inhalation/deposition (see
      Chapters 11 & 12)
Acid flux (= acid and      Hours to      10-10  km
 S02 mass concentration     years
 x deposition velocity)
Mass concentration,        Hours to      10-10  km
 particle size               days
 distribution, and
 composition

Atmospheric burden         Decades       Global
 of particle mass,
 particle size, and
 composition

Mass concentration,        Hours to      < 1 cm
 particle size              years
 distribution,
 composition
                                             2-6

-------
                                        SOURCES
                                                DISPERSION
                                                   AND
                                              TRANSPORTATION
      so2
      - PHYSICAL
       PROPERTIES
      - CHEMISTRY
          / i   I i
     DRY
REMOVAL
       ECO-SYSTEMS
                   TRANSFORMATIONS
WET
REMOVAL
                       EFFECTS
                                                                       1
     PARTICULATE MATTER
     - PHYSICAL
      PROPERTIES
     - DYNAMICS
     -CHEMISTRY
     DRY
REMOVAL
WET
REMOVAL
       ECO-SYSTEMS
  Figure 2-2. Inter-relations of pathways, processes, and properties of sulfur oxides and particulate mat-
  ter and effects.
                                             2-7

-------
     This  chapter discusses  the state  of knowledge of  physical  properties,  chemistry,  and
gas-to-aerosol transformations.  The  dry and wet removal  pathways are discussed in Chapter 6,
which  also  addresses  modeling  of  atmospheric  dispersion,   transport,  transformation,  and
removal.
2.3  PHYSICS AND CHEMISTRY OF SULFUR OXIDES
     Knowledge of the physics and chemistry of sulfur oxides is necessary for designing satis-
factory  samplers  and monitors,  understanding relationships between  sources  and  effects,  and
understanding important tropospheric processes, e.g., chemical transformations and deposition.
     In  Section  2.3, the physical properties  and reaction chemistry of sulfur oxides  in  the
gas and  solution phases  are reviewed.   Current  knowledge in  these  important areas  is  sum-
marized at the end of each of the subsections.
2.3.1  Physical Propertiesof Sulfur Oxides in the Gas Phase
     The  four known monomeric sulfur oxides are sulfur monoxide  (SO),  sulfur dioxide  (S0?),
sulfur trioxide  (SO,),  and disulfur monoxide  (S^O).  Of  these,  only SO, is present at signi-
ficant concentrations  in  the gas phase of the troposphere.   S0~ is emitted directly into the
atmosphere by combusion and manufacturing sources and is formed in the atmosphere by oxidation
of SOy-   However,  because of its high affinity for water (H^O), it reacts within milliseconds
to form  sulfuric  acid (H2SO,).   Polymeric sulfur oxides  are  known to exist, but they are not
stable in the presence of H,0 vapor and are not found in the atmosphere.
     Since the  standard enthalpy of  formation AH°  of  S02 is -70.9  kcal/mole  (25°C),  S02 is
thermodynamically  stable  with regard  to its  formation  from  the  elements  (Glasstone,  1947).
SO- is  capable of  being  oxidized to SO- (AH° = -94.4  kcal/mole), which yields  H_SQ,  in  the
atmosphere (the  important tropospheric  reactions are discussed in Sections  2.3.3  to 2.3.5).
S02 is  also   capable  of being reduced by reaction  with  H,S  to form elemental S (the  Clauss
Reaction).  This  reaction  is important commercially but is not thought to be important in the
troposphere for removing S09; however, it is the likely formation pathway for formation of the
elemental S found in urban particles.   The physical  properties of SQ_, including its molecular
structure and bonding,  vapor pressure of liquid  and solid phases, electromagnetic absorption
(ultraviolet, visible, and infrared) spectra, and thermodynamic constants are well established.
Extensive descriptions  and  references to original work are given by Schenk and Steudel  (1968)
and Schroeter (1966).
     The physical properties of gaseous S02 are well known.
2.3.2  Solution Physical Properties
     Knowledge of  the physical  properties of dissolved S02 solution species  and sulfates is
required for  sampler design,  interpretation  of laboratory measurements  of  S02 oxidation,  and
modeling SO, oxidation in particles, fog, and rain.
           £                                                                               y-
2.3.2.1  Sulfur Dioxide—SO^ dissolves in H2G to form these species:   S02«H20, HS03, and S03 .
Although  the  formation of sulfurous  acid, H2S03$  is often postulated instead  of SQ2-H20, it
                                             2-8

-------
has not been observed (Lyons and Nickless, 1968).  The electronic absorption spectra (Hayon et
al.,  1972),  redox potentials  (Valensi  et al.,  1966), and  structure and  bonding  (Lyons and
                       «        *?»•
Nickless, 1968) of  HSO_  and S0»  are known.   The formation of these  specie's  in water^occurs
through  the  equilibrium  reactions  given in  Table 2-3.   Eigen et al. (1961)  measured -'the
forward (k+,) and reverse (k ,) rate constants at 20°C for reaction:
          S02-H20  *  H+ -t- HSOg                                                '   '  (2-1)'
and found that
               k+1  =  3.4 x JLO^'1                                               ,  ;
               k.j  =  2 x loW1,                .            '                    "   ""b
which are in  good agreement with the thermodynamic value of K., in Table 2-3.  These measure-
                                                              "-I-                     ,
ments are  important because  they demonstrate that the SCU'hLO -  HS03  reaction will  achieve
equilibrium within  1  |js  of a perturbation.   The rate  constants k ? and k_?  for the reaction

          HSQ~  $  H+ + SO^"                                                '   •     ' (2-2)
                                                                                   i     - J,,f~ >  ~*
are unknown.   It  is reasonable that the  value of the protonation rate  constant (k_«) is  less
than  the theoretical  diffusion limit (~5 x  10   Ms),  but greater than k ,.   The expected
                                          3 -1                     2-
range of k+2  is  therefore (0.008-2) x  10 s   ,  which means that SOt  will achieve equilibrium
concentration  within  0.5-125  ms of  a  perturbation.  Thus,  the equilibrium  distribution of
                      2-
SOp'HpQ, NSC,, and  SO-  is expected to be  achieved with a relaxation time of 0.5 to 125 ms.
This  time is  too  short to impact SO,  formation rates in particles, mists, and rain; that is,
equilibrium conditions  can be  assumed  to be continuously satisfied  in  these liquid systems.
However, the  relaxation times  need to be  considered in  interpreting  the kinetics  of rapid
oxidation^  such as measured in flash photolysis  and flash radiolysis experiments.
                                              ,.     o—                             4.          '< f
     An important feature of the SQ^'HJ) - HSO,  - S03  system is the influence of H  in govern-
ing the distribution of these species and the ratio of S0?-- .. to total dissolved S(IV) specfes
concentrations, as  shown  in  Figure 2-3.   The oxidation  rate of this system  is  often pH de-
pendent, indicating different oxidation rates for the three species.  The oxfdation reactions
are discussed in Section 2.3.4.
                                                                                               i
     Sulfite  ion  forms stable  complexes  with many metal  ions, especially those in Periodic
Group  VIII  (Lyons  and Nickless, 1968),   The possible  formation  and significance' of stable
transition metal  ion-sulfite complexes have been suggested (Eatough  et al.,  1978; Hilton et
al.,  1979),  but  contrary evidence  has  been  reported (Dasgupta et  al.,  1979).   At this time,
the  issue  is  unresolved (Eatough  et  al.,  1979).   The  formation  of the  stable  complex di-
chlorosulfonatomercurate ion  is the basis of the West-Gaeke method for  determining S02 in the
air (see Chapter 3).
      The physical  properties  of dissolved  SO- and  its  water-association products  are well
known.
                                             2-9

-------
                   TABLE 2-3.  DILUTE SULFUR DIOXIDE-WATER SYSTEM
	Reaction	Constant (25°C)
  S09, , + H90,0,  5  S09-H90                               H  =  0.0332
    £\.y J    *•" \.-**.jF       £.  €m
          S02-H20  *  H+ + HSO~                           KA1  =  1.39 x 10"2
                                                         pKA1  =  1.86
             HSOg  ?  H"1" + S02~                           KA2  =  4 X 10"8
                                                         pKA2  =  7-4°
Notes:
1.    H   =  Henry's law constant (dimensionless)
          =  ^^2(a"i mo^ar concentration)/(SO--HpO molar concentration)
     Source of value:  Hales and Sutter (1973)
2.   K«,  =  dissociation constant, mole/liter
                             -H^]  (for dilute solutions)
     where
      a.  =  activity of species i, mole/liter
     [i]  =  concentration of species i, mole/liter
     Source of value:  Huss  and Eckert (1977)
3.   K«2  =  dissociation constant, mole/liter
                                (for dilute solutions)
     Source of value:  Salomaa et al. (1969)
4,   The ratio of dilute concentrations of S0?r  v to the total quantity dissolved
     in water is given by:
                                             2(g)
           CS(IV)]
                         [S02-H20] + [HS033 +  [SOg  I
                                        2-10

-------
            '  MOLAR CONCENTRATION  \  \
           .*           RATIO          \  »

           /    I    I     I         I    IV  \l
           1     234567    8
   10'° —
   10
,-7
                                                                   Z
                                                                   o
                                                                   LU
                                                                   O
                                                                   Z
                                                                   o
                                                                   o
                                                                   o
                                               10   11   12
Figure 2-3.  The  distribution of species for the SO2-H2O—HSO3~—SO32"
system as a function of pH. Also, the ratio of the concentrations of SO2(g) to
the total quantity dissolved in water is shown; see Table 2-3, Note 4.
                                  2-n

-------
2.3.2.2   Sulfur Tri oxide and  Sulfur ic Acid-- Know! edge  of the properties of SO,  and  H^SO.  are
important  for the  design   of samplers  and monitors,  and  for  understanding  the  behavior  of
ambient particles, fogs, and rain.
     SO, has a high affinity for water and is not present at significant concentrations in the
atmosphere.  Free SO-  molecules  quickly react with water molecules and droplets to form H?SO,
water  solution  droplets.   The vapor pressure  of S03 over  H^SO. water  solutions is  extremely
low; the  vapor  pressure of hLQ over hLSO* water solution  is an important parameter governing
nucleation of particles  and the  size and pH of water droplets in the  atmosphere.   (See dis-
cussion in Section 2.4.4.)
     H2$0.  is  the only  important strong sulfur oxy acid  in  the troposphere.   Its solution
properties are well  known.   H^SO. is a strong dibasic acid that reacts with water:
                                                                                     (2-3)
           HSO~  *  H* + SO^"                                                        (2-4)
For  water systems  likely to be  present in  the troposphere, the  first dissociation  can  be
considered to  be  complete.   The pK. of  the  first dissociation is -3 and of the second disso-
ciation is ~2 (Robinson and Stokes, 1970).  The relative strengths of acids likely to be found
in tropospheric particles are shown in Table 2-4.  Thus, for pH >3,  the hLSO. - H00 system can
                                       +       p-                '£.**£.
adequately be  described in  terms  of H   and SO,  ;  for lower pH's, it  is  often  necessary  to
consider the presence of HSO».
     Sulfuric acid is a strong dibasic acid; however, it is not a very strong oxidizing agent.
Metals below  hydrogen in  the electromotive  series  are not oxidized by cold  concentrated  or
dilute HnSO*.   The properties of  H^SO,  are  well known (Cotton and  Wilkerson,  1967;  Robinson
and Stokes, 1970; Gillespie, 1968).
     Host  sulfate salts are  soluble;  the  only important exceptions  in the  troposphere  are
CaSO.  and PbSO..   The properties  of tropospheric  aerosols  are  influenced  by  NhLHSO.  and
(NH4)2S04 (see Section 2.4.4).
2.3.3  Gas-Phase Chemical  Reactions of Sulfur Dioxide
     The  chemical  transformation  of  sulfur dioxide  in the  atmosphere  has been  studied  ex-
tensively  over the past  20 years.  Recent  reviews  (Calvert et al., 1978;  Middleton  et a!.,
1980;  and Holler,  1980),  which consider analysis of  laboratory and field  data as well  as
theoretical studies,  indicated  that S0? oxidation may proceed through  both yas-  and  liquid-
phase ractions.  The oxidation of SOg in the atmosphere is of considerable importance,  in that
it  represents  a  major pathway  for particle production  through  the formation  of  sulfates.
Although  the  mechanism of SO- oxidation is  not completely understood,  it  appears to  proceed
via  four  pathways:   homogeneous  gas  phase  reactions;  heterogeneous gas-solid  interface  re-
actions;  and  catalyzed  and  uncatalyzed  liquid phase  reactions.   Homogeneous gas phase  re-
actions are by far the most extensively studied and best understood quantitatively.
                                             2-12

-------
TABLE 2-4.  RELATIVE STRENGTHS OF ACIDS IN WATER SOLUTION (25°C)

Acid Dissociation
HI 2 H* ^
HBr * H* H
HCL ? H* H
H2S04 ? H+<
HN03 * H+<
H20-S02 2 H* H
HSO~ ? H* H
H3P04 * H+H
HF J H*H
HN02 ? H* ^
H2C03 '? H* H
H2S * H4" H
HSO" ? H+ H
HCN +• H* H
NH^ ^ H* ^
HCO; ? H+-
H20 ? H* -
- r
H Br"
^ CL~
h HS04
3
hHS03
H SflJ"
H H2p0;
H F"
h N02
3
i- HS"
H SO|"
i- CN"
- NH3
^ cof
i- OH"
pKft (= -Log KA)
-10
-9.5
-7
-3
-1.3
1.86
1.92
2.12
3.2
3.3
6.38
7.0
7.4
9.21
9.25
10.3
14.0
Reference
Cotton & Wilkerson (1980)
Cotton & Wilkerson (1980)
Cotton & Wilkerson (1980)
Robinson & Stokes (1970)
Robinson & Stokes (1970)
Huss & Eckert (1977)
Cotton & Wilkerson (1980)
Cotton & Wilkerson (1980)
Cotton & Wilkerson (1980)
Cotton & Wilkerson (1980)
Cotton & Wilkerson (1980)
Cotton & Wilkerson (1980)
Salomaa et al. (1969)
Cotton & Wilkerson (1980)
Cotton & Wilkerson (1980)
Cotton & Wilkerson (1980)
Cotton & Wilkerson (1980)
                              2-13

-------
     The  homogeneous  gas-phase chemistry  of oxidation  in the  clean  and the polluted tropo-
sphere  is reviewed in  this section.   The current  status of knowledge  is presented for the
elementary  oxidation  reactions of SO,  and' the importance  of volatile  organic -and nitrogen
oxides  as  generators  of free radical oxidizers.  This review will show that the photochemical
oxidation  of  SO,, may  be a  significant  pathway  for tropospheric sulfate  formation.  The three
most important  oxidizers  of SOp are:   hydr*oxyl radical, HO; peroxyl radical, H0?; and methyl-
peroxyl radical,  CH^O,-   At this   time^  only  the  reaction rate constant  for HO is well estab-
lished.  The pathways of formation of the  oxidizer radicals for the unpolluted troposphere can
be  explained  in  terms  of  the   photochemistry  of  the  NO-CH.-CO-O.,  system.    In  polluted
atmospheres,  volatile organics  and  oxides of -nitrogen  act  together to  produce additional
radicals  and  accelerate overall radical production.   There is also evidence  that a dark re-
action among 0.,, alkenes, and SO,  is effective in-oxidizing-SOp,
2.3.3.1   elementary Reactions—The elementary chemical  reaction's of SOp  in air'have been the
subject  of intense investigation.   Studies prior  to 1965  have been critically  reviewed by
Altshuller  and  Bufalini  (1971), and more  recently by Calvert et al.  (1978).   Calvert et al.
(1978) systematically examined the rate constants and significance of SOp elementary reactions
known  to  occur  in  the  troposphere.    Identified  as generally  unimportant  reactions  were:
photodissociation; photoexcitation; reaction with singlet  delta oxygen, 09('A ); re'action with
                 3                                                        ^9
atomic oxygen, 0( P); reaction with ozone, 03; reaction with nitrogen oxides (NO-, N03, NpO,-);
reaction  with tert-butylperoxyl  radical,  (CH3)3C02;  and  reaction  with acetylperoxyl radical,
CH-COO-.   The only SQp  reactions  in  the  troposphere that were  identified as important were
those due  to  HO, HOp, and CH^Op.   The  rate constants recommended by Calvert et al. (1978) for
these three reactions are given in Table  2-5.   More recent work is in conflict with the rate
constants  for HO- and CH3Qp that  have  been recommended by Calvert et al.  (1978).  Graham et
al. (1979)  and  Burrows et al.  (1979)  have reported rate  constants for  the HOp reaction that
are much lower than that recommended by Calvert et al. (1978).  Also, Sander and Watson (1981)
have reported a  rate  constant for the  CHgOp reaction that is much lower  than that recommended
by Calvert et al.  (1978).  These  more  recent values are  given  in Table  2-5.  The reasons for
the discrepancies  in  these two rate constants 'are unknown, and there is  no basis to recommend
preferred  values.
     Although the  dark reaction of SOp  +  03  is too  slow  to  be  important in the troposphere,
the addition of  alkenes greatly enhances the-oxidation rate.  The experimental work of Cox and
Penkett U971a,b; 1972) and flcNelis et  al.   (1975) has been reviewed and reevaluated by Calvert
et al.  (1978).   The reaction system is too complex to discuss here, but  Calvert et al. (1978)
report  results  of their calculations for  total alkenes =  0.10 ppm, [0,]  = 0.15 ppm, and [$02]
= 0.05 ppm; they estimated that the disappearance rate of  S0?  is 0.23 and 0.12 percent/h at 50
and 100 percent relative humidity (25°C),  respectively.   The  reaction mechanism for the 0, •*•
                                             2-14

-------
            TABLE 2-5.   RATE CONSTANTS FOR HYDROXYL, PEROXYL, AND METHOXYL RADICALS
            Reaction
                            Second grde$ rate..  ,
                          constant,  cm-mole  s
  Source
     HO + S0£ •» HOSO£
     H0£ + S02 •+ HO
S02 •* CH30 + S03
                           (1.1 ±0.3) x 10
                                                        -12
                           >(8.7 ± 1.3) x 10

                                 v-18
                                                         -16
                                           x 10
                                        <2 x 10
                                               -17
                                        (5.3 ±2.5) x 10

                                        5 x 10" 17
                                                        -15
Calvert et al. (1978)
Calvert et al. (1978)
                                                      Graham et al. (1979)
                                                      Burrows et al. (1979)
Calvert et al. (1978)
                                                      Sander and Watson
                                                       (1981)
       Temperature = 25°C.
alkene +  S02 system  is  not  known,  but studies  by Niki et  al.  (1977) and Su  et al.  (1980)
indicate that the reactive species may be the biradical (RCHOO) formed by the decomposition of
the original molozonide,
     In summary:
     1.   Three  gas-phase  tropospheric oxidation  reactions  have  been  identified  as  being
possibly important:
          a.   HO radical.  The rate constant appears to be well established.
          b.   H02 radical.  The rate constant is not well established.
          c,   CH302 radical.  The rate constant is not well established.
     2.   SO, + 0~ + alkenes may be an important dark reaction.
2.3.3.2  Tropospheric Chemistry. of Sulfur Dioxide Oxidation— The chemistry of the clean tropo-
sphere and  its  mathematical  simulation have been studied extensively by Levy (1971), Wofsy et
al. (1972),  Crutzen  (1974),  Fishman and Crutzen (1977), Chameides and Walker (1973, 1976) and
Stewart et al. (1977).
     The  photochemistry  of  the  unpolluted  troposphere  develops   around a  chain  reaction
sequence  involving NO,  CH, ,   CO  and 0,.   The photochemical  reaction  chain sequence  in the
                                                                1
troposphere is initiated by HO formed from the interaction of 0(0), the product of photolysis
of 0, in the short wavelength portion of the solar spectrum, with HJ3.
               hv(A. < 310 nm) •* 0(AD)

                  0(1D) + H20 * 2HO
                                                         (2-5)

                                                         (2-6)
                                             Z-15

-------
     The HO produced reacts with CH» and CO present  in  the  clean  troposphere,  resulting in the
generation of peroxyl radical species, H02, CH^Og.
          HO + CH4 •» CH3 + H20                                         (2-7)
           HO !+ CO •»• H + C02                    _                       (2-8)
      CH3 + 02 + M -» CH302 + M                                         (2-9)
        H + 02 + M -> H02 + M                                           (2-10)

     In turn, the  peroxyl  radicals participate in a chain  propagating sequence which converts
nitric oxide  (NO) to nitrogen  dioxide (NOp) and in the process produces additional  hydroxyl
and peroxyl radical species.
                CH302 + NO •» CH30 + N02                                (2-11)
                  H0£ + NO -» HO + N02                                  (2-12),.
                 CH30 + 02 -» H02 + H2CO                                (2-13)
     H£CO + hv(A < 370 nm) •* H + HCO                                   (2-14)
                  HCO + 0£ •+ H02 + CO                                  (2-15)
                 H202 + HO •+ H20 + H02                                 (2-16)
     The major chain terminating steps include:
              HO + N02 + M •»• HON02 + M                                 (2-17)
                 H02 + H02 •»• H202 + 02                                 (2-18)
     The reaction  sequence for 03 production  involves  converting NO  to  N02  at a  rate suffi-
ciently  high  to  maintain  a  N02/N0  ratio to  sustain  the  observed background levels of  0,.
                  H02 + NO •»• HO + N02                                  (2-12)
                  N02 + hv •»• NO + 0                                    (2-19)
                0 + 02 + M-»03 + M                                    (2-20)
                   NO + 03 •*• N02 + 02                                  (2-21)
                   HO + CO -» H + C02                                   (2-8)
     In general, reactions (19) through (21) govern  the 03  concentration  levels present in the
sunlight  irradiated well-mixed  atmosphere at  any  instant  and  to a  first  approximation  the
steady state relationship, Leighton (1961),
               (N02) k,g
               	f   J-J  =  (n \
               (NO) k21      tu3j
                                             2-16

-------
provides an accurate estimate of 0, given the ratio of (NO?)/(NO) and k,q/k?,.   The photolytfc
rate constant k-,g is directly related to the integrated actinic solar flux over the wavelength
range 290 - 430 nm.
     The paths for CU net destruction in the troposphere include the reaction sequence
          H02 + 03 -» HO + 202                                         (2-22)
           HO + 03 -» H02 + 02                                         (2-23)
     Hydroxyl  radical  abundances predicted  by the tropospheric photochemical  models,  10  to
  c          «. o
10  molec  cm  ,  are in qualitative agreement with recent measurements by Davis et al. (1976),
Perner  et  al.  (1976),  and Campbell  et  al.  (1979)  and inferred HO  levels  based on measured
trace gas abundances in the troposphere by Singh (1977).
     In  the case  of  the  chemistry  of polluted  atmospheres,  extensive  discussions  on the
mechanism  of photochemical  smog and  its computer  simulation  have  been presented by Demerjian
et al.  (1974), Calvert and McQuigg (1975), Niki et al. (1972), Hecht et al. (1974), and Carter
et al.  (1979).
     Perturbations  of  the photochemical  oxidation cycle within the  atmosphere introduced by
anthropogenic emissions are predominately caused by two classes of comppunds, volatile organic
compounds  and  nitrogen oxides.   The  reaction  chain sequence  discussed  earlier for the clean
troposphere  has  now been  immensely complicated by  the addition of scores of volatile organic
compounds which participate in the chain propagating cycle.  Figure 2-4 depicts a schematic of
the  polluted atmospheric  photooxidation  cycle.   The  addition of volatile  organic compounds
(VOC) in the atmosphere introduces a  variety of new peroxyl radical species.
     In  its  simplest   form,  the  photochemical  oxidation  cycle in  polluted  atmospheres is
governed by the following basic  features.   Free  radical attack on atmospheric VOC's is  ini-
tiated  by  a  select group of  compounds,  which are  for the most part activated by sunlight.
Formaldehyde  and  nitrous  acid, in particular, show  high potential  as free radical initiators
during  the sunrise  period.   After initial  free  radical attack, the VOC's decompose through
paths resulting  in  the production of peroxyl  radical  species (HO,,, R0?, R'0», etc.) and  par-
tially  oxidized  products,  which in themselves may be  photoactive radical-producing compounds.
The peroxyl  radicals react with NO, converting it to N02, and  in the process produce hydroxyl/
alkoxyl  radical  species (OH,  RO, R'O, etc.).  Alkoxyl radicals can be further  oxidized, form-
ing  additional  peroxyl  radicals and  partially oxidized products, thereby completing the inner
loop reaction  chain illustrated in Figure 2-4; or they may attack, as would be the major  path
for  hydroxyl  radical,  the VOC pool present  in the polluted atmosphere, thereby completing the
outer loop  reaction chain.  The resultant  effect in either case  is the conversion of NO to NG*2
with a  commensurate oxidation  of  reactive  organic carbon, shortening of the hydrocarbon chain,
and production of organic  particles,  C02 and HgO.
     The  complex mixtures  of  organic compounds  present in  the polluted atmosphere react at
different  rates depending  upon their  molecular structures, resulting  in varying yields of  free
radical  species,  ozone, N02,  peroxyacetyl nitrate  (PAN), and  other partially oxidized organic
products as  a  function  of  VOC  composition  and VOC-NO   levels.
                                                    A
                                             2-17-

-------
 FREE RADICAL INITIATORS

      NO2 + hv
      03 + hv
      MONO + hv
      RCHO + hv
      PAN + hv/AT
      03 +C-C
VOC = VOLATILE ORGANIC COMPOUNDS
PAN - PEROXYACETYL NITRATE
        Figure 2-4.  Schematic of the polluted atmospheric photooxidation cycle.
                                       2-18

-------
     Hydroxyl radical reactions  seem to be the dominant  mechanism by which CO, hydrocarbons,
N02 and SO, are consumed in the atmosphere (Niki et al., 1972; Demerjian et al., 1974; Calvert
et al.,  1978).   This highly  reactive  transient  species, quite contrary  to  its  organic free
radical counterparts, does  not show appreciable change in  concentration with atmospheric VOC
and NO  variation, a result readily explainable upon review of the free radical production and
      X
consumption  sources.   In the  case  of HO, ambient concentration conditions which enhance its
production tend  also to  consume the radical  at  an  equivalent rate.  The result is  a faster
cycling in the VOC-NO  oxidation chain (i.e., increased chain lengths) but very little pertur-
                     A
bation  in the  HO  steady  state  concentration.   In  contrast,  organic  free  radicals,  mainly
peroxyl species,  are  consumed by alternate pathways  which  are less competitive and result in
increased steady state concentration.
     Applying this  basic  knowledge  of the photochemistry of  the lower atmosphere, Calvert et
al,  (1978)  determined theoretical  rates  of  SO, oxidation via attack  of various  free radical
species whose concentrations were estimated from computer simulations of the chemical  reaction
mechanisms for clean and polluted atmospheres.
     Based on  limited rate constant data for the SO,,-free radical  reactions, Calvert deter-
mined  that HO  dominated  the rate of SO, oxidation in the clean troposphere,  while in polluted
atmospheres  the  rate  of  SO- oxidation showed equivalent contributions from HO, HO,, and CH,0,
radicals.  Figure  2-5 depicts  the  estimated  time  dependent rates  of SO,, oxidation  by free
radical species  in  a polluted air mass.  Recent laboratory measurements suggest that the rate
of reaction  of  SO,  with HO,  and  CHqO?  may  not  be  as  great as  estimated by Calvert et al.
(1978)  (see  discussion   in  Section  2.3.3.1).   During July  at mid-northern latitudes, typical
rates  of  S0? oxidation  were  of  the order of  1.5  percent/h and  4.0  percent/h  for clean and
polluted  atmospheres,  respectively.  The major difference  in rates  was  a   result of higher
concentration  levels of  free radicals in  the hydrocarbon rich  polluted atmospheres.   In  a
                                                                                            2-
similar manner,  Altshuller  (1979)  predicted the rates of homogeneous oxidation of S0? to SO-
in the clean troposphere using concentration predictions of the pertinent free radicals from a
two  dimensional  global  model  by Fishman  and  Crutzen  (1978).   A sample  result from this study
showing the  latitudinal  and seasonal dependence of the  rate  of S0? oxidation is presented in
Figure  2-6,  the  variability  in rate being  predominantly  due  to  availability of ultraviolet
solar  intensity  which drives the free-radical production process.  The solar radiation depend-
ence  of SO,  conversion rate  has  also been observed  in field measurements within power plant
plumes  (Husar  et al. ,  1978), but  should be  viewed  cautiously in  light  of  the complicating
factors  introduced by  the  dispersion  and  local  chemistry of the  primary  source emissions.
     The  most  important  impact  on   SO, homogeneous  gas phase reactions has  come from recent
experimental determinations of the  reaction  rate constants of SO, with HO,  by Graham et al.
(1979)  and by Burrows  et al.  (1979) and S02  with  CH302 by Sander  and Watson (1981).   As a
result of these  recent  determinations, H0? and CH,0, must be considered questionable as con-
tributing sources to the oxidation of  S02 in  the atmosphere.   Therefore,  in the theoretical
                                             2-19

-------
                 30           60            80

                    IRRADIATION TIME, min
120
Figure 2-5.  The  theoretical rate of reaction (percent per hour) of
various free-radical species with SO2 is shown for a simulated sunlight-
irradiated (solar zenith angle of 40°) polluted atmosphere. The initial
concentrations (in ppm) were as follows: 862, 0.05; NO, 0.15; NC^,
0.05; CO, 10; CH4,1.5; CH20,0; CH3CHO, 0. The relative humidity
was 50 percent, and the temperature was 25° C.

Note:  The rate constants for HO£ and CHgO2 radical reactions with
SO2 are not well established. See Table 2-4 and  its discussion.

Source: Calvertetal. (1978).
                              2-20

-------
                                 O  JULY,HO,HO2,CH3O2
                                 •  JULY, MO
                                 D  JANUARY, HO, HO2,CH3O2
                                 •  JANUARY, HO
                            LATITUDE, °N
Figure 2-6.  Percentage conversion at mid-day of sulfur dioxide to
sulfate by HO and by HO, HO2. and CH3O£ radicals as function of
°IM latitude in summer and winter.

Source:  Altshuller (1979).
                                 2-21

-------
estimates of S02 oxidation rates, by Calvert et al. (1978), and by Altshuller (1979), only the
hydroxyl radical portion  of the contribution is now accepted as established, in view of these
recent  experimental  rate constant  determinations.   This  results  in maximum  established  SCL
oxidation rates of  the order of 1.5 percent/h  for both clean and polluted atmospheres during
July at mid-northern  latitudes,  a factor of 2.5  less  than previous theoretical estimates for
polluted atmospheres.   The revised  rate is equivalent  to a  diurnally  averaged rate  of the
order 0.4 percent/h.   Field measurements on the  rates  of  SO* oxidation, discussed in Chapter
6j indicate that maximum  S0? oxidation rates of the order of 10 percent/h are typical of many
atmospheric pollution scenarios.  Our present knowledge of homogeneous SOp gas-phase reactions
does not sufficiently  account for the rates observed.   Smog chamber studies have demonstrated
that some species other than HO radical oxidizes  S02  (Kuhlman et al.,  1978;  McNelis et al.,
1975).   Alternate  homogeneous gas  reaction  oxidation  pathways are being studied  (Su et al.,
1980),  but certainly  the  role of heterogeneous and liquid phase SQp oxidation pathways should
not be overlooked in attempts to resolve this discrepancy.
     In summary:
     1.    HO radical  dominates the  gas-phase  oxidation of  SOp in the  clean  troposphere.   A
          typical  rate  is on the order  of  1.5  percent/h at noon during July  at mid-northern
          latitudes.
     2.    HO radical  accounts for  about 1.2 percent/h  of the SO- oxidation  in the polluted
          troposphere.  The combined contribution of H0« and CH^O, radical reactions may be as
          great as  about 2.8  percent/h,  but their  rate constants are  not well established.
2.3.4  Solution-Phase Chemical Reactions
                                                  2-
     The reactions  of  the aqueous 502^20^503-50.3  system  is  important to understanding the
processes of H,SO*  formation  in tropospheric particles, mists,  fogs,  and rain.  This section
reviews the oxidation  reaction of dissolved SQp species, including the auto-oxidation, metal-
ion catalyzed oxidation, carbon catalyzed oxidation, and reactions with the dissolved oxidants
N02, 03, and HgOg.
     The state  of   knowledge  of aqueous oxidation  rates  of dissolved SQp,  HSO-, and SO,   is
inadequate for  simple  systems and is extremely poor (or nonexistent) for complex systems that
include dissolved nitrogen  and carbon compounds.   Unfortunately, most of the  studies are not
definitive because  the  investigators:   (1)  did not provide sufficient descriptions of experi-
mental  procedure (especially the purification of the water and reagents), (2) did not select a
proper  reactor  design, and (3) worked  at  concentration levels that were  orders of magnitude
greater  than  possible  for ambient  atmospheric  aqueous  systems.   Trace quantities  (at  the
part-per-billion level) of  catalytic metal  ions are capable of enhancing the reaction veloci-
ties by orders  of  magnitude  over  the  auto-oxidation rate, while similar  trace quantities of
organics inhibit the rate.  The characteristics  of the chemical reactor govern the range of
the  half-life   that can  be investigated and may influence the  observed rate  of  oxidation.
                                             2-22

-------
Two-phase air-water reactors (e.g., bubblers and supported droplets) may have reaction charac-
teristics that  are dependent upon:  (1)  the  mass  transfer rate of  the reactants through the
air-water interface, and (2) the mixing rates within the gas and water phases (Carberry, 1976;
Freiberg  and  Schwartz, 1981).   Unless an adequate characterization  of the two-phase reactor
was performed,  it is  not  recommended  that the implied elementary  rate constant be accepted,
although  in many cases the  results may  be correct.   Supported droplets may present an addi-
tional problem:   radical chains are efficiently terminated at liquid-solid  interfaces, thereby
reducing  the  observed rate.   Therefore,  supported  droplet  measurements  are  not defensible
unless it is  established  that the oxidation is not a free-radical mechanism.  Several notable
reviews  of the  oxidation  of dissolved S0? and its hydration products  in  simple systems have
been published recently (Schroeter, 1963; Hegg and Hobbs, 1978).
     This review will  show that:
     1.   The auto-oxidation (uncatalyzed) reaction  is very slow compared  to  the other re-
          actions.
     2.   Mn(II)  and Fe(III) are  significant catalysts for the  oxidation.   The kinetic rate
          expression  is in  doubt  for the  Mn(II)  reaction, but  several independent investi-
          gators agree on a  rate expression for Fe(III).
     3.   The catalytic effectiveness  of these ions  is unknown:  Cu(II), V(¥),  V(IV), Ni(II),
          Zn(II), and  Pb(II).
     4.   Elemental  carbon (soot) with  a water film may be  an effective oxidation catalyst.
     5.   Dissolved  HN02  and 03  oxidation rates  are known and  appear to  be  too  low  to be
          effective.
     6.   The kinetics of  the dissolved H202 oxidation of dissolved  S02 species  are known and
          appear to be effective for forming SO.   in  particles, mists,  fogs,  and  rain.
2.3.4.1   S(IV)-00  -  H.,0 System—The  simple S(IV)  - 0»  auto-oxidation (see glossary) has been
                 iL.     "£,, "	"	  	""'"	                        
-------
              TABLE 2-6.  INVESTIGATIONS OF S02 - 02 AQUEOUS SYSTEMS

Investigators
Bigelow (1898)
Titoff (1903)
Lumiere and Seyewetz (1905)
Milbaur and Pazourek (1921)
Reinders and Vies (1925)
Haber and Wansbrough- Jones (1932)
VoVfkovick and Belopol'skii (1932)
Backstrom (1934)
Fuller and Crist (1941)
Riccoboni et al. (1949)
Abel (1951)
Winkelmann (1955)
van den Heuvel and Mason (1962)
Schroeter (1963)
Schwab and Strohmeyer (1965)
Rand and Gale (1967)
Scott and Hobbs (1967)
McKay (1971)
Miller and de Pena (1972)
Brimblecombe and Spedding (1974a)
Beilke et al. (1975)
Horike (1976)
Larson (1976)
Huss et al. (1978)
Larson et al. (1978)
Type of system
Bubbler
Bulk
Bulk
Bulk
2-phase bulk
Bulk
Bubbler
Theoretical
Bubbler
Bulk
Theoretical
Bulk
Supported droplet
Bubbler
Bulk
Bulk
Theoretical
Theoretical
Supported droplets
Bubbler
Bulk
Bubbler
Bubbler
Bulk
Bubbler
*
Comment
1,2,3
2,3
2,3
2,3
1,2,3
2,3
1,3
--
1
2,3
--
2
1,3
1
2
2,3
—
__
1
1
2
1,3
1
1,3
I

1.   Incompletely characterized 2-phase system; results may be incorrect since
    the investigators did not completely account for mass transfer rate or
    demonstrate that no concentration gradients existed.

2.   Purity of water is uncertain; results cannot be considered reliable.

3.   Rate expression not reported.
                                       2-24

-------
Termination
                   SO. + inhibitor -» (non-reactive species)           (2-28)
                 radical + radical •* (non-reactive species).          (2-29)
     Brimblecombe and Spedding (1974b) propose an alternative scheme that does not include the
SCL radical -ion; in their scheme, equation (2-26) is replaced by:
                        S0~ + SOg" + SOj + SOg"                       (2-30)
                                   •* 2 soj~                           (2-31)
and equation (2-28) is absent,
     Hegg and  Hobbs  (1978) have discussed most of the investigations identified in Table 2-6,
and they summarized the rate expressions, rate constants, and important features of the studies.
The observations can be classified into three types of rate expressions:

     1.   The type first reported by Fuller and Crist (1941),
                        2-
                    d[SO»  ]                  n co     ?_
                    - gf-  =  (ka + kb [H ]U'bU) [SOf ]             (2-32)

     2,   The type first reported by Winkelmann (1955),
                    d[So!~]           „
                    - gj-  =  kc [SO^ ]                             (2-33)
3.    The type observed by Beilke et al.  (1975).
                  o|~]          + _0 ,6   2_
                  gf-  =  kd [H+] °'16[S02 ]                    (2-34)
                    d[So~]
     The values  that have been reported for k  and k. for Type 1 reaction (equation 2-32) are
                                              oE      D
shown  in Table 2-7.   Schroeter (1963) has argued that his values obtained* for air (20 percent
Op)  are in  excellent  agreement with those  obtained  by Fuller and Crist (1941)  for pure Qy,
assuming a first-order dependence of the rate on dissolved [0?].  In fact, no study has demon-
strated an [Op]-dependence other than zero; hence, these two studies differ by a factor of ~5.
     The values  reported for k  for Type  2  reaction  (equation 2-33)  are shown  in Table 2-8.
These  values, many  of which  were  obtained by  Beilke and  Gravenhorst (1978) through  a re-
analysis of  the  original investigators'  reported data and expressions, are in good agreement.
     For Type 3  reaction (equation 2-34), Beilke et al. (1975) observed a value of krf = 1.2 x
10"4 sec'V"16  (pH  = 3-6, T = 25°C).  It  is presently unresolved as to which type of rate ex-
pression is  correct.  Doubt is cast on  "Type  3" found by Beilke et al. (1975) because of the
                                             2-25

-------
                TABLE 2-7.  VALUES3 OF ka AND kb FOR REACTION TYPE 1

Investigator
Fuller and Crist (1941)
Schroeter (1963)
K
McKay (1971)°
Larson et al. (1978)
k^sec"1
1.3 x 10"2
2.9 x 10"3
_•}
1.3 x 10 4
4.8 x 10"3
k^sec'V*
6.6
32

57
8.9
pH
5.1 - 7.8
7.0 - 8.2

5.1 - 7.8
4-12
T,°C
25
25

25
25
a Adapted from Hegg and Hobbs (1978)

  McKay used a more recent value for K.n(=6,26 x 10  M) in expressing k.  for Fuller
  and Crist's (1941) data.            w                               D
                    TABLE 2-8.  VALUES3 OF k  FOR REACTION TYPE 2

Investigator
Winkelmann (1955)
Schroeter (1963)b
Scott and Hobbs (1967)c
Hi Her and de Pena (1972)
Brimblecombe and.
Spedding (1974a)a
kc, sec'1
3. 5x10" 3
M6~0.6)xlO~3
1.6x10" 3
3xlO"3
(3.7-0.6)xlO~3
pH
7
7-8
•v.6.2 - 6.9
2-4
4-6
T,°C
25
25
25
25
25

3  Adapted in part from Beilke and Granvenhorst (1978) and Hegg and Hobbs (1978).

b  Determined by Beilke and Gravenhorst (1978) from Schroeter1s (1963) data for
   pH = 7-8.
c  Determined by Scott and Hobbs (1967) from Van den Heuvel and Mason (1963) data.

   Determined by Beilke and Gravenhorst (1978) by transforming the rate constants
   reported for pH = 4-6 by Brimblecombe and Spedding (1974a).
                                        2-26

-------
use of a  plastic vessel  that could have  introduced trace organic inhibitors into the system.
All of  the other  studies  (yielding  "Types  1 and  2")  were performed  witn  two-phase systems
whose mass transfer properties were insufficiently reported.
     The auto-oxidation  is  inhibited  by trace concentrations of organic species.   The classes
of  organic species  capable of  serving as  inhibitors  include alcohols,  glycols,  aldehydes,
ketones, phenols,  amines,  and  acids.   Backstrom (1934) first demonstrated that the inhibition
of sulfite oxidation can be expressed as:

                     ]                        2_
               	g—  =  {A/(B + m)} k35 [SO* ]                          (2-35)

where:
     koc  =  the uninhibited rate constant
     A,B  =  constants that are functions of the inhibitor
       m  =  molar concentration of the inhibitor.
The influence  of inhibitors on the rate has been extensively studied by Schroeter (1963), and
more  recently  by Altwicker (1979).   According to Schroeter (1963), A and B are usually on the
order  of  10   molar,  which means  that inhibitor concentrations greater  than  10   molar are
effective.  The  form of the rate equation (Equation 2-35) suggests that the mechanism involves
a bimolecular  reaction between an inhibitor molecule and a radical in the chain.
     In summary:
     1.   The  auto-oxidation reaction is very slow.
     2.   The  rate is extremely sensitive to the presence of catalysts and inhibitors.
     3.   The  rate is first order in sulfite.
     4.   No reaction mechanism has been satisfactorily demonstrated to account completely for
          the  observations  of  the more reliable studies  (e.g.,  the  dependence of the rate on
          [H*]°-5 found by Fuller and Crist, 1941 and by Larson et al., 1978).
2.3.4.2   S(IV)- catalyst  - 00  -  H00  System—It  is well  established 'that some metal cations
                             4       2-
catalyze  the  oxidation of  HSO, and SO,  by molecular 0~.  Of particular interest to the issue
of  atmospheric  sulfate  formation  in particles,  mist,  fog,  and rain is possible  catalytic
activity  of Mn(II),  Fe(III),  Cu(II),  Ni(II), V(IV),  and V(V).   General features of the cata-
lyzed  reaction  include:    (a)  inhibition  by  oxidizable  organic  molecules,  (b) inhibition by
metal ion-complexing molecules (inorganic and organic), (c) exhibition of an induction time of
several  seconds  to  several  minutes,  (d)  detection  of  metal  ion -  S(IV) complexes,  (e) no
dependence of  rate on dissolved 0? concentration,  and  (f) dependence of  the  rate  on the in-
verse  of  the  initial H  concentration  (i.e., the  rate is independent  of  pH  change  after the
reaction  has  been initiated).   While the catalytic  reaction  mechanisms are  unknown, they are
thought to be a modification of the  initiation step of the auto-oxidation free radical mecha-
                                                   +                                  —9
nism  (Equations 2-24  through  2-29);  instead of M  being  a  trace  concentration (<10   M) of
                                             2-27

-------
metal  ion  or a  reactive wall, it i-  a  reagent present at concentrations >10    M.   The rate
expressions  for  the  various  catalysts  have different  forms,  suggesting different  types  of
initiation  mechanisms (e.g.,  simple  redox  reactions  or the  formation of  stable,  reactive
complexes).  The agreement  among  independent investigators is  generally poor,  indicating the
likelihood of mass  transfer limitations of the rate or the presence of contaminants.   A large
percentage of  the  investigations  were conducted with  two-phase reactors for  which  the mass
transfer characteristics  were  not adequately reported; therefore, those results, which may be
correct, are  not considered .for  estimating  the elementary rate  constant  and for determining
the  reaction  order.  Also,  results  of investigations using supported  droplets may  be biased
due to radical chain termination at the liquid-solid interface.
     The Mn(II)  catalyzed  reaction  kinetics have  been  investigated for over  75 years.   The
                                        9-
studies pertinent to the formation of SO-  in the troposphere are presented in Table 2-9.  One
of  the  first  critics  of  Mn(II)  catalysis  studies  was  Titoff (1903),  who  remarked:   "In
Bigelow's  (1898) work the reaction occurred between  two phases,  and the retardation could be
determined by a change in the,boundary layer or by a decrease in the solution rate of oxygen."
Unfortunately, that comment applies  to all  but  three  of the Hn studies shown in Table 2-9,
which are:  Hoather and  Goodeve (1934), Neytzell-de Wilde and Taverner  (1958), and Coughanowr
and Krause  (1965).   It is odd that none of these investigators presented rate expressions and
rate constants derived from their data.  Instead, they left to the reader the task of extract-
ing that information.  Estimates of their rate expressions are presented in Table 2-10.  There
is  agreement  that  the Mn(II)  catalyzed rate  is independent of dissolved 00,  S00,  HSO~, and
  9*.                                                                        *-     *-     *^
S03  concentrations.
     Clearly, Hoather and  Goodeve (1934)  and  Coughanowr and Krause  (1965) are in agreement.
However, Neytzell-de  Wilde  and Taverner (1958) observed a first-order dependence on [Mn(II)].
There seems to be  no basis to discount  any  of the three  investigations, yet it appears that
serious errors may  have  been made.  The results  of Neytzell-de Wilde and Taverner (1958) are
slightly preferred  because:   (1) they measured  the' rate"'df!"~disappearance of  S(IV)  by direct
chemical means,  and (2) the period  of observation (10-100 minutes)  of  the  experimental  runs
were sufficiently long that it is reasonable that the rate of oxidation was measured after the
establishment of the radial chains, and not during the induction period.
                                                                                     2-
     The Fe(III)  catalyzed reaction studies that are pertinent to the formation of SO^  in the
troposphere are  identified  in  Table  2-11.   The only studies not using two-phase systems (sub-
ject  to mass  transport  limitations)  are those of  Neytzell-de Wilde  and  Taverner (1958),
Karraker (1963), Brimblecombe  and Spedding (1974a), and  Fuzzi  (1978).   Hegg  and Hobbs (1978)
have  pointed  out that Karraker (1963)  did  not investigate the  catalyzed oxidation  in which
dissolved  02  is  the  oxident,  but instead  investigated the redox system  associated  with the
couple Fe(III) + e   •* Fe(II)  in  an  oxygen-free system.   Thus, Karraker's work is not appli-
                                                                  2-
cable.  Neytzell-de Wilde and Taverner (1958) reported that the SO,  formation rate was second
order  for  [S(IV)],  but Karraker (1963)  has  reanalyzed  their data and  has shown instead that
                                             2-28

-------
         TABLE 2-9.   INVESTIGATIONS OF S02  - MANGANESE - 0£ AQUEOUS SYSTEM
     Investigators
  Type of System
Comment
Titoff (1903)
Johnstons (1931)
Hoather and Goodeve (1934)
Bassett and Parker (1951)
Johnstons and Coughanowr (1958)
Neytzell-de Wilde and Taverner (1958)
Johnstone and Moll (1960)
Coughanowr and Krause (1965)
Bracewell and Gall (1967)
Matteson et al. (1969)
Cheng et al. (1971)
Bulk
Bubbler
Bulk
Bulk
Supported droplet
Bulk
Free droplets
Bulk and flow
Bubbler
Free and supported droplets
Supported droplets
    2 •
   1,2
    2
    2
   1,2
    2
    2
    2
    1
    3
    1
 1.  Incompletely characterized 2-phase sys.tem; results may be incorrect since
     the investigators did not completely account for mass transfer rate or
     demonstrate that no concentration gradients existed. ,
2.   Rate expression not reported.
3.   Results are biased due to continued reaction of (supported) droplets on
     filter of sampler; rate expression cannot be considered reliable.
                                        2*29

-------
         TABLE 2-10.  RATE EXPRESSION FOR THE MANGANESE-CATALYZED OXIDATION
           Expressiona'b'c
                                                                  Investigators
           ~
           =  44
                                    n   + n
                             [S(IV)]U [H f
 3-4
Adapted from
 Hoather and
 Goodeve (1934)
           =  1.7 x 10 J CMn(II)] [S(IV)]
                                         °
    dt
           =  8
                          [S(IV)]
                                                       -2.2
-3-4
Adapted from
 Neytzell-de Wilde
 and Taverner
 (1958)

Adapted from
 Coughanowr and
 Krause (1965);
 dependence on
 pH not reported
a.

b.
     The units are:   liter, mole, second.
                                                       -.0,
     Concentrations shown with zero power (e.g., [S(IV)] ) indicate that the
     investigators found the rate to be independent of those species.   Note that
     any concentration to the zero power is equal to unity.

c.    The term [H+1   indicates that the rate is dependent only on the inverse of the
     initial H  ion concentration; changes in H
     in progress do not affect the rate.
                                                 concentration after the reaction is
                                        2-30

-------
            TABLE 2-11.  INVESTIGATIONS OF S02 - IRON - 02 AQUEOUS SYSTEM

Investigators
Reinders and Vies (1925)
Bassett and Parker (1951)
Higgins and Marshall (1957)
Johnstone and Coughanowr (1958)
Junge and Ryan (1958)
Neytzell-de Wilde and Taverner (1958)
Johnstone and Moll (1960)
Danilczuk and Swinarski (1961)
Karraker (1963)
Bracewell and Gall (1967)
Brimblecombe and Spedding (1974a)
Brimblecombe and Spedding (1974b)
Freiberg (1974)
Lunak and Veprek-Siska (1975)
Barrie and Georgii (1976)
Fuzzi (1978)
Type of System
Bulk
Bulk
Bulk
Support droplet
Bubbler
Bulk
Free droplets
Bulk
Bulk
Bubbler
Bubbler
Not reported
Theoretical
Flow
Supported droplet
Bulk
*
Comment
2
2
2
1
1,2
2
2
2
3
1
1
4
--
5
1
™* *"*

*
 1.   Incompletely characterized 2-phase system; results may be  incorrect  since
     the investigators did not completely account for mass transfer  rate  or
     demonstrate that no concentration gradients existed,

2.    Rate expression not reported.    '•••-. i«    -i' i  *.:..».>

3.    0?-free system; results not applicable to tropospheric SO,, oxidation.

4.    Insufficient details reported to determine if the results  should be
     considered reliable.

5.    Photochemical initiation.
                                        2-31

-------
the order  Is  unity.   NeytzeTl-de Wilde and Taverner  (1958) did not present a-rate expression
and constant  for the  Fe(III)  system.  An  estimate derived from their  paper  is presented in
Table  2-12.    Brimblecombe and  Spedding  (1974a)  reported  a rate  expression  and  constant
measured at a constant pH = 4; unfortunately, they used a plastic reaction vessel, which could
have released organic inhibitors into the system, causing the rate to be diminished.  (At pH =
4, their  rate is  0.25 of that  of Neytzell-de  Wilde and Taverner, 1958, and  0.1  of that of
Fuzzi,  1978.)  Fuzzi  (1978)  did  not  note the  similarity of his  observations to  those of
Neytzell-de Wilde  and Taverner  (1958), especially  the dependence of the  rate  on the initial
inverse H  concentration for  pH < 4.0.   Fuzzi's (1978) rate expression  has been modified by
incorporating  the  dependence  on  [H  ]   and is  presented  in  Table 2-12.  Note that Fuzzi's
(1978) modified rate constant is 2.5 times greater than that of Neytzell-de Wilde and Taverner
(1958), which  is  good agreement for  this type  of measurement; these two studies appear to be
definitive for the  Fe(III) system, and there is no basis to prefer one over the other.  Fuzzi
(1978) has clearly  demonstrated  the change in the reaction order of [$(!¥)] from 1 to 2 as pH
increases from 4  to 5.  The change  in  kinetics  is due to  the  formation of colloidal Fe(OH),
for pH  >  4,  which  provides an, explanation for the disagreement  among  earlier  investigators.
 •p    •>•»•».',,* J"M*~ i£i  ""'.  ' ""*  *     ' '    '
Because of the formation  of  the  Fe(OH)-  colloid,  it  is unlikely that a meaningful Fe(III)
catalyzed  rate  expression for  use in  tropospheric  sulfate formation can be  stated  for con-
ditions in which pH > 4.

                 TABLE 2-12.   RATE EXPRESSION FOR THE IRON-CATALYZED OXIDATION
                Expression9'
 pH
Investigators
         dt
                M
                =  0.04 [Fe(III )] [SUV)]
                =  100 [Fe(III)] [S(lv)]
                ~
         dt     =0.1 [Fe(III)] [S(IV)]
         dt
<4
 Adapted from
  Neytzell-de
  Wilde and
  Taverner (1958)
 Adapted from
  Brimblecombe and
  Spedding (1974a)
 Adapted from
  Fuzzi (1978)
       The units are:  liter, mole, second.
     K            4- — 1
       The term [H ]   indicates that the rate is dependent on the inverse of the
                   •!•                                +
          initial H  ion concentration; changes in H  concentration after the reaction
          is in progress do not affect the rate.
                                             2-32

-------
     The  Cu  catalyzed  reaction kinetics  have  been  described  in the  early work  of  Titoff
(1903).    The  pertinent investigations  are  identified in  Table  2-13.   As with the  Mn  and Fe
studies,  most  of  the  Cu studies  were performed  with  incompletely characterized  systems.
Fuller  and  Crist (1941)  point  out that the prior work  is unreliable  because of the  likely
presence of contaminants.   However, the investigations of Fuller and Crist (1941)  were carried
out in  a  two-phase  reactor whose mass transfer characteristics were not completely described;
no one  since  has conducted a more  definitive  study of this stystem.  The reagent concentra-
tions used  by Barron  and O'Hern (1966) are orders of  magnitude too large,  and  the pH range
(>8)  used by  Mishra  and  Srivastava (1976)  is  not  applicable.   For that  reason,  no rate ex-
pression  can  be  recommended  as  reliable for use in calculating sulfate formation  rates due to
Cu catalysis  in the troposphere.   Also, because of interference from the electric  motors in
sampling  devices,  reported airborne  Cu concentrations may  be .unreliable (Patterson,  1980).
     Vanadium catalysis  has  been  reported  in  only one study (Bracewell  and Gall,  1967);  a
bubble  reactor  was  used,  and  its  mass  transfer characteristics  were  inadequately reported.
Therefore, no  rate  expression  can be  recommended  as reliable.    However,  Bracewell  and Gall
(1967)  did observe  qualitatively " that "V~W was 'Srdefs crf'yagnituVfess1 effective 'than Mn and,
Fe; unfortunately,  they did  not study V(IV).  It is likely'that V(V) catalysis is unimportant
for sulfate formation  in  the troposphere.  "There are  also no  definitive studies  for Cr(III),
Ni(II),  Zn(II),  and Pb(II), but from  the  qualitative  work of Bracewell  and Gall  it appears
that these catalytic reactions  are unimportant.            ,           .
     Barrie and Georgii (1976)  have demonstrated qualitatively that Mn(II) and Fe(III) exhibit
a synergistic rate for the catalysis of S(IV) oxidation.   Their rate expression cannot be con-
sidered reliable, since they use a supported droplet.
     In summary:
     1.    S(IV)  oxidation  rates  are significantly increased t?y Mn(II) and'Fe(III).   There is
          serious doubt  regarding  the rate  expression  for Mn(II),  but  the  agreement among
          independent studies is much better for Fe(III).
     2,    These  systems  are presently inadequately  characterized:   Cu(II), V(V),  Ni(II),
          Zn(II), Pb(II), and Cr(III).  No studies of V(IV) have been reported.
     3.    There  are no quantitative studies of metal ion-metal  ion synergism.
     4.    The ability  of  atmospheric  organic compounds to inhibit the  catalysis is unknown.
     5.    All  studies have  been  performed in  the absence of  HCO.,; however, the  reactions
               so  + HcOg -» Hco3
                HO + HCOg -* HC03 + OH                     ,             .
          may  be  important.   It  is possible  that such reactions may occur,  and  if, so, they
          would prevent the oxidation  radical chain  from  establishing, since  HCOg  is not a
          powerful oxidizer (Hoigne and Bader, 1978).
          In general,  the  rate  expressions for catalytic oxidation to form HgSO^ are not well
          established.
                                             2-33

-------
          TABLE 2-13.   INVESTIGATIONS OF S02 - COPPER - 02 AQUEOUS SYSTEMS

Investigators
Titoff (1903)
Reinders and Vies (1925)
Alyea and Backstrom (1929)
Johnstone (1931)
Albu and von Schweinitz (1932)
Fuller and Crist (1941)
Riccoboni et al. (1949)
Bassett and Parker (1951) . >.,.- ._A«,,>.
Higgins and Marshall (1957)
Johnstone and Coughanowr (1958)
Junge and Ryan (1958)
Barren and O'Hern (1966)
Bracewell and Gall (1967)
Cheng et al. (1971)
Veprek-Siska and Lunak (1974)
Barrie and Georgii (1976)
Huss et al. (1978)
Mishra and Srivastava (1976)
Type of System
Bulk
Bulk
Bulk
Bubbler
Bulk
Bubbler
Bulk
., Bulk
Bulk
Supported droplet
Bubbler
Flow
Bubbler
Supported droplet
Flow
Supported droplet
Bulk
Flow
*
Comment
2
2
2
1
2
1
2
2
2
1
1
—
1
1
2
1
2
*"•*

 1.  Incompletely characterized 2-phase system; results may be incorrect since
     the investigators did not completely account for mass transfer rate or
     demonstrate that no concentration gradients existed.

2.   Rate expression not reported.
                                        2-34

-------
2.3.4.3   S(IV)  -  Carbon  Black -  0,  - H,0—The catalysis of  the  oxidation of dissolved S00 by
       '   """"**	              '" '	 ' ' £.	——^—                                                 ^
carbon particles  suspended in  the water,has been  studied by  Chang  et  al.  (1979) and by  Eatough
et  al.  (1979).  -It was  found  by Chang et al.  (1979)  that the oxidation  rate of  dissolved SOp
.species was:

          -   d[S(IV)]   =   k36[C]  [02]°-69[S(IV)]° expC-E^/RT)             (2-36)

with  an activation  energy  of E = 11.7  kcal/mol  over -the pH  range of 1.45  to 7.5  for the
carbon studied, which was Nuchar-190.   (The investigators demonstrated that Nuchar-190  behaved
                                                                                  •  •         5
similarly to  soot from acetylene and  natural,gas flames.)   An average value of k = 1.17 x 10
mol '   x liter '  /g-sec was reported.  «T.he  rate-limiting  step has  been suggested to be the
formation of  an  activated complex  'between-molecular  oxygen and the  carbon surface (Chang et
al.,  1979;.Eatough et al., 1979).
      Chang  et al. (1979)-have estimated that for 10 ug of their fine carbon soot suspended in
0.05  g of liquid water  and  dispersed  in 1 m  of  air,  the atmospheric sulfate production would
be  about 1 ug/hr.   Heavy hydrocarbons are ads'orbed on the  surfaces  of atmospheric soots and
may inhibit the carbon-surface catalyzed oxidation of  dissolved SQp,  At  this time,- it  remains
to  be demonstrated that the  laboratory soots used by Chang et al. (1979)  correspond to'those
                                                                                         3
present  in  the atmosphere  or that the suspension  of soot  at  ambient  levels  (<10 ug/m ) in
aerosols, cloud droplets, or rain is similar  to the  laboratory system,
2-3.4.4   S(IV)  -  Dissolved  Oxidants - H,,0--Hydrogen peroxide, 0,  and NOp may be important in
the oxidation of SOp in aqueous aerosols  and  fogs.   These  compounds are  not highly reactive
with  SO- in  air, but their reactivity is enhanced  in the liquid phase.   Again,  caution is
advised  in  accepting the results of studies of  two-phase  systems in which the  investigators
have  not completely  accounted for  the possibility of the  mass  transport limitation  of the
oxidation rate.   Therefore,  only the  recent results for  single-phase  systems are discussed
here.
      Martin et al. (1981) used a stopped-flow reactor  to investigate  the  kinetics of oxidation
of  aqueous  SO^ species  by  aqueous  NO,  NO™,  and  NO,.   Over the pH range of 0.6  to 3.2, they
found*for NO  and  N03 that the  disappearance of  S(IV)  species is:

          -   dCS(IV)]    =  k3?[NO or NQ~]  [S(IV>]                           (2-37)

                   k3?   =  0.01  mole  I'V"1.

                                                                                         9-
Hpwever,  for  the  same conditions,  the reaction with NOp is  rapid  and the formation of  SO. can
be  expressed  as:

            d[S04~]            4-05
           	3?	 =  k38CH J    [HN02  + N02]  [S02*H2° *  HS03]           (2"38)
                  k3g =  142 (liter/mole)1'V1
                                              2-35

-------
The N09 is reduced quantitatively in this reaction to N90.  Martin et al. (1981) also observed
                                                               9+
that this  reaction  is not catalyzed by  Fe(III),  Mn(II),  or VO  .   It is unlikely that tropo-
spheric HN02 concentrations are high enough for this reaction to be important for HUSO, forma-
tion.
     The  oxidation  of dissolved  S02 by  03  has been  investigated  with  stopped-flow systems.
Penkett (1972)  and  Penkett et al. (1979) have interpreted their work in terms of a decomposi-
tion of 0- to initiate a  free-radical  chain  reaction involving OH,  HSOo,  and  HSOr radicals,
after Backstrom (1934).  Penkett et al .  (1979) suggested that the rate expression is
                                       +  ,
                    •• 1
where  k-g  = 71  sec  .   Erickson  et  al.  (1977) reported the  fractional  contributions  of the
oxidation  of  the three  sulfur oxide  species  by ozone  at  various pH  values;  their rate ex-
pressions are:
                   =  k40[SVH20:i [03]
            dt     =  k41[HS03] [03]                                       (2-41)

          d[S02~]           2_
            dt     =  k42tS03 1 t°3^                                       <2-42>

where  k4Q  = 590  liter/mol -sec,  k41 =  3.1 x  10  liter/mol -sec,  and  k42 =  2.2  x 10  liter/
mol -sec.
     Penkett et  al.  (1979)  used a stopped-flow reactor to determine the kinetics of oxidation
of dissolved S02  species  by H202-  It was  found that the rate  of sulfate formation is given
by:
          d[S02~]                   _    +
            dt/    =  k43[H202] [HS03] [H ] + k43a[H202] [HS03] [HA]       (2-43)
where k43 =  2.6 x 10  liter /mol2-sec, with k43 and k43g being the third-order rate constants
for  the  catalysis by  hydronium  ions  and  proton-donating buffers  (HA),  respectively.   At
pH £ 4,  it  is  found  that  ^43^43  >  3200.    Therefore,  the  second term  is  probably  not
important for  acid aerosols  and fogs.   It  is  of  great significance  that  the reaction rate
increases as  the solution becomes  more acidic,  which  is in contrast  to  aqueous  oxidation by
metal ions and by 03.   The activation energy and the effect of ionic strength on the reaction
have been measured by  Penkett et al. (1979).  Dasgupta (1980) has criticized the presentation
of  Penkett  et al.  (1979);  use  of the  rate  expression  (Equation  2-43) takes  into account
Dasgupta1 s (1980) points.  Martin and Damschen (1981) have found that:
                                             2-36

-------
                                [S0-H0]/{0.1 + [H ]}                     (2-44)
            dt         4422     22

                     4 ~1
where k». =  7.2 x 10 s  ; their expression is applicable over the range 0 dissolves in the water,
                NH3(g) l NH3(aq)              +
     2.    The dissolved NI-L/  ^ reacts with H , which raises the pH

                NH3(aq) + H+ ^ <
     Therefore,  the  ambient pathways  of  auto-oxidation,  Mn(II)-  and Fe(III)-catalyzed oxida-
tion, and 03 oxidation  would  have  their rates enhanced by  absorption  of NH^.   However, the
ambient pathways of H,,Q2 and HN02 would have their rates  retarded by NH3 absorption.  The  rate
for soot would  not be influenced.
     NH3  can play other important roles.  Reinders and Vies  (1925) observed  qualitatively  that
Cu(II) was complexed by NH3 and  rendered  noncatalytic.  At high pH's (>9) where  NHg,   ^ is the
dominant  form,  NhU may be oxidized by 03  and free-radicals (Hoigne and Bader,  1978).
     In summary, the role of NH~  is  explained in terms  of its  influence  on the pH of  the water
system; NHL  is  not a catalyst.
                                              2-37

-------
2.3.5  Surface Chemical Reactions
     Industrial emissions  of solid particles  (e.g.,  fly ash) and  fugitive  dust (e.g.,  wind-
blown soil  and minerals)  provide a  solid-surface  that may  chemisorb  SCL  and  yield sulfate
ions.  This section  will  review investigations of  the  S02 oxidation on the surfaces of metal
oxides,  fly ash,  charcoal, and  soot.   Reaction  kinetics  have not  been  reported,  but  two
general types  of  processes have been recognized:   a capacity-limited reaction for S02 removal
and a catalytic SCL  oxidation  process.  The initial  contact  of SOp with the solid produces a
rapid  loss  of  SO,  from  the  gas phase;  the   reaction rate  decreases with  time.    For  the
capacity-limited  reaction,  the rate  slowly approaches  zero;  for the  catalytic precess,  the
rate levels off for a time and then approaches zero.  The latter phenomenon is attributed to a
pH decrease caused by hLSO, formation.
     Urone et  al.  (1968)  and Smith et al.  (1969) found a number of solids to be effective in
removing SCL.   In Urone's  studies,  S02 was admitted  to a flask containing  a  powder that  was
allowed to  react  with  no  mixing, and the product and remaining SO- were determined.   Only the
average reaction  rates  can be  calculated from  these  experiments;  more  importantly,  with this
experimental procedure  the rates may be diffusion-limited.  The highest  rate determined  was
for  SO,  with  Fe,SO,;   the  value was  >75 percent  per  minute.  Other  materials found  to  be
slightly less  reactive than Fe^SO, were Fe304,  PbO, PbOp, CaO, Al-03.  The rate for the ferric
oxide experiment was for 20 mg of Fe203 in a 2-liter flask; the Fe-0, concentration would thus
be 10  ug/m •  Assuming a direct proportionality between rate and particle concentration,  the
SO,  removal rate  in the  atmosphere would be calculated to be 0.04 percent per hour for  100
    3
ug/m  of particles with the same reactivity as  ferric oxide.  However, since the mass transfer
characteristics of the  reactor  were not reported, these results cannot be considered reliable
for estimating rates.
     Smith  et  al.  (1969)  did  not  focus  on  sulfate  formation kinetics;   instead,  they
illustrated through  a   novel experiment  the ability  of solid particles to  adsorb S0? and  to
release SO, during passage  through a tube with  a wall that adsorbs SO,.   They measured  the
number of SQy  monolayers  absorbed on  suspended Fe,0,  as a function of SO,  partial  pressure.
The  monolayer  coverage data reported in their Table  I  are  irr error by a  factor of 100  too
                                                                     •y
large; e.g., the number of monolayers at 1.13 ppm should be 0.38 x 10 .
     Chun and  Quon (1973)  measured the reactivity of  Fe203  to S02, using  a  flow  system  in-
volving a filter  containing suspended particles.   They  determined  a removal rate constant of
        "~^     —i     —i
9.4 x 10    ppm   min   (-din  8/dt), where 0  is  the fraction of surface  sites  available  for
reaction.   Extrapolating  this  to an  atmospheric  particle concentration of  100  yg/m  with  an
                                                                      o
equivalent  reactivity  and an  SO,, concentration  of 0.1 ppm (260 yg/m  ),  the  data project  an
atmospheric removal rate of 0.1 percent per hour.
     Stevens et al. (1978) report total iron concentrations in six U.S.  cities ranging between
0.5 and 1.3 yg/m  .   Other species such as manganese, copper, or vanadium had total  concentra-
                              o
tions  usually  below  0.1   jjg/m  .   Thus  actual  ambient  air concentrations  are  approximately
                                             2-38

-------
l/50th those assumed by the authors in the above papers.   A reactive particle concentration of
      Q
2 ug/m   would yield  a  predicted  SO,  removal  rate  of  no more  than 0.002  percent  per hour.
Therefore, surface  reactions  are  probably not important except in sources prior to or immedi-
ately after emission.
     The most comprehensive  studies  to date on S0? removal by pure solids were made by Siege!
et  al.  (1974)  and Judeikis  et  al.  (1978).   A  tubular  flow reactor,  in which  solids  were
supported on  an  axial  cylinder,  was used to measure reactivities of MgO, F^o^S' ^2^3' ^n^?'
PbO, NaCl, charcoal,  and fly ash.  They  found  that  the  rates of  SO,  removal  diminished with
exposure  until  the solids completely  lost  ability to react with  SO,.   The  relative humidity
was important in determining the  total capacity  for S0?  removal, but not the initial rate of
uptake;  total capacity  increased  as relative humidity increased.   The capacity for SO, could
be  extended  by   exposure  to  NH,.   This type of behavior  is consistent with  the formation of
H?SQ. on the surfaces.
     Because  carbonaceous matter  is  so  common in  ambient  air paniculate  samples, various
studies  have  been  made  of the S0?  removal rate  by carbon.   A comparison  of  the results is
difficult because  of  the varieties of carbon available for study, such as activated charcoal,
graphite, acetylene flame products, and combustion products of diesel oil and heating oil.  In
regard to  investigations  that deal with the gas-solid reaction of S0~ with carbon, Novakov et
al.  (1974) performed  laboratory  experiments  that  showed that  graphite and  soot  particles
                                                                          2-
oxidize SOg  in  air.  The soot exposed to humidified air  produced more SO.  than that exposed
only  to  dry air.   For downtown Los  Angeles,  they observed a  strong  correlation between con-
                                       2~
centrations  of  ambient carbon  and SO. , which  supports  their  hypothesis  that carbon (soot)
                                            2-
oxidation of SO, is the major pathway  for SOI  formation (see Section 2.3.4.3).
     Tartarelli   et al.  (1978) studied  the  interaction  of  SO, with  carbonaceous  particles
collected  from  the flue  ducts of  oil-burning power  stations.   They concluded that the amount
of  adsorption is increased by the presence of  oxygen and water  in  the  gas  stream.   Reaction
rates were not determined in this  study.
     Liberti  et  al. (1978) studied the absorption and  oxidation of S0£ on various particles,
including  soot  from an  oil  furnace  and  various  atmospheric  particulate samples.   They con-
cluded that the main  interaction between  the  S0? and PH is  adsorption,  with most catalytic
reactions  occurring at  high  temperatures, near the  combustion source.  Their experiments with
atmospheric  particulate  samples  led  them to the  conclusion  that any heterogeneous nonphoto-
chemical  sulfate formation  is  strongly  dependent on the reactivity  of  the particle surface,
and hence the history (aged, freshly emitted), of the aerosol.
     In summary:
     1,    Surface  reactions   are  capacity-limited.   Those  that  involve catalysis  in liquid
          films  can be extended by the absorption of NH~.
     2.    The initial rates may be large, but quickly approach zero.
                                             2-39

-------
     3.    Except for  the carbon  (soot)  reaction, solid  surface reactions do not  seem  to  be
          effective pathways for hLSO. formation in the troposphere.
2.3,6  Estimates of SQgOxidation
     At this point  it is interesting to compare the rates of SO, oxidation by the more impor-.
tant reactions  identified  in  the previous sections of  this  chapter.   The important reactions
for gas-phase and aqueous-phase oxidation are listed in Table 2-14, and rates of SO, oxidation
for an  assumed set of  conditions are present.   These  calculations ignore the nonhomogeneous
nature of  the  troposphere  and assume that  all  of the reactants are well  mixed.   (The more
general  case is treated in Chapter 6.)
            TABLE 2-14.   ESTIMATES OF S02 OXIDATION RATES IN WELL-MIXED TROPOSPHERE



I.



II.








Reaction
Gas Phase
HO radical
H02 radical
CHgOa radical
Aqueous Phase, pH:
Mn(II) catalysis
Fe(III) catalysis
C (soot) catalysis
03 (40 ppb)
03 (120 ppb)
H202 (1 ppb)
H202 (10 ppb)
_-i
Rate, % h """

0.3 - 1.3
0.4 - 2.0
0.3 - 1.5
135
1E-1 1E+1 1E+3
5E-5 5E-1 5E+3
3E+1 3E+1 3E+1
2E-8 2E-6 2E-4
6E-8 6E-6 6E-4
2E-2 3E-2 3E-2
2E-1 2E-1 3E-1
Discussion
Section

2.3.3.2
2.3.3.2
2.3.3.2

2.3.4.2
2.3.4.2
2.3.4.3
2.3.4.4
2.3.4.4
2.3.4.4
2.3.4.4

*
Comments

a
a,b
a,b

b.c.d
c.e.i
f,i
c,g
c,g
c,h
c,h





,i







NOTE:
*
a.
b.
c.
"E" denotes "exponent to

Typical range for daytime
This reaction rate is not
Assumed that liquid water
the base 10" (e.g., 3E-1

at northern midlatitudes
= 3 X 10"1)

during the summer.






well established; see discussion section.
volume of aerosol = 50 x
10"12m3/m3, [50,1
= 10
ppb
          (or 26 ug/m3).
      d.  'Assumed that Mn(II) mass concentration = 20 ng/m3; also, the Mn(II) is assumed
          to be uniformly dissolved in the liquid water of the aerosol (CMn(II)] = 8.9 x
          10  M).  Rate calculation used the expression of Neytzell-de Wilde and Taverner
          (1958); see Table 2-7.
      e.  Assumed that Fe(III) mass concentration = 2 M9/m3"» also, the Fe(III) is assumed
          to be uniformly dissolved in the liquid water of the aerosol ([Fe(III)] =0.9 M).
          Rate calculation used the expression of Neytzell-de Wilde and Taverner (1958);
          see Table 2-9.
      f.  Assumed that C mass concentration = 10 |jg/m3 and behaves as the soots studied by
          Chang et al. (1979), whose expression was used for this calculation (Equation 2-36).
      g.  Rate calculation was based on Equation 2-39.
      h.  Rate calculation was based on Equation 2-43.
      i.  Influence of inhibitors has been ignored, but they are likely to suppress the
          rate by orders of magnitude.

                                             2-40

-------
                                                                                            2
     For this comparison,  it  has been assumed that the S0? concentration is 10 ppb (26 [jg/m )
                                                                                    -12  3  3
for all  of the  reactions  and the  liquid water  content of the aerosol  is  50 x 10    m /m .
     The  gas-phase  rates  were  calculated  from  the  material  in  Section 2.3.3.2, while  the
aqueous-phase rates were derived from Sections 2.3.4.2-4.  Unbased assumptions include:
                                                    3
     1,   The ambient mass  concentration  of 20 ng/m  for Mn is reasonable, but: (a) it is not
          known if the predominant form is Mn(II), and (b) it is unlikely that Mn is uniformly
          distributed and dissolved.   Inhibitors  have been ignored, but  they likely  suppress
          the rate by orders of magnitude.
     2.   Likewise, the ambient concentration of 2 (ag/m  for Fe is reasonable, but:  (a) it is
          not known  if Fe(III)  is  the predominant  form, and  (b)  it is  unlikely  that Fe is
          uniformly distributed  and  dissolved.   Inhibitors have been ignored, but they likely
          suppress the rate by orders of magnitude.
     3.   There is no basis to assume that the rate equation observed for laboratory-generated
          carbon (soot) applies to atmospheric carbon.  Inhibitors may be important.
     4.   The rates for  the H0? and CI-LOp  reactions  recommended  by Calvert et al. (1978) are
          not well established.
     It is very  likely that the inhibitor-free rates estimated for Mn(II) catalysis, Fe(III)
catalysis, and C  (soot)  catalysis are gross overestimates.  Also, the HO, and CH.,0? rates may
be too high.
     Uncritically accepting all  of  the rates, at  a  pH = 3, and [H^O^] = 10 ppb, the SO^ con-
version  rate would  exceed 40 percent/h.   However,  if  only  the well-established rates  are
considered, the S02 conversion rate becomes vL.l percent/h.
     In summary:
     1.   The gas-phase  reaction rate of HO  and  the aqueous-phase reaction  of H?0p  are well
          established, but  are  expected  to account  for only  about 1.1  percent/h (under the
          conditions given in Table 2-14).
     2.   The  Mn(II),  Fe(III),  and C  (soot)  catalyzed  reactions  have  sufficient  rates to
          dominate S0? oxidation in the troposphere, but  the  assumptions discussed above may
          not be reasonable.
2.4  PHYSICS AND CHEMISTRY OF  PARTICIPATE MATTER
     Knowledge of  the  physics and chemistry of particulate  matter is necessary for design of
satisfactory  samplers  and monitors,  understanding  the  relationships  between  sources  and
effects, and understanding important processes in the troposphere that involve chemical trans-
formations and removal.
     In Section  2.2,  the global cycle and  annual  budget for sulfur were  presented to aid in
establishing  the goals" and  limitations  of this document's treatment  of  sulfur oxides.  That
discussion is incomplete, in that particulate matter  is  related to cycles of numerous elements
and  their interactions.   Among  the most important cycles of  elements are:  sulfur, nitrogen,
carbon, hydrogen,  boron,  oxygen, sodium, aluminum, silicon, phosphorous, chlorine, potassium,
                                             2-41

-------
calcium, vanadium, manganese,  iron,  mercury, and lead.  Since  it  is beyond the scope of this
document to deal with the details of these cycles, a perspective can be obtained form a budget
estimate of  the particulate  mass  injected  into  the troposphere.    The  estimate by  Hidy and
Brock (1971) of  the  daily particulate mass  emitted  or  formed in the troposphere is presented
in  Table  2-15.  Globally,  the anthropogenic  contribution  is about  6 percent;  however,  the
nonhomogeneous distribution and  the  type of emissions pose  serious  problems.   (See Chapter 4
for a discussion  of  sources in the  United  States.)   Figure 2-2 showed the general  interrela-
tions of  pathways,  processes, and  properties  of  sulfur oxides  and particulate  matter and
effects.   Section  2-3  treated the  SO,  physical properties  and chemistry  (including  trans-
formation chemistry), which are indicated in Figure 2-2.   Section 2.4 will discuss the physics
and chemistry  of  particulate  matter that are  related to  particle  properties,  single-particle
dynamics of motion, formation and growth, and aerosol system dynamics.
     The budget  in Table 2-15 indicates that secondary particulate matter formation dominates
the rates.   This important source will be discussed in Section 2.4.
2.4.1  Definitions
     Aerosol science  spans chemistry, physics,  engineering, meteorology,  and  the  biological
sciences.   Unfortunately, the lack of communication among workers in these diverse disciplines
has impeded the unification of their ideas.   One of the results has been a lack of universally
accepted definitions of the terms "aerosol"  and "particle" and the  terms for classification of
aerosol  systems.  The  definitions  used here are  consistent  with general  usage by atmospheric
scientists.
Particle:   Any object having definite physical  boundaries in all  directions, without any limit
     with respect to  size (Cadle,  1975).  In practice,  the particle size range of interest is
     used to define  "particle."   In atmospheric sciences, "particle" usually means  a solid or
     liquid subdivision  of  matter  that has  dimensions greater than  molecular  radii (~10 nm);
     there is also not a firm upper limit, but in practice it rarely exceeds 1 mm.
Aerosol:   A  disperse system  with  a  gas-phase medium and  a solid or  liquid  dispersed phase
     (Fuchs, 1964).  Often,  however, individual  workers modify the definition of "aerosol" by
     arbitrarily  requiring limits  on individual particle  motion  or  surface-to-volume ratio
     (e.g., see Hidy and Brock, 1970).  Aerosols are formed by (a)  the suspension of particles
     due to grinding  or atomization, or (b)  condensation  of supersaturated  vapors  (Fuchs,
     1964).
An  aerosol  is not the halocarbon vapor  used  as the  propellant in  pressured  cans (commonly
referred to  as "aerosol  cans" and  "aerosol bombs").   Improper use of the  term "aerosol" by
marketers  of  foams,  gels,  sprays,  etc., has  caused the lay public  to  associate incorrectly
environmental issues of suspended particulate matter with the issue of halocarbon impact Dn the
stratospheric ozone layer.  In the context of this document, the term "aerosol" is not related
to the impact of halocarbons on the stratospheric ozone layer.
                                             2-42

-------
  TABLE 2-15.   ESTIMATE OF GLOBAL TROPOSPHERIC PARTICULATE MATTER PRODUCTION RATES3
          Source                               •        % by Weight of Totalb


     A.   Natural Sources

         1.   Primary

             Windblown dust                                   9.3
             Sea Spray                                       28
             Volcanoes                                        0.09
             Forest Fires                                     3.8

         2.   Secondary

             Vegetation                                      28
             Sulfur Cycle                                     9.3
             Nitrogen Cycle                  .                14.8
             Volcanoes (gases)                                0.009
                                          SUBTOTALS:
a Source:   Hidy and Brock (1971)

  Production rate = 10.7 x 10  metric tons/day

  Not 100% because of round-off errors.
                                          SUBTOTALS:          93


     B.   Manmade Sources

         1.   Primary

             Combustion and Industrial                        2.8
             Dust from Cultivation                            0.009

         2.   Secondary

             Hydrocarbon Vapors                               0.065
             Sulfates                                         2.8
             Nitrates                                         0.56
             Ammonia                                          0.028
                                              TOTAL:          99C
                                        2-43

-------
     Traditionally, workers  in various scientific fields have  classified  the aerosol systems
to  reflect  their origin, physical state,  and  range  of particle size.   The  meanings of these
classifications  are not  universally  accepted;  however, the  following- definitions  'are con-
sistent with general usage by atmospheric scientists.
Particulate Mass;   A generic classification in which  no distinction is made  on  the basis of
     origin,  physical  state,  and range  of particle  size  (Dennis,  1976).    (The  term "par-
     ticulate" is an adjective, but often it is incorrectly used as a noun.)
Dust:   Dispersion  aerosols  with  solid  particles  formed by  comminution or disintegration,
     without  regard to  particle  size (Fuchs,  1964;  Dennis,  1976; Hidy and  Brock,  1970).
     Typical  examples  include (a) natural  minerals  suspended  by the action  of  wind, and (b)
     solid particles suspended during  industrial grinding, crushing, or blasting.
Smokes:  Dispersion aerosols  containing  both  liquid and solid  particles  formed by condensa-
     tion  from supersaturated  vapors (Fuchs,  1964; Hidy and  Brock, 1970).  Generally,  the
     particle  size  is  in the range of 0.1 urn to 10 pm.  A typical example is the formation of
     particles due to incomplete combusion of fuels.
Fumes:  Condensation aerosols  containing  liquid or  solid particles  formed by condensation of
     vapors produced by  chemical  reaction of gases or sublimation (Dennis, 1976').  Generally,
     the particle size  is  in  the range  of 0.01  ym to 1 urn.   Distinction  between the terms
     "smokes" and "fumes" is often difficult to apply.
Mists;  Suspension  of liquid  droplets formed  by condensation  of  vapor  or  atomization;  the
     droplet  diameters  exceed 10 iim  and  in general  the pariculate  concentration is not high
     enough to obscure visibility (Hidy and Brock, 1970).
Fogs:   Same as "mists",  but the  particulate  concentration is  sufficiently high  to 'obscure
     visibility (Hidy  and Brock,  1970).   [Dennis (1976) proposes  alternate definitions that
     distinguish "mists" and "fogs" on the basis of particle size.]
Haze:  An aerosol that impedes vision  (Dennis,  1976) and may consist of a combination of water
     droplets, pollutants, and dust (Hidy and Brock,  1970).                     ,
Smog:   A  combination of  "smoke"  and  "fog."   Originally, this  term referred to episodes in
     Great  Britain  that were  attributed  to coal  burning during  persistent foggy conditions
     (Chambers, 1976).   In  the United States "smog" has  become associated with urban aerosol
     formation during periods of high oxidant concentrations.
Cloud:  A  free aerodisperse  system  of any type  having a definite  size and  form and without
     regard to particle size (Fuchs,  1964).
Primary particles (or  primary  aerosols):   Dispersion aerosols  formed from particles that are
     emitted directly  into  the air and that do not change form in the atmosphere (MAS, 1977).
     Examples include windblown dust and ocean salt spray.
Secondary particles  (or  secondary aerosols):   Dispersion aerosols that form in the atmosphere
     as a result  of chemical reactions, often involving gases (HAS, 1977).  A typical example
          2-
     is SO.  produced by photochemical oxidation of SO,,.
                                             3-44

-------
     In addition to  classifying  aerosol  systems by their  properties  (origin,  physical  state,
size), systems are classified  according  to the performance  characteristics  of the sampler or
analyzer.   Some of the more common classifications used are the following:
Aitken nuclei:   Those particles  and ions measured  by  means  of an  instrument  in  which  water-
     vapor is made to condense on particles by  supersaturating  the vapor (White and Kassner,
     1971).  In  order to  eliminate condensation on light ions, the supersaturation should not
     exceed 270  percent (White  and  Kassner,  1971).   The term "condensation nuclei"  is often
     used synonymously with the term "Aitken nuclei."
Total suspended  particulate  (TSP) mass:   The particulate  mass that is collected by the high-
     volume sampler.   (The system is classified in terms of the operational characteristics of
     the sampler.  See the discussion in Chapter 3.)
Coarse and fine particles:  These two fractions are usually defined in terms of the separation
     diameter  of a  sampler.   For the  dichotomous sampler  (see  Chapter 3),  the separation
     diameter has usually  been set at 2.5 urn.   Thus, for the dichotomous sampler, the "coarse
     particles" are those collected by the sampler with aerodynamic diameters greater than 2.5
     pm; the "fine particles"  are those collected by the sampler with diameters less than 2.5
     [jffl.   (NOTE:  separation diameters other than 2.5 urn have been used.)
     Additional  definitions  that  relate  to particle  size,  particle  size  distributions,  and
particle motion  will be  provided in the  context of the material  discussed .in the following
sections.
2.4.2  Physical Properties of Gases and Particles
     To understand the behavior of an aerosol, it is necessary to know the physical properties
of  the gases  and particles.   Such  knowledge is  necessary to designing  particle samplers,
understanding  the  effects of  aerosols  (e.g.,  loss of  visual  quality),  understanding aerosol
processes  (such  as coagulation,  growth,  deposition), and modeling the effects and dynamics of
aerosols.
2.4.2.1  Physical Properties of  Gases—An  aerosol consists  of two  principal  components:   the
gas-phase medium and  the solid or  liquid dispersed phase.  The behavior of aerosol systems can
be described in terms of the behavior and interaciton of these two components.
      For tropospheric aerosols,  the gas  of interest is "air."  The  molecular and fluid pro-
perties of air are  well established and  will  not be reviewed here  (see  Hirschfelder et al.,
1954;  Bird et al.,  1960).  The  fluid  motion of air,' especially  laminar  flow, is adequately
understood.   Presently,  turbulent flow  is formulated  in  statistical  descriptions, and often
the  flow  fields  in  complex  geometry cannot be  satisfactorily predicted.   This limitation in
theory  has seriously  affected our ability to describe  the  tropospheric  microscale motion of
particles  with  diameters  greater than  10  (jm; specific  problems  include the performance of
particle samplers and the  formulation of particle dry deposition.
                                             2-45

-------
     In summary:
     1.   The physical properties are adequately known.
     2.   Laminar flow of air is adequately understood.
     3.   Turbulent  flow  must be described  in terms of a random  fluctuating component;  this
          limitation seriously affects our ability to'describe particle motion, especially for
          particles with diameters greater than 10 pm.
2.4.2.2  Physical Properties of Particles—The physical properties of particles that influence
behavior are  divided into three types (Billings and Gussman,  1976):   physical configuration,
bulk material properties,  and surface properties.
2.4.2.2.1  Physical  configuration.   The  shape, structure,  and density are physical configura-
tion properties  that are  very important parameters  in  the  equations of motion for particles.
     The shape  of particles  is  highly varied.  Tropospheric particles have  been  reported to
have the following types of shapes:   spherical, irregular,  cubical, flake, fibrous, and conden-
sation floes.  Particle shape is related to source type, as shown in Table 2-16.

                       TABLE 2-16.  PARTICLE SHAPES AND SOURCE TYPES3

               Shape                                        Examples

             Spherical                                   Smoke, pollen, fly ash
             Irregular                                   Cinder
             Cubical                                     Mineral
             Flakes                                      Mineral, epidermis
             Fibrous                                     Lint, plant fiber
             Condensation floes                          Carbon, smoke, fume

        a Whitby et a!., 1957.

     The physical  dimensions of  particles are  usually expressed  in  terms of  an equivalent
statistical diameter.  For such a measure to be meaningful  for nonspherical particles, it must
be  applied  as an  average to a  statistically significant number  of particles (Cadle, 1975).
For sizing collected particles,  the most widely used "diameters" for irregular particles are:
     1,   Martin's  diameter:   The distance  between  opposite sides of  the particle,  measured
          crosswise  of  the particle,  on a  line that bisects the projected  area  and that is
          parallel  to  a  reference  line.   (For examples,  see Cadle, 1975;  McCrone and Delly,
          1973.)
     2.   Feret's diameter:   The  distance  between  two tangents  on  opposite sides  of the par-
          ticle and parallel to a reference line.  (For examples, see Cadle, 1975;  McCrone and
          Delly, 1973.)
     3.   The maximum  horizontal  intercept:   The  longest diameter  from edge  to  edge of the
          particle, parallel to the reference  line (McCrone and Delly,  1973).
                                             2-46

-------
     4,    The projected  area  diameter (British  Standard method);   Found  by comparing  the
          projected area  of the particle  with the  areas of  reference  circles on  an  ocular
          graticule (McCrone and Delly, 1973).
     These four "diameters"  are used in powder technology.   However, they are less useful  for
applications relating  to  particle  dynamics.   They describe  the  shape in terms of one  or  two
dimensions (i.e., projected  surface  area).   The dynamics of particle motion are formulated in
terms of  the diameters of  three-dimensional  spheres.   The  relation between the  "diameters"
measured for projected surface areas of irregular particles and the "diameter" meaningful  for
particle  motion  and  light scattering  is  not obvious.   For regular-shaped  particles  (e.g.,
cubes, cylinders, oblates),  Fuchs  (1964)  has derived dynamic  shape  factors that permit their
representation as equivalent spheres.
     The  density  (mass per unit volume) of  particles is important  because  it affects  motion
and behavior.  The  density  of particles that are spheres, cubes, and other regular geometries
is the same  as  the density of the bulk material.   However,  many particles are agglomerates of
smaller particles of  various composition.   A  large  percentage  of  the volume of such agglome-
rates is  voids  or  air-filled pores.   Such  a structure  has the appearance  of a cluster of,
grapes.    The sum of  the  volume of  the small  solid/liquid  particles plus  the void  volume is)
defined as  the  "apparent" volume of the agglomerate.   The  "apparent density" of agglomerates
is defined  as the  ratio  of solid/liquid mass  to  the apparent volume (Fuchs, 1964),  and it i«
often 2-10 times lower than the density of mass that excludes the pore volume (Hesketh,  1977)J
     Because many tropospheric particles  are irregular or agglomerates  and have unknown dem
sity, it  is common  practice to represent the shape, structure, and density of particles in
terms of  dynamically equivalent  spheres  of unit  density.   Hence,  the  following definition!;
Aerodynamic  diameter:   The diameter  of a sphere  of unit density (1 g/cm  )  that  attains  the
same terminal velocity at low Reynolds number  in  still  air as the actual particle under con-
sideration.
2.4.2.2.2   Bulk material  properties.    The bulk material properties that affect  aerosol  be-
havior  include  chemical  composition,   vapor pressure,  hygroscopicity and  deliquescence,  and
index of  refraction.   These properties are  of interest because they control  (a) the physical
state and growth,  and (b)  the scattering  and absorption of  light  by tropospheric particles.
     The  chemical  composition of tropospheric particles  will  be discussed  briefly in Section
2.4.5 and  in more  detail  in Chapter 5.  It  is sufficient to point out here that the particles
                                                                        2-    +       +
with  diameters  less  than approximately 2.5 urn  contain  most  of the SO. , H ,  and  NH. and a
significant  fraction  of  the  NCL and Cl ;  therefore, these  particles interact with  H^O vapor
much more strongly than larger particles (Meszaros,  1971; Char!son et al., 1978).
     The  most important  systems  are those  of HgSO^,  NH-HSO^,  and (NH^SO^.  The  most  im-
portant aqueous  systems  are those containing  H+s  NH*,  So|  , N03, and Cl  since these species
are usually  present in sufficient mass  to control the liquid water concentration and the phase
transition  points  of particles as a function  of relative humidity.  Presently, phase diagrams
                                             2-47

-------
for  this multi component  system are  not available  for  conditions relevant  for tropospheric
                                                 +   +   2-
particles.  However,  the  phase diagram for the H -NI-L-SO. -HJ3 system at equilibrium is shown
in  Figure  2-7.    On  this  diagram,  the  dry,  pure crystals  (NH,)2S04,  (NH.)3(HSO.)2,  and
(NH.)HSO. are indicated as points A, 8, and C, respectively.  If letovicite, (NH4)3H(S04)2, is
exposed  to  relative  humidities that start at 0  and increase to 100 percent, its behavior can
be described  in  terms of the  locus BO in Figure 2-7.  From  0  percent r.h. , the salt immedi-
ately enters the 3-phase zone consisting of (NH.KHCSQ.)-, (NH^pSO., and some liquid solution
of H+, NH4, and SO. .  At point D, the locus intersects a phase boundary for (NH4)3H(S04>2 and
a  partial  deliquescence occurs.   Between the  point  D and  the intersection with  curve EE,,
solid  (NH.^SO.  remains;  however,  at  the  intersection  of  EE,,   a  second and  complete de-
liquescence  occurs.   From EE-,  to point  0,  only the  solution phase  is  present.   In similar
fashion,  equilibrium trajectories  for  salts  subjected  to  varying composition  and relative
humidity  can be  traced.  The locus  EE2 demonstrates  the  dependence of  the  complete de-
liquescence point  r.h.  on  the weight percent  H»,SO  •  as  the system's acid composition changes
from  0  to 35 percent,  the complete deliquescence  point  r.h.  changes  from 80  to 39 percent.
Thus, it is obvious that NI-U plays a key role in governing the phase transition points.
     The H2S04,  NH4HS04, (NH4)3H(,S04)2> and (NH4>2S04 particle systems have been characterized
recently by  Char! son  et al.  (1978), Tang et  al.  (1976),  Tang and Munkelwitz (1977), and Tang
(1980a).   Charlson et al.  (1978)» used an apparatus in their  studies  that  measured the light
scattering coefficient  of  the aerosol  as a function  of  the relative humidity.   They obtained
good agreement between  theory and experiment in observing  the hygroscopic  behavior of H?S04.
However, they observed no deliquescence point  for  NH4HS04 particles,  and  one  at ^38 percent
r.h. for (NH4).jH(S04)p particles.   The bulk salts have a deliquescence point at 39 percent and
69 percent r.h., respectively.   They suggested that  the  deliquescence point for NH4HS04 par-
ticles may not have been exhibited because the  initial  droplets which were only  dried to 15
percent  r.h.  entered  a hysteresis  loop, forming  supersaturated  solution droplets.   Their
observation  of  a  growth point for  (NH4)3H(S04)2  particles at  ^38 percent r.h.  but not the
deliquescence point at  68  percent r.h.  is difficult to explain, but it may have been due also
to the particles forming supersaturated droplets and entering a hysteresis loop.  If that were
the case, then the true deliquescence point at 69 percent r.h.  would not be exhibited, and the
observed transition  at ~38  percent may be  an efflorescence  point.   Efflorescence points of
solution droplets  exposed  to decreasing  relative humidity are not  sharp and usually occur at
relative humidities more than 30 percent below the deliquescence  point.   Tang (1980a) used a
system in which  the salt aerosol  was first dried,  and then passed through a controlled humi-
dity chamber; the  particles  were  sized with a  single-particle optical analyzer.   His data on
deliquescence points  and hygroscopic growth  agreed  well  with theory, as  is shown in Figure
2-8, and he concluded that for the NH4HS04-H2Q systems that the equilibrium size of mixed-salt
droplets may be  adequately predicted from bulk solution properties.   (The Kelvin effect, which
                                             2-48

-------
   100
8
 M
*"?r
z
o
Hi
    30
    20
    10
                                  30        40

                                WEIGHT % H2SO4
70
 Figure 2-7. Solubility diagram for the H+—NH4*-SO42-—H2O system at
 equilibrium (30°C).
 A = solid phase of (
 B = solid phase of (
 C = solid phase of (NH4)HSO4
    = liquid solution phase
 a^ = activity of water
    = fractional relative humidity
 y =  mole fraction of (IMH^gSC^
 The numbers in parentheses are the fractional relative humidities for the complete
 deliquescence points that are indicated.
 Source: Tang (1980a).
                                     2-49

-------
PO
                         2.4
                         2.2
                        2.0
                     UJ
                     a   1.8
                     u
                     N
                         1.6
                     o   1.4

                     E
                         1.2
                         1.0
                          50
 COMPOSITION, wt%
 (NH4)2S04   H2S04

 • 100
 A 88.3
 O 83.3
	 0
     60              70              80

              RELATIVE HUMIDITY, %
90
                     Figure 2-8.  Growth  of  H*—NH4*—SQ42~  particles  as a function of a relative
                     humidity. The solid curves represent theory and the points are experimental
                     observations.
                     Source; Tang(1980a).

-------
may limit growth, is discussed in Section 2.4.4.)  Tang (1980a) also demonstrated the hysteresis
phenomenon for  (NH.^SCL  particles;  it can be seen  in  Figure 2-9 that as  the  particles  were
subjected to  decreasing relative humidity,  solid  crystals did not form  at the deliquescence
point (79.5 percent r.h.), but formed at about 30 percent r.h, (in agreement with the observa-
tions of Orr et al., 1958).
     The  deliquescence  and efflorescence  points of salt  particles relevant to environmental
investigations are presented in Table 2-17.
            TABLE 2-17.  DELIQUESCENCE AND EFFLORESCENCE POINTS OF SALT PARTICLES6

Composition
NaCl
KC1
NaCl-KCl
(NH4)2S04
(NH4)3H(S04)2
NH,HSO,,
4 4
2NH4NQ3-(NH4)2S04
Deliquescence, % r.h.
75.7
84.3
73.8
79.5
69.0
39.0

56.4
Efflorescence, % r.h.
43
53
38
•v36




       Adapted from Tang (1980a).
     It will be pointed out in Section 2.4.5.2 and in Chapter 5 that the composition of tropo-
spheric particles  may consist of many species.  In general, it is not known if the individual
particle  is essentially  a  pure compound  in a population  of  particles  of various compounds
(external  mixture) or  if the individual  particle  contains a  mixture  of compounds (internal
mixture)  (Winkler, 1975).  The composition of the individual particles governs their size as a
function  of  the  relative  humidity.   For  two-component droplets  smaller than 0.05  (jm,  the
surface curvature  affects the vapor pressure of the droplets, and their size is related to the
relative  humidity  through the modified Kelvin-Gibbs equation (Nair and Vohra, 1975):
                ln(p/pQ) -  In
                             2Mcr  ri    x dp   3x da-,
                             RTpT Li   p dx " 25 dxj
(2-47)
where
p  = vapor pressure over the curved surface
p  = vapor pressure over a flat surface
a° = surface tension
M  = molecular weight of water vapor
p  = density
x  = solute mass fraction
                                              2-51

-------
    2.4
 o

51
13
 *
111
o
01
N



a
u

oc

2
    2.0
    1.8
    1.6
    1,2
    1.0
      20
                    THEORETICAL



                    EXPERIMENTAL
                        .•Op-
                             	n	—.
                30
40
50        60        70



 RELATIVE HUMIDITY,%
                                                                    80
90
100
  Figure 2-9. Condensational growth and evaporation o

  humidity at 25°C.



  Source: Tang(1980a).
                                                             particles as a function of relative
                                              2-52

-------
     R  =  universal gas constant
     T  =  temperature
     r  =  radius
    a   =  activity of water
     w
     Nair and Vohra  (1975)  added the terms in  the  square brackets on the right side of Equa-
tion  2-47  to  correct  the  Kelvin-Gibbs  equation  for  the variation  of surface  tension  and
density  as  a  function of  concentration.   The  behavior  of particles  that are  external  and
internal mixtures  can  be  described  by Equation 2-47.   However, the  Nair and  Vohra  (1975)
modification has been  recently challenged (Renninger et  al.,  1981)  on the basis that nuclea-
tion  theory  may have  been  used  incorrectly  to derive the thermodynamic expression Equation
2-47.   Presently,  the  issue has not been resolved  (Doyle,  1981) and the validity of Equation
2-47  is  not  known.   In general, the activity  of water (a ) is known only for pure salt solu-
tions,  with  notable exceptions,  such  as ocean water, which makes  the  equation (if correct)
immediately applicable only to external mixtures.  There have been attempts (Fitzgerald, 1975;
Hanel and Zankl,  1979; Sangster and Lenzi, 1974) to calculate a  for multicomponent electro-
lyte  solutions  without resorting to detailed  theory.   Hanel  and  Zankl (1979)  compared  the
results  for  two  mixture rules:  (a) the  mass  of water condensed on  the mixed electrolyte is
equal to the sum of the masses  of  water condensed  on the  separated pure components, and (b)
the  practical  osmotic  coefficient is  equal  to the molality-weighted  practical  osmotic  co-
efficient of the  separate^  pure components.  The first mixture  rule gave the better acuracy,
but  to  limit errors to 10-15  percent,  Hanel  and Zankl advised that it be used only for water
activities  larger than 0.85  to  0.9,  which corresponds  to relative  humidities  of 85  to 90
percent.  Thus,  for  known  internal mixtures,  the   change  in  droplet size  as a  function of
relative humidity can be calculated adequately only for high humidities  (greater than 85 to 90
percent).
      Ideally,  the growth  behavior  (especially at  deliquescence points)  of particles as  a
function of  relative  humidity offers a means of distinguishing internal/external mixtures and
of identifying the predominant salts.  However, both the likelihood that atmospheric particles
also  exhibit hysteresis by  forming supersaturated salt solutions and our inability to predict
accurately  the  activity, of water for  multicomponent  mixtures  for  relative humidities below
about 85-90 percent militate against this approach.
                                                                      -19       -12
     The growth  of H?SO.  droplets  with  dry  mass  in  the range of  10    to 10   g and their
concentrations (normality)  are shown in Figure  2-10,  which is based on Equation 2-47 and the
data  given  in  Table 2-18 (Nair  and  Vohra,  1975).   However, as previously pointed out in this
section, the validity of Equation  2-47 has been questioned;  thus,  the  information in Figure
2-10  must also  be questioned  until the issue of the validity of Nair and Vohra1 s (1975) modi-
fication is resolved.  As Figure 2-10 shows, in the absence of atmospheric NHn, H^SO^ droplets
are  highly acidic.  For example, at relative humidities <40 percent, the concentration of H£SO
                                             2-53

-------
    100
n   80 —
£
Q
LU


1

Ul
cc
    20  —
      0	
   Figure 2-10. The equilibrium size of su If uric acid solution droplets as a function of relative
   humidity. The mass of sulfuric acid (from 10"^ to 10"^ gram) is indicated on the growth
   curves (solid). The normality (N) of the solution is marked for selected values from 1 N to
   21 N (dashed lines).
                                            2-54

-------
                  TABLE 2-18.   SULFURIC ACID SOLUTION VALUES (25°C)

Mass %
x
0.5
1.0
5.0
10.0
20.0
25.0
40.0
50.0
66.0
85.0
P
g cm"
1.000
1.004
1.030
1.064
1.136
1.175
1.299
1.391
1.560
1.773
dp/dx
x 103
8.1
7.4
6.7
7.0
7.5
7.9
8.8
9.8
11.3
8.7
a
dyn cm
72.0
72.1
72.3
72.6
73.7
74.3
75.8
76.8
75.2
68.7
do/dx
x 102
12.0
8.6
5.5
8.0
11.4
11.3
10.0
6.4
-30.0
-72.0
aw
0.998
0.996
0.982
0.958
0.880
0.823
0.555
0.340
0.075
0.001
H20/H2S04
mole ratio
1081. 0
537.8
103.2
48.9
21.7
16.3
8.1
5.4
2.8
1.0
Normality
0.102
0.206
1.03
2118
4.65
6.00
10.7
14.2
21.0
30.6
Source:  Adapted from Nair and Vohra (1975)
                                        2-55

-------
is >14N.  Increases in the relative humidity up to 100 percent cause great reductions in H,SOA
                                              -16
concentrations for  droplets  with dry mass >10   g.   At  100 percent r, h. , these droplets grow
without bounds  and become infinitely  dilute.   However,  droplets with  HpSO.  dry mass <10   g
have their  growth  restricted by the Kelvin-Gibbs effect, which is expressed in the right-hand
side of Equation 2-47.  Such droplets have a critical radius at relative  humidity >100 percent
                                                                                    ~19
that prevents  unbounded growth.   As an  example,  the droplet  with dry  mass  of 10   g H?SO.
attains a radius of 0.004 urn and  concentraiton  of 6N at 100 percent  r.h.;  it has a critical
radius of M).005 pm which can be attained at ~114 percent r.h.  As long as the relative humid-
ity does not exceed 114 percent, the droplet1 s size is governed by the portion of the curve to
the left of the critical radius (0.005 pm); if the relative humidity becomes >114 percent, and
the critical  radius  is  reached,  the  droplet  will then  grow without  bounds  as long  as the
relative humidity  is  XIOO  percent.  For ambient air and breathing, the relative humidity does
not exceed percent.  Thus, from Figure 2-10, it is easy to see that small H9SO, droplets (with
             -16
dry mass  <10   g)  might pose serious  threats  to health, as  hypothesized by Stauffer (1974).
If inhaled, such droplets would attain H7SO. concentrations of 2 to 6N for dry masses of 10"
     -19
to 10   g, respectively, if not neutralized by NH.,.
     Cl  and NO, may be' displaced as HC1 and HNO- from solution droplets  or thin aqueous films
on particles  by HpSO,.   Although the  order of  the relative acid  strengths  is  HC1 > HoSO. >
HN03 (see Table  2-4), their vapor pressures do  not have the same order.  The acid-gas vapor-
pressures over aqueous solutions are given by;

PHC1      =                               '
PHN03
            5 x 10"6 atm M~2 (aH+)(ac1-)
            2 x 10"7 atm M~2 (aH+)(aNQ-)
                                      +3
PH SO    =  % SO ^aH SO * = % SO )(K1,H SO *  (aH+)(aHSQ-)            (2-50)
  24         2424       24      24             4
         =  3 x 1Q~U atm M~2 (aH*)(aH*)(aSo2"^
                                    i      _^-a
     The Henry's Law constant H,,»-|(= atm M  ) was calculated from the free energy change for
HC1,  £ HC1, y (Cotton and Wilkinson, 1980).  The value of HHNQ (= 4.8 x 10"6 atm M"1) has
been reported by Schwartz and White (1981).  Only recently has H,,SO», * vapor pressure become
available for aqueous solutions (Roedel, 1979; Chu and Morrison, 1980).  The value of HH cn
        -R      -i                •                                                     n?iU4
(=3 x 10   atm M  ) was estimated from Roedel 's (1979) data for 82 percent (mass) H2S04> Thus,
from these vapor pressure equations it can be seen that the vapor pressure of HpSO,/ % will
                                             2-56

-------
be orders  of magnitude less than  HMO-,  -,  and  HC1, , for  tropospheric  particles.   In acidic
solutions, hLSQ.  is effective  in  volatilizing HNCL  and HC1 ,  causing these  acids  either to
reside  in the  gas , phase  or to  recondense on  less  acidic  particles (Harker et  al.,  1977;
Hitchcock et al, , 1980).  .                   ,  -                ,
     The  influence  of  NO,  is an important consideration.  Tang (1980b) has performed detailed
                •*         J
qualitative calculations of  the partial  pressures of NH-, and HNO^ over the NH--HNO--HoSO»-H?0
system at 25°C.  His calculations for the partial pressures PNUO and -Pimno over this system at
85 percent relative humidity^are shown  in  Figure  2-11.   The solution composition ranges from
pure H2S04 ([NH*]/[SC)|~] = 0)  to pure, (NH4)2$04  ([NH*]/[$o|~]  = 2).- The PRN03  has almost a
linear dependence  on  [NH.]/[SO. ],  while P^LIO^S not sensitive to this parameter, but depends
only on  solution  pH.   Figure 2-11 demonstrates the  complex interactions among NH,, HNOo, and
HpSO..   He also studied the effects, of relative humidity and pH on these partial pressures and
deduced  that:   (1) the HNO- .partial  pressure depends strongly on both  the relative humidity
and  droplet pH,  and  (2)  the , NH,  partial  pressure varies only  slightly with  humidity but
                +
inversely with H  concentration.  Tang (1980b) remarked that the strong dependence of the HMO,
partial  pressure  on relative  humidity may  affect  the nitrate content  of 'particles that are
sampled, leading to biases in the> determination of ambient NCU.
     Charlson et  al.  (1978)  have reviewed the potential  use of the difference in the indices
of  refraction  for  differentiating  ambient  particles  of HpSO, and  (NH.)?SO».   The  ratio  of
backward  hemispheric  scattering to the total  scattering was measured near St.  Louis in 1973
during  periods  that were  classified  as  H2$Q.  or (NH.^SQ. dominated events.   While the ob-
served  differences  were  generally in  the  right direction,  the average value of the backward/
total  scatter ratio did  not agree with prediction,  which led Charlson et  al.  (1978) to con-
clude  that  refractive index  is too complex  a variable  to be used as an  analytical tool for
                               2-
differentiating the types of SO,  systems.
     In summary:
     1.   Bulk  material  properties are  adequate in  most cases for  describing the  state  of
          tropospheric particles.
     2.   However,  there is a paucity of thermodynaraic data to permit prediction of deliquesc-
          ence and  hygroscopic behavior and  vapor pressures of multicomponent systems, especi-
          ally for  relative humidities below about 90 percent.
2.4.2.2.3   Surface properties.   The  surface properties  of particles  provide  means of detec-
tion,  measurement,  and collection; and may increase persistence of droplets in trie atmosphere.
Some of the more important surface properties are:  electrostatic charging, adhesion, and the
influence of surface films.                                           "  "
     A number  of  identifiable  mechanisms  can  lead to' electrostatic  charging  of, particles
including contact charging, photoionization,  field emission charging, and gaseous ion capture.
For  practical  applications in  the  troposphere,  gase'ous  ion capture  is  the most important of
these  mechanisms.   Contact also produces charging,  as  in  the  triboelectric  charging of dust
                                             Z-S7

-------
   1000
    100
•a
a

LU
cc


1
Ul
K
a.
re

<
a.
10
    0.1
   0.01
0.5           1.0



 MOLAR RATIO CNH^
                                            1.5
                                                    2.0
   Figure 2-11. NHo and HNOg partial pressures as a function of

   droplet's nitrate fc|yOo"^ anc* su'^1*6 ^~^ concentrations

   at 85% relative humidity, 25°C.




   Source:  Tang(1980b).
                                    2-53

-------
blown  by wind  along the  earth's surface.   Bursting  bubbles may  produce charged  sea  spray
aerosols over the ocean.
     Reviews of  experiment and  theory for gaseous  ion capture  by aerosol particles may be
found  in papers  by Bricard  (1977)  and  by  Whitby  and Liu  (1966).   Two  recognized  capture
mechanisms are field charging and diffusion  charging.   Field  charging denotes the process in
which an  ion  is  captured by a particle through the influence of an external electrical  field.
Diffusion charging  is a process in the absence of an external electrical field.  Field charg-
ing of  aerosol particles is used in particle control technology for the operation of electro-
static  precipitators.   Diffusion charging is  employed  in classification  or sizing of aerosol
particles according to their electrical mobility.
     The  rate  of gaseous  ion capture by  an  aerosol  particle depends upon a  number of  para-
meters  including the particle  size and  shape,  the  dielectric constant  of the particle,  the
number  of charges  already on the particle, and the mean free path and mobility of the gaseous
ion, plus, for field charging, the external electrical  field strength.
     The charging of an aerosol  has been shown by Boisdron and Brock (1970) to be a stochastic
process.  Inherent difficulties  in the use of particle charging as an aerosol  detection method
have  been shown by  Marlow (1978a  and 1978b) and by Porstendorfer and  Mercer (1978).   These
studies  indicate that the polydispersity of  the  aerosol, the dielectric constant of the par-
ticles,  and  humidity in the presence of trace gases lead to uncertainties  in aerosol particle
charges.
     Particles are removed from  the troposphere by diffusion to or impaction against surfaces.
Particles also collide  with each other and stick together.   The forces of adhesion that hold
particles to  surfaces and to each other include electrostatic forces, capillary forces in the
presence  of  a liquid,  and  London-van  der Waals forces.  In  general,  for uniform conditions,
the  efficiency  of  particle removal from  surfaces  by air flow decreases  as the particle size
decreases for  dry,  solid  particles (Corn, 1976),   While  the types of  forces  are known,  the
magnitude of these forces usually cannot be predicted precisely.
     The  influence  of  surfacefl1ms  on aqueous  droplets  has been recognized  for many  years
(Bradley, 1955;  Eisner  et al., 1960).  Chang  and Hill (1980) have reviewed  some of the studies
on  droplet  stabilization by surface films.   They  have  also demonstrated that the products of
the  reaction between  0,  and 1-decene  in humid  air  contain species  that  adsorb  on  water
droplets  and retard the evaporation  rate,  Chang and  Hill  (1980)  suggest that photochemical
reactions may produce  similar  products  that  would retard the evaporation  of  urban fogs,  and
perhaps extend  their duration by hours.  However, they report no kinetic data.  Eisner et al.
(1960)   investigated  the  kinetics  of evaporation  of  droplets  with fatty  alcohols on  the
surface.  They were able to  increase  the  lifetime of an evaporating  10-pm droplet only by  a
factor  of about  250, which corresponds to about 2.5 minutes.  (Another likely cause of stable
fogs  is the  formation  of  supersaturated  droplets.)  At this  time,   the  influence of photo-
chemically  produced organic condensates on the  kinetics of droplet evaporation is not known.
                                             2-59

-------
2.4.3  pynarm' csof SI ngl e Part id es         ,-.,'••*.
     The behavior af atmospheric aerosols depends upon" the physical" properties' of the suspend-
ing  gas,  the  particles,< gas-particle interactions,  particle-particle interactions, and !the
fluid motion of  the  gas.   Knowledge'of  these properties  and interactions is essential  to our"
understanding  of  atmospheric phenomena,  our  ability to  formulate predictive models of pol-
lutant particle concentrations  and effects,  and our ability  to  sample and measure particles.
In this section,  the conditions will be  presented  for  which aerosol  systems can be described
in terms of the dynamics of single particles.
     According to Hidy and Brock (1970),  particles may be considered to be independent of each
other and the  dynamics  of single particles may be applied if the conditions in Table 2-19 are
satisfied.    If coagulation or  deposition are important processes, and the stystem satisfies
all but the  last condition,  they  r.efer  to the  system as being  in  the "quasi-single particle
regime."  As  is  seen  in  Table  2-19,   the conditions  for the  "single particle  regime"  are
generally satisfied for the troposphere.• ,           ,,        ,       ,  ,

                    TABLE 2-19.   CONDITIONS FOR THE SINGLE-PARTICLE REGIME3

Conditions
1.
2.
3.
4.

5.
n
X
R
n

Q
i
G
i

i
1
/n
n
n
i
,1
G
1/3
1/3
1/3L

Q

j
VI
K/kT
« 1
« 1
« 1
« 1

« 1
for
for
for
for

for
Range in Troposphere
• > <••(,*•• %
all, i , , :, , - ., ~ 10"13 to 10"19
all i ~ 10~3 to 10"5
all i,j " ~ 10"1 to 10"7
all i ~ 0

all i,j * ^ • - ~ 10"2 to 10"6 ''

a
b
From




Hidy

and Brock (1970).



        n.
        Ri
       Lvi
                                              19  "3
=  number concentration of air molecules (~ 10  cm  )
=  mean free path of air molecules-(~ 10   cm)
=  radius of particle, cm          ,          .  •         »
=  characteristic distance (cm) associated with change in numb,er concentration,
   temperature, and velocities
=  electrostatis charge, esu
         K   =  Debye reciprocal length (effective distance of Goulombic interaction), cm
         k   =  Boltzmann' constant (= 1.38 x 10    erg/K-mole)
         T   =  temperature, K
                                                                                         -1
                                             2-60

-------
     The dynamics  of single  particles  include  sedimentation,  impaction,  diffusion,  coagula-
tion, electrodynamics, and  filtration (Fuchs, 1964; Hidy and Brock, 1970; Friedlander, 1977).
In  general,  the  dynamics  can be  described adequately  for all particle  sizes in calm air.
According to  Fuchs (1964), complete  (~99 percent) entrainment of  particles  by eddy  fluctua-
tions occurs if

          t/tL < 0.02                                            (2-51)
where
          T  =  particle relaxation time
                  g
         V   =  initial velocity of particle in the absence of external forces, cm/sec
         V   =  terminal (steady) settling velocity of particle in still air, cm/sec
          s>
          g  =  gravitational constant
         t,   =  the Lagrangian period of eddy fluctuations, sec
             =  0.5rtX/u.
                       A
          X  =  scale length of eddy fluctuations, cm
         u,   =  root mean square (rms) eddy velocity corresponding to fluctuations on a scale
                < X, cm/sec.
     Using an  rms  eddy velocity of 30 cm/sec, which is a typical value ~1 'm above the ground,
and X ~  40 cm  at  X/u. ~ 1;  in this case, t/t.  ~  0.01 for a particle  (unit  density)  with a
      ~"               A                         U
diameter  of  10 urn.  That means particles with diameter < 10 urn will  be  >99  percent density
entrained in the atmospheric eddy fluctuations.  As the  rms eddy velocity increases above 30
cm/sec and the  particle approaches nearer to the surface or a large obstacle,  t., decreases, t
remains constant, and thus the particle diameter corresponding to ~99 percent entrainment must
decrease  from  10 urn.   It is important to note that larger particles with diameters >1 mm have
< 2 percent entrainment in  the eddy fluctuations for these conditions (Fuchs,  1964), which is
adequate  justification  for  ignoring the influence of  atmospheric  turbulence on their motion.
Thus, the following practical limits for considering atmospheric motion can be stated:
     a.   Diameter  <10 pm:   The  particles  follow the eddy motion  with 99 percent entrainment
     b.   10 Mm < diameter < 1 mm: The particles  lag behind the eddy motion
     c.   Diameter > 1 mm:  The particles do not  follow the eddy motion.
These practical limits suggest that in turbulent  atmospheres the dynamic behavior of particles
with  diameter  <10  urn  can be  described  in terms of viscous  flow  mechanics with superimposed
eddy flow.  The dynamic behavior of particles with diameters in the range 10 ym to 1 mm cannot
be  similarly  described (see Soo, 1967*,  Fuchs, 1964), and  satisfactory approaches are still
fertile areas for research.   The inability to describe completely the motion of such particles
in turbulent air has had an  inhibiting effect on  the design and use of aspiration samplers for
particles with diameter >10 jim (May et  al.,  1976).   A practical method  to reduce biases in
aspirating samplers caused  by  the inertial  effects  of flowing large particles  is to sample
                                             2-61

-------
1soklneticany,  which means  that  the air streamlines neither converge nor diverge upon enter-
ing the  sampler.   Isokinetic sampling is attempted by matching the inlet flow velocity of the
sampler  to  the  local  air  flow velocity.   However,  even under  ideal  conditions, the  probe
extends  a  disturbance  upwind of  the  inlet,  causing  entrance biases  for  particles  having
appreciable  inertia  (Fuchs,  1964).   In  the  turbulent atmosphere,  it is  not  practical  to
attempt isokinetic sampling which would require:  (a) a fast-response realigning inlet that can
maintain its axis  parallel  to the local, rapidly fluctuating wind vector, (b) a fast-response
pumping  system that  can  maintain the inlet  flow  speed  equal  to that of the wind's, and (c) a
thin-walled  inlet  (Belyaev  and  Levin,  1974).   The most common atmospheric  particle  samplers
(e.g., high  volume samplers, dichotomous samplers; see Chapter 3)  are  operated anisokinetic-
ally.   For  real,  fluctuating turbulent  atmospheres,  actual  trajectories  for  unit  density
particles greater  than ~10  urn diameter  cannot  be calculated; instead,  the trajectories are
estimated by  ignoring turbulence.  While turbulent wind  tunnel tests of  sampler inlet effi-
ciencies can  be performed  under  steady  conditions  (Wedding and Weigand, 1980;  Liu  and Pui,
1981; McFarland et al., 1977), the turbulence in the wind  tunnels in general is much lower than
atmospheric  levels.   The degree  of  correspondence between wind-tunnel  characterizations and
performance of aspirating inlets in turbulent atmospheres  is not known;  however, for particles
in the  low portion  (10-15 urn) of the  nonviscous flow range, significant  differences  in the
inlet's  entrance efficiency  are  unlikely.  For particles  in the complete nonviscous range (10
urn -  1  mm),  no  practical inlet has been demonstrated at atmospheric turbulent conditions.  To
date, it appears that the only reliable atmospheric particle size data for diameter >10 urn has
been  obtained with Rotorod and similar samplers that draw the impactor stages through the air
instead  of aspirating  air through a  fractionating  device  (May et al., 1976;  Noll  and Pilat,
1971; Johnson, 1976).
     The  influence of  the  variety  of elevations  and  temperatures  in  the United States  on
particle  formation,  growth,  and motion  must be considered.   The  important  mass transport
parameters for air that influence particle dynamics and that change with altitude and tempera-
ture  are atmospheric pressure (or density)  and  air viscosity.   Typical  values  for  several
elevations and temperatures  are shown in Table 2-20.
     As  shown in Table 2-20, viscosity is  independent  of  altitude over the range considered.
Within the contiguous  48 States,  the acceleration  due  to  gravity (g) varies within the range
                         p
of 9,790 to 9,809 m/sec ;   this variation  is  so  small  that it will  have  significant effects
only  on precise  fundamental  investigations.   The  general  shift  in particle behavior  as  a
function of  air  temperature, pressure,  and  viscosity is  given in Table 2-21.  The dependence
of particle mechanics on these variables  is well  known;  for many instruments and samplers that
use  particle mechanics  for  sizing  and  separating,  the effects  of changes  in  temperature,
pressure, and viscosity must be considered.
2.4.4  Formation andGrowth  ofParticles
      Particles  are formed by  two processes:   (1) grinding or atomization of matter,  and (2)
nucleation of supersaturated vapors.   The particles formed in the first process may be emitted

                                             2-62

-------
                        TABLE 2-20.   MASS TRANSPORT PARAMETERS FOR AIR

Elevation, km
0
1.52
(5,000 ft)
3.05
(10,000 ft)
Standard
Pressure3
kPa
10.33
(1.00 atm)
8.56
(0.83 atm)
7,07
(0.68 atm)
Density ,
-30°C
1.56
1.30
1.06
kg m
20°C
1.29
1.07
0.88
Molecular
Free Path
-30°C
54.0
64.8
79.2
Mean-
, nm
20°C
65.3
78.8
95.4
Viscosity0,
-30°C
1.54
1.54
1.54
_-i _-i
kg m sec
20°C
1.81
1.81
1.81

a  Fairbridge (1967)
b  Weast (1976)
c  Bretsznajder (1971)
NOTE:  -30°C = -22°F;  20°C = 68°F.

directly into  the atmosphere.   However,  the particles  formed in the  second  process usually
result from  reactions  of gases in the atmosphere to yield compounds with low vapor pressures;
when  such  species reach  sufficiently high supersaturation, they nucleate  to  form particles.
The  dynamics  of nucleation,  which are still  understood only  incompletely, have  been exten-
sively reviewed by Hidy and Brock (1970),  who discuss the two types described below:
     1.   Homogeneous nucleation  is  the  formation of particles by the molecular agglomeration
          of supersaturated  vapors  in the absence of foreign  particles  and ions.   Important
          examples  include the  formation  of  particles  by H2SQ.  molecules produced  by  the
          reaction of HO radical  with S02, and  carboxylic  acids  formed by the reaction of 0,
          and olefins.
     2.   Heterogeneous  nucleation  is  the condensation of molecules of a supersaturated vapor
          onto  foreign   particles  or  ions.   Important  examples  include the  condensation of
          hydrocarbon vapors  onto Pb halide and carbon particles during cooling of automobile
          exhaust, and the condensation  of HpSO. molecules onto fly ash during the cooling of
          plumes form power plants burning fossil fuels.   Heterogeneous nucleation occurs when
          foreign nuclei  are  plentiful and may suppress the critical supersaturation pressure
          below the critical value required for homogeneous nucleation.
     Particle growth in  the atmosphere occurs through gas-particle interactions, which will be
discussed  in this section,  and particle-particle  interactions (coagulation),  which are well
understood and mentioned above in Section 2.4.3.
                                             2-63

-------
   TABLE 2-21.   DEPENDENCE O'F PARTICLE BEHAVIOR ON AIR TEMPERATURE, PRESSURE,  AND VISCOSITY
      Behavior
Dependence for variation within the ranges given in Table 2-17
 Temperature, T        Pressure, p        Viscosity, n
  Nucleation rate3



  Condensation growth rate3

    a.   Nonvolatile species


    b.   Volatile species
  Sedimentation, velocity in
       calm air
  Impaction parameter, and
        stop distance

  Diffusion coefficient
  Electrical mobility
increases in com-
plex fashion as
T increases
increases as a
function of T
.1.5
complex; may de-
crease due to in-
crease in partial
vapor pressure of
volatile species

dependence appears
through n; as T
increases, n
increases
none
increases in a
complex fashion
as T increases

dependence appears
through n;, as T
increases, n
increases
        probably none
                                         none
        depends on p
depends on p
                                  -1
                   none
                           none
        none for viscous   depends on
        flow; slight de-
        crease for non-
        viscous flow (i.e.,
        diameter = 100 urn)
        as p increases
                                                                                     -1
        none for
        viscous flow

        decreases in a
        complex fashion
        as p increases

        increases in
        complex fashion
        as p increases
                   depends on
                   depends on
                   depends on
                                                                                     -1
                                                     -1
                                                                                     -1
a.   See Hidy and Brock (1970) for general discussion.

b.   See Fuchs (1964) for equations with dependence on T, p (or mean free path length), and
    n-
                                             2-64

-------
     Gas-particle interactions include  the  absorption and the adsorption of  pollutant  gases,
such as  SO,,  NOp,  hydrocarbons,  0.,, and H2CL,  followed by their chemical  reactions  to yield
products such as SO*  ,  NO.,,  and organic compounds.   Also  included  is the condensation  of low
vapor-pressure molecules formed  in  gas-phase reactions, such as HpSO,  and  organic compounds.
An important  limitation  on  the accumulation of chemical species on  submicrometer particles is
the Kelvin-Gibbs effect;  The vapor pressures of the  solvent  and  solute (or surface-absorbed
species)  are  increased  as  surface  curvature  is  increased.   For a condensing  species  being
formed by  gas-phase  reactions,  there will  be a minimum particle size below which condensation
will not occur; this value is determined in part by the supersaturation reached by the species
(Friedlander, 1977).
2.4.4.1  Growth Dynamics--Knowledge of the mass transfer process is  necessary to understanding
the growth of particles due to gas-particle  reactions in the troposphere  and  in laboratory
studies.   Since  the dynamics  of transfer  of  gases  to particles has been  presented"in great
detail by Hidy and Brock (1970),  only general features will be reviewed here.   The controlling
mass transfer processes can be identified through their relaxation times (see glossary), which
are:

     a.   the interparticle diffusion relaxation time, T

          lpp  -  ^i.air"1                                  <2-52>
          which is a measure of the time required for a cloud of particles to attain a uniform
          vapor concentration of species i, where only molecular diffusion is important.

     b.   the diffusion to a single particle, t

          *sp  =  •fy.alr"1' Kn<<1                            (2'53)
      or
           Tsp  =  Rv."1, Kn « I                                 (2-54)
          which is a measure of the time required for transport of species i to a single
           particle in a  stagnant gas, where only molecular diffusion is important.

     c.    the momentum to a single particle, T
           tm  =  R u^1                                          (2-55)
           which is the characteristic time  for transport on momentum to a particle by fluid
           motion.
           tsp  • V1  =  Pe                                      (2-56)
           where Pe is the Peclet number and is given  by
           Pe  =  fluid velocity x R x D.   .Jl,                  (2-57)
                                        1 9 31 I
                                             2-65

-------
     The Peclet  number  represents the ratio of convective and diffusion transport of gases to
a particle.   When this  value is «  1,  the mass  transport relaxation times  are governed by
molecular diffusion;  however,  for larger values, fluid convection, is  significant and t   and
T   will be reduced,
     where
                                                      -3
               N  =  particle number concentration, cm
                                                                          2
          D.  .   =  gaseous diffusion coefficient of species i in air, cm /sec
           I, ai r
               R  =  radius of particle, cm
              v.  =  the molecular mean speed of species i in air, cm/sec
              u   =  the relative speed between the particle and air, cm/sec
              Kn  =  Knudsen number = mean free path of air/R.

                                                                            ft  ~~"%
For particles with radius R < lOpm and' a-maxfmum number concentration N < 10 cm   , the relaxa-
tion times are
                    Tsp $ Tm ^ *pp * 10~3sec»
which  has  this  meaning:   The  magnitude of  a perturbation  in  the concentration  of gaseous
species i will be reduced by 1/e at-time ='t.  Thus, in the case stated above, which is reason-
able for lower  tropospheric aerosols, stationary conditions in the gas phase will.be achieved
               -3
in less than 10   sec.
     The relaxation  time T, for adsorption-and,solution at clean surfaces may be on the order
                           a
of that for relaxation  of internal  molecular energy, which  is  very  small  compared to T
However, as  mentioned  in  2.4.2.2.3,  surface  films- may  significantly retard  mass transfer
between the  air and  aqueous  phase.  The  mechanism and  the magnitude of  the effect  is  not
established; probably,  the  phenomenon is due to the organic molecules in the film aligning to
present their hydrophobic (paraffinie) parts to«the air interface.  Thus, the droplet takes on
a "paraffinic"  type surface,  which will exhibit  a much  lower condensation coefficient than
clean water  surfaces  to polar molecules such as HpO, NHj, and S0_.  The only relevant studies
of the  influence of organic surface  films on  T   have been conducted for the mass transfer of
                                                Q
HgO molecules  from  the  aqueous solution droplets  to  air  (see Section 2.4.2.2.3).  The signi-
ficance of  such films  in reducing T   for SOv,  and NH, have not  been  studied.   However,  the
                                      a         L.        &
effect  has  been proposed  by  Junge and  Scheich (1971) to explain  their  observations  of  the
simultaneous presence of H?SO,  droplets and NH, in London.  If organic surface films are able
to reduce T   to >5 sec  for absorption  and  subsequent reaction of  NH-,  in  the breath by H«SOA
           a     ^                                                   o                    £.  *T
droplets,  then their inhalation and deposition in the lungs may pose serious health threats to
humans.
     For chemical  reactions occurring  in  aqueous  droplets.,  the  diffusional  transfer  of  the
reactants must  be considered.   The diffusion" relaxation  time for  reacting  species  i  in  the
absence of-internal circulation is (Lamb, 1945):
                             p
                    Ton  =  £ /D.     (with no internal circulation)       (2-58)
                     •*
-------
where
          £  =  the thickness of the aqueous film on the particle, or the
                diameter of the aqueous droplet
      D.     =  diffusion coefficient of species .i in the aqueous,solution.
        i ,aq
     Since  atmospheric particles  are normally  not, subjected ,to forces  that  would maintain
oscillation or deformation of small droplets,-.internal .circulation needs to be considered only
for droplets with diameter > 10 jjm.  For droplets with internal circulation;

                  ~4        ~1
        T£D  =  ^  cm sec ^       (with internal circulation).       (2-59)

     Thus,  the  diffusional .relaxation  time of reactants for  likely  tropospheric droplets is
expected to be !„„ < 0.1 sec.
     For  aqueous  particles with  diameters less  than  1 pm,  the mass-  transport  processes of
reacting gases have relaxation times T < 10   sec, except possibly for t   for transfer through
                                       ""*                                a
an organic  film-air,, interface.   If the reactant half-lives are much less  than the diffusional
relaxation  time,  then  the droplets may be viewed as homogeneous  reactors  whose feedstock rate
is controlled by  the mass transfer through the interface.  Otherwise, the effects of chemical
gradients  in  the droplet  and in  the surrounding air  must be  included  (Satterfield, 1970).
2.4.4.2   Sulfuric Acid-Water Growth Dynamics—The growth rate of 100  percent HpSO. submicro-
meter  droplets  suddenly  exposed  to  humidified  gas  streams  has been measured  (Carabine and
Haddock,  1976);   it  was  found  that  the  growth  ended when  the water  vapor  equilibrium was
reached in  <6  sec.   Calculations that  ignore  the dissipation of the heat of  dilution of the
                                    -5       -2
droplet predict  growth times  of 10    to  10    sec.   Gentry and  Brock (1968) have performed
mass/heat  transfer  calculations for  0.1 (j™ diameter,  l.OM H-SCK droplets and found that the
heat transfer dominated the growth rate,  causing it to be much  less than  expected from assum-
ing  isothermal conditions.   Azarniouch et al.  (1973)  have  performed similar calculations for
supermicrometer H-SO.  droplets  and have also deduced that heat transfer significantly reduces
the  growth rate.   Thus,  it  is  reasonable  to expect that  inhaled concentrated H2$0^ droplets
may  require a  period of <6  sec. to attain their terminal size (governed by the relative humi-
dity in the lungs) and  final dilution.
2.4.4.3   Dynamics of  Growth by Chemical Reaction—Few  studies   have  been  reported  on  the
chemical  reaction rate for gases and particles that are of  interest in  the lower troposphere.
Attempts  to measure the SO^  oxidation  rate in free and  supported  droplets were discussed in
Section  2.3.4.    Because  of  the attendant  problems of  the inability  to separate collected
droplets  from  the outflow stream of  the reactor  for free-droplet studies  and possible radical
termination at solid  surfaces  for supported  droplets, those  studies  cannot  be  accepted as
reliable  for estimating mass transfer rates.  At  this time,  the  only reaction kinetics studies
for  gases and suspended droplets  that  are applicable to the  lower  troposphere are those for

                                             2-67

-------
the  NH~<- N  - HgSO^  droplet  system  (Robbins  and  Cadle,  1957;  Cadle  and Robbins,  I960;
Huntzicker et  al.,  1980).  'Using 98 percent and  12  percent H^SO, droplets with diameters 0.2
to  0.9  pra,  Robbins  and Cadle  (1957)  observed that the  reaction of NH., with the  98 percent
HgSO, droplets was  not  diffusion controlled, whereas the  reaction  with 12 percent H^SO, was.
They interpreted the slower reaction rate for the 98 percent H?SO. droplets (10 percent of the
collisions of  NH3 were  effective) in terms of surface adsorption followed by the formation of
an  activated  complex  that transmitted  the NH~  through  the  interface.   For the  12 percent
HgSO., 100 percent  of  the NH3 collisions were effective.   They did not consider the influence
on  growth  time of  heat transfer  and formation  of solid (NH4)2S04 as a surface crust; both of
these should be  expected  to  be  important  in  reducing the growth  time  of  concentrated H2SO,
droplets.  Huntzicker et al.  (1980) investigated the reaction for 0.3 to 1.4 (jm H2SO, droplets
at  8  to  80 percent relative humidity;  they found that the rates were between 21 to 70 percent
of  the diffusion-limited  theoretical  rate.   They interpreted this range of rates, in addition
to  the observed  decrease  in  rate as a function of time for low relative humidity, to indicate
the accumulation  reaction product  on  the  surface.   Possible heat transfer  efforts  were not
considered.  Cadle  and  Robbins  (1960)  also observed a  reaction  between N02 and Nad solution
droplets that was too fast to measure;  the products were NaN03 droplets and HClx  *.
     No  kinetic  studies  relevant  to  tropospheric  conditions have  been reported  for gases
reacting with  solid particles.   In Section 2.3.4, the  works  of Smith et al.  (1969), Chun and
Quon  (1973),  Siegel, et al.  (1974), and Judeikis  et al.  (1978) were mentioned.  The work of
Judeikis and coworkers  is of  most interest  because their  reactor design permitted  them to
estimate the effectiveness of  gas  collisions  for reacting.   They  investigated  the collision
effectiveness  $  of S02,  N02,  03, and CO  on  a variety of materials,  including  metal oxides,
salts, charcoal, cement,  fly  ash, and sand.   S02 and  N02 exhibited  high reaction  rates with
most materials.  Unfortunately,  the highest collision efficiency leading to reaction that they
                         -4
could measure  was $ <  10  .   (The value for a water molecule condensing on a water droplet is
$ < 10   ;  for such  values of $  > 10   ,  there is  little  difference for gas-particle reaction
rates in the lower troposphere if ~10   <$
-------
     In summary:
     1.    The transfer of reactive  gases  to lower tropospheric and  urban  particles is known,
          with the exception of the possible role of organic films.
     2.    The growth rate of  hLSQ,  droplets suddenly exposed to high humidity has been demon-
          strated by  experiments  and theory to  be  dependent on the heat  transfer  rate.   The
          nonisothermal  growth rates are  several  orders of magnitude greater  than  those pre-
          dicted assuming heat transfer to be unimportant,
     3.    Few  gas-particle  chemical reactions  relevant  to the  lower  troposphere  have  been
          reported,  which is hindering  the testing of theory.
     4.    No quantitative investigations  of S02  or NH3 desorption from  particles under condi-
          tions relevant  to the  lower  troposphere have been  reported.   One  study indicates
          that S0? may be carried (as sorbed species) on particles through  tubes with reactive
          walls in mcuh greater quantities than gaseous S0? can penetrate.
2.4,5  Characterization ofAtmospheric  Aerosol
     Significant  advances  have  been made  in  the  past decade in  regard  to  elucidating  the
nature of  the tropospheric particle size,  area, volume, and mass distribution functions  and
the chemical  composition.   This  section  will  discuss  general  aspects of distribution func-
tions,  the  observed  behavior  of  urban  particle  distributions,  and  chemical  composition.
Evidence will  be  presented  for the existence of multimodal mass distributions and the differ-
ence in composition of  urban  coarse and fine particles separated at about  2 urn, but it is not
a sharp division  of the chemical  composition.   One  of  the major reasons  for selecting 2.5 pm
as  the  separation point  is  that it occurs  in the  region  of the minimum  between the accumu-
lation (fine)  mode  and  the  coarse particle mode  in volume (or mass)  distributions of urban
particles.    However,  nonurban  particles  may  not  have volume  (or  mass)   distributions  that
resemble those  of urban  environments;  in  such  cases,  a  separation at 2.5 pm may not yield
differences in composition (see Figure  1-1),
2.4.5.1  PistributIon—The multimodal  nature (to be discussed below) of tropospheric particle
surface and  mass  distribution functions remained unrecognized  until the early 1970's largely
because of the methods used to present number,  size and mass distribution  data.
     Tropospheric particles are polydisperse.  For reason of convenience,  particle number (N),
area  (A),  volume (V),  and  mass  (M) concentration data are  usually expressed  in  terms  of a
mathematical  distribution  function of  diameter  (D).   Such  functions  are  ordinarily charac-
terized by two parameters.   The  fraction  of  the total number of particles  having diameters
which lie between D and dD is:

                              dN  =  f(D)dD                      (2-60)
with the normalization condition:
                      / °° f(D)dD  =  1.                          (2-61)
                                             2-69

-------
The curve  representing  the function f(D) is called the "number frequency distribution" or the
"number  differential"  curve.   Similarly,  the  area and  volume frequency  distributions  are:
and
dA  =  f(D2)dD                                    (2-62)

dV  =  f(D3)dD                                    (2-63)
where the proper normalization is taken into account.
     Prior to the  1970's,  atmospheric scientists employed two  predominant  types of frequency
                                                      2           3
distributions  [i.e.,  functional  forms of  f(D),  f(D ),  and f(D )].   They were  (a)  Junge's
(1955) power law  distribution for particle size and (b) the log-normal distribution for mass.
     Investigators who  used  instrumentation to measure  particle size spectra usually reported
their data in terms of Junge's (1955) power law, which is:

               {}}  =  AD~k                                       (2-64)

where A  and  k  are constants.  Clark and Whitby (1967) found that this power law was a reason-
able fit  to  the number distribution, but it was an inadequate model of the surface and volume
distribution.
     Also, investigators who used cascade impactors to determine the mass distribution spectra
usually  reported  their  data  in $erms of the  log-normal  distribution (see Fuchs, 1964; Cadle,
1975).   While  multimodal  log-normal  distributions  can  be  recognized from  standard plots of
data on  log-normal graph  paper,  three effects  combined to mask the  multimodal  character of
urban particle  mass  distribution:   (1) the cascade  impactors had  serious (and unknown) inlet
biases  against  particles  larger than  about  5 pm,  (2)  the cascade  impactors did  not  have
operational  characteristics  that  permitted mass  fractionation below about  0.2 \im,  and  (3)
particle  bounce distorts  the mass  distribution  in cascade impactors.   A warning  by Fuchs
(1964) that the log-normal  distribution is an adequate model  only if the particles are sampled
perfectly and the sampler provides adequate fractionation points went unnoticed.
     Whitby  et  al. (1972) were  responsible for  a major advance in  recognizing that Junge's
(1955) power law  and the  log-normal  distribution function were inadequate  models  of experi-
mental data  for urban  aerosols.   Instead  of  seeking other  functional  forms  to  express  the
number,  area,  and volume  distribution,  they plotted dN/d log  D,  dA/d log D,  and  dV/d log D
versus log D.  The result is seen in Figure 2-12.  This  type of plot has a convenient feature:
The area  under  the curve  is proportional to  the quantity  (N, A, or V) between two diameters.
The particle volume between the diameters D-^ and D2 is:

                        log D?
        V(D,, D?)  =  J      ^ (dV/d log D) d log D              (2-65)
                        log D
                                             2-70

-------
in
 o
tr
ui
CD
          12
     —   1.0
        JH 0.8
         Q.
        Q

        9

     r— 3 0.6
        (J
        < 0.4
     —   0.2
	     6
   J^  4
   D
   ca
   _O
   LU


   I
             . —     1
                               I  11 Mllll    TTT
                                           NUMBER

                                      	__ SURFACE

                                      — • — VOLUME
                          0.01
                          01           1


                       PARTICLE DIAMETER,/
                                                                    \ —

                                                                  .-1 H.L.
                                                                10
        Figure 2-12. Frequency plots of number, surface, and volume distribu-
        tions for 1969 Pasadena smog aerosol.


        Source:  Whitby, in National Academy of Sciences (NAS), Airborne
        Particles, (1977).
                                     2-71

-------
For small values of A log  D, Equation 2-47 becomes
               V(Dr D2)  =  (AV/A log D) x A log D    m                    (2-66)
where (AV/A log D) is the average value in the interval between log DI and log D«.
     The peaks  in  the three types of  distribution  plots  are called modes.  As  is  evident in
Figure 2-12,  there  is usually one number mode, one or two surface modes, and two volumemodes
for urban aerosols; sometimes an additional volume mode is observed in the range from 0.005 to
0.05 urn when  a strong source of fresh nuclei is close to the sampling site.   In strict usage,
a  mode  is a  single point  which  is a  maximum in a  frequency  distribution;  however,  aerosol
investigators  have  often used  it to  represent  the integral of  the  distribution between the
minima on each side of the maximum.   For  example,  the^ particle volume distribution in Figure
2-12 has maxima (modes)  at 0.3 and  8  urn.   However,  the integral  of  the particle volume fre-
quency distribution  with a  maximum at 0.3  urn and minima  at ~0.02 nm and ~2  pm is commonly
called the accumulation  mode.   The use of "mode" to  denote integrals over specific limits is
unfortunate,  but  now so  common that  change  is unlikely.   A reader must determine from the
context of an article whether the author  uses  "mode"  to mean a  single  point  or an integral.
     The urban  particle  volume  (mass)  distribution generally has 2 or 3 modes.   The integrals
associated with these modes are:   (1) the coarsemode,  which usually extends from ~2-3 urn to
~100 urn,  and  has  a maximum  in the range 5-50  pm; (2) the  accumulation mode,  which  usually
extends from  ~0.02  to ~l-5 urn, and  has a  maximum in the range 0.1-1.0 |jm;  and (3) the Aitken
(or nuclei) mode, which is sometimes observed near a strong source of fresh nuclei and extends
from ~0.005 to 0.05 urn.   These modes  result  from the differences in major source types.  The
coarse mode consists  mainly of  primary particles such as mineral  dust.   The accumulation and
Aitken modes  consist  of  primary combustion smoke and secondary particles.  The limitations of
particle growth processes  cause these  particles to "accumulate" in the size range from 0.2 to
2  urn.  The minimum between the coarse mode and the accumulation mode is generally not at zero
on the distribution  curve because the mineral dust size extends below 1 (jm,  and the growth of
secondary particles extends above 3 |jm.  Also, the minima  may shift in the region of 1-5 pm.
In spite  of  this  behavior, the  minimum between  the  coarse  and  accumulation modes  offers an
attractive position  to fractionate  particle  samples  for  health  and welfare  considerations.
However, it must be recognized that this mode of the urban particle volume (or mass) distribu-
tion is an idealization.   The chemical  separation between the accumulation mode (integral) and
the coarse mode (integral)  is not sharp.  Figure  2-13 shows the  more general  picture  of the
volume (or mass)  distributions  for a variety  of  types of locations and conditions that range
from  urban  to  mountain   background.   It  is  important  to   note  the influence  of  combustion
sources, secondary  sources,  natural  sources,  windspeed, and  turbulence  on  the  volume distri-
bution.  It is obvious that the distributions observed at a site will be a composite of these,
and  it  may exhibit  a large  degree  of fluctuation among the components.  Thus,  the relative
                                             2-72

-------
f\J
 I
            irf*
            10"
E
o
o»   1°2
o

                                 -FINE

                         I  I IliilJ  I  I I
         COARSE-

            L
                                                                   x
                                                                   0
                             %
                             2

                             O

                             tu
                             O

                             oc
                      10
                        ,-2
                       10
                         ,-1
10U
10'
102
10J
                             PARTICLE DIAMETER (D),/m>
                                                                10S
                                                                10"
                                                                       10
                                                                       10
                                                                       ID
                                                                         '2
                                                                       10
                                                                 '3
                                     11  I III

                                     -    b
                                               11 nun)  11 uniij  i i iuiii|  i  Mi]
                                                                FOREST FIRE:
                             —   "CLEAN"
                               COMBUSTION
                             -    //'
                                                                                                        /
                                                                                                        MECHANICAL-
                                                                                                          /     \
                                                                                                         /FLYASH\
                                                            U/S
                                                                                                 / 1 1  '
                                                                                                 /fit

                                                                                               /I

                                                                                             x  '/  '
                                                                                           >•' / /  /
                                     /  /

                                                                                     -FINE-
                                                       -j	COARSE	»~ —
                                  i mill  111 mill  111 mill  i  i nun!  i iimiil  i i  mi
10"
10'
102
                                                                                 PARTICLE DIAMETER (D),/um
          Figure 2-13a,Idealized size distribution for particles found in
          typical urban aerosols (mainly from anthropogenic sources)
          under varying weather conditions.  Note bimodal distribu-
          tion under usual conditions and shift in distribution (in-
          creasing fine-mode particles, decreasing coarse-mode par-
          ticles) under stagnation (1) and serious "smog" conditions,
          (2), respectively.
          Source:  Adapted from Slinn (1976).
                                                                 Figure 2- 13b,Idealized size distribution for atmospheric par-
                                                                 ticles from anthropogenic sources, showing fine particle con-
                                                                 tributions from "clean" high-temperature combustion, and
                                                                 coarse particle contributions from "dirty" fly ash sources,
                                                                 forest fires, and crushing and grinding operations.  Note
                                                                 change in distribution near sources (1) and at increasing
                                                                 distances (2,3,4) from sources.

                                                                 Source;  Adapted from Slinn (1976).

-------
IN*
I
             10°
             10"
             103
.a

I
         01
         o
u

D

P   10°


O

I   10'1
a



    10'2
            10
              ,-3
                =l  MNlllj  i I  Illllj  I f11 III! j  Milltllj  | I  ||llll|  I I

                ~*   c
                                             HURRICANE
                                    //"V-
         ,^-V   \  \\    i
                   NORTH-ATLANTIC
                    i mill  1
         E_j	COARSE—	»

\ii ii i mill  111 mill 111 mill  11
              10'1
               10
                        ,-2
10
,-1
10U
               10
,1
      10*
                            PARTICLE DIAMETER (D),Aim
                                     a
                                     O)

                                                                   Ul
                                104
                                                               105
                                                               10"
                                                               1015
                                     = l llllll]  I IIHH!) II 11 lilt)  I 11 Hill]  I II Hlli

                                     M*_*  J

                                                                DUST STORM f

                                                                           I
                                                                    o
                                                                    E
                                                                    5   102
                                                                        10'
                                          n-1
                                                                       10"'
                                                                ,-3
                                                                       10
                                                               10"
                                                                                                                 I
                                                                                  /
                                                                                 /,
                                                                                //
                                                                                                             I  I IIS
                                                                                         \
                                                                                         \ —

                                                                           ATYPICAL CONTINENTAL
                                                                           ~ (WITHIN MIXED LAYER)
                                                                                  I-
                                                                                  !~
                                                                                  f|

                                                                                  Iz
                                                                             .'ill
                                                                             \ll  -
                                                                                                          MOUNTAIN
                                                                                                         ELEVATIONS
                                                                                                     (ABOVE MIXED LAYER)
                                                I II
                                               I I
                                                          BACKGROUND

                                                              -FINE—j—COARSE	

                                                          il  1111 mi!  11 n mil  i 11 mill
                                                                        linn
                                                                          10
                                                                   .-3
10''
10
                                                             ,-1
10'
                                                             ,0
10
                                                                               PARTICLE DIAMETER (D),M«n
           Figure 2-13c.ldealized size distribution for atmospheric par-
           ticles from natural sources in a marine setting. Note, in com-
           parison to typical background levels over open ocean, in-
           creasing levels of coarse-mode particles ranging from those
           found in sea spray (1,2) to the extreme cases of storms (3)
           and hurricanes (4).

           Source:  Adapted from Slinn (1976).
                                                                Figure 2-13d.Idealized size distribution for atmospheric par-
                                                                ticles from natural sources in a continental setting. Note, in
                                                                comparison to usual background profiles over typical con-
                                                                tinental and high-elevation mountain areas, increasing con-
                                                                tributions of coarse-mode particles from wind-blown dusts
                                                                (1,2,3), ranging to the extreme case of a dust storm (4).
                                                                Source:  Adapted from Slinn (1976).

-------
magnitudes of the  Aitken,  acculation,  and coarse mode  integrals  will  vary in an uncorrelated
manner, as  will  the  degree  of goodness  of  the chemical separation of  these  mode integrals.
     Samplers have been  devised to collect the particles  into  two size fractions (coarse and
fine) with a separation diameter in the range of 1-5 pm.  The coarse fraction is the mass with
aerodynamic  diameter  between the  separation  and the inlet  cutoff diameters.   Therefore,  the
coarse fraction  is the  portion of the  coarse  mode that lies  between  these  aerodynamic dia-
meters.   The fine fraction  is  the mass  with aerodynamic diameter less than  the separation
point; it usually  corresponds  approximately  to the  sum of  the  accumulation and Aitken modes.
A variety of separation  diameters in the range  of  1-5  urn and  inlet cutoff diameters greater
than 10 urn have been used.
     Multimodal distributions  generally are  observed for urban  aerosols but  may not be de-
tected in other cases, e.g., marine environments or areas dominated by a strong source.
     In summary:
     1.   The particle volume  (or mass) frequency function (AV/A log D versus log D) is often
          multimodal.   The fine-volume  fraction may have two or  more  modes at ~0.02 and ~0.2
          urn.  The coarse fraction generally has one mode within the range ~5 - 50 Mm-
     2.   The types of sources  that contribute particles to the  fine  and to the coarse frac-
          tions are relatively well known.
     3.   The  particle  volume  frequency  functions  for  the fine and for the  coarse fractions
          often behave independently.
2.4.5.2   Composition  of  Particles—Upon  elucidating the multimodal behavior  of particle dis-
tributions  through the  use of the forms AN/A log  0,  AS/A log  D,  and AV/A log D,  it  was
recognized that the chemical composition of urban particles in the coarse fraction is differ-
ent  from  that in  the  fine fraction (with a separation diameter of 1-3 pin).  Evidence is cited
                                                                              2-    +   +
in Table 2-22, which shows that tropospheric secondary particles containing SO, , NH«, H , and
organics  are in the  fine  fraction, while  primary  particles consisting principally of basic
minerals  are in the  coarse  fraction.    The  composition of the fine and coarse fractions are
shown  in  Figure  2-14, which is an  idealization of the bimodal  mass distribution.  As pointed
out  in Section 2.4.5.1, the accumulation and coarse mode integrals are not sharply divided.
Thus,  the  separation  point for the fine and coarse modes does  not neatly divide*the particles
by chemical  composition.
     Investigations of  the  chemical  composition of  the fine  and  coarse particles for urban
aerosols  indicate  that  chemical species may be distributed mainly  in the fine or coarse frac-
tion,  or  both,  as is  shown  in Table 2-22.  The major components of the  fine fraction of urban
                 o-    +     -
particles are S0| , NH», N03,  Pb compounds, elemental  C (soot), and condensed organic matter.
In  Sections  5.5  and 5.6, the composition  of the fine  and coarse  fractions and their acidity
characteristics are discussed in more detail.
2.4.5.2,1   Elemental  carbon  (soot) and organics.   The carbon in fine particles consists of an
elemental  component (such as  graphite or soot)  and an organic  component of low  volatility.
                                             2-75

-------

-------
             TABLE 2-22.   CLASSIFICATION OF MAJOR CHEMICAL SPECIES ASSOCIATED WITH
                                   ATMOSPHERIC PARTICLES
                                                      Both Fine
     Fine Fraction        Coarse Fraction        and Coarse Fractions    Variable

     S0^~, C (soot),     Fe, Ca, Ti,  Mg,               NQ~, Cl~          In, Cu, Ni,  Mn
     organic (con-       K, P0^~, Si, Al,                                 Sn, Cd, V,  Sb
     densed vapors),     organic (pollen,
     Pb, NH*, As,        spores, plant parts),
          -j.^"
     Se, H , acids       bases
     Sources:   Lee and Goranson (1976); Patterson and Wagman (1977); Durham et al.  (1975);
               Rahn et al.  (1971); Akselsson et al.  (1975); Hardy et al.  (1976); Gladney
               et al.  (1974); Lundgren and Paul us (1975); Lee et al. (1968); Lee et al.
               (1972).

There are  significant differences  in  the  optical  properties of elemental  and  organic  carbon
components.  Elemental carbon  is  formed during the combustion  of  fossil  fuels  and is emitted
as primary particles  (~0.1  urn),  which strongly absorb  light.   The organic component consists
of primary hydrocarbons emitted  in combustion  exhaust and of  secondary organics  formed  by
photochemical   reactions.    These  primary  hydrocarbons  and secondary  organic  vapors  either
nucleate or  condense  on existing aerosols.   They  do not strongly absorb  light,  but do con-
tribute to light scattering in urban hazes.
     There are only limited data on the mass ratio of elemental/(primary + secondary organics)
for  a few cities.  Appel  et  al.  (1978,  1979)  found  that  for a, 4-day  period in  July 1975
elemental  carbon was  the  most abundant  carbon  species  in  Pasadena,  Pomona,  and Riverside.
Also, the  concentration  of  secondary organic  carbon  was  usually twice that of primary  hydro-
carbons.   Of  the secondary  organics,  hexanedioic  and pentanedioic acids were  among the most
abundant products;  most likely  they were oxidation  products  of cyclohexene  and cyclopentene
emitted  by motor  vehicles.   The composition of  the  organic  component retained  on filters
varied  with  the length of the sampling  period.   The retention  of  less  polar organics  (e.g.,
hydrocarbons)  was  favored by  longer sampling time,  apparently  because of  adsorption of such
organics on previously  collected material.  From total  carbon,  benzene  soluble organics, and
hydrogen analyses  of  fine  particles collected  in  Denver in November 1971,  it was estimated
                                            3
that  the elemental  carbon was 2.3-3.6  |jg/m  for the episode days  observed;  using Pb concen-
tration  as a  tracer,  it was suggested that in November 1973 in  Denver the elemental carbon in
                                  •>
fine  particles  was  1.7 - 4.4 jjg/m   (Durham et al. , 1979).  Also, for Denver in November 1973,
                                             2-77

-------
Pierson and Russell (1979) estimated from high-volume samples the total elemental carbon to be
2.9 - 27.6 ug/m3.
     Although  atmospheric measurements of carbon-containing particles  are less complete than
those of  sulfates, available  results suggest that  carbon-containing  particles in many loca-
tions,  both urban  and  nonurban,  are  the second  most abundant fine-particle species after
sulfates.    At  some western  urban  locations where SO   emissions  have  been small, carbon-con-
                                                     )\
taining aerosols  have made the largest contribution to fine-particle mass.  The concentration
of  primary carbonaceous  particles  is likely  to have been  even higher  in  the past  in the
Eastern United States when  coal  was  more  widely  used as a  fuel.   With the  growing use of
industrial coal and wood  combustors for home heating, carbonaceous particle concentrations are
likely to increase.
     Grosjean  (1977)  has  extensively  reviewed the  methods of primary and secondary organic
particle  identification,  and the  physical  and chemical aspects  of  their formation.  Primary
organics  emitted into  the  atmosphere by  industrial   sources,  motor  vehicles,  agricultural
activities, and natural sources include:  linear  and branched alkanes and alkenes, substituted
benzenes  and  styrenes, quinones,  acridines,  quinolines,  phenols,  cresols,  phthalates, fatty
acids,  carbonyl   compounds,  polyaromatic  hydrocarbons, terpenes, and  pesticides.   Secondary
organic particles are formed by tne  oxidation reactions of the  primary  organics,  ozone, and
nitrogen oxides.  Typical products that have been identified are:  aliphatic organic nitrates,
dicarboxylic acids, benzoic  and phenylacetic acids, and terpene products such  as pinonic acid
(Grosjean  and  Friedlander,  1975;  Miller  et al.,  1972;  Schuetzle  et a!., 1975).   By using
computer-controlled high-resolution  mass spectrometry  and thermal  analysis  Schuetzle et al.
(1975) and  Cronn  et al.   (1977) obtained diurnal  variations of primary and secondary organics
from 2-hour size-resolved samples.
     In an attempt  to  understand  the  atmospheric oxidation  pathways that yield secondary
organic particles,  simple mixtures  have been investigated  in  laboratory  chamber studies.  As
discussed  in  more  detail by  Grosjean  (1977),   the  following trends  have been  observed by
chamber researchers:   (a)  most paraffins  do not  generate aerosols during  irradiation,  (b)
acetylenes do  not form aerosols,  (c)  all other  unsaturated compounds with six or more carbon
atoms can form organic aerosols, (d) cyclic olefins and diolefins form more aerosol than their
1-alkene analogs, (e) conflicting results have been reported on the aerosol-forming ability of
aromatics,  (f)  carbonyl   compounds  do  not  generate  aerosol,   and  (g)   mechanical  stirring
inhibits particle formation.   Cyclic  olefins are the most efficient class of organic particle
precursors, due mainly to their high  gas-phase  reactivity  and their ability to form nonvola-
tile dicarboxylic acids.
     The  chemical composition  of  organic particles  generated in smog chambers is  not well
established for  suspected important  aerosol  precursors.    Functional  group analyses  for the
products of olefins, benzene and benzene-substituted compounds, and terpenes that have reacted
with ozone  and nitrogen  oxides show  that  the bulk consists  of  highly oxygenated compounds,
                                             2-78

-------
which Include carbonyls,  carboxylic acids, and nitrate esters.  Only a few studies of species
Identification have  been  reported.   Detailed aerosol product identification has been reported
for the  ozone-1-butene  reaction (Lipeles et al., 1973); the NO -toluene, NO -cyclohexene, and
NO -crpinene photoreactions  (Schwartz, 1974), and  the  NO -cyclopentene,  NO -cyclohexene, and
1-7-octadlene  photoreactions  (Grosjean,   1977),   Good  agreement was  indicated  by  Grosjean
(1977) with  Schwartz (1974)  for the NO -cyclohexene photoreaction,  except  that Grosjean ob-
                                        J\
served hexanedioic acid to be the major product (not reported by Schwartz).  It is significant
that most  of  the polyfunctional compounds identified (see the cited papers and the HAS report
for details) have also been Identified as important constituents in ambient aerosols.
     The secondary particles formed form alkenes having seven or more carbons (cyclic olefins,
diolefins, and terpenes)  grow into the light-scattering range and produce appreciable visibi-
lity reduction.   For example, particles formed from cyclic olefins and diolefins have particle
sizes between 0.1  and 0.3 (jm.   For  such  systems,  the gas-to-particle conversion process con-
sists of the  formation  of the supersaturation of the gas phase and subsequent condensation on
preexisting particles.
     The  rates of  conversion of precursor organic  vapors to organic particles in Los Angeles
have been  estimated  to  average  1 to  2 percent per hour.  This moderate rate of conversion Is
consistent with  the  observation (see Grosjean,  1977) that  organics  account for an important
fraction  of the  fine particles  under conditions of intense  photochemical activity, while only
a small  part of the precursor organic vapors are converted to particulate matter.
2.4.5,2.2  Nitrates.  Nitrogen oxide gases are oxidized in the atmosphere to yield HNO,, which
accumulates as nitrate  in both  fine  and  coarse  particles.   Because the topics related to the
transport  of  nitrogen oxides and their transformation to gaseous and particulate nitrates are
discussed  in  the document Air Quality  Criteria  for Oxides of Nitrogen (U.S. EPA, 1982), they
will  not  be  repeated  here.   (These topics  include  visibility, environmental  transport and
transformation,  and  acidic  precipitation.)   Atmospheric  nitrates  most  likely  result  from
photochemical  reactions  involving  the  oxidation  of  NO and  N0?  to yield HNO.,  and organic
nitrates  (Demerjian  et  al.,  1974).   The  measurement of  ambient nitrate particles  has  been
recognized to be subject to significant  sampling  errors, which are discussed in Chapter 3.
     In  summary:
     1.   The  composition  of the  coarse fraction  of  continental tropospheric  particles is
          dominated by primary minerals.
     2.   The  composition  of  the  fine   fraction  of  continental  tropospheric  particles  is
                                                                        2-     -    +    +
          dominated  by secondary  particles  that consist mainly of  SO,  ,  NO,,  NH»,  H , and
          organics, plus  primary elemental soot.
     3.   The fine fraction is often acidic, and the coarse  fraction  is often basic.
     4.   The  chemical  pathways  for forming organics and NO- particles  are not fully under-
          stood.
                                             2-79

-------
2.4.6  Particle-Size Spectrum Evolution
     The evolution  of the  atmospheric particle-size distribution  spectra can  be  related in
principle to pollution  sources  through the "general dynamic equation" (GDE).   The form of the
GDE is complex, requiring the application of sophisticated numerical techniques to obtain solu-
tions through  the  use of large computers.  Since  the  application of the GDE requires the de-
tailed  knowledge  of all  processes  significantly  influencing particle  formation,  growth,  and
removal, its  main application  to date has been- to, simu-late simple model systems.   In  a  few
cases,  simulation  results have  been  compared to  smog chamber  and atmospheric observations.
While this type  of research is active, progress has been slowed due to the lack of knowledge
of the  important pathways and  chemical  rates  for  forming sulfate,  nitrate,  and organic par-
ticles  in  the atmosphere.  Another limitation  is  imposed by the incomplete  knowledge of  the
relation between atmospheric  turbulence and particle dynamics.   In this section, the GDE will
be presented and recent studies identified in which it has been applied.
2.4.6.1  General DynamicEquation (GDE)—The evolution of the particle size/composition distri-
bution  spectra  is  given by the nonlinear, partial  integro-differential equation (Brock, 1976;
Friedlander, 1977; Gelbard and Seinfeld, 1978):

                    8n./8t + ¥-n. v  =  ^
               *       K         K~"~
                                     + I SN   -  7-cnk                               (2-67)

where

n.,   =    the number concentration of particles of type "k" at a specific point r in.space at
 K        time t, (r,t)

v    =    the fluid velocity

K    =    eddy diffusivity tensor

SnC^    rate of input at (r,t) of k-type particles from primary source P
 r

eN.  =    rate of production of particles at (r,t) by homogeneous nucleation of the i-th
  1       chemical species

c    =    particle velocity resulting from external force field (such as gravity).
                                             2-80

-------
In words, the equation has the following meaning:
     rate of change in        particle transport
     composition or In   +    with velocity v
     size distribution        (advection)
                    particle transport       rate  of change (growth)
               =    due to dispersion    +   due to condensation
                    (convection)             and chemical reactions
                                             in particles
                    rate of change           rate  of change
               +    due to coagulation   +   due to input from
                                             primary sources
                    rate of change due       rate  of change from
               +    to homogeneous       +   external force field
                    nucleation               (sedimentation).

2.4,6.2  Application of the GDE—Through the use of the GDE, it is possible to describe simul-
taneously and  quantitatively particle  interactions,  gas-to-particle  conversions,  nucleation,
and sedimentation.  The  GDE can be used as  a  module in atmospheric K-theory transport/trans-
formation models  that may include:  (a) primary  point,  line,  and area sources, and  (b)  gas-
phase photochemistry,  producing low-vapor  pressure products  that  condense on  existing  par-
ticles or nucleate new ones.  Because of computer limitations  and lack of knowledge of all  the
pathways, the  GDE  has  usually  been  applied  by  making many  simplifications.   Many of  the
studies and the features of the GDE that they retained are shown in Table 2-23.   These studies
have been classified as (a) Growth Laws, (b) Complex Simulations, (c) Comparisons with Chamber
Observations, and (d) Comparisons with Atmospheric Observations.
     The "Growth Laws" studies deal with the differences in evolution of the size distribution
due to condensation, chemical reactions on the particle surface, and chemical reactions in the
particle.  While the results of these calculations are enlightening, caution must be exercised
in using them to infer atmospheric gas-to-particle pathways.  The remaining processes that are
ignored may  not  be stationary (especially "sources") and may  significantly influence the  par-
ticle size evolution.
     The  "Theoretical  Complex  Simulations"  studies  are  those in  which  the  investigators
realized that  processes  other than "growth" of particles  are  important to the size distribu-
tion  evolution.   These  studies demonstrated  through  theoretical  calculations the  relative
roles of the processes considered in  influencing  the  evolution.   Again,  caution is suggested
in  using  the results  to  infer atmospheric pathways from simulations that  do  not  incorporate
processes known to be  important in the atmosphere.
     The "Comparisons with Chamber Observations" studies are useful for developing the growth,
coagulation, and nucleation components of the investigators' GDE model.   Table 2-23 references
only those  studies  in which the investigators predicted the particle size evolution using the
GDE.  Takahashi  (1970),  using  a non-chemically reactive system, obtained  good  agreement for

                                             2-81

-------
     Investigators
                           TABLE 2-23.   APPLICATION OF GDE TO DESCRIBE PARTICLE SIZE EVOLUTION
Condensation   Coagulation   Sources   Nucleation   External Forces   Advection  Convection
A.  GROWTH LAWS
  Brock (1972)                        V
  Seinfeld & Ramabhadran (1975)       V
B.  THEORETICAL COMPLEX SIMULATIONS
  Huang et al.  (1970)
  Burgmeier et al.  (1972)
  Wadden et al.  (1974)
  Burgmeier & Blifford (1979)
  Chu & Seinfeld (1975)
  Ramabhadran et al.  (1976)
  Middleton & Brock (1976)
  Kiang & Middleton (1977)
  Middleton & Brock (1977)
  Sheih (1977)
  Suck et al. (1977)
  Suck & Brock (1979)
  Crump & Seinfeld (1980)
  Tambour & Seinfeld (1980)
  Tsang & Brock (1982)
                               V

                               V

                               V
                                                                                                         (continued)

-------
                                                     TABLE 2-23 (continued)
DO
OJ
     Investigators            Condensation   Coagulation   Sources   Nucleation   External Forces   Advection  Convection

C.   COMPARISONS WITH CHAMBER
         OBSERVATIONS
  Takahashi (1970)                               V           V
  Heisler & Friedlander (1977)        V
  Gel bard & Seinfeld (1979)           V          V           V           V
  McMurry (1980)                      V          V           V

D.   COMPARISONS WITH ATMOSPHERIC
           OBSERVATIONS
  Husar et al.  (1972)                 V
  Heisler et al. (1973)               V
  Husar & Whitby (1973)               V
  Gartrell & Friedlander (1975)       V
  Heisler & Friedlander (1977)        V
  Suck et al.  (1978)                  '                      V                          V              V           V
  Eltgroth & Hobbs (1979)             V          V           V                          V              V           V
  McMurry et al. (1981)               V
  McMurry and Wilson (1982)           V

-------
the  evolution of  the  size distribution  in a  continuous-stirred  tank reactor.   Gelbard  and
Seinfeld  (1979)  were able  to predict  the  condensation growth  of  pre-existing particles  for
HcMurry's  (1977)  SO- photochemical  oxidation/particle evolution measurements,  but they were
unable  to predict  new  particle  formation  using the  theory  of binary  nucleation.   McMurry
(1980)  has reported  good agreement between modeling  results  and the photochemical oxidation/
particle  evolution measurements  of Clark (1972).  He  considered the condensation of monomers
on  small  clusters of  molecules  and  cluster  coagulation.   Heisler  and  Friedlander  (1977)
s'tudied the gas-to-particle conversion of organics in photochemical  smog.   They concluded that
their  chamber results  for particle  evolution  are  explained by  the gas-phase  oxidation  of
organic  gases to  form  low-vapor pressure molecules that  condense  on pre-existing particles.
Their mechanism  included the  Kelvin effect, which restricted" the  growth  to particles above a
critical  particle  size.   Heisler and Friedlander (1977) reported that this particle evolution
mechanism  yielded computed  particle size  distributions  which  agreed well  with atmospheric
measurements.  Unfortunately,  to date,  GDE modeling  comparisons with observations  for more
complex chemically reactive aerosol systems have not been reported.
     The  "Comparisons with  Atmospheric  Observations" studies referenced in Table 2-23 include
only those  for which the  investigators  predicted the particle  size  evolution  using  the GDE.
Host of the  studies  included only  particle  evolution  due to condensation  growth and are  not
reviewed  here.   Suck et al.  (1978) have used a 3-dimensional K-theory model with primary area
sources to describe  the transport and dry  deposition  of dust in Maricopa  Co.,  AZ.   They  re-
ported  good  agreement  between predicted and observed  suspended  mass concentrations.   McMurry
et al.  (1981) have used a 1-dimensional model  to infer the evolution of  the averaged cross-
wind particle size distribution.   For the data they analyzed, they reported that about 80 per-
cent of the particle volume formation could be accounted for by condensation growth.  The most
elaborate application of the  GDE has been  by  Eltgroth and Hobbs (1979) for  the evolution  of
particle  size in  coal-fired  power plant  plumes.   They have  combined a  trace-gas chemistry
scheme  for  SO ,  NO , hydrocarbons, and  oxidants (35 reactions) with  a particle scheme (GDE)
              t\    if\              *
including  condensation, coagulation,  gas-particle  reactions,  and  sedimentation.  These  two
schemes were  used in  a K-theory dispersion model  to predict the 3-dimensional concentrations
and  size  distributions.  The model  predicted  the essential features  of  the  plume reactions,
including enhanced reactivity at the outer boundaries.   They concluded that diffusion, coagula-
tion, sedimentation, and condensation  growth all are important to the particle size distribu-
tion evolution.
     The  GDE offers  an attractive pathway to develop atmospheric  aerosol  models  based  on
physico-chemical  processes.   An  active  area of  research is  to  formulate  such models on large
computers  and reduce  their  size  through  simplified  representation  of parameters to  obtain
versions  that can be  operated on smaller, more available computers.  This derivative approach
defines  the  useful range of  parameters in relation to  the  phenomenologically correct parent
model.   The alternate pathway of directly formulating reduced parameter models to fit a

                                             2-84

-------
limited number and type of field observations has been used, but those studies will  not be re-
viewed here.   It is  difficult  to define the  useful  range of the parameters  of  such models.
In Summary:
     1.   The evolution of atmospheric particle-size distribution spectrum can be described by
          the GDE.
     2.   The application of  the  GDE requires detailed  knowledge  of  all  important processes.
          Often, such information is not known.
     3.   The GDE is suitable for use as a module in K-theory type dispersion models.
     4.   Most applications of the GDE have been made with extensive simplifications; however,
          comparisons with observations  of  several  smog chamber and  atmospheric  studies have
          indicated good agreement.
     5.   The application of the GOE to relate sources and particle size distributions for the
          ultimate use in planning control strategies is an active area of research.
                                              2-85

-------
2.5  REFERENCES

Abel, E.  Theory of the oxidation of sulfite to sulfate by oxygen.  Monatsh.  Chem.  82:815-834,
     1951.                                                                          ~~

Akselsson,  R.,  J.  W.  Nelson,  and J.  W.  Winchester,  Proton  scattering for analysis of  atmo-
     spheric particulate matter.  In:  Proceedings  of the Conference on Nuclear  Cross Sections
     and Technology, NBS Special Publication 425.   Bull. Am.  Phys. Soc. 2:484, 1975.

Albu, H.  W.,  and H. D. von Schweinitz.   Autoxidations.  Vol.  V.   Formation of dithionate  by
     the  oxidation  of aqueous sulfite solutions.   Ber.  Dtsch.  Chem.  Ges. 65:729-737,  1932.

Altshuller, A.  P.   Model  predictions of  the  rates of homogeneous oxidation  of  sulfur  dioxide
     to sulfate in the troposphere.  Atmos. Environ. 13:1653-1662, 1979.

Altshuller, A.  P.,  and J.  J.  Bufalini.   Photochemical aspects  of  air pollution:   A  review.
     Environ.  Sci. Techno!. 5:39-64, 1971.

Altwicker,  E.  R.   Oxidation  of sulfite  ion  in aqueous solution.   In:  Control  of  Emissions
     from Stationary Combustion Sources:   Pollutant Detection  and Behavior in the Atmosphere.
     AICHE Symp. Ser. 75 (188):145-150, 1979.

Alyea, H. N.,  and H.  L. J.  Backstrom.   The inhibitive action  of alcohols  on the  oxidation  of
     sodium sulfite.  J. Am. Chem. Soc. 51:90-109,  1929.

Appel, B. R.,  E. M.  Hoffer, E,  L. Kothny, S. M. Wall, M. Haik,  and R.  L.  Knights.   Diurnal and
     spatial variations of  organic  aerosol constituents in  the Los Angeles  Basin.   In:   Pro-
     ceedings:  Carbonaceous Particles in the Atmosphere, National Science  Foundation and Others
     Berkeley,  California,   March 20-22,   1978.    T.  Novakov,  ed.,  LBL-9037,  University  of
     California, Lawrence Berkeley Laboratory, Berkeley, CA,  June 1979.   pp.  84-92.

Appel, B, R. ,  E.  M.  Hoffer, E., L. Kothny, S. M. Wall, M. Haik,  and R.  L. Knight.   Analysis  of
     carbonaceous material  in  southern  California atmospheric  aerosols.  II.   Environ.  Sci.
     Techno!.  13:98-104, 1979.

Azarnioueh, M. K., A. J. Bobkowicz, N. E. Cooke, and E. J. Farkas.  Growth  of a  sulphuric acid
     droplet exposed to water  vapor.   Can. J. Chem. Eng. 51:590-5, 1973.

BackstrSm,  H.  L.  J.  The chain mechanism in the  auto-oxidation of  sodium sulfite  solutions.
     Z.  Phys.  Chem.  625:99-121,  1934.

Barrie,  L.  A.,  and  H.  W.  Georgii.  An experimental investigation of the  absorption of  sulphur
     dioxide  by water drops  containing heavy metal  ions.   Atmos.  Environ. 10:743-749,  1976.

Barren,  C.  H.,  and  H.  A.  O'Hern.  Reaction kinetics of sodium  sulfite  oxidation by the rapid-
     mining method.   Chem.  Eng.  Sci.  21:397-404, 1966.

Bassett,  H. ,   and   W.   G.   Parker.    The  oxidation  of  sulphurous   acid.   J.   Chem.   Soc.
     1951:1540-1560, 1951.

Belike,  S., and G.  Gravenhorst.  Heterogeneous sulfur dioxide  oxidation  in the  droplet phase.
     Atmos. Environ. 12:231-9,  1978.

Beilke, S., D. Lamb, and J.  Milller.   On the uncatalyzed oxidation of atmospheric SO,  by oxygen
     in aqueous systems.  Atmos. Environ. 9:1083-1090, 1975.
                                             2-86

-------
Belyaev, S. P.  and  L.  M. Levin.  Techniques for collection of representative aerosol samples.
     Atmos. Environ. 5:325-38, 1974.

Bigelow,   S.   L.    Katalytische   Wirkungen  auf   die   Geschwindigkeit   der   Oxydation  des
     Natriumsulfits durch den Sauerstoff der Luft.  [Catalytic action on the rate of oxidation
     of sodium sulfite by the oxygen in the air.]  Z. Phys. Chem. 26:493-532, 1898.

Billings,  C.  E.,  and  R.  A. Gussman.  Dynamic behavior of aerosols.  In: Handbook on Aerosols.
     R. Dennis, ed. ,  Report No. ERDA TID-26608,  U.S.  Energy Research and Development Admini-
     stration, Washington, DC, 1976.  pp. 40-65.

Bird,  R. B.,  W.  E.  Steward, and  E.  N.  Lightfoot.  Transport Phenomena.  John Wiley and Sons,
     Inc.,  New York, NY, 1960.

Boisdron,  Y.,  and J.  R.  Brock.   On the stochastic  nature of  the acquisition of electrical
     charge and radioactivity by aerosol particles.  Atmos. Environ. 4:35-50, 1970.

Bracewell,  J.  M,, and D.  Gall.  The  catalytic oxidation of sulphur  dioxide in solution at con-
     centrations occurring  in  fog droplets.  In:  Air Pollution, Proceedings of the Symposium
     on  the  Physico-chemical Transformation  of Sulphur  Compounds  in  the Atmosphere and the
     Formation  of  Acid  Smogs,   Mainz,  Germany,  June  1967.    Organization  for  Economic
     Co-operation and Development, 1967.  pp. 17-26.

Bradley, R. S.  The rate of evaporation of microdrops in the presence of insoluble monolayers.
     J. Colloid Sci. 10:571-575, 1955.

Bretsznajder, S.  Prediction of Transport and  Other  Physical  Properties of Fluids.  Pergamon
     Press, New York, NY, 1971.

Bricard, J.  Physique des aerosols.  2eme Partie,  Nucleation, Condensation, Ions,
     Electrisation,  Proprietes   Optiques.    [Physics  of  aerosols.   Part  2.   Nucleation,
     Condensation,   Ions,   Electrification,  Optical   Properties.]   Rapport  CEA-R-4832(2),
     Institut de  Protection et de Surete Nucleair Departement de Protection, Centre d'  Etudes
     Nucleaires de Fontenay-aux-Roses, Gif-sur-Yvette, France, 1977.

Brimblecombe, P., and  D. J. Spedding.   The  catalytic oxidation of micromolar aqueous sulphur
     dioxide.  I.  Atmos. Environ. 8:937-945, 1974a.

Brimblecombe, P., and D. J. Spedding.  The reaction order of the metal  ion catalyzed oxidation
     of sulphur dioxide  in  aqueous solution.  Chemosphere 3:29-32,  1974b.

Brock,  J.  R.   Condensational growth of  atmospheric  aerosols.   J.   Colloid Interface Sci., 30:
     32-36, 1972.

Burgmeier,  J.  W.  and  I. H.  Blifford,  Jr.   A  reinforced coagulation-sedimentation model for
     stratospheric aerosols.  Wat. Air Soil Pollut. ,  5_:133-147, 1975.

Burgmeier, J. W., I. H.  Blifford,  Jr., and 0. A. Gillette.  A reinforced coagulation-sedimenta-
     tion  aerosol model.  Wat.  Air Soil  Pollut., 2:97-104,  1972.

Burrows, J.  P.,  D.  I. Cliff, 6. W.  Harris, B.  A. Thrush, and J. P. T.  Wilkenson.  Atmospheric
     reactions  of  the H02  radical  studies by laser magnetic resonance  spectroscopy.  Proc.  R.
     Soc.  London.  A368:463-481.  1979.

Cadle,  R.  D.  The Measurement of  Airborne Particles.  John Wiley and Sons, Inc., New York, NY,
     1975.

Cadle,  R.   D.,  and  R.  C.  Robbins.  Kinetics of  atmospheric  chemical  reactions  involving
     aerosols.  Discuss.  Faraday  Soc. 3£:155-61,  1960.


                                             2-87

-------
Calvert,  J,  G., and  R.  D. McQuigg.   The computer simulation of  the  rates and mechanisms of
     photochemical  smog  formation.   In:   Proceedings of  the Symposium  on Chemical  Kinetics
     Data  for  the  Upper  and  Lower  Atmosphere,  Stanford  Research  Institute,   Warrenton,
     Virginia, September 15-18, 1974.  Int. J. Chem.  Kinet, Symp.  (1):113-154,  1975.

Calvert,  J.  G., F.  Su,  J. W.  Bottenheim, and  0. P. Strausz.   Mechanism of  the  homogeneous
     oxidation  of  sulfur  dioxide   in   the  troposphere.   In;   Sulfur  in  the   Atmosphere,
     Proceedings  of the  International  Symposium, United  Nations  Environmental  Program and
     Others,  Dubrovnik,  Yugoslavia,  September 7-14,  1977.   Atmos. Environ. 12:197-226,  L978.

Campbell,  W.   J.,  J.  C.  Sheppard, and  B.  f. Au.   Measurement of  hydroxyl  concentration in
     boundary  layer  air  by  monitoring carbon  monoxide  oxidation.   Geophys.   Res.   Lett.
     6:175-178, 1979.

Carabine,  M.  D. and  J.  E. L,  Haddock.   The  growth  of  sulphuric  acid aerosol particles when
     contacted with water vapor.   Atmos.  Environ.  10:735-42, 1976.

Carberry,  J.   J.   Chemical and  Catalytic Reaction  Engineering.-  McGraw-Hill,  New York, NY,
     1976.  pp. 194-305.

Carter, W.  P.  L.,  A.  C.  Lloyd,  J. L. Sprung, and J. N. Pitts, Jr.  Computer  modeling of smog
     chamber  data:   progress  in validation of a  detailed mechanism for  the photooxidaticn of
     propene and n-butane in photochemical smog.   Int. J. Chem. Kinet. 11:45-111, 1979.

Chambers,  L.  A.  Classification  and extent  of  air  pollution  problems.   In:   Air  Pollution.
     Volume I.  A. C. Stern, ed.,  Academic Press,  New York, NY,' 1976.  pp.  1-21.

Chameides, W.  L.,  and J.  C. Walker.  Photochemical theory of tropospheric  ozone.   J.  Geophys.
     Res. 78:8751-8760, 1973.

Chameides, W.  L.,  and J.  C. Walker.   A  time-dependent photochemical  model for ozone  near the
     ground.  J. Geophys. Res. 81:413-420, 1976.

Chang,  D.  P.   Y.,  and R.  E.  Hill.   Retardation  of  aqueous  droplet  evaporation by air pol-
     lutants.   Atmos. Environ. 14:803-807, 1980.

Chang,  S.-G.,  R.  Brodzinsky,  R. Toossi,  S. S. Markowitz, and J. Novakov.   Catalytic oxidation
     of S02 on carbon in aqueous  solutions.  In:  Proceedings:  Carbonaceous  Particles  in the
     Atmosphere, National  Science Foundation and Others,  Berkeley, California, March  20-22,
     1978.  T. Novakov, ed., LBL-9037, University  of  California, Lawrence Berkeley  Laboratory,
     Berkeley, CA, June 1979.   pp. 122-130.

Charlson,  R.   J.,   A.  H,  Vanderpol,  D.  S.   Covert, A.  P.  Waggoner,   and  N.  C. Ahlquist.
     H2S04/(NH4)2S04 background  aerosol:   optical  detection in the  St.   Louis  region.    Atmos.
     Environ.  8:1257-1267, 1974.

Charlson, R. J., D. S. Covert, T.  V.  Larson,  and A. P. Waggoner.   Chemical  properties  of trop-
     ospheric  sulfur aerosols.   Jji:   Sulfur in  the Atmosphere,  Proceedings  of  the  Inter-
     national  Symposium, United Nations Environment Program and Others, Dubrovnik,  Yugoslavia,
     September 7-14, 1977.  Atmos. Environ. 12:39-53, 1978.

Cheng,  R.  T.,  M.  Corn, and J.  0.  Frohliger.   Contribution to  the reaction kinetics  of water
     soluble aerosols and S0? in air  at ppm concentrations.  Atmos.  Environ. 5:987-1008, 1971.

Chu, K.  S.  and R.  Morrison.  The  vapor  pressure  and latent  heat  of aqueous sulfuric acid at
     phase  equilibrium.    In:    Environmental   and  Climatic  Impact of  Coal  Utilization,
     [Proceedings of the Symposium], Williamsburg, Va., April 17-19, 1979.  J.  J. Singh  and A.
     Deepak, eds., Academic Press, NY, 1980.  pp.  293-307.
                                             2-88

-------
Chun, K. C. ,  and J.  E. Quon.  Capacity of ferric oxide particles to  oxidize  sulfur dioxide  in
     air.   Environ. Sci. Technol. 7:532-538, 1973.

Clark, W.   E.   Measurements of aerosols produced by the photochemical oxidation of S02  in  air.
     Ph.D.   thesis.   Mechanical Engineering Department,  University  of Minnesota, Minneapolis,
     MM, 1972.

Clark,  W.   E. ,  and  K.  T.  Whitby.   Concentration  and  size  distribution measurement  of
     atmospheric aerosols  and a test of the theory of self-preserving size distributions.   J.
     Atmos. Sci. 24:677-687, 1967.

Corn, M.   Nonviable  particles in the air.   In:   Air Pollution.  Volume  I:   Air  Pollution and
     Its Effects.  A. C. Stern, ed., Academic Press, New York,  NY, 1976.   pp. 47-94.

Cotton, F.  A., and G. W. Wilkinson.  Advanced Inorganic Chemistry:  A Comprehensive Text.  4th
     Edition.  John Wiley and Sons,  Inc., New York, NY, 1980.

Coughanowr, D. R. , and F. E. Krause.  The reaction of S09 and 05  in aqueous solution  of MnSOA.
     Ind.  Eng. Fund.  4:61-66, 1965.                     ^       ^                             ^

Cox, R. A. ,  and S. A. Penkett.   Oxidation  of atmospheric S0?  by-products of the ozone-olefin
     reaction.  Nature (London) 230:321-322, 1971a.

Cox,  R.  A.,  and  S.  A.  Penkett.   Photooxidation of  atmospheric SO,.    Nature   (London)  229:
     486-488, 1971b.

Cox, R. A., and S. A. Penkett.  Aerosol formation from sulfur dioxide in  the  presence of ozone
     and olefinic hydrocarbons.  J.  Chem. Soc. Faraday Trans. I 68:1735-1753, 1972.

Cronn D. R. ,  R.  J. Charlson,  R.  L.  Knights,  A.  L. Crittenden,  and  B.  R. Appel.  A  survey  of
     the molecular nature  of  primary and  secondary  components  of particles in urban  air  by
     high-resolution mass spectrometry.  Atmos. Environ. 11:929-937,  1977.

Crump,  J.  G.  and  J.  H. Seinfeld.   Aerosol  behavior in the  continuous  stirred  tank reactor.
     AIChE J., 26:610-616,  1980.

Crutzen, P. J.  Photochemical reactions initiated by and influencing ozone in  unpolluted tropo-
     spheric air.  Tell us 26:47-57,  1974.
                                                                         *3 *3 r\
Danilczuk,  E,,  and  A.  Swinarski.   The  complex ion  [Fe  (111)  (SO™) ]     .  Roczniki Chem.
     35:1563-1572, 1961.                                            d n

Dasgupta,  P.  K.   The  importance of  atmospheric ozone  and  hydrogen  peroxide  in  oxidizing
     sulphur  dioxide  in  cloud and rainwater—further discussion.  Atmos.  Environ. 14:620-621,
     1980.

Dasgupta,  P.  K. ,  P.  A.  Mitchell,  and  P.  W. West.   Study of transition metal  ion  - S(IV)
     systems.  Atmos. Environ. 13:775-782, 1979.

Davis,  D.  D. ,  W. Heaps, and  T.  McGee.   Direct measurements of natural tropospheric  levels  of
     hydroxyl  via  an aircraft-borne tunable dye  laser.   Geophys. Res.  Lett. 3:331-333, 1976.

Demerjian  K. ,  J.  A.  Kerr, and  J.   Calvert.  Mechanism  of  photochemical  smog formation.  In;
     Advances  in Environmental Science and Technology.   Volume  4.    J.   N,   Pitts  and R.  J.
     Metcalf,  eds., John Wiley and Sons, Inc., New York, NY, 1974.  pp. 1-262.

Dennis, R., ed.   Handbook  on Aerosols.   Report No. ERDA TID-26608,  U.S. Energy Research and
     Development Administration, Washington, DC, 1976, 142 pp.
                                              2-89

-------
Doyle, G.  J.   Response to  the  comment on self-nucleation  In  the  sulfuric acid-water system.
     J. Chem.  Phys. 75:1585-1586, 1981.

Durham, J.  L.,  R.  K.  Patterson, and R. G. Draftz,  Carbon  in Denver's urban plume.  In:  Pro-
     ceedings:   Carbonaceous Particles  in  the  Atmosphere,  National  Science  Foundation and
     Others, Berkeley,  California,  March 20-22,  1978.  T.  Novakov,  ed. ,  LBL-9037, University
     of  California,  Lawrence  Berkeley  Laboratory,  Berkeley,  CA, June  1979.    pp.  102-106.

Durham, J.  L.,  W.  E.  Wilson,  T.  G.   Ellestad,  K.  Willeke, and K. T.  Whitby.   Comparison of
     volume and mass  distributions  for Denver  aerosols.   Atmos. Environ.  9:717-722, 1975.

Eatough, D. J.,  W.  P.  Green, and  L.  D.  Hansen.   Oxidation  of  sulfite  by activated charcoal.
     In:   Proceedings:  Carbonaceous  Particles in the Atmosphere, National Science Foundation
     and Others, Berkeley California, March 20-22, 1978.  T. Novakov, ed., LBL-9037, University
     of  California,  Lawrence  Berkeley  Labojatory,  Berkeley,  CA, June  1979.    pp.  131-132.

Eatough, D. J.,  T. Major, J. Ryder, M. Hill, N. F. Mangelson, N. L. Eatough, and  L. D. Hanson,
     R. G. Meisenheimer, and J. W. Fischer.  The formation  and  stability  of sulfite species in
     aerosols.   In:   Sulfur in  the Atmosphere,  Proceedings of the  International Symposium,
     United Nations  Environmental  Program and Others,  Dubrovnik,  Yugoslavia,  September 7-14,
     1977.  Atmos. Environ. 12:263-272, 1978.

Eatough, D. J.,  L. D. Hansen, P. K. Dasgupta, P.  A. Mitchell, and  P. W. West.  Discussions of:
     "Study of  transition  metal  ion -  S(IV)  systems" by  Dasgupta  et  a!.,  Atmos.  Environ.
     13:775-782.  Atmos. Environ. 13:1725-1729, 1979.

Eigen, M.,  K.  Kustin,  and  G.  Maass.   Die Geschwindigkeit  der  Hydration  von  S05 in Wassriger
     LSsung,  Z. Phys.  Chem. 30:130-136,  1961.                                  ^

Eisner, H.  S.,  B.  W.  Quince, and  C.  H.  Clark.  The stabilization of water mists by insoluble
     monolayers.  Discuss.  Faraday Soc. 30:86-95, 1960.

Eltgroth, M.  W. and  P.  V. Hobbs.  Evolution  of particles  in  the plumes of coal-fired power
     plants—II.  A numerical model and comparisons with field  measurements.  Atmos. Environ.,
     13:953-975, 1979.

Erickson, R.  E.,  L.  M. Yates,  R.  L.  Clark,  and C. M. McEwen.  The reaction of sulfur dioxide
     with  ozone  in  water  and  its  possible  atmospheric  significance.   Atmos.  Environ.
     11:813-817, 1977.

Eriksson,  E.    The yearly circulation   of  chloride  and  sulfur  in  nature;  meteorological,
     geochemical, and pedological implications.  Tell us 11:375403, 1960.

Eriksson,  E.   The yearly  circulation of  sulfur in  nature.   J.   Geophys. Res.   68:4001-4008,
     1963,

Fairbridge, R.  W.,  ed.   The Encyclopedia  of  Atmospheric  Sciences and Astrogeology.  Encyclo-
     pedia of Earth  Sciences Series,  Volume  II.  'Van Nostrand Reinhold  Publishing Corp., New
     York, NY, 1967.

Fishman, J., and P. J.  Crutzen.  A numerical study of tropospheric photochemistry using a one-
     dimensional model.  J. Geophys.  Res. 82:5897-906, 1977.

Fishman, J.,  and P.  J. Crutzen.  The distribution of the hydroxyl radical in the troposphere.
     Atmospheric  Science  Paper  No.  284, Colorado  State  University, Fort Collins, CO, 1978.
                                             2-90

-------
Fitzgerald, J. W.   Approximate formulas for the  equilibrium  size of an aerosol particle as a
     function of  its  dry  size and composition  and  the ambient relative humidity,  J. Applied
     Meteorol, 14:1044-9,  1975.

Freiberg, J.  Effects  of  relative humidity and temperature on iron-catalyzed oxidation of SOp
     in atmospheric aerosols.  Environ. Sci. Techno!. 8:731-734, 1974.

Freiberg, J.  E.,  and S.  E.  Schwartz.   Oxidation of SOp  in  aqueous droplets:  mass-transport
     limitation   in  laboratory  studies   and  the   ambient  atmosphere.    Atmos.  Environ.
     15:1145-1154, 1981.

Friedlander,  S. K.   Smoke,  Dust and Haze.   Fundamentals  of  Aerosol Behavior,  John Wiley and
     Sons, Inc., New York, NY, 1977.

Friend,  J.  P.   The  global  sulfur  cycle.   In:   Chemistry  of  the Lower  Atmosphere.   S.  I.
     Rasool, ed., Plenum Press, New York, NY, 1973.  pp. 177-201.

Fuchs, N. A.  The Mechanics of Aerosols.  Pergamon Press, New York, NY, 1964.

Fuller,  E.  C.,  and  R.  H.  Crist.   The rate of oxidation of sulfite  ions  by oxygen.   J. Am.
     Chem. Soc.  63:1644-1650, 1941.

Fuzzi, S.   Study  of iron  (111) catalyzed sulphur dioxide oxidation in aqueous solution over a
     wide range of pH.  Atmos. Environ. 12:1439-1442, 1978.

Gartrell, G., Jr.  and S. K.  Friedlander.  Relating particulate pollution to sources:  The 1972
     California aerosol characterization study.  Atmos. Environ., 9:279-299, 1975.

Gelbard, F. and J. H. Seinfeld.  The general dynamic equation for aerosols = Theory and appli-
     cation to  aerosol formation and growth.    J.  Colloid Interface  Sci., 68:363-382, 1979.

Gentry,  J.  W. and  J.  R.  Brock.   A study  of  the lifetime of  aerosol  particles.   J. Colloid
     Interface Sci. 27:691-701, 1968.

Gillespie,  R. J.   Sulphuric acid as a solvent system.   In:   Inorganic Sulphur Chemistry.  G.
     Nichless, ed., Elsevier Publishing Co., New York, NY, 1968, pp. 563-586.

Gladney,  E. S. ,  W.  H. Zoller, A. G.  Jones, and G. E. Gordon.  Composition and size distribu-
     tions  of atmospheric  particulate matter  in  the Boston  area.   Environ.  Sci.  Techno!.
     8:551-557, 1974.

Glasstone,  L.   Thermodynamics for  Chemists.   Van Nostrand,  Princeton,  NJ, 1947.   pp. 72-73.

Graham,  R.  A.,  A. M. Winer, R. Atkinson, J. N.  Pitts, Jr.  Rate constants  for the  reaction of
     H09 with HO,,  SO,, SO  N,0, trans-2-Butene, and 2,3,Dimethyl-2-Butene  at 300 K.  J.  Phys.
     ChSm. 83:1563-1567,  1979T

Granat,  L., H. Rodhe, and R. 0. Hallberg.   The  global sulfur  cycle.  In:  Nitrogen, Phosphorus
     and  Sulfur-Global Cycles.  SCOPE  Report 7.   Ecol. Bull.  (22):89-134, 1976.

Grosjean,  D.   Aerosols.  In:   Ozone  and  Other Photochemical  Oxidants.   National  Academy of
     Sciences, Washington,  DC, 1977.   pp. 45-125.

Grosjean,  D. ,  and S.  K.  Friedlander.  Gas-particle distribution  factors for  organic and  other
     pollutants  in the Los Angeles atmosphere.  J.  Air  Pollut.  Control Assoc. 25:1038:1044,
     1975.
                                             2-91

-------
Haber, F. , and 0. H. Wansbrough-Jones.  Autoxidation  (VI) action of  light on  sulfite  solutions
     in absence and presence of oxygen.  Z. Physik. Chem. 618:103-123,  1932.

Hales, J. M. ,  and S.  L. Sutter.   Solubility  of sulfur dioxide in low  concentrations.  Atmos.
     Environ. 7:997-1001, 1973.

Hanel, G.  and 8.  Zankl.   Aerosol  size and  relative humidity:   water uptake by mixtures of
     salts.   Tellus 31:478-486, 1979.

Hardy, K.  A., R.  Akselsson,  J.  W.  Nelson, and J. W.  Winchester.   ElementaJ constituents of
     Miami aerosol  as  a function of  particle  size.   Environ. Sci.  Techno!.  10:176-182,  1976.

Harker, A. B.,  L.  W.  Richards, and W.  E. Clark.  The effect  of atmospheric S02 photochemistry
     upon observed nitrate concentrations in aerosols.  Atmos. Environ.  11:87-91,  1977.

Hayon, E.,  A.  Treinin, and  J.  Wilf.   Electronic  spectra,  photochemistry,  _and  autoxidation
     mechanism  of the  sulfite-bisulfite-pyrosulfite  systems.   The SQ?,  SO,,  SO.,  and SO^
     radicals.  J. Am.  Chem. Soc. 94:47-57, 1972.                       L     s    <\          a

Hecht, T. A.,  J.  H. Seinfeld, and M. C. Dodge.  Further development of generalized mechanisms
     for photochemical  smog.  Environ.  Sci. Technol.  8:327-339, 1974.

Hegg, D,  A.,  and P. V. Hobbs.  Oxidation of sulfur dioxide in aqueous  systems with particular
     reference to the atmosphere.  Atmos. Environ. 12:241-253, 1978.

Heisler,   S.   L.  and S.  K.  Friedlander.    Gas-to-particle  conversion  in  photochemical  smog:
     Aerosol   growth laws  and  mechanisms   for  organics.   Atmos.   Environ.,  11:157-168,  1977.

Heisler,  S.  L.,  S.  K.  Friedlander, and R.  B.  Husar.   The  relationships of smog  aerosol  size
     and  chemical   element  distributions  to   source  characteristics.    Atmos.   Environ.,
     7:633-649, 1973.

Hesketh,  H.  E.   Understanding  and Controlling Air Pollution.  Ann  Arbor Science Publishers,
     Ann Arbor, MI, 1977.

Hidy, G.  M.,  and J. R. Brock.   An assessment of the global  sources of tropospheric  aerosols.
     In:   Proceedings  of  the  Second  International Clean  Air  Congress,   International Union of
     Air Pollution Prevention Associations, Washington, DC, December 6-11, 1970.   H.  M. Englund
     and W.  T. Beery,  eds., Academic  Press  Inc., New  York, NY, 1971.  pp. 1088-1097.

Hidy, G.  M., and  J.  R. Brock.   The Dynamics of Aerocolloidal Systems.  Pergamon Press, New
     York, NY, 1970.

Higgins,  W.  C.  E.,  and J. W.  Marshall.   Equivalence  changes in oxidation-reduction  reactions
     in solution:   some  aspects of the oxidation of  sulphurous  acid.    J. Chem. Soc. 447-458,
     1957.

Hilton, C. M.,  J.  J.  Christensen, D.  J.  Eatough,  and L. D.  Hansen.  Fe (III) - S(IV)  aerosol
     generation and characterization.  Atmos. Environ. 13:601-5, 1979.

Hirschfelder, J. 0., C.  F.  Curtiss,  and R.  B.  Bird.   Molecular theory of  gases  and liquids.
     John Wiley and Sons, Inc., New York, NY, 1954.

Hitchcock, D. R.,  L.  L.  Spiller, and W.  E. Wilson.   Sulfuric and aerosols and HC1 release in
     coastal  atmospheres:   evidence  of rapid formation of sulfuric  acid particulates.  Atmos.
     Environ. 14:165-182, 1980.

Hoather,  R.  C., and C.  F. • Goodeve.   The  oxidation  of sulphurous  acid.   III.   Catalysis of
     manganous sulphate.  Trans. Faraday Soc. 30:1149-1156, 1934.


                                             2-92

-------
Hoigne, J., and  H.  Bader.   Ozonation of water:  kinetics of oxidation of ammonia by ozone and
     hydroxyl  radicals.  Environ. Sci. Techno!. 12:79-84, 1978.

Horike, N.  R.   The  Rate of Oxidation  of Aqueous Solutions of  Sodium Sulfite:   Influence of
     Temperature, Oxygen Partial  Pressure,  and  pH.   MS Thesis,  University  of Washington,
     Seattle,  WA, 1976.

Huang,  C.  M. ,  M.   Kerker,  and  E.  Matijevic.   The  effect  of Brownian  coagulation,  gradient
     coagulation, turbulent coagulation and wall losses upon the particle size distribution of
     an aerosol,   J. Colloid Interface Sci., 33:529-538, 1970.

Huntzicker, J. J. ,  R.  A. Gary,  and  C.  -S.  Ling.   Neutralization of  sulfuric acid aerosol by
     ammonia.   Environ. Sci. Techno1. 14:819-824, 1980.

Husar, R. B. ,  K.  T. Whitby, and B.  Y.  H. Liu.  Physical mechanisms governing the dynamics of
     Los Angeles smog aerosol.   J. Colloid Interface Sci., 39:211-224, 1972.

Husar, R. B.  and K. T. Whitby.  Growth mechanisms and size spectra of photochemical aerosols.
     Environ.  Sci. Technol., 7:241-247, 1973.

Husar, R.  B.,  D.  E.  Patterson,  J. D. Husar, N. V. Gillani, and W. E. Wilson.  Sulfur budget of
     a power  plant plume.   Iri:   Sulfur  in  the Atmosphere,  Proceedings of the  International
     Symposium,  United   Nations  Environment  Program  and  Others,  Dubrovnik,  Yugoslavia,
     September 7-14, 1977.  Atraos. Environ. 12:549-568, 1978.

Huss,  A.,  Jr., and  C.  A.  Eckert.   Equilibria and  ion activities  in aqueous sulfur dioxide
     solutions.  J,  Phys. Chem. 81:2268-2270, 1977.

Huss,  A., Jr.,  P.  fa  Lim,  and C.  A. Eckert.  On the  uncatalyzed  oxidation of sulfur (IV) in
     aqueous solutions.  J.  Am.  Chem. Soc. 100:6252-6253, 1978.

Johnson,  D.  B.  Ultragiant urban aerosol  particles.   Science  (Washington,  DC) 194:941-942,
     1976.                                                                        ~"~

Johnstone,  H,  F.   Metallic ions as  catalysts  for the removal  of  sulfur dioxide from  boiler
     furnace gases.   Ind. Eng.  Chem. 23:559-561, 1931.

Johnstone, H.  F., and D. R.  Coughanowr.  Absorption of sulfur dioxide  from the air.  Ind.  Eng.
     Chem. 50:1169-1172, 1958.

Johnstone,  H.  F. ,  and  A.  J.  Moll.  Formation  of  sulfuric acid  in  fogs.   Ind.  Eng.  Chem. 52:
     861-863,  1960.

Judeikis, H. S., T.  B. Steward,  and  A. G. Wren.  Laboratory studies of heterogeneous reactions
     of S00.  Atmos. Environ. 12:1633-1642, 1978.
          £.                    "~
Junge, C.  E.   The size distribution and aging of natural aerosols  as  determined  from electric
     and optical data on the atmosphere.  Am. Meteorol. Soc. J.  12:13-25, 1955.

Junge, C., and T. G. Ryan.  Study of the S0? oxidation in solution  and its role in atmospheric
     chemistry.  Q. J. R. Meteorol.  Soc. 84T46-55, 1958.

Junge,  C.,  and G.  Scheich.  Determination of the acid  content  of  aerosol  particles.   Atmos.
     Environ.  5:165-75,  1971.

Karraker,  D.  G.   The  kinetics  of  the reaction  between  sulphurous acid  and ferric ion.  J.
     Phys. Chem. 67:871-874, 1963.
                                              2-93

-------
Kellogg, W. W.,  R.  D.  Cadle, E. R. Allen, A. L. Lazrus, and E. A. Mar-tell,  The sulfur cycle.
     Science (Washington, DC) 175:587-596, 1972.

Kiang,  C.   S.  and  P.  Middleton.   Formation  of secondary  sulfuric  acid aerosols  in urban
     atmosphere.  Geophy. Res. Lett., 4:17-20, 1977.

Kuhlman, M. R., D. L. Fox, and H. E. Jeffries.  The effect of CO on  sulfate aerosol formation.
     Atmos. Environ. 12:2415-2423, 1978.

Lamb, H.  Hydrodynamics.  Dover, NY, 1945.

Larson, T.  V.  The  Kinetics of  Sulfur Dioxide Oxidation by Oxygen and Ozone in Atmospheric
     Hydrometers.  Ph.D. Thesis, University of Washington, Seattle,  WA,  1976.

Larson, T.  V., N. R. Horike, and H. Harrison.  Oxidation of sulfur dioxide by  oxygen  and ozone
     in aqueous  solution:   a  kinetic study with significance  to atmospheric rate processes.
     Atmos. Environ. 12:1597-1611, 1978.

Lee, R. E.,  Jr.,  and S. Goranson.   National  Air Surveillance  Cascade Impactor Network.   III.
     Variations in  size of airborne particulate matter over three-year period.  Environ.  Sci.
     Techno!.  10:1022, 1976.

Lee, R. E.  ,  Jr., S. S. Goranson, R. E.  Enrione, and G. 8. Morgan.   National  Air  Surveillance
     Cascade Impactor Network.  II.  Size distribution measurements  of trace metal components.
     Environ.  Sci. Techno!.  6:1025-1030,  1972.

Lee, R. E.,  Jr.,  R. K.  Patterson,  and  J. Wagman.  Particle-size distribution of  metal compo-
     nents in  urban air.  Environ. Sci. Technol. 2:288-290, 1968.

Leighton, P. A.  Photochemistry of Air  Pollution.  Academic Press, New York, NY, 1961.

Levy,  H.,  II.  Normal  atmosphere:   large  radical  and formaldehyde  concentrations predicted.
     Science (Washington, DC) 173:141-143, 1971.

Liberti, A.,  D.  Brocco,  and M.  Possanzini.   Adsorption  and  oxidation  of  sulfur dioxide  on
     particles.   In:   Sulfur in  the Atmosphere, Proceedings  of the  International Symposium,
     United  Nations Environment  Program and  Others,  Dubrovnik, Yugoslavia,  September 7-14,
     1977.   Atmos. Environ.  12:255-261, 1978.

Lipeles M.,  C. S. Burton, H. H.  Wang,  E. P.  Parry, and  G.  M.  Hidy.   Mechanisms  of  Formation
     and  Composition  of  Photochemical  Aerosols.    EPA-650/73-036.   Rockwell  International
     Science Center, Thousand Oaks, CA, 1973.

Liu,  B.  Y. H.  and D.  Y.  H. Pui.   Aerosol  sampling  inlets  and inhalable particles.  Atmos.
     Environ.  15:589-600, 1981.

Lumiere, A.  L., and A.  Seyewetz.   Sur 1'antioxydation des solutions de sulfite de  sodium  et
     sur  les  antioxydants.   [On  the  antioxidation  of sodium sulfite  solutions  and on the
     antioxidants,]  Bull. Soc. Chim. France  33:444-454, 1905.

Lunak,  S.,  and J.  Veprek-Siska.  Photochemical  autooxidation of  sulphite catalyzed  by  iron
     (III)  ions.  Coll.  Czech.  Chem. Commun.  41:3495-3503, 1975.

Lundgren,  D.  A. ,  and J. J.  Paulus.  The  mass  distribution of  large  atmospheric particles.   J.
     Air Pollut.  Control Assoc. 25:1227-1231,  1975.

Lyons,  D.,   and  G.  Nickless.    The  lower  oxy-acids  of  sulphur.    In:   Inorganic  Sulphur
     Chemistry.   G.  Nickless, ed., Elsevier  Publishing Co., New  York, NY, 1968.   pp. 509-533.
                                              2-94

-------
Marlow, W. H.  Unipolar aerosol diffusion charging.  Part  I.  Particle dielectric  constant  and
     ion mobility distribution effects.  J. Colloid Inter.  Sci. 64:543-548, 1978a.

Marlow, W. H.   Unipolar aerosol diffusion  charging.   Part II.   Ion and aerosol polydispersi-
     ties:   the  "diffusion charging  mobility analysis"  hypothesis.   J.  Colloid  Inter. Sci.
     64:549-554, 1978b.

Martin, L. R. ,  and D. E.  Damschen.   Aqueous  oxidation of  sulfur dioxide  by  hydrogen  peroxide
     at low pH.  Atmos. Environ. 15:1615-1621, 1981.

Martin, L. R.,  D.  E. Damschen, and H. S. Judeikis.  The reactions  of nitrogen  oxides  with  SO,
     in aqueous aerosols.  Atmos. Environ.  15:191-195, 1981.

Matteson,  M.  J., W.  Stober,  and H.  Luther.   Kinetics of  the  oxidation  of sulfur dioxide  by
     aerosols of manganese sulfate.   Ind. Eng. Chem. Fundam. 8:677-687, 1969.

May, K.  R. ,  N.  P.  Pomeroy, and  S.  Hibbs.   Sampling techniques for large  windborne particles.
     J. Aerosol Sci. 7:53-62, 1976,

McCrone, W.  C. ,  and J. G. Delly.  The Particle Atlas.  2d  Edition.  Volume I:   Principles  and
     Techniques.  Ann Arbor Science Publishers, Ann Arbor,  MI, 1973.  pp.  264-266,

McFarland, A.  R.,  J.  B.  Wedding  and J.   E.  Cermak.   Wind  tunnel evaluation  of a  modified
     Anderson  Impactor and an  all-weather sampler  inlet.   Atmos. Environ.  11:535-42,  1977.

McKay,  H.  A.  C.   The atmospheric  oxidation of  sulphur  dioxide in  water  droplets in  the
     presence of ammonia.  Atmos. Environ.  5_:7-14, 1971.

McMurry, p.  H.   On the Relationship  Between  Aerosol  Dynamics and  the Rate of  Gas-to-Particle
     Conversion.   Ph.D.  Thesis,  California   Institute  of Technology.    Pasadena,  CA,  1977.

McMurry, P.  H,   Photochemical  aerosol formation from  sulfur  dioxide:   A  theoretical  analysis
     of smog chamber data.  J. Colloid Interface Sci., 78:513-527,  1980.

McMurry,  P.  H., D.  J. Rader, and  J.  L.  Stith.   Studies  of  aerosol  formation in power  plant
     plumes—I.  Growth  laws  for secondary aerosols  in  power plant plumes:   Implications  for
     chemical conversion mechanises,  Atmos.   Environ., 15:2315-2327, 1981.

McMurry, P, H. and J. C. Wilson.  Growth laws  for the  formation of  secondary  ambient aerosols:
     Implications  for  chemical  conversion mechanisms.    Atmos.  Environ.,  16:121-134,  1982.
                                                                                   W
McNeil's,  D.  N. ,  L.  A.  Ripperton,  W.  E.  Wilson,  P. L.  Hanst,  and  B.  W.  Gay.   Gas  phase
     reactions of  ozone and olefin in the  presence  of sulfur dioxide.  In:   Removal  of Trace
     Contaminants from Air, A Symposium at  the 168th Meeting, American Chemical  Society,  1974.
     ACS Symp. Ser.  (17):187-200, 1975.

Meszaros,  E.   The  size  distribution  of water soluble particles  in the atmosphere.   Idojaras
     75:308-314, 1971.

Middleton, P.  B. ,  and J.  R.  Brock.   Simulation  of  aerosol  kinetics.   J.   Colloid Interface
     Sci., 54:249-264, 1976.

Middleton, P.  B., and  J. R. Brock.  Modeling  the urban aerosol.  J. Air  Pollut.  Control  Assoc.
     27:771-775, 1977.

Milbaur, J. ,  and J. Pazourek.   Oxidation  of  sulfites in concentrated solutions.   Chem.  Listy
     15:34-38, 1921.
                                              2-95

-------
Miller,  J.  M. ,  and  R.  G. de  Pena.   Contribution of scavenged  sulfur dioxide to the  sulfate
     content of rain water,  J. Geophys. Res. 77:5905-5916, 1972.

Miller, M. S., S.  K. Friedlander, and G. M. Hidy.  A chemical element  balance  for the Pasadena
     aerosol.  J.  Colloid Interface Sci. 39:165-176, 1972.

Mishra,  G.  C.,  and  R.  D. Srivastava.   Homogeneous kinetics  of potassium sulfite oxidation.
     Chem. Eng,  Sci. 31:969-971, 1976.

Holler,  D.   Kinetic  model  of atmospheric  S0?  oxidation  based  on  published  data.   Atmos.
     Environ. 14:1067-1076, 1980.              i

Moss, M.  R.   Sources of sulfur  in  the  environment;  the global  sulfur cycle.   In:   Sulfur  in
     the  Environment.   Part I:   The Atmospheric  Cycle.   J.  0.  Nriagu, ed. ,  John Wiley and
     Sons, Inc., New York, NY, 1978.  pp. 23-50.

Nair, P.  V.  N.,  and K. G. Vohra.   Growth of aqueous sulphuric  acid droplets  as a  function  of
     relative humidity.  J. Aerosol Sci. 6:265-271, 1975.

National Research Council.  Subcommittee on Airborne Particles,  Airborne  Particles, University
     Park Press, Baltimore, MD, 1977.

National Air  Pollution  Control Administration.   Air Quality  Criteria for Particulate  Matter.
   -  NAPCA-Pub-AP-49, U. S. Government Printing Office, Washington, D.C., 1969.

National  Air Pollution  Control  Administration.   Air  Quality  Criteria for  Sulfur  Oxides.
     NAPCA-Pub-AP-50, U.S. Government Printing Office, Washington, D.C.,  1970.

Neytzell-de Wilde, F. G., and  L. Taverner.  Experiments relating to the possible production  of
     an  oxidizing  acid leach  liquor by  auto-oxidation  for  the extractions of uranium.  In:
     Proceedings of the 2d  United Nations  International  Conference  on  the Peaceful  Uses  of
     Atomic  Energy, Volume  3,  Geneva,  Switzerland,  September 1-13,  1958.   United Nations,
     Geneva, Switzerland, 1958.  pp. 303-317.

Niki, H.,  E.  E.  Doby, and B.  Weinstock.   Mechanisms of smog reactions.  Adv. Chem. Ser. 313:
     16-57, 1972.

Niki,  H.,  P. D.  Maker,  C. M.  Savage,  and L. P.  Breitenbach.   Fourier transform IR spectro-
     scopic  observations  of  propylene  ozonide   in  the  gas  phase  reaction of  ozone-cis-
     butene-formaldehyde.  Chem. Phys. Lett. 46:327-330, 1977.

Noll,  K.   E.  and M.  J.  Pilat.   Size  distribution of  atmospheric  giant  particles.   Atmos.
     Environ. 5:527-540, 1971.

Novakov,  T.,  S.  G.  Chang,  and  A.  B. Harker.  Sulfate  as  pollution  particulates:   Catalytic
     formation on carbon (soot) particles.  Science (Washington,  DC) 186:259-261, 1974.

Orr,  C.,  F. K.  Hurd,  and W.  J. Corbert.   Aerosol  size  and relative  humidity.   J.   Colloid
          Interface Sci. 13:472-482, 1958.

Patterson,  R.   K.    Aerosol  contamination  from high-volume  sampler  exhaust.   J.   Air  Poll.
     Control Assoc., 30:169-171, 1980.

Patterson,  R. K.,  and J.  Wagman.   Mass  and composition of an  urban aerosol as a  function  of
          particle  size for  several visibility   levels.   J.  Aerosol Sci.  8:269-279,  1977.

Penkett,  S.  A.   Oxidation  of S02  and  other atmospheric gases  by ozone in aqueous solution.
          Nature (London) Phys. Sci. 240:105-106,  1972.
                                             2-96

-------
Penkett, S. A.,  F.  J.  Sandalls, and  B.  M.  R. Jones.  PAN measurements  in  England—analytical
     methods and results.  VDI-Ber. 270:47-54, 1976.

Penkett, S. A.,  B.  M.  R. Jones, K. A. Brice, and A. E. J. Eggleton.  The importance  of  atmos-
          pheric ozone and  hydrogen peroxide in oxidizing  sulphur dioxide in  cloud  and rain-
          water.  Atmos.  Environ. 13:123-137, 1979.

Perner, D.  D.,  D.  H.  Ehhalt,  H.  W.  Paetz,  U.  Platt,  E.  P.  Roeth,  and A.  Voltz.   Hydroxyl
          radicals in the lower troposphere.  Geophys. Res. Lett.  3_:466-468, 1976.

Pierson, W.  R.  ,  and P.  A.  Russell.   Aerosol  carbon in  the Denver  area in November 1973.
          Atmos. Environ. 13:1623-1628s  1979.

Porstendorfer, J.,  and T.  T. Mercer.   Influence of  nuclei concentration and humidity upon  the
          attachment rate  of atoms  in   the  atmosphere.   Atmos.  Environ.  12:2223-2228,  1978.

Rahn,  K.  A.,  R.  Dams,  J.   A.  Robbins,   and  J.  W. Winchester.   Diurnal variations of aerosol
     trace element concentrations as  determined by nondestructive  neutron activation  analysis.
     Atmos. Environ. 5:413-422, 1971.

Ramabhadran, T.   E. ,  T. W.  Peterson,  and J.  H.  Seinfeld.   Dynamics of  aerosol  coagulation  and
     condensation.  AIChE J., 22:840-851, 1976.

Rand, M. C.,  and S. B.  Gale.  Kinetics  of the oxidation of sulfites by  dissolved oxygen.   In;
     Principles  and Applications of Water Chemistry.  S. D. Faust  and J. V. Hunter, eds., John
     Wiley and Sons, Inc.,  New .York,  NY, 1967.  pp.  380-404.

Reinders,   W., and S. I.  Vies.   Reaction velocity of oxygen  with  solutions of some  inorganic
     salts.  Reel. Trav.  Chim. »44:249-268, 1925,

Renninger,  R.  G. ,  F.  C.  Hiller, and  R.  C.  Bone.   Comment on self-nucleation  in the  sulfuric
     acid-water system.   J.  Chem. Phys.   75:1584-5, 1981.

Riccoboni,  L.,  A.  Foffani,  and E.  Vecchi.  Studies  of the chemical kinetics of the process of
     autoxidation in dilute liquid phases.    Note  II.   Autoxidation of sodium  sulfite.   Gazz.
     Chim.  Ital. 79:418-442, 1949.

Robbins, R. C.  and R. D,  Cadle:  Kinetics of  the reaction between  gaseous ammonia and sulfuric
     acid droplets in an aerosol.  J.  Phys.  Chem. 62:469-71,  1958.

Robinson,   E.,   and  R.   C.   Robbins.   Sources,  abundance,   and  fate  of  gaseous  atmospheric
     pollutants.   SRI  Project  Report  PR-6755,  American  Petroleum Institute, New  York,  NY,
     1968.   '

Robinson,   E,,  and R. C.  Robbins.   Gaseous  sulfur pollutants from urban and natural  sources.
     J. Air Pollut. Control Assoc.  20:233-235, 1970.

Robinson,   R.  A.,   and  R.   H.  Stokes.   Electrolyte  Solutions.    2nd  Edition.   Butterworths,
     London, England, 1970.

Roedel, W., and W. Junkerman.  High volume sampling  of sulfur dioxide for isotropic studies by
     gas absorption in small droplets.   Atmos. Environ. 12:2399-2402, 1978.

Roedel, W.  Measurement  of  sulfuric  acid saturation vapor pressure; implications for aerosol
     formation by heteromolecular nucleation.  J. Aerosol Science,  10:375-386,  1979.

Salomaa,  P.,  R.  Hakala,  S.  Vesala, and  T. Aalto.   Solvent  deuterium  isotope  effects  on
     acid-base reactions.   III.  Relative acidity constants of inorganic oxyacids  in  light  and
     heavy water.  Kinetic  applications.  Acta Chem. Scand. 23:2116-26,  1969.


                                              2-97

-------
Sander, S.  P.,  and R,  T. Watson.   A kinetics study of the reaction of S09 with CH-09.  Chem.
     Phys. Lett. 77:473-475, 1981.                                        ^        d l

Sangster, J.  and  F.  Lenzi.   Choice of  methods  for the prediction of  the  water activity and
     activity coefficient for  multicomponent^aqueous solutions.  Can. J. Chem. Eng. 52:392-6,
     1974.

Satterfield,  C.  N.  Mass Transfer in  Heterogeneous .Catalysis.   M.I.T.   Press,  Cambridge, MA,
     1970.

Schenk, P.  W.,  and  R.  Steudel.   Oxidesr of« su.lphur.   In:   Inorganic  Sulphur Chemistry.  G.
     Nickless, ed., Elsevier Publishing Co.,  New York, NY, 1968.  pp. 365-418.

Schmidkunz,  H.  Chemiluminescence of .Sulfite Axidation.  Ph.D. 'Theses, Joljann Wolfgang Goethe
     University, Frankfurt, Germany, 1963.

Schroeter,  L.  C.    Kinetics  of  air, oxidation  of  sulfurous  acid  salts.   J.  Pharm.  Sci.
     52:559-563, 1963.

Schroeter,  L.  C.   Sulfur  Dioxide:  .Application  in ,f.oods>,  Beverages,  and  Pharmaceuticals.
     Pergamon Press, Oxford, England,  1966.

Schuetzle,  D.,  D.  Cronn,  A.  L.   Crittenden, and  R.  J.  Charlson.   Molecular ,composition of
     secondary  aerosol   and  its , possible origin.   Environ.   Sci.  Techno!.  9_;838-845,   1975.

Schwab, G.  M.,  and M.  Stwahmeyer..   On the kinetics of autoxidation of benzaldehyde by sodium
     sulfite in the absence of light.  Z.  Physk. Chemie, New Series 7_:132-157, 1965.

Schwartz, S. E.  and W.  H. White,  Solubility  equilibria of the  nitrogen oxides and oxyacids in
     dilute aqueous  solution.   In:   Adv-  Environ.  Sci. Eng. J.  R. Pfafflin and E. N. Ziegler,
     eds., Gordon and Breach, New York, NY,  in press, 1981.

Schwartz, W., P. W. Jones, C. J. Riggle, and  D. F.  Miller.  Chemical Characterization of Model
     Aerosols.   EPA-650/3-74-011,  U.S.   Environmental  Protection  Agency,  Research Triangle
     Park, NC, August 1974.

Scott,  W. D.,  and  P.  V.  Hobbs.   Formation  of  sulfate  in water  droplets.  J,.  Atmos.  Sci.
     24:54-57, 1967.

Seinfeld, J.  H.  and T.  E. Ramabhadran.  Atmospheric aerosol growth by heterogeneous condensa-
     tion.  Atmos. Environ., 9:1091-1097,  1975.

Sheih,  C.  M.   Mathematical  modeling  of   particulate  thermal  coagulation  and  transport
     downstream of an urban area source.   Atmos.. Environ., 11:1185-1190,  1977.

Siege!, S.,  H.  S.  Judeikis,  and  C.  C.  Babcock.   The Role of Solid-Gas Interactions in Air
     Pollution.   EPA-650/3-74-Q07, U.S. Environmental Protection Agency,  August 1974.

Singh,  H. B.  Atmospheric halocarbons:  evidence in favor of  reduced average  hydroxy! radical
     concentration in the troposphere.   Geophys. Res. Lett. 4:101104, 1977.

Smith,  B. M.,  J.  Wagman,  and  B.   R.  Fish.  Interaction  of  airborne  particles  with gases.
     Environ. Sci. Techno!.  9:558-562, 1969.

Soo,  S.  L.   Fluid Dynamics of Multiphase Systems.   Blaisdell  Publishing  Co.,  Waltham, MA,
     1967.

Stauffer, D.   On  the theory  of  lung deposition  of .very small H90-H,S04 aerosols.   Health
     Physics 26:365-366. 1974.


                                              2-98

-------
Stevens,  R,   K. ,  T.  G.  Dzubay,  G.   Russwurm,   and  D.  Rickel.   Sampling  and  analysis ot
     atmospheric sulfates and  related species.   In:  Sulfur in the Atmosphere, Proceedings of
     the  International  Symposium, United  Nations Environment  Program and Others, Dubrovnik,
     Yugoslavia, September 7-14, 1977.  Atmos. Environ.  12:55-68, 1978,

Stewart,  R.  W., S.  Hameed,  and  J.  P.  Pinto.  Photochemistry  of  tropospheric ozone.  JGR J.
     Geophys.Jies.  82:3134-3140, 1977.

Su,  F.,  J.  G.  Calvert,  and  J.  H.   Shaw.   A FT-IR  spectroscopic study  of  the  ozone-ethene
     reaction mechanism in 02-rich mixtures.  J.  Phys. Chem. 84:239-246, 1980.

Suck, S. H. and J.  R.  Brock.   Evolution of atmospheric aerosol particle size distributions via
     Brownian coagulation:  Numerical simulation.  J. Aerosol Sci., 10:581-590, 1979.

Suck, S.  H., E. C.  Upchurch, and J.  R.  'Brock.   Dust transport  in Maricopa County., Arizona.
     Atmos. Environ.,  12:2265-2272, 1978.

Suck, S.  H.,  P.  B.  Middleton, and J.  R.  Brock.   On the multimodality of  density  functions of
     pollutant aerosols.  -Atmos. Environ., 11:251-255, 1977.

Suck, S.  H., E. C.  Upchruch, and J.  R.  Brock.   Dust transport  in Maricopa County,, Arizona.
     Atmos. Environ. 12:2265-2271, 1978.

Takahashi,  K.   Changes  in particle  size distribution  of  aerosols  flowing  through vessels.
     Tech. Repts.  Eng. Res. Inst., Kyoto Univ., No. 149, 10 pp., 1970.

Tambour,  Y.  and J.  H.  Seinfeld.   Solution of the  discrete  coagulation  equation.   J. Colloid
     Interface Sci., 74:260-272, 1980.

Tang, I. N.  Phase transformation and growth  of aerosol particles composed of mixed salts.  J.
     Aerosol Sci.  7:361-371, 1976.

Tang, I.  N.   Deliquescence properties and particle size change of hygroscopic aerosols.  In;
     Generation of  Aerosols and  Facilities  for  Exposure  Experiments,  K.  Willeke,  ed., Ann
     Arbor Science Publishers, Ann Arbor, MI, 1980a.  pp. 153-167.

Tang, I.  N.   On the  equilibrium partial  pressures of  nitric acid and  ammonia in the  atmos-
     phere.  Atmos.  Environ. 14:819-828, 1980b.

Tang, I.  N.,  and  H. R. Munkelwitz.  Aerosol  growth studies.  III.  Ammonium bisulfate aerosol
     in a moist atmosphere.  J. Aerosol Sci.  8:321-330, 1977.

Tartarelli,  R.,  P.  Davini,  F. Morel!i,  and P.  Corsi.  Interactions between SO, and carbon-
     aceous  particulates.   lr\:   Sulfur  in the  Atmosphere,  Proceedings of  the International
     Symposium,  United   Nations  Environment  Program  and  Others,  Dubrovnik,  Yugoslavia,
     September 7-14, 1977.  Atmos. Environ. 12:289-293, 1978.

Titoff, A.   Contributions to the knowledge  of  negative  catalyses  in  a homogenous system.  Z.
     Phys. Chem. 45:641-683, 1903.

Tsang, T. H. and J.  R. Brock.  Aerosol coagulation  in the plume from  a cross-wind  line source.
     (Accepted for publication by Atmospheric Environment, 1981).

Drone,  P.,  H.  Lutsep,  C.  M.  Noyes, and  J. F. Parcher.   Static studies  of sulfur dioxide
     reactions  in air.  Environ.  Sci. Techno!. 2:611-618, 1968.

U.S. EPA.   Air Quality Criteria  for  Oxides of Nitrogen.  Draft final, EPA-600/8-82-026.   U.S.
     Environmental  Protection  Agency, Research Triangle  Park, NC,  September 1982.
                                             2-99

-------
Valensi, G., J. Van Muylder, and M. Pourbalx.  Sulphur.   In:  Atlas of  Electrochemical  Equili-
     bria in Aqueous Solutions.  M. Pourbaix, ed., Pergamon  Press, Oxford,  England,  1966.   pp.
     545-553.

van den  Heuvel,  A.  P., and B. J. Mason.  The formation of ammonium sulphate  in water  droplets
     exposed  to  gaseous  sulphur dioxide and  ammonia,   Q.  J.  R. Heteorol.  Soc. 89:271-275,
     1963.

Veprek-Siska, J., and  S.  Lunak.  The  role  of  copper  ions in copper  catalyzed autoxidation of
  .   sulfite.  I. Naturforsch 29b:689-690, 1974.

Vol'fkovick, S.  I.,  and A. P.  Belopol'skii.   Oxidation of  sulfites.   Report No.  1.   J.  Appl.
     Chem. 5:509-528, 1932.

Wadden,  R.  A., J.  E.  Quon  and H. M.  Hulburt.   A  model of a growing, coagulating  aerosol.
     Atmos. Environ., 8:1009-1028, 1974.

Weast, R. C., ed.  Handbook of Chemistry and Physics.   57th  Edition.  CRC Press,  Cleveland OH,
     1976.

Wedding,  J.   B. • and  M.   Weigand.   Sampling  effectiveness  of  the  inlet   to the dichotomous
     sampler.  Environ. Sci. Techno!.  14:1367-70, 1980.

Whitbys  K. T.,  A.  B. Algren,  R.  C.  Jordan, and  J.  C.  Annis.   The American Society  of Heating
     and  Ventilating  Engineers  and  airborne  dust survey.   J.   Air  Pollut.  Control  Assoc.
     7:157-165, 1957.

Whitby,  K. T.,  and  B.  Y. H. Liu.  The electrical behavior of aerosols.  In;  Aerosol  Science.
     N. Oavies, ed., Academic Press, New York, NY, 1966.  pp. 59-86.

Whitby,  K. T.,  R.  B. Husar, and B.  Y. H. Lui.   The aerosol size distribution of Los Angeles
     smog.  J. Colloid Interface Sci.  39:177-204, 1972.

White,  D.  R.,   and  J,   L.  Kassner,  Jr.    Experimental  and  theoretical   study   of  the  sign
     preference in the nucleation of water vapor.  J.  Aerosol Sci. 2:201-206, 1971.

Winkelmann,  D.   Die elektrochemische  Messung der  Oxydationsgeschwindigkeit von  NapSQ-j  durch
     gelosten Sauerstoff.  [The  electrochemical measurement  of  the rate of  oxidation or Na?SO-
     by free oxygen.]  I. Elektrochem. 59:891-895, 1955.                                   ^

Winkler,  P.   Chemical  analysis of Aitkin particles  (<0.2 |jm radius) over  the Atlantic Ocean.
     Geophys. Res. Lett., 2:45-48, 1975.

Wofsy, S.  C.,  J.  C.  McConnell,  and  M.  B.  McElroy.  Atmospheric  methane, carbon  monoxide,  and
     carbon dioxide.  J.  Geophys. Res. 77:4477-4493, 1972,
                                             2-100

-------
                       3.   TECHNIQUES FOR THE COLLECTION AND ANALYSIS OF
                  SULFUR OXIDES, PARTICULATE MATTER, AND ACIDIC PRECIPITATION

3.1  INTRODUCTION
     The 1969 and  1970  Air Quality Criteria documents  for  particulate matter (PM) and sulfur
oxides (SO ), respectively,  (National  Air Pollution Control Administration,  1969,  1970)  pro-
          A
vided a reasonably  thorough  review of measurement  techniques  available  at that time.   Subse-
quent advances  in measurement  technology  for  these pollutants have resulted  in   several  new
techniques and more information on the quality of data collected by older methods.   This chap-
ter provides  a  review  and,  where possible, a critical assessment of both  the  earlier tech-
niques used  in  historical monitoring  efforts  and  the  newer techniques on  which  much  of the
information gathered in the next few years will be based.
     Methods were  selected for inclusion in this review based primarily on frequency of their
use in past  or  current  studies.  These include routine monitoring applications used in demon-
strating compliance with air quality standards; in support of effects studies, especially epi-
demiology; and in examining long-term trends for the evaluation of control strategy effective-
ness.   More widely  used research measurement methods that have been used to collect important
ancillary data,  such as particle  size  distributions for aerosols, are also discussed  but in
less detail.
     Measurement techniques for SO ,  PM, and acidic precipitation are governed by the chemical
and physical properties of the substances to be measured.  Since the chemistry and physics of
SO  and ~PM are discussed in detail in Chapter 2, and those of acidic precipitation in Chapters
6-8,  only  the measurement methods per £e  are discussed in this  chapter.   Chemical  analysis
methods for  PM  and acidic precipitation for constituents such as sulfates are described fol-
lowing the sections  on  methods of sample collection.   The  relationship of particles to visi-
bility and their related measurements are discussed in Chapter 9.
     Discussion  of each  sampling and analytical  method covered  in  this chapter  includes  a
general description, a  discussion of the utility and  applicability of the method, and, where
information  is  available,  a  critical assessment of the method's capabilities.   The capabili-
ties  described  include  accuracy, precision, measurement  range,  sensitivity  to interferences,
and reliability.  The last parameter (reliability) is strongly influenced by competency of the
operator and  completeness  of accepted procedure documents.   Except in very specific cases, it
is  difficult to evaluate these  factors  and make conclusions about the  general  usefulness of
the method.   Hence,  an assessment of  the  quality of historical data  based  on reliability of
the method alone is virtually  impossible.  Many important earlier studies did not collect cer-
tain  quality  assurance  information now shown to be important in field monitoring (Von Lehmden
and Nelson,  1977).    In other cases, supporting data were collected, but are no longer avail-
able.   Therefore,  critical assessment of  the  methodology will focus on  those areas  that are
the most  important  to  the general  usefulness of the method, except  in  cases where specific
problems of a selected study were quantified in the open literature.

                                           3-1

-------
3.2  MEASUREMENT TECHNIQUES FOR SULFUR DIOXIDE
3.2.1  Introduction
     Atmospheric SO   originates  from both natural and  manmade  sources.   Sulfur diaxidertS09)
                   "•                                                                        &>
is the predominant  SO  in the atmosphere. This section discusses commonly used techniques for
                      >/\
determining atmospheric concentrations of S0_.
     Manual methods  for determination of S02 are those  in  which  sample collection, prepara-
tion, and  analysis,  or some  combination thereof, are  performed by  hand.    Automated methods
are those in which sample collection and analysis are performed continuously and automatically
by  devices generally referred to as continuous analyzers.
     This section briefly  describes each method, emphasizing measurement principle and method
characteristics, such as detection limits and interferences.  Sample collection and method ca-
libration are  discussed for both manual,and automated methods.  The most widely used manual
methods are discussed first.   Sulfation  methods are presented last because they measure "sul-
fation rate"  rather  than ambient  SO,  concentration  per  se.   The  discussion  of automated
methods follows a semichronological  order, with earlier continuous analyzers described first.
Much of  the descriptive  information in  the section is  based  on a  review  by  Tanner et al.
(1978).   Also discussed in this  section are various continuous analyzers designated by EPA as
equivalent methods for the measurement of atmospheric S0« to determine compliance with Nation-
al Ambient Air Quality Standards (NAAQS).
3.2.2  Manual  Methods
3.2.2.1  Sample Collection—A  number of  methods use aqueous solutions  for collection of S0?.
The efficiency  of mass transfer of  SO,  from air to the solution phase depends  on the gas-
liquid contact time,  diffusion coefficients of S02 in the gas and liquid phases, bubble size,
concentration of SO,,  and  solubility of S0? in solution.   Calvert and Workman (1960) describe
a method to predict the efficiency of various bubbler designs in collecting S02.  Their method
is predominantly qualitative,  but  it can serve as a useful guide.   The most  efficient designs
include that  of Wartburg  et  al. (1969);  the Greenberg-Smith impinger  (Smith  et al., 1961);
midget impingers  (Jacobs  et   al.,  1957);  Drechsel  bottles  (British Standards  Institution,
1963); and packed columns  (Bostrom,  1966), which are useful where low flow rates are involved.
In using such devices,  care must be taken to prevent carryover of solution at high flow rates
and to compensate for solvent losses by evaporation.
     Collection efficiency depends  in part on the solution in which SQ« is actively dissolved
and stabilized.   One  current  method involves stabilization  of  S0«  as the sulfite anion in an
aqueous solution of sodium or potassium tetrachloromercurate, with which the  sulfite anion com-
plexes.   To prevent  conversion of  sulfite to sulfate, the temperature of the collecting solu-
tion must be maintained below 20°C.   Failure to maintain temperature control  of samples during
collection,  shipment,  and storage  leads  to underestimation of SOp levels in  the atmosphere,
particularly during summer months.   Another approach involves collection  in  aqueous solution
and conversion to the sulfate anion by oxidizing agents such as H-,0,,.    Although stabilization
                                             3-2

-------
of SOp as  the  sulfate anion can be  effective,  some of the soluble sulfate in the atmospheric
aerosol is collected (unless removed by a particle filter) and added to the sample; thus,  dis-
crimination between SO- and sulfate may be impossible.
     Several methods employ alkaline solutions for the absorption of SO,,.   Although their  col-
lection efficiency  is quite  high,  alkaline  solutions  rapidly oxidize the  collected sulfite
anion to sulfate unless some means are available for the direct complexation and stabilization
of the sulfite anion.
     Another collection technique uses filter papers or tapes impregnated with an alkaline re-
agent  such as  potassium  hydroxide,  triethanolamine,  or  potassium carbonate,  together  with
small  amounts  of glycerol  as an humectant  (Lodge et  a!.,  1963; Huygen, 1963;  Pate et  a!,,
1963; Forrest and Newman, 1973).   The collected SO, is supposed to be maintained as a sulfite,
but it may be  oxidized to sulfate.  Although laboratory tests have shown that such oxidation
can be negligible, field tests have produced very erratic results.  Typical PM contains traces
of transition  metal   ions,  which  promote  rapid  oxidation of  sulfite  to  sulfate.   Prefilters
should be used  to  eliminate PM.    Oxidation of  the  collected  sulfite  to sulfate  prior  to
analysis  is  also recommended.  As  an  alternative, an analytical  technique  that measures the
sum of sulfite and sulfate may be employed.
     Some of the earlier methods for estimating ambient S0? concentrations (sulfation methods)
are based on the reaction of SO^ with lead dioxide to form lead sulfate (Wilsdon and McConnell,
1934).  The SO^  is  stabilized in the form of a sulfate, eliminating the problem of oxidative
conversion;  however,  any  particulate  matter containing sulfate  species  that  comes into  con-
tact with the collection surface will lead to errors.
     Occasionally, samples  of ambient  air are collected in a gas-tight syringe or other suit-
able  container  for   later  analysis.   The  reactivity  of  S09  is a  major  problem,  however.
                                                                                             ®
Natusch et al.  (1978) have reported extensive adsorption losses of SO,, on thick-walled Mylar
                ®        ®
laminates, Tygon  , Teflon  , and stainless steel container walls.
3.2.2.2   Calibration—The  relationship  between true pollutant  concentration and the measured
value  by  any  method is determined by calibration.  For methods that measure relative exposure
to sulfur  species (e.g.,  sulfation methods), no calibration is usually attempted.  With these
methods,  use of  uniform reagents, equipment, and  procedures  is essential to compare exposure
data over time and space.  Methods involving  direct collection of air samples for later analy-
sis or collection of the S02 in an air sample by absorption or adsorption require calibration
of both the sample volume measurement and the analytical measurement.
     Devices used for  sample volume measurement generally are calibrated against reliable vol-
ume  standards.    The analytical   measurement often  is  calibrated  statically,  using  a known
amount of  the  sulfite or sulfate anion in solution.  Static calibration is a rapid and simple
method  for checking  the  analytical  procedure,  but does not  subject  the overall measurement
method to  scrutiny  since the process of  S02 collection is circumvented.   Dynamic calibration
of these  methods has  an advantage  over  the static approach because  it scrutinizes the total
                                             3-3

-------
measurement,  but  it  is  time consuming  and  therefore  not  used routinely.   This approach,
described  in  more detail  in Section  3.2.3 on  automated  methods,  uses synthetic  atmospheres
containing the pollutant in  known concentrations to define the response of the method.
3.2.2.3  Measurement Methods—This  section deals with the principal manual methods for deter-
mining S02 in the air.
3.2.2.3.1   Colori'metric method:   pararosaniline.   The West-Gaeke method  is  probably the most
widely  used colorimetric  procedure  for   SCL  determination  in  ambient air  (West  and Gaeke,
1956).  It  is  also the basis of the EPA reference method for measurement of S0? in the atmos-
phere (U.S. Environmental  Protection Agency,  1979).  In the West-Gaeke method, air is bubbled
into fritted bubblers containing 0.1 M sodium  tetrachloromercurate (TCM) solution, which forms
a stable complex with SO,,.  This complex,  which resists air oxidation, was thought to be the
dichlorosulfitomercurate  (II)  ion.    Recently,  however,  Dasgupta  and  DeCesare  (1981)  have
clearly demonstrated  that the SO., group  is bonded to mercury through  the  sulfur atom rather
than through one of  the oxygen atoms  and  that the complex is actually a monochlorosulfonato-
mercurate  (II)  ion.   The SO^-TCM  complex  is  reacted  with  acid-bleached  pararosaniline and
formaldehyde to  form  red-purple pararosaniline  methanesulfonic  acid.   The  optical absorbance
of the  solution  is measured spectrophotometrically at  560  nm and is, within limits, linearly
proportional to  the concentration of SO,,.   The method is applicable to the measurement of SO,,
in  ambient air  using  sampling  periods  from 30  minutes  to 24 hours.   The lower  limit  of
detection of S09 in  10 ml of TCM absorbing solution is approximately  0.5  ug,  representing a
                             3
concentration of 13 |jg 50,,/m  (0.005 ppm)  in an air sample  of  38.2  liters.   Ozone, nitrogen
dioxide, and heavy metals were negative interferents in early versions of this method.
     An improved version of  the West-Gaeke method was adopted by the EPA in 1971 as the refer-
ence  method for determining atmospheric  S0?  (U.S.  Environmental  Protection  Agency,  1979).
Several  important  parameters were optimized,  resulting  in  greater  sensitivity and reproduci-
bility,  as  well  as adherence to Beer's  Law throughout a greater working  range.   In the EPA
method,  S02  is  collected  in impingers containing 0.04 M  potassium  tetrachloromercurate.   A
20-minute wait before  analysis  allows ozone,  a potential interferent, to decompose.  Sulfamic
acid is then added, followed by a 10-minute wait, to -remove interference from nitrogen oxides.
Interference by  heavy  metals is eliminated  by use of phosphoric acid  in the  dye reagent and
the disodium  salt of  ethylenediaminetetraacetic acid  (EDTA)  in the  TCM absorbing solution.
The complex  is  then  reacted with a purified pararosaniline dye reagent  and formaldehyde  to
form the colored pararosaniline  methanesulfonic acid.  Absorbance is measured at 548 nm.  Ac-
curacy depends on  rigid control  of  many critical  variables:  pH, temperature, reagent purity,
color development  time,  age  of  solutions, and concentrations of some atmospheric interferents
(Scaringelli et  al.,  1967).  Because temperature  affects  rate of color  formation and color
fading,  a constant-temperature bath  is recommended for maximum precision.   Highly purified re-
agents,  especially the pararosaniline dye, are vital for acceptable reproducibility.  The pre-
cision of the "EPA reference method  analytical  procedure  was  estimated using standard sulfite
                                             3-4

-------
samples (Scaringelli et a!., 1967) and reported to be 4.6 percent at the 95-percent confidence
level.  The  lower  limit of detection of  S0«  in  10 ml of TCM  absorbing solution was 0.75 ug,
                                           tf
representing a concentration of 25 ug S02/m  (0.01 ppm) in an air sample of 30 liters.
     A collaborative study (McCoy et a!., 1973) of the 24-hour EPA reference method indicated
the  following:  method  repeatability (day-to-day variability within  an  individual  laboratory)
varies linearly with S02  concentration from ± 18 ug/m  (0.007 ppm) at concentration levels of
100 ug/m  (0.04 ppm) to ± 51 ug/m  (0.019 ppm) at concentration levels of 400 jjg/m3 (0.15 ppm);
method  reproducibility  (day-to-day variability  between  two  or more  laboratories)  varies
linearly with S09 concentration from ± 37 ug/m3 (0.014 ppm) at 100 ug/m3 to ± 104 ug/m3 (0.040
                   3
ppm)  at  400 ug/m .   The  method  has  a concentration-dependent  bias.   This  bias  becomes
                                                          3
significant  (95-percent confidence  level) at the 400 pm/m  level.  Observed values tend to be
lower than the expected S0? concentration level.
     Results of the  above collaborative study and other investigations (Blacker et al., 1973;
Bromberg  et  al. ,  1974; Foster and  Beatty,  1974)  suggest that pararosaniline  methods  tend to
underestimate S0? concentrations by 5 to 20 percent.   In the Bromberg study, simulated 24-hour
bubbler samples were analyzed by 134 laboratories throughout the United States.  Observed neg-
                                                  3                                     3
ative biases ranged  from  -3 percent for a 45 ug/m  sample to -16 percent for a 767 ug/m  sam-
ple,  but  reasons   for  the negative  biases  have  not  been  determined.  Based  on the Bromberg
study results,  EPA recommended  that intralaboratory quality control  programs  be upgraded and
improved in laboratories that routinely analyze SO?-TCM samples.   EPA also recognized the need
for and promoted development of standard reference samples for use in laboratory quality con-
trol programs.
     More recent  information, on  the reliability of pararosaniline  analytical  procedures has
been obtained through EPA's ambient air audit program.   In this program, freeze-dried mixtures
of  sodium  sulfite  and  TCM  are  sent  to  participating  laboratories  for analysis.   These
simulated  field samples  represent  ambient S09 concentrations ranging from  about 10  to 200
    3
pg/m   (0.004 to 0.076  ppm).  EPA audit  results  from 1976-1978 summarized by  Bromberg  et al.
(1979, 1980) indicate  no  apparent problems with bias  (accuracy)  in  the analytical portion of
the pararosaniline methods.
     Subsequent to promulgation of  the  S0« reference  method, effects of  temperature  on the
method have  been studied  (Kasten-Schraufnagel  et al. , 1975; Sweitzer,  1975).    Fuerst et al.
(1976) showed that collected SO^-TCM samples decay at a temperature-dependent rate.  Table 3-1
indicates that  sample  collection at 25°C results  in a  1.1 percent loss in SO, during the 24-
hour sampling period, but further exposure of the collected sample for 4 days at this tempera-
ture  leads to  a 10 percent loss in S0?.  Significant decay can occur during collection of am-
bient samples and  during  shipment and storage of collected samples when TCM solutions are ex-
posed  to  temperatures  above  20°C.   Under typical  field conditions, temperature  exposure is
quite often  extreme, especially during the summer months at sites with relatively little pro-
tection from the elements (e.g., rooftops).
                                             3-5

-------
                  TABLE 3-1.  TEMPERATURE EFFECT ON COLLECTED S09~TCM SAMPLES
                                    (EPA REFERENCE METHOD)      *

°c
15
20
25
30..
35
40
op
59
68
77
86
95
104
' •
At end
of sampling
99.8
99.6
98.9
97.4
95.1
87.6
Percent

2
99.0
97.8
94.4
87.4
74.1
50.8
SOp remaining
Days of Exposure
4
98.2
96.1
90.2
78.5
57.9
29.5


• 6
97.4
94.3
86.1
70.4
45.2
17.2

     Source:  Fuerst et al. (1976).

     Measures to  minimize  these temperature effects have  been  investigated by Martin (1977),
who recommends use of thermostatted shelters to house sampling equipment during sample collec-
tion.  The  temperature of  samples  during shipment can be  controlled  with cold-pack shipping
containers.    When samples  are  stored before  being analyzed, refrigeration  at 5°C minimizes
further decay.   Temperature control  procedures are currently  being  incorporated  in the EPA
reference method.
     A variation  of the pararosaniline  method that eliminates the  use of the toxic mercuric
chloride has recently been  reported (Dasgupta et a!., 1980).  In this method a dilute solution
of  formaldehyde,  buffered at pH 4  with  potassium hydrogen phthalate, is  used to collect and
stabilize  atmospheric  SO,  as  hydroxymethanesulfonic  acid.   Sulfite,   liberated  from  the
compound by base, is added  to acidic pararosaniline for color development.  The optical absor-
bance of the  colored solution is measured  at  580 nm.   The method is comparable to the estab-
lished  pararosaniline  methods  in  absorption  and  recovery  efficiency,  sensitivity,  and
precision.   No unusual  interferences  are observed  due  to  0.,, N0?, and transition metal ions,
except Mn  (II).   The  stability of collected  S02  samples  is  significantly greater  than that
observed for samples collected in TCM.  The reported decay rates are 0.033 percent/day and 0.3
percent/ day at  room  temperature  and  37°C  respectively.   No photochemical  degradation of
collected samples was observed after 8 hours of exposure to bright sunlight.
     Under the provisions  of EPA's "Ambient Air  Monitoring Reference  and Equivalent Methods"
regulations  (U.S.  Environmental   Protection  Agency,  1979b),  two  additional  pararosaniline
methods  have been  designated  as  equivalent  methods  (U.S.  Environmental  Protection Agency,
1975).  These methods are identified as:
                                             3-6

-------
     1.   EQS-0775-001, "Pararosaniline Method for" the Determination of Sulfur
          Dioxide in the Atmosphere - Technicon I Automated Analysis System."
     2.   EQS-0775-002, "Pararosaniline Method for the Determination ofSulfur
          Dioxide in the Atmosphere - Technicon II Automated Analysis System."
These methods employ the same sample collection procedure used in the EPA reference method and
an automated analytical measurement based on the colorimetric pararosaniline method.
3.2.2.3.2  Titrimetric method:   hydrogen peroxide.  The British Standard Method uses a single-
day  or 7-day  sampling instrument  for the  measurement  of smoke  and SO,  (British  Standards
Institution, 1963).  Air is drawn through a filter paper and into a Drechsel bottle containing
~0.3 percent hydrogen peroxide solution adjusted to pH 4.5.  The H^Op oxidizes the atmospheric
S02  to HoSO,,  which  is  subsequently titrated  with  standard sodium  borate using  the  mixed
indicator of the  British  Drug House (gray at pH 4.5).  The method is capable of measuring SO,
                                           3
concentrations from about 25 to 25,000 jjg/m  (0.01 to 10 ppm) using a 24-hour sampling period.
     Since  the  method measures total  acidity rather than S0? specifically,  any  strong  acids
that are  collected  produce positive errors.  Normally the concentration of such substances is
low relative to that of SO,, and the measurement is generally accepted as a good approximation
of the actual  S02 concentration.   Ammonia will neutralize the HpSO, and give negative errors.
When the presence of ammonia is suspected, a portion of the absorbing solution can be analyzed
for dissolved ammonia and the S0? measurement adjusted accordingly.
     An instruction  manual  on  the use of the hydrogen peroxide method in the British National
Survey was  issued in 1966 (Warren Spring Laboratory, 1966).   The manual discusses the quality
of  the water  used  for reagent preparation  and states  that  it need  not  be free  of  carbon
dioxide.   Martin and Barber (1971), however, reported that use of water rich in carbon dioxide
can  lead  to  significant  negative errors   in  the  method.   During   sample  collection  and
subsequent  standing,  sufficient carbon dioxide can be evolved from the absorbing solution to
cause  low titers  and,  on  some  occasions, to  result even in alkaline solutions.   The instruc-
tion manual  also  discusses the problem of alkaline contamination in the glassware required in
the  method.   The Drechsel  bottles  used  during  sample collection  and  sample  storage bottles
need to be  conditioned with absorbing reagent prior .to use.   Likewise, alkaline contamination
in  other  glass parts  of  the  sampling apparatus can  lead to underestimation of ambient SO™
levels.
     Evaporation  of  absorbing  reagent during sampling can result in overestimation of ambient
S02  levels  with  the hydorogen peroxide method  (Fry,  1970).   If  evaporation occurs,  the pH of
the  solution is  lowered  and  a portion  of  the standard  alkali  added during the subsequent
titration is required to  compensate for  this  effect  alone.   The effect is  likely to be more
prevalent in the  summer months and can lead to overestimation of S0« levels by up to about 15
    3
M0/m   (0.006 ppm).   Fry reports that this source of error can be overcome either by making up
the absorbing solution to its original volume prior to the titration or by making a mathemati-
cal  correction  to the titration result based  on the  final volume of absorbing solution after
collection.

                                             3-7

-------
     Uncertainty  in  the tltration endpoint and  rounding-off  of the volume of alkali required
in the  titration  to the nearest  0.1  ml  each introduce errors  of  up  to about ±5 ug/m  (0.002
ppra) (Warren Spring  Laboratory, 1975).  Other potential sources of error in the method include
inaccurate  air sample  volume  measurements and  conversion  of  ambient  SO* to  sulfate  on the
smoke filter used in the sampler.
     Warren Spring Laboratory (1962) has reported the reproducibility of the hydrogen peroxide
method,  based  on results  from five  comparative studies using duplicate  sampling apparatus.
The  coefficients  of variation  were on the  order  of 15 to 20  percent  for SO, concentrations
                                 3
ranging from about 15 to 250 (jg/m  (0.006 to 0.095 ppm) and 5 to 10 percent for concentrations
                                   o
ranging  from about  100 to 800  ug/m   (0.033  to  0.305 ppm).    The same  reagents and analytical
apparatus were used to  service  all the samplers  in each study,  thus obviating a further poten-
tial source of error.
     In  a more recent investigation (Barnes, 1973),  duplicate  S0? measurements were obtained
with the British Standard Method at a residential site where ambient levels were low (18 to 84
|jg/m )  (0.007  to  0.032 ppm).    Nineteen sets of observations were made from two samplers with
a common inlet using the same supply of reagents and glassware  and a further 18 sets of obser-
vations  were  made using a  separate supply for  each.   Differences  in  measured concentrations
using the two  samplers on individual occasions ranged up to 31 percent of the mean of the two
separate values.  Most of  these differences, however,  did  not  exceed 13 percent of the mean.
Titration error was  cited  as  the single most common source of variation between the samplers
in  these experiments.   An error in  titration  of  0.1 ml would  result  in  an error  in the
                                      3
measured S0?  concentration of  7  ug/m  (0.003 ppm).   When measuring  low concentrations,  such
errors  could   represent a  difference  of  100  percent  from  the  true  concentration.   Barnes
concludes from these observations that measurement of  low  SO,  concentrations with the method
require great care on the part of the operator,  more than might be expected of most operators.
3.2.2.3.3  lodimetric methods.   Several iodimetric methods have been used for measuring S02 in
the  atmosphere.   In one version,  an absorbing  solution containing soluble  starch,  potassium
iodide,  dilute H^SO,,  and standard 0.01 N  iodine  solution  is  prepared  (Katz,  1950).   SO^ in
the  air sample reacts  with this 8 x  10    N iodine  solution to decolorize  the  blue iodine-
starch  complex.   The  reduction  in  color  intensity  is measured spectrophotometrically.   The
range of applicability is  25  to 2600 ug S0?/m  (0.01 to 1 ppm), depending upon the volume and
concentration of  absorbent  solution  and the volume of air sampled.  In a modification of this
method,  the excess  iodine is  titrated with  a standard thiosulfate  solution (Katz,  1969).
     Oxidizing gases  interfere to  give low results;  reducing  agents  interfere  to  give  high
results.  Interference  from high concentrations of  nitrogen oxides  or 03 can be  removed by
introducing hydrogen  into  the  air sample and passing  the mixture  over a platinum catalyst at
100°C (Bokhaven and Niessin, 1966).
     In another version, air  is bubbled through a sodium hydroxide solution that absorbs SO™
(Jacobs, 1960).  After acidification of the solution, the liberated sulfurous acid is titrated
                                             3-8

-------
with standard iodine solution, using starch as an indicator.  Because sulfite oxidizes to sul-
fate in the alkaline absorbent solution, samples cannot be stored.  Oxidizing agents, NO?J and
0_  interfere,  resulting  in an  underestimation of the  S0? concentration.   Hydrogen sulfide
(H?S)  and  other reducing  agents result  in  an overestimation.   For an 850-liter  air sample
                                                                                 1
collected  at 30  liters/minute,   the  lower  limit  of  detection  is  25  ug SOp/m   (0.01 ppm)
(Terraglio and Manganelli, 1962).
3.2.2.3.4   Impregnated  filter paper methods.   Filter papers,  impregnated  with alkali  plus a
humectant to keep  them  moist,  will absorb S02  from  air samples  (Lodge  et al.,  1963; Huygen,
1963; Pate et al., 1963; Forrest and Newman,  1973).   Two solutions commonly used to impregnate
papers are  a mixture of 20 percent potassium  hydroxide  and 10 percent triethanolamine, and a'
mixture  of  20  percent  potassium  carbonate and 10  percent glycerol.  The  treated  filters are
inserted into  filter holders, and air  is aspirated through them.   An  untreated  prefilter is
generally  recommended  to  remove  particulate  matter.   Absorbed S0?  can  be extracted from the
papers and  determined  colorimetrically  by the West-Gaeke method.  The alkali must be neutral-
ized exactly to  attain  the proper acidity prior to color development.  Alternatively, the ex-
tract solution may be treated with an oxidizing agent, such as l-LOp  to convert sulfite to sul-
fate,  followed  by a  sulfate  analysis  (Johnson and Atkins, 1975;  Forrest and Newman, 1973).
     Efficiency of SO,  absorption is better than 95  percent under  average weather conditions
but  decreases rapidly  below 25-percent RH and above 80-percent RH.  The error may be minimized
by using two filter papers in series (Forrest and Newman, 1973).  Elimination of glassware and
reagents during  sampling  removes  the  possibility of  spillage or  breakage during transport.
Sampled papers may be stored conveniently for long periods before being analyzed.
     Sulfur dioxide S02 may be sampled on Whatman No. 17 filter papers impregnated with tetra-
chloromercurate TCM solution containing mercuric chloride, sodium chloride, ethyl alcohol, and
glycerol in water (Axelrod and Hansen, 1975).  Sampled filters are extracted with TCM, and the
West-Gaeke  procedure generally  follows.   Capacity of the 47-mm filters is 13 mg of SO™, after
which  collection  efficiency decreases.   Samples 'collected  at  very low RH (10 percent) cannot
be  stored  more  than  1  day  before  exhibiting  losses.   Filters sampled at 40 percent RH may be
safely stored for 1  week.  No interference  is  observed for N02  and H2S,  but 0- at 175 yg/m
(0.09 ppm)  causes negative errors.
     A method that uses nondispersive X-ray fluorescence  to measure ambient S0? collected on
sodium carbonate-impregnated membrane filters has been developed  by  Hardin and Shleien (1971).
After  collection,  the  sample  filter is irradiated with  a one millicurie  iron-55 source.  The
resulting 2.3 kev sulfur X-rays are counted by a proportional counter with a beryllium window.
A minimum  detectable quantity of 30 M9 S09  can be detected by  the counter,  equivalent to 25
     3
(jg/m  (0.01 ppm)  using a  collection time of  1 hour and a  sampling rate of 20 liters/minute.
Chlorine gas is collected  to a significant degree and since  its characteristic X-ray cannot be
distinguished from that of sulfur, it may interfere to produce elevated results if not removed
prior  to encountering the  treated filter.
                                             3-9

-------
3.2.2.3.5   ChemiTuminescencemethod.    The  basis  for  this  method  is the  chemiluminescence
produced when  a sulfite  solution is oxidized  (Stauff and  Jaeschke,  1975).  Ambient  S02 is
absorbed in  50 ml  of tetraehloromercurate solution to  form the monochlorosulfonatomercurate
ion.   Five milliliters  of  2 x  10"5  N  KMn04 in  10"3 N  H£S04  is  added.   Oxidation  of  the
absorbed sulfite is accompanied by a chemiluminescence, which is detected by a photomultiplier
tube.  The total light yield,  measured  by a photon counting system, is proportional  to  the
                                     3                 3
oxidizable sulfite.   By  sampling 1 m  of air, 0.5,|jg/m  (0.2 ppb) of S02 may be detected with
an error of less than 10 percent.
3.2.2.3.6   Ion exchange  chromatographic method.   Small  et al.   (1975)  have described  an  ion
exchange chromatographic  system  that  separates ionic species and  effectively  neutralizes  the
eluant, allowing  a  conductometric measurement of  the ion.  A commercial instrument  based on
the  above  system  is now available (Dionex Corporation, 1975) for use in trace anion analysis.
In  this  system,  a strong base anion  exchanger of low capacity, agglomerated  onto  a  surface-
sulfonated DVB  resin,  is  used as the analytical  column.  This is followed by a high capacity,
strong acid  exchange  column that converts  the  eluant  (typically 0.003 M  NapC03  + 0.024  M
NaHCO,)  into   a  nonconducting carbonic acid  solution,  after  which  the  separated  ions  are
     «5
monitored  with  a  high sensitivity, multirange conductivity meter.  Although the method is not
totally free  from  ambiguity, careful  selection of  eluant  and  ion  chromatographic exclusion
steps can  effectively separate ionic species of interest.
     Mulik et  al.   (1978)  have  developed a  method for collection and  ion  exchange chromato-
graphic analysis of atmospheric SO,.   The method uses dilute (0.6 percent) HgOg to collect the
ambient S02.   The resultant sulfate  ion is  analyzed by ion exchange  chromatography.   When  a
prefilter  is  used  in  the sampling train to remove  aerosol sulfates, there  are no  apparent
interferences.    Collection 'efficiency  is approximately  100 percent over  the  range  of  the
method, 25 to 1300 ug S02 /m3 (0.01 to 0.5 ppm).
     A  novel   approach for  the  ion  chromatographic  determination of  atmospheric  SQ2  has
recently been  suggested by  Dasgupta  (1981).   Sulfur  dioxide is  collected  and  stabilized as
hydroxymethanesulfonic  acid  in   a  dilute  solution of  formaldehyde  buffered at  pH  4  with
potassium  acid phthalate  (KHP).  The sample  is  analyzed by an  ion  chromatographic procedure
using  KHP  as the eluant.   The hydroxymethanesulfonate ion elutes as a very sharp peak and the
analysis is facilitated  by the  fact that both  the sample and the eluant have the  same ionic
background of KHP, thus minimizing any undesirable phthalate peak or solvent dip.
3.2.2.3.7  Sulfation methods.  Sulfation methods  are based  on the reaction of gaseous S02 in
air  with lead  dioxide (Pb02) paste to form lead sulfate (PbSO^).  They are cumulative methods
for  estimating average concentrations over extended periods.  In the lead dioxide gauge method
(Department of Scientific  and Industrial Research, 1933)  and the lead candle method (Wilsdon
and  McConnell, 1934),  the  paste is prepared  by mixing  Pb02,  gum tragacanth,  alcohol,  and
water.  The  paste  is  applied to a piece of  cotton  gauze wrapped around a  cylinder  10 cm in
circumference  and  10 cm high.   After drying,  the  cylinder is exposed to the  atmosphere  in  a
                                             3-10

-------
sheltered  location.   After exposure,  the gauge  and  sulfated  lead  dioxide are  treated with
sodium carbonate  solution,  and  the  dissolved sulfate  is then  determined  gravimetrically or
turbidimetrically.  Measurements with  the method are reported  as  sulfation rates (mg S0,/100
  2 "                                                                                     o
cm /day).  In the sulfation plate method (Huey, 1968), a similar paste containing glass filter
fibers is  poured into  a  plastic petri  dish 48  mm in diameter.  After drying,  the  plate is
exposed to the atmosphere and analyzed for sulfate.
     Sulfation methods  have the  advantage of being inexpensive, but their accuracy is subject
to many  physical  and chemical variables  and interferents.   For example, the  rate  of sulfate
formation  is proportional  to  atmospheric SO, concentration up to 15-percent conversion of the
lead dioxide (Wilsdon and McDonnell,  1934).  Reaction rate increases with temperature and with
humidity.   Other  factors  affecting  rate  of sulfation  are  purity  of  the  lead  dioxide,  its
particle size  and shape,  wind  velocity,  and  shape of the shelter  (Bowden,  1964).   Positive
errors are contributed by hydrogen sulfide and sulfate aerosols.  Methyl mercaptan is a poten-
tial negative interferent.
     Huey  et al.  (1969) compared the  sulfation  plate  method  with the sulfation candle method
at some  250 sampling sites nationwide.  A correlation  coefficient of 0.95 was obtained, con-
firming  that  both methods  are  measuring  the  same species.   The  results also indicated that
sulfation plates are 10-percent less reactive than sulfation candles.
     Various attempts have been made to correlate sulfation methods with more specific methods
for estimation  of S0? concentrations.   In 1962,  as part of the establishment of the British
National Survey,  measurements with  the lead dioxide gauge were compared to simultaneous meas-
urements with the hydrogen peroxide  method (Warren Spring Laboratory, 1967).   The correlation
between  829 pairs of results  from 20  sites  over 4 years was highly significant, showing that
both methods were predominantly affected by the same pollutant, SOp.   The Warren Spring Labor-
atory  concluded,   however,  that there  was  no  generally applicable  conversion factor  for
relating  lead  dioxide  and  hydrogen  peroxide  results.   The conversion  from  lead dioxide to
hydrogen peroxide  reading was  not recommended except to give a rough indication of the levels
of concentration concerned.
     Stalker et al.  (1963) compared  the  lead  dioxide  method  and the pararosaniline method to
measure  SO™ at  123 stations in  Nashville, Tennessee.   The lead dioxide method was considered
good for estimating  mean  SO,  levels in communities during months with arithmetic mean concen-
                             3
trations of at  least 65 ug/m  (0.025 ppm).  The reliability of these mean estimates was esti-
mated to be within ± 25 percent.  Seasonal effects were noted, however, and the lead peroxide
estimates  of  SO,  (using  an average factor  of 0.031  for conversion of sulfation rate in mg
           2
S03/100  cm /day to S02 concentration  in  ppm)  during  the spring season of low SOg levels were
about twice as high as simultaneous 24-hour colorimetric measurements of SOg.
     Huey  et al.  (1969) compared ambient S02 measurements by conductometric, coulometric, and
colorimetric methods with sulfation results.   They concluded that sulfation data in mg SO.,/100
  2
cm /day  could  be  converted  to  S02  concentrations in ppm by multiplying by  0.03.   They also
                                             3-11

-------
determined that  95 percent of the  time  this  approximation from a single sulfation value will
lie within a factor of about 3 of any single measurement using the other techniques.
3.2.2.3.8  Other manual methods.  Other manual methods that have been used for the measurement
of  ambient concentrations  of  S0?  include the  barium perchlorate-thorin  titrimetric method
(Fritz  and Yamamura,  1955), the  barium sulfate  turbidimetric  method (Volmer  and Frohlich,
1944),  the barium chloranilate  colorimetric  method  (Bertolacini  and Barney,  1957),  and the
silica gel reduction method (Stratmann, 1954).
3.2.3  Automated Methods
3.2.3.1  Sample Collection—In continuous S0_ analyzers, sample collection is an integral part
of the  total  automated measurement process.  The sample line leading from the sample manifold
to the  inlet  of the analyzer should be constructed of an inert material such as Teflon .  The
sample line dimensions (length and  internal diameter) should be selected to minimize the resi-
dence time without creating a  significant  pressure drop between the  sample  manifold  and the
analyzer  inlet.   The  use  of an  inert  particle filter  at the  inlet  of the  analyzer should
depend  on  the analyzer's  susceptibility to  interference,  malfunction,  or damage  due to PM.
Heavy loading of PM on the filter may lead to erroneous SO, measurements; therefore, it may be
necessary to change the filter frequently.
3.2.3.2   Calibration—The  relationship  between true pollutant concentration  and the response
of a  continuous  analyzer is best determined  by  dynamic  calibration.   In dynamic calibration,
zero  air  and  standard  atmospheres containing known concentrations  of  S0? are introduced into
the analyzer  to  define the analyzer response over  the  full measurement range.  Dynamic cali-
bration  provides  evidence  that all  components  of the  instrument are  functioning properly.
     Standard atmospheres  required  for  calibration purposes may be generated using permeation
tubes (Q'Keeffe and Ortman, 1966), (i.e., sealed Teflon  tubes containing liquified gas).   Gas
diffuses through the walls at a low, constant  rate at constant temperature.   The gas  is then
diluted with  zero air  at  accurately known  flow'rates to  obtain S0?  concentrations over the
required  range.    Permeation  tubes  with  certified permeation  rates  are available from the
National Bureau  of Standards  (NBS)  as Standard Reference Materials (SRM's) or from commercial
suppliers.  Dynamic calibration  may also be carried out  using  known concentrations of SO^ in
high-pressure  cylinders.   To  ensure stability,  they are usually  prepared in  high  concen-
trations and  dynamically diluted to the desired level.  Traceability  of  such standards to NBS
SRM's may be established by the gas standard manufacturer or by the user.
     Static calibration  techniques  are  possible for  several  of the continuous  S02 analyzers
described  below.   Static calibration  introduces a stimulus to  measure  instrumental response
under no  sample  air  flow conditions. Typical  stimuli  are electrical signals, solutions chemi-
cally equivalent  to  the pollutant,  or solutions  producing comparable physical  effects upon
properties by which  the  pollutant is detected,  such  as  optical  density or electrical  conduc-
tivity.   Static  calibration is a rapid  and  simple method for checking  various  components of
the instrument, but does not scrutinize total  instrument performance.
                                           3-12

-------
3.2.3,3   Measurement Methods — The  principal  automated  methods  (continuous  analyzers)  for
determining S0? in the air are discussed in this section in chronological order.
3.2.3.3.1   Conductometric  analyzers.   Conductometric  analyzers were  the  first  commercially
available  instruments  for  continuously monitoring  atmospheric SO,.    They  are still  used
today.    In  their operation,  air  is brought  into contact  with an absorbing  solution,  which
dissolves SCL.  The ions formed by SOp dissolution increase the conductivity, which is propor-
tional   to  the  concentration.   The  absorbent may  be  either deionized  water   or  acidified
hydrogen peroxide solution.   When  water is used,  conductance is increased by formation of the
sulfite and bisulfite ions:
                        S02 •*• H20 t SQ2-H20 S H+ + HSO~                         (3-1)

                        HSO~ ? H+ + S0~2
Hydrogen peroxide solution oxidizes S0? to form sulfuric acid:
                        S02 + H202 •> H2S04 t H+ + HS04                          (3_2)
Conductance is measured by a pair of inert (platinum) electrodes within the cell.   To increase
accuracy, comparison  is  made  to a reference cell, which measures conductance of unused absor-
bent.  The response characteristics of Conductometric analyzers include lower detection limits
                                                  3
ranging  from  0.005 to  0.04 ppm  (10  to 100  jjm/m ),  lag times (time  interval  from  change in
input  concentration  to  change in output signal)  ranging from 5 to 200  seconds,  and response
times  (time  interval  from change  in  input  concentration  to 90  percent of  maximum output
signal) ranging from 1 to 4 minutes (Lawrence Berkeley Laboratory, 1972).
     The major disadvantage of conductometric analyzers is  their  susceptibility  to  interfer-
ence  by  any  species  that  either  forms   or  removes   ions  from  solution  and  changes  the
conductivity of  the solution.   The degree of  interference  depends on humidity,  temperature,
S0? concentration, and the particular instrument.   The worst interferents are chlorine, hydro-
chloric acid, and ammonia (Rodes et a!., 1969); nitrogen dioxide and carbon dioxide interfere
to a  lesser extent.   Airborne particles, especially oceanborne salt aerosols, are potentially
damaging.  Several  methods have  been used to  minimize these  problems.   Chemical  scrubbers,
which selectively remove gaseous interferents, have been incorporated into some Conductometric
analyzers.  Particle filters have also been employed.
3.2.3.3.2  Colon' metric analyzers.  Colorimetric analyzers are based upon reaction of SOg with
solutions of organic  dyes to  form colored species.   Optical absorbance of the resulting solu-
tion, measured spectrophotometrically,  is  within  limits linearly  proportional  to the concen-
tration of the colored  species in accordance with  Beer's  Law.   Most instruments use modified
versions of the  manual  pararosaniline method developed  by  West and Gaeke (1956).  Automation
of the  West-Gaeke method per  s>e does not  ensure  a practical continuous monitoring instrument
since some solutions require daily preparation.

                                             3-13

-------
     The  response  characteristics  for  some commercially  available  instruments  include lower
                                                             3
detection limits ranging from 0.002 to 0.01 ppm (5 to 25 jjm/m ), lag times ranging from 0.6 to
25  minutes,  and response  times ranging  from  5 to 30  minutes  (Lawrence Berkeley Laboratory,
1972).  Advantages  of these  instruments include  good  sensitivity and,  with proper control,
good  specificity.    Interferences  by  nitrogen  oxides  may be  controlled by using  a reagent
containing sulfamic  acid.   Heavy metals may be complexed  with  EDTA in the scrubbing solution
or with phosphoric acid in the dye solution.  Ozone interference may be controlled by use of a
delay coil downstream from the absorber to allow time for ozone to decay, but this results in
longer lag and  response times.  Major disadvantages of these instruments are the need for re-
agent and pump tubing replacement and frequent recalibration.
3,2.3.3.3  Coulometric andamperpmetric analyzers.   Coulometric  analyzers  are  based  on  the
reaction  of  S02 with  a halogen,  formed directly by electrolysis of  a  halide  solution.   The
current  necessary to replace  the depleted  halogen is  proportional to the  amount  of  SO^
absorbed in the solution, and hence to the SO, concentration in the air.
     In one common coulometric system, an inner chamber, into which air is introduced, is con-
tiguous with an  outer chamber (Treon  and  Crutchfield,  1942).   Both contain a  solution of
potassium bromide and bromine in dilute  sulfuric acid.   Potential difference between chambers,
relative to  a  reference potential, is measured by  the  reference electrodes.  As absorbed SO,
reduces the  Br2  concentration  in  the  inner  chamber,  the amplifier produces a  current to
restore the Br, content in the inner chamber until the potential difference is again zero.   In
a second  system,  the change in  halogen  concentration  is detected as  a  current  change rather
than a potential difference.  The cell is filled with a potassium iodide solution, buffered to
pH 7.  At the  platinum anode,  a  constant  current source continuously generates iodine, which
is  subsequently reduced at  the cathode.   An  equilibrium  concentration  of  iodine  is estab-
lished,  and  no current  is generated  at  an  activated-carbon  bipolar  reference  electrode,
connected in parallel.   Reaction with SO,, decreases the equilibrium concentration of iodine,
which  cannot transport  the charge generated  by the  constant-current  source.   Part  of  the
current is diverted through  the reference  electrode;  this flow  is proportional  to the  SO,
concentration in the air sample.  The response characteristics of modern coulometric analyzers
include lower  detection limits  ranging from 0.002  to  0.05 ppm (5 to  130  jjm/m ),  lag times
ranging from 2  to  120  seconds,  and  response times ranging  from  2  to 5  minutes (Lawrence
Berkeley Laboratory, 1972).
     Interferent species are  those able to oxidize halides,  reduce  halogens, or complex with
either.  They consist primarily of sulfur compounds (hydrogen sulfide, mercaptans, and organic
sulfides and disulfides) with  sensitivities comparable to or greater than that of SOg.  Other
potential  interferents, at lower sensitivities, are ozone,  nitrogen oxides, chlorine, olefinic
hydrocarbons, aldehydes, benzene, chloroform, other nitrogen- or halogen-containing compounds,
and other  hydrocarbons  (DeVeer et a!.,  1969; Schulze,  1966;  Thoen et a!., 1968; Washburn and
Austin, 1952).   Interferences can be minimized by selective filters, which are sometimes built
                                             3-14

-------
Into  the  instrument  or offered  as optional  features.   For example,  a heated  silver  gauze
filter js reported to remove hydrogen sulfide, ozone, chlorine,  nitrogen oxides, carbon disul-
fide, -ethylene;  aldehydes,  benzene,  and  chloroform, but will not  remove  mercaptans (Philips
Electronic Instruments, undated).
     Minimal  maintenance  is the  major  advantage  of a coulometric  analyzer  (reagent may need
only monthly replacement; electrodes  may  require annual  cleaning).   Also,  reagent consumption
is negligible because of halide regeneration, and evaporated water is replaced by condensation
from air or from a reservoir.
3.2.3.3.4  Flame photometric analyzers.  The  flame  photometric  detector (FPD) is based on the
                 	—.           ^
measurement of the  band emission  of excited  S?  molecules during passage of sulfur-containing
compounds through a hydrogen-rich (reducing) flame.   The emitted light passes through a narrow-
                                                  *
pass optical filter, "which  isolates the 394  nm  $2  band, and is detected by a photomultiplier
tube (PMT).  PMT  output is  proportional to  the  square of the sulfur concentration; hence, an
electronic system to "linearize" output is a desirable feature.   Application of the FPD to the
detection of  SQ~ was  first made by  Crider  (1965)  and  analyzers using FPD  have been widely
accepted for ambient  S0?  monitoring.   The response characteristics of continuous flame photo-
metric SO, analyzers  include  lower detection  limits  ranging  from 0.002 to 0.010 ppm (5 to"25
    3
fjm/m ),  lag  times  ranging  from  1  to  5  seconds, and  response  times  ranging  from 10  to 30
seconds (Lawrence Berkeley Laboratory, 1972).
     Although the  FPD  is insensitive to  nonsulfur  species,  it  will  detect sulfur compounds
other  than  S0».   Particle  filters  will   remove  troublesome  aerosol sulfates  and selective
filters may be  used to reduce1  interference  from  other gaseous  sulfur compounds (e.g., an H?S
filter is used on most commercial  instruments).  Interference by C02 can be minimized by main-
taining ambient levels of C0? in the calibration and sample matrices.
     Gas chromatographs with  flame  photometric detectors (GC-FPD)  are  also  available commer-
cially.   GC-FPD  can   separate  individual  sulfur  compounds and  measure  them  individually
(Stevens et  al,,  1971).  The temporal  resolution  of GC-FPD data,  however,  is  limited by the
chromatographic elution time of SO- and other gaseous sulfur compounds.
     Disadvantages  of  FPD  systems  include the  need for a source  of compressed hydr^g'-<\ and
sensitivity to all  sulfur compounds.   Advantages of FPD systems  include low maintener e, gocd
sensitivity, very  fast response,  and good selectivity for  sulfur compounds.   No reagents are
necessary, other than compressed hydrogen.
3.2.3.3.5    Second-derivative  spectrometric  analyzers.    The  second-derivative  spectrometer
processes  the  transmission-versus-wavelength function   of  a  spectrum to  produce a signal
proportional to the second derivative of this function (Hager and Anderson, 1970).  The signal
amplitude  is proportional  to  the  concentration of the gas in  the absorption  path.   These
instruments  center  on the shape  characteristics  rather  than  basic  intensity  changes of
molecular band  spectral  absorption.   The slope and  curvature characteristics are often large,
specific,  and  independent  of  intensity.   Because  these shape characteristics  are large but
specific to individual compounds, resolution  of component gases is possible.

                                              3-15

-------
     In the operation of a second-derivative spectrometer, radiant energy from a UV or visible
source  is  directed into a monochromator, where  it is dispersed by a grating to provide mono-
chromatic  light  to the sample cell. The wavelength of this light is modulated with respect to
time  in a sinusoidal  fashion by  an oscillating entrance slit.  The angular  position of the
grating sets  the center wavelength  (299.5 nm) coming out of the monochromator into the multi-
pass  cell.  The  sample  is continuously  drawn  through the  cell by  a  pump.   Output  from the
              4
photomultiplier  tube  is electronically  analyzed  to develop  the   second  derivative  of the
absorbance.
     Sensitivity is enhanced because the output is an AC signal of known wavelength and phase,
adaptable  to high-gain  electronic  amplification.   Uniqueness  of   the  curvature of  a given
molecular  band  enables  this  type  of  instrument  to  be  highly   specific.   A  theoretical
assessment  by Ratzlaff  and  Natusch (1977) indicates  that precision may  be a  problem with
spectrometric techniques  of  this type.   Measurements are  independent of sample flowrate, but
relatively  high  flowrates  (4  liters/minute)  are  necessary to  achieve reasonable  response
times.  The response characteristics for one commercially available instrument include a lower
                                     3
detection  limit  of 0.01 ppm (25 um/m ), lag  time of I minute, and response time of 8 minutes
(U.S. Environmental Protection Agency, 1979a).
3.2.3.3.6   Fluprescence analyzers.    Fluorescence  analyzers  are based  on  detection  of the
characteristic fluorescence  released by  the sulfur dioxide molecule when  it is irradiated by
ultraviolet  light  (Okabe et  a!,,  1973).  This  fluorescent light is also  in  the ultraviolet
region  of  the spectrum,  but at a  different wavelength  than  the incident  radiation.   Wave-
lengths between 190 and 230 nm are used for excitation and the fluorescent wavelengths usually
monitored  are between 300  and 400  nm.  In  this  region of the  spectrum, there  is relatively
little quenching of the fluorescence by other molecules  occurring  in  ambient air.  The light
is detected by a photomultiplier tube that, through the use of electronics,  produces a voltage
proportional to the light intensity and SO, concentration.  The fluorescent light reaching the
photomultiplier  tube is  usually modulated  to  facilitate the  high degree  of  amplification
necessary.   Some analyzers mechanically  "chop"  the incident irradiation before  it enters the
sample cell.  Other instruments  electronically  pulse the  incident light source at a constant
rate.  The response characteristics of fluorescence analyzers  include  lower detection limits
of 0.005  ppm, lag  times of  about  30 seconds,  and response times  of  about  5  minutes (U.S.
Environmental Protection Agency, 1979a).
     Potential interferences  to the  fluorescence technique include any species  that either
quenches or exhibits  fluorescence.   Both water vapor and  oxygen strongly  quench the fluores-
cence of SOp at some wavelengths.  Water vapor can be removed by a dryer within the instrument
or the  water interference  can be  minimized  by careful  selection  of  the  incident radiation
wavelength.  The effect  of  oxygen quenching can  be  minimized  by maintaining identical oxygen
concentrations in the calibration and sample matrices.
                                             3-16

-------
     Aromatic hydrocarbons such  as  naphthalene exhibit strong fluorescence  in  the same spec-
tral  regions  as  S0?  and are  major interferents. These  aromatics must  be  removed from  the
sample gas stream by  an appropriate scrubber  upstream  of  the sample cell.   The scrubbers  may
operate at ambient  or elevated temperature.   Certain elevated-temperature scrubbers, however,
have  the  potential  for  converting  ambient FLS  (which normally  does  not interfere with  the
fluorescence technique)  into  SOy-   In  these cases, the hydrocarbon  scrubber must be preceded
by a scrubber for H?S.
3,2.3,3.7  Other automated methods.   Other automated  methods (continuous analyzers) that have
been  used  for the  measurement  of ambient concentrations  of  SO,  include:   voltammetry (Chand
and Marcote, 1971); correlation spectroscopy (Barringer Research, Ltd.,  1969;  Moffat  et a!.,
1971); and differential lidar (Johnson et a!., 1973).
3.2.3.4  EPA Designated Eguivalent Methods—-Under provisions  of  EPA's  "Ambient Air Monitoring
Reference and Equivalent Methods"  regulations (U.S.   Environmental Protection  Agency,  1979b),
several commercial  continuous  analyzers  have been designated as equivalent methods for deter-
mining compliance with  National  Ambient  Air Quality  Standards for SO^.   These analyzers have
undergone  the  required  testing  and  meet  EPA's performance  specifications  for  automated
methods, summarized in Table 3-2.  A list of S0? analyzers designated from the promulgation of
the regulations  in  1975 to  December 31,  1980, is given in Table 3-3.   Information on designa-
tion  of  these analyzers  as  equivalent methods  may  be obtained  by writing  the Environmental
Monitoring Systems Laboratory, Methods Standardization Branch (MD-77),  U.S.  Environmental Pro-
tection Agency,  Research Triangle Park, North Carolina 27711.
     Review of performance  data submitted in  support  of  the designations listed in Table  3-3
indicates that these  modern  analyzers  exhibit performance better than that specified in Table
3-2.  For the analyzers tested, noise levels were typically 3 ppb or less.  The zero drift  re-
sults  (12-  and  24-hour)  were all   less  than  5  ppb  and typically less than 3  ppb.  The span
drift  results (at 20  and 80  percent  of  the  full scale range of  0 to 0.5 ppm) were all less
than  5 percent  and  typically 2 to  3  percent.   The precision results (at 20 and 80 percent of
the full  scale  range  of 0 to  0.5 ppm) indicate a typical precision of 1 to 2 ppb.  Lag times
were typically less than 1 minute.   Response times (rise and  fall times) for the various types
of  analyzers were typically as follows:   flame  photometric,  1 minute or less; fluorescence, 5
minutes; coulometric,  3 minutes;  conductometric, 0.5 minute; second-derivative spectrometric,
8  minutes.   For  analyzers   of  the  same  type  (e.g.,  flame photometric),  interference test
results  for  a given  potential  interferent were somewhat variable.  The  concentration of  S0?
during  the  tests was  0.14 ppm  and  the interferent concentrations were as  indicated in Table
3-4.   The  interference equivalent  for each  interferent must not exceed ±20 ppb and the total
interference  equivalent (sum of the absolute  values of  the individual  interference  equiva-
lents)  must  not exceed 60  ppb.  Interference equivalents  of 5 ppb or less were obtained in
each  case  except for the following:  flame  photometric-negative  CO-  interference equivalents
                                             3-17

-------
TABLE 3-2.  PERFORMANCE SPECIFICATIONS FOR EPA EQUIVALENT METHODS  FOR  S02
                            (CONTINUOUS ANALYZERS)3

Performance parameter
Range
Noise
Lower detectable limit
Interference equivalent
Each interferent
Total interferent
Zero drift, 12- and 24-hour
Span drift, 24- hour
20 percent of upper range limit
80 percent of upper range limit
Lag time
Rise time
Fall time
Precision
20 percent of upper range limit
80 percent of upper range limit
Units
ppm
ppm
ppm
ppm
ppm
ppm
percent
percent
minutes
minutes
minutes
ppm
ppm
Specification
0-0.5
0.005
0.01
±0.02
0.06
±0.02
±20.0
±5.0
20
15
15
0.01
0.015

Note:  1 ppm S02 = 2620 \jtg/m .
aSource:  U.S. Environmental Protection Agency (1979b).
                                    3-18

-------
      TABLE 3-3.   LIST OF EPA DESIGNATED EQUIVALENT METHODS  FOR SO
                         (CONTINUOUS ANALYZERS)

Designation
number
EQSA-1275-005
EQSA-1275-006
EQSA-0276-009
EQSA-0678-010
EQSA-0876-011

EQSA-0876-013
EQSA-0877-024
EQSA-0678-029
EQSA-1078-030
EQSA-1078-032
EQSA-0779-039
EQSA-0580-046
EQSA- 1280-049
Manufacturer
Lear Siegler
Meloy
Thermo Electron
Philips
Philips

Monitor Labs
ASARCO
Beckman
Bendix
Meloy
Monitor Labs
Meloy
Lear Siegler
Model
SM1000
SA185-2A
43
PW9755
PW9700

8450
500,600
953
8303
SA285E
8850
SA700
AM2020
Measurement principle
Second-derivative spectrometric
Flame photometric
Fluorescence
Coulometric
Coulometric
•B
Flame photometric
Conductometric
Fluorescence
Flame photometric
Flame photometric
Fluorescence
Fluorescence
Second-derivative spectrometric

The four digits in the middle of each number indicate the month and
year of designation.
                                  3-19

-------
                           TABLE 3-4.   INTERFERED TEST  CONCENTRATIONS  {PARTS PER MILLION)® USED IN THE TESTING
                                                     OF  EPA EQUIVALENT  METHODS FOR SO
Analyzer type
Flame photometric (FPD)
Gas chroaatography-FPD
Spectrophotonetric-wet
chemical (pararosaniline
reaction)
Electrochemical
Conductivity
Spectrophotometric-gas phase
Hydro-
chloric Am-
acid morn a
,.
—
0,2 0.1C
0.2 O.lc
0.2 O.lc
—
Hy-
drogen
sulfide
0.1
0.1
0.1
0.1
_.
--
Sulfur
dioxide
0.14d
0.14d
0.14d
0.14d
0.14d
0.14d
Nitro-
gen Nitric
dioxide oxide
..
—
0.5
0.5 0.5
0.5
0.5 0.5
Carbon Eth-
di oxide ylene
780
750
750
0.2
750
-.
Carbon
m- Water mon-
Ozone Xylene vapor oxide
20,000C 50
20,000C 50
0.5
0.5 -- 20,000C
--
0.5 0.2
 Concentrations of interferent" listed must be prepared and  controlled  to ±  10  percent of  the  stated value.
 Analyzer types not listed will be considered by the EPA Administrator as special  cases.
cDo not mix with pollutant.
 Concentration of pollutant used for test.   These pollutant concentrations  must be prepared to ±  10 percent  of  the  stated  value.
eSource:  U.S. Environmental Protection Agency (1979b).

-------
of about  10  ppb were typical; coulometric-positive 0,, interference equivalents" of about 8 ppb
were typical.
     As part of required  equivalency testing by manufacturers,  all  continuous  S0? analyzers
designated by  EPA  as equivalent methods have  demonstrated  a consistent relationship with the
reference method.  A consistent relationship is demonstrated when the differences between (1)
measurements made  by the test analyzer, and (2) measurements made by the reference method are
less than or equal to the allowable  discrepancy  specifications prescribed in the equivalency
regulations, when  both methods simultaneously measure S0? concentrations in a real atmosphere.
All the equivalent methods listed in Table 3-3 have demonstrated this consistent relationship
with the  reference method and the observed differences between simultaneous measurements were
generally well within the required specifications.
     A comparison  study using EPA designated equivalent methods for SCU was recently conducted
by  EPA  in an  urban/industrial/commercial  area of Durham,  North  Carolina  (U.S.  Environmental
Protection Agency, 1979a).   Eight continuous S0? analyzers  were  compared  over 150 days under
more  or   less  typical   air  monitoring  conditions.   During  the  study,   the  analyzers
simultaneously  measured  ambient  air  sampled  from a common  manifold.   The  ambient sample was
occasionally augmented with artificially generated pollutant to allow for analyzer comparisons
at higher concentrations.   A statistical comparison of hourly averages for each test analyzer
with  the average  of the  hourly  averages  (for corresponding  hours) from  the  other  test
analyzers is  presented  in Table 3-5.   Each test  analyzer is identified in the table by manu-
facturer, model  number,  and measurement principle.  The data clearly indicate that these con-
tinuous SCL analyzers are capable of excellent performance  (high correlation with one another,
small mean differences).
3.2.4  Summary
     Methods  for measuring of  SO- can  be  classified as:   (1) manual  methods,  which involve
collection of  the  sample over a specified time period and  subsequent analysis by a variety of
analytical techniques,  or (2) automated methods,  in which  sample collection and analysis are
performed continuously and automatically.
     In  the. commonly  used  manual  methods,  the  techniques  used  for the  analysis  of  the
collected  sample  are based  on  colorimetric,  titrimetric,   turbidimetric,  gravimetric,  X-ray
fluorescent, chemiluminescent, and ion  exchange chromatographic measurement principles.
     The  most  widely used  manual  method  for the  determination  of atmospheric  SO,  is  the
pararosaniline  method developed by West and Gaeke.  An  improved  version of this colorimetric
method, adopted as the EPA reference method in 1971, is capable of measuring ambient S0? con-
                              3
centrations  as  low as 25 yg/m   (0.01  ppm),  with sampling  times ranging from 30 minutes to 24
hours.  The  method has acceptable specificity  for  S02>  but samples  collected in tetrachloro-
mercurate  (II) are  subject to a  temperature-dependent  decay, which can  result in an under-
estimation  of the ambient SO, concentration.  Temperature control  during sample collection,
shipment, and  storage effectively minimizes this  decay problem. A variation of the
                                             3-21

-------
                  TABLE 3-5.  COMPARISON OF EPA DESIGNATED EQUIVALENT METHODS FOR SO, (CONTINUOUS ANALYZERS)0
Analyzer
Measurement
principle
Correlation
coefficient
Mean
difference
ppb
Std. dev.
of diff.
ppb
Max. abs.
diff.
ppb
No. of abs.
diff.
>20 ppb
No. of data
pairs
Meloy Monitor Labs
SA185-2A 8450
Flame Flame
photometry photometry
0.999 0.999
-3.695 -0.006
3.925 4.555
19.3 19.2
0 0
3302 3186
Heloy
SA285E
Flame
photometry
0.999
-0.251
3.243
15.5
0
3306
Thermo Electron Beckntan Lear Siegler
43 953 SH1000
Second-
derivative
Fluorescence Fluorescence spectronetry
0.997 0.998 0.936
-0.177 5.108 4.924
8.300 6.901 20.712-
29.9 25.4 100.9
49 21 427
2170 1594 1820
Philips
PW9755
Couloietry
0.998
5.775
4.631
25.6
13
3070
Bendix
8303
Flame
photometry
0.998
-3.278
4.392
21.0
1
1984
a8etween subject analyzer and average of other test analyzers (for corresponding hours).
 Source:  U.S.  Environmental Protection Agency (1979a).
Note:     1 ppb S02 = 2.62 vg/m3.

-------
pararosaniline  method  uses  a buffered  formaldehyde  solution  for  sample collection  and  is
reported to be less susceptible to the temperature-dependent decay problem.
     In Great Britain,  a titrimetric method based on collection of SOp in dilute hLOp followed
by titration  of the resultant HpSO. with  standard  alkali  is the standard method  and  is used
extensively.  Although simple  to  perform,  the method  requires  long  sampling times (24 hours)
and is  subject  to  interference from atmospheric acids and bases.   Additional sources of error
with this method  include evaporation of reagent during  sampling,  titration  errors, and alka-
line contamination of glassware.
     Methods that collect SO, with alkali-impregnated filter papers for subsequent analysis as
sulfite or  sulfate by a variety of techniques have  been developed.  Most  of these  methods
involve an  extraction  step  prior to analysis, although nondispersive  X-ray fluorescence has
been used for the direct measurement of S0? collected on sodium carbonate-impregnated membrane
filters.
     Two of the most sensitive methods  available  use measurement principles  based on chemi-
luminescence  and  ion  exchange  chromatography.   In  the   chemiluminescence   method,  SOp  is
absorbed in a tetrachloromercurate solution and subsequently oxidized with potassium permanga-
nate.    The  oxidation  of  the  absorbed  SO,  is  accompanied  by a  chemi luminescence  that  is
detected  by  a  photomultiplier   tube.   One  method,  using  ion  exchange  chromatography  to
determine  ambient   levels  of SQp that  have been absorbed into dilute  HpQp and  oxidized  to
sulfate, has been developed.   Another ion chromatographic approach, using a buffered formalde-
hyde absorbing reagent,  has been reported.
     Sulfation methods,  based on reaction of SQp with lead dioxide paste to form lead sulfate,
have  commonly been used to  estimate  ambient SOp  concentration  over extended  periods.   The
accuracy of sulfation methods is subject to many physical and chemical variables and interfer-
                                     2
ents.    Sulfation  rate  (mg SQ.,/100 cm /day)  is  commonly  converted to a rough  estimate of SO,
concentration (ppm) by multiplying the sulfation rate by the factor 0.03.
     Automated  methods for measurement of ambient levels of SO, have gained widespread use in
the air-monitoring  community.   Some  of  the  earliest  continuous  SO- analyzers  were based on
conductivity and coulometry.   These  first-generation analyzers were subject to interference by
a wide  variety of  substances  present in  typical  ambient  atmospheres.   However,  more recent
commercially  available  analyzers  using  these measurement  principles exhibit improved speci-
ficity  for  SOp through  the  incorporation  of sophisticated chemical and physical scrubbers.
Early continuous colorimetric  analyzers  using West-Gaeke type reagents and having good sensi-
tivity  and acceptable  specificity  for  SO,  were   fraught  with various  mechanical problems,
required frequent calibration, and never gained widespread acceptance.
     Continuous  SOp analyzers  using  the  techniques  of flame photometric  detection, fluor-
escence, and second-derivative spectrometry have been developed over the past 10 years  and are
commercially  available   from  a  number  of air-monitoring  instrumentation  companies.   Flame
photometric detection of ambient  SO, is based  on measurement of the band emission of excited
                                             3-23

-------
 *
$2 molecules  formed from sulfur species  in  a  hydrogen-rich flame.   The FPD analyzers exhibit
high sensitivity  and fast response, but must be used with selective scrubbers or coupled with
gas chromatographs when high specificity is  required.
     Fluorescence analyzers  are based on detection  of  the  characteristic  fluorescence of the
SOg molecule  when it is irradiated by  UV light.   These analyzers have acceptable sensitivity
and response  times, are insensitive  to sample flowrate, and require  no  support gases.   They
are subject  to interference  by water  vapor (due to quenching  effects)  and certain aromatic
hydrocarbons,  and therefore must  incorporate  some  means  to minimize  these species or their
effects.
     Second-derivative  spectrometry is  a highly specific technique  for measuring atmospheric
S0o>  and  continuous  analyzers based  on  this  principle  are  available  commercially.   The
analyzers are insensitive to sample flowrate and require no support gases, but relatively high
sample  flowrates  are  required  to achieve  reasonable  response times.   Excessive electronic
noise and inherent lack of precision can be  problems with these analyzers.
     Continuous analyzers  based on  many of the  above  measurement  principles (conductivity,
coulometry,  flame  photometry,  fluorescence,  and second-derivative  spectrometry)  have  been
designated by EPA as equivalent methods for  the measurement of SQ~ in the atmosphere.  Testing
of these analyzers by the manufacturers prior to designation demonstrated adequate performance
for use when an  EPA reference or equivalent  method is desired or  required.   EPA testing of
these methods verified their  performance and  has  also demonstrated  excellent comparability
among these designated methods under typical monitoring conditions.
3.3  PARTICULATE MATTER
3.3.1  Introduction
     As  described in Chapter 2, particulate matter  (PM) suspended  in  ambient air presents a
complex multiphase  system  consisting  of a  spectrum of aerodynamic  particle sizes from below
0.01 micrometer (urn) up to 100 urn and  larger.   Fine particles  (below  ~ 2  pm) tend to remain
suspended  in  air unless  removed  by  external  processes such  as rainfall.   Coarse  particles
(above  ~  2   urn)  have  appreciable settling velocities and  tend to   settle  unless  kept  in
suspension by high windspeeds-or turbulence.  The sources and characteristics of the particles
in both size ranges are generally  quite different, and depending on the objectives of the sam-
pling, measurements are often made that consider only a selected size fraction.  Samplers used
to identify fine  and coarse particle fractions typically are  designed to have inlet and sub-
stage cutpoints   that  are  as  sharp as  possible.   Samplers used  to simulate  the deposition
pattern of particles in the respiratory system have well-defined  but  more gradual cutpoints.
Lippmann (1970) summarized  samplers and deposition  patterns in  the  1-10 (jm range proposed by
several  organizations.  As Figure  3-1 shows these include models of the American Conference of
Governmental   Industrial Hygienists (ACGIH),  British  Medical Research Council (BMRC), and U.S.
Atomic  Energy Commission  (called the  "Los Alamos"  curve)   (Lippmann,  1970).   Miller  et  al.
(1979) proposed a sampler cutpoint of  15 pm related to respiratory system deposition but did
                                           3-24

-------
   100
 UJ
 oc
 O
O


O


p

cc
h-
Ul
Z
IU
a.
    40
    20
      02468


                 DIAMETER UNIT DENSITY SPHERE, pm


Figure 3-1.  Respiratory deposition models used as patterns for

sampler outpoints.

Source: Lippmann (1970).
                               3-25

-------
not recommend a  desirable  cutpoint sharpness.  Particle deposition  in  the respiratory system
is discussed in more detail in Chapter 11.
     The aerodynamic diameter  is  one of the most important physical parameters when consider-
ing  particle  deposition  in   the  atmosphere  or  the  human   respiratory  system.   Suspended
particles  are  rarely  spherical,  and characterizing  particle  size with  a  single  physical
dimension is often difficult.   Aerodynamic diameter is not a direct measurement of size but is
the equivalent diameter  of a  spherical particle of unit density that would settle at the same
rate.   This definition inherently considers such factors as particle density and shape without
requiring their  direct measurement.   Aerodynamic  diameters are  used in  this chapter unless
stated otherwise.   Sampling methods  using collection  or  separation techniques  based on the
inertia or  settling  properties of particles are classified according to the aerodynamic size.
In general, all  sampling methods  that draw the particles  into an inlet or opening perform an
aerodynamic size segregation.   However, particles with unusual  geometries,  such as long fibers
may not be  separated as  effectively as more spherical particles, since the orientation of the
fiber at the point of separation has a substantial  impact on the effective diameter.
     As Fuchs (1964) pointed out,  particle size distributions can be examined in several ways.
Separate distributions  of volume, surface area, and  number of particles, as  shown in Figure
3-2, can  be measured to provide  detailed information especially useful  in  studying particle
transport  and transformation.   The  particle  size  distribution by  mass  is  perhaps  the  most
important  characteristic of  an aerosol  to  consider with the  majority of current sampling
methods.    A mathematical  integration of  the mass distribution  function  over  the effective
aerodynamic collection  range  of  the sampler directly  provides the  total mass  collected per
unit volume of  air sampled.   This information can be obtained indirectly from volume, surface
area, or number distribution,  but an estimate of the average particle density must be included
in the calculations.
     The most common aerosol  measurement  made  in  conjunction with  health and welfare effect
studies is the mass concentration measurement. Direct measurement of the mass concentration is
made by collecting particles on a substrate, such as a filter,  gravimetrically determining the
mass of the particles,  and dividing the mass by the volume of air sampled.  Ideally, the par-
ticles reaching  the ^substrate  have been  segregated  by an efficient sampling mechanism that
provides a  defined  portion of the  ambient  size distribution  of  particles to  be collected.
Airborne [Particles  (National  Research Council,  1979) stated that ". . .  integral methods used
are always  sensitive to  the modification  of  the size distribution by the sampling inlets and
transport  lines  used  in the  technique."   This reference  notes that  Lundgren  (1973),  using
special samplers to  produce mass  size distributions  as  shown  in Figure 3-3, showed that most
mass  sampling  methods  truncate  the  true  ambient  particle  distribution,  thereby  giving
concentrations  less  than  those  actually  existing.    If these   less-than-perfect  samplers
operated  consistently in  all   conditions,  the mass  collected  would always  be  a consistent
proportion  of the  true ambient size  distribution,  assuming  a  constant distribution function.
                                             3-26

-------
 o
 *"
 X
  a
 O
 K
 1U
 eo
      I	   1.2
      —   1.0
h-Ji 0.8
HgO.6
   3
   ui
— < 0.4
   u.
   cc
      —   0.2
              o
              x
              ~"Q.
              Ul
              1

                                              l
                                                        ——. NUMBER
                                                        _ —_ SURFACE
                                                        — »— VOUIME
                                                                          A
                               0.01
                                         0.1              1

                                      PARTICLE DIAMETER, urn
                                                                       10
Figure 3-2. Plots illustrating the relationship of particle number, surface area, and volume
distribution as a function of particle size.
Source: Whitby (1975).
                                               3-27

-------
     180 —

     165

     150

     135
m
 1   «0
 a.
 Oft  105
 a
 3    go
 I
 ^    75
60 —

45

30
15
 0
 0.01
           MASS CONCENTRATION £lg/m3
           Q—.—i,,..   192»
              . _ __    93
                —    71
           0_.._    60
           «<_..._    26
           •FUGITIVE DUST EPISODE
0.1
1.0
10
100
                                PARTICLE DIAMETER (Dp),/um
  Figure 3-3. Typical ambient mass distribution data for particles up to 200 ^m.
  Source:  Lundgren (1973).
                                                                                    1000
                                         3-28

-------
•However,  it  has been determined that  many  sampler inlets are substantially affected by wind-
speed  and,  in  some  cases, wind  direction.   Some of the  commonly  used PM samplers employing
direct mass measurement techniques  include the TSP Hi-Volume Sampler, the dichotomous sampler,
cascade  impactors, and cyclone samplers.
     Mass concentrations of particles  can be estimated using methods that do not employ direct
weighing.   These indirect measurements  use analytical  techniques  other than direct weighing
for  assessing  integral properties  (other than  mass)  of  particles.   Typically,  an empirical
relationship  with a  gravimetric method is developed and  pseudo-mass concentrations reported in
lieu of  the integral property measurement.  Seta-ray attenuation by the particles on a filter
and  optical  reflectance of the darkening  of a filter by  collected  particles  are examples of
indirect measurement techniques.   Li  situ methods, which  examine particles still suspended in
the  airstream,  include a wide variety of techniques such  as the light-scattering measurements
of  the  integrating  nephelometer  and  the  size classification capability  of optical particle
counters.
     Analytical  measurement  of  the  chemical  composition of particles  can  be  strongly in-
fluenced by  the  sampling  method.   Surface measurements,  such as  X-ray fluorescence spectro-
scopy, require  a  filter that  retains particles on  the surface rather than allowing penetration
as  into  a  fiber filter.   The  composition and impurities in the  collection  substrate can be
critical, especially in the analysis of  trace  elements.  Selected substrates can also interact
with  ambient gases  to produce artifact particulate matter.   Section  3.3.4 contains descrip-
tions  of the most  common analytical  methods and Section 3.3.5.  briefly  discusses particle
morphology measurements by microscopic examination.
     Measurement  technology  for  aerosols  has advanced significantly  in  the past  10 years,
especially  in the area of size-specific measurements for  larger particles.  Before the advent
of  specially  designed  wind  tunnels  into  which  specific  aerosol  sizes and  types  can be
injected, determination of sampling accuracy (effectiveness) under conditions  similar to field
sampling had  rarely  been attempted.  For these tests, effectiveness  is  measured as the percent
of  the  particle  mass reaching the collection substrate  of  the sampler  compared  to results
obtained by  isokinetic sampling in  the wind tunnel.
     It  has been recognized  (Cermak,  1974) that the length and time scales of fluid motion in
a  wind tunnel can be  vastly  different from those existing in the atmospheric boundary layer.
While  ultimate agreement between the  performance  characteristics of an inlet  for large parti-
cles tested in a wind tunnel  and in  the atmosphere must  be  substantiated,  the use of a  wind
tunnel provides a tool to characterize  inlets over a wide range of test parameters under  con-
trolled  conditions.   Tests conducted  in the  atmosphere  will  only provide benchmarks for  com-
parison  because of  the lack  of  consistency of test conditions.   The comparability of artifi-
cial  and  real  particles  and the  turbulent  macroscale  on the  motion of particles  must be
addressed when considering wind tunnel  simulation.  While no models exist for predicting the
behavior of  large  particles  in  the   atmosphere,  such models  are  not  necessarily required to
                                              3-29

-------
substantiate  the  validity of wind  tunnel  data.   Wedding et al.  (1980)  developed a rigorous,
analytically  based  model for  predicting accurately  the  results of wind  tunnel  data  for the
Sierra dichotomous  sampler  inlet.   The turbulence of  the  flow approaching the sampler is not
considered by  this  model,  which describes the behavior of particles after entering the inlet.
In both  the wind  tunnel and  the  atmosphere, the  sizes  of the  turbulent eddies are  usually
larger than  the characteristic dimension of  the  inlet opening and are  much  greater  than the
particles' stopping distance.   Even  though  atmospheric eddies are much  larger  than those in
the wind  tunnel,  the inlet performances for the particle types tested could be expected to be
similar  in  either  regime.   Wind  tunnel test results, therefore,  provide controlled  testing
environments  for  intercomparison  of inlets,  but to date have not yet been used to predict the
mass concentrations realized in atmospheric situations.
     Researchers  such  as McFarland and Ortiz  (1979),  Wedding  et< al.  (1980),  and Liu  and Pui
(1980) have designed and built such test facilities for -characterizing aerosol samplers.   From
these tests,  it  is  now recognized that ambient  windspeed and direction  can  have a  profound
effect on particle  sizes reaching the point  of collection or  measurement within the  sampler.
Without  knowledge of  these  and related sampler characteristics, an accurate interpretation of
the  aerometric data  is  impossible.   This  section  describes  important  characteristics  for
commonly used sampler types, so that the usefulness of aerometric data discussed  in subsequent
chapters can be assessed.
3.3.2  Gravimetric PM MassMeasurements
     Techniques that employ  direct gravimetric weighing of particles collected on a substrate
are  discussed  separately  here.    Sampling  techniques  that  fall  into  this   category  are
extractive  rather  than  i_n  situ,   in  that the  particles  are  removed  from the  airstream for
subsequent analysis.   Typically the  ambient air  is  drawn into an inlet,  transported to the
collection substrate,  often after  one or more  stages of particle size separation, and then
deposited on  a substrate  by either  filtration  or impaction.    In addition to  the effect of
internal  separation  stages,  the particle size range collected by a filtration sampler depends
on other parameters such as inlet  geometry,  internal  wall losses, and  the efficiency of the
filter material.   The high-volume  sampler defined in the previous Air Quality Criteria for
Participate Matter  (National Air  Pollution Control Administration, 1969) and in the reference
method for TSP (U.S.  Environmental Protection Agency,  1979c)  was considered to have  captured
all  sizes  of  particles  up  to  100  inn  (aerodynamic  diameter).   However,  recent  sampler
characterization  testing by Stevens  and  Dzufray (1975), Wedding et al.  (1977),  and McFarland
and Rodes (1979)  has  shown that the  gable  roof used  as an  inlet  and  weather shield  actually
provides a D™ (the particle size at which 50 percent of the particle mass is passed on to the
filter)  of only 25  to 50 ym,  depending  on the windspeed.   As  shown by  the data of McFarland
and Rodes in  Figure 3-4, the sampling effectiveness of the hi-vol sampler for large particles
is  substantially  affected   by ambient  windspeed.   Lundgren   (1973)  has  examined the  mass
distribution of large particles up to 200 jjm  in the atmosphere as shown  in Figure 3-3.
                                             3-30

-------
  2
  i.
  w

  u
      100
      80
      60
      40
       20
 WIND
SPEED,
 km/hr
                                 50
                           2     45
                           8     44
                          24     30
                               \
             I    I  t  I  I I
                       I
I    I    I
                6     8   10            20           40

                    AERODYNAMIC PARTICLE DIAMETER,/urn
                                    60
Figure 3-4.  Sampling effectiveness of a Hi-Vol sampler as a function of wind
speed. Sampler rotated at 1 RPM and operating at 1A m3/min.

Source:  McFarland and Rodes (1979).
                             3-31

-------
Comparison of hi-vol sampler collection efficiency data in Figure 3-4 with these particle size
distributions  shows that  the hi-vol  sampler does not  provide a  true  measure of  the large
particles in the  atmosphere.   Because particle mass increases as a cubic function of diameter
for particles with constant density, the sampling of large particles must be treated carefully
when considering a broad size distribution.
     Size-specific  sampler  inlets  designed to limit the particles collected to a certain size
range  are  a relatively  new technology  for  particles  larger than 10 urn.   Since these larger
particles are  difficult to transport  in quantity,  a  sharp cutoff for large  particles is not
easily obtained except at high sampler flowrates using multiple stages of separation.  The ef-
ficiency of  a single  stage inlet  designed  in 1977  (Stevens and  Dzubay,  1978;  see Appendix
                                                                                         q
Figure 3A-1) to  provide a  15 urn  cutoff  for a  low flowrate sampler operating  at  1.0 m /hr,
(Wedding et  al.  1977),  is  shown  in  Figure  3-5.   Note  that the  D™ for  this  inlet  is very
windspeed dependent and varies  from  9 to  22 urn.   More advanced" inlets  (see Appendix Figure
3A-2 for diagram) for this flowrate have been designed by Wedding (1980)  and Liu et al. (1980)
and have  reduced windspeed sensitivities and sharper  cutpoints,  as shown  in  Figures  3-6 and
3-7, respectively.  The  geometric  standard deviations of the sampling effectiveness curves (a
measure commonly reported  in  the literature  of the sharpness of the size cut-off and denoted
as  o )  for  these  inlets  vary  from  approximately   1.2  to  1.5  as compared  to  an  ideal
step-function inlet with a cr  of 1.0.   The use of a  should be interpreted as only an estimate
                            y           *           y
of  the slope,  since  it implies that  the  effectiveness versus particle  size relationship is
log-normal,   which  is  rarely  the  case.   Values  of  o  typically  are  determined   either  by
dividing the particle  diameter associated  with an effectiveness of 84 percent into the D™ or
by dividing the D™ into the diameter corresponding to 16-percent effectiveness.
     After particles  pass  through the  sampler inlet they  caft be  lost from the  flowstream
before collection or measurement by attraction to or impaction on the internal surfaces of the
sampler.  Minimizing internal  loss,  especially for larger particles, requires careful design
of  the sample  transport system geometry as  well  as consideration of factors  such  as  surface
charge  dissipation.   Wedding  et  al.  (1977)  reported internal  wall  losses  in a  prototype
size-specific  sampler  to  exceed 40 percent for particles  greater than 15  urn.   Loo  et  al.
(1979) reported that  recent improvements in the dichotomous sampler reduced internal particle
losses to less than a few percent.'
3.3.2.1  Filtration Samplers—The most commonly used method for direct gravimetric measurement
involves  collection of the  particles  suspended  in  a   known  volume   of  ambient  air on  a
preweighed filter.   The size distribution of particles reaching the filter are affected by the
characteristics of  the inlet, the transport system,  and the separation  stages,  operating at
the sampler  flowrate.   The performance  of  a sampler  is  also  substantially  affected  by the
filter characteristics.  The efficiency of the filter medium used can influence the total mass
collected if very  small  particles  are not retained on the filter,  or if very large particles
bounce  from  the  filter  to  subsequent stages.   The collection efficiencies  over a range of
                                             3-32

-------
                                      152
    100
LU

Ul



6
u.
LU
     80
     60
     40
     20
  (SIERRA 244EINLET)

AVERAGE OF ALL TESTS

     O  5 km/hi


     A  15km/hr


     O  40km/hr
                                        I  I   I  M
                                                 1I   I
             3         5     7      10      15


           AERODYNAMIC PARTICLE DIAMETER, M
                                                              20   25 30
          Figure 3-5. Sampling effectiveness of the dichotcmous sampler inlet as a
          function of wind speed.

          Source: Wedding et al. (1980).
                                    3-33

-------
CO
W
UJ

IU



I
IU
u.
U.
UI

(9
Z
    120
    110
    100
                           AERODYNAMIC DIAMETER, (Jtm



            Figure 3-6. Sampling effectiveness of the Wedding IP inlet.

            Source: Wedding et al. (1980).
                                       3-34

-------

-------
particle sizes  for  a wide variety of filter materials, face velocities, and effective porosi-
ties  have  been determined  by Liu et  al.  (1978a)  for clean filters and by  John and Reischl
(1978) for exposed filters.  Appendix Table 3A-1 tabulates selected fractional efficiency data
                                                                        (8
for the  commonly  used TSP hi-vol sampler glass fiber filter, the Teflon  membrane filter used
by the  dichotomous  sampler, and the cellulose  fiber filter material (Whatman No.  1) used by
the BS  sampler.  The  latter filter shows  some inefficiency at  the  smallest particle sizes,
                                                        /a
while the  glass fiber and nominal 2 [jm  porosity  Teflon  filters are highly efficient for all
particle sizes.   The relationship  of flowrate through  the  filter to the pressure drop across
it  is also  a  very important  mechanical  consideration  since  this determines  the available
operating  flowrate   range  for  a vacuum pump  of  any  given  size.  Membrane  filter samplers,
because  of the  rapid increase in pressure drop as  particles deposit, require lower flowrates
than  fiber filter samplers.  This results  in substantially less PM  being  collected  during a
sampling interval and necessitates the use of a much more sensitive weighing device (balance).
3.3.2.1.1  TSP  high-volume sampler.  The  hi-vol  sampler is the EPA reference method  for TSP.
                                                        3
It is intended to operate at flowrates from 1.1 to 1.7 m /min, drawing air through a 200 x 250
mm (8 x  10 in) glass fiber  filter.  The mass of particles collected on the filter is deter-
mined from the  difference between weights before and  after exposure.   The mass concentration
is averaged over  the sampling interval and  is  normally expressed in (jg of mass collected per
 3                     3
ra  of air sampled ((jg/m ).
     Although materials such as quartz fiber can be used, glass fiber is by far the most com-
monly used filter  medium for  this  sampler and  is nearly 100-percent  efficient  for  0.3 urn
particles (Liu et al., 1978a).  As noted by Friedlander (1977), this size particle is  the most
difficult to  capture,  since the collection of smaller and larger particles is accomplished by
diffusion and impaction, respectively.   This filter material is not prone to rapid overloading
as is a  membrane substrate and  permits  sampling  over  24-hr periods in ambient TSP concentra-
                                   2
tions in excess of  300 to 400 ug/m .  Glass fiber filters, although available in a variety of
types, do  not  generally  provide a  chemically  inert surface, and  the  surface impurities and
basic pH may  interfere with some measurements.   The fibrous nature of  the  filter also makes
surface  measurements,  such  as  X-ray fluorescence,  impractical except  for  high  atomic number
elements such as lead.
     Ine hi-vol  is  relatively simple to operate  and reasonably inexpensive to purchase.  The
original method description in  the  Federal  Register  (U.S.  Environmental  Protection Agency,
1979c) was recognized  to  be  an  inadequate description  of  the procedure,  and a much more
detailed document was  prepared  by EPA (Smith and Nelson,  1973) to improve the quality of TSP
data.
     As  shown in Appendix  Figure  3A-3,  the inlet  is  formed  by the overhang  of  a gable roof
which serves  as a rainshield for the  filter.   The  inlet effectiveness, as already discussed,
does not produce  a  sharp particle size  cutoff  and  is sensitive to windspeed.  The collection
efficiency of the hi-vol  is also affected by sampler orientation (i.e., it is somewhat sensi-
tive to  wind  direction) as described by Wedding  et al. (1977).  The average sampler flowrate

                                             3-36

-------
is determined either by averaging single measurements before and after collection using an ex-
ternal flowmeter or  by integration of a  flow  recorder trace.   The effect of sampler flowrate
on the sampling  effectiveness  for large particles  as  shown in Figure 3-8 is not substantial;
however,  use of a flow controller provides the most accurate sampler performance.
     The absolute accuracy  of  ambient particle measurements such  as  those made by the hi-vol
sampler cannot be determined  directly with current  technology.  On the other hand, estimates
of components of the overall  accuracy can be  determined,  including the collection effective-
ness of the  sampler  inlet and filter media and the accuracy of the  flow measurement system.
Two  commonly used  flow measurement  devices  on  hi-vol  samplers   are  the rotameter  and the
orifice meter  with  a  pressure recorder.   The  rotameter is used  to measure  the  initial and
final  flowrates  from which  an  average  is  calculated.   The pressure  recorder  provides  a
continuous  trace of  the  orifice  pressure  drop that  can  be  integrated  for a  more accurate
measurement.  Smith  et al.  (1978) using hi-vol samplers with both types of devices noted that
the pressure recorder produced smaller errors (2 to 4 percent) when compared with a reference
flow device than the rotameters (6 to 11 percent).
     The precision  of the  hi-vol  sampler as determined from  collocated  sampler measurements
under field  conditions and  expressed by the coefficient of variation (cv) have been reported
by several investigators.   McKee et al.  (1971) determined the cv for a measurement by a single
analyst to  be  3.0  percent while the  same measure  among multiple  analysts  in  a collaborative
test was 3.7 percent.
     The design  of  the gable  roof provides  a settling chamber  above the  filter for larger
particles blown  in  during periods when the  sampler is  not  operational.   McFarland and Rodes
(1979) have determined  this   deposition  experimentally as a  function  of particle  size and
ambient  windspeed.    Interpreting these  relationships, however,   requires  knowledge  of the
existing ambient size distribution of particle mass.   For  a typical  distribution, the amount
of mass added to a hi-vol  sampler filter during 5 days of exposure when it was not operational
was  predicted  to be  6 to  8 percent.   This effect  has been measured in  a  field situation by
S.ides and Saiger (1976) and Lizarraga-Rocha (1976), who measured weight increases from 3 to 12
percent.   Errors from this effect can be  reduced  by equipping the sampler  with a mechanical
device that keeps  the filter  covered  during  nonsampling  periods.  Timely  installation and
retrieval of filters will  also minimize the problem.
     As shown  by Coutant  (1977), Spicer and Schumacher (1979) and Appel et al. (1979), arti-
fact PM can be formed  by, oxidation of acidic gases (e.g., SO^, NOp) or by retention of gaseous
nitric acid on the  surface of  alkaline  (e.g., glass fiber)  filters and other filter types.
The  effect  is a surface-limited reaction and,  depending  on  the  concentration  of the acidic
gas,  should be especially  significant early  in  the  sampling  period.   The magnitude of the
resulting error depends on such factors as the sampling period, filter composition and pH, and
the RH.  The magnitude and the significance of artifact mass errors are variable and dependent
                                              3-37

-------
     100
     80
  I   60


 tf
 tu
 z
 01


 p   40

 u
      20
           I  i  i  I  I   I  I  I  I  1   1  I  I  I  I   I  i  I  I  I
                                                           O--
           I  I   I  I  I  I  I   1  I  1   I  I   I  I  I   I  I   I  I  I
                    OS           1.0           1.5




                      VOLUMETRIC FLOW RATE, m3/min
2.0
Figure 3-8.  Effect of sampler flow rate on the performance of a Hi-Vol for

30 nm particles at a wind speed of 8 km/hr.



Source:  McFarland and Rodes (1979).
                                   3-38

-------
on local conditions.   Excluding the uncertainty associated with  the collection and retention
of  organic  particulate  matter with  appreciable  vapor  pressure,   artifact  mass  primarily
reflects the sum  of the  sulfates and nitrates formed by filter surface reactions with S03 and
nitric  acid  gas,  respectively.   The ambient concentration  of SO,  is  primarily dependent on
fossil  fuel  combustion,  while  the nitric  acid  concentration   is  dependent  on  atmospheric
photochemistry and,  possibly, reactions in suspended water droplets (Orel and Seinfeld, 1977).
A laboratory study  by Coutant (1977) reported artifact  SO,,  for 24-hour samples from 0.3 to 3
    3                                                 3
ug/m ,   Appel  et  al. (1978)  observed up  to 5  pg/m  artifact  SO.  on alkaline  glass  fiber
                                                            3
filters  in  24  hr  laboratory exposures,  and  up  to 3.2 ug/m   artifact  sulfate  in atmospheric
trials  at  two  California  sites.   Stevens et al.  (1978) similarly  found 2.5  pg/m  average
artifact SO,  based on sampling at  8  sites around  St.  Louis, Missouri; and Rodes  and  Evans
                     3
(1977) noted 0.5 pg/m  artifact sulfate in West Los Angeles,  California.
     Artifact particulate nitrate values on glass fiber filters ranging from 1.9 to 26.4 ug/m
(mean  10.6  ±6.9 ug/m ,  n = 13),  were  reported by  Spicer  and  Schumacher  (1979) in Upland,
California.  These  values were  obtained  by comparison  with  nitrate  concentrations measured
simultaneously with quartz  fiber  filters.   The  likelihood of negative sampling artifacts on
quartz fiber filters, as discussed below,  make these artifact nitrate measurements upper limit
values  only.   Appel et al.  (1980)  reported that artifact particulate  nitrate  on glass  fiber
filters  is  limited  only  by the gaseous nitric acid concentration.  Such filters approximated
total  inorganic  nitrate  samplers,  retaining  both particulate nitrate  and HNO.,  even when the
latter  was  present at very  high  atmospheric  concentrations  (e.g., 20  ppb).  Nitric acid was
found  to represent from approximately 25  to 50  percent of  the total inorganic  nitrate at
Pittsburgh,  Pennsylvania,  and  Lennox and  Claremont,  California.   Based on  an estimate of the
most  probable  24-hour artifact  sulfate error (3  (jg/m ), and  of the  most  probable artifact
particulate  nitrate (8.2  ug/m  in the Los Angeles, California, Basin and 3.8 pg/m  elsewhere)
typical  errors  in mass due to  SO,  plus nitrate  artifacts are estimated at  11.2 ug/m  in the
                              3
Los Angeles Basin and 6.8 pg/m  elsewhere.
     Nitrate salts  can rapidly be  lost from inert filters (e.g., Teflon, quartz) by volatili-
zation  (Appel  et al.,  1980; Forrest  et  al., 1980),  and by  reactions with acidic materials
(Harker  et  al.,  1977;  Forrest et  al.  1979).    Loss  of  atmospheric  nitrate  from glass  fiber
filters  occurs  slowly.  For example Smith et  al.   (1978) observed a  25-percent decrease in
nitrate  over 3 months in  storage  at room temperature accompanied by  a corresponding loss of
ammonium ion.   Colovos et al.   (1977) noted loss of up to 1.5 pg/m  NH, after 30 days storage.
Immediate analysis  after collection would minimize the significance of  such loss.
      In  general,  the hi-vol sampler data have been shown to be reproducible (3 to 5 percent),
if  an  orifice  meter  and  flow recorder  are  used.   The sampling  effectiveness  for larger
particles  is windspeed dependent and, based on the data in Figures 3-3 and 3-4, the effect of
windspeed could be  estimated to produce as much as a  10-percent day-to-day variability for the
same  ambient concentration  for typical  conditions.   The effect of  the sums of the reported
                                             3-39

-------
positive and negative artifact related to the glass fiber filter could be expected to add 6 to
7 M9/m  to the collected mass.
3.3,2.1.2  Dichotomous sampler.  The dichotomous sampler collects two particle size fractions,
typically 0  to  2.5 urn and 2.5 to about 15 jam, the latter cutoff depending on the inlet.  This
bimodal collection approximately separates  the fine  particles  from the  coarse particles as
described in Chapter 2 to assist in  the identification of particle  sources.   Since  the fine
and  coarse  fractions collected  in many  locations  tend to be  acidic  and basic, respectively,
this separation also minimizes potential particle interaction after collection.
     The  particle separation  principle  used by  this sampler  was  described  by  Hounam and
Sherwood (1965) and  Conner (1966).   As illustrated in a simplified fashion in Appendix Figure
3A-4, the separation principle involves acceleration of the particles through a nozzle, after
Which 90 percent  of  the flowstream is drawn off at right angles.  The small particles  follow
the  right  angle  flowstream,  while  the  larger  particles, because  of their inertia,  continue
                                                      ©
toward the collection nozzle.  A separate 37 mm Teflon  filter is used for each fraction.  The
sharpness of separation is  shown  in  Figure 3-9 from  data  by Loo et al.  (1976)  for  a design
cutpoint  at  2.5   urn.   Although  cutpoints  below  2.5  pm are  mechanically  impractical  with
dichotomous  separators,  3.5  urn units  are  available  commercially  with  equivalent  cutpoint
sharpness.   A 2.5  urn cutpoint  for the  separator was recommended by  Miller,  et.  al.(1979)
because  it   provided   good   chemical   separation  between   size   fractions,   satisfied  the
requirements  of  health  researchers,  and  was  mechanically  practical.   Inherent  in  the
dichotomous  separation  technique is  a  contamination  of  the coarse  particle fraction  with a
small percentage  of  the fine particles  in  the total  flowstream.   This  is  not considered a
substantial  problem for mass measurements and a simple mathematical correction as described by
Dzubay et al. (1977)  can be applied.
     Teflon  membrane filters  with  porosities as large as  2.0 (jm can be  used  in the sampler
and have been shown to have essentially 100-percent collection efficiency for 0.3 urn particles
(Liu et al . , 1978a).   Filters with smaller porosities,  such as 0.5 and 1.0 urn, a^e also highly
efficient,  but  are prone  to much more  rapid clogging  as  loading  increases,  accompanied by
                                                                                          3
rapid  decreases  in sampler flowrate.  Because the  sampler  operates at a flowrate  of  1 m /hr
(16.7  i/min)  and collects sub-milligram quantities  of particles, a microbalance with  a  1 ug
resolution is required for filter weighing.  Removal of the stickier fine particles causes the
collected coarse particles to  have a greater tendency to fall  off the filter  if  care is not
taken during filter handling and shipments (Shaw et al . , 1979).
     Dichotomous samplers are significantly more complicated to operate than single size frac-
tion  samplers  and  therefore are  more  prone to  operator  errors. As  with the low  flowrate
cyclone  samplers,  the  small mass  collected on  each filter requires  careful  weighing  on  a
microbalance to provide reproducible results.  The Beckman  inlet currently available for this
sampler is  shown in  Appendix Figure 3A-1.   Testing  has  shown that this inlet,  as  well as the
essentially identical Sierra  inlet, are significantly windspeed sensitive,  as shown in Figure
                                             3-40

-------
  120
                                  SEPARATOR EFFICIENCY
                           3     4567      10
                              PARTICLE SIZE (DJ.um
20
Figure 3-9. Separator efficiency and wall losses of the dtchotomous sampler at 2 5 p.m.
Source: Loo et al.  (1976).
                                     3-41

-------
3-5.   As the  windspeed  changes,  the D™  changes,  resulting  in  variable collection  of the
larger  particles.   As  noted  earlier, inlets  have  recently been  developed  that have sharper
cutpoints and are less windspeed dependent.
     Automated  versions  of  this  sampler  can automatically  change the  sampler  filters  to
provide  unattended  operation.   Depending on atmospheric concentrations, short-term samples of
as  little as  4 hours  are possible  with  the automatic  samplers  to provide  diurnal  pattern
information.   The mass collected during such short sample periods, however, is extremely small
                                                                *
and high  variability  of the results  could  be  expected.   With an inlet sampling effectiveness
at  15  km/hr as  described in  Figure  3-5,  the total  mass collected would be  5  to 10 percent
lower than the concentration collected by an ideal (a  = 1.0) inlet for a typical ambient size
                                 3                   "
distribution such  as  the 60 |jg/m   case  shown in Figure 3-3.   The  overall  reproducibility of
dichotomous mass measurements  is somewhat dependent on  the  care  taken during filter handling
and weighing, but could be expected to be about ± 10 percent.
3.3.2.1.3  Cyclone  samplers.   Ambient cyclone samplers are  simple  to  operate  and only moder-
ately  complex  to  build.    Lippmann and  Chan  (1979)  summarized   the  available  cyclones for
ambient particle sampling below 10 jam and noted that the separation effectiveness of cyclones
can be  designed  to match respiratory deposition curves closely (see Figure 3-1).  The cyclone
separation principle  can  be applied to larger particle  cutpoints,  as  demonstrated by Wedding
et al.  (1980)  for  a 15 urn  sampler  inlet.   The small size of  some cyclones  makes them useful
for  personnel  dosimetry  sampling,  if a  suitably  small  pump and flow  control  system are
employed.  The  Dorr-Oliver hydroclone,  which is 10 cm  in  length and  10  mm inside diameter,
matches  the  ACGIH  curve  (American  Industrial  Hygiene Association, 1970) and  can  be  used for
personal  sampling.   This cyclone  has also  been  used in ambient  field studies  including the
Harvard 6-City Study (Lioy, et al., 1980).
     A cyclone sampler used in the Community Health Environmental  Surveillance Studies (CHESS)
(Barnard, 1976)  is shown in  Appendix Figure 3A-5.   This sampler,  as  characterized in Figure
3-10, provides a relatively sharp separation  with a D™ of 3.5 p.m.  The inlet of the sampler
is the  cyclone  inlet,  and a single 0 to  3.5 urn particle fraction is collected on the filter.
The filter medium used in the CHESS network was glass fiber.
     At an operational  flowrate of 9.0 liter/minute, a typical fine fraction concentration of
       a
30 ug/m  would result  in  the collection of  only  390 pg  of  PM  on  the  filter in 24 hours.  At
this level,  Barnard (1976) determined the  reproducibility of this sampler to be 13  percent.
The effectiveness of  the  cyclone inlet for  particles  3.5 urn and smaller should be nearly 300
percent.  Use  of the glass fiber filter  would cause artifact mass problems similar  to those
identified with the hi-vol sampler.
     Lippman (1970) discussed  the  effect of  sample flowrate  on  the   performance  of cyclone
samplers.   Knight  and  Lichti  (1970)  compared the performance  of  the  10mm cyclone sampler to
that of  horizontal  elutriators and noted that the  results  were comparable if the appropriate
flowrates were  used.   Cap!an et al.  (1977)  noted that five different  flowrates,  from 1.4 to
                                             3-42

-------
100
90
80

S 70
g
i
FECTIVENESS,
en e>
o o
u.
UJ
0 40
13
a.
< 30
20
10
0
(
I ! I I I I I I _J 	 £
-«•**"" ""
-
X
ox
/
/
o
1
- /
I
1
1
(61111111
J12 3456789 1(
                           AERODYNAMIC PARTICLE SIZE,microns




Figure 3-10. Sampling effectiveness for the 3.5-jum outpoint CHESS cyclone sampler.



Source: Barnard (1976).
                                             3-43

-------
2.8  liter/minute,  have been  recommended  by researchers since 1962 in  order  for this cyclone
sampler to  meet  the ACGIH curve.  They also noted that these small samplers are unaffected by
ambient  air  velocity,  dust   loading,  mass  loading,  orientation,  or  aerosol charge.   The
reproducibility  of  this  sampler  has not  been  given in  the literature,  but  the  low sampler
flowrate and  proportionately  small  aerosol mass collected may result in values greater than ±
10 percent.
     Collection of the larger particles excluded by a cyclone sampler on a removable subsbrate
is difficult, but alternative approaches, such as that designed by John et al, (1978) shown in
Appendix Figure 3A-6, are available to provide a "total" sample dependent on the effectiveness
of the  inlet.  The  efficiency data  for  this cyclone  as a  function of  sampler flowrate are
shown in  Figure 3-11  and indicate  that  sharp cutpoints are possible  with current state-of-
                                    ©
the-art units.   A  neutral pH Teflon  filter  medium  was recommended to minimize artifact mass
formation.   The inlet normally used for this sampler is the dichotomous sampler inlet shov/n in
Appendix  Figure 3A-1.  This  inlet was  designed  to  operate at 16,7  £/min.    The  windspeed
influence  on  sampling effectiveness  is  shown  in  Figure 3-5,   Reproducibility data for this
sampler are not available but would be expected to be approximately 10 percent.
3,3.2.1.4   Hi-vol sampler withsize selective inlet.  To  meet the monitoring requirements for
Inhalable Particles (IP)  as proposed by Miller et al. (1979), EPA commissioned the design of a
size-selective inlet  for existing TSP hi-vol samplers to provide a single 0 to 15 urn particle
size fraction.   This  inlet is mounted on  a conventional  hi-vol  sampler in place of the gable
roof inlet.   (See Appendix Figure 3A-7.)  It has been tested by McFarland and Ortiz (1979) and
has  an  inlet  effectiveness as shown in Figure 3-12 and a sensitivity to windspeed as shown in
Figure 3-13.  Dry particle bounce and reentrainment were also reported to be insignificant at
                             3
the sampler flowrate of 1.1 m /min.
     The filter  materials used  are  the  same  as  those  used  for TSP  hi-vol  samplers, thereby
presenting  the  same potential  for artifact mass  formation.   This sampler, as  with any size
fractionating device, is somewhat sensitive to sampler flowrate for larger particles, as shown
in Figure  3-14.   However, these  data  suggest that special flow controlling  measures are not
necessarily required  to  maintain  consistent collection efficiencies over a  range  of sampler
flowrates.
     The  inlet  effectiveness  data shown  in Figure  3-12  would indicate  reasonably accurate
particle collection with minimal  windspeed dependence.   The influence of artifact mass on the
                                                                3
total mass  collected  could be expected to add about 6 to 7 (jg/m .   The reproducibility should
be similar to the 3 to 5 percent of the TSP hi-vol.
3.3.2.1.5  Elutriator samplers.  The British Medical Research Council (BMRC) (Orenstein, 1960)
defined  a   respiratory  system particle  deposition curve  (see  Figure 3-1).   This  deposition
curve is defined by the performance of a horizontal elutriator consisting of multiple parallel
plates  (Hamilton and  Walton, 1961).   A  schematic  diagram  of this  elutriator is  shown in
Appendix Figure  3A-8.   This sampler has  been  used in Great  Britain  for  ambient air sampling
                                             3-44

-------
                   AERODYNAMIC DIAMETER, fig

Figure 3-11. Fraction of methylene blue particles deposited in a cyclone
sampler as a function of the aerodynamic particle diameter. Curves are
labeled with flowrate in liters/min.

Source; John et al. (1978).
                                3-45

-------
    100
     80
     60
en
sa
ill
Z
u
ui
ui
     40
     20
                     I    I    I   I  I  I  I
       2            4       6     8   10            20            40


                    AERODYNAMIC PARTICLE DIAMETER, Mm



Figure 3-12, Sampling effectiveness for the size-selective inlet Hi-Vol sampler.



Source; McFarland and Ortiz (1970).
                                       3-46

-------
    20
    15 —
I

 o
 IB
O

LU
N
O
a.



o
    10
                                             I
_L
       04            8           12           16           20




                                      WIND VELOCITY, km/hr





   Figure 3-13. Effect of wind speed upon outpoint size of the size selective inlet.



   Source. Me Far I and and Ortiz (1979).
                          24
                                            3-47

-------
S.
ta
ui
z
    100
     80
     60
o

S    40
     20
          !   I   I  I  I   i   I  I  i  I  I  I   I   I  i  i   I  I   I   I
      nl  I   I  I  I   I   I  I  I   I  I  I  I   I  I  I   I  I  I   I   I
                    0.8           1.0           1.5



                     VOLUMETRIC FLOW RATE,m3/min
2.0
  Figure 3-14.  Effect of sampler flow rate on the sampling effectiveness of

  the size selective inlet Hi-Vol for a particle size of 15.2 jim and wind speed

  of 2 km/hr.



  Source:  McFarland and Ortiz (1979).
                                 3-48

-------
and during mining  operations  in the United States  as an occupational exposure sampler.   Corn
et al.  (1967)  successfully used a horizontal elutriator  to  collect ambient particles below 3
urn selectively  for optical examination on glass slides. Hamilton and Walton (1961) noted that
reentrainment  of  coarse particles can  be  a  problem in an elutriator  if mechanical  vibration
exists.  Because current ambient or wind tunnel  test data on these samplers are not available,
the reproducibility or accuracy cannot be estimated.
3.3.2.2  Impactor  Samplers—Impactor samplers provide a means of dividing an ambient particle
sample  into  subfractions  of specific particle  sizes.   As shown in Appendix Figure  A-9  for a
cascade impactor the jet of air is directed toward a collection surface, which is often coated
with an adhesive or grease to enhance collection.  Large, high-inertia particles are unable to
turn  with  the  airstream  and  consequently  impact against  the collection  surface.   Smaller
particles follow  the airstream  and  can be  directed  either to another  stage  of impaction or
collected on a  filter.    Use  of multiple  stages, each  with  a different  particle  velocity,
provides collection of  particles in  several  size  ranges.  Particle size  distributions  are
constructed using  impactor sampler data.  (See Figure 3-3.)
     Impactor  samplers  use removable impaction  surfaces  for collecting particles.   Impaction
substrates are  weighed  before  and after exposure and typically are metal foil plates or glass
fiber filters.  The selection and preparation of these substrates have a significant effect on
the impactor  performance.   Improperly  coated or overloaded surfaces can cause particle bounce
to lower stages,  resulting in substantial cutpoint shifts  (Dzubay et al.,r!976).  Marple and
Willeke (1976)  showed  the effect of various impactor substrates on the sharpness of the stage
cutpoint.   Glass fiber substrates can also cause particle bounce and are subject to the forma-
tion of artifact particles similar to those on hi-vol  sampler filters.
3.3.2.2.1  Cas cade impactors.   Cascade impactors typically have 2 to 10 stages, and commercial
                                                            3
low-volume version flowrates range from about 0.01 to 0.10 m /minute.   Lee and Goranson (1972)
                                         2
modified a commercially  available 0.03 m /minute  low-volume  impactor  sampler and operated it
          o
at 0.14 m /minute to  obtain  larger mass  collections on each  stage.   Cascade impactors have
also been designed to mount on a hi-vol sampler and operate at flowrates as high as 0.6 to 1.1
 o
m /minute.   A hi-vol sampler with a single impactor stage, shown in Appendix Figure 3A-10, was
                                                                                     o
used  in the  Community  Health  Air Monitoring  Program (CHAMP) and  operated  at 1.1 m /minute.
     Particle  size cutpoints  for each  stage are 'dependent primarily  on  sampler geometry and
flowrate.    The  smallest  particle  size  cutpoint  routinely  used is  approximately  0.3  urn,
although special low-pressure impactor samplers, such as that described by Hering et al. (1978)
are available  with cutpoints  as small  as 0.05 urn.  A high-efficiency filter typically is used
after  the  last impaction  stage  to .collect the  small  particles ,not impacted previously.   The
masses  collected  on each stage plus the  backup filter mass collection are often reported, as
shown  in  Figure 3-15  from data  by  Lee (1972).'   This  cumulative  d-istribution format permits
determination  of  the Mass Median Diameter  (MMD), at which  point 50 percent  of the mass is
smaller than the indicated size.  Use of straight  line plotting techniques (as shown in Figure
                                             3-49

-------
    10,0
     5.0
w

O
DC
U
CC
111

UJ
s


5
UJ

o

CC

IX
     2.0
     1.0
     0.5
     0.2
     0.1
           i      i      i       i     r~
                   CHICAGO, III. 1970 AVERAGE
            I       I      I
       12     5     10      20   30  40  50  60   70   80     90    95     98  99


                      CUMULATIVE PERCENT MASS <£  PARTICLE DIAMETER


      Figure 3-15. An example of a mass size distribution obtained using a cascade impactor.


      Source: Lee (1972).
                                           3-50

-------
3-15) implies  a log  normal  mass distribution,  which can result  in  misinterpretation  of the
mass median diameter.
     Cascade impactors are not normally operated in routine monitoring networks because of the
manual labor required  for sampling and analysis.  Although  impactor  sampling systems are not
extremely complex,  careful operation is required to obtain reliable data, especially if coated
collection  surfaces  are   used.   Analysis  of constituents  other  than  mass  to obtain  size
distributions  of  species such  as  sulfates  are possible,  but  require careful  analytical
techniques  or  compositing  by stage  with  other samples  to obtain  an  adequate quantity  of
material   for  analysis.   Impactor  stages  that use  grease coatings may  prove undesirable for
certain analyses because the grease may interfere with the method.  Natuseh and Wallace (1976)
investigated the errors associated  with impactor sampling and  concluded that even under very
unfavorable conditions the HMD can be determined to well within 25 percent of the true value.
     The inlet characteristics of most impactors have not been determined, resulting in uncer-
tainty about the size  range  of particles sampled.  McFarland (1980) examined the inlet of the
NASN low-volume (0.14 m /min) cascade impactor and determined that particles larger than 10 urn
were unlikely  to reach the collection stages because of substantial wall losses.  Willeke and
McFeters  (1975)   characterized   the   CHAMP  hi-vol   sampler  inlet  under  static  windspeed
conditions, as  shown in  Figure 3-16.   If the characteristics of the impactor inlet are known,
the total mass collected  by  the sampler can  be  used for comparison with  other  similar size-
specific measurements.
     The  particle  separation  of  an impactor  stage can  be very sharp, and mathematical  models
are available to permit stage sizing at selected cutpoints.  The single impaction stage of the
CHAMP hi-vol sampler designed to be 3.5 \an was  characterized by Ranade and Van Osdell  (1978)
and demonstarted a reasonable agreement with theory (Figure 3-lf).  Note, however,  that solid
particles  above 5  urn  deviate  from  the  relationship,   indicating  possible particle  bounce
effects.
3.3.2.2.2   Rotary  inertial impactors.  Whereas  cascade  impactors draw the fluid stream to the
impactor  surfaces,  rotary inertial  impactors move  the  impactor  surfaces  through  the fluid.
The impactor surfaces, which are spun by an electric motor,  are coated with a sticky film and
collect  particles  with  diameters  greater  than about  1  ura  by inertial  impaction.   The
collection  efficiencies  of such  samplers,  which  include the Rotorod (Balzer,  1972)  and the
Noll rotary impactor (Noll, 1970), are based on the Stokes number (Fuchs, 1964).
     Using  rectangular glass  slide collectors, at a rotational speed of at least 35 m/sec (78
mph), Noll  (1970)  measured efficiencies of 85 to 100 percent for particles from 6 to 108 urn.
The  collection  efficiencies  are  windspeed  independent if  the  rotational  velocity  of the
impactor  surfaces  is large compared to the ambient windspeed.
     Microscopic counting techniques  are  used to determine the  particle  distribution  on the
collector.  The  sample volume is the  volume  swept by the stage during the sampling interval.
Variations  of this  design principle  include exposed wires  and plates extended  from moving
vehicles  such  as boats and airplanes.

                                              3-51

-------
co
a
Ul
a
3


I
Ul
CO
Ul
_J
u
p
e
<
a.
1
o
<
DC
U.
    100
    90
    80
    70
60
50
40
30
    20
     10
     20.0
                   25.0
30.0
35.0
                  40.0
                  AERODYNAMIC DIAMETER, fim


 Figure 3-16.  Fractional particle collection of the CHAMP
 fractionator inlet at a sampler flow rate of 1133 liters/min.
 under static windspeed conditions.

Source: Willeke and (WcFeters (1975).
                             3-52

-------
 o


 a.
100





 90





 80





 70





 60
z
ui

0
E  50
u.
ui


2  40

U
IU


3  -
u


    20




    10
             CALCULATED FROM MARPLi'S THEORYH970)
                                                          AMMONIUM FLUORESCEIN

                                                          (SOLID)
                                                       O DI-OCTYL PHTHALATE

                                                          (LIQUID)
                                                  I    I    I   I   I   I
       1                  2          3      456789 10




                                 AERODYNAMIC DIAMETER, pn




Figure 3-17. Efficiency of the single impaction stage of the CHAMP Hi-Vol sampler.



Source: Ranade and Van Osdell (1978).
                                                                           15
20
                                          3-53

-------
     The advantage  of  this  type of sampler is that it eliminates inlet biases associated with
aspirated samples,  especially  for particles > 10  urn  in  turbulent air.  Because of the amount
of  labor  required  to  measure  large-particle  size distributions, they are  not practical  for
routine use in networks.
3.3.2.3   Dustfall  Sampling—Since very  large suspended  particles have  appreciable  settling
velocities, they are  collected by deposition in a dustfall container and weighed as described
by the American  Society for Testing and Materials (1981a).   Although a cylindrical jar might
be expected to collect the equivalent of the dust content of an air column of its own diameter
extending to the  top  of the atmosphere, the  aerodynamic  effects of the jar, the angle of ap-
proaching windflow, the mounting brackets for the jar, and adjacent structures tend to compli-
cate the collection pattern.  As noted by Nader (1958), it is difficult to interpret the mean-
ing of dustfall  data  and the  significance  of  correlations with other measurements.  The most
recent evaluation   of  available dustfall  measurement techniques was  reported by  Kohler  and
Fleck (1966),  who  noted  that  typical coefficients  of variation were  less  than  ±10 percent.
3.3,3  Non-gravimetric Mass Measurements
     A variety  of  particle measurement  techniques other than  direct  weighing are available.
Many of these  techniques collect the particles on a filter substrate, followed by an analysis
that  measures  an  integral  property  of the  deposited  particle other  than the  total  mass.
Examples include light reflectance,  light  transmittance, and beta-ray attenuation.   Other i_n
situ measurements  are  also  used, which do not deposit the particles on a filter but measure a
characteristic  of  the  suspended  particles,  such  as  light  scattering.    Host  of  these
alternative methods are  less'  expensive  per  sample  and  provide more  rapid  collection  and
analysis  of data   than  gravimetric analysis.   Some  measurements  are  not   generally  useful
because they  depend  heavily   on site-dependent  particle characteristics,  such as  color or
density.   In most  cases, a  scientifically-based physical model  relating the integral measure-
ments  to  mass  is  not  available,  thereby providing  no  basis  for  regression  analysis.   A
site-by-site best-fit regression must then be considered, which provides questionable accuracy
in  predicting  the  true  mass  concentration,  especially  if  the  composition  of the  local  PH
changes with time.
3.3.3.1  Filtration and  Impaction  Samplers—Samplers  in this category collect  particles  on a
substrate and then  use an alternative analytical technique as a surrogate to direct weighing.
3.3.3.1.1  British Smoke Shade sampler (BSS).   The design of the currently used BSS sampler is
based  in  part  upon early  work  by Hill  (1936),, who  used transmitted  light to  assess  the
darkness of  the stain resulting  from  particle collection on the filter  paper.   This sampler
draws an airstream  upward through an inverted funnel  and 3 meters of nominal one-quarter-inch
diameter plastic  tubing  to an  inverted filter  holder containing  a Whatman  No.  1 cellulose
fiber  filter.   As  noted earlier,  the  Whatman filter  medium has been  shown to  be  somewhat
                                             3-54

-------
inefficient when sampling  very small  particles (Liu et a!., 1978a).   A schematic diagram of a
version of this sampler designed to collect sequential samples for 8 days is shown in Appendix
                                   *
Figure 3A-11.   A  bubbler is  often used downstream  of the  filter holder  for  subsequent SCL
measurements.   The sampler  is operated at approximately  1.5  liters/minute, which is verified
by a  dry test meter built  into  the  sampler.   The filter  holder  can  be 25, 50,  or  100  mm in
diameter  to  collect  a  spot  of  the proper  darkness  range for  subsequent measurement  by
reflectance.
     Since  the  early  1900's,  many  studies  have  been  conducted   in  order   to  establish
relationships between  smoke shade reflectance readings and gravimetric measurements.   A 1964
study supported by the OECD established the currently  used relationships between smoke shade
reflectance measurements and  gravimetrically determined  particulate  concentrations.   These
data  were  accepted  by  the  WHO  (1976)  and compiled  into a  standard  operating  procedure for
reporting smoke shade  measurements in equivalent ug/m .  These equivalent mass concentrations
are not  determined by weighing  the smoke  shade  sampler filter but through comparison  with a
collocated  gravimetric  sampler.   The gravimetric  measurements  that  were  made   for  OECD and
compared to the smoke  shade measurements were called "high-volume sampler" readings, but were
not taken  with the  U.S.  TSP  hi-vol  sampler.   The OECD  gravimetric  "high-volume sampler" as
described  by  the   British  Standards  Institution  (1964)   operates  at   approximately  60
                                     2
liters/minute, compared  to  the 1.5 m /minute of the U.S. hi-vol  sampler.   The OECD hi-vol was
designed to be  aerodynamical^ similar to the smoke shade unit but has not been characterized
for  aerosol  collection  effectiveness.    Prior  to  1964,  various  calibration  curves  were
                                                                            3
published relating smoke shade reflectance to mass estimates in nominal pg/m  units, but.it is
difficult to compare these  earlier relationships to the OECD version.
     McFarland  (1979)  examined  the  aerosol  collection properties of  the  smoke  shade sampler
and  produced  the  effectiveness  plot  shown  in  Figure  3-18, which   shows  that  the D™ for
particles  reaching  the  filter  (entire  system)  is only  about 4.5 jjm.   A comparison  of the
"entire  system"  data  with  the  deposition models  given in  Figure  3-1 shows a  coincidental
agreement within  1  urn  of -the ACGIH  and Los  Alamos curves.   Most  large  particles  are  either
rejected at the  inlet  or lost in the inlet line, although some particles as large as 10 pm do
reach the filter.   Because  the size range of particles collected by the smoke shade sampler is
substantially  less  than that  collected by  the TSP  hi-vol  sampler,  results of comparisons
between the methods  could be expected to vary.
     Ball  and  Hume   (1977)  and Waller  (1963)  noted  that  consistent  relationships can  be
developed, but are site, season, and particle-source dependent.  Lee et al.  (1972) noted, from
collocated  TSP  hi-vol  and  smoke shade  sampler comparisons made at various sites in England,
that  the  overall  correlation coefficients between these measurements  for all sites was 0.618.
However,  the  individual  coefficients   ranged  from  0.936  (good correlation)  to  0.072 (no
correlation).   Bailey  and Clayton (1980) showed that smoke shade measurements correlated more
closely  with  soot carbon content than  with gravimetric mass.  Recent work by Edwards  (1980)
                                           3-55

-------
   100
    80
&
tf
ui
ui
o
iu
u.
u.
UI
1
    SO
    40
    20
                 1      I     I    I    I   I   1  1
                           "O
                               I

                        INLET ALONE
                         O 2 km/hr
                         • 8 km/hr

                      ENTIRE SYSTEM
                         A 2 km/hr
                              I
J	I
                              5       7       10                         30

                             AERODYNAMIC PARTICLE DIAMETER, ftm
Figure 3-18. Sampling effectiveness of the inlet alone and through the entire flow system of the
British Smoke Shade sampler.
Source; McFarland (1979).
                                         3-56

-------
has  shown  that  smoke  shade  reflectance  measurements  can  be  related  to  the  absorption
coefficient of  the atmosphere.   This  work also  showed that smoke  shade measurements can be
converted to approximate COH measurements made by the American Iron and Steel Institute (AIS!)
tape  sampler  (see  Section 3.3.3.1.2)  using  the absorption  coefficient  relationships.   As
several investigators  noted,  (e.g.,  Lodge,  et al. 1981), if a relationship could be developed
between optical  and gravimetric  measurements, it would  be site  specific but still  variable
because of seasonal  and  long-term  differences  in  the  sources  of collected  particle size
fractions  and  their   carbon  content.   The  smoke  shade  sampler  is  relatively  simple  and
inexpensive to  use  for routine monitoring.   The British Standards Institution (1964) reported
the reproducibility of collocated smoke shade sampler measurements as 6 percent.   The accuracy
of a  given  relationship's  predicted mass concentrations  from  reflectance measurements cannot
be  discussed  in  general terms  because of -the  previously mentioned  confounding influences.
Subsequent use of smoke shade results elsewhere in this document (Chapter  14) will discuss the
accuracy of the measurements relative to the specific studies evaluated there.
3.3.3.1.2  Tapesampler.   A  variation of the  optical measurement  of',€pot darkness is the use
of  a  continuous filter  tape  and an  automatic tape  advancing system.   A sampler; using this
approach,  developed by Hemeon (1953) for the  AISI,  samples at a'flowrate of approximately 7
liters/minute using Whatman  No.  4 filter paper and collects particles on  a 25 mm filter spot.
                                                                                 f *
The spot darkness is read either by a transmittance or reflecta'nce measurement.  Transnffttance
measurement is most popular in the United States.
     The AISI tape  sampler typically collects particles  in selected time intervals of 1 to 4
hours, and  then advances to an unexposed clean portion of/the tape.  Optical measurements are
referenced to an  unexposed filter area and can  be made either external  to the  sampler after
sample collection or with a continuous readout self-contained in the sampler.
                                                       •6
     Transmittance  measurements  are  converted  to  optical  density  through  a Beer's  Law
relationship and then  to CoH units per  1000  linear  feet of air sampled.  A CoH  is defined as
the quantity  of PM on the paper  tape" that  produces a change in optical  density  of 0.01.  The
alternate  Reflectance  Unit of Dirt Shade (RUDS)  is  equivalent to 0.1 CoH units  per 1000 feet
(American Society for Testing and Materials, 1981b).
     As stiown  in  Appendix Figure 3A-12, this  sampler uses  a funnel  inlet  and a small diameter
transport tube  nearly identical to the British Smoke Shade  sampler.  Although the .two samplers
operate at different flowrates,  the  particles reaching the filter  tape  could be expected to
have a size range similar to that illustrated  in  Figure 3-18.
     The utility of the sampler to estimate mass concentrations has been  investigated by many
researchers, usually  in comparison with the TSP  hi-vol  sampler.   Since  these two samplers do
not  collect similar particle  size  ranges,  such comparisons could be  expectedvto vary unless
only  a small  proportion of coarse  particles  are present.   Regan  et  al. (1979,)  as  well  as
others  have shown with  field  data  that the correlation  improves  substantially  when the tape
                                             3-57

-------
sampler data are  compared  with smaller particle fractions such as the 0-2.5 urn fine fraction.
As noted  in  the  discussion of the smoke shade sampler, the accuracy of a relationship between
AISI readings and mass concentration for a given data set is difficult to predict.
3.3.3.1.3  Beta-ray attenuation.  Beta-ray attenuation is another technique for estimating the
mass of particles collected on a filter, without direct weighing.  An exposed filter is placed
between a low-energy beta-ray source, such as   Ni,   C, or    Pm, and a beta detector is used
to measure the amount of attenuation caused by the particle as compared to a clean filter.   A
set  of  gravimetrically prepared  standards  are used  to relate the results to  units  of mass.
This  method  is  useful because  it  can be  automated  to  handle  a large  number  of  samples
(Goulding et al., 1978;  Loo et  al.,  1978).   Real  time mass measurements  are also feasible
(Macias and Husar, 1976).
     Investigators (Macias and Husar, 1976; Goulding et al., 1978) have studied the dependence
of the  beta-ray  absorption  coefficient on  elemental  composition of the  sample.   Goulding et
al.  (1978)   found the  dependence on  composition to  be reasonably small  for the  ranges  of
average compositions that occur in aerosol samples.  In a recent inter!aboratory comparison of
aerosol sampling  and measurement methods (Camp et al., 1978), it was demonstrated that labora-
tory beta gauge measurements of ambient aerosols collected at one site by dichotomous samplers
on Teflon  filters compared  favorably in precision and  accuracy with concurrent gravimetric
analyses.  Principal sources of error in the method are possible changes in the orientation of
the  filter substrate  between the pre-  and  post-sampling measurements  and changes in attenua-
tion  because  of  absorption of  water  from  the  atmosphere by  the  filter  material  or the
collected particulate matter (Lawrence Berkeley Laboratory,  1975).
     Jaklevic  et al.   (1981),  after an extensive evaluation of  the technique,  have concluded
                                                     2
that  the  precision is typically better  than  5 ug/cm  for a variety of sample types, assuming
reasonable  care  in  laboratory' manipulation.   The  precision  and  accuracy  are significantly
better  for   samples of fine particles  only,  since one of the major causes  of deviation from
ideal behavior is the presence of  a significant  number of  larger particles.  The possibility
of  automating the process makes it particularly attractive  for laboratories handling  large
numbers of samples.  Especially considering the latter advantage,  it is not seriously inferior
to  gravimetric techniques.   In  some cases  it may be  superior  in  precision and accuracy to
weighing  very lightly  loaded  samples  by  hand on a microbalance,  since  it  is insensitive to
such  errors  as those caused by loss  of  small fragments  from  the  edges  of the  filters.
3.3.3.1.4 Piezoelectric microbalance.   The piezoelectric microbalance technique collects par-
ticles  on an oscillating quartz  crystal, either  by impaction or  electrostatic precipitation.
The  frequency change  of the crystal  oscillation  is proportional  to the mass  collected and the
rate of change   in frequency  is  proportional  to  the mass concentration (Woods, 1979).  Advan-
tages of the  piezoelectric  detection principle,  as-noted by  Lundgren et al. (1976), include
extreme sensitivity  and real time  response.   The technique can also  be applied in a multiple
stage impactor form,  using  crystals as the collection  plates.   This  approach provides  rapid
determination  of  particle  size distributions.
                                             3-58

-------
     Disadvantages of this  detection  principle include the need  for  frequent cleaning of the
crystal and severe  interference  from  large RH changes and nonlinearity of crystal response to
large particle concentrations.   As  shown by the data of Lungren et al. (1976) in Figure 3-19,
the effect of  humidity  depends on particle type  and is,  therefore, nearly impossible to pre-
dict.   Daley and  Lundgren  (1974) studied the  potential errors  in piezoelectric detection and
noted that although  not currently used as a routine monitoring method, it can be used sucess-
fully for short-term studies when realistic operating limits are observed.
3.3.3.2   In  Situ Analyzers—Instead  of  collecting  particles  on a  filter  before  analysis,
certain aerosol characteristics can be examined while the particles are still suspended in the
airstream.
3.3.3.2.1   Integrating nephe1ometer.    The  integrating  nephelometer  measures  the  optical
scatter  caused  by   suspended  particles  in  the  airstream.    The  initial  designs  for  this
technique were  made by Beuttell  and  Brewer (1949) and subsequently  improved by Ahlquist and
Charlson  (1967).   The  nephelometer is  calibrated using the  known  scattering coefficients of
gases such as Freon 12 (CCl-F,).   Light scattering is at a maximum for particles in the 0.3 to
0.8 urn size  range as shown in Figure 3-20.   Accumulation mode particles are the primary catise
of light  scattering,  which is only slightly affected by particles in the nucleation or coarse
particle  modes  (Waggoner,  1973).   Visual  range  can also be calculated from  the  scattering
coefficient using a relationship developed by Koschmieder (1924).
     Correlation  of  optical  scattering  with hi-vol  suspended particle  mass would be possible
only  at  sites  where  coarse  particle  mass  concentrations  are  low  or  correlated  with
accumulation mode particle  mass.  From theoretical  models  (Waggoner  et al.,  1973),  there
should be a  high site-independent correlation between  fine particle  mass loading and optical
scattering.   Waggoner and  Weiss  (1980)  and Groblicki et al.  (1980) have reported correlations
above 0.95 between  gravimetric fine particle mass and nephelometric measurement of scattering
extinction.   Their measurements found the same ratio of scatter to fine mass at an urban and a
rural  Colorado  site,   as  well  as  at  a Pacific  coast  site  in Washington.   Nephelometric
scattering extinction appears to be a useful indicator for fine particle mass but will provide
erratic results  when compared to any particle measure  that  contains  coarse particle mass.  A
comprehensive discussion of tha nephelometer as a visibility monitor is contained in Chapter 9.
3.3.3.2.2  Condensationnuclei counter.   The  condensation  nuclei counter  measures  the total
light  scatter  of submicron particles whose size has been increased  by condensing  vapor onto
their surface  in a  cloud chamber.  This  device  is of interest in determining the number con-
centration of  particles in the nuclei  mode but  is not normally  used for particle sizes above
about  0.5 urn  (Perera  and Ahmed,  1978).   It is  often used  in  conjunction with prior size
separation stages to obtain  a particle size distribution for  submicron  size particles.   The
condensation nuclei counter is rarely used for routine monitoring.
                                             3-59

-------
                                                                      URANINE
            10
20    30     40    BO     60    70     80


        RELATIVE HUMIDITY, percent
90
 Figure 3-19. Response of a Piezoelectric Microbalance to relative humidity for
various particle types.

Source: Lundgren et al. (1976).
                                           3-60

-------
       10
  .3 3-


 £^"5

 •F  ~
*  E
                \   •  '  i  ' '"I      I   '   '  I ' '"tT
SCATTERING   /      \



            I         »
           /          %

                \    I  l  I | | |yf      I   I   t  II IIII
        10"2    2       5    10"1    2
                                                       2    4.00
                               DiAMiTER,  pm




Figure 3-20  Light scattering and absorption expressed per unit volume of aerosol.




Source: Charlson et al. (1978).
                                   3-61

-------
3.3.3,2.3  Electrical aerosol analyzer  (EAA).   The electrical aerosol  analyzer,  as described
by Whitby and  Clark (1966),  measures the electrical mobility of particles as related to their
size.  This device  provides  a detailed number size distribution over the range of 0.01 to 0.5
urn, approximately (Liu  and Pui,  1975).   The analyzer must be empirically calibrated to obtain
the relationship to  aerodynamically sized particles.   The size range sampled does not include
the  entire'0  to 2.5 (jm  fine  fraction  and, comparision between the  measurements  must include
extrapolations.
3.3.3.2.4  Diffusion battery.  The  diffusion battery, as described fay Sinclair et al. (1979),
is a  set of  parallel  tubes or plates  through-which  the airstream  flows  to produce selected
differential  particle  removal  as  a-function  of  parti die size-by diffusion  to the walls.   A
condensation nuclei  counter  is  us'ed as the particle counter.  This diffusion separation prin-
ciple is useful in the 0.01 to. 0.3 (jm .range.
3.3.3.2.5   Optical  particle counters.   Optical  particle  counters  direct  the  -flow  stream
through  a  small nozzle  into a narrow  collimated  light  beam so that  the  light  scatter from
single  particles  can  be..measured.   The  signal  produced  by this  scatter  is  mathematically
related  to the geometric size of a spherical  particle of a  specified  refractive index which
scatters  an   equal   amount of  light.   These  devices  are   sensitive  to  geometric  particle
diameters from  about 0.5 to 10 pm  (Whitby and Willeke, 1979).  Calibration with monodispersed
particles  is  required.   For  sampling  of  particles  larger  than  10 (jm,  modification  of
commercially available  devices  is  required.   Mass concentrations for specific size ranges can
be estimated  by selecting  an  appropriate particle density.  These  devices can  be  used for
routine  operations,  but their usefulness in estimating mass  concentrations is limited by the
accuracy and  consistency  of  the  selected average  particle  density  and  index  of refraction.
3.3.3.2.6  Long path optical measurement. >Long-path (typically >1 km) optical measurement de-
vices  for ambient   air  are  available.   They examine  one  of  several  possible  aspects  of
visibility over a defined distance.  Transmissometers measure the  attenuation  of transmitted
light resulting from scattering  and absorption in  the  atmosphere.   These devices are similar
to their in-stack counterparts,  requiring either a light source and receptor or  light source,
retro-reflector, and receptor  at   separate  locations.   Telephotometers measure  the contrast
caused by brightness differences between a distant object and its surroundings.   These devices
appear to have  promise  as visibility monitors  (see Chapter 9),  but estimates of ambient mass
concentration have not been made from their data.
3.3.4  Particle Composition
     Particles collected  from  ambient air contain a wide range of metallic elements and inor-
ganic and organic compounds.  Their  identification and determination usually  involves the col-
lection  of the particles  upon a substrate (e.g., glass fiber  filters in a high-volume sampler)
with subsequent chemical  analysis   in a  laboratory.   Most methods  for analyzing the  inorganic
fraction  of  particulate  matter  have  focused on  elemental and  ionic  composition.   Atomic
absorption spectrometry has been  the technique  most used  for  the determination of metallic

                                            3-62

-------
elements, although multielement analytical tools, such as optical emission spectroscopy, X-ray
fluorescence spectroscopy, and  neutron  activation analysis, have been successfully applied to
                                                                                 +    -?
the analysis of  elements.   For the most  part,  inorganic io-nic species (e.g., NH., SO. ,  NO-,
etc.) have been  analyzed  with wet-chemical spectrophotometric techniques.   The  organic frac-
tion  of  PM  contatns aliphatic  and  aroma-tic  hydrocarbons, acids,  bases,  and  other  organic
compounds,  such   as  those  containing  nitrogen.   Methods  for  analyzing organics  generally
involve solvent  extraction,  some  form of chromatographic separation  and  detection based on a
selected physical or chemical property of the specific compound.
     Due to the complex chemical composition of the atmospheric particles and the wide variety
of compounds likely  to  be present, it is not practical to review all the possible methods for
their analysis and characterization.  Only methods pertinent to the primary objectives of this
document are reviewed here in detail.   Primary emphasis is placed on methodology for measuring
particulate  sulfur  compounds with  lesser emphasis on  metallic  elements and other inorganic
ionic  species.   Detailed  information  concerning  the  analysis   of airborne  particles  is
contained in a recent monograph edited by H. Malissa (1978), among others.
3.3.4.1  Analysis of Sul fates—Analytical techniques for determining trace amounts of sulfates
in clean,  uncomplicated solution  matrices are numerous  (Forrest and Newman, 1973).  Howev,er,
application  of these techniques  to complex atmospheric particles  is  not so straightforward,
Quantitative transfer  from the collection medium and homogeneous  dispersion  in the analysis
medium—without   contamination,    chemical    alteration,    or   cotransfer   of   analytical
interferents—is required.
     A detailed  critical  review  of the state of analytical methodologies  for aerosol sulfur
compounds  has  been  compiled  by Tanner  et al.  (1978}.   Tanner's  review  includes  methods for
total aerosol  sulfur,  for total water-soluble sulfates, and  for quantitative differentiation
of  aerosol  sulfur compounds  of various  oxidation  states,  as well as a  definitive review of
methods  for  speciation of aerosol  sulfate.  Much of  the discussion  in this  section is taken
from Tanner's  review, with emphasis on the more widely used methodology.   Where information is
available, a critical assessment of each method's capabilities is provided.
3.3.4.1.1  Total water-soluble  sulfates.   A comprehensive review  of  wet  chemical  methods has
been  compiled  by Hoffer and Kothny (1974),  providing  background information on the principal
methods  for  determinating trace sulfates  in aqueous extracts  of particulate matter collected
on  filters.   Sulfate measurements  made with these methods, particularly when applied to the
analysis  of samples  collected with  alkaline filter  media,  are  vulnerable  to error  due to
"artifact  sulfate"  formation caused by the absorption and subsequent oxidation of ambient SOy
in  the  presence  of  the basic components  of the filter media.  With the use  of common glass
fiber filters  under  normal  high-volume sampling  conditions, this  error  has been estimated to
range from 0.3 to >3 ug/m   (0.1  to > 1 ppb), depending on the  ambient SO, levels at the time
of  sampling  (Coutant,  1977;  Pierson et al., 1980).  This potential error should be considered
when  assessing  data collected  using  any of the  methods for  water-soluble  sulfates discussed
below.
                                             3-63

-------
     Currently,  the  methods most  widely:employed for  soluble  sulfate determinations are the'
BaSO.  turbidimetric,  methylthymol  blue, thorin,  and  ion exchange chromatographic procedures'.
These  methods  have  inherent analytical* precisions,1 accuracies, detection limits, linear work-
ing ranges, and other operational characteristics'which will be discus'sed in the following sec1*'
tions.  Each method  has an analytical 'lower  detection  limit (LDL) and working range for sul-
fate in the aqueous extract expressed in'jJg/ml.  The'corresponding detection limit and working '
                                                                      3
ranQe  of  each  method for measuring sulfate  in ait4 expressed in  ug/m  depend  upon  the analy-
tical  LDL and the sample size (volume of air sampled, fraction of filter taken for extraction,
etc.)-  For example, if an analytical technique'having a working  range of 1 to 10 yg SQ./ml is
                                                                                     •    3
used  to  measure sulfate  from a typical hi-vol'-sample  (20.3 cm x 25.4  cm  filter,  1.4 m /min
flow  rate,  24-hr sample,  1.9  cm x 20.3s cm  strip extracted in 50 ml),  the  working 'range for
                                             Q                     "
sulfate in  air would be D.3"to '2.9>ug SQ./rh .'  l£ should'be "noted that the upper range limit
can usually be "extended to higher concentrations"by  dilution of the aqueous extract prior to
analysis,                             •
     3.3.4.1.1.1  BaSO*  turbidiroetry.'   Sulfate in the aqueous extract from a particulate sam-
ple* is precipitated by addition of barium chloride.  The resulting"BaSO. turbidity is "measured
spectrophotometrically or'hephelometrically  and compared to a standard curve prepared by mea-
suring the  absorbance  of standard solutions of sulfate.   Numerous versions and modifications
of methods  based  on  this principle appear in the 1'iterature (Kol'thoff'et a!., 1969; Techm'cori
Corp., 1959;  American Pubfic Health Association,  1971; Appel et a!.,  1979a).   Appel,  et al.
                                                                  *             »  »
(1979a) described and  evaluated  a  procedure  applicable  to  the measurement  of  sulfate in
aqueous extracts from 24-hour hi-vol particulate samples.  They reported an analytical working
                                                   3
range from 10 to 70 ug SO^/ml (2.9 to 20.3 ug S04/m  for a typical hi-vol sample), an accuracy
within 4  percent, and  an average* precision  of 3.8 percent '(coefficent  of  variation)  of the
working range.  They  reported that extract background  turbidity  and color interfere with the
procedure but  are minimized by means of blanks.  Sulfur compounds converted to sulfate by air
oxidation also  interfere.   The apparatus required for  turbidimetric  sulfate determination is
relatively  inexpensive  and if proper care is taken, the procedure is  capable of producing re-
liable data.
     3.3.4.1.1.2  Methylthymol  blue  (MTB).   A  reagent  containing equimolar  amounts of barium
ions and MTB,  at a pH of 2.8, is added to the aqueous extract from a  particulate sample.  Sul-
fate in the solution is precipitated as BaSO. and the'pH of the solution -is raised to 12.4 by
addition  of NaOH.  The remaining barium combines with the anionic MTB and leaves an amount of
free MTB  equivalent  to the sulfate.   The MTB is measured spectrophotometrically at 460 nm and
compared  to a  standard  curve of  absorbance  versus concentration.   Lazrus  et  al.  (1966) de-
scribed an  automated  version of this method  and  reported that the reagent is oxidized in air
when made alkaline,  thus limiting the use of the method to a closed system.
     An evaluation by  Appel  et al.  (1979a) of  automated MTB methods examined two procedures;
one covering an analytical range of" 0 to 100 ug SO./ml developed by Midwest Research Institute
(HRI)  (Bergman  and  Sharp,  1979) for EPA and  the other'covering  a range  of 0 to 10 ug S04/ml
                                             3-64

-------
developed by  Colovos et  al.  (1976).   The  results  for the MRI procedure  indicated  a working
range of 17 to  90 M9 S0,/ml  (5.0  to  26.5 ug SQ./m  for a typical hi-vol sample), an accuracy
of 1.06  (mean observed/  theoretical)  with an average  coefficient of variation of 2.8 percent
for analyses  of filter strips spiked  with  known amounts of sulfate.   Similarly,  the results
for the  Colovos  et al.,  (1976) procedure indicated  a  working range of 2 to 10 ug S0«/ml (0.6
to 2.9  g SO^/m   for a typical  hi-vol sample),  an  accuracy  of  0.98 and  a  precision  of 1.3
percent.  The samples  must be treated with an  ion  exchange resin to  remove  metal  ions which
may also react  with MTB.   No  significant sources of interference were  found in  this evalua-
tion.   The automated versions of the MTB procedure are widely used and are capable of produc-
ing reliable results.  Large sample loads can be analyzed in relatively short periods of time.
The equipment is relatively expensive, but can be used for other analyses.
     3.3.4.1.1.3  Thorin.   Titrimetric methods for sulfate using barium ion and thorin indica-
tor for  visual  or  spectrophotometric  detection  of  the endpoint  are  popular (Akiyama,'-1957;
American Society  for Testing and  Materials,  1981c;  Bakas,  1956; Oubois et al.,  1969;  Fritz,
1955 and 1957;  Menis,  1958; Rayner, 1966).   These procedures provide for titration of aqueous
sulfate with a solution of barium ion to precipitate barium sulfate (BaSO.).  When the sulfate
is completely reacted, excess barium complexes with thorin to produce a pink color indicating
the endpoint  of.the titration.   The samples  must be treated with a  cation exchange resin to
remove metal ions that also complex with thorin.
     A recent modification of this technique by Brosset and Perm (1978) allows rapid determi-
nation  of  sulfate  by  employing an  automatic pipetting system.   Aqueous   sulfate extract is
treated  with  a  solution  containing an amount of barium in excess of  the  anticipated sulfate
and BaSO. is precipitated.  Then, a solution of thorin indicator is added,  which combines with
the remaining barium to form a colored complex.  The absorbance of the solution is measured at
520 nm  and  compared to a  standard curve  obtained from sulfate  standards.   The absorbance of
the solution  is  inversely proportional to the sulfate concentration.  This procedure has been
evaluated by  Appel  et al.  (1977) who reported an effective working range of  3 to 13 ug SO./ml
                     3
(0.8 to  3.8 ug SO./m  for  a typical hi-vol sample), an accuracy of 1.04 (mean observed/theore-
tical),  and a precision  of 5 to 9 percent  (coefficient of variation).  No significant source
of  interference  is  reported, but  the  samples must be corrected  for background turbidity and
color.  The Brosset  modification employs an relatively expensive automatic pipet.
     3.3.4.1.1.4   Ion  exchange  chromatography.   The principle of the  ion exchange  chromato-
graphic  technique is  described  briefly under  Section 3.2.2.3.6.   Stevens  et al.  (1978) de-
scribed  the use of this technique  for analysis of sulfate as well as other  ions.  Appel et al.
(1979a)  evaluated a  procedure for  sulfate analysis using a system manufactured by Dionex Corp.
(1975).  This procedure  showed  a  working  range of 7 to  130 ug  SO./ml,  an accuracy of 1.08
(mean  observed/theoretical),  and  a  precision of  6.2  percent  (coefficient of variation).  A
small  interference  from  nitrate ion was also reported.  Apparatus for this procedure is rela-
tively  expensive  and requires a skilled  operator.   Nevertheless,  the procedure is considered
to be reliable and  specific.  Other ionic species can  be determined simultaneously.
                                              3*65

-------
3,3.4.1,2  Total aerosol sulfur.   Nearly  100 percent of aerosol sulfur mass is present in the
form of  sulfate  (Forrest and Newman, 1973a).  This was experimentally demonstrated by Stevens
et al.  (1978), who also showed that most of the data on air-borne sulfur concentrations can be
accurately described  as total  sulfur  calculated as sulfate or  total  soluble  sulfate.   X-ray
fluorescence  is  the primary  and  most  practical technique for measuring  total  aerosol  sulfur
collected on  filters.   This  technique  can  be used to  analyze  many  elements  besides sulfur.
It  is  nondestructive  and  can  be automated to  facilitate the  analysis  of large  numbers  of
ambient  aerosol  samples.  A  particulate  sample, collected on an  appropriate  filter (usually
      ©
Teflon )  is  irradiated  with  photons (X-rays,  gamma rays, etc.),  protons, or  other charged
particles, and  the intensity of  the fluorescent X-rays induced is measured as  a function of
wavelength or energy  to determine the  amounts  of the  constituent elements present.  Qualita-
tive and  quantitative  analysis  can be  obtained  when the  system is properly calibrated.   This
calibration step is difficult,  since few standards of known elemental composition are availa-
ble  in disks  of  known thickness  in   an appropriate  matrix (Adams  and Van  Grieken,  1975),
However,  recent  work  by Dzubay  et  al. (1977)  has shown that  calibration standards can  be
prepared to an accuracy of ± 5 percent.
     The most extensive set  of aerosol  sulfur data were reported by Stevens et al.  (1978) and
Loo  et al.  (1978)  using a energy nondispersive X-ray fluorescence  spectrometer designed by
Goulding  and  Jaklevic  (1973).   Stevens et al.  (1978)  reported sulfur and 18 other elements
from dichotomous   samplers  operated in New York City, New York;  Philadelphia,  Pennsylvania;
Charleston, West Virginia;  St.  Louis,   Missouri;  Portland, Oregon; and  Glendora, California.
Loo  et al.  (1978)  reported  sulfur  concentrations determined from  samples collected over  a
2-year period from a  network of  10  automated  dichotomous  samplers operated in and around St.
Louis,  Missouri,  during the  Regional Air  Pollution Study (RAPS),  They  reported a detection
                     2.                                                                       3
limit  of 0.034 \ig/cm   of  filter, which  corresponds to a concentration value  of  <0.1  ug/m
sulfur for a  2-hour sample collected at 50  liters/minute  on  a 37-iran filter and is adequately
sensitive for a 1-hour time discrimination at ambient sulfur levels.
     Proton-induced  X-ray  emission   spectroscopy,   which   has  the   advantages   of  lower
bremsstrahlung background and focusing properties of the exitation beam,  is a useful tool when
short-time resolution  of ambient sulfur  levels is desired (Johansson et al.,  1975;  Courtney
et al., 1978).  However, it  takes substantially more  energy  to  produce  an X-ray with charged
particles than with photons, and  in some cases vaporization or decomposition of the sample may
occur  (Shaw and Willis, 1978).   A related approach to nondestructive aerosol  sulfur analysis
based  on  cyclotron in-beam gamma-ray spectroscopy  has  been reported  by  Macias (1977).   Gamma
rays induced by proton or  a-irradiation are  detected  by  a Li-drifted Ge detector and used to
determine S and other light elements such as  Mg and C in aerosol samples.   This technique is
less sensitive for S than X-ray emission methods.   In spite of the aforementioned limitations,
                                            3-66

-------
it is clear that these induced X-ray and gamma-ray methods will continue to be important tools
for determining total sulfur in large numbers of ambient aerosol samples.
     Other techniques that  have  been applied to the  determination  of total sulfur in aerosol
particles include:  (1) electron  spectroscopy for chemical analysis (ESCA) (Novakov, 1973; and
Novakov et al.,  1974);  (2)  various applications of  flame photometric detectors (FPD) (Crider
etal.,  1969;  Kittelson  etal., 1978;  Huntzicker eta!., 1976,  1977, 1978;  Tanner etal.,
1978,  1980);  and,  (3)  an isotope tracer  technique  using     Ag  tracer  (Forrest  and Newman,
1977).   ESCA  is  sensitive  to  surface  composition   of  samples,  which is  an advantage  for
surface-oriented studies  but  not  for ambient aerosol samples  whose  elemental  composition is
likely to  be heterogeneous.   A  comparison  of  ESCA  to  wet chemical sulfate  measurements by
Appel et al.  (1976) showed agreement only within a factor of two.   Direct flame photometry has
potential as  a sensitive total  aerosol sulfur  analyzer, but-its  application  is  complicated
because S0~ must be  removed and the FPD response varies with the chemical form of the aerosol
sulfate.   Recent work by Huntzicke'r et al.  (1978)  and  Tanner et al.  (1980)  has  shown that
direct flame  photometry can  not only  provide  a  sensitive  total  aerosol  analysis  but, when
combined with  thermal  volatilization, can  provide semicontinuous measures  of H^SO,, ammonium
sulfate,  and metal  sulfates.
3.3.4.1.3  Sulfuric aciddetermination.  Most of the efforts to determine the species composi-
tion  of  sulfate  in airborne particles have concentrated on development of a specific analyti-
cal method for H?SO. in air.  Despite the substantial efforts of several groups, the existence
of free  aerosol  H2SO, in the  ambient atmosphere  has been unequivocally established in only a
few  cases.   Interference problems  and difficulties  in  sample preservation  have  contributed
markedly to  the lack of  valid HpSCL measurements.  The  procedures  discussed  below have been
applied primarily by research analysts, are vulnerable to error both in sampling and analysis,
and are not generally applicable to routine monitoring.
      Procedures for determining  H2SO. and other sulfate species include thermal volatilization
and  solvent  extraction techniques,  gas phase  ammonia (NH-)  titration,  infrared  and visible
spectrometry,  flame  photometry,  and  electron microscopy.  The determination  of H?SO.  by its
selective  thermal   volatilization  from  filters  has   been   reported  by  several  workers
(Scaringelli  and Rehme,  1969;  Dubois et al.,  1969a;  Maddalone et  al., 1975; Thomas  et al,
1976;  Leahy  et al.,  1975).   This technique generally suffers  from poor H^SO. recoveries, poor
reproducibility, and  interferences from ammonium  sulfate salts.  The most  successful approach
to thermal volatilization of H?S(L. in  ambient  aerosol  samples was reported by Mudgett et al.
 ,                                                    si
(1974).  Aerosol samples  are collected on Fluoropore  filters, the HgSO. subsequently volati-
lized by passage of heated  (~150°C),  dry  N? in the reverse direction  through  the filter and
released HpSO.  determined with a  flame  photometric  detector.   Lamothe and Stevens  (1976) re-
ported that  laboratory aerosol  samples of as little as 0.25  pg H?SO»  may  be determined with
reasonable  precision.   However,  serious  difficulties   were  encountered  in  removing HgSO^
quantitatively  in  the presence  of ammonium  bisulfate (NH.HSO,).   They observed that HnSQ, is

                                             3-67

-------
totally removed  at 180°C,  but NH4HS04 is also partially volatilized; at 140°C» NH4HS04 is not
volatilized but H^SO. is incompletely volatilized.
     A solvent extraction  procedure to remove collected hUSO. aerosol selectively in the pre-
sence  of  other aero,sol  sulfates was  first reported by Barton  and  McAdie (1971).   They con-
cluded that  aerosol  collectfon on Nuclepore  filters,  followed  by extraction with 2-propanol
for subsequent analysis by the chloranilate procedure, was selective for airborne H^SCK.  Sub-
sequent work  by  Barton and McAdie (1973) reported reduction of  interference by buffer control
of  the 2-propanol extract  and also  reported development of  an automated instrument for the
extraction  procedure.   Leahy  et al.  (1975)  reported,  however, that  2-propanol will  also
extract NH^HSO.  quantitatively and partially extract other  bisulfates  and that it should not
be  considered a  selective  extractant for  H?SO..  They  demonstrated  that benzaldehyde  is  a
selective extractant  for  H,SO. in the presence of bisulfates  and sulfates.  Subsequent radio-
                                                     35
chemical  experiments  (Tanner  et  al.,  1977) with  H9  SO.  have  established that  H9SO.  may be
                                                           •     ft            ®
reproducibly  removed  from  a variety  of  filter media  (Mitex  ,  Fluoropore,  H,P(K-treated
quartz) with  recoveries  varying from 75 percent  for 10 ug H~SO,  samples to 95 percent for LOO
pg  samples.   The  selectivity of the benzaldehyde extraction technique has been confirmed in a
study  by  Barrett  et al.  (1977) employing laboratory  aerosol  H?SO*  samples as low as  5 ug in
the presence of bisulfate and sulfate.
     As discussed briefly in  the previous  section,  systems  for measuring  ambient  levels of
H,SO. ahd other sulfate aerosols by flame photometry have been developed recently by Huntzicker
et  al.  (1978)  and Tanner et  al.  (1980).   A heated denuder for  SO-  removal  allows the direct
measurement \)f total sulfate aerosol and selective thermal volatilization allows the discrimi-
nation of semicontinuous measurements  of H^SO.  ammonium sulfates  and  metal  sulfates.  Con-
version 6ff  the H2SO. or the  aerosol  sulfates to ammonium sulfate  [(NH,)2SO,]  by addition of
NH-  v'i initiates the problem of FPD  response variations  with the  chemical  form of the aerosol
sulfate.
     Since -atmospheric sulfate »can be associated with  various cations,  the compounds of sul-
fate can ^sometimes  be inferred by measuring  the  cation.   If  the ions in  a  series of samples
are  measured  and Ithe  ammonium (NH.)  content is highly correlated  with the  sulfate content,
                  *                        4*
then  it  can be  inferred  that vari6us NH-  salts  of H2$0- are probably  present.   Brosset and
Perm  (1978)  and  Stevens  et al.  (1978)  described  a Gran titration procedure  for hydrogen ion
(H  )  and  a procedure  using  ion  selective  electrode for NH,  in aqueous extracts of aerosols
                    i".                                       T"
collected  on Teflon   filters.   Stevens  et  al.  (1978)  applied such techniques  to  aerosols
collected at  Research Triangle Park, North  Carolina,  during the summer of 1977  and  1978 and
                                                        •f-4-                  4-       4-
found  a stoichiometric balance between sulfate ion  (S04  )  and the  sum of H  and NH4 ion con-
centrations.  The acidity was found to range from none  [(NhL^SQ.J to that of NhLHSQ,.
     Dzubay (1979)  developed  and used a sensitive radiolabeling  technique for measurements of
acid  sulfate  aerosols.   Liu  et al.  (1978b)  described  a new  technique  that  uses  an aerosol
                                             3-68

-------
mobility chromatograph 'for  the  detection of sulfuric acid aerosols.  Several semiquantitative
methods  for  estimating  sulfate  species have  been  investigated.   These  include gas  phase
ammonia titration  techniques  (Dzubay et al. , 1974),  humidographic  techniques  (Covert et al.,
1980), methods based  on  infrared spectroscopy (Blanco et  al.,  1968, 1972; Cunningham et al.,
1974; Cunningham and  Johnson,  1976), and microscopy techniques (Heard and Wiffen, 1969; Lodge
et al., 1960; Frank and Lodge, 1967; Mamane and de Pena, 1978).
3.3.4.1.4  Filter sampling problems  related to sulfate analysis.  Ideally,  the measurement of
aerosol sulfate species  in  PM requires that the  sulfate-containing particles  from the air be
quantitatively collected on a filter surface that does not permit chemical or physical trans-
formations and that does not  lead to  spurious  sulfate particle formation from SO, present in
the gas stream.  The  particles  must then be transferred to the analysis medium under the same
constraints.   Sampling  for airborne  sulfate  is especially  difficult, since  acidic sulfate
species react with many  common  filter materials (e.g., "neutral" glass fiber filters and many
plastic filters—Nuclepore  , Acropore  ,  Millipore )  as well as basic particles in the sample-
The result  is  neutralization  of the -acid sulfate  and alteration of the composition from that
extant in ambient air.  Many of the historical data on sulfate species are questionable due to
insufficient consideration of the above sampling difficulties.
                                             id
     Several  filter materials made of Teflon , have  been  found to  be  inert and  suitable for
nonreactive  collection of aerosols,  including acid sulfates.   The most widely used are backed
Teflon   membranes, Fluoropore,  and  Mitex .   A modified quartz   filter  material   has  been
developed (Tanner  et  al., 1977) from which impurities are removed by preheating to 750°C, and
reactive basic sites  are removed by treatment with  hot,  concentrated phosphoric acid.  After
rinsing and  drying, the  quartz filters may  be  used  for high-volume, high-efficiency particle
collection  without interfering  with  acid  determinations  of the  collected particles  at the
fractional microequivalent level.
     Two  additional   problems  have  been  identified   in filter  sampling for airborne sulfate
analysis.   Sulfur dioxide may be converted to sulfate by adsorption  on and catalytic oxidation
by the filter material (Lee and Wagman, 1966) and/or  by previously collected particles (Coffer
et al., 1974).   Recent  studies  by  Forrest  and  Newman (1973a)  seem to indicate  that active
catalytic sites on the filter material are the likely culprits.  Experiments by Tanner et al.
(1978)  with high- and low-level S0?-spiked  ambient air  passed through preloaded  and clean
hLPQ.-treated  quartz  filters  at  high  and  low  linear  flow  velocities failed  to  find any
evidence of  artifact  sulfate  formation for  this  filter material.   This work was confirmed b>
the  work  of  Pierson  and coworkers  (1976)  from whose  data it  is  clear that  the low sodium
content of  the Pallflex GAO  quartz  is  the probable reason  for the  negligiblv  low artifact
sulfate formation.
     A second problem  results from potential neutralization of acidic sulfate particles by NH,
in the  gas  stream  traversing the filter.  Neutralization by NFL and oxidation of SO^ may both
                                             3-69

-------
be  reduced  by "diffusion-denuding"  the stream  of  these gases  (without removing significant
particles), but this is a cumbersome process, especially for high-volume sampling.
     The most  effective  way to -reduce  particle  formation  and  transformation reactions on the
filter is to  collect the smallest airborne  particle  sample that, is compatible with available
analytical  methods.   Minimum-quantity  sampling  also  reduces collision-induced  interaction of
particles on  the  filter  surface and thus real-life chemical inhomogeneities in ambient parti-
cles are  more likely  to be  unaffected by  the  sampling process.   The  only way to eliminate
confounding interparticle  interactions  on filters totally is to determine the sulfate in situ
without filter collection.
3.3.4.2   Ammonium and Gaseous Ammonia  Determination—Gaseous  NhU   and   ammonium  ion  (NH.)
measurements  are  important  in  understanding  speciation  of sulfate  in airborne  particles.
Ambient ammonia is  by far the most important neutralizing agent for acid sulfate; its concen-
tration is  directly  related to the chemical form of sulfate in the ambient air.   Measurements
of  ammonia  along  with the set of species which, by mutual interaction, determine the chemical
form of sulfate, are crucial to understanding such pervasive problems as acid rain, visibility
degradation and  the effects  of dry  deposition.   Ammonium ion is  found predominantly in the
optical-scattering size range or below  and is presumed to be secondary in origin, being formed
in  the neutralization  of acidic sulfate particles.   The high  correlation of NH, content with
soluble sulfate,  in  both urban (Tanner et a!., 1977a) and rural  (Tanner et a!.,  1977) aerosol
samples,  and the identification by X-ray diffraction of  (NH4)2S£L in dried aqueous extracts of
airborne particles would tend to confirm the above hypothesis (Brosset et al.,  1975).
     Ammonium  ion in particulate matter is almost always determined by collection on filters,
extraction  into an  appropriate leach solution,  and determination  by one of two methods.   The
first is  a concentration  measurement  by an ion-selective electrode  sensitive  to  either NH*
(Beckman  electrode)  or  NH~  (Orion or- Markson  electrodes).   The  limit of  detection  is
determined  by  the equilibration  time of the  electrode, a representative value being  5  to 7
minutes  for 20 ppb  NH,  concentration  in water  (Eagan  and DuBois,  1974;  Gilbert and Clay,
                       ^                                                                     4.
1973).   This  is marginally sensitive for  high-volume samples  of rural  ambient  air where NH,
                          3
may be as  low as  0.3 ug/m  .   A  later development,  the air gap electrode (Ruzicka and Hansen,
1974; and Ruzicka et al., 1974), eliminates the problems of electrode contamination by sensing
of  the NH_-water  equilibrium across  an air gap between the analyte solution and the electrode
surface.
     The second  commonly used  method  for NH.  traces in  aqueous  solution is  the indophenol
colorimetric method  based on  the  color-producing reaction of phenol  and hypochlorite in the
presence of NHg.  Modifications  most analytically useful for determination of  NH, in aqueous
leaches were  reported by Bolleter et al.  (1961) and  by Tetlow and  Wilson  (1964).   Automated
procedures  have  been proposed  by  Lazrus et al.  (1968)  and by  Keay  and Menage (1970).   The
latter method is capable of a lower detection limit of 0.05 (jg/ml (as nitrogen),  requires  only
a few minutes of analysis time, and has a minimum sample volume of 2 ml.

                                            3-70

-------
     Methods for determination  of free atmospheric NH~ can be divided into direct methods and
methods in which NH,  is first immobilized on acid-treated filters, leached, and determined as
  +
NH* by one of the methods described above.  Ammonia has been analyzed directly by quantitative
conversion to nitric  oxide  (NO) over a hot catalyst and determination by chemiluminescence of
the  NO (Hodgeson  et al.s   1971.  and Baumgardner  et  a!., 1979).   This method  is  marginally
                                                                          3
sensitive enough for  ambient levels of NH, (limit of detection = ~ 1 )jg/m ) and must be care-
fully  zeroed  in the  NhL scrubber  mode  to eliminate interference  from  atmospheric oxides of
nitrogen (NO ).  Filter  pack methods using either KHSOA (Eggleton and Atkins, 1972) or oxalic
            />                                          T*                                     »
acid impregnants have  been  used to collect ambient  levels  of NH~, but  they  are fraught with
blank  and contamination  problems  and may  not collect  ambient levels of 0.5 to 5 ppb NH-, with
reproducible  efficiency  under  commonly  observed  conditions  of  temperature  and  relative
humidity.   Shendrikar and Lodge (1975) applied the ring oven technique to the determination of
ammonia collected  from  ambient air on filter  tapes  impregnated with  ethanolic oxalic acid.
The range for quantitative determination is reported to be 0.1 to 1.00 pg of ammonia.
     It has also been proposed that gaseous NH3 could be determined at or below ambient levels
by  gas phase   reaction  with  HC1  vapor.   The  resultant ammonium, chloride  (NH.C1)  aerosol
particles would be  measured by a condensation nuclei  counter (CMC).  Unfortunately, there are
several difficulties that severely limit the usefulness of this technique.  The concentration
of  HC1 and  the relative  humidity  must   be  carefully controlled  to  attain proportionality
between number of particles and NH-, concentration.  In addition, it is necessary to provide an
ionization source (a corona discharge or a UV light source) in the airstream just prior to HC1
vapor  addition  in  order to  approximate  precise, proportional  CNC response.   However,  this
method  has  potential   for extremely high  sensitivity  and real  time  operation.   McClenny and
Bennett (1980)  developed a  semi-real time detection technique  for ambient NH, based on inte-
                             8
grative collection  on Teflon   beads,  followed  by thermal desorption and  detection by either
chemiluminescence or photoacoustics.   Perm (1979) and Braman and Shelley (1980) reported col-
lection of NH~  on  diffusion tubes.  Perm  used  oxalic  acid as a  coating which is rinsed from
                                                           +
the  tube  at the end  of a 24-hour  run and analyzed  for NH. by  ion selective electrode tech-
niques.  Braman and Shelley used a  tungsten  oxide coating for 20 minute samples and released
the  NH- into  a  chemiluminescence analyzer by thermal  desorption.  Hoell, et al. (1980) deter-
mined  vertical  concentration profiles  by interpretation  of  infrared solar  spectra obtained
with a heterodyne  radiometer.   Abbas and Tanner (1981) reported work on the continuous deter-
mination  of  gaseous  NH-, using  fluorescence  derivatization.   These  recent  advances  in the
development of new techniques for measuring NH., will  be helpful in determining the role of NH,
in conversion of H?SO. to less harmful materials.
3.3.4.3  AnalysIs o'f Nitrates—Nitrate analyses  have  been performed routinely for many years,
and  a  large  number of chemical methods have been  reported.  In typical monitoring for nitrate
in  air, a  portion  of  a  particulate filter is subjected to aqueous,extraction and  the water-
soluble nitrate is analyzed by one  of the methods  discussed below.

                                             3-71

-------
3.3.4.3.1   Measurement techniquesfor nitrates.   The  oldest  procedures, for analyzing nitrate
used brucine  (Intersociety Committee, 1977b) or phenol  disulfonic  acid (Intersociety Commit-
tee, 1977c).  Many other methods for  analyzing nitrates have been reported, including: (1) the
nitration  of  chromotropic  acid  (West and Ramach.andran, 1966) and  coumarin analogs (Laby and
Morton,  1966;  Skujins, 1964); (2) the quenching of the fluorescence after  nitration of fluor-
escein  (Axelrod  et  al. 1970); (3) reduction  with  Devarda alloy to ammonia (Kieselbach, 1944;
Richardson, 1938);  and (4) the use of  ion-selective  electrodes  (DiMartini, 1970; Driscoll et
al., 1972;  Gordievskii et al.,  1972).   Microscopic  techniques  also  allow identification and
size estimation of individual nitrate particles (Bigg  et al., 1974).
     One of the  most extensively used techniques  to  analyze nitrates in atmospheric particu-
late extracts involves reduction of the nitrate to nitrite by zinc (Chow and Johnstone, 1962),
cadmium  (Morris  and Riley, 1963; Strickland  and Parsons,  1972;  Wood et al., 1967), or hydra-
zine (Mullin  and  Riley,   1955).   Measurement of  the nitrite produced  is accomplished  by  a
sensitive diazotization-coupling reaction  CSaltzman,  1954).   Automated versions of this tech-
nique provide much  better results because critical reduction  parameters  such as temperature,
surface contact area, and  reaction time can be precisely controlled (Technicon, 1973).  Lazrus
et al.   (1968)  used  an automated system,  in  which  nitrate was reduced to  ammonium and deter-
mined  by the  indophenol  method.   Another  extensively used  technique  to  analyze  nitrate in
atmospheric particulate matter  extracts  involves the  nitration  of  xylenols and separation of
the nitro-derivative  by extraction or distillation.   A  comparison  of a 2,4-xylenol procedure
(Intersociety Committee,  1977d)  with the automated copper-cadi urn  reduction and diazotization
method  in  samples collected near high density vehicular traffic,  demonstrated a negative in-
terference in the former up to a factor of 3  (Appel et al. 1977#).   ,
     Small  et al. (1975)  report an application of ion exchange chromatography to the measure-
ment of a  wide  variety of  cations  and anions  including the nitrate and  nitrite ions.   The
novel  feature of  this method is the use of a second ion exchange "suppressor" column (after a
conventional separating  column)  that  effectively  eliminates the ions of  the eluting medium.
Since the  chromatographically separated  species  of interest  leave  the  suppressor column in a
background  of  deionized  water,  concentration  determinations may  be  made  by  a  simple and
sensitive  conductometric   technique.   Mulik  et  al.   (1976)  report the  application of  this
technique  to  measurement   of  water-soluble  nitrate on hi-vol  filters.  The separator column,
containing  a  strong basic  resin,  separates  anions in a background of  carbonate eluant.   The
suppressor column,  containing a  strong acid  resin, converts  the sample ion and the carbonate
eluant  to  nitric and  carbonic  acid,  respectively.  Since carbonic acid  has  low conductivity
and partially  decomposes  to carbon  dioxide and  H-0, the  nitrate  ion alone  is  effectively
measured in  a conductivity detector.   Under the  experimental conditions,  sensitivity  of 0.1
ug/ml   was  reported.   The  relative  standard deviation  was  1 percent  (95-percent confidence
                                             3-72

-------
level)  for  10  replicate  injections  at  the  5  itg/ml  level.    At this  concentration,  no
interferences  were  found  from fluoride,  chloride,  nitrite,  sulfite,  sulfate,  silicate,  or
carbonate.   Positive  interferences  were found  for  bromide and  phosphate,  but  the authors
suggest techniques for eliminating these.
     In other  work, Glover and Hoffsommer (1974) and Ross  et al, (1975) reported a technique
for  assay of  aqueous  nitrate and  nitrite  extractions by  conversion to  nitrobenzene.   Both
techniques involve the  nitration  of benzene in the  presence of H?SO, to form nitrobenzene, a
relatively stable compound,  followed by gas chromatographic analysis.  Careful calibration is
required  in  both procedures, since  a significant  fraction  of the nitrobenzene  formed  may be
lost to  the  acid layer.   Ross et  al.  (1975)  recommended a  calibration  procedure,  whereby a
standard  is  subjected  to  the same procedures as the unknown, while Glover and Hoffsommer used
internal  calibration with  added  nitrotoluene.   The lower detection limits reported by Ross et
                                  —1 ?
al.  (1975) are in the range of 10    g  nitrobenzene in a 1 |jl sample.  Conversion efficiences
for KN03, KN02, and  HNO_ were reported as 90.3 ± 7.9, 100.4 ± 4.2, and 99.9 ± 5.2 percent, re-
spectively.  Glover  and Hoffsommer report similar recovery rates for KNQ~ and KNO?.
3.3.4.3.2   Filter sampling problems related to nitrate analysis.   Serious difficulties  have
been  reported  to be  associated  with the  routine  analysis  of nitrates  in  PM collected using
glass fiber  filters.   In  a study of  nitrate  in  auto exhaust, Pierson et al.  (1976) reported
that  glass  fiber filters  collected  about  twice  as  much   nitrate  as  quartz  fiber filters.
Nitrate also was  found on glass fiber  filters  that were inserted downstream of either quartz
or  glass   fiber  primary  filters,  providing  additional evidence  of artifact  formation  from
gaseous  constituents.   Spicer  (1976)  reported  that  glass  fiber filters  completely removed
gaseous nitric acid (HN03)  in low  concentrations  in  gas  streams,  while  Teflon   and quartz
filters showed no corresponding effect.  O'Brien et al. (1974) described very unusual particle
size distribution determinations for photochemical aerosol collected in the Los Angeles Basin.
This  study used a cascade  impactor,  and all  particle  size  fractions were  collected on glass
fiber  filters.   The authors  speculated that conversion of  gaseous  nitrate precursors  on the
filter masked the true nitrate size distribution.   Qkita et al. (1976) reported that untreated
glass fiber  filters collect nitric acid vapor with a highly variable collection efficiency (0
to  56  percent),  suggesting erratic nitrate artifact  formation in urban atmospheres containing
HN03.
     In  an  intensive  laboratory  investigation  of  interferences in atmospheric particulate
nitrate sampling, Spicer et al, (1978) concluded that all five types of glass filters investi-
gated exhibited  serious  artifact formation due to  collection of  gaseous HN03 and, to a small
extent, N0?  as nitrate.   Cellulose  acetate and  nylon filters were  also reported to collect
                                                                           (8
HNQ.,.  Negligible interferences  were reported for polycarbonate and Teflon  filters.  Collec-
tion of N0?  on quartz fiber filters  varied with the  filter type, with ADL Microquartz showing
the  least effect.   Artifact nitrate  formed on the  Gelman AE filter was calculated to be less
than  2  (jg/m   during  a standard  24-hour hi-vol  measurement.  This  estimate  was derived from

                                             3-73

-------
                                  3                      3
drawing  air  samples of  about 1 m  containing  4512 pg/m  (2.4 ppm) N02  through the filters.
The relative humidity was 30 ± 10 percent.
     Spicer and Schumacher (1977) also reported the results of a comparison of nitrate concen-
trations  in  samples collected on  various filter types  in Upland,  California,  during October
and November 1976.  During the experiment, meteorological conditions varied from warm and hazy
to  hot,  dry,  very clean desert wind  conditions.   Nitrate analyses were  performed  by ion ex-
change chromatography.   All  filter types used had comparable particle collection efficiencies
according to the manufacturers'  specifications.   The ratio of nitrate collected on glass fiber
filters  to  that  collected simultaneously  with identical hi-vol  samplers on  quartz filters
ranged from 2.8 to 49.
     More recently, Appel and coworkers (1979, 1980) conducted several  studies bearing on both
positive  artifact  formation and  loss of nitrate  from a variety of filter media.   They con-
cluded that gaseous HNQ, is the principal source of artifact nitrate formation; NO, collection
only became substantial at high ozone levels.  Ambient particulate nitrate values (at San Jose
and Los  Alamitos,  California) differed by up to  a factor of 2.4 depending upon filter medium
and  sampling  rate,   in  contrast  to  the  much  larger  differences  reported  by  Spicer  and
Schumacher,  1977.  They  also  reported that at low HNO~ levels, nitrate on glass filters indi-
cated  (within  3  percent)  total  nitrate,  (i.e.,  particulate nitrate  plus  HNQ.,  rather than
particulate nitrate  alone).   They  concluded that  the degree of error associated with glass
fiber  filter  media could  be  expected to vary  with location,  time of year, and time of clay,
paralleling changes in HNO-, levels.
     Laboratory studies  by  Harker  et al.  (1977) and ambient  studies  by Pierson et al.  (1980)
have suggested that  reaction  of particulate nitrates  with acidic  particulate sulfates (e.g.,
H?SOfl) can  result in negative errors in the determination of nitrate  on filters.   Formation
and subsequent loss of gaseous HNO, was presumed to be the mechanism.   Recent studies by Appel
and Tokiwa (1980) support these observations and indicate that atmospheric particulate nitrate
                           00
levels obtained with Teflon  filters may be only a small fraction of the true values.  Similar
studies with Gelman A glass fiber filters showed insignificant nitrate losses.
     Another mechanism  for the  loss  of  nitrate  from particulate  samples collected on inert
filters  is the dissociation  of NHJKU and  loss  of the resulting HN03.   The equilibrium vapor
pressures of NH, and HNQ-, above solid NH.NO, are appreciable and very sensitive to temperature
(Stelson  et al.,  1979).  In  a laboratory study, Appel  et al.  (1980)  observed losses of up to
                                                                                             ®
50 percent of  the particulate nitrate when  NHL-  and HNOv-free air was passed through Teflon
filters  loaded with about  200  ug  NH4N03 (<0.5 pm .particle  size) at  20 liter/minute  for  6
hours.   These  results  suggest  that  the  volatilization of NH.NO,,  can be  a major  source of
                                                               ©
negative  error  in  sampling  particulate nitrate  with  Teflon    filters.   The presence  of
relatively high NH-  and HNO, levels  in ambient air or high humidity may decrease  the  error,
while elevated temperatures should increase it.
                                             3-74

-------
     These results point to the conclusion that most of the existing data on urban ambient ni-
trate concentrations are  of  doubtful  validity.  Furthermore, it is unlikely that any of these
data can be  corrected  even if mechanisms for artifact formation or nitrate loss are clarified
in the future,  since  HNO«, which appears to play a significant role in each mechanism,  is not
routinely monitored.
3.3,4.4  Analysisof Trace Elements—Over the years, a variety of techniques have been applied
to  analysis  for  elements.  Presently,  the most  commonly used techniques  use some type  of
spectroscopic detection.   By definition,  these techniques respond only to the presence of the
elements in  the  PM and do not provide  information concerning the chemical compounds present.
For the most part,  techniques  for analyzing trace elements have not provided information con-
cerning the  oxidation  state  of the elements,  although  Braman et al.  (1977) reported attempts
at such analysis for arsenic.                                     ,
3.3.4.4.1  Atomic absorption spectrometry.   Atomic  absorption spectrometry  (Morrison,  1965)
has been widely employed for quantitative analysis of a large number of elements in particles.
In  principle,   a  beam  of  light of  a wavelength that  is  characteristic  of an  electronic
transition for the element  of interest is made to traverse a region of space with a constant
intensity and to  impinge  on a detector.  The  element of interest is atomized in a portion of
the beam of  light.  The amount of light absorbed by the atoms of interest in the sample can be
related to  the  amount  of that element present.   Any element can be determined if  a lamp is
available to produce the characteristic light.
     A variety  of techniques can be  used  for  atomizing and introducing the  element into the
light beam.   Generally, a flame or a heated carbon rod atomizer is used.  Flame techniques are
most commonly used for atmospheric PM.  An  extract  of the PM is prepared  and aspirated into
the flame, which volatilizes the sample and produces a sufficiently large population of ground
state  atoms  for  absorption.   An example  of  this  kind of  application  is the  EPA  reference
method for  lead  (U.S.  Environmental  Protection Agency, 1979d).   If the concentration  of the
element of  interest is  too  low  for  flame  application,  however,  or  if an extremely limited
amount of sample  is available,  an electrically heated atomizer can be used to volatilize atoms
into the light beam.   In this application, solutions can be used, or a small portion of soiled
filter without  any other  preparation may be examined directly.  In the latter case the filter
substrate must  be  oxidizable  and  the  representativeness of the  sample  may  be  questioned.
     Properly applied,  atomic  absorption spectrometry is  generally specific  for the analysis
of  the  elements  desired.   The instrumentation  can  be  inexpensive relative  to other instru-
mental techniques for the analysis of trace elements and is  generally  available  as standard
equipment in most analytical  laboratories.   However,  it can only  analyze one element  at a
time.  Additional elements must  be determined  serially.   This can be  a  severe disadvantage
when a number  of elements need  to  be determined on the same sample, both from the standpoint
of  the resources required to obtain the  information  and the  limitations  of the  volume of
                                            3-75

-------
extract available  to  perform the analysis.  Although  the  technique is useful, there are sig-
nificant  interference problems  in  some  determinations.   Spectral interferences  from other
elements  absorbing at the  same wavelength  can  be a  problem but  can  usually be  avoided  by
judicious  choice  of wavelengths.   However,  interference  effects  on the  element of interest
caused by  other  substances  present in the material  introduced into the spectrometer, refered
to as matrix  effects,  can be much  more  difficult to resolve.  These effects  can differ sig-
nificantly,  ranging from  the  effect  of  viscosity on  the amount  of  material  which  can  be
atomized  to  effects due  to the presence  of refractory  compounds  containing the  element  of
interest,  which  may  not be completely  volatilized at  the  flame or  atomizer  temperature.
Matrix  effects  can  adversely affect  the  concentration of  atoms   in the  beam and  result  in
significant  errors  in  the  measurement  of an  element.   The  literature is   replete  with
discussions  of  these difficulties  and provides both  general  and  specific  techniques  for
overcoming them.   One of the more convenient ways to keep abreast of developments in this area
is through the use  of a continually  updated bibliography with convenient indexing,  such  as
that provided by  "Atomic Spectroscopy." (S.  Slavin, ed.)  Bibliographies for this publication
are routinely updated and appear each January and July.
3.3.4.4.2  Optical emission  spectrometry.  Optical  emission spectrometry is a technique which
can  simultaneously  determine  the  amounts  of  numerous  elements.  The   advantage  of  this
technique  are obvious  in   situations  of limited  sample  availability or  limited time  and
resources  with which  to  do a measurement.   Conventional arc or spark-excited optical emission
spectrometry  has   been  used  extensively  on  atmospheric PM  (Scott et a!., 1976).   In most
applications of this  technique  an  extract of the PM is excited by a  spark or arc discharge.
This  decomposes  any  substances present  and excites  the  atoms to other than  their ground
electronic states.  In  the de-excitation to the ground state, light is emitted at a character-
istic wavelength.   The  intensity of the light emitted is an indication of the quantity of the
element  present.   Most  conventional  optical  emission  spectrometers   are  capable  of  simul-
taneously  analyzing 20 to 30 elements.
     Conventional arc or spark-excited optical emission spectrometers were never very popular,
partly  because  of  cost, partly because  the photographic  readout was complicated  and gave
rather  approximate answers,  and  partly  because  of high  detection limits  for a  number  of
species.   The development   of   optical  emission  spectrometers based   on  plasma  excitation
(Boumans  and DeBoer,  1975}  has resulted  in significant  improvements.   Although  there  are
several  kinds of  plasma  excitation, the commercially available optical emission spectrometers
with  inductively  coupled argon plasma excitation  has  proven most  advantageous.    In  this
technique,  an extract  of the  sample  of PM  is  aspirated  into  an  inductively  coupled argon
plasma whose  very  high temperature decomposes the materials and excites the atoms.  The light
emitted when  these excited species fall back to  the ground state  is collected  and monitored
just as before.   However,  this  approach has  numerous  advantages  not available with the older
excitation techniques.  The technique is capable of using the same  acid extract used in atomic

                                             3-76

-------
absorption.    It  is more  free  of matrix  effects than  atomic  absorption;  it  requires,, for. a
single  multielement determination  on  a  given  sample, about  the  same amount  of time  and
solution as a  single  element determination on atomic  absorption;  it usually has a much wider
linear range than  atomic  absorption;  and it has detection limits equal to or lower than flame
atomic absorption  (Fassel,  1978).   If an acid  extract of atmospheric PM  is  to be analyzed,
inductively coupled  argon plasma optical  emission  spectrometry  is usually  the technique of
current choice.
3.3.4.4.3   Spark  source  mass spectrometry.    Spark  source -mass   spectrometers  are  high-
resolution  magnetic  sectoring  mass  spectrometers  which  usually  use  photographic, emulsion
detection and  very high resojution  densitometry for quantitative  analysis.  They are uncommon
and  very expensive.   The material  to be analyzed is  incorporated into two  small  (usually
graphite) electrodes, which  are placed in the spectrometer with a well-controlled gap between
them.  A spark is  passed across the  electrodes,  vaporiz-ing  them and ionizing the material, in
them.   The  ions  are subsequently  led  into; the  mass  analyzer  portion of  the spectrometer
(Ahern, 1972).
     The electrodes used  in  the spark  source mass  spectrometer can be fabricated either fr,om
PM  that  has been  separated  from a  filter or. an  extract of  the  PM.   The technique 'is  not
suitable for the  generation  of  large data bases because only a few  samples can be analyzed-on
any  given  day.  The  time required  to  prepare the  instrument  and to obtain a  set of spectra
necessary for quantitation is substantial.  Double  ionization of elements is common and so are
ionized oligomers  of  carbon.  Therefore,  high resolution, detection  and complex interpretation
are  the  rule  rather than the exception.   The advantage of spark  source  mass  spectrometry is
that  it  can simultaneously estimate  the quantity of all nonvolatile-elements in the periodic
table and can do so with roughly equal sensitivity.                 „
3.3.4.4.4  Neutron activation analysis.   Neutron  activation  analysis (Morrison, 1965) implies
a variety of  distinct  procedures, all of which produce  unstable atomic nuclei .which then,emit
high-energy radiation  or particles.  The  intensity  of  a specific emission  and  its energy are
monitored as an indicator of. the element and its quantity.
     Instrumental  thermal neutron activation  analysis  is the technique  most commonly applied
to atmospheric  PM.  With  this  approach, a nuclear  reactor is  used  to produce neutrons, which
bombard the samples and produce the unstable nuclei.  The emitted gamma radiation is detected
by a Li-drifted Ge detector whose output is processed to produce the gamma-ray spectrum of the
irradiated PM.  The method  has  low detection limits, can simultaneously determine up to,about
25 elements in a given sample,  and particulate matter can be analyzed directly as received,-on
a very  small  portion  of the filter surface.   The-technique has been successfully applied with
the  glass  fiber  used  in hi-vol samplers (Lambert  e± a!.,  1979).   The 'time  required  for
analysis is short. However,  data are usually-not available until 2 to 3 weeks after the sample
                                            3-77

-------
is  irradiated  because there  is a  significant delay period  between the  irradiation  and the
collection  of  gamma-ray  spectra  for certain  long-lived isotopes.  Some  important elements
(e.g.,  S  and Pb) are not  practically measured using this method.   These  limitations  and the
need  for  very  complex,  highly specialized and  expensive equipment are the main disadvantages
of the neutron activation technique.
3.3.4.4.5   X~ray f1uorescencespectrpmetry.   X-ray  fluorescence  spectrometry  is  a  multi-
element, nondestructive technique that can simultaneously determine numerous trace elements in
PM,  directly on  the filter media.   It involves the excitation of  tightly  bound electrons in
the  atoms  by an  X-ray  generator  arjd observation  of  the X-ray  emissions  that occur  as the
de-excitation of the electrons proceeds (Dzubay, 1977).
     X-ray  fluorescence  spectrometers  use  either  energy-dispersive  or wavelength-dispersive,
detection.  Spectrometers using energy-dispersive detection simultaneously collect all  emitted
quanta with  a  silicon-lithium detector and, through  subsequent processing, can analyze about
30   elements.    Spectrometers   using  wavelength-dispersive  detection   monitor  carefully
preselected wavelengths that are characteristic of  the de-excitation emissions of the,elements
of interest.  With wavelength-dispersive detection, about 20  elements can be determined simul-
taneously on a  single sample and  interelement  effects  are minimal due to the high resolution
capability  of  the  instrumentation.   With  energy-dispersive  detection,  all  wavelengths are
simultaneously collected  and interelement  corrections must  be handled  in  the data reduction
process.
     Good detection  limits  and  the ability to  handle  a sizable number of samples nondestruc-
tively  with  minimal  sample preparation  are  clear  advantages  of the  X-ray  fluorescence
technique.   In  order to analyze the  sample  directly,  however, it  must  be  of uniform  surface
texture,  and it is  best if the particulate  layer is very thin.   This  obviously places some
limitations  on the  kind  of sample that can be analyzed without preparation.  Even in the most
ideal  samples,   concern  with   special  corrections  must  exist  (Gould  et  al.,  1976).   The
techniques  has  been  applied extensively  to  analysis  of  filters  from  dichotomous samplers
tDzubay and Stevens, 1975).                                                 ~
3,3.4.4.6  Electrochemical Methods.   Electrochemical  methods  have been used  to  a limited ex-
tent  to  determine a  small  number  of elements in  airborne particles.   These methods  include
potentiotnetry with  ion selective  electrodes,  polarography,  and  anodic stripping voltammetry
(Morrison,  1965).  Electrochemical  techniques  have few  advantages for  airborne particulate
analysis, aside  from their  low initial  capital equipment cost compared  to, other techniques.
While the methods  are usable (Ryan and Siemer, 1976) there appears to be fairly little use of
such techniques at present, except in the area of ion-selective electrodes.
3.3.4.4.7  Chemical methods—In the past, many classical wet  chemical procedures were employed
for trace element  analysis of airborne particles.  In general, a col or-forming reagent was in-
volved.   The amount of the given element present is determined by the extent of color develop-
ment.  Perhaps the best known of these procedures is based on the use of diphenylthiocarbazone

                                             3-78

-------
(dlthizone) as the  colorlmetric  reagent for lead (Snell,  1978).   Wet chemical procedures are
labor intensive and  slow  compared with spectral techniques.  Sample preparation and interfer-
ences are  also usually  a  problem.  As  a  result,  laboratories with heavy sample loads tend to
use instrumental  methods,  particularly spectral techniques,
3.3.4.5   Analysis of Organic Compounds—Numerous  papers   have   appeared  dealing  with  the
characterization of  organic compounds  in airborne particles.   The following  discussion was
taken primarily  from the  monograph  edited by  H.  Malissa  (1978)  and  describes the principal
methods used in this field and some typical examples mentioned in the literature.
     Organic  compounds  significantly  contribute  to  total  PM   in  urban aerosols.   Organic
contents of  up to  43  percent of the total particle  mass  have  been reported  (Hidy,  1975).
Characterization of  organic compounds  in ufban aerosols  generally  involves  trace separation
and identification by gas and liquid chromatography, with detection methods having sensitivity
in the  nanogram  range.   The sample amount needed to allow analysis of substances in parts per
million  concentrations   is in  the milligram  range  (Cautreels   and  Van  Cauwenberghe,  1976;
Ketseridis, et a!., 1976).  Usually, high-volume samplers with glass fiber filters are used to
provide the needed sample sizes.
     One of the earlier simple and extensively used methods for estimating the organic content
of PM was  called "benzene soluble organics."   Filters  were simply refluxed with benzene in a
soxhlet  extractor  for  several   hours.   The benzene  was  evaporated, and the  weight  of the
residue was measured and reported.  Benzene soluble organic data were recorded in the National
Aerometric Data  Bank for  many years.   Because of the purported hazard of benzene, this method
has  not  been used  on  a  national   scale  by   any  single  laboratory for  roughly  a decade.
Extraction efficiencies of 25 different solvents and 24 binary  mixtures were investigated by
Grosjean (1975).   Grosjean determined that extraction with benzene or other nonpolar solvents
usually  leads  to  serious underestimation  of  aerosol  organics,  especially  of  the  polar
secondary  (photochemical)  products like  carbonyl  compounds,  organic  nitrates, or carboxylic
acids.   The  use  of  binary mixtures  for extraction or a nonpolar  and a polar  solvent for
successive  extractions  was  strongly recommended.   This   leads  to a  higher  organic  carbon
         •i
extraction  efficiency  (in comparison  to  benzene  as solvent) than with  both  single polar and
nonpolar solvents.
     In  the  area  of compound-specific analysis,  a large amount  of  early  work techniques
focused  on the  measurement of  polycyclic  aromatic hydrocarbons  (PAH).  Numerous measurement
techniques  have  been proposed to analyze quantitatively for many of the polycyclic aromatics.
Chromatographic  techniques (Sawicki,  1964; Thomas et  al.,  1960) were  used in  much  of the
earlier  work; more  recently,  frozen  solution  fluorimetry (Bacon  et al.,  1978)  and matrix
isolation  spectroscopy  (Wehry and  Mamantov,  1979) have been explored.   High-pressure liquid
chromatography  (HPLC)  is  a promising technique for separation  of high  molecular weight PAH.
The  development  of bonded octadecylsilyl (ODS) columns  of  micro  particle  size allowed Fox and
                                             3-79

-------
Stal.ey  (1976)  to accomplish the near baseline  separation of the benzo[a]pyrene (BaP) arid its
isomer, benzo[e]pyrene.  A significant increase in sensitivity over other methods was achieved
by  use  of fluorescence  spectroscopy for on-line detection.  Perhaps  the  most extensive data
base  has  been  concerned  only  with BaP   (Swanson  et  a!.,  1978)  utilizing  a  thin  layer
chromatographic technique with fluorescence detection.  Gas chromatographic (GC) separation of
organic extracts  of  airborne particles requires the  application  of preseparation steps, such
as  thin layer  chromatography (Zoccolillo et a!., 1972) or liquid-liquid extraction (Cautreels
and Van Cauwenberghe, 1976; Ketseridis et al. , 1976).  Primary extraction is generally carried
out by  means  of single solvents such as benzene, cyclohexane," or others.  A typical procedure
including solvent extraction for preseparation is described by Ketseridis et a1.(1976).
     The application  of  gas chromatography coupled with  mass spectrometry for the analysis of
benzene"extractable  compounds  in airborne  particles is  described  in  detail  by Cautreels and
Van Cauwenberghe  (1976).   This work  led to  the identification of more  than  100 compounds in
urban aerosols.  The benzene-extractable compounds (5.8 percent of total particles) were sepa-
rated into  neutral,  acidic, and  basic substances.   The  acidic fraction was  converted to the
methylated derivatives  for  GC  analysis.   In  the   neutral  fraction,  22  saturated  aliphatic
hydrocarbons, 36 polynuclear hydrocarbons, and 13 polar oxygenated substances were identified.
In  the  acidic  fraction,  19 fatty acids  and  19  aromatic  carboxylic acids were identified.  In
the basic fraction, 15 peaks of nitrogen-containing analogs of the PAH were identified.
     Interest  in  the  organic content of atmospheric  particles  ranges  from particulate carbon
(Rosen  and  Novakov,  1978) to  any  other  possible organic substance.   A  variety of techniques
have been used in an effort to  solve this  problem (Fox  and Jeffries, 1979).    However, it is
clear that  this area  is  large, exceedingly complex,  and will  need a great  deal  of develop-
mental effort.
3.3.4,6   Analysis of Total  Carbon and Elemental Carbon—The  most 'widely  used  technique  for
measurement of total carbon  in  collected PM  is  by combustion to  carbon  dioxide  followed by
detection of  this gas;  or,  after  C0?  to CH*  reduction, of an  equivalent quantity  of CH,.
Several  commercial instruments  are  available for this purpose.   Johnson and Huntzicker (1979)
have designed  an  instrument for analysis of organic and  elemental carbon o'n filter samples of
ambient particles.   The  organic  carbon  is  volatilized  or pyrolyzed  in an inert atmosphere,
oxidized to C02, and converted to methane, which is detected with a flame ionization detector.
Elemental  carbon  is  determined  after oxidation to C0? and  chromatographic separation of the
CO- from 0-.   Cadle  et al.  (1980) have discussed a similar system using non-dispersive infra-
red detection of the C0? and automation of the analysis procedure.
     Nondestructive analysis of carbon in collected PM involving the interaction of light with
the particles  (i.e., changes  in reflection, transmission,  and  absorption of the filter) is
used to  estimate elemental carbon  loading.   Delumyea et al. (1980) used  a reflectance tech-
nique having  a tungsten  filament lamp  as  a light  source.   Lin et  al. (1973)  described an
                                                                            ®
apparatus to  integrate light  scattered  by particles collected on  Nuclepore   filters.   Since

                                             3-80

-------
elemental carbon is the major absorbing species in ambient PM, this technique has been used by
other  researchers  to  estimate the  elemental  carbon  content of  collected  PM.   Yasa  et  al.
(1979)  have  shown a  linear  relationship  between  photoacoustic measurements  and  optical
attenuation  measurements  of  the  type  done by  Rosen  et al.  (1978),  to establish  that  the
optically absorbing component  of  urban aerosol particles  is  graphitic  carbon.   Photoacoustic
measurements  have  also  been  used  for in-line measurements  of diesel  particulate emissions
(Faxvog  and Roessler,  1979).   (See  Chapter 9 for  a more thorough discussion of  the role of
light absorbing particles in visibility reduction.)
     Macias et  al.  (1978)  used the gamma-ray  analysis  of light elements (GRALE) technique to
measure carbon in ambient aerosols deposited on quartz or Teflon filters.  The GRALE technique
involves  the in-beam  measurement of  gamma rays  emitted from  an aerosol  sample  during  the
inelastic scattering of 7-MeV protons accelerated in a cyclotron.
     Calcium  carbonate  also is commonly  found in airborne particles.   Mueller  et al.  (1971)
described a technique  to  distinguish  carbonates  from  elemental  carbon  by  acid evolution of
co2.
3.3.5  Particle Morphology Measurements
     Visual examination of  particles  collected on a filter or impaction substrate can provide
extremely useful  information  concerning  the sources and  transport of  airborne  particles.   A
reticle-equipped light microscope can  be used to  examine  particles  larger than about 0.5 urn.
Use of transmission and scanning electron microscopes can  improve the resolution for particles
as  small  as 0.001  (jm.   The effective ranges of microscopes and their utility are described by
McCrone  (1973)  and shown  in Appendix Figure 3A-13.  Particle size distributions by number can
be generated using statistically valid counting procedures.   By applying an average density an
estimate of the size distribution by mass can be made.
     Microscopic identification  and analysis  requires a  high  degree of skill  and experience
plus  extensive  quality  assurance to  provide  meaningful  information.   Critical   factors  in
effective  use of  these methods  are  the selection  of  sampling  substrates,  allowable particle
loadings,  and sample handling.   In  addition,  particle  interactions  and structure changes on
the collection  surface  must be minimized if accurate size distributions and characterizations
are  to be  obtained.   In  a  study  of  ambient  particles collected on hi-vol  filters, Bradway
et al.  (1976)  examined  the ability  of multiple  microscopists to characterize  particles in
specific  categories.   Significant  misidentification  and misassignment problems  were  noted,
which  made  it difficult to compare results.  Multiple microscopists and blind replicates were
recommended as  standard procedures for quantitative optical characterization studies.
3.3.6  Intercomparison of PM Measurements
     The  intercomparison  of particle  sampling methods  is not  straightforward  because  of the
complex  nature  of  ambient particulate matter.  As  noted earlier  in this chapter, gravimetric
mass measurement methods can differ dramatically in the  particle size ranges collected and the
                                             3-81

-------
sensitivity of the  sampler to external factors, such as windspeed.  Also, nongravimetric mass
measurement methods examine  only  specific  portions  of the particle  size spectrum  and,  in
addition, measure selected integral properties of particles rather than mass.
     Comparison  of  gravimetric  mass  measurement  methods  can  be made  in the  field  through
collocated sampling or  in the laboratory by examining the sampling effectiveness of the inlet
or substage for  various particle sizes.  Because it is difficult to simulate the character of
real  suspended  particles, however,   the  final  intercomparison  test  must  be performed  at
selected field locations.  The choice of the number and types of field locations is important,
since the  local  particle  sources can  have  a substantial impact  on  the  sampler performance,
especially if coarse particles dominate the size distribution.
     The  most  recent  and  comprehensive  intercomparison  of   gravimetric  mass  measurement
particle samplers was reported by Camp et al. (1978).  Eleven different types of samplers were
compared for mass and  other analyses  including sulfates, nitrates, and elemental composition.
The  most  salient  observation  of the study was  the  difficulty  of  intercomparison  of  the
samplers because of differences in inlet or substage particle size cutpoints.   Since there is
no reference sampler,  all measurements of the same size fraction were averaged as a comparison
measurement.    Some  duplicate  samplers were  present,  permitting  reproducibility measurements.
The  coefficient  of  variation  for  the  automated dichotomous samplers was  determined to  be 11
percent for the  coarse  fraction and 3  percent  for  the fine fraction.  The  same values  for a
manual  dichotomous  sampler were  18  percent  and  1  percent, respectively.   The  high-volume
impactor used  in the  CHAMP network gave values of 15 percent and 5 percent, respectively, for
the  equivalent size fractions.   The results from this study should be considered "best case,"
since the  sampler   operations were monitored continuously by highly  skilled  individuals.   In
some  cases,  the  operators  were the developers  of  the sampling method.   It  is expected that
routine field  sampling by  less qualified personnel would produce  larger variabilities.   The
reproducibility of  certain chemical analyses  were  reported  to be better than the  mass mea-
surements,  such  as elemental  sulfur,  which  averaged  ±3  percent  for  all   size  fractions.
Overall,  the  study showed that comparable  results  could be  obtained  by  different  particle
samplers if appropriate quality assurance steps were  taken  and  identical size fractions were
compared.
     Miller and DeKoning (1974) compared the TSP high-volume sampler with several commercially
available cascade impactors.  None of the impactors gave results comparable to the high-volume
sampler, but  the two types of  samplers did correlate reasonabley well.   The  agreement  among
cascade impactors for  mass median diameter  (HMD) was very poor.   The HMD  often  differed by
more than a factor of 2.
     Comparison of  gravimetric  mass  measurements with indirect  mass  measurements  should only
be attempted to  determine correlation  or to test a  physical  model relating the measurements.
The  literature contains many  intercomparison studies attempting to relate the TSP hi-vol with
surrogate techniques such  as:   the British Smoke Shade sampler (Commins and Waller,  1967; Lee

                                             3-82

-------
et a!.,  1972;  Pashel and  Egner,  1981);  the integrating.nephelometer  (Charlson  etal.,  1968;
Kretzschmar, 1975);  and the  AISI  tape sampler  (Lee  et al., 1972; Ingram  and  Golden,  1973).
Comparisons are  also available between other  direct  and indirect mass  measurements,  such as
the dichotomous  sampler with  the  AISI tape sampler  (Regan  et al, 1979) and with  the  nephe-
lometer  (Waggoner  and Weiss,  1980).   In many  cases, a simple regression  was  fitted between
measurements rather than attempting to establish a physical  basis for the comparison for test-
ing with empirical  data.   Most of these comparisons  were attempts to demonstrate the useful-
ness of  a  nongravimetric  sampling method  for  predicting  the mass concentrations  that would
have been  measured by  a  gravimetric mass  method.   However,  there  is  currently  no indirect
technique  that  has gained acceptance  as  a general-purpose  surrogate  for direct mass concen-
tration measurements.
     Mulholland et al. (1980) compared the estimated mass concentrations calculated from Elec-
trical  Aerosol  Analyzer (EAA)  measurements with direct gravimetric  analyses.    It was noted
that the  errors are  in the  ±  20  to 30  percent range for  spherical  particles,  and for non-
spherical particles  the errors are as high as ± 60 percent.  Therefore, great care should be
taken  in  attempting to predict gravimetric mass concentration  from  nongravimetric particle
measurements.
3.3.7  Summary
     Particulate matter suspended  in ambi'ent  air contains a  range of  particle  sizes  and
shapes.  Separating particles according to aerodynamic size  groups particles that behave alike
in more  situations of interest to human  health  and welfare (other than visibility) than  any
other measure of particle  size.  Samplers can be designed  to  collect specific  size fractions
or match specific  particle deposition patterns through carefully designed inlets and substage
fractionators.    Mass concentration  derived  from  gravimetric analysis is  the most  common
measure  of PM.     High-volume  samplers,  dichotomous samplers,  cascade  impactors,  and cyclone
samplers are the most common examples of this type of measurement.   Carefully collected size
distributions of  ambient  particle mass  have  shown that most  particle samplers  underestimate
the concentration  of particles  in the air because of external  factors  such as windspeed or
because  their  particle transport  systems are  not  effective  for  the larger particle  sizes.
     Mass  concentrations   can  be  estimated  using methodologies that  measure an  integral
property of particles  such as optical reflectance.  Empirical  relationships between mass con-
centrations and the integral measurement have been developed and are used to predict mass con-
centration.  Without  a  valid physical  model relating  the measurements plus empirical data to
demonstrate the  model,  these  techniques  have a limited ability  to  estimate mass  concentra-
tions.    These  conditions  are  poorly  met  in  the  case of  reflectance or  transmission tape
samplers,  fairly well met  in the  integrating nephelometer,  and  very  well met in the case of
beta-ray attenuation analysis.
     Sampling accuracy  is  difficult  to determine directly,  since the measurement requires  the
production of very accurately known  concentrations of particulate matter of a wide variety of

                                            3-83

-------
sizes.   Instead,  accuracy  is  estimated by  determining the  accuracy  of  sampler  flow rate
measurement or control, and inlet sampling effectiveness.  These separate measurements provide
a means of intercomparing methods in the absence of a reference measurement technique.  Recent
interest in larger particle sampler cutpoints (e.g., 15 pm)  have resulted in wind tunnel test
procedures for  particles that  determine sampling  effectiveness  under controlled conditions.
Such  measurements  have added  significantly to  the  ability to  estimate particle sampling
accuracy.
     Recent evaluations show that  the high-volume sampler  collects a  smaller particle size
range than that  reports in the 1969 criteria document (National Air Pollution Control Admini-
stration, 1969).  The  sampling  effectiveness of the hi-vol inlet also is wind speed-sensitive
for  larger (>10  pm)  particles.   Sampling  biases caused  by  typical  wind  speeds  could  be
expected  to  cause  no  more  than  a 10-percent  day-to-day variability  for the  same ambient
concentration.    The  hi-vol  is  one of  the  most reproducible particle  samplers  currently in
use, with a coefficient of variation of 3 to 5 percent.  A significant problem associated with
the glass  fiber  filter used on  the  hi-vol  is  the formation of artifact mass  caused  by the
                                                                        2
presence of acid  gases in the air.  These artifacts can add 6 to 7 pg/m  to a 24-hour sample.
     The  dichotomous  sampler was  designed to  collect the fine  and coarse  aiftbient particle
                                                                                  ®
fractions, usually  providing a  separation at  2.5  pm.  This sampler  uses Teflon  filters to
minimize  artifact mass formation and  is  available in versions for  manual or automatic field
operation.  The earlier inlets used with this sampler were very windspeed dependent, but newer
versions  are  much  improved.    The  dichotomous  sampler  collects  submilligram quantities  of
particles because of  low sampling flowrate and requires microbalance analyses, but is capable
of reproducibilities of ± 10 percent, or better.
     Cyclone  samplers  with cutpoints  in the vicinity of 2  urn  have  been  used  for  years  to
separate the  fine particle  fraction.   A recent development has coupled the cyclone sampler to
a sampler  inlet  to  give a 15 pm  cutpoint.   Cyclone samplers can be designed to cover a range
of  sampling  flowrates  and are  available in a  variety of physical  sizes.  A  10mm version is
available  for  personnel dosimeter  sampling.   Cyclone sampling  systems  could  be expected to
have coefficients of variations similar to that of the dichotomous sampler.
     The Size Selective Inlet  (SSI) hi-vol collects samples containing particles less than 15
Miu for  comparison with TSP.   This sampler is identical to the TSP hi-vol except for the inlet
and is expected to have the same basic characteristics.
     Cascade  impactors have  been used  extensively to  obtain  mass distribution  by particle
size.    Because  care 'must  be exercised  to  prevent errors, such  as those caused by particle
bounce  between  stages,  these  samplers  are  normally not  operated as  routine  monitors.   A
comparison of impactors  showed  inconsistencies  in  the MMD  and  in  total   mass collections
compared with the hi-vol.
     Samplers that  derive mass  concentrations  using analytical techniques other than  direct
weight  have been  used extensively.   One of the  earliest was  the British Smoke Shade sampler,

                                            3-84

-------
which measures the reflectance of particles collected on a filter and uses empirical relation-
ships to predict mass concentration.  These relationships have been shown to be more sensitive
to carbon concentrations than mass; hence, they are very difficult to interpret as total mass.
Another optical technique, the AISI tape sampler, uses transmittance instead of reflectance to
predict  mass.   It  has been  shown  that this  sampler  correlates favorably  with gravimetric
measurements  limited  to the  smaller particle sizes.  Several  researchers  have also reported
good correlation between the integrating nephelometer and gravimetric fine particle mass.  The
EAA, however,  was shown to have difficulties in reliably predicting gravimetric mass measure-
ments.                                             -  -
     Optical particle  morphology  techniques  are very useful for identifying the character and
sources  of  collected particles.  Researchers have noted,  however, that  these techniques are
dependent on the skill of the microscopist and stressed the need for careful quality assurance
procedures.
     An  extensive number  of analytical techniques are available  for  analyzing particles col-
lected  on  a suitable  substrate.   Many  of  the analytical  techniques, such as those for ele-
mental  sulfur,  have been  demonstrated to be  more precise than  the  analyses for gravimetric
mass  concentration.   Methods  are  available  to provide reliable  analyses for  sulfates,  ni-
trates,  organic fractions,  and elemental composition (e.g., sulfur,  lead,  silicon).  Not all
analyses can be performed on all particle samples because of factors such as incompatible sub-
strates  and inadequate sample  size.   Misinterpretation of analytical results can occur when
samples  have not been  appropriately segregated by  particle size  and when  artifact  mass is
formed  on  the substrate  rather than  collected in  a  particle  form.   Positive  artifacts are
particularly likely  in sulfate and nitrate determination, and negative nitrate artifacts also
occur.
     Sampling technology  is available to meet  specific  requirements,  such as providing sharp
cutpoints,  cutpoints which  match  particle deposition models, separate collection of fine and
coarse  particles,  automated sample  collection capability, collection  of at  least milligram
quantities  of  particles,  minimal  interaction of the substrate with  the collected particles,
ability  to  produce particle  size distribution data, low purchase  cost,  and  simple operating
procedures.    Not  all  these  sampling requirements may  be needed  for  each measurement study.
Currently,  there is no single sampler which meets all requirements, but samplers are available
that can meet most typical requirements.
3.4  MEASUREMENT TECHNIQUES FOR ACIDIC DEPOSITION
3.4.1  Introduction
     Studies  designed  to monitor precipitation first appeared in the  literature around the
turn of  the century.  Many  small-scale networks were organized in the United States and Europe
between  the 1920's  and the 1950's.   Knight  (1911),  Wilson (1926), and Collison and Mensching
(1932)  reported the  earliest U.S. precipitation chemistry studies at local sites.  The physi-
cal  size of the networks changed during the 1950's  from single or dual site studies to large-

                                            3-85

-------
scale national/International  studies.   During the 1950's, Barrett and Brodin (1955) organized
a European monitoring network and Junge and Gustafson (1956) established the first continental
U.S. network.  The World Meteorological Organization (WHO) reported (1971) the formation of an
international network to monitor global trends and changes in the chemical composition of acid
precipitation.  In addition to the existing WHO network, various local, regional, and national
acid precipitation studies currently operate in the industrialized nations.
     Ecologists and  biologists were  among  the first scientists to become concerned with the
causes and effects  of acid precipitation.  These  two  groups were responsible for most of the
acid precipitation studies conducted before 1970.  During the period 1974-1976, multidiscipli-
nary study  of acidic  precipitation  phenomena expanded.   International  scientists  from every
scientific field  focused their  attention on  the  potential  affects  of acidic precipitation.
Symposia, committees,  and groups were  organized  to examine every" facet  of  past and on-going
studies  (Dochinger  and  Seliga,  1976;  Kronebach,  1975; Likens  et  al., 1972;  Likens,  1976).
Network  siting,  sampling procedures  and analysis schemes were  critically reviewed.  Special
committees were formed to develop techniques to describe statistically the quality of the pre-
cipitation chemistry  data being  reported by the  various  international  laboratories.   During
this period,  concern  about the effects of particulate deposition on vegetation increased.  As
a result, improved wet/dry collection devices were designed to give a better understanding of
total acidic deposition problems.
     Many new studies resulted  from  this increased emphasis on acidic deposition,  including
the  development  of  a long-term  continental  U.S. monitoring  network (Galloway  and Cowling,
1978).    Comprehensive reviews  of past and current  studies aife provided by Niemann  et al.
(1979); Kennedy (1978); and the U.S.  Atomic Energy Commission (1974).
3.4.2  U.S.Precipitation Studies
     Past U.S.  precipitation chemistry studies can  best  be described as  ad hoc (U.S.  Atomic
Energy Commission, 1974).   New studies were randomly developed without adequate consideration
of  either  past or current  proposals.   General characteristics  of past  U.S. studies include:
     1.   Overall  study objectives varied among projects.
     2.   Pre-1970 studies were short-term, lasting only 1 to 2 years.
     3.   Sites were  randomly selected  at  locations, of convenience.   Siting  with  respect to
          program objectives or standard siting criteria were rarely considered.
     4.   Sampling/storage procedures  and the extent of sample chemical analysis varied among
          studies.
     Each of these deviations in study design/protocol obviously affects the existing data and
precludes any simplistic consolidation or correlation of past study  results.  Of these varia-
tions,  the differences  in sampling/storage procedures  are the  most  difficult to resolve.  In
general, sampler  collection efficiencies, sample  representativeness,  and  sample integrity at
the time of  chemical  analysis (i.e.,  does  the sample reflect what was collected in the field
or have chemical  changes occurred?) can only be speculative for the earlier  studies.

                                             3-86

-------
     Since 1900, techniques used to collect and store precipitation samples have changed.   The
most significant change in wet-only or bulk precipitation sampling is in the collector itself.
In earlier  reports,  monthly bulk  samples  were manually collected in  glass devices.   Plastic
devices became  the collection  medium during 1950-1960.  Galloway  and  Likens (1976) note that
plastic collectors  are preferred  today  for inorganic  species, whereas  glass  collectors are
currently  used  for  organic sampling.   Automated wet/dry  samplers  are replacing  the  manual
wet-only  and  bulk collection  techniques.   The height  of the  collection  device above  ground
level  has varied throughout  the years.   Earlier studies  placed  collectors at  ground  level
while  current studies  commonly place the  collector  1 to 3 meters above  ground level.   Other
study-dependent sampling variations observed in wet-only and bulk sampling procedures include:
filtering versus  non-filtering of  the sample before chemical analysis; the use of biocides to
preserve the sample and retard biological growth;  and the storage techniques used after sample
collection and  before  chemical analysis.  Although the degree to which each of these sampling
and  storage differences  affect the sample remains unanswered, it is generally agreed that the
study data will  be related to the study  sampling techniques employed.
     When  reviewing  past  studies,  the  extent to  which these  common variations  biased the
resulting data  must be  determined.   For  example,  the effects  of  sample evaporation,  splash
contamination,  loss  of initial  (usually most concentrated)  rainfall,  and contamination from
insects,  leaves,  etc.,  have been commonly reported.   How the authors addressed these problems
differ.   Some  deleted the  questionable  data,  others  did  not,  and  still  others  stated that
these effects would be averaged out over the length of the study.
     The data analyst must also know if  the samples were filtered before analysis.  Past study
data  indicates  that the  inclusion of  particulate  "wash-out"  material  in bulk  and  wet-only
samples as well  as dry deposition in bulk samples changes the overall chemistry of the sample.
Several studies routinely filtered the sample before analysis. ' This filtration of particulate
matter, depending on  technique and actual time when filtering occurred, could possibly change
the  resulting sample chemical  composition and related analytical data.   Whether glass or plas-
tic  collection  devices  were used could  also affect the data.  Galloway and Likens (1976) note
the  leaching  of  inorganic  species into and out  of  glass collection devices and  the loss of
organic  species with plastic  devices.   Other authors  (Kadlecek and  Mohnen,  1976; Norwegian
Institute for Air Research, 1971) report similar findings.  Metal ion losses in dilute samples
have  been  repeatedly  reported.   To  minimize  this  potential  metal  loss, various  authors
acidified  a  representative aliquot  of  the  sample  immediately  following  collection.   Again,
others  did  not.   Sample storage techniques also  vary among studies.  Larger networks usually
kept  the  sample  in  a cool, dark  place.  Some  smaller networks  either  froze the  sample or
refrigerated the  sample  at 4°C.  Galloway and Likens (1976) indicate that  significant changes
do not occur  when samplers are  stored at  4°C.   Unfortunately, this storage technique is cost
prohibitive  for large national/international  networks.   How  long  a sample  is stored  before
                                             3-87

-------
chemical analysis  may also affect its integrity.  A review of the literature reveals that the
reported length  of time from sample collection to sample analysis has varied by as much as 60
days.  Some  authors used chemical biocides to  retard  green algae growth in samples collected
in  warmer  climates  because the  presence  of  algae  changes  the sample  chemical  composition.
Obviously,  the storage  technique affects  sample integrity at the time  of chemical analysis.
     Although  procedures for collection of dry deposition are not as well documented, similar
sampling variations  are expected depending on technique (i.e., dust-fall buckets, ambient air
monitors, or  automated dry-only collection devices).  Siting  of  the dry deposition collector
is  crucial;  in particular,  the  height  of the collector above ground  level.   Dry deposition
samples 0 to 1 meter above ground level have been reported to be heavily influenced by contri-
butions from the local terrain, as well as bird and vegetation contamination.
     Data from special  studies  is also available for summation.  Behrmann  (1975), Pickerel!
et al. (1979),  Anderson (1978),  Cooper et al.  (1976),  and Gatz et  al.  (1971)  report special
one-of-a-kind  samplers  to  monitor the  change in the  chemical  composition  of precipitation
events at single sites.  Special  sampling  procedures have been developed for the collection of
snow (Galloway and Likens, 1976;  Hagen et  al., 1973), fog (Waller, 1963; Mrose, 1966), indivi-
dual  raindrops and canopy  throughfall  (McColl and Bush, 1978).   Understanding these special
sampling techniques is essential before these data are compiled or summarized.
     The data analyst  must  also consider the various collection  periods  reported  in  past
studies.   Over the  years,  bulk  samples  have  typically been collected  on  a  monthly basis.
However,   wet  deposition  collection  periods  have  ranged  from   event  sampling  to  monthly
sampling.  Although  clearly  defined  in terms of  showers and  thunderstorms,  the definition of
the  beginning  and ending  of an  event  during a  large  frontal system  varies  from  author to
author.  Monthly  sampling  is  common  in  larger  networks  designed  to  monitor the  trends in
chemical  composition over  time. • Daily or more frequent sampling  is  typically  reported at
individual  sites  with the  objective  to determine exact chemical loadings at specific sites.
Weekly  sampling  is  currently  recommended  by  Galloway  and  Cowling (1978)  as  the  minimum
allowable sampling frequency  to  obtain   usable  wet deposition  results.   Dry deposition is
commonly collected on a 1-to 2-month  basis,  as recommended by Galloway  and Cowling (1978).
     Each  of  the   sampling  and  storage  variations  addressed above can affect the  sample
integrity and  resulting data.   In addition,   site  meteorology and  collector efficiency  also
affect the  sample.  Summers  and  Whelpdale (1976)  stress  the existing need  to document  both
scavenging and collection mechanisms involved with acid precipitation.  Initial reviews of the
most  commonly used  collectors have  been conducted and are  provided  by Niemann  (1979)  and
Galloway and  Likens  (1976).   Additional  comprehensive  collector evaluations, including  dry
deposition collector efficiencies  and species  collected,  along with  a reevaluation of  the
meteorological mechanisms  involved  in  acidic  precipitation processes are  needed.   Before  any
past data summarization  can be developed, a careful analysis of the sampling and storage  pro-
cedures and collection mechanisms must be  performed.

                                            3-88

-------
3-4.3  Analytical Techniques
3.4.3.1   Introduction—Analytical  methodologies currently  employed to  analyze precipitation
samples are, for  the most part, state-of-the-art freshwater or natural water analytical tech-
niques.   Typical  procedures are presented  in Methods of Air  Sampling and Analysis (Interso-
ciety Committee, 1977a),  Standard Methods for the Examination of Water and Waste Water (Ameri-
can Public Health Association,  1971) and U.S.  Environmental  Protection Agency's "Methods for
Chemical Analysis of Water and Waste" ( U.S. Envrionmental Protection Agency, 1979d).  Theore-
tical  detection limits and  the quality  of the data in terms  of  precision/accuracy for each
technique are specified in the literature.  Rainwater, however, is a dilute solution of chemi-
cal species, and  represents extremely pure freshwater or natural water.  Chemical analyses on
rainwater yield results at or below the published analytical detection limits (Miller and High-
smith,  1976).   Added  precautions  must  be  taken  to minimize  field/laboratory contamination
(Likens,  1972),  which  is  analytically  indistinguishable  from,  and  sometimes  larger  than,
natural rainwater species  contributions.   To preclude changes  in  the  chemical  composition of
the sample,  analyses should occur  within 24 hours after sample collection.  Although operator
and instrumentation dependent, laboratory analyses should meet or exceed the analytical preci-
sions and accuracies presented in Table 3-6.  Operator and instrumentation biases must be mini-
mized  through  supporting  internal  and external quality assurance  programs.  Before 1975,, no
mechanism existed to externally evaluate the quality of the precipitation chemistry data being
reported  by  the  international  precipitation laboratories.    WHO  instituted such  an interna-
tional  quality  assurance  program  in 1975.   Potential  errors  in past data have subsequently
been  noted by  Ridder  (1978), Galloway et. a!,, (1979), Tyree et al. (1979), and Tyree (1981).
3.4.3.2  Analysis of Acidic Deposition Samples
3.4.3.2.1   Sample preparation.   Wet  deposition samples  are  allowed  to equilibrate  to room
temperature  before chemical  analyses.   Sample pH  and  conductivity  are  initially measured.
Filtering  or centrifuging of the  sample  may follow.  A representative  portion of the sample
may  then  be acidified (HN03) to  preserve the  metal  ion concentrations.   Between analyses,
samples are  either stored in a dark,  cool  place or at 4°C.   Dry  deposition samples are dis-
solved  with  a  known quantity of distilled water (typically 50 ml).  Analytical  procedures for
these  dissolved dry deposition  samples are  identical  to  the wet deposition analytical proce-
dures described below.
3.4.3.2.2  Volume.  Direct volume measurements, accuracy + 3 percent,  are made  on the wet depo-
sition  sample with Class A graduated cylinders.  Care must be taken to ensure  that t,he sample
is not  contaminated by the  labware used in this procedure.  Indirect volume techniques include
weighing  the collection  container  before and after  the  sampling period or. measuring the col-
lection in a standard  rain gauge (cylinder, tipping bucket, or weighing).  Standard rain gauge
accuracies are  +  0.02  in or better, depending on the manufacturer.  The weighing rain gauge is
                                            3-89

-------
             TABLE 3-6.  RECOMMENDED PHYSICAL/CHEMICAL PARAMETERS FOR ANALYSIS

Parameter
Volume (field)
pH (field + lab)
Conductivity
(field + lab)
so4
N03
Cl
NH4
K
Ca
Na
Mg
Acidity
Unit of
report
inches
PH
|jS/cm
mgS/1
mgN/1
mgCl/1
mgN/1
mgK/1
mgCa/1
mgNa/1
mgMg/1
ueq/1
Expected
range
0.00-10.00
2.00-9.00
0.1-200.0
0.1-10.0
0.1-10.0
0.1-10.0
0.05-10.0
0.01-5.00
0.01-5.00
0.01-10.00
0.01-2.00
1.0-500.0
Suggested
precision
±.02"
±.01pH
±5%
±2%
±2%
±2%
±3%
±2%
±2%
±5%
±2%
±10%
Suggested
accuracy
+.02"
±.01pH
±5%
±2%
±2%
±2%
±2%
±2%
±2%
±3%
±2%
±5%
preferred since it offers minimal evaporation loss and a higher degree of reliability over the
tipping bucket rain gauge during intense storms.
3.4.3.2.3 jaH.  The pH is the measurement of the hydrogen ion activity (i.e., pH = -log  [H+]),
commonly  referred  to as  the  free acid content of  the solution.   The pH of  a typical  United
States wet  deposition  sample  ranges between 4.0  and  5.0 pH units.  Samples collected in more
arid regions may range as high as pH 8.0, whereas samples collected in the Northeastern United
States typically range from pH 3.5 to pH 4.5.  Precipitation pH measurements were not reported
prior to 1962.  Instead, methyl orange indicator solution (endpoint pH = 4.4) (Skoag and West,
1965) was  added to  assess  the  sample  acidity.   Cogbill and Likens  (1974)  and Granat (1972)
indirectly  calculated  and  reported  the pH's  of  precipitation samples  taken prior  to  1962.
Currently,  pH is electrometrically  determined with  a standard pH meter  in  conjunction with
either glass/reference calomel electrodes or a combination electrode at 25°C.   Three certified
buffer solutions are  used to  calibrate the  pH  meter/electrode  system in the pH range 3.50 to
7.50.  Direct pH measurements are made  on  a representative portion  of  the sample.   Measure-
ments  of pH  are dependent on  operator techniques  and  the  condition of  the electrode(s).
Galloway  et al.  (1979);  Ridder  (1978); and Tyree et al.  (1979)  noted  potential  sources of
                                             3-90

-------
errors  in  pH measurements.   With  proper care  and quality equipment, pH  results  with a pre-
cision  of  ± 0.02  pH  units and  accuracy of  ±  0.05 pH  units should  be obtainable (American
Public Health Association, 1971).
3.4.3.2.4   Conductivity.   The specific  conductance of a wet  deposition sample indicates the
capability  for  that sample to carry an electrical current.  Conductivity  is  related to the
total concentration  of dissolved ions,  is  directly determined on the sample with a standard
Wheatstone  bridge  in  conjunction with a calibrated conductivity cell at 25°C.   Wet deposition
sample conductivities  normally range from ca. 10 to 200 uS/cm.  Daily calibration of the con-
ductivity meter and  cell  with freshly prepared standard KC1 solution is required for accurate
measurements.  Operator techniques, the condition of the conductivity cell, and the quality of
the standard KC1 solutions determine the quality of data reported.  Under careful supervision,
conductivity measurements  with precision  and accuracy of  5  percent are obtainable (American
Public Health Association, 1971).
3.4.3.2.5   Acidity.   An  acidity measurement indicates  the  capacity  of the  wet deposition
sample  to  donate  protons from  both strong  and weak acids.   Numerous techniques  have been
reported  to measure  sample  acidity.   Each  technique  can  yield  a  slightly  different result
(Tyree, 1981).  Acidity  values from -200 ueg/liter to +200 ueg/liter are routinely reported.
The precision and  accuracy of any acidity  measurement dependents on the analytical technique
employed  and the  ability of the operator  to  standardize  and  titrate minute  quantities of
highly  diluted  strong  base.   Tyree (1981)  notes  that  the operator is the key to good acidity
measurements.   Both  the presence  of substances  such  as carbon  dioxide,  aluminum,  iron, and
ammonium, and the method endpoint pH influence the  results.
     3.4.3.2.5.1   p_H.   See  Section 3.4.3.2.3.   Depending on the actual  sample  pH,  acidity
based on  a pH  measurement yields either the  strong acid proton component (sample pH <4.5) or
the  strong acid component  plus  some undetermined contribution  from  the  weak  acid component
(sample pH  >4.5).   Acidity based on pH alone  is not considered conclusive.
     3.4.3.2.5.2  Titrimetric.  Various  titrimetric endpoint procedures  are available.  ,,
     Phenolphthalein Endpoint.   The rainwater  sample  is  titrated with standardized  0.01 to
0.001 N strong  base  (NaOH) to the phenolph\halein  endpoint (8.0 to 9.6  pH units per Skoag and
West, 1965).  Precision  and accuracy (+ 5  percent) depend on the ability of the operator to
standardize and deliver minute quantities of  dilute base and the  repeatability of the endpoint
color (American Public Health Association, 1971).
     WHO.   A prescribed  quantity of strong acid  (H^SO»)  is added to the sample,  lowering the
     immammr-r-                                            £  i\,
sample  pH to less than 4.0 and thus removing  the C02-  The  sample  is titrated against standar-
dized base until  the sample  pH, monitored via a pH meter,  reads  pH 5.6.  Precision (+ 0.02 pH
units)  and accuracy (+ 0.05 pH  units) are  dependent on  the  quality of  the pH meter and  stan-
dardized  base,  the condition  of  the electrodes, and operator  technique.
     AmericanPub!ic Health Association.   Strong acid is  added  to  lower  the  sample pH  below
4.0.   Hydrogen  peroxide is then added.   The  sample is boiled  to  eliminate COg, cooled to  room
temperature and titrated electrometrically with  strong  base  to  pH 8.3  (EPA method endpoint =
                                            3-91

-------
8,2).  The  U.S.  Environmental Protection Agency  (1979) reports a standard  deviation  of 1 to
2 mg  CaCO,/liter  and bias  of ca. +  20 percent  for  sample measurements in the  10  to 120 mg
CaCQ,/liter range.
     Likens (1972).  Nitrogen is bubbled through the sample to eliminate any CO, interference.
Samples are titrated  to pH 9.00 with standard base.  The accuracy of the pH meter is reported
to be + 0.03  pH  units (1972).  Following this  technique,  Hendry and Brezonik (1980) titrated
samples to pH 7.00 endpoint.
     Coulometric/Potentiometric.   The sample  is titrated with cathodically generated hydroxyl
ion  (OH )  [i.e.,   (-)  reference  electrode/test solution/glass  electrode (+)]  as outlined by
Liberti et  al.  (1972)  and Askne et al.  (1973).   Gran plots (Gran,  1952)  are  interpreted to
determine the strong  and weak acid contributions.  Liberti et al. (1972) report a + 5 percent
standard  derivation  and  0.1 mg/liter  ^SQ.  sensitivity  when  analyzing strong  acids.   The
Norwegian Institute for  Air Research (NAIR)  (1971) reported  a  2 to 5 ueq acidity/liter stan-
                                                -4   -5
dard  deviation  in rainwater  samples  having  10   -10    acidity  concentrations.   Askne et al.
(1973)  observed   "exact  agreement"  in  strong acid  analyses,  but  only  "reasonable-to-good
agreement" with samples  containing  various  concentrations of strong and weak acids.   Tyree et
al.  (1979) and Krupa  et al. (1976) also observed difficulties in determining strong/weak acid
contributions in rainwater samples using this technique.
     3.4.3.2.5.3   Ion Balance.  Granat  (1972) and Cogbill  and  Likens  (1974) reported a tech-
nique to  calculate sample  pH's  based  on  the total  ionic strength.   In this  technique, the
charge difference  in  favor of the anion concentration  is  related to the sample hydronium ion
concentration.  Possession  of accurate  analytical  data for  the  individual  principal  ions is
essential.  The overall  precision and accuracy of this technique is no better than the summa-
tion of precisions and accuracies of the analytical  methods  used to determine the individual
ionic species.  Tyree  et al.  (1979) states that  this technique could possibly be used to de-
termine the strong acid contribution in samples with observed pH 5.6 or below.
3.4.3.2.6  Sulfate (S0^.~).  Analytical procedure for sulfate analysis are described in Section
3.3.4.1.1.  Typical wet deposition samples contain 0.1 to 5.0 mg SO."/liter.
                         .1                                          '
3.4.3.2.7  Ammonium  (NH. ).  Ammonium  concentration of 0.1  to  1.0 mg N/liter  are normally
observed  in  wet  deposition  samples.   Two techniques (ion  selective electrode  and indophenol
colorimetry)  are   discussed  in  Section  3.3.4.2.   Manual Nesslerization  techniques  (American
Public Health Association,  1971)  are commonly used.  The  Nessler reagent is thoroughly mixed
with the  sample  (ca.  30 minutes).  The  characteristic  yellow color is photometrically deter-
mined (425 mu with 1 cm  cell path).  Analyses  of samples containing 0.2 mg N/liter typically
produce results with + 0.12 mg N/liter standard deviation and + 18 percent bias (U.S. Environ-
mental Protection Agency, 1979).
3.4.3.2.8  Nitrate (NO,  ).   The  rainwater sample's nitrate (normal concentration range 0.1 to
5.0 mg  N/liter)  is  quantitatively  reduced to  nitrite by  the  addition  of  hydrazine sulfate
                                             3-92

-------
(Kamphake et al., 1967)  or by passing the  sample  through a copper-cadmium column (U.S.  Envi-
ronmental Protection Agency, 1979d).   The addition of sulfanil amide and N-(l-napthyl)-ethylene
diamine  dihydrochloride  yields  a highly  colored azo dye measureable  colorimetrically  at 520
nm.    Automated  techniques  (U.S.  Environmental  Protection  Agency,  1979c)  minimize  operator
error  and  increase  ,the  sample  throughput.   A  second analysis without  the  nitrate reduction
step is  required  to  correct for sample nitrite  concentration.   Precision and accuracy of + 5
percent are expected with  samples  above 1  mg N  per liter.   Butler et al. (1978) and Tyree et
al.  (1979) report comparable  sensitivity,  precision, and accuracy using an automated 1C tech-
nique (see Section 3.3.4.2.2.4).
3.4.3.2.9  Chloride  (Cl ).   Various manual and automated  procedures  are used  to determine
chloride in rainwater in the concentration range 0.1 to 10.0 mg Cl /liter with precision ca. +
0.2 mg Cl/liter.  The  WMO  method adds mercuric  nitrate and diphenylcarbazone-bromophenol blue
to the sample forming  mercuric  chloride.   The excess  mercury complexes with the indicator to
form  a blue-violet  dye measured photometrically at  525  nm.   Zall et al.  (1956)  displace the
thiocyanate ion (SCN )  in Hg(SCN)? and form HgCl~.   In the presence of excess iron, the highly
colored  dye   [Fe  (SCN)+]  is formed and  can  be  photometrically  measured  at  460 |jm.   The
automated ferricyanide  procedure (U.S.  Environmental  Protection Agency,  1979d)  is preferred
over the manual methods  since operator/standard solution errors  are  minimized.   Automated 1C
techniques (Butler et  al.,  1978; Tyree et al.,  1979 and Section 3.3.4.1.1.4) yield comparable
results.
3.4.3.2.10   Fluoride (F ).   Fluoride  in  wet  deposition  (range  0.01  to 0.1'mg  F/liter)  is
generally determined by  the ion selective electrode technique.  The condition of the fluoride
ion selective electrode is critical.   Analysis  of synthetic samples containing 0.85 mg F/liter
yielded  results with a 3.6 percent relative standard deviation and 0.7 percent relative error
(American Public  Health Association,  1971).   Automated  1C  techniques  (Butler  et al. ,  1978;
Tyree et al., 1979;  and Section 3,3.4.1.1.4) yield similar results.
3.4.3.2.11 Trace Metals.   Techniques  used to determine trace metal concentration in rainwater
are described in  Section 3.3.4.3.   Observed metal  concentration  ranges generally approximate
the lower detection limit for flame atomic  absorption metal analysis.
3.4.4  Interlaboratory Comparisons
     WMO (1975) instituted an international inter!aboratory program to describe the quality of
the wet  deposition  chemistry data being reported by the various WMO laboratories.  Participa-
tion  in  this program  is voluntary.   The results of  three  comparisons  on synthetic rainwater
samples  (WMO,, 1976; Thompson,  1978; WMO,  1980) have been  reported previously.   The  United
States Department of Energy (DOE) sponsored a similar round  robin (Battelle Pacific Northwest
Laboratories, 1979)  on  both simulated and  composited rainwater samples.  Tyree et al.  (1979)
and Tyree (1981) discuss the WMO and DOE results.
     Three physical  analyses  (pH,  conductivity, and  acidity) are most frequently reported and
compared.  Conductivity  and  pH  data have  been  extracted from the three WMO reports  and are

                                             3-93

-------
                                   TABLE  3-7.   RESULTS OF HMO  INTERCOMPARISONS OH SYNTHETIC PRECIPITATION SAMPLES
pH (pH Units)
Sample _k
Session type
1976



1978



1980



A
B
C
D
A
B
C
D
71
72
,73
"74
x
5.45
5.53
1.53
5.56
5.66
5.77
5.60
5.65
4.21
4.02
5.58
3.91


.74
.76
.52
.41
.49
.51
.44
.39
.16
.11
.19
.16
.j
Nd
17
18
17
18
25
25
25
25
26
26
26
24

High X
6.40
7.22
6.20
6.10
6.54
7.07
6.58
6.64
4.55
4.17
6.02
4.30

Sample
Low x Session type
3.65 1976
3.85
4.10
4.20
4.61 1978
4.65
4.54
4.79
3.80 1980
3.72
5.17
3,60
A
B
C
D
A
B
C
0
71
72
73
a74

X
6.9
17.3
56.3
109.3
6.7
17.9
52.4
103.6
29.1
80.4
195.9
'64.4
Conductivity {yS/cm)

°x
4.3
4.9
4.8
8.1
3.0
8.8
12.0
23.9
4.0
9.1
21.6
8.0

N
17
18
17
18
25
25
25
25
25
25
25
23


Acidity (peq/1)
Sample
High x Low x Session type x 0X N
22.0
28.0
62.0
119.0
19.0
57.0
67.5
132.0
37.3
95.6
224.0
81.5
3.9 1976
20.0
44.0
84.0
3.9 1978
9.0
19.1
37.9
18.3 1980
57.3
132.0
48.0
NOT REPORTED BY WHO



NOT REPORTED BY WHO



71 70.0 36.7 22
72 106.3 43.6 22
73 25.3 63.7 18
a74 10.9 16.1 22

High x Low x








206.0 18.9
260.6 29.2
202.0 -37.0
68.0 4.7
 Sample contains only
b  = sample mean
 a  = standard deviation =

/
        _
   = number of laboratories.

-------
provided in Table 3-7.  WMO did not summarize the acidity measurement results in the first and
second analysis sessions.  The synthetic samples used in the first two sessions contained only
weak acids.  The laboratory results indicated that the WMO laboratories, as a whole, could not
perform acidity  measurements on  samples  containing  only  weak acids.   Four  new samples were
used in the  third  WMO analysis session.  Three of the third analysis session samples (samples
71,  72,  73)  contained  both  weak  and strong  acids.   Sample  type  74  contained  only  strong
sulfuric acid.  The results of the acidity measurements reported in the third analysis session
have  also  been  extracted.    Table  3-8  lists  the among  laboratories percent  coefficient  of
variation  (%  cv)  by  session for'those chemical  analyses  routinely  performed  by the  WMO
participants, where
                      cv =
VI(di"5)2 '       x 100%                  (3-3)
   N-l
Improvement in  the  analysis for a given  constituent  by the WMO laboratories, as a whole,  is
indicated by  decreasing  percent cv.  from session 1 to session 3.  In general, the WMO labora-
tories showed improvement from session 1 to session 3.
     WMO and  DOE  intercomparison results highlight the  difficulties  encountered -in analyzing
dilute precipitation samples.  When comparing data from various studies, the data analyst must
include the appropriate biases resulting from the laboratory's sampling and storage techniques
as well as the ability of the laboratory to perform chemical analyses on the sample.

                    TABLE 3-8.  COEFFICIENTS OF VARIATION OF WMO INTERCOMPARISONS ON
                                 SYNTHETIC PRECIPITATION SAMPLES
                      Among laboratories percent coefficient of variation (% cv.)
                                           % cv by session
         Constituent            1976             1978             1980
PH
Conductivity
SO.
NHA
NO,
Cl3
Ca
K
Ma
Na
10
15
13
38
79
58
27
32
21
29
8
22
24
32
64
24
25
30
8
27
3.4
12
34
33
74
25
25
22
99
19
                                            3-95

-------
3.5  REFERENCES

Abbas,  R.  and  R.   L.  Tanner.   Continuous  determination  of  gaseous ammonia  in  the ambient,
     atmosphere using fluorescence derivatization.  Atmos. Environ. 15:277-81,  1981.

Adams, F. C. , and R. E. Van Grieken.  Absorption correction for X-ray fluorescence analysis of
     aerosol  loaded filters.  Anal. Chem. 47:1767-1773, 1975.

Ahern, A. E.   Trace Analysis by Mass Spectrometry,  Academic Press, New York, NY, 1972.

Ahlquist, N. C., and R.  J. Charlson.   A new instrument for evaluating  the visual  quality of
     air.  J. Air Pollut. Control Assoc. 17:467-469, 1967.

Akiyama, T.   Direct titration of sulfate.  Kyoto Yakka Daigaku Gakuho 5:48-50,  1957.

American Industrial  Hygiene Association.   Guide for respirable  dust sampling. Am.  Ind. Hyg.
     Assoc,  J.  31:133, 1970.

American Public Health Association.   Standard Methods for the Examination  of Water and Waste-
     water,  13th ed.  American Public Health Association, Washington, DC, 1971.

American Society  for Testing and  Materials.   Standard method for collection  and analysis of
     dustfall (settleable  particulates):  ASTM D 1739-70.  In:  Annual Book of  ASTM Standards.
     Part 26:   Gaseous Fuels;  Coal and Coke;  Atmospheric Analysis.   American Society for
     Testing and Materials, Philadelphia, PA, 1981a.  pp. 521-525.

American Society  for Testing and  Materials.   Standard  method for particulate matter in the
     atmosphere (optical density of filtered-material):  ASTM  D  1704 78.    In:   Annual Book of
     ASTM Standards.   Part 26:   Gaseous Fuels; Coal and Coke; Atmospheric  Analysis.  American
     Society for Testing and Materials, Philadelphia, PA, 1981b.   pp. 513-520.

American Society for Testing and Materials.  Standard test methods  for sulfate  ion in
     water:    ASTM   D  516-80.   Method  C  -  volumetric   method.    In:   Annual  Book  of ASTM
     Standards.   Part  31:   Water.   American Society for  Testing  and Materials, Philadelphia,
     PA, 1981c.   pp. 602-605.

Anderson, D. E.  Continuous Measurement of  pH  Using  An  Automatic  Precipitation Collector and
     Analyzer.   University of Maryland, Publication 78-178, College Park, MD, May 1978.

Appel, B. R., and Y. Tokiwa.  Atmospheric particulate nitrate  sampling errors due to reactions
     with particulate and gaseous strong acids.  Atmos. Environ.  15:1087-1089,  1981.

Appel, B. R., E.  M. Hoffer, E.  L.  Kothny, and S. M. Wall.  Interference in 2,4-xylenol proce-
     dure for nitrate  determination in atmospheric aerosols.   Environ.  Sci.  Techno!. 11:189,
     1977a.

Appel, B. R., E. L. Kothny, E. M. Hoffer, and J. J. Wesolowski. Comparison  of Wet Chemical and
     Instrumental  Methods   for  Measuring  Airborne  Sulfate.   EPA-600/7-77-128,' U.S.  Environ-
     mental  Protection Agency, Research Triangle Park, NC, November 1977.

Appel, B. R., S.  M. Wall,  Y. Tokiwa,  and M. Haik.  Interference effects in sampling particu-
     late nitrate in ambient air.  Atmos. Environ. 13:319-325, 1979.

Appel,  B.  R.,  S.  M.  Wall,  Y.  Tokiwa,  and M.  Haik.   Simultaneous  nitric acid,  particulate
     nitrate  and  acidity  measurements  in ambient  air.   Atmos.  Environ.   14:549-554,   1980.
                                             3-96

-------
Appel, B.  R. ,  Y.  Tokiwa, S. M. Wall, E. M. Hoffer, M. Haik, and J. J. Wesolowski.  Effects of
     Environmental Variables  and Sampling  Media on  the Collection  of  Atmospheric Sulfates.
     Final Report  to  the California Air Resources Board, Contract #5-1032, Air and Industrial
     Hygiene Laboratories, Calif. Dept. of Health Services.  Berkeley, CA, 1978.

Appel, B.  R.,  E.  M.  Hoffer, M,  Haik,  W.  Wehrmeister, E. L. Kothny,  and J. J. Wesolowski, Im-
     provement  and  Evaluation  of  Methods  for  Sulfate  Analysis,   EPA-600/4-79-028,  U.S.
     Environmental Protection Agency, Research Triangle  Park, NC, April 1979a.

Appel, B. R., J. J. Wesolowski, E, Hoffer, S, Twiss, S.  Wall, S. G. Chang, and V. T. Novak. An
     intermethod comparison of X-ray photoelectron spectroscopy (ESCA) analysis of atmospheric
     particulate matter.  Int.  J. Environ. Anal. Chem. 4:169-181, 1976.

Askne, C., C. Brosset, and M.  Fern.   Determination of the Proton-Donating  Property of Airborne
     Particles. NL Publication B157, Goteborg, Sweden, 1973.

Axelrod,  H.  D. , and S.  G.  Hansen.   Filter sampling method  for  atmospheric  sulfur dioxide at
     background concentrations.  Anal. Chem. 47:2460-2461,  1975.

Axelrod,  H.  D. ,  J.  E.  Bonelli,  and J.  P.  Lodge,  Jr.   Fluorimetric determination  of trace
     nitrates.   Anal.  Chim. Acta 51:21-24, 1970.

Bacon, J.  C.,   M.  E.  Edelson,  and  G.  J.  Small.  Fluorescence  line narrowing spectrometry in
     organic glass containing  PPB  levels of  polycyclic aromatic  hydrocarbons.   Anal. Chem.
     50:1394-1397, 1978.                                   -

Bailey, D. L. R. , and P. Clayton.  The Measurement of Suspended Particulate and Carbon  Concen-
     trations in the Atmosphere Using Standard Smoke Shade Methods.   Warren Springs Laboratory
     Report LR325(AP), Stevenage, Hertfordshire, England, 1980.

Ball, D. J., and R. Hume.  The relative importance of vehicular and domestic  emissions  of dark
     smoke in  Greater  London  in the mid-1970's, the significance of  smoke shade measurements,
     and  an  explanation  of the  relationship  of smoke  shade  to gravimetric measurements of
     particulate.  Atmos. Environ. 11:1065-1073, 1977.

Balzer,  J.   L.   Inertia!  and  gravitational  collectors.   In;   Air  Sampling Instruments for
     Evaluation  of Atmospheric  Contaminants,  4th  ed. ,  American  Conference on Governmental
     Industrial Hygienists, Cincinnati, OH, 1972.  pp. 16-17.

Bakas, E.  Fast titrimetric method for the determination of  sulfate ions.  Magy. Kern.
      Foly. 62:135-139,  1956.

Barnard, W. F.  , ed.  Community Health Environmental Surveillance Studies (CHESS) Air Pollution
     Monitoring  Handbook:   Manual  Methods.   EPA-600/1-76-011,  U.S.  Environmental Protection
     Agency, Research Triangle Park, NC, January 1976.

Barnes, R. A.   Duplicate measurements of low concentrations  of smoke  and sulphur dioxide using
     two  "National  Survey" samplers  with a  common inlet.  Atmos.  Environ.  7:901-904, 1973.

Barrett,  E.,  and  G.  Brodin.  The acidity of  Scandinavian  precipitation.   Tellus 7(2):25-57,
      1955.

Barrett,  W.  J. ,  H. C.  Miller, J. E.  Smith,  Jr.,  and C. H. Gwin.   Development of a Portable
     Device  to  Collect   Sulfuric   Acid   Aerosols.    EPA-600/2-77-027,  U.S.  Environmental
     Protection Agency,  Research Triangle Park, NC, February 1977,

Barringer  Research Limited.   A  report to  Department  of  Health,  Education,  and Welfare of
     Optical Measurements  of  S0« and NO, Air Pollution  Using Barringer Correlation Spectrome-
     ters.   Technical  Report  TR-69-112r,  Barringer  Research,  Ltd., Rexdale,  Ont.,   Canada,
     December  1969.
                                             3-97

-------
Barton, S. C. , and H. G. McAdie,  A specific method for the automatic determination of ambient
     HjSCL aerosol.   In:   Proceedings  of the 2nd  International  Clean  Air Congress, Interna-
     tional Union  of Air Pollution Prevention Associations,  Washington,  D.C. ,  December 6-11,
     1970,  H. M.  Englund  and  W.  T.  Berry,  eds, , Academic  Press, New  York,  NY, 1971.  pp.
     379-382.

Barton,  S. C. ,  and  H.  G.  McAdie.    An automated  instrument  for monitoring  ambient  H«SO,
     aerosol.   In:   Proceedings of the Third  International  Clean Air Congress,  Internationa*
     Union of  Air  Pollution Prevention Associations, Dusseldorf , Germany, October 8-12, 3973.
     Verein Deutscher Ingenieure, Dusseldorf, Germany, 1973.  pp. C25-C27.

Battelle  Pacific  Northwest  Laboratories.   MAP3S Precipitation  Chemistry Network.  PNL-2829.
     U.S. Department of Energy, Washington, DC,  1979.

Baumgardner,  R. E.  , W. A. McClenny, and R. K. Stevens.  Optimized Chemi luminescence System for
     Measuring Atmospheric  Ammonia.    EPA-600/2-79-028,  U.S.  Environmental Protection Agency,
     Research Triangle Park, NC, 1979.

Behrmann,  J.  J.    Design and Testing  of a  Continuing  Monitor  of  Atmospheric  Precipitation.
     Master's  Thesis,  North  Carolina State University,  Department of  Chemical  Engineering,
     Raleigh, NC, 1975.

Bergman,  F.  J. ,  and  M.  C,   Sharp.   Measurement  of  Atmospheric Sulfates:   Evaluation  of the
     Methylthymol  Blue  Method  Using  an Autoanalyzer II.   EPA  Contract  No.  68-02-1728, U.S.
     Environmental  Protection Agency, Research Triangle Park, NC, March,  1979.

Bertolacini,   R.  J. ,  and J.  E.  Barney.   Colorimetric  determination  of  sulfate with  barium
     chloranilate.   Anal. Chem. 29:281, 1957.

Beuttell.  R.  G. , and A.  W.  Brewer.   Instruments for the measurement of  the  visual  range.  J.
     Sci. Instrum.  Phys. Ind. 26:357-359, 1949.

Bigg,  E.  K. ,  A.  Ono,  and  J.  A.  Williams.  Chemical  tests  for individual  submicron aerosol
     particles.  Atmos. Environ. 8:1-13, 1974.

Blacker, J. H. ,  R.  G.  Confer, and R. S. Brief.   Evaluation of the  reference method for deter-
     mination  of  sulfur dioxide  in  the  atmosphere (pararosaniline method).   J. Air Pollut.
     Control  Assoc. 23:525-527, 1973.

Blanco, A. J. , and G.  B. Hoidale.  Microspectrophotometric technique for obtaining the intra-
     red  spectrum  of microgram  quantities  of atmospheric  dust.   Atmos.  Environ.  2:327-330,
     1968.

Blanco, A. J. , and R. G. Mclntyre.   An infra-red  spectroscopic  view  of atmospheric particu-
     lates over El  Paso, Texas.  Atmos. Environ.   6:557-562, 1972.

Bokhoven,  C. ,  and  H.  J.  Niessin.  The continuous  monitoring of traces  of SO-  in air  on the
     basis of  discolouration of  the starch-iodine reagent with  prior  eliminations  of inter-
     fering compounds.  Air Water Pollut. 10:233-243, 1966.

Bolleter,  W.  T. ,  C.  J.  Bushman, and P.  W.  Tidwell.   Spectrophotometric  determination  of
     ammonia as indophenol.   Anal. Chem. 33:592-594, 1961.
Bostro'm, C.  E.   The absorption of low concentrations (pphm) of hydrogen sulfide in a       ^
     suspension  as  studied  by an isotopic tracer method.  Air Water Pollut. 10:435-441, 19i5B.

Boumans, P.  W.   J.  M. ,  and  F.  J.  Le Boer.  Studies  of an inductively-coupled high-frequency
     argon plasma for optical  emission spectrometry -  II.   Compromise conditions for simul-
     taneous multi-element analysis.  Spectrochem.  Acta. 308:309-334,  1975.

                                             3-98

-------
Bowden, S.  R,   Improved lead dioxide  method of assessing sulphurous  pollution of the atmos-
     phere.  Air Water Pollut.  8:101, 1964.

Bradway,  R.  M., F.  A.  Record,  and  T.  G.  Pace.  Application of  polarizing microscopy to the
     characterization of  ambient suspended  particulate.   In:   Proceedings Annual Meeting of
     the  Federation  of  Analytical   Chemists and  Spectroscopists  Society,  Philadelphia, PA,
,  ;   1976. Available from Standard Oil Company, Cleveland, OH, 1976.

Braman, R.  S.,  and T.  J.  Shelley.   Tungsten VI Oxide for Pre-Concentration and Determination
     of Gases and  Particulate Ammonia and Nitric Acid in Ambient Air.  EPA-600/7-80-179, U.S.
     Environmental Protection Agency, Research Triangle Park, NC, 1980.

Braman, R. S., D. L.  Johnson, C. C.  Foreback, J. M. Ammons, and J.  L. Bricker.  Separation and
     determination  of  nanogram  amounts  of  inorganic arsenic  and  methylarsenic compounds.
     Anal. Chem. 49(4):621-625,  1977.

British Standards Institution.   Methods for  the Measurement of Air  Pollution.   Part 3:  Deter-
     mination of Sulphur  Dioxide.   B.S. 1747, British Standards Institution, London,  England,
     1963.

British Standards Institution.   Methods for-the Measurement of Air  Pollution.   Part 2:  Deter-
     mination of Concentration of Suspended Matter.  B.S. 1747, British Standards  Institution,
     London, England, 1964.

Bromberg,  S. M.,  G.  G.  Akland,  and  J.  C.  Puzak.   Survey of  laboratory performance.   Analysis
     of simulated  ambient S00  bubbler samples.  J. Air Pollut. Control Assoc.  24:1073-6  1973,
  ,   1974.                   l                                                   ~

Bromberg,  S. M.,  R.  L.  Lampe,  and 8.  I,  Bennett.   Summary of Audit Performance:  Measurement
     of S02, NQ2, CO, Sulfate, Nitrate, Lead, Hi-Vol Flow Rate - 1977.  EPA-600/4-79-014, U.S.
     Environmental Protection Agency, Research Triangle Park, NC, February  1979.

Bromberg,  S. M.»  R.  L.  Lampe,  and B.  I.  Bennett.   Summary of Audit Performance:  Measurement
     of SO,,  NQp, Sulfate,  Nitrate, Lead,  Hi-Vol  Flow Rate -  1978.   EPA-600/4-80-017, U.S.
     Environmental Protection Agency, Research Triangle Park, NC, June 1980,

Brosset,  C., and M.  Ferm.   Man-made airborne  acidity and its determination.   Atmos.  Environ.
     12:909-916, 1978.

Brosset,  C.,  K. Andreasson,  and M.   Ferm.   The nature and possible  origin of acid particles
     observed at the Swedish west coast.  Atmos. Environ. 9:631-642, 1975.

Butler, F. E.,  R. H. Jungers, L. F.   Porter,  A.  E. Riley, and  F. J.  Toth.  Analysis of  air par-
     ticulates  by  ion  chromotography:   comparison with accepted  methods.   In:   Ion Chromato-
     graphic Analysis of Environmental Pollutants.  E. Sawicki, J.  Mulik, and E. Wittgenstein,
     eds., Ann Arbor Science, Ann Arbor, MI, 1978.- pp. 65-76.

Cadle, S.  H., P. J. Groblicki, and D. P. Stroup.  An Automated Carbon Analyzer  for Particulate
     Samples.   Research Publication  GMR-3306,  ENV f77,  General  Motors Research Laboratories,
     Warren, MI, 1980.

Calvert,  S., and W. Workman.  Estimation of  efficiency for bubbler-type gas absorber.  Talanta
     4:89-100,  1960.

Camp,  D.  C. ,  A. L. Van Lehn, and B.,W. Loo.  Interconoparison of Samplers Used  in  the  Determi-
     nation  of  Aerosol  Composition.    EPA-600/7-78-118,  U.S.  Environmental  Protection Agency,
     Research Triangle Park, NC, July 1978.
                                             3-99

-------
Caplan,  K.  J.,  L,  J.  Doemeny,  and S.  D. Sorenson.   Performance characteristics of the 10mm
     cyclone  respirable mass  sampler:   Part  II.  Am.  Ind.  Hyg.  Assoc.  J.  38:162-173, 1977.

Cautreels, W.,  and  K.  Van Cauwenberghe.   Determination of organic compounds  in  airborne par-
     ticulate  matter by  gas/chromatography  - mass  spectrometry.   Atmos.  Environ,,  10:447,
     1976.

Cermak, J.  E,   Applications  of fluid mechanics  to wind engineering,  Freeman  Scholar Lecture.
     Presented  at Winter  Annual  Meeting of the American Society  of Mechanical  Engineers.  New
     York, November 17, 1974.

Chand, R.,  and  R.  V. Marcote.  Evaluation of  portable  electrochemical  monitors and  associated
     stack sampling  for  stationary source' monitoring.  Presented at  the  68th  National  Meeting
     of the  American Institute of Chemical Engineers,  Houston, TX, February 28-March 4, 1971.

Charlson, R. J.,  N.  C.  Ahlquist,  and H.  Horvath.   On  the generality of  correlation of atmos-
     pheric  aerosol  mass concentration  and light  scatter.   Atmos.  Environ.   2:455-464, 1968.

Charlson, R. J.,  A.  P.  Waggoner,  and J. F. Thielke.  Visibility  Protection for Class I Areas:
     the  Technical   Basis.   CEQ517714498, Council  on  Environmental  Quality,  Washington,  DC,
     August 1978.

Chow, T.  J., and M.  S.  Johnstone.   Determination of  nitrate  in seawater.   Anal.  Chim. Acta
     27:441-446, 1962.

Coffer, J. W.,  C.  McJilton,  and R.  J.  Charlson.   SCL  oxidation  to sulfate due to particulate
     matter  on a  high  volume air sample filter.   Paper no.  ANAL  102.   Presented at 167th
     National Meeting, American Chemical Society, Los Angeles, CA, April  1-5,  1974.

Cogbill,  C.  V.,  and G.  E.   Likens.   Acid precipitation  in the northeastern United  States.
     Water Resour. Res.  10:1133-1137, 1974.

Collison, R.  C., and J.  F.   Mensching.   Lysimeter  investigations.   II.   Composition  of rain
     water at Geneva, N.Y. for a ten-year period.  N.Y. Agr.  Expt. Sta.,  Tech.  Bull. 193:1-19,
     1932.

Colovos, G., M.  R.  Panesar,  and E. P. Parry.   Linearizing the calibration curve  in  determina-
     tion of sulfate by the methylthymol blue  method.   Anal.  Chem. 48:1693-1696,  1976.

Colovos, G., E.  P.  Parry, and C.  E.  Rodes.    Comparison of the degradation of ammonium ion on
     high volume  glass  fiber filters and  on  membrane filters.    In:   The Los  Angeles'Catalyst
     Study  Symposium,  U.S.   Environmental  Protection Agency, Raleigh, NC,  April   12-13, 1977.
     EPA-600/4-77-034, U.S.  Environmental  Protection Agency, Research Triangle Park, NC, June
     1977.  pp. 265-279.

Commins, B.  T.,  and R.  E. Waller.   Observations  from  a ten  year study of pollution at a site
     in the city of London.   Atmos. Environ. 1:49-68, 1967.

Conner,  W.  D.   An  inertial-type  particle  separator  for  collecting  large samples.   J.  Air
     Pollut. Control Assoc.  16:35-38, 1966.

Cooper, H. B.  H.,  J. A. Lopez, and J. M. Demo.  Chemical composition of  acid  precipitation in
     central Texas.   In;   First  International Symposium on  Acid  Precipitation and  the Forest
     Ecosystem, U.S.  Department  of Agriculture and  Others.   Columbus, Ohio,  May  12-15, 1975.
     Water, Air, Soil Pollut. 6:351-359, 1976.

Corn,  M.,  F.  Stein, and  N. Esmen.  The Shape  of Atmospheric  Particles:   Progress  Report.
     Public Health  Service Research Grant AP00431-02,  University of  Pittsburgh,  Air Pollution
     Laboratory, Pittsburgh,  PA,  November 1967.

                                             3-100

-------
Courtney, W. J.,  S.  Rheingrover, J. Pilotte,  H.  C.  Kaufmann, T. A. Cahill, and J. W. Nelson.
     Continuous  observation  of particulates  during  the General  Motors  Sulfate Dispersion
     Experiment.  J.  Air Pollut. Control Assoc.  28:224, 1978,

Coutant,  R,  W.   Effect  of  environmental  variables  on  collection  of  atmospheric  sulfate.
     Environ. Sci. Techno!. 11:873-878, 1977,

Covert, D. S. ,  A.  P.  Waggoner,  R.  E. Weiss,  N.  C. Ahlquist, and R. J. Charlson.  Atmospheric
     aerosols,  humidity, and visibility.  In:  The Character and Origins  of Smog Aerosols.  G,
     M. Hidy,  P.  K.  Mueller,  D. Grosjean,  B.  R.  Appel, and J.  J.  Wesolowski, eds. , J. Wiley
     and Sons,  New York, NY, 1980.   pp.  559-581.
Crider, W. L.   Hydrogen flame emission spectrophotometry  in monitoring air for sulfur dioxide
     and sulfuric acid aerosol.  Anal.  Chem. 37:1770-1773, 1965.

Crider, W. L.,  N. P.  Barkley,  M. J, Knott, and R. W. Slater, Jr.  Hydrogen flame chemilumines-
     cence  detector  for  sulfate  in aqueous solutions.   Anal.  Chim. Acta  47:237-241, 1969.

Cunningham,  P.  T.,  and  S. A.  Johnson.  "Spectroscopic observation of  acid sulfate in atmos-
     pheric particulate samples.  Science (Washington, D.C.) 191:77-79, 1976.

Cunningham,  P.   T. , S,  A.  Johnson, and R. T. Yang.  Variations in chemistry of airborne parti-
     culate material  with particle size and time.  Environ. Sci. Techno!. 8:131, 1974.

Daley,  P.  S.,  and D. A.  Lundgren.   The performance of piezoelectric  crystal  sensors used to
     determine  aerosol  mass concentrations.   Presented at American Industrial Hygiene Confer-
     ence, Miami, May 12, 1974.

Dasgupta,  P.  K., and  K.  B.  DeCesare.   Stability of  sulfur dioxide  in formaldehyde and its
     anomalous behavior in tetrachloromercurate (II).  Anal. Chem., in press 1981.

Dasgupta,  P. K.,  K.  DeCesare, and J. C.  Ullrey.   Determination of atmospheric sulfur dioxide
     without tetrachloromercurate  (II)  and the mechanism  of the Schiff reaction.  Anal. Chem.
     52:1912-22, 1980.

Delumyea,  R. G.,  L,  C.  Chu, and E. S.  Macias,  Determination of elemental carbon  component of
     soot in ambient aerosol samples,  Atmos. Environ. 14, 647-652, 1980.

Department of Scientific and Industrial Research.  The Investigation of Atmospheric Pollution,
     1931-1932.  18th Report.   H. M. S. 0., London, 1933.

De  Veer,  S,  M.,  H.  J.  Brouwer, N. V.  Philips, and  H.   Zeedijk.  A remotely controlled SO,
     monitor for  operation in air pollution  networks.   Presented at the 62nd Annual Meeting,
     Air Pollution Control Association, New York, NY, June 1969.  Paper no. 69-6.

DiMartini,  R.   Determination  of nitrogen dioxide  and nitric oxide  in  the parts per million
     range in  flowing  gaseous mixtures by means of the nitrate specific-ion electrode.  Anal.
     Chem. 42:1102-1105, 1970.

Dionex  Corporation,   Dionex Analytical  Ion Chromatographs.  Technical Brochure, Dionex Corp.,
     Palo Alto, California, 1975.

Dochinger,  L.  S., and R.  A. Seliga, Eds.  Proceedings of  the First International  Symposium on
     Acid  Precipitation  and the Forest Ecosystem,  U.S.  Department of Agriculture  and Others,
     Columbus,  Ohio, May  12-15,  1975.   NEFES/77-1,  U.S. Department  of Agriculture, Forest
     Service, Upper Darby, PA, August 1976.

Driscoll,  J. M. ,  A.  W. Berger,  J. M. Becker, J. T. Funkhouser, and J. R. Valentine.  Determi-
     nation  of  oxides  of nitrogen  in  combustion effluents  with  a  nitrate  ion  selective
     electrode.  J. Air Pollut.  Control Assoc. 22:119-122, 1972.

                                             3-101

-------
Dubols, L.,  C.  J.  Baker, T. Telchman, A. Zdrojewski, and J.  L. Monkman.  The  determination of
     sulphuric acid in air:  a specific method,  Mikrochim.-Acta  2:269-279,  1969.

Dubois, L.,  R.  S.  Thomas, T. Teichman,  and J. L. Monkman.   A general  method of  analysis for
     high volume air samples.  I.  Sulphate and sulphuric acid.   Mikrochim.  Acta.  6:1268-1275,
     1969.

Dzubay, T. G.,  ed.  X-ray Fluorescence Analysis  of Environmental  Samples.   Ann Arbor  Science,
     Ann Arbor, MI, 1977.

                                                                  14
Dzubay, T.  G.  Aerosol  acidity determination by reaction  with    C  labelled amine.   Atmos.
     Environ. 13:1209, 1979.

Dzubay, T.  G., and  R.  K.  Stevens.   Ambient air  analyses  with dichotomous sampler and  X-ray
     fluorescence spectrometer.  Environ. Sci. Tech. 9:663-668, 1975.

Dzubay. T. G.,  H.  L.  Rook,  and  R.  K.  Stevens.  A chemiluminescent approach to measurement of
     strong  acid  aerosols.   In:   Analytical   Methods  Applied to  Air Pollution Measurements.
     R. K. Stevens and  W.  F. Herget, eds., Ann Arbor Science Publishers, Inc., Ann  Arbor, MI,
     1974.  pp. 71-83.

Dzubay, T. G. ,  L.  E.  Mines, and R.  K.  Stevens.   Particle bounce errors  in  cascade  impactors.
     Atmos. Environ.  10:229-234, 1976.

Dzubay, T.  G., P. J.  Lamothe,  and  H.  Yoshuda.   Polymer films  as  calibration standards for
     X-ray fluorescence analysis.  Adv. X-Ray  Anal. 20:411,  1977.

Eagan, M.   L.,  and  L.  Dubois.  The determination of ammonium ion  in airborne particulates with
     selective electrodes.  Anal. Chim. Acta 70:157-167, 1974.

Edwards, Joseph Q.,  A Comparison of the Bristish  Smoke  Shade Method  and  the Integrating  Plate
     Method,  Masters Thesis,  .Department   of  Civil  Engineering,  University  of Washington,
     Seattle, WA, 1980.

Eggleton,   A.  E.,  and  D.  H. Atkins.   Results of the Tees-side  investigation.   U.K. Atomic
     Energy Authority, Research Group Report AERE-R6983, 1972.

Fassel, V.   Quantitative  elemental  analyses by plasma  emission spectroscopy.   Science (Wash-
     ington, D.C.) 202:183-191, 1978.

Faxvog, F. R. , and D. M. Roessler.   Optoacoustic Measurements of  Diesel Particulate  Emissions.
     Research  Publication  GMR-2987,  General  Motors Research Laboratories,  Warren,  MI,  1979.

Perm,  M.   Method  for  determination  of atmospheric  ammonia.  Atmos.  Environ.  13:1385-1393,
     1979.                                                                       ~~

Forrest, J.,  and  L.  Newman.   Ambient air monitoring  for sulfur  compounds—a  critical  review.
     J. Air Pollut. Control Assoc.  23:761-768, 1973.

Forrest, J., and L. Newman.  Sampling and analysis of atmospheric sulfur  compounds for isotope
     ratio studies.  Atmos. Environ.  7:561-573, 1973a.

Forrest, J.,  and  L.  Newman.   Silver-110 microgram sulfate analysis for the  short  time resolu-
     tion of ambient levels of sulfur aerosols.  Anal. Chem.  49:1579, 1977.
                                             3-102

-------
Forrest, J., R.  L.  Tanner, D. Spandau,  T.  D'Qttavio, and L. Newman.  Determination of atmos-
     pheric  HMO., with NaCl-impregnated  filters  at high  volume  flow  rates.   In:   Current
     Methods to  Measure Atmospheric  Nitric Acid and Nitrate Artifacts.   R.  K. .Stevens, ed.,
     EPA-600/2-79-051, U.S. Environmental Protection  Agency, Research Triangle  Park, NC, March
     1979.   pp. 51-61.

Forrest, J. , R.  L.  Tanner, D. Spandau,  T.  D'Qttavio, and L.  Newman.   Determination of total
     inorganic  nitrate utilizing  collection  of nitric  acid  on  sodium  chloride-impregnated
     filters.  Atmos. Environ., 14:137-144, 1980.

Foster,  J.   E.  ,  and  G.  H. Beatty.   Inter!aboratory Cooperative  Study  of the Precision and
     Accuracy of the Measurement of Sulfur  Dioxide Content in the Atmosphere Using ASTM Method
     D2914.   ASTM  Data  Series  Publication  DS-55-S1,  American   Society  for  Testing  and
     Materials, Philadelphia, PA, 1974.
Fox, D. L., and H. E. Jeffries.  Air pollution.  Anal. Chem. 51(5):27R, 1979.

Fox, M.  A., and  St.  W.  Staley.   Determination  of polycyclic  aromatic  hydrocarbons in atmos-
     pheric  particulate matter by  high pressure  liquid  chromatography  coupled with fluores-
     cence techniques.  Anal. Chem. 48:992, 1976.

Frank,  E.  R. ,  and J.  P.  Lodge,  Jr.  Morphological  identification  of airborne particles with
     the electron microscope.  J. Microsc.  6:449, 1967.
                                                                                         f
Friedlander, S.  K.    Smoke,  Dust,  and  Haze.   Wiley  Interscience Publication,  New  York, NY,
     1977.   p.  154.

Fritz, J. S. , and S. S. Yamamura.  Rapid microtitration of sulfate.  Anal. Chem. 27:1461-1464,
     1955.

Fritz,  J.  S. ,  S. S.  Yamamura, and  M.  J. Richard.   Titration  of sylfate  following separation
     with alumina.  Anal. Chem. 29:158-161, 1957.

Fry,  J.  D.   Determination  of Sulphur  Dioxide  in  the  Atmosphere  by Absorption  in Hydrogen
     Peroxide   Solution—the   Effect   of   Evaporation   of   the  Solution  During  Sampling.
     SSD/SW/M.246, Central  Electricity Generating Board, Southwestern Region,  England, April
     1970.

Fuchs, N. A.  The Mechanics of Aerosols, Pergamon Press,  New York,  NY, 1964.  p. 64.

Fuerst,  R.  G, ,  F. P. Scaringelli,  and  J.  H.  Margeson.   Effect of  Temperature  on Stability  of
     Sulfur Dioxide Samples Collected by the Federal  Reference  Method.  EPA-600/4-76-024, U.S.
     Environmental Protection Agency, Research Triangle Park, NC, May 1976.

Galloway,  J. N. , and E. B. Cowling.   The effects of precipitation on aquatic  and terrestrial
     ecosystems:  A proposed precipitation  network.   J. Air Pollut.  Control Assoc. 28:229-235,
     1978.

Galloway,  J. N., B.  S. Cosby, and  G.  E. Likens.  Acid  precipitation:   measurement of pH and
     acidity.  Limnol.  Oceanogr. 24(6):1161, 1979.

Galloway,  J. N.,  and G. E.  Likens,  Calibration  of collection procedures  for the determination
     of precipitation chemistry.  In:   First International Symposium on Acid Precipitation and
     the Forest Ecosystem, U.S.  Department of  Agriculture and Others,   Columbus, Ohio, May
     12-15,  1975.  Water Air Soil Pollut. 6:241-258,  1976.

Gatz,  D. f., and A. N.  Dingle.  Trace  substances in  rain  water:   concentration  variations dur-
     ing convective rains,  and their  interpretation.  Tellus 23:14-17, 1971.
                                              3-103

-------
Gilbert, T.  R.,  and A, M.  Clay.   Determination of ammonia  in  aquaria and in sea water  using
     the ammonia electrode.  Anal. Chem. 45:1757-1759,  1973.

Glover, D, J., and J.  C. Hoffsommer.  Gas chromatographic  analysis  of  nitrate  and  nitrite ions
     in microgram  quantities by conversion  to  nitrobenzene.  J.  Chromatogr. 94:334-337,  1974.

Gordievskii, A. V., A. Ya,  Syrchenkov, V. V.  Sergievskii,  and M.  I.  Savvin.  Nitrate-selective
     membrane electrodes.   Elektrokhim. 8:520-521, 1972.

Gould, R.  W.,  C.  S.  Barrett,  J.  B.  Newkirk, and C. 0.  Ruud, eds.   Adv.  in X-ray  Anal. |9:305
     et seq., 1976.

Goulding, F. S. , and J. M.  Jaklevic.  X-ray  Fluorescence Spectrometer  for Airborne Particulate
     Monitoring.  EPA-R2-73-182, U.S. Environmental Protection  Agency,  Research  Triangle  Park,
     NC, April 1973.
Goulding,  F.  S.,  J.  M. Jaklevic, and B. W.  Loo.  Aerosol  Analysis  for the Regional Air Pollu-
     tion  Study.   EPA-600/4-78-034, U.S.  Environmental Protection Agency,  Research Triangle
     Park, NC, July 1978.

Gran,  G.   Determination  of the  equivalence  point in potentiometric  titrations.   Part  II.
     Analyst 77:661-671, 1952.

Granat, L.   Relation  between  pH and  the  chemical composition in  atmospheric  precipitation.
     Tell us 24(6):550, 1972.

Groblicki, P. J.,  G.  T.  Wolf,   and  R.  J.  Countess.  Visibility Reducing  Species in the Denver
     "Brown  Cloud,"   Part   I,   Relationships  Between  Extinction  and  Chemical  Composition.
     General Motors Research Laboratories publication #GMR-2417,  Env # 81, MI, 1980.

Grosjean, D.  Solvent  extraction and organic carbon determination  in  atmospheric particulate
     matter:  the  organic extraction-organic carbon analyzer  technique.   Anal.  Chem. 47:797,
     1975.

Hagen, A., and A.  Langeland.  Polluted snow  in  southern Norway  and  the effect  of the meltwater
     on freshwater and aquatic  organisms.  Environ. Pollut.  S:45-57, 1973.

Hager, R.  N.,  Jr.,  and R.  C.  Anderson.   Theory of the derivative  spectrometer.  J. Opt.  Soc.
     Am.  60:1444-1449, 1970.

Hamilton, R.  J. ,  and W.  H.  Walton.  The  selective sampling of respirable dust.   In:  Inhaled
     Particles and Vapours.  C. N.  Davies,  ed.,  Pergamion Press,  London, England, 1961.   pp.
     465-475.

Hardin, J.,  and  B.  Shleien.  Determination  of  sulfur dioxide in  ambient  air by  non-dispersive
     X-ray fluorescence.   In:   Proceedings of  Symposium on  Nuclear Techniques and Measurement
     and  Control  of  Environmental   Pollution.   International  Atomic  Energy  Agency, Vienna,
     Austria, 1971.

Harker, A,, L. Richards, and W. Clark.   Effective atmospheric S0? photochemistry upon observed
     nitrate concentrations.  Atmos. Environ. 11:87-91,  1977.

Heard, M.  J. ,  and R.   D.  Wiffen.  Electron  microscopy   of  natural aerosols and the identifica-
     tion of particulate ammonium sulphate.  Atmos. Environ. 3_:337-340, 1969.

Hemeon, W. C.  L.,  G.   F.  Haines, Jr.,  and H. M. Ide.   Determination of haze and smoke concen-
     trations by filter paper samples.   Air  Repair 3:22-28,  1953.
                                             3-104

-------
Hendry,  C.  D.,  and  P.  L.  Brezonik.   Chemistry  of  precipitation at  Gainesville, Florida.
     Environ.  Sci, Techno!. 14(7):843-49, 1980.

Hering, S.  V. , R. C. Flagan, and S. K. Friedlander.  Design and evaluation of new low-pressure
     impactor.   I.  Environ. Sci. Technol. 12:667-73,  1978.

Hidy, G. M.   Summary  of the California aerosol  characterization experiments.  J. Air Pollut.
     Control Assoc. 25:1106, 1975.

Hill, A. S. G,   Measurement of the optical densities  of smokestains on filter papers.  Trans.
     Farad. Soc. 32:1125-1131, 1936,

Hodgeson, J. A.,  J.  P.  Bell, K.  A.  Rehme,  K. J. Krost, and R. K,  Stevens.  Applications of  a
     Chemiluminesence Detector  for the measurement of total oxides of nitrogen and  ammonia in
     the atmosphere.   Presented at  Joint Conference  on  Sensing of Environmental Pollutants,
     AIAA,  New York, NY, 1971.  Paper no.  71-1067.

Hoell,   J.  M.,  C.  N.  Harwarel, and  B.  S. Williams.   Remote  infrared  heterodyne  radiometer
     measurements of atmospheric ammonia profiles.  Geophys. Res. Lett., 7.313, 1980.

Hoffer,  E.,  and E. L.  Kothny.   A Micromethod for Sulfate  in  Atmospheric Particulate Matter.
     Report No.  163, Air and Industrial Hygiene Laboratory, Berkeley, CA, July 1974.

Hounam,  R.  E.,  and  R.  J.  Sherwood.   The cascade centripeter:   a  device  for determining the
     concentration  and  size distribution  of  aerosols.   Am.  Ind.  Hyg.  Assoc.  J. 26:122-131,
     1965.

Huey,  N. A.   The lead  dioxide  estimation  of  sulfur oxide  pollution.  J.  Air Pollut. Control
     Assoc. 18:610-11, 1968.

Huey,  N.  A.,  M.   A.  Wallar, and C.  D.  Robson.   Field  evaluation of an  improved  sulfation
     measurement system.  Presented at the 62nd Annual Meeting, Air Pollution Control Associa-
     tion,  New York, NY, June 22-26, 1969.  Paper no.  69-133.

Huntzicker, J. J., L. M. Isabelle, and J.  G. Watson.   The continuous monitoring of particulate
     sulfate by  flame photometry.   In:  International Conference on Environmental Sensing and
     Assessment,   vol.  2.   Institute of Electrical and Electronics Engineers, Inc.,  New York,
     NY, 1976.  Paper no. 23-4.

Huntzicker, J.  J., L. M. Isabelle, and J.  G. Watson.   The continuous monitoring of particulate
     sulfur compounds by flame  photometry.  Presented  at  the  173rd National Meeting, American
     Chemical  Society, New Orleans, LA, March 20-25, 1977.  Paper no. ENVR-39.

Huntzicker, J.  J. , R. S.  Hoffman,  and C.  S.  Ling.  Continuous  measurement and speciation of
     sulfur containing aerosols by flame photometry,   Atmos. Environ. 12:83-88, 1978.

Huygen, C.   The  sampling of sulfur dioxide in air with impregnated filter paper.  Anal. Chim.
     Acta 28:349, 1963.                           }

Ingram, W., and J, Golden.  Smoke  curve calibration, J. Air Pollut. Control Assoc. 23:110-115,
     1973.

Intersociety Committee.  Methods of Air Sampling and Analysis, 2nd  ed.  American  Public Health
     Association, Washington, DC,  1977a.

Intersociety Committee.   Tentative method of analysis for  nitrate in atmospheric particulate
     matter (brucine  method).   In:   Methods  of Air Sampling  and Analysis,  2nd ed.  American
     Public Health Association, Washington, DC, 1977b.  pp. 521-523.


                                              3-105

-------
Intersociety  Committee.  Tentative  method  of  analysis for  total  nitrogen  oxides  as nitrate
     (phenoldisulphonic  acid method).   In:   Methods  of Air  Sampling and  Analysis,  2nd ed.
     American Public Health Association, Washington, DC, 1977c.  pp.  534-538.

Intersociety Committee.  Tentative  method of  analysis  for  nitrate in atmospheric participate
     matter  (2,4-xylenol method  (1)).   In:   Methods  of Air  Sampling and  Analysis,  2nd ed.
     American Public Health Association, Washington, DC, 1977d.  pp.  518-520.

Jacobs,  M.  B.   The  Chemical Analysis  of Air  Pollutants.   Wiley-Interscience,  New York, NY,
     1960.

Jacobs,  H.  B., H.  H.  Braverman, and S.  Hochheiser.   Ultramicrodetermination of sulfides in
     air.  Anal. Chem. 29:1349-1359,  1957.

Jaklevic, J.  H.,  R.  C.  Gatti,  F.  S.  Goulding,  and B. W.   Loo.   A  p-gauge method applied to
     aerosol samples.  Environ. Sci.  Techno!.  15:680-686, 1981.

Johansson,  T.  B., R.  E. Van Grieken,  J.  W.  Nelson,   and J.  W.  Winchester.  Elemental  trace
     analysis  of  small  samples  by  proton induced  X-ray  emission.    Anal.  Chem. 47:855-860,
     1975.
John,  W. G.,  and G.  P. Reischl.   Measurements of the filtration  efficiencies of selected
     filter types.  Atmos.  Environ.  12:2015, 1978.

John, W., G. Reischl, and J. Wesolowski.   Size  Selective Monitoring Techniques for Particulate
     Matter.   Report No. ARB  A5-00487, California  Department of Health,  Air and Industrial
     Hygiene Laboratory, Berkeley, CA,  February 1978.

Johnson,  D.  A., and  D.  H.  F.  Atkins.   An airborne system for the  sampling and analysis of
     sulphur dioxide and atmospheric  aerosols.  Atmos.   Environ. 9:825-829,  1975.

Johnson,  R.  L.,  and J.  J. Huntzicker.   Analysis of  volatilizable and elemental  carbon in
     ambient aerosols.   In:   Proceedings:   Carbonaceous Particles in the Atmosphere, National
     Science Foundation  and Lawrence Berkeley  Laboratory,  Berkeley,  California, March 20-28,
     1978.  T.  Novakov,  ed., LBL - 9037,  Lawrence Berkeley  Laboratory, University of Califor-
     nia, Berkeley, CA, June 1979.  pp.  10-13.

Johnson,  W.  B., R.  J.  .Allen,  and  W.  E.  Evans.   LIDAR Studies of  Stack  Plumes in Rural and
     Urban Environments.   EPA-650/4-73-002, U.S.  Environmental Protection  Agency, Meteorology
     Laboratory,  National  Environmental Research Center, Research Triangle Park, NC, October
     1973.

Junge, C. E.,  and P. E. Gustafson.   Precipitation sampling for chemical analysis.  Bull. Ara.
     Meteq.ro!.  Soc. 37: 244, 1956.

Kadlecek, J. A.,  and V.  A.  Mohnen.   Time dependence of the pH of rain.   |n:   Proceedings of
     the  First International Symposium on Acid  Precipitation  and the Forest Ecosystem, U.S.
     Department of  Agriculture and  Others, Columbus,   Ohio, May  12-15,  1975.  L.S. Dochinger
     and T.  A.  Seliga, eds., NEFES/77-1, U.S.  Department of Agriculture, Forest  Service,  Upper
     Darby, PA, August 1976. p. 207.

Kamphake, L., S, Hannah, and J. Cohen.  Automated analysis for  nitrate by hydrazine reduction.
     Water Res. 1:205-16, 1967.

Kasten-Schraufnagel, P., D.  L. Ehman, and  D. J. Johnson.  Telmatic Study Phase II:  Effects of
     Collection and  Handling Conditions on the Stability of the Dichlorosulfitomercurate (II)
     Complex Formed  in  Sampling for  S0? by the Modified West-Gaeke Method.   Texas Air Control
     Board,  Air Quality Evaluation Division, January 1975.
                                             3-106

-------
Katz, M.  Photoelectric  determination of atmospheric sulphur dioxide employing dilute starch-
     iodine solutions.  Anal. Chem. 22:1040-1047, 1950.

Katz, M.  Measurement of Air Pollutants—Guide to Selection of Methods.  World Health Organi-
     zation, Geneva, Switzerland, 1969.

Keay, J. ,  and  P.   M.  A.  Menage.   Automated  determination of  ammonium and  nitrate  in soil
     extracts by distillation.  Analyst 95:379, 1970.

Kennedy, V.   Research and Monitoring  of  Precipitation Chemistry in the U.S.  1/m Present and
     Future Needs, U.S. Geological Service, Reston, VA, July 1978.

Ketseridis, G., J.  Hahn, R.  Jaenicke, and  C.  Junge.   The organic constituents of atmospheric
     particulate matter.   Atmos. Environ. 10:603, 1976."

Kieselbach, R.  Microdetermination of nitrates by the  Devarda  method.   Ind.  Eng. Chem. Anal.
     Ed. 16:764-766, 1944.

Kittelson,  D.  B. ,   R.  McKenzie, M. Vermeersch,  f.  Dorman,  D.  Pui, M.  Linne, B.  Liu, and K.
     Whitby.   Total  sulfur aerosol concentration with an electrostatically pulsed flame photo-
     metric detector system.   Atmos, Environ. 12:105, 1978

Knight,  G.  and Lichti,  K.   Comparison of  cyclone and horizontal  elutriator size selectors.
     Am. Ind. Hyg.  Assoc. J.  31:437-441, 1970.

Knight,  N.   Nitrogen in rain and snow.  Proc. Iowa Acad. Sci. 18:75-77,  1911.

Kohler,  A., and W.  Fleck.  Comparative dust-precipitation and dust-concentration measurements,
     Staub Reinhalt.  Luft 26(3):21-28. 1966.

Kolthoff,  I.  M. ,  E.  B.  Sandell,  E.  J.  Meehan,  and S.  Bruckenstein.   Quantitative Chemical
     Analysis, 4th ed., MacMillan, Toronto, Canada, 1969.  pp. 992-994.

Koschmieder, H. Beitr. Phys.  Freien Atmos.  12:33-55, 171-181, 1924.

Kretzschmar,  K.   Comparison  between  three  different methods for  the  estimation of the total
     suspended matter in urban  air.  Atmos. Environ. 9j931-935, 1975.

Kronebach, G. W.  WHO letter 31,675/R/lOP,  Report of the Expert Meeting  on Wet and Dry Deposi-
     tion (Toronto, 17-21 November 1975)(WADEM), December 12, 1975.

Krupa,  S.  V., M.  R.  Coscio, Jr.,  and F.  A. Wood.  Evidence  for multiple hydrogen-ion donor
     systems in rain.  In:  First  International Symposium on Acid Precipitation and the Forest
     Ecosystem, U.S.  Department  of  Agriculture and Others, Columbus,  Ohio,  May 12-15, 1975.
     Water Air Soil Pollut. 6:415-422, 1976.

Laby, R. H., and T. C. Morton.   Estimation  of nitrates  by nitration of 7-hydroxy-4, 8-dimethyl
     coumarin.  Nature (London)  210:298-299, 1966.

Lambert, J.  P.  F.,  and  F. W. Wilshire.  Neutron activation analysis for simultaneous  determi-
     nation  of  trace  elements  in ambient air collected  on  glass-fiber filters.  Anal. Chem.
     51:1346-1350,  1979.

Lamothe, P.  J. ,  and R.  K. Stevens.   Sulfuric acid analysis using  low temperature volatiliza-
     tion.   Presented at the 171st National Meeting, American Chemical  Society, New York, NY,
     April 4-9, 1976.  Paper  no.  ENVR-38.
                                              3-107

-------
Lawrence Berkeley Laboratory,  Instrumentation for Environmental Monitoring,  Air.   LBL-1, Vol.
     1, Part 1, University of California, Berkeley, CA, May 1972.

Lawrence Berkeley Laboratory.  Instrumentation for Environmental Monitoring,  Air.   LBL-1, Vol.
     1, Part 2, University of California, Berkeley, CA, July  1975.

Lazrus, A.  L.,  K.  C.  Hill, and J. P. Lodge.  In:  Proceedings of  the  1966  Technicon  Symposium
     on Automation in Analytical Chemistry.  Mediad,  Inc., New York, NY,  1966.   p.  291.

Lazrus, A.,  E.  Lorange,  and J. P. Lodge, Jr.  New automated  microanalyses  for  total  inorganic
     fixed  nitrogen  and  for sulfate  ion in water.   In:  Trace  Inorganics in  Water,  R.   F.
     Gould, ed., American  Chemical  Society, Washington,  DC,  1968.  pp.  164-171.   Advances  in
     Chemistry Series 73.

Leahy, D., R. Siegel, P.  Klotz, and L. Newman.  The separation and characterization of sulfate
     aerosol,  Atmos. Environ. 9:219-229, 1975.

Lee,  R.  E.   Size  distribution  of  suspended  particulates  in air.   Research/Development
     23(6):18-21, 1972.

Lee, R.  E., and S.  Goranson.  National  Air Surveillance Cascade Impactor Network.   I. Size
     distribution measurements of suspended particulate matter in  air.  Environ.  Sci. Techno!.
     6:1019-1024, 1972.

Lee, R. E. ,  Jr.,  and J.  Wagman.  A  sampling anomaly  in  the  determination  of atmospheric sul-
     fate concentration.   Am. Ind. Hyg. Assoc. J. 27:266-271, 1966.

Liberti, A.,  M. Possanzini,  and  M.  Vicedamini.   Determination  of the nonvolatile acidity  of
     rainwater by a coulometric procedure.  Analyst 47:352, 1972.

Likens, G.  E.  Acid precipitation.  Chem. Eng. News 54(48):29-44,  1976.

Likens, G.  W., F.  H.  Bormann, and  N.  M.  Johnson.    Acid  rain.   Environment  14:33-40, 1972.

Lin, C.  I., M. Baker, and  R.  J.  Charlson.   Absorption coefficient  of atmospheric aerosol: a
     method for measurement.  Appl.  Opt.  12, 1356-63,  1973.

Lioy, P. J., J. G.  Watson, and J.  D. Spengler.  APCA  Specialty Conference Workshop  on Baseline
     Data  for  Inhalable  Particulate  Matter.  J.   Air  Pollut.  Control Assoc.  30:10, 1980.  pp.
     1126-1130.

Lippmann, M.  Respirable dust sampling.   Am. Ind. Hyg. Assoc. J. 31:138-159,  1970.

Lippmann,  M.,  and  T.  L.  Chan.   Cyclone  sampler performance,  Staub Reinhalt.    Luft 39:7-11,
     1979.

Liu, B, Y,  H., and D. Y.  H.  Pui.   On the performance of the Electrical  Aerosol  Analyzer.   J.
     Aerosol Sci. 6:249-264, 1975.

Liu, B. Y.  H,  D.  Y.  H.  Pui,  K.  L.  Rubow, and G. A.  Kuhlmey.  Research on  Air  Sampling  Filter
     Media: Progress  Report - Grant Report R804600,  University of Minnesota, Minneapolis, MN,
     May 1978a.

Liu, B. Y.  H. , D.  Y. H.   Pui,  K.  T.  Whitby, D.  B.  Kittelson, Y.  Kousaka,  and  R. L. McKenzie.
     The aerosol mobility chromatograph,  a new  detector for sulfuric acid aerosols.   Atmos.
     Environ. 12:99,  1978b.
                                             3-108

-------
Liu, B.  Y.  H.,  D.  Y.  H.  Pui,  and B. W.  Loo.   A Sampling  Inlet  for Inhalable Particles (in
     press).   Particle Technology Laboratory, University of Minnesota, 1980.

Liu, B. Y. H.,  and D.  Y. H. Pui.  Aerosol sampling inlets and inhalable particles.  In:  Pro-
     ceedings:   Advances  in Particle Sampling and Measurement,  U.S.  Environmental Protection
     Agency,  Daytona Beach, FL,  October 1979.  EPA-600/9-80-004, U.S. Environmental Protection
     Agency,  Research Triangle Park, NC, January 1980.  pp. 382-408.

Lizarraga-Rocha, J. A.   The Effect of Exposure of Filters During Non-operating Periods  in the
     High Volume Sampling Method.  M.S. Thesis, University of North Carolina, Chapel Hill, NC,
     1976.

Lodge,  J. P., Jr.,  A.  P. Waggoner, D.  T.  Klodt, and C.  N.  Grain.  Non-health effects of air-
     borne particulate matter.  Atmos. Environ. 15:431-482, 1981,

Lodge,  J.  P., Jr., J.  Ferguson,  and B. R. Havlik.  Analysis of micron-sized particles:   Deter-
     mination of sulfuric acid aerosol.  Anal. Chem.  32:1206-1207, 1960.

Lodge,  J.  P., Jr., J. B.  Pate,  and  H.  A.  Huitt.   The use  of  impregnated filters to collect
     traces of gases in the atmosphere.  Am. Ind. Hyg. Assoc. J. 24:380-387, 1963.

Lodge,  J. P., Jr.,  K.  C. Hill,  J. B. Pate, E. Lorange, W. Basbergill, A.  L. Lazrus, and G. S.
     Swanson.  Chemistry of United States Precipitation:  Final Report on  the National Precip-
     itation  Sampling Network.   National  Center for Atmospheric Research, Boulder, CO,  August
     1968.

Loo, B. W., J. M. Jaklevic, and F. S. Goulding.  Dichotomous virtual  impactors for  large-scale
     monitoring  of airborne  particulate matter.   In:  Fine Particles:   Aerosol   generators,
     measurement,  sampling,  and  analysis.  B. Y.  H.  Liu, ed., Academic  Press,  New York, NY,
     1976.

Loo, B. W. ,  R.  S.  Adachi, C. P.  Cork, F. S. Goulding, J. M. Jaklevic, D.  A. Landis, and W. L.
     Searles.  A  second  generation dichotomous sampler for large scale monitoring  of airborne
     particulate matter,  LBL-8725,  Lawrence Berkeley  Laboratory,  Berkeley,  CA,  January 1979.

Loo, B. W, ,  W.  R.  French,  R. C.  Gatti,  F.  S. Goulding,  J.  M.  Jaklevic,  J. Llacer, and A. C.
     Thompson.   Large scale measurement of  airborne particulate  sulfur.   Atmos.  Environ.
     12:759-771, 1978.

Lundgren, D.  A.   Mass Distribution of Larger Atmospheric Particles.   Ph.D., Thesis, University
     of Minnesota, Minneapolis,  MN, 1973,

Lundgren, D.  A. ,  L.  D.  Carter,  and P. S. Daley.   Aerosol mass measurement using piezoelectric
     crystal   sensors.   |n:  Fine Particles:   Aerosol  Generation, Measurement,  Sampling, and
     Analysis.  B. Y.  H.  Liu, ed.  Academic Press, New York, NY, 1976.  pp. 485-510.

Macias, E.  S.   Instrumental  analysis of  light  element  composition  of atmospheric aerosols.
     In:   Methods  and Standards  for Environmental Measurement.   W.  Kirchhoff,  ed. , National
     Bureau  of  Standards Special  Publication 464.   Department of Commerce,  Washington, DC,
     1977.  pp.  179-188.

Macias, E. S. ,  and R.  B. Husar.    Atmospheric  particulate mass measurement with beta attenua-
     tion mass monitor.   Environ. Sci. Techno!. 10:904-907, 1976.

Macias, E. S. , C. D. Radcliffe,  C. W. Lewis, and C. R. Sawicki.  Proton induced X-ray analysis
     of  atmospheric  aerosols for carbon,  nitrogen,  and  sulfur  composition.    Anal.  Chem.
     50:1120-1124, 1978.
                                             3-109

-------
Maddalone, R.  F.,  G,  L. McClure,  and  P.  W,  West,  Determination of  sulfate  by  thermal  reduc-
     tion of perimidylammonium sulfate.  Anal. Chem. 47:316, 1975,

Malissa, H.  Analysis of Airborne Particles by Physical Methods.  CRC  Press,  Inc., Cleveland,
     OH, 1978.

Mamane, Y.,  and R.  G.  de  Pena.   A  quantitative method for the detection of individual  sub-
     micron size sulfate  particles.   In:  Sulfur  in  the Atmosphere, Dubrovnik,  Proceeding of
     the International  Symposium,  Yugoslavia,  September 7-14,  1977.  Atmos. Environ. 12:69-82,
     1978.                               ,                                            ~

Harple, V. A.,  and K. Willeke.  Impactor design.  Atmos. Environ. 10:891-896,  1976.

Harple, V. A.,  and K, Willeke.  Inertia!  Impactors.   Aerosol  Measurement, University Presses
     of Florida, Gainesville, FL,  1979.

Martin, A., and  F.  R. Barber.  Some measurements of loss of atmospheric  sulphur dioxide  near
     foliage.   Atmos. Environ. 5:345-52, 1971.

Martin, B. E.   Sulfur Dioxide Bubbler Temperature Study.  EPA-600/4-77-040, U.S.  Environmental
     Protection Agency, Research Triangle Park, NC, August 1977,

McClenny,   W.  A., and C.  A.  Bennett,  Jr.  Integrative  technique for detection of atmospheric
     ammonia.   Atmos. Environ. 14:641-645, 1980.

McColl, J. G.,  and D. S. Bush.  Precipitation and throughfall  chemistry  in  the  San Francisco
     Bay Area.   J.  Environ.  Qua!.  7:352-357, 1978.

McCoy,  R.  A.,  D.  E.  Camann, and  H.  C.  McKee.  Collaborative  Study   of Reference Method for
     Determination  of  Sulfur Dioxide   in  the  Atmosphere  (Pararosaniline  Method)  (24-Hour
     Sampling).  EPA-650/4-74-027, U.S.  Environmental  Protection  Agency,  Research Triangle
     Park, NC,  December 1973.

McCrone, W. C.,  and J.  G.  Delly.  The  Particle Atlas  Edition  Two.   Volume I:   Principles and
     Techniques.  Ann  Arbor Science Publishers, Inc.,  Ann Arbor, MI, 1973.   p.  259.

McFarland, A.  R.  Wind Tunnel Evaluation  of British Smoke Shade  Sampler:  Test Report.  Air
     Quality Laboratory Report, 3565/05/79/ARM, Texas A&M University, College  Station, TX, May
     1979.

McFarland, A.  R.,  and C. A. Ortiz.  Aerosol  Characterization of Ambient  Particulate Samplers
     Used  in  Environmental  Monitoring  Studies:   Progress  Report.   Texas A&M  University
     Research Foundation, College  Station, TX, October  1979.

McFarland, A.  R.,  and C. E.  Rodes.   Characteristics of aerosol samplers  used in ambient air
     monitoring.   Presented   at   86th  National  Meeting,  American  Institute  of  Chemical
     Engineers, Houston, TX, April 2, 1979.

McFarland, A.  R. ,  C.  A.  Ortiz, and  C.  E.  Rodes.   Evaluation  of  the  NASN  Cascade Impactor
     Ambient Air Monitor.   Report  3565/07/78/ARM, Texas A&M,  Air  Quality Laboratory, College
     Station,  TX, 1978.  Atmos. Environ., in press, 1980.

McKee,  H.  C.,  R.  E.  Childers, and 0.  Saenz,  Jr.  Collaborative Study  of  Reference Method fpr
     the  Determination of  Suspended  Particulates  in  the Atmosphere  (High  Volume  Method).
     APTD-0904,  U.S.  Environmental Protection Agency,  Research  Triangle  Park, NC,  June 1971.

Menis, 0., D.  L. Manning, and R.  G. Ball.  Automatic spectrophotometric titration of fluoride,
     sulfate, uranium, and thorium.  Anal. Chem. 30:1772-1776,  1958.


                                            '3-110

-------
Miller, ,F. J. ,  D.  E.  Gardner, J. A. Graham, R. E. Lee, Jr., W. E. Wilson, and J. D. Bachmann.
     Size considerations for  establishing a standard for inhalable particles.  J. Air Pollut.
     Control  Assoc. 29:610-615, 1979.

Miller, J. M. ,  and V.  R. Highsmith.   The WMO  Regional Precipitation Chemistry Network  in the
     United States.   NOAA-TM-ERL ARL-60,  U.  S.  Department  of Commerce,  National Oceanic and
     Atmospheric Administration, Silver Spring, MD, October  1976.

Miller, K. ,  and H. W.  DeKoning.   Particle sizing instrumentation.   Presented at 67th  Annual
     Air  Pollution Control  Association meeting,  June 9, 1974, Denver,  CO.   Paper no.  74-48.

Moffat, A. J. ,  J.  R.  Robbins, and  A.  R.  Barringer.   Electro-optical sensing of  environmental
     pollutants.  Atmos. Environ. E>:511-25, 1971.

Morris, A. W. ,  and J.  P. Riley.  The  determination  of nitrate in seawater.  Anal. Chim. Acta
     29:272-279, 1963.

Morrison, G.   H.,  ed.   Trace Analysis:   Physical  Methods.   John Wiley and Sons,  New York, NY,
     1965.  pp. 245-263, 271-310, 325-360, 377-425.

Mrose, H.  Measurements of pH,  and chemical  analyses of rain-  snow-,  and fog-water.   Tell us
     18:266-70, 1966.

Mudgett,  P. S. ,  L. W.  Richards, and J.  R.  Roehrig.   A new  technique to measure  sulfuric acid
          in   the  atmosphere.   In:   Analytical Methods Applied  to Air Pollution Measurement.
          R.  K. Stevens and W. F. Herget, eds., Ann Arbor Science  Publishers,  Inc., Ann  Arbor,
          MI, 1974.

Mulholland, G.  W. , D.  Y. H.  Pui, A.  Kapadia,  and B. Y. H.  Liu.   Aerosol number  and mass con-
     centration measurements: - A comparison of the EAA with other measurement techniques.  J.
    , Colloid Interface Sci.   77(l):57-67, 1980.

Mulik, J. , R.  Puchett,  D.  Williams, and  E.  Sawicki.   Ion chromatographic analysis of sulfate
     and  nitrate in ambient aerosols.  Anal. Lett. 9(7): 653-663,  1976.

Mulik, J. D. , G. Todd, E. Estes, R.  Puckett, E. Sawicki, and D. Williams.  Ion Chromatographic
     Determination  of Atmospheric  Sulfur Dioxide.  In:  Ion Chromatographic Analysis  of Envi-
     ronmental  Pollutants.   E.  Sawicki,  J. D.  Mulik, and  E. Wittgenstein,  eds.,  Ann Arbor
     Science Publishers, Inc., Ann  Arbor, MI,  1978.  pp. 23-40.

Mueller,  P.  K. ,  R. W.  Mosley, and  L.  B.  Pierce.  Carbonate and  noncarbonate  carbon  in  atmos-
     pheric particles.   In:   Proceedings of the  2nd  International Clean Air  Congress,  Inter-
     national  Union of  Air  Pollution  Prevention  Associations, Washington,  DC, December 6-11,
     1970.   H.  M.   Englund  and W.  T.  Beery,  eds., Academic Press,  New  York,  NY, 1971.  pp.
     532-539.

Mullin,  J.  B. , and J.  P.  Riley.  The spectrophotometric determination  of nitrate in natural
     waters,  with  particular reference to seawater.  Anal. Chim.  Acta 12:464-480, 1955.

Nader,  J. S.    Dust-retention  efficiencies of dust  fall collectors.   J.  Air Pollut. Control
     Assoc. 8:35-38, 1958.

National  Research  Council,  Subcommittee  on Airborne  Particles.   Airborne Particles.  Univer-
     sity Park Press, Baltimore, MD, 1979.

National  Air  Pollution  Control Administration.   Air  Quality Criteria for Particulate Matter.
     NAPCA  Publication  No.  AP-49,   U.S.  Department  of   Health,   Education,   and   Welfare,
     Washington, DC, 1969.


                                             3-111

-------
National Air Pollution Control Administration.  Air Quality  Criteria  for  Sulfur Oxides,  NAPCA
     Publication No. AP-50, U.S. Department of Health,  Education,  and Welfare, Washington, DC,
     1970.

Natusch, D.  F.  S.,  and J. R. Wallace.  Determination of  airborne  particle  size distributions:
     Calculations of  cross sensitivity and discreteness  effects in cascade impaction.  Atmos.
     Environ. 10:315-324,  1976.

Natusch, D.  F.  S.,  C. F.  Bauer, and A. Loh.  Collection  and analysis of  trace elements in the
     atmosphere.  In;  Air Pollution Control, Part III.   W.  Strauss,  ed., John Wiley  and Sons,
     Inc., New York, NY, 1978.

Niemann, B. L., J. Root, N. Van Zwalenburg, and A. L. Mahan.  An Integrated Monitoring Network
     for Acid Deposition:  A Proposed Strategy.   Report R-023-EPA-79, Teknekron Research  [nc.,
     Berkeley, CA, November 1979.

Noll,  K,   E.   A  rotary  inertia!  impactor for   sampling  giant particles  in the  atmosphere.
     Atmos. Environ. 4:9-20, 1970.

Norwegian  Institute  for Air  Research (NIAR).   LRTAP  5/71.  Coulometric Titration  of Strong
     Acid  in Precipitation, 1971.

Novakov, T.   Chemical  characterization of atmospheric  pollution particulates by photoelectron
     spectroscopy.  In:   Proceedings  of the 2nd  Joint  Conference  on Sensing of Environmental
     Pollutants, Instrument Society of America, Washington,  DC, December  10-12, 1973.  p. 197.

Novakov, T. ,  S.  G.  Chang, and  A.  B.  Harker.  Sulfates as pollution particulates:   catalytic
     formation on carbon (soot) particles.  Science (Washington, DC)  186:259-261,  1974.

O'Brien, R.  J. ,  J. R. Holmes,  R.  J.  Reynolds,  J.  W.   Remoy, and  A.  H.  Bockian.   Analysis of
     photochemical aerosols in Los Angeles basin  according to particle size.  Presented at the
     67th  Annual  Meeting, Air  Pollution Control Association,  Denver,  Colorado,   1974.  Paper
     no. 74-155.

Okabe,  H., P.  L. Splitstone,  and  J.  J.  Ball.   Ambient and souce   S0»  detectors based on  a
     fluorescence method:  J. Air Pollut. Control Assoc.  23:514-516,  1973.

O'Keeffe,  A.  £., and  6.   C.  Ortman.   Primary standards  for trace gas  analysis.   Anal. Chem.
     38:760-63,  , 1966.

Okita,  T. , S.  Morimoto,  M.  Izawa,  and S.  Konno.   Measurement  of gaseous and  particulate
     nitrates in the atmosphere.  Atmos. Environ. 10:1083-1089, 1976.

Orel, A.  E.  and  J.  H.  Seinfeld.  Nitrate formation  in  atmospheric  aerosols.   Environ.  Sci.
     Techno!. 11:1000-7,  1977.

Orenstein, A.  J., ed.   Proceedings of the Pneumoconiosis  Conference,  the  University of Wit-
     watersand, Johannesburg, South Africa,  February 9-24,  1959,  Little, Brown, Boston, 1960.
     p.  619.

Pashel,  G. E., and  D.  R.  Egner.  A comparison of ambient suspended particulate matter concen-
     trations as measured  by the British smoke sampler  and the  high volume  sampler at 16 sites
     in the United States.  Atmos. Environ. 15:919-27,  1981.

Pate, J.  B., J.  P,  Lodge, Jr., and M.  D.  Neary.  The use  of  impregnated  filters to collect
     traces of gases in the atmosphere.  Anal. Chim. Acta 28:341,  1963.
                                             3-112

-------
Perera, F. P. ,  and  A.  K. Ahmed.   Respirable  Particles:   Impact of Airborne  Fine  Participates
     on Health and the Environment.  National Resources Defense Council,  Inc., Washington, DC,
     October 1978.

Philips Electronic Instruments.  Coulometric amperometric analyzers.  Bulletin.  WA-4-C13, un-
     dated.

Pickerell, D.,  T. Hook,  T.  Dolzine, and  J.  K.  Robertson.   Intensity-weighted sequential sam-
     pling of precipitation:   a technique for monitoring  changes  in storm chemistry during a
     storm.   Ion Chromatogr. Anal. Environ. Pollut. 2:289,94 1979.

Pierson, W.  R.,  R. H. Hammerle, and W.  W. Brachaczek.  Sulfate formed by  interaction of  sulfur
     dioxide with filters and aerosol deposits.  Anal. Chem. 48:1808-1811, 1976.

Pierson, W.  R., W.   W.  Brachaczek, T.  J.  Korniski,  T.  J. Truex,  and  J.  W.  Butler.  Artifact
     formation  of  sulfate,  nitrate, and hydrogen ion on  backup  filters:  Allegheny Mountain
     experiment.  J. Air Pollut. Control Assoc. 30:30-34, 1980.

Ranade, M. B. ,  and  D.  VanOsdell.  Quality  Assurance  Evaluation of the CHAMP Aerosol Sampler.
     Project Report 43U-1487-59, Research Triangle Institute, Research Triangle  Park, NC, July
     1978.

Ratzlaff, K.  L., and D. F. S.  Natusch.   Theoretical assessment of  precision in dual wavelength
     spectrophotometric measurement.  Anal. Chem. 49:2170-2176, 1977.

Rayner, A. C.   A rapid titrimetric method  for sulfate in lead dioxide exposed  to atmospheric
     sulfur pollution.  J. Air Pollut.  Control Assoc. 16:418-421,  1966.

Regan,  G.  F.,  S. K.  Goranson,  and  L.   L.  Larson.   Use of  tape  samplers as fine  particulate
     monitors.  J. Air Pollut. Control  Assoc. 29:1158-1160,  November 1979.

Richardson,  H.  L.   The nitrogen cycle in grassland soils; with special reference  to the Roth-
     amsted Park grass experiment.  J.  Agric, Sci. 28:73-121, 1938.

Ridder, T. B.   On the Chemical  Composition  of Rain Fall:   Comparison of Measurement Results.
     KNMI-WR-78-4,  Royal Netherlands  Meteorological  Inst., DeBilt,  The  Netherlands,   1978.

Rodes,  C.  E. ,  and  G.  F.  Evans.   Summary  of  LACS Integrated Measurements.   EPA-600/4-77-034,
     U.S. Environmental Protection Agency, Research Triangle Park, NC, June 1977.

Rodes,  C. E., H.  F. Palmer,  L.  A.  Elfers, and C. H.  Morris.   Performance characteristics of
     instrumental  methods  for  monitoring  sulfur  dioxide.   J.   Air  Pollut.  Control   Assoc.
     19:575-584, 1969.

Rosen,  H. , and  T.  Novakov.    Identification  of  primary particulate carbon and sulfate species
     by ramon spectroscopy.   Atmos. Environ. 121:923-926,  1978.

Rosen,  H., A. D. A.  Hansen, L. Gundel, and T. Novakov.  Identification of the optically  absorb-
     ing component in urban aerosols.  Appl. Opt. 17:3859-61, 1978.

Ross,  W.  D.  , G.  W.  Buttler,  T.  G. Duffy,  W.  R.  Rehg,  M. T.  Wininger, and  R.   E. Sievers.
     Analysis for aqueous  nitrates  and  nitrites  and gaseous  oxides  of nitrogen  by electron
     capture gas chromatography.   J, Chromatogr.  112:719-727, 1975.

Ruzicka,  J. , and E.  H.  Hansen.   New  potentiometric  gas sensor.   Air-gap  electrode.    Anal.
     Chim. Acta. 69:129-41, 1974.
                                              3-113

-------
Ruzicka, J., E. H. Hansen, P. Bisgaard, and E.  Raymann.   Determination  of  the  ammonium  content
     in waste  waters  by means of  the  air-gap electrode.  Anal.  Chim.  Acta. 72:215-219,  1974.

Ryan, M.  D.,  and  D.  D.  Siemer.   Anodic stripping  voltammetry of atmospheric samples.   Pre-
     sented  at 3rd Annual  Meeting Federation  Analytical Chemistry and Spectroscopy Society,
     1976.

Saltzman, B.  E.  Colorimetric microdetermination of  nitrogen  dioxide  in the  atmosphere.   Anal.
     Chem. 26:1949-1955, 1954.

Sawicki, E.  Separation and analyses of polynuclear  aromatic  hydrocarbons  present  in the  human
     environment.  Chemist. Analyst 53:24-26, 28-30, 56-S2, 88-91,  1964.

Scaringelli, F.  P.,  and K. A. Rehme.  Determination of  atmospheric concentrations  of sulfuric
     acid  aerosol  by  spectrophotometry,  coulometry,   and  flame  photometry.    Anal.   Chem.
     41:707-13, 1969.

Scaringelli,  F.  P.,  B.  E.  Saltzman,  and S.  A.  Frey.   Spectrophotometric determination of
     atmospheric sulfur dioxide.  Anal. Chem. 39:1709-1719, 1967.

Schulze, F.  Versatile combination ozone and  sulfur dioxide analyzer.   Anal. Chem.  38:748-752,
     1966.

Scott, D. R., W.  A. Loseke, L. E. Holboke, and  R. J. Thompson.  Analysis of  atmospheric parti-
     culates   for   trace   elements  by  optical   emission  spectrometry.    Appl.  Spectrosc.
     30(4):392-405, 1976.

Shaw, R. W.,  T. G. Dzubay, and W. J. Courtney.  Reproducibility of  mass measurements by dicho-
     tomous sampler.  Technical Note.  U.S. Environmental  Protection  Agency, 1979.

Shaw, R,  W.,  and  R.  D. Willis.   X-ray emission  analysis:   sample losses during  excitation.
     In;   Electron Microscopy  and X-Ray Applications.   P.   A.  Russell and A.  E. Hutchings,
     eds., Ann Arbor Science Publishers, Inc.,  Ann Arbor, MI,,1978.

Shendrikar, A. D.,  and J.  P. Lodge, Jr.  Microdetertnination  of ammonia by the ring oven  tech-
     nique and its application to air pollution studies.   Atmos.  Environ.  9:431, 1975.

Sides, J.  0.,  and H.  F. Saiger.   Effects  of Prolonged  Static  Exposure of Filters to Ambient
     Air on High  Volume Sampling Results.  Kansas State Department of  Health and  Environment,
     Bureau of Air Quality and Occupational Health, Topeka, KS, May 1976.

Sinclair, D., R.  J. Countess, B. Y. H. Liu, and D. Y. H.  Pui.   Automatic analysis  of submicron
     aerosols.  I_n:  Aerosol  Measurement,  a Workshop, University of  Florida, Gainesville,  FL,
     March 24-26,  1976.   D.  A.  Lundgren, M.  Lippmann,  F. S.  Harris, Jr., W.  E.  Clark,  W. H.
     Marlow,  and  M.  D. Durham,  eds. ,  University  Presses of Florida,  Gainesville, FL,  1979.
     pp. 544-563.

Skoag,  D.  A.,  and D. M. West.   Analytical  Chemistry.    Holt,  Rinehart  and Winston, New  York,
     NY, 1965.

Skujins, J. J.   Spectrophotometric determination of nitrate  with 4-methylumbel!iferon.   Anal.
     Chem. 36:240-241, 1964.

Small, H., T.  S.  Stevens, and W.  C.  Bauman.   Novel ion exchange chromatographic  method  using
     conductimetric detection.  Anal. Chem. 47:1801-1809,  1975.

Smith,  A.  F., D.  G.  Jenkins,  and D.  E.  Cunningworth.   Measurement  of  trace  quantities of
     hydrogen sulfide in industrial atmospheres.  J. Appl. Chem.  11:317, 1961.


                                              3-114

-------
Smith, F., and  A.  C.  Nelson, Jr.  Guidelines  for Development of a Quality Assurance Program.
     Reference Method for  the Determination of Suspended Particulates in the Atmosphere (High
     Volume Method).  EPA-R4-73-028b,  U.S.  Environmental Protection Agency, Research Triangle
    'Park,' NC, June 1973.

Smith, F.,  P.  S.  Wohlschlegel,  R. S.  C.  Rogers, and D. J.  Mulligan.   Investigation of Flow
     Rate Calibration Procedures  Associated with the  High  Volume Method for Determination of
     Suspended Particulates.  EPA-600/4-78-047, U.S. Environmental Protection Agency, Research
     Triangle Park, NC,  June 1978.

Snell, F. D.   Photometric and Fluorimetric Methods of  Analysis Metals, Part I.  John Wiley and
     Sons,  New York, NY, 1978.  p. 1-25.

Spicer, C.  W.   The Fate of Nitrogen  Oxides in the Atmosphere.   EPA-600/3-76-030.  U.S. Envi-
     ronmental Protection Agency, Research Triangle Park, NC, March 1976.

Spicer,  C.   W.,  and  P.  M.  Schumacher.   Interferences  in sampling  atmospheric particulate
     nitrate.  Atmos.  Environ. 11:873-876, 1977.

Spicer,  C.  W.,  and P. M.  Schumacher.   Particulate nitrate:   laboratory and  field studies of
     major sampling interferences.  Atmos, Environ. 13:543-552, 1979.

Spicer, C.  W., P. M. Schumacher, J. A.  Kouyoumjian, and  D.  W. Joseph.  Sampling and Analytical
     Methodology for  Atmospheric  Particulate  Nitrates.   EPA-6QO/2-78-067, U.S.  Environmental
     Protection Agency,  Research Triangle Park, NC, April 1978.

Stalker, W. W., R.  C.  Dickerson, and G. D. Kramer.  Atmospheric sulfur dioxide and particulate
     matter:    a comparison  of  methods of  measurements.   Am.  Ind.  Hyg.  Assoc.  J.  24:68-79,
     1963.

Stauff,  J.,  and W. Jaeschke.   A chemiluminescence technique for measuring atmospheric trace
     concentrations of sulfur dioxide.  Atmos. Environ.  9:1038-1039, 1975.

Stelson, A. W., S.  K.  Friedlander, and  J. H. Seinfeld,   A note on the equilibrium relationship
     between  ammonia  and  nitric  acid and  particulate  ammonium nitrate.   Atmos.  Environ.,
     13:369, 1979.

Stevens, R. K, , and T. G. Dzubay.  Dichotomous Sampler - A  Practical Approach to  Aerosol Frac-
     tination   and Collection.    EPA-600/2-78-112,   U.S.   Environmental  Protection  Agency,
     Research Triangle Park, NC, June  1978.

.Stevens, R.  K., T. G.  Dzubay, G.  Russwurm,  and D. Rickel.   Sampling and analysis of atmos-
     pheric  sulfates  and related species.  In:   Sulfur  in the Atmosphere, Proceedings of the
     International  Symposium, United  Nations,  Dubrovnik,  Yugoslavia,  September 7-14, 1977.
     Atmos. Environ. 12:55-68, 1978.

Stevens, R.  K.,  J. D. Mulik, A. E. O'Keeffe,  and K. J.  Krost.  Gas chromatography of reactive
     sulfur gases  in air at the parts-per-billion level.  Anal. Chem.  43:827-831, 1971.

Stevens, R.  K.,  A. R. McFarland,  J.  B. Wedding.  Comparison  of  a virtual dichotomous sampler
     with  a  Hi-Volume  sampler.   (Environ. Sci.  Res.  Lab.,  EPA, Research Triangle Park, NC,
     Prepr. Pa. Natl.  Meet,, Div.  Environ.  Chem., Am.  Chem.  Soc.  1976.

Stratmann,  H.   Microanalytical   method  for   determining   sulfur  dioxide in  the atmosphere.
     Microchim. Acta, 668-78, 1954.

Strickland,  J.  D.  H., and  T.  R.  Parsons.  A  Practical  Handbook  of Seawater  Analysis, 2nd ed.
     Bulletin 167,  Fisheries  Ressearch Board,  Ottawa,  Canada,  1972.


                                              3-115

-------
Summers, P. W., and D. H. Whelpdale.  Acid precipitation  in Canada.   Ini:   First  International
     Symposium on  Acid  Precipitation and the Forest  Ecosystem, U.S.  Department of Agriculture
     and Others, Columbus,  Ohio, May 12-15, 1975.   Water, Air, Soil  Pollut. 6:447-455,  1976.

Swanson, D.,  C.  Morris,  R.  Hedgecoke,  R.  Jungers,  R.  J. Thompson,  and J.  E.   Bumgarner.  A
     rapid  analytical procedure for the  analysis of  BaP in  environmental  samples.  Trends
     Fluoresc. 1:22-27, 1978.

Sweitzer,  T.  A.   The  evaluation of gas  bubbler  field  performance.  Presented at the  32nd
     Annual Meeting  of the  East Central Section,  Air  Pollution  Control  Association, Dayton,
     OH, September 17-19, 1975.

Tanner, R.  L.,  J.  Forrest, -and  L.  Newman.   Determination of atmospheric gaseous and particu-
     late sulfur compounds.   In:  Sulfur in the  Environment:   Part  I, The Atmospheric Cycle.
     J. 0.  Nriagu, ed., John Wiley & Sons, Inc., New  York, NY,  1978.   pp. 371-452.

Tanner,  R.  L.,  T. D'Qttavis,, R. Garber,  and L,  Newman.  Determination  of ambient aerosol
     sulfur using  a  continuous  flame photometric  detection  system.   Part I, sampling system
     for aerosol sulfate and sulfuric acid.  Atmos. Environ. 14:121-127,  1980.

Tanner, R.  L.,  R.  Cederwall, R.  Garber,  D.  Leahy, W. Marlow,  R.  Meyers, M.  Phillips, arid  L.
     Newman.   Separation  and analysis  of aerosol  sulfate species at ambient concentrations.
     Atmos. Environ. 11:955-966,  1977.

Tanner,' R.  L., R. W., Garber, W.  H. Marlow, B. Leaderer, D. Eatough, and M.  A. Leyko.  Chemical
     composition and  size  distribution of sulfate  as a function of  particle size in New York
     City  aerosol.   Presented at the Industrial  Hygiene Conference, New  Orleans,  Louisiana,
     May 22-27, 1977a.  Paper no. 140.

Technicon  Corporation.   Sulfate  Method  VIb Via  Turbidimetry.   Technicon  Corporation, Tarry-
     town, NY, 1959.

Technicon Industrial Systems, Method 100-70W, Tarrytown,  NY, 1973.

Terraglio,  F.  P.,  and R.  M. Manganelli.   Laboratory evaluation of sulfur  dioxide methods and
  v the influence of ozone oxides of nitrogen mixtures.   Anal. Chem,  34:675-677, 1962.

Tetlow, J. A.,  and A. L. Wilson.  An absorptiometric method for determining  ammonia in boiler
     feed-water.  Analyst 89:453-465, 1964.

Thoen, G.  N.,  G. G. Dehaas, and  R. R. Austin.  Instrumentation  for quantitative measurement  of
     sulfur compounds in kraft gases.  Tappi 51:246-249,  1968.

Thomas, J.  F.,  B. D.  Tebbens,  E.  N.  Sanborn,  and J.  M.  Cripps.   Fluorescent  spectra  of
     aromatic  hydrocarbons  found in  polluted  atmosphere.   Int.  J.  Air  Pollut.  2:210,  1960.

Thomas, R.  L.,  V.  Dharmarajan, G. L. Lundquist, and  P. W. West.  Measurement of  sulfuric acid
     aerosol,  sulfur  trioxide,  and  the total sulfate content of the  ambient  air.  Anal.  Chem.
     48:639-642, 1976.

Thompson, R. J.  Second Analysis  on References Precipitation Samples.  U.S. Environmental  Pro-
     tection  Agency,  private  communication 78,  Research Triangle  Park,  NC, May  22,   1978.

Freon, J.  F.,  and  W.  E. Crutchfield, Jr.  Rapid turbidmetric method  for  determination of sul-
     fates.  Ind. Eng. Chem., Anal.  Ed. 14:119-121, 1942.
                                             3-116

-------
Tyree, S. Y.  Rainwater acidity measurement problems.  Atmos. Environ. 15:57-60, 1981.

Tyree, S. Y.,  M.  Ballinger, S, E. Delos, and J. M. Stouffer.  Analysis of Rainwater for Acci-
     dents  -  Standards for and Methods  of.   EPA Grant  Report  R8Q449801.   College of William
     and Mary, Williamsburg, VA, 1979.

U.S.  Atomic Commission,  Division of  Biomedical  and Environmental Research.  Chemist-Meteoro-
     logist Workshop,  Fort Lauderdale,  FL,  January  14-18,  1974.   Available  as WASH-1217-74
     from National Technical Information Service, Springfield, VA.

U.S.  Environmental  Protection  Agency.   Ambient  air  monitoring  reference   and  equivalent
     methods.   Notice  of equivalent  method  designation.   Fed.  Regist.  40:34024-34025, 1975.

U.S.  Environmental  Protection  Agency.  National   Primary  and  Secondary Ambient  Air Quality
     Standards.  Appendix A -  Reference method  for the  determination of sulfur  dioxide in the
     atmosphere (pararosaniline method).  40 CFR 50:6-11, July 1,  1979.

U.S.  Environmental  Protection Agency.   Summary of  Performance Test Results  and Comparative
     Data for  Designated Equivalent  Methods for  S0«.   QAD/M-79.12, Environmental Monitoring
     Systems Laboratory, Research Triangle Park, NC,^December 1979a.

U.S.  Environmental  Protection  Agency.   Ambient  Air  Monitoring  Reference   and  Equivalent
     Methods.  40 CFR 53:952-986, July 1, 1979b.

U.S.  Environmental  Protection  Agency.  National   Primary  and  Secondary Ambient  Air Quality
     Standards.  Appendix B - Reference method for  the determination of suspended particulates
     in the atmosphere (high volume method).   40 CFR 50:12-16, July  1, 1979c.

U.S. Environmental Protection Agency.  Methods for  Chemical Analysis of Water and Wastes.  EPA
     600/4-79-020, U.S.  Environmental Protection Agency, Environmental Monitoring and Support
     Laboratory, Cincinnati, OH, March 1979d.

Volmer,  W. , and F.  Z.  Frohlieh.   Determination of small  quantities  of  sulfate by turbidity
     measurement.  Z. Anal.  Chem. 126:401-404, 1944.

von  Lehmden,  D.  J. ,  and C.  Nel.son.   Quality Assurance  Handbook  for Air Pollution Measurement
     Systems, Volume I  - Principles.  EPA-600/9-76-005, U.S. Environmental  Protection Agency,
     Research Triangle Park, NC, May  1977.

Waggoner, A.  P.,  and R. E. Weiss.   Comparison of  fine  particle  mass  concentration and light
     scattering in ambient  aerosol.   Atmos. Environ. 14:623-626,  1980.

Waggoner, A. P., M. B.  Baker, and R.  J. Charlson.   Optical absorption by atmospheric aerosols.
     Appl.  Opt. 12:896, 1973.

Waller, R.  E.  Acid droplets in town  air.  Air Water Pollut. 7:773-8, 1963.

Waller, R.  E., A. G.  F. Brooks, and J. Cartwright.  An electron microscope study of
     particles in town air.   Air Water Pollut. 7:779-86, 1963.

Warren Spring  Laboratory.   Measurement of Atmospheric Smoke  and  Sulphur Dioxide:  Reproduci-
     bility of Results.  Stevenage, England, August 1962.

Warren Spring  Laboratory.   National  Survey of Smoke and Sulphur  Dioxide:  Instruction Manual,
     Stevenage, England, 1966.

Warren  Spring  Laboratory.   The  Investigation  of  Atmospheric Pollution  1958-1966.   Thirty-
     second Report.  HMSO,  London,  England, 1967.


                                             3-117

-------
Warren Spring Laboratory.  National Survey of Air  Pollution,  1961  -  1971.   Vol.  5:   Scotland  -
     Northern  Ireland.   Chapter  15:    Accuracy of  Data  and Representativeness  of  National
     Survey Data.  Stevenage, England, January  1975.   pp.  111-118.

Wartburg, A.  F.,  J.  B.  Pate, and  J.  P.  Lodge, Jr.  An  improved gas sampler for air pollutant
     analysis.  Environ. Sci. Technol. 3:767-768,  1969.

Washburn, H.  W. , and  R.  R.  Austin.   Air Pollution.  In:   Proceedings of the  U.S.  Technical
     Conference on Air Pollution.  McGraw-Hill, New  York,"NY,  1952.

Wedding, J.  B.  Ambient Aerosol Sampling:  History,  Present  Thinking,  and  a Proposed Inlet for
     Inhalable  Particulate  Matter.   Presented  at  the  73rd  Annual  APCA  Meeting,  Montreal,
     Canada, June 1980.

Wedding, J. B.,  A.  R.  McFarland, and J. E. Cernak.  Large particle  collection  characteristics
     of ambient aerosol samplers.  Environ. Sci. Technol.  11:387-390,  1977.

Wedding, J. B.,  M.  Weigand, W. John, and S. Wall.   Sampling effectiveness  of the  inlet to the
     dichotomous sampler.  Environ. Sci. Technol.  14:1367-1370, 1980.

Wehry, E. L.,  and G.  Mamantov.  Matrix  isolation  spectroscopy.  Anal.  Chem.  51(6):  643A-636A,
     1979.

West, P. W.,  and G.  C.  Gaeke.  Fixation of sulfur dioxide as  disulfitoroercurate (11) and  sub-
     sequent colorimetric estimation.  Anal. Chem. 28:1816-1819, 1956.

West, P. W., and T.  P.  Ramachandran.  Spectrophotometric determination of  nitrate  using chrom-
     otropic acid.  Anal. Chim. Acta 35:317-324, 1966.

Whitby, K. T., W. E.  Clark,  V. A. Marple, G. M. Sverdrup,  G.  J. Sem, K. Willeke, B.  Y.  H.  Liu,
     and D.  Y.  H.  Pui.   Characterization  of California  aerosols—I.  Size distributions of
     freeway aerosol.   Atmos. Env. 9:463-482, 1975.

Whitby,  K.  T.,  and W.  E.  Clark.   Electrical aerosol  particle counting and  size  distribution
     measuring system for the 0.015 to 1.0 mm size range.  Tellus. 18:573-586,  1966.

Whitby, K. T.  and B.  Y.  H. Liu.  Advances in Instrumentation and Techniques for Aerosol  Gener-
     ation and Measurement.  Particle Lab. Pub. No.  216; avaiable  from Mech.  Eng.  Dept., Univ.
     of Minn., Minneapolis,  MN, 1973.

Whitby,  K. T.,  and  K.  Willeke.  Single  particle optical  counters:   principles and field  use.
     In;  Aerosol  Measurement,  a  Workshop.    University of  Florida,  Gainesville,  FL, March
     24-26,  1976.  D. A.  Lundgren, M. Lippmann, F.  S. Harris, Jr.,  W.  E.  Clark, W.  H.  Marlow,
     and  M.  D.  Durham,  eds. ,  University  Presses  of  Florida,  Gainesville,  FL,  1979.   pp.
     145-182.

Willeke, K. ,  and  J.  J.  McFeters.   Calibration  of the CHAMP  Fractionator,  Particle Technology
     Publication No.  252, University of Minnesota, Minneapolis, MN,  March  1975.

Wilsdon, B.  H., and F.  J. McConnell.  The measurement  of atmospheric sulfur pollution by means
     of lead peroxide.  J. Soc. Chem. Ind., 53:385-81,  1934.

Wilson, B. D.  Nitrogen  and  sulfur  in rainwater  in New York.   J. Am. Soc.  Agron, 18:1108-1112,
     1926.

Wood, E. D. ,  F.  A.  J.  Armstrong, and F. A. Richards.  Determination of nitrate in seawater by
     cadmium-copper reduction to nitrate.  J.  Mar.  Biol.  Assoc. U.K.  47:23-31,  1967.
                                              3-118

-------
Woods,  D.  C.   Measurement of  particulate aerosol  mass  concentration  using  a piezoelectric
     crystal microbalance.   In:  Aerosol  Measurement, D. A.  Lundgren,  and M.  Lippmann, eds.
     [Workshop].  University Presses of Florida, Gainesville,  FL, 1979.  pp. 119-130.

World Health Organization.  World Health Organization  Selected  Methods  of Measuring Air Pol-
     lutants.  WHO Offset  Publication No.  24, World Health Organization, Geneva, Switzerland,
     1976.

World Meteorological  Organization,   Manual  No.  299,  WMO Operations  Manual  for Sampling and
     Analysis  Techniques for  Chemical  Constituents  in  Air and Precipitation.   WMO Geneva,
     Switzerland, 1971.

World Meteorological  Organization.  Letter  PR-2569.  The Operation of  the Precipitation Refer-
     ence Laboratory for WMO Members, June 30, 1975.

World Meteorological   Organization.   Environmental  Pollution  Circular No.  6—Results  of the
     First  Analysis   of  Reference  Precipitation  Samples.   Geneva,  Switzerland,  October 25,
     1976.

World Meteorological  Organization.   Report  of  the  Third Analysis  on Reference Precipitation
     Samples.  Geneva, Switzerland, April 10, 1980.

Yasa, Z. , N.  M.  Amer, H. Rosen, A. D. A. Hansen,  and  T. Novakov.  Photoacoustic investigation
     of urban aerosol  particles.  Appl. Opt. 18:2528,  1979.

Zall, D. M., D. Fisher, and M.  Q.  Garner.  Photometric  determination of chlorides in water.
     Anal.  Chem. 28:  1665-1668, 1956.

Zoccolillo,  L.,  A.  Liberti,  and D. Brocco.  Determination of  polycylic  hydrocarbons in air by
     gas chroroatography  with high  efficiency packed  columns.  Atmos.  Environ,  6:715, 1972.
                                             3-119

-------
                                     APPENDIX 3A
FILTER:  Gelman Type A, glass fiber

AP, cm Hg .
V, cm/sec
D , pm
0.035
0.10
0.30
1.0
FILTER: Ghia S2
1
11.2

<0.0001
<0.0001
<0.0001
<0.0001
37PJ 02, teflon
1.5
16.9
PENETRATION
<0.0001
<0,0001
<0.0001
<0.0001
membrane, 2.0 nn
3
32.7

<0.0001
<0.0001
<0.0001
<0.0001
n pore
10
108

0.0008
0.00054
<0. 00007
<0.0002


&?, cm Hg
V, cm/sec
Dni Mm
P
0.035
0.10
0.30
1.0
FILTER: Whatman
1
23.4

<0.0002
<0. 00006
<0. 00007
<0. 00007
No. 1, cellulose
3
64.1
PENETRATION
0.0011
0.00008
<0. 00007
<0. 00009
fiber
10
187

0.0005
<0. 00024
<0. 00022
<0. 00008










AP, cm Hg
V, cm/sec
Dp, Mm
. 0.035
0.10
0.30
1.0
1
6.1

0.56
0.46
0.16
0.019
3
17.4
PENETRATION
0.52
0.43
0.044
0.034
10
47.6

0.34
0.13
0.0049
0.0044
30
102

0.058
0.0071
0.00051
0.00042
Table A-l.  Fractional penetration by particle size and face velocity for three
            selected filter types (Liu et al., 1978a).
                                        3-120

-------
SCALE, in  SCALE, cm
      0
      3-
      4-J
         -6
         -8
         -10
        FLOW
BUG SCREEN
16X16 MESH
                                                                  SHIELD
                                         FLOW TO DICHOTOMOUS SAMPLER
                 Figure 3A-1, Early inlet for the dichotomous sampler.

                 Source: Stevens and Dzubay (1978).
                                      3-121

-------
     COLLECTING SURFACE-~
      8 DIRECTIONAL.
         VANES
           I

ENTRANCE   I
  PLANE •—
                             I

                     44»H  I i I
                     f
                              a
                                                    • INLET
                                                    HOUSING
                                  i PARTICLE PATH
                               EXIT PLANE
Figure 3A-2. Wedding IPM inlet, section view, not to scale.
Source: Wedding (1980).
                           3-122

-------
                FLOW
                    FILTER/
            FLOW
            CONTROLLER
            FLOW
            RECORDER
                                     INLET COVER
Wt.~65 Ibs.
  Figure 3A-3. TSP Hi-VoI.
              3-123

-------
                           16.67 1/min

                                 TOTAL F LOW, Q
                                        SEALED HOUSING
   ACCELERATION
       NOZZLE
FRACTIONATION
     ZONE

  COLLECTION
    NOZZLE
         LARGE PARTICLE
             FLOW.fQ	
    LARGE PARTICLE
  COLLECTION FILTER"

1

r

               SMALL PARTICLE
              	FLOW(l-f)Q
                SMALL PARTICLE
              -COLLECTION FILTER
                        TO F LOWM ETE R
                          AND PUMP
                           1.67 l/min
TO F LOWM ETE R
   AND PUMP
    15.0 l/min
                   Figure 3A-4. Dichotomous sampler separator.

                   Source: Loo etal. (1979)
                                  3-124

-------
  MAST SUPPORT
AND VACUUM LINE
                    CRITICAL ORIFICE
                     9.0 LITER/WIN.
                                                                  RUBBER
                                                                  VACUUM HOSE
CONNECTIONS
                                                          CYCLONE SEPARATOR
                Figure 3A-5.  Chess cyclone sampler and shelter assembly.

                Source:  Barnard (1976).
                                       3-125

-------
                                            AIB INLET
                                                   TO PUMP
                                                          47 mm
                                                          AFTER-FILTER
                                                     CYCLONE
                   47 mm
            TOTAL FILTER
                         TO PUMP

 Figure 3A-G. Assembly for sampling with a total filter and cyclone in parallel,

, Source: John et al. (1978)
                                   3-126

-------
                                       INLET
                        FILTER
                   FLOW
                   CONTROLLER
                   FLOW
                   RECORDER
                                                  FLOW
                                               STANDARD HI-VOL
                                                   SAMPLER
Figure 3A-7.  Size-Selective Inlet (SSI) Hi-Vol.
                   3-127

-------
                                       AIR FLOW LINES
                                                               DUST FREE AIR
                                                              PENETRATING
                                                              DUST CLOUD
              DUST TRAJECTORIES
                                  (a)
  75cm
                                            DUCT FLOOR AREA 8400 cnT
                                            AIR FLOW RATE 76 I/mm
         24 TRAYS
                                  (W
Figure 3A-8. The horizontal elutriator designed to match the BWIRC deposition curve.

Source: Hamilton and Walton (1961).
                                    3-128

-------
 STAGE 1 <
 STAGE 2<
STAGE
 AFTER
 FILTER
NOZZLE


JET EXIT
 PLANE
                                                          IMPACT1ON
                                                            PLATE
                                                         FILTER
                          TO VACUUM PUMP


            Figure 3A-9. Schematic diagram of a cascade impactor.

            Source:  Marple and Willeke (1979).
                               3-129

-------
LARGE PARTICLE
 FRACTIONATOR
  IMPACTOR UNIT-—

           SPACER
                       HANDLE

                       IMPACTOR
                        PLATE
\    \
           FLOW SENSOR
                                             COVER
                                             X
                                                      .
                                                       INLET
                    T.
                                                     WATER DRAIN
                     ,GLASS FIBER
                      IMPACTION SURFACE
                                                FINAL FILTER
                                          VACUUM PUMP
         Figure 3A-10. Cross section schematic of the CHAMP aerosol sampler.

         Source: Ranade and Osdell (1978).
                                3-130

-------
                                                   ROTARY SEQUENCING VALVE
                                                       SUCTION PUMP
                          Schematic arrangement of sampling apparatus.
                          Sampler capable of sight-day sequential operation.
Figure 3A-11. British smoke shade sampler.
          3-131

-------
PUSH TO TEST VALVE
                              J~A
    THREi WAY VALVE
                                      
-------
    10"3g, mg
                                               SEASAND-
                                          HUMAN HAIR-
o
p
cc
<
a.
LL
O

CD
O
O
 10"9g, ng
1(T12g,pg
    10"15g,fg
    10'18g, ag
         RAGWEED POLLEN.

       POTATO STARCH-

     CORN STAHCH-

RED BLOOD fELLS •

   BACTERIA-
           10A
                 100A

                0.01 fan
     lOOnm             0.01mm

     0.1 jim     1 fim      10 nm

                I         I   '
  LOG PARTICLE DIAMETER (p = Ig/oc)
0.1 mm
                                                                        10
1.0 mm
            TRANSMISSION ELECTRON MICROSCOPE'
                                                  STEREOBINOCULAR
                                                     MICROSCOPE
                                             MONO OBJECTIVE
                                           OPTICAL MICROSCOPE
                        SCANNING ELECTRON MICROSCOPE
                CHOICE OF MICROSCOPE FOR PARTICLE SIZE MEASUREMENT


   Figure 3A-13. Relationship between particle size, diameter and number of
   atoms for the light and electron microscope range.

   Source: McCrone and Delly (1973).
                                         3-133

-------

-------
                                 4.   SOURCES AND EMISSIONS

4.1  INTRODUCTION
     This chapter  highlights the magnitude and characteristics of  natural  and anthropogenic
sources and emissions  of  particulate matter and sulfur oxides.  Natural emissions are defined
as those not caused by human activities, such as volcanoes and the biosphere.  Manmade sources
include stationary  point  sources (e.g.,  utility power plants,  industrial  facilities,  etc.),
fugitive  industrial  and  non-industrial  sources  (e.g.,   roadway dust),  and  transportation
sources (e.g., vehicle  exhausts).   Each of these emissions categories is discussed further in
this chapter.
     Chapters  4,  5,  and 6  present  the  information concerning  the  relationship  between
emissions and ambient air concentrations.   The information in Chapter 4 concerning the sources
and emissions of particulate matter and sulfur oxides relates directly to the chapters in this
document  which  discuss  pollutant  effects  on  visibility,   acidic   deposition,  and  health.
Chapter 5 summarizes measured ambient pollutant concentrations and characteristics.   Chapter 6
presents  what is  known  about  the  complex processes  that  alter  and  disperse the  emitted
substances as they move through the atmosphere.
     The issue of  the relationship  between emission  intensity and possible effects on humans
is  important,  but  the proximity  of  emissions  to  humans can  often be more  important  than
relative  intensity.   For example,  mass  emissions  from residential  fuel  combustion  (home
heating) and transportation  sources  are minor on a national  level.  Since they are emitted in
highly  populated areas and   close  to  ground  level,  however, they  are  more likely to  affect
human  health and welfare.   On the other  hand,  dust from  unpaved roads appears significant in
many areas, but  unpaved roads are more prevalent in rural areas, and their  influence tends to
be highly localized.  Conversely, although some natural source emissions can be fairly intense
(volcanic ash or sulfur from marshlands, for example), their effects are lessened, in gereral,
because  they  tend  to  be  distributed  fairly  broadly  nationwide.   Consequently,  simple
comparisons of total national tonnages of manmade vs natural  emissions will  seldom reflect the
impact  that  localized manmade sources can  have  on  an area's air quality.   For these reasons,
certain manmade  emission  sources,  particularly stationary  point  sources,  have  been given a
greater share of the attention in this chapter.
     A number of other  issues are discussed here briefly or not at all.  Predictions of future
emissions  trends   have  not  been  presented  because  of  the  complexities  of  supporting
assumptions.   Documents  in  which   adequate   discussions  of  assumptions  can  be  found  are
referenced.  In many cases, data on the particle size distribution and chemical composition of
particulate  emissions  are  incomplete  or  inadequate.   Available  data  have been  briefly
summarized.   Also, discussion of  the  effects  of control devices on emission  particle  size
distributions  has  been  limited.   Documents  that  discuss  these  effects  thoroughly  are
referenced.

                                              4-1

-------
4.2  DATA SOURCES AND ACCURACY
     The most important information presented in this chapter concerns emission quantities and
characteristics.   Though  this  information  was  gathered  from  the  best  and  most  recent
literature  available,  problems are  still  apparent.   Specifically,  estimates  of  emission
quantities vary, as do those of emission characteristics.
     Emission  quantities  are  typically estimated  using  emission factors.   (The  impossible
alternative would  be direct  measurement of pollutants  from each  emission point.)  Emission
factors relate  the quantity  of pollutants emitted  to  an indicator  of  activity  such as pro-
duction  capacity  or quantity  of  fuel  burned.   Because emission  factors  are statistical
averages,  they do  not precisely  reflect  emissions from  individual  sources.   When  a large
number  of  sources  are  considered,  however, a  reasonable estimate of  total  emissions can be
obtained.  Therefore, the percentage error associated with emission estimates decreases as the
geographic  area studied increases  (from the local  to  the  regional  to  the  national scale).
     Table  4-1  illustrates  the  variability  of  specific  emission  estimates  that can  be
associated with the use of  different estimation approaches.  In the table, "estimates" refers
to National Air Pollutant Emission Estimates, 1970-1978  (U.S. Environmental Protection Agency,
1980a).  Emissions totals in that document were obtained from one calculation performed at the
national level  by  use of total national activity levels and national average emission factors
for each source category.   "NEDS" (National Emissions Data System) refers to the  1977 National
Emissions  'Report  (U.S.  Environmental  Protection  Agency, 1980b).   NEDS  national  emissions
totals were obtained by adding the emissions from individual facilities and may be affected by
mistakes,  biases,   and  omissions.     Therefore,  the   "Estimates"  are  judged  by  the  U.S.
Environmental  Protection  Agency  to  be  the  most  reliable  national  estimates  presently
available.   The NEDS State emissions totals, presented in Section 4.5.2 are the best available
at the  State  level.    However,  neither of these totals  should be regarded as precise, for the
reasons given above.  Nevertheless,  such emission estimates do  provide useful indications of
the relative contributions from various source categories.

                TABLE 4-1.   TWO EPA ESTIMATES OF 1977,-EMISSIONS OF
             PARTICULATE MATTER AND SULFUR OXIDES (10° METRIC TONS PER YEAR)

Source Category
Fuel combustion
Industrial processes
Solid waste disposal
Parti cul ate
Estimates
4.8
6.4
0.5
Matter
NEDS
3.
3.
0.


Sulfur
Estimates
6
9
4
22
4
0
.2
,2
.0
Oxides

NEDS*-
22.
5.
0.
,1
.1
,0
                                              4-2

-------
     Attempts to  obtain more  than  rough estimates  of fugitive  industrial  and nonindustrial
participate  emissions  also   present   problems.    Industrial  process  fugitive  particulate
emissions, or  process fugitives, include  most industrial  particulate emissions  not  passing
through  a stack  or  another  identifiable  emission  point.   Process  fugitive emissions  are
difficult to  estimate because  of the  lack  of engineering  data  and  adequate  information on
emission  factors.   In one reference (Zoller  et  al., 1978) these emissions  were  estimated at
3.4  x  10   metric  tons  per  year.   However,  the  U.S.  Environmental  Protection  Agency's
"Estimates"  include  "rough  estimates  of  fugitive  particulate emissions  from  industrial
processes."  Therefore,  the particulate emission  estimates  alluded  to under  the industrial
processes category  discussed  in this .chapter probably include part of the emissions listed as
process fugitives in Table 4-4.
     Estimates  of  nonindustrial  fugitive  particulate emissions  vary quite  significantly in
some cases.   Cooper et  al.  (1979)   estimated  annual emissions from  entrainment  of dust from
unpaved  and  paved  roads  at 290  x  10   metric tons and  7.2  x 10  metric  tons, respectively.
The U.S.  Environmental  Protection Agency (1980b) estimated emissions from the same categories
at 35 x 10  metric tons and 4.7 x 10  metric tons,  respectively.   The differences can probably
be attributed to the use of different assumptions and methods of calculation.
     Finally, emission estimates by particle size and chemical composition also vary depending
on the  specific  information and estimation method used.   Most of these estimates are based on
data from emissions sampling and analysis studies.   Although these studies probably exhibit a
high  degree  of  accuracy on  a  case-by-case  basis,  data  from a  small  number of individual
sources  should  not  be  used to  make generalizations.  Emission  characteristics, as  well  as
emission  quantities,  are highly  dependent on a  number of  source-specific factors,  such as
source  and fuel  characteristics and operating conditions.   For example, the size distribution
of particulate  emissions  from a given utility boiler can be altered significantly by changing
the boiler load.   Therefore,  the emission characteristics from a particular source could vary
from the information presented in this chapter.
4.3  NATURAL SOURCES AND EMISSIONS
     Knowledge  of natural  sources  and emissions  of particulate  matter  and sulfur, including
sulfur  oxides,  is  important  for understanding air pollution.   Baseline  concentrations  in
continental and marine  air  represent natural exposure levels and thus provide a reference for
comparing concentrations in air polluted by emissions from manmade sources.
     Significant  natural  sources of particulate  matter and  sulfur,  including reduced sulfur
which can  become  oxidized to sulfur oxides in the atmosphere (see Chapter 6), are terrestrial
dust, sea  spray,  the biosphere, volcanoes, and wildfires.   Estimates of emissions from these
natural  sources  in the U.S. are  described  in more detail in subsequent  sections.   Table 4-2
presents a summary  of natural source emission totals and characteristics.
                                              4-3

-------
4.3.1  Terrestrial Oust
     Terrestrial  dust  is transferred to the  atmosphere  by the action of wind- on the earth's
soils and  crustal  materials.   Theoretical  and experimental  studies  (Gillette, 1974) indicate
that sand  grains,  produced by the weathering  of  rocks and soils and moved by wind, cause the
pulverization of soil minerals, as in sandblasting, to produce particles.  These particles may
become airborne and may be transported through the atmosphere for considerable distances.  For
example, dust from the Sahara Desert may be carried by air currents across the Atlantic Ocean
as far as Florida and Barbados (Delany et a!., 1967; Junge, 1957).
     The amounts  of  global terrestrial  dust  have  been estimated at 180 x 10  metric tons per
                                                   fi                      *
year (Robinson  and Robbins,  1971) and 100-500 x 10  metric  tons per year  (National Research
Council  1979).   Calculations by  Vandegrift et al.  (1971),  based on  soil  conservation data,
resulted in estimated U.S. natural dust emissions of 57 x 10  metric tons per year.
     Terrestrial  dust  in  the atmosphere  is  composed primarily  of  seven  major  elements,  -
silicon,  aluminum, iron,  sodium, potassium,  calcium,  and  magnesium;  organic  material;  and
trace elements  (Miller et al., 1972).  The  major  elements are present  in  aerosol  samples to
nearly  the  same   extent  as  in  earth  crustal  material  (Miller  et  al.,  1972;  Lawson  and
Winchester,  1979a).   Atmospheric   concentrations  of  many  trace  elements,  however,  are
10- to 1000-fold  higher  than would  be  expected from  physical dispersion  of soil materials.
These  anomalous trace  element enrichments  have  been  observed in  many parts of  the  world,
including  northern Canada (Rahn, 1974),  the  South  Pole  (Zoller  et al., 1974),  and South
America  (Adams  et al.,   1977).   Table  4-3  summarizes geometric  mean enrichment factors,
relative to  aluminum,  for various elements  according to Rahn's  compilation  of all published
data up to 1976 (Rahn,  1976).
     The  atmospheric  enrichment  sources  of  these elements are  unknown,  but  transport from
polluting  industries  (Rahn,  1974),   natural  rock volatility  (Goldberg, 1976),  and biogenic
emanations (Barringer, 1977)  have all been suggested.   In general,  not enough is known about
element  ratios  in the  natural  atmosphere  to detect  a pollution  component,  thus  a high
anomalous  content in particulate  material  cannot be  related arbitrarily  to air pollutants.
     Most  terrestrial  dust particles are  greater than  2 pm  in  diameter  (U.S. Environmental
Protection Agency, 1979).   The major  element constitutents  of  terrestrial   dust  also occur
principally as  coarse  particles.   Size-fractionated particle  samples  indicate that more than
90 percent of the mass occurs on the first three impactor stages, representing particle sizes
of  >4,  4-2,  and  2-1 urn aerodynamic diameter  (Winchester et  al.,  1979).   The  low relative
abundance of submicron  silicon,  iron, and other major dust  constituents reflects the greater
*
 Includes unknown amounts of indirect manmade contributions.
                                              4-4

-------
                  TABLE 4-2.   SUMMARY OF NATURAL SOURCE PARTICULATE AND SULFUR EMISSIONS"

Estimated U.S. <
(106 metric ton;
Source category Parti cul ate
Terrestrial dust 57
f*
Sea spray 5.5
Biosphere 20
Volcanoes Variable
Wildfires 0.5 - 1.0
TOTAL -V-84+


^missions
; per year). Parti cul ate characteristics
Sulfur Size range data Chemical composition
10% <1 fjrn Al, Ca, Fe, K, Mg, Na,
Si, organics, trace
elements
22% <3 (jm Seawater, organics
1.2 - 5.5 Unknown Organic aerosols, trace
metals
Variable ^5% <1 urn Al , Ca, Fe, K, Mg, Si,
Trace elements
80% <1 |jm Organics, trace minerals
., * A _
 All data are referenced in the text.
 When oxidized, one metric ton of sulfur equals 2 metric tons of sulfur dioxide, SO,,.
clncluding 0.7 x 10  metric tons of sulfate aerosol.
 Predominantly reduced sulfur compounds.

-------
              TABLE 4-3.  AEROSOL ENRICHMENT FACTORS RELATIVE TO Al
                    EFa                 Elements
                    0.7-7               Li, Na, K, Rb
                                        Be, Mg, Ca, Sr, Ba
                                        Sc, Y, lanthanides
                                        Al, Ga, Tl
                                        Si, Ti, Zr, Hf, Th, U
                                        Mn, Fe, Co, Nb
                                        F, P
                    7-70b               Cr, Cs, V, W, B, Ni,  Ge
                    70-400b             H, In, Cu, Mo, Bi, Zn, As
                    400-4000b           I, Hg, S, Cl, Au, Ag, Sn, Sb,
                                        Pb, Br, Cd, Te, Se, C, N
aGeometric means of element ratios to Al, relative to geochemical average earth
 crustal material.
                              EF=  (e1emen^A1 Aerosol
                                    (el ernent/A 1) c r us t
 Anomalously enriched elements arranged in order of increasing EF.
Source:  Based on Rahn (1976).
                                          4-6

-------
amount  of  energy  needed In  order  for  fine  particles to  be  generated from  soils by  the
wind-driven sandblasting  mechanism.   This  energy  is normally not provided  by the atmosphere
near the ground.
4.3.2  Sea Spray
     Aerosol droplets  are generated at the  ocean  surface  by the action  of  wind,  principally
through  a  process whereby  air bubbles  become entrained  and rise to  burst at the  surface.
Robinson and Robfains  (1971)  estimated global  emissions  of particles  from sea  spray  at  900 x
10   metric tons  per year,  including 120  x  10  metric  tons  of  sulfate  aerosol   per  year
(Eriksson,   1959,   1960;   Robinson  and Robbins,  1968).   Assuming 10 percent  of  the  annual
production penetrates continental  areas  (Eriksson, 1959),  and assuming the impact on the U.S.
                                                                          3                  3
is  proportional  to  the  ratio  of the U.S.  to  global  coastline (12 x 10   miles:   200 x 10
miles),  approximately  5.5 x 10   metric  tons  of  sea spray  particulate per  year (including
0.7 x 10  metric tons of sulfate aerosol  per year) impact on U.S. coastal areas.
     Sea spray is  composed  of seawater,  organic materials, and surface-active materials which
may  be  concentrated  into the 0.05  to 0.5  um thickness of bubble surface  (Maclntyre,  1974).
The  surface-active material  may  be  of natural  or pollutant  origin  and may  include organic
molecular  films  and  organic  and  inorganic particles  including viruses, bacteria,  and other
microscopic organisms (Blanchard  and Parker,  1977; Duce and  Hoffman, 1976),   Such materials,
by  becoming  components  of sea  spray aerosol  droplets, may be carried  through the atmosphere
far  from the point  of  origin.   The  potential  for virus transfer from  coastal  waters  to the
atmosphere and transport  by  winds inland to  inhabited areas  has been demonstrated (Baylor et
al., 1977), although  the process is not clearly understood.
     Because of  differences  in the  mechanics of  droplet  formation,  differences  ih chemical
composition may exist  (Berg  and Winchester, 1978).  For example, some droplets may or may not
contain  the  surface-active  particulate matter scavenged from the water  column by the  rising
bubbles.   Chloride,  bromide,  and  iodide  also  may  be  present  (Moyers  and Duce, 1972a,b),  The
general  processes through  which   sea spray  droplets are  formed and  transported  have been
described  by  a number  of authors,  including  Blanchard and  Woodcock (1957)  and  Wallace and
Hobbs  (1977).  The size distribution of sea  spray particles  by weight percent, as documented
by  Taback  et al.  (1979), is  as follows:   >10 um,  24 percent; 3-10 (am, 54 percent; 1-3 urn, 20
percent; <1 um, 2 percent.
4.3.3  Biogenic Emanations
     Plants  emit  particulate  matter in  the  form  of  organic aerosols,  trace  metals,  and
nutrients.   Global  emissions  of  volatile  organic compounds  released  from plants  have been
estimated  at 200 x 10  metric tons  per year  (Went,  1960).  The U.S.  total  would  probably be
less than  20 x 10 metric tons per year.  Isoprene derivatives such  as terpenes, caroteniods,
and  other compounds  are believed to predominate and are likely  to be partially   oxidized,
resulting  in blue  haze  and  submicron condensation  nuclei  (Went, 1960; Went  et  al.,  1967;
Rasmussen  and Went, 1965; Schnell and Vali, 1972,  1973).
                                              4-7

-------
     Trace  metals  have  long  been known  to occur in  fluids  secreted  by plants.   Radiotracer
strontium is transferred from plant foliage to the atmosphere, presumably in particles (Moorby
and Squire, 1963) which may be affected by electric fields (Fish, 1972).  Transpiration caus.es
the  transfer of both cations and  anions to  the  atmosphere (Nemeryuk,  1970).   Twenty-seven
trace elements  have  been identified in exudates from  coniferous trees (Curtin et al., 1974).
Radiotracer experiments  using zinc and lead show that particles greater than 5 pm in diameter
contain  most  of the  metals released  (Beauford  et al., 1975,  1977).   Sulfur,  potassium, and
phosphorus  have  also  been associated  with  tropical  forests and  occur  in large  aerosol
particles (Lawson  and Winchester, 1979b).  The metal  content of plant-derived aerosols is so
high that  several  investigators  suggest  that  it might serve as an  indicator for geochemical
prospecting (Barringer, 1977; Curtin et al., 1974).
     The terrestrial  and marine  biospheres,  while not direct  sources  of  sulfur  dioxide, are
significant  sources  of  reduced   sulfur  compounds.    Volatile   reduced  sulfur compounds  are
released to the atmosphere via microbiological  processes  and may become  oxidized  to  S0? and
sulfate.   The  compounds  released  included hydrogen  sulfide  (H?S),  dimethyl  sufide  (DMS),
dimethyl  disulfide   (DMDS),  carbon  disulfide  (CS2),  carbonyl  sulfide  (COS),  and  methyl
mercaptan  (CH-SH)  (Lovelock  et  al.,  1972;  Rasmussen,  1974;  Lovelock,  1974;  Adams  et  al.,
1979a).   Transformation  to  H-S and thence to SCL and SOT is predicted (McElroy et al., 1980),
as well  as direct oxidation to S02 and/or SO,.
     The classic method of estimating biogenic sulfur emissions has  been to identify the net
difference  between input  from known  sources  and removal  by  scavenging processes as  being
indicative  of an  unmeasured source, namely a widespread source  in the biosphere.   A number of
previous estimates of global  emissions of reduced sulfur compounds range from 64 x 10  metric
tons per year (land)  and 27 x 10  metric tons per year (ocean) (Robinson and Robbins, 1968) to
3 x 10  metric  tons  per year (land) and 34 x 10  metric tons per year (ocean) (Granat et al.,
1976).   Granat's estimate,  scaled down to the U.S.,  would result in 0.2 x 10  metric tons per
year (land) and about 2-5 x 10  metric tons per year (ocean) (based on Galloway and Whelpdale,
1980).   These estimates, however,  were derived indirectly as balances for other sulfur fluxes.
     Results  of  recent field  monitoring  studies  conducted by Maroulis  and Bandy  (1977),
McClenny et al.  (1979), and Adams  et al. (1979a) yield slightly different estimates.   Adams et
                                                                — o  ™ 1
al. (1979b) calculated  a mean annual sulfur  flux  of  0.02  g S m  yr  , weighted over a number
of  Eastern  U.S.   soil  types,  including marshes.   The entire U.S.  would   probably  average
                _p  -i
0,02-0.05 g S  m  yr    (Adams,  1980).  When  these numbers are  applied  to the  entire  earth's
land area (about  56  million square miles) the result is about 3-7 x 10  metric tons per year.
The land area of the U.S. (3.6 million square miles)  emits about 0.2-0.5 x 10  metric tons per
                              -2  -1
year based  on 0.02-0.05 ig S m  yr  .  The impact of marine biogenic activity would be limited
primarily to  coastal  areas.   Sulfur  emissions from  marine biogenic activity are probably cm
the order  of 1 x 10   metric  tons  per  year (based on Galloway and Whelpdale,  1980).   These
lower estimates of  sulfur emissions on a national  scale do not preclude significant localized
                                              4-8

-------
biogenic sulfur emissions,  especially  in areas where wetlands  are  prevalent (Henry and Hidy,
1980).   Marshes  and tidal  flats  may  have high local sulfur-gas  production  (Hitchcock,  1976;
Hitchcock et  al.,  1980),  but  since the  total area  in  the U.S.  covered by  such  features is
relatively small, the  contribution  to  total background sulfur is modest.   Thus, the inventory
of  terrestrial  biogenic  sulfur  emission  should  not overemphasize  the  wetland  areas  as  a
source.  In the western  U.S.,  more arid soils would be expected to have a much reduced sulfur
emission rate, but detailed study in this area is lacking.                                ,'
4.3.4  VolcanicEmissions
     Emissions  from volcanic  eruptions  and  fumaroles may  contribute to global  atmospheric
background levels of particulate  matter and sulfur.  Volcanoes  are one of the few sources of
atmospheric particles  and sulfur whose  effects  can be felt at great  distances.   Plumes from
volcanoes intense enough  to inject material into  the upper troposphere or lower stratosphere
(about  10  to 15 miles above  the earth's surface)  have  been  tracked  great distances before
removal (Fegley et  al.,  1980).  The famous eruption  of  Krakatoa in 1883 injected enough dust
and sulfur into the stratosphere to cause brilliant sunsets thousands of kilometers away and a
global reduction of incoming solar radiation (Wexler, 1951a,b).
     Until recently, volcanic activity has been relatively insignificant in the United States.
The Mt. St.  Helens  eruption was only the second this century in the contiguous United States.
Mt. Lassen, California,  in 1915,  was the first.   About 20 volcanic eruptions have occurred in
Hawaii and Alaska since 1900.
     With the 1980 eruption of Mt. St.  Helens in the Pacific northwest, considerable attention
has been focused on the potential impact of volcanoes on the atmosphere and air quality.  Away
from the  immediate downwind area,  volcanic  impacts  can  probably be related  to  a cycle that
starts with  injection  into the stratosphere of  dust  and  sulfur gases, oxidation and reaction
of  the  sulfur to  form particulate compounds,  and  finally  injection of the particles into the
troposphere, where they are scavenged.   Since injection from the stratosphere occurs mainly in
low pressure systems,  it is likely that precipitation scavenging predominates.
     The  average  global  volcanic emission  rates of particles  and  sulfur  compounds have been
estimated  by  a number of  investigators.   Robinson and Robbins  (1971) estimated the average
global emission rate of small particles  (the persistent fraction) at 3.6 x 10  metric tons per
year.   Airborne  measurements   and  observations  made during  the  1976 eruption  of  the St.
Augustine volcano (Alaska)  led to particulate emissions estimates for a 1-year period for that
particular volcano  of  6  x  10   metric  tons  for particles of 0.01 to  66 (jm in size and 0.25 x
10  metric tons for particles  0.01 to 5 (am in size  (Stith et al., 1978).
     Estimates  of  global  volcanic  sulfur emissions,  as  documented by Granat et al. (1976),
range  from  0.75 to 3.75 x  10   metric  tons  per year.  Emissions of SO, for a 1-year period at
St. Augustine were  estimated at  0.1 x  10  metric tons (or  0.05  x  10  metric tons of sulfur)
(Stith et al.,  1978).  The  St.  Augustine volcano also emitted lesser quantities of H,S.   Stith
                                                                      £             ^
et  al.  (1978)  estimated  global  volcanic emissions of H2S  at  1 x 10  metric tons per year.
                                              4-9

-------
     Particles collected  from the St. Augustine eruptions were composed primarily of silicon,
aluminum, magnesium, calcium, and iron.  Trace amounts of potassium, titanium, and sulfur were
also present (Stith et a!.,  1978).   Samples of Mt, St. Helens ash  contained mainly silicon,
aluminum,  iron,  calcium,  sodium,  magnesium,  and  potassium.   Titanium,  phosphorus,  and
manganese, as well  as traces of sulfur, chlorine, strontium, barium, vanadium, zirconium, and
zinc (among others) were also found (Fruchter et al., 1980).
     Based on the  St.  Augustine particulate emissions  (6  x 10  metric tons total, 0.25 x 10
metric tons  less  than 5 urn), less than  5  percent of the  particles  were  smaller than 5 urn in
size (Stith, et  al.,  1978).   According to preliminary airborne studies of Mt. St. Helens ash,
significant  amounts  of  particulate  matter  between 1  and  2  (jm  have been  emitted  to  the
atmosphere (Hobbs  et  al., 1981).  Other preliminary studies of Mt. St. Helens  ash  place the
fraction less than 3.5 urn at around 2 percent (Fruchter et al., 1980).
4.3.5  Wildfires
     The three major types of large scale fires are:  wildfires, prescription fires in natural
areas,   and  agricultural  burning.   The  latter  two types  are exclusively  caused  by human
activities.  Wildfires, defined  by the U.S. Department  of Agriculture Forest Service as "any
fire that burns  uncontrolled in  vegetative or  associated flammable  material,"  are treated
here, as  in  the literature,  as  a  natural  emission source  even though man's  activities cause
about 90  percent of their total  number; only 10 percent are truly  "natural,"  resulting from
lightning (U.S.  Department of Agriculture,  Forest Service, 1979).
     Wildfire particulate emissions calculations typically have been  based on  three numbers:
wildfire  acreage,  fuel  burned per acre, and  emissions per unit mass  of  fuel.   Robinson and
Robbins (1971) estimated  yearly  particulate emissions from  torest  fires  in the U.S. as 0.7 x
10  metric tons,  based  on 4.5 x  10   acres  burned, 18 tons of fuel per acre,  and 17 pounds of
particulate  per  ton  of  fuel.   Yamate  (1973)  arrived  at  0.5  x 10   metric tons  per year,
assuming numbers similar to those used by Robinson and Robbins.
     In recent research,  however,  particulate emissions per  unit  mass of fuel  were estimated
at 17-67  pounds  per ton (GEOMET, 1978)  and  80  pounds per  ton  (Radke  et  al., 1978), based on
airborne  sampling  studies   in   Oregon  and  Washington.   Because  emissions  from  fires  are
dependent on fuel  conditions  and fire behavior  (GEOMET,  1978),  the estimates should probably
be averaged.   If  an emission rate of 40 pounds per ton, a U.S.  wildfire acreage of 3.15 x 10
in 1977 (U.S. Department  of  Agriculture, Forest Service, 1979), and a U.S. average of 17 tons
of fuel per  acre  (Yamate, 1973) are  assumed,  U.S.  particulate emissions from wildfires total
1.0- x 10  metric tons per year.
     Chemical  analysis   of   particulate  matter  from  temperate  forest  burning  indicates
approximately 50 percent  benzene-soluble organic matter, 40  percent elemental  carbon,  and 10
percent mineral  matter (Ryan  and McMahon, 1976).   Another analysis suggests 55 percent tar, 25
percent soot, and  20  percent ash (Vines et al., 1971).   About 80 percent of the mass of smoke
particles from forest fires  is  less than 1 urn in diameter, with the average size being 0.1 pm
(GEOMET, 1978;  Radke et al.,  1978; Vines et al., 1971).

                                              4-10

-------
     Wildfires contribute  varying  amounts of  other  pollutants  to  the atmosphere.   Carbon
monoxide and hydrocarbons are the most significant.   Wildfires are not, however, considered to
be a source of sulfur oxides (Radke et al. , 1978; Yamate, 1973; Vines et al., 1971).
4.4          SOURCES AND EMISSIONS
     A number of definable source categories emitting particulate matter and sulfur oxides can
be attributed solely to man and are the subjects of this section.   Representative estimates of
emissions  from  these  source categories  in the  United States are  summarized in Table  4-4,
Manmade emissions of  particulate matter result primarily from  stationary  point sources (fuel
combustion  and  industrial  processes),  industrial  process  fugitive  particulate  emission
sources, nonindustrial fugitive sources (e.g., roadway dust from paved and unpaved roads,  wind
erosion of cropland),  and  transportation  sources (e.g., automobiles).   The  data in Table 4-4
show  that nonindustrial fugitive  emissions are  significant  on a  mass basis.   However,  the
relative impact of these emissions is somewhat lessened by their coarse size and the fact that
fugitive dust sources (e.g. unpaved roads) are more prevalent in rural areas.

                 TABLE 4-4.  SUMMARY OF ESTIMATED ANNUAL MANMADE EMISSIONS (1978)

Source category
Stationary point sources
Industrial process fugitives
Nonindustrial fugitives
Transportation sources
TOTAL
Emissions (10
Particulate matter
10,5
3.3a
110-370
1.3
-125-385
metric tons)
Sulfur oxides
26.2
_
0.8
27.0

          NOTE:  Approximately half of the 3.3 x 10  metric tons of particulate matter
                 from process fugitives are probably included in the 10.5 x 10  metric
                 tons from stationary point sources.  See Section 4.2 for explanation.
         Source: U.S. Environmental Protection Agency, 1980a

     Manmade  emissions  of  sulfur oxides  result  almost exclusively  from  stationary  point
sources.   The  combustion  of  fossil  fuels  by electric  utilities  causes  most  sulfur  oxide
emissions.  Transportation  sources also contribute a small  amount of sulfur oxide emissions.
4.4.1  Historical Emission Trends
     Economic  conditions  and the  degree  to  which air pollution control  devices  are  used are
the two  factors having the most  impact  on  emissions totals, especially from stationary point
sources (fuel combustion and industrial processes).  Economic conditions affect the amounts of
goods produced  and,  therefore, the amounts of emissions generated.  The economics of relative
fuel  prices  also  affect  emissions;   that  is,   higher  prices on  oil  and natural gas  cause
increased  use of coal,  which generally  emits more particulate matter  and sulfur oxides per
                                              4-11

-------
unit energy  than  oil  or natural gas.   Increased use of control devices has resulted from the
enactment  of regulations such  as  New  Source  Performance Standards  and State Implementation
Plans.
     Historical trends  in  emissions of particulate  matter (not including fugitive emissions,
which have not been documented) and sulfur  oxids are shown  in Table 4-5.  Data for the years
1940, 1950,  and 1960  are  from  the U.S.  Environmental Agency  (1978b);  data  for  1970 through
1978 are  from  the U.S. Environmental Protection Agency (1980a).  Emissions estimates from the
latter are considered  more accurate.  It should be  noted that  local emission trends might not
necessarily coincide with national  emission trends.
     Nationwide  emissions   of  particulate  matter   (not  including  fugitive   emissions)  have
generally  decreased since  1950  after  a  slight  increase from  1940  to  1950,   These emissions
have  resulted  primarily  from  stationary  fuel  combustion   (utility  and  industrial)  and
industrial processes.   Particulate  emissions from stationary fuel combustion decreased fairly
consistently from 1940 to 1978.  From 1940 through the early 1970's this decrease was probably
due to "increased  use  of oil  and natural  gas.   Even though the oil  embargo of 1973-74 caused
increased  use  of  coal, conservation  efforts by industry  and  the installation  of control
equipment resulted  in further reductions in particulate emissions through 1978.
     Industrial process emissions  of  particulate  matter increased  from  1940 to  1960,  then
declined steadily through  1978.   Increases  were attributed to  expanding production; decreases
were attributed to  installation of controls.
     Nationwide  emissions   of  sulfur  oxides  have  increased  overall   since  1940.  As  with
particulate  matter,  stationary  fuel   combustion   (primarily  utility  and  industrial)  and
industrial  processes  (primarily  ore  smelting)  have  been   the  main  contributors.    Coal
combustion was the  largest stationary fuel combustion source, although coal use by industrial,
commercial/institutional, and residential users has  declined, corresponding with a decrease in
sulfur oxide emissions from  those  categories.   Increased coal  use  by  electric utilities has
snore than  offset this  decrease.   Sulfur oxide emissions from  electric  utilities  account for
more than  half the total  emissions.   Flue  gas  desulfurization  (FGD) systems  have seen only
limited use  to date and have not  had a  major  impact on emissions.  About 11  percent  of U.S
coal-fired electrical  generating capacity  is  presently  fitted with FGD  (U.S.  Environmental
Protection Agency, 1980c).
     Increased  industrial   production   caused  most  of  the  sulfur  oxide  emission increases
through 1970.   Since  that time,  however,   significant emission  reductions  from  nonferrous
smelters and sulfuric acid plants have occurred.   For smelters, byproduct recovery of sulfuric
acid has  significantly reduced  sulfur oxide emissions.  Sulfur  oxide  emissions  from copper,
lead and zinc  smelters have decreased from  4  x  10  metric tons per year in 1970 to about 2 x
10  metric tons per year in 1978.
     Future  emission   trends   are   subject  to  a  number of  assumptions  concerning  economic
climate, fuel  use,  environmental  policy,  and control  technology.   These  considerations are
beyond the scope of this document.   (See U.S. Department of Energy, 1978; 1979.)

                                              4-12

-------
                        TABLE 4-5 (a).gNATIONAL ESTIMATES OF PARTICIPATE EMISSIONS'1
                                   (10  metric tons per year)

SOURCE CATEGORY
Stationary fuel
combustion
Industrial processes
Solid waste disposal
Transportation
Miscellaneous
TOTAL

1940
8.7
9,9
0.5
0.5
5.2
24.8

1950
8.1
12.6
0.7
1.1
3.7
26.2

1960
6.7
14.1
0.9
0.6
3.3
25.6

1970
7.2
12.8
1.1
1.1
1.0
23.2

1975
5.1
7.4
0.5
1.0
0.6
14.6

1978
3.8
6.2
0.5
1.3
0.7
12.5
                     Table 4-5 (b).  NATIONAL ESTIMATES OF SULFUR OXIDE EMISSIONS
         	(10  metric tons per year)	

         SOURCE CATEGORY	1940     1950     1960     1970     1975     1978
         Stationary fuel              15.1     16.6     15.7     22.7     20.9     22.1
           combustion
Industrial processes
Solid waste disposal
Transportation
Miscellaneous
TOTAL
3.4
0.0
0.6
0.4
19.5
4.1
0.1
0.8
0.4
22.0
4.8
0.0
0.5
0.4
21.4
6.2
0.1
0.7
0.1
29.8
4.5
0.0
0.8
0.0
26.2
4.1
0.0
0.8
0.0
27.0

         aDoes not include industrial process fugitive particulate emissions, and non-
          industrial fugitives from paved and unpaved roads, wind erosion, construction
          activities, agricultural tilling, and mining activities.
          Includes forest fires, agricultural burning, coal refuse burning, and structural
          fires.
         Sources;  U.S. Environmental Protection Agency (1978b)
                   U.S. Environmental Protection Agency (1980a)

4.4.2  Stationary Point Source Emissions
     In this analysis of sources and characteristics of particulate and sulfur oxide emissions
from  stationary  point sources,  the  two major  source  categories  are  fuel  combustion  and
industrial  processes.   A third but  minor category is solid waste disposal.   Table 4-6 lists,
calculated  estimates  of  1978  emissions  from  these  source  categories.    Based on  these
estimates,  fuel   combustion  contributed  36  percent of  the particles  and 84 percent  of the
sulfur  oxides  emitted by  stationary point sources in 1978.   Industrial  processes  emitted 59
percent of  the  particulate matter and  16 percent  of the sulfur oxides.  Solid waste disposal
contributed 5 percent of the particulate  matter.
                                              4-13

-------
              TABLE 4-6.   1978 ESTIMATES OF PARTICULATE AND SULFUR
                 OXIDE EMISSIONS FROM STATIONARY POINT SOURCES
Source category
Fuel combustion
Utility
Coal
Oil
Gas
Industrial
Coal
Oil
Gas
Other fuels
Commercial /Institutional
Coal
Oil
Gas
Residential
Coal
Oil
Gas
Industrial processes
Metals
Iron and steel
Primary smelting
Iron foundries
Other
Mineral products
Cement
Asphalt
Lime
Crushed rock
Other
Petroleum
Refining
Natural gas production
Chemicals
Sulfuric acid
Other
Other
Grain processing
Pulp and paper
Other
Solid waste disposal
TOTAL
Emissions
Parti cul ate
Matter


2,350
140
10

700
90
40
280

20
60
10

20
20
30


830
480
140
120

780
150
150
1,340
910

70
0

0
190

730
240
60
500
10,460
(10^ metric tons)
Sulfur oxides


15,900
1,720
0

1,890
1,150
0
150

40
900
0

60
260
0


110
1,960
0
0

670
0
0
0
30

900
140

220
0

0
80
0
0
26,180
Primarily wood/bark waste.
 Source:   U.S.  Environmental  Protection Agency (1980a).
                                     4-14

-------
     An  unknown  percentage  of particulate  matter  and  sulfur oxides  is emitted  as  primary
sulfates.  Primary sulfates  consist  of gaseous sulfur trioxide (SO,),  sulfuric acid (H2S04),
and particulate sulfates.  These  primary sulfates are of  increasing concern because of their
potential impacts, especially  on  health, but estimates of primary sulfate emission quantities
from major sources have  not  been generated to  date.   Secondary sulfates can be formed in the
atmosphere,  following oxidation of  S0»,  hours or even days  after its release.   The principal
reactions are  described  in  Chapter  2;   field  measurements  attempting  to  trace  the  actual
production  and  fate  of  secondary  sulfates  under  the  complex  influences  and  diverse
combinations of meteorological  variables  are discussed in Chapter 6.
     Varying  amounts  of  particulate  matter  and  sulfur  oxides  are   emitted  in  different
geographic regions of the  United  States.  Table 4-7  presents  State and regional estimates of
1979 population, PM  and  SO  emissions Abased on the 1977 NEDS inventory), emission densities,
                           )\,                                     ,
and percentage contributions  to total U.S. point source emissions.  Based on this information,
Regions  III  through  VI  accounted for over  70 percent  of the particulate  matter and sulfur
oxides emitted by  stationary sources in  the U.S.   In Region III, utility and industrial fuel
combustion contributed most of the particulate matter and sulfur oxides.  The mineral products
industry also contributed heavily to particulate emissions.  In Regions IV and V, utility fuel
combustion and  the mineral  products  industry contributed most of  the  particulate emissions,
while  utility  fuel combustion contributed  most of  the  sulfur oxide emissions.   The  mineral
products industry  and total  fuel  combustion caused most of the particulate emission in Region
VI.  The primary metals  and  petrochemical industries, along with fuel combustion, contributed
most of the sulfur oxides emissions in Region VI.
     In  other regions grain  processing (Region VII)  and mineral  products (Region IX)  emitted
large  amounts  of particulate  matter.   Fuel combustion  (Regions  II and  VII) and  the  primary
metals industry (Region IX) contributed significant amounts of sulfur oxides.
                                                              i
     Several  factors affect  the  quantity  and  characteristics   (size  and composition)  of
particulate matter emissions  from  stationary  sources.  Examples  of such  factors are source
type,  operating  conditions  and   practices,  fuel   characteristics  (if  the  source  is   a fuel
combustion source),  and  type of  emission control  equipment, if any.  The chemical composition
of  emitted  particles can  determine possible  reactions  that occur during  transport  and the
final  effects  upon receptors  (see Chapters  5  and 6).  Particle  size affects  suspension time
and  transport  distance  and  is also an  important factor  in determining any  possible health
effects  (see Chapters 11-14).
     Table  4-8  presents  a  summary  of particle  size  and  chemical   composition  data  for
uncontrolled particulate emissions from stationary sources.  These data demonstrate the strong
influence of  control  devices  on  the particle size  distribution  of  emissions.    Table 4-9
illustrates that,  for coal-fired  boilers, most control devices are more efficient at removing
larger particles.  Therefore,  even  though the total  mass  of smaller particles decreases, the
percentage  increases.   [Refer to U.S.  Environmental Protection Agency (1980b)  for  further
                                              4-15

-------
TABLE 4-7.  STATE-BY-STATE LISTING OF TOTAL PARTICULATE ANOULFUR OXIDE EHISSIOHS
                      FROM STATIONARY POINT SOURCES (1977),a
                         POPULATION, AND DENSITY FACTORS
Region and state
Region I
Connecticut
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont
TOTAL
Percent of U.S.
Region II
New Jersey
New York
Puerto Rico
Virgin Island
TOTAL
Percent of U.S.
Region III
Delaware
Population
(1000 's)

3,099
1,091
5,774
871
935
487
12,257 *
5.5

7,327
17,748
2,712
62
27,849
12.6

582
District of Columbia 674
Mary! and
Pennsylvania
Virginia
West Virginia
TOTAL
Percent of U.S.
4,144
11,862
5,032
1,821
24,143
10.9
Total garea*"

5,009
33,215
8,257
9,304
1,214
9,609
66,608
1.8

7,836
49,576
3,435
133
60,980
1.7

2,057
67
10,577
45,333
40,817
24,181
123,032
3.4
Population
density,
(people/mi )

619
33
704
94
770
51
184
-

935
358
790
466
457
-

283
10,060
392
259
126
77
196
.
Parti cul ate matter
Total emissions
(10 aetric tons)

17.6
37.7
38.2
7.7
1.7
2.2
105.2
1.51

46.8
171.6
55.2
11.3
284.9
4.08

31.0
1.7
40.3
652.9
101.9
174,4
1,002.2
14.36
Emission
density,
(tons/mr)

3.5
1.1
4.6
0.8
1.4
0.2
-
-

6.0
3.5
16.1
84.6
-
-

15.1
25.1
3.8
14.4
2.5
7.2
-
-
State
emissions
(X of U.S.)

0.25
0.54
0.55
0.11
0.02
0.03
.
-

0.67
2.46
0.79
0.16
-
-

0.44
0.02
0.58
9.36
1.46
2.50
-
-
Sulfur oxides
Total emissions
(10 metric tons)

53.4
115.2
206.7
100.8
10.2
0.9
487.3
1.85

213.7
749.7
280.1
3.6
1,247.2
4.73

115.5
15.0
289.2
2,149.5
372.0
1,084.7
4,025.8
15.28
Emissions
density „
(tons/Mi

10.7
3.5
25.0
10.8
8.4
0.1
-
~

27.3
15.1
81.5
27.4
-
-

56.2
223.5
27.3
47.4
9.1
44.9
-
-
State
emissions
) (% of U.S.)

0.20
0.44
0.78
0.38
0.04
<0.01
-
-

0.81
2.85
1.06
0.01
-
-

0.44
0.06
1.10
8.16
1.41
4.12
-
-

-------
TABLE 4-7. (continued)
Participate matter
Region and state
Region IV
Alabama
Florida
Georgia
Kentucky
Mississippi
North Carolina
South Carolina
Tennessee
TOTAL
Percent of U.S.
Region V
Illinois
« Indiana
i- Michigan
*"* Minnesota
Ohio
Wisconsin
TOTAL
Percent of U.S.
Region VI
Arkansas
Louisana
New Mexico
Oklahoma
Texas
TOTAL
Percent of U.S.
. _ Population
Population Total area density
(1000's) (mi ) (people/nrr)

3,742
8,594
5,084
3,498
2,404
5,577
2,918
4,357
36,174
16.4

11,243
5,374
9,189
4,008
10,749
4,679
45,242
20.5

2,186
3,966
1,212
2,880
13,014
23.2S8
10.5

51,609
58,560
58,876
40,395
47,716
52,586
31,055
42,244
383,041
10.6

56,400
36,291
58,216
84,068
41,222
56,154
332,351
9.2

53,104
48,523
121,666
69,919
267,338
560,550
15.5

73
147
86
87
50
106
94
103
94
-

199
148
158
48
261
83
136
-

41
82
10
41
49
41
-
Total emissions
(10 metric tons)

248.
182.
88.
369.
135.
159.
79.
132.
1,395.
20.

448.
313.
209.
116.
529.
231.
1,848.
26.

105.
295.
77.
71.
367.
917.
13.

5
5
6
1
8
6
1
3
5
0

5
4
3
5
3
8
8
49

8
7
6
0
1
2
14
Emission
density
(tons/mi )

4.82
3.12
1.50
9.14
2.9
3.02
2.6
3.13
-
-

8.0
8.6
3.6
1.4
12.8
4.1
-
-

2.0
6.1
0.6
1.0
1.4
-
-
State
emissions
(X of U S. )

3.56
2.62
1.27
5.29
1.95
2.29
1.13
1.90
.
-

6.43
4.49
3.00
1.67
7.58
3.32
-
-

1.52
4.24
1.11
1.02
5.26
-
,
Sulfur oxides
Emissions State
Total emissions density, emissions
(10Jinetric tons) (tons/mi ) (% of U.S.)

898.
861.
597.
1,442.
211.
510.
260.
1,135.
5,917.
22.

1,462.
1,572.
1,064.
231.
2,840.
568.
7,740.
29.

103.
290.
511.
91.
1,141.
2,138.
8.

0
6
5
4
1
6
7
3
1
46

1
9
0
6
6
8
1
38

1
4
8
1
8
3
12

17.4
14.7
10.1
35.7
4.4
9.7
8.4
26.9
-
-

25.9
43.3
18.3
2.8
68.9
10.1
.
-

1.9
6.0
4.2
1.3
4.3
-
-

3.41
3.27
2.27
5.48
0.80
1.94
0.99
4.31
-
-

5.55
5.97
4 04
0.88
10.78
2.16
-
-

0.39
1.10
1.94
0.35
4.33
-
-

-------
TABLE 4-7.  (continued)
Particulate matter
Region and state
Region VII
Iowa
Kansas
Missouri
Nebraska
TOTAL
Percent of U.S.
Region VIII
Colorado
Montana
North Dakota
South Dakota
Utah
Wyoming
5 TOTAL
Percent of U.S.
Region IX
Arizona
California
Hawaii
Nevada
Guam
TOTAL
Percent of U.S.
Population
(1000 's)

2,896
2,348
4,860
1,565
11,669
5.3

2,670
785
652
690
1,307
424
6,528
3.0

2,354
22,294
897
660
85
26,317
11.9
Population
Total area density Total emissions
(mi4) (people/mr) (Mr metric tons)

56,290
82,264
69,686
77,227
285,467
7.4

104,247
147,138
70,665
77,047
84,916
97,914
581,927
16.1

113,909
158,693
6,450
110,540
212
389,880
10.8

51
29
70
20
41
-

26
5
9
9
15
4
11
-

21
140
139
6
401
68
-

192.0
157.8
112.4
110.1
572.3
8.20

34.2
22.8
19.2
37.3
72.3
144.4
330.2
4.73

40.1
185.7
15.8
85.3
4.5
331.5
4.75
Emission State
density emissions
(tons/mr) (.% of U.S.)

3.4
1.9
1.6
1.4
-
-

0.3
0.2
0.3
0.5
0.9
1.5
-
-

0.4
1.2
2.5
0.8
21.3
-
-

2.75
2.26
1.61
1.58
-
-

0.49
0.33
0.28
0.53
1.04
2 07
-
-

0.58
2.66
0.23
1.22
0.06
-
-
Sulfur oxides
Emissions State
Total emissions density, emissions
(104metric tons) (tons/nr) (X of U.S.)

288.9
157.4
1,296.1
41.5
1,783.9
6,77

91.4
177.7
97.3
34.6
173.6
146.4
721.0
2.74

1,097.9
498.0
55.6
295.2
53.6
2,000.4
7.59

5.1
1.9
18.6
0.5
-
-

0.9
1.2
1.4
0.4
2.0
1.5
-
-

9.6
3.1
8.6
2.7
252.9
-
-

1.10
0.60
4.92
0.16
-
-

0.35
0.67
0.37
0.13
0.66
0.56
-
-

4.17
1.89
0.21
1,12
0.20
-
-

-------
                                                              TABLE 4-7,  (continued)
Region and state
Region X
Alaska
Idaho
Oregon
Washington
TOTAL
Percent of U.S.
U.S. TOTALS
. Population
Population Total area density
(1000' s) (rnr) (people/mi^)

403
878
2,444
3,774
7,499
3.4
220,951

589,757
83,557
96,981
68,192
838,487
23.0
3,620,000

1
11
25
55
9
-
61
Particulate matter
Emission State
Total emissions density emissions
(104 metric tons) (tons/mi ) (% of U.S.)

12.7 <0.1 0.18
15.1 0.2 0.22
67.1 0.7 0.9S
96.2 1.4 1.38
191.0
2.74
6,978.7
Sulfur oxides
Emissions State
Total emissions density, emissions
(10 metric tons) (tons/mi ) (% of U.S.)

5.6
34.2
23.3
217.5
280.7
1.07
26,341.8

<0.1 0.02
04 0.13
0 2 0.09
3.2 0.83
-
-


   "Source;   U.S.  Department of Commerce (1979).
   cSource:   The World Almanac and Book of Facts 1980 (1980).
Ja.

*"*

-------
                        TABLE 4-8.   EXAMPLES OF UNCONTROLLED PARTICULATE EMISSION CHARACTERISTICS
I
f\3
o

Particle size data3
(weight % less than stated size)
Source category 15 urn 2.5 urn 1.0 urn
Fuel combustion
Utility
Coal 15-90 5-70 1-15
Oil 95 70-95 5-20
Industrial
Oil -- — 65-95
Gas — — 100
Commercial/Institional/
Residential
Oil
Chemical composition data
Major elements
and compounds
A1,Ca,Fe,Si,
sul fates .organics
Al,Ca,Fe,Mg,Na,
sul fates, organics
Al,Fe,Mg,Si,
sul fates, organics
Cl.Na, sul fates,
organics
Al,Ca,Mg,Zn,
sul fates
Trace elements
(less than 1% by weight)
As.BjBa.Be.Cd.CljCo^r,
Cu,FfHg,K,Mg,Mn,Na,Nf,
P,Pb,S,Se,Ti,V,Zn,Zr
As,Ba,Br,Co,Cr,Cu,K,
Mn,MosNi,Pb,Se,SrJi,V
As,Ba,Ca,Cd,Co,Cr,Cu,Hg,
K.Mo.Ni.Ph.Se.Sr.Ti.V.Zn
As,Ba,Cd,Cr,Cu,Hg,K,
Ni.Pb.Sb.C
          Gas
100
Cl,Na,sulfates
organics

-------
                                           TABLE 4-8.   (continued)
                            Particle size data
Source category
(weight % less  than stated  size)
15 |jm       2.5 (Jin    1.0 pm
                                               Chemical  composition data
                                                           Major elements
                                                           and compounds
                           Trace elements
                       (less than 1% by weight)
Industrial  processes
  Metals
     Iron and steel
   Primary aluminum      90

   Primary copper         —


   Primary lead          —

   Primary zinc          —

   Iron foundries      70-95

Mineral products
   Cement               80



   Asphalt            10-15


   Lime                  —


   Gypsum                --
            35-99     30-95



               75     35-45

            20-95        70


               80

            90-98

            65-90
               30
                                                 65


                                               5-30
                                      1-2      < k


                                    25-50         5


                                     --          20
A1,C,Ca,Cr,Fe,K,Mg,    Ag.As.Br.Cd.Cs.Cu.F.I,
Mn,Pb,Si,Zn,           Mo.Ni.Rb.Se.Sn.Sr.V.Zr
sul fates,organ!cs

Al,C,Ca,F,Fe,Na

Cu.Pb.S.Zn             Ag,Al,As,Cd,Hg,Sb,Se,
                       Si.Te

Pb,Zn                  As.Cd.SeJe

Cd,Fe,Pb,S,Zn          Cu.Hg.Mn.Sn




Al ,0,03,01 ,K,Mg,       Ag.Ba.Cd.Cr.Cu.F.Fe^n,
Na.Si.carbonates,      MosNi,Pb,Rb,Se,Ti,Zn
sulfates

Al.C.Ca.Fe.K.Mg,       Ag,As,Ba,CrJi
Si,sulfates

Ca,Fe,Mg,Se,Si,
carbonates

Al.C.CA.Mg.Na,         As,Ba,Br,Cd,Cl,Cr,Cu,
sulfates               Fe^.Mn.Mo.Ni.Pb.Se,
                       Sr.Y.Zn

-------
                                            TABLE 4-8.   (continued)

Source category
Crushed rock
Petroleum
Particle size data3
(weight % less than stated size)
15 urn 2.5 urn 1.0 urn
1-2
50-90
Cheiical coiposition data
Major elements
and compounds
Ca.Si.P
Asphalt, coke dust,
Trace elements
(less than 1% by weight)
'§a,Cu,Fe5K,Mn,Sr
Chemicals
  sulfuric acid

Others
40-95
10-55
                                                              sulfuric acid mist,
                                                              flyash,  soot
                                                             Sulfuric acid mist



-p.
1
ro
PO
Grain processing
Pulp and paper


Solid waste disposal
Incinerators

15
90-95



45

1 0
70-80



35

Organics
Ca,Mg,Na, carbonates,
sul fates


..

b
Since a number of references were cited, some characterizing different processes, discrepancies
may exist in the ranges shown.

Elements and compounds listed were included in at least one of the references cited.
Sources:   Surprenant et al.  (1979);  Taback et al.  (1979);  U.S.  Environmental  Protection Agency
          (1980c); U.S. Environmental  Protection Agency (1980d);  Plemons and  Parnell  (1981).

-------
                                  TABLE 4-9.  SIZE-SPECIFIC PARTICULATE  EMISSIONS  FROM

                                                    COAL-FIRED BOILERS

Control device
ESP
Wet scrubber
Fabric filter3
Inlet size distribution
(uncontrolled)
(Mass percent less than)
15 p* 2.5.|jm
15-50 5-20
30-95 10-70
55-65 20-45
Outlet size distribution Removal efficiency (%)
(controlled)
(Mass percent less than)
15 Mm 2.5 |jm 15 pi
70-95 15-70 65-99+
80-95 50-90 75-95

-------
discussion  on  the effects  of control devices on  emissions  characteristics.]   Therefore,  the
application  of the  particle  size  percentages  representing uncontrolled emissions  shown  in
Table 4-8  to the  emission  quantities from  controlled sources  listed in Tables  4-6 and  4-7
would probably result in an underestimation of the finer particle fractions.
     As  a  further example  of the difference between  controlled  and uncontrolled conditions,
control  devices  have helped  reduce  the  mass flow  of particulate  emissions  in California's
South Coast Air Basin by 95 percent or more from what prevailed under uncontrolled conditions.
However, over 90 percent of the remaining emissions (both point sources and miscellaneous area
sources) have particle sizes less than 10 pm (Taback et a!., 1979).
     A  final  point  with  respect  to  Table  4-8  is  that  the  particle size and  chemical
composition  data  represent  an overall  source  category.   Therefore,  in the   iron  and steel
industry,  for  example,  not  all  of  the  many different processes  emitting particulate matter
would  necessarily  have  emissions  exhibiting  the  exact-  characteristics  shown.   Further
information can be obtained from the cited documents4<
     The same  factors mentioned  earlier may affect the quantity and characteristics of sulfur
oxide emissions.  By  volume,  over 90 percent of total national  sulfur oxide emissions are in
the form of sulfur  dioxide,  SO,,.   Primary sulfates  account for most of the other 10 percent.
Little  is  known about  primary-'sul fates, but combustion of  coal  and oil is  thought to be a
major source.  Primary  sulfates  are of increasing  concern  because of their potential impacts
on visibility, acidic deposition, and health.
4.4.2.1   Fuel Combustion—Stationary  fuel   combustion  includes  all   boilers,  heaters,  and
furnaces   found   in  utilities,   industry,   and   commercial/institutional   and  residential
establishments.   In  the utility  and industrial  sectors,  coal,  and to a lesser  degree,  oil
combustion  contribute  most of the  particulate  and  sulfur  oxides  emissions (see  Table 4-6),
Oil combustion  causes most  of these emissions  from commercial/institutional  and residential
establishments.
     Coal  is  a   slow-burning  fuel  with  a  relatively  high ash  content.    Coal  combustion
particles  consist primarily- of   carbon,  silica,  alumina,   and  iron  oxide  (See  Table 4-8).
Particulate  sulfates  and trace elements  are  also included.   A large  percentage of the trace
elements in  raw coal  remains in the  solid waste or bottom ash, as  shown in Table 4-10 (based
on 1974 emissions data).   The roughly 940,000  metric tons of trace  elements  emitted to  the
atmosphere represent about 15 percent of total particulate emissions.
     Uncontrolled, the quantity  and  particle size distribution of coal fly  ash depend on  the
amount  and  type of  coal  burned,  the  unit type,-and the ash content of the coal. - Cyclone  and
pulverized-coal furnaces,  typically  used  in utility  boilers, discharge  finer particles than
stoker-fired boilers, used  mainly by industry.   The  combustion  of low-ash coal produces less
particulate  matter  than  the combustion  of  high-ash  coal.   High-sodium  lignite  causes less
combustion  particulate  formation  than does low-sodium lignite  (U.S. Environmental Protection
Agency, 1977).
                                              4-24

-------
           TABLE 4-10.   TRACE ELEMENT AIR EMISSIONS VS. SOLID WASTE;   PERCENT FROM CONVENTIONAL
                        STATIONARY FUEL COMBUSTION SOURCES, AND TOTAL  (METRIC TONS PER YEAR)

Air emissions (fly
Element
As
Ba
Be
B
Br
Cd
C1
Cr
Co
Cu
F
Fe
Pb
Mn
Hg
Ni
Se
Ti
U
V
Zn
Zr

Util
90
88
89
90
84
61
83
84
63
72
83
77
92
89
81
60
85
89
86
63
89
78

Indust
8
9
9
9
13
21
13
11
23
16
13
20
7
10
14
21
13
9
10
20
10
20

Com/Inst
2
3
2
1
2
18
2
5
14
12
2
3
1
1
3
19
2
2
4
17
1
2

ash)
Res Total
<1 2,990
<1 2,770
<1 240
<1 4,990
1 6,080
<1 300
2 644,100
<1 1,630
<1 460
<1 2,540
2 33,570
<1 154,200
<1 1,180
<1 4,630
2 50
<1 7,350
1 790
<1 56,250
<1 1,540
<1 9,980
<1 2,090
<1 2,090
939,820
Solid waste (bottom
Util
89
83
83
85
0
83
0
75
69
78
0
87
81
98
78
81
76
83
84
84
83
86

Indust
9
13
12
13
0
14
0
12
9
12
0
9
15
1
20
12
22
13
13
12
13
11

Com/Inst
1
2
2
1
0
1
0
5
8
4
0
2
2
1
2
3
1
2
1
1
2
1

ash)
Res
1
2
3
2
0
2
0
8
14
6
0
2
3
<1
<1
4
1
2
2
2
3
2


Total
12,250
15,970
740
16,240
0
110
0
5,040
1,920
4,280
0
1,369,900
2,530
12,520
10
4,700
370
178,700
4,510
8,450
6,890
17,240
1,662,370

Source:   Surprenant et al.  (1976).

-------
     In the  combustion of most coals  (most  commonly bituminous),  more than 90 percent of the
coal sulfur is converted to gaseous S0?; about 1 to 2 percent of the emitted sulfur oxides are
in the form of primary sulfates (Homolya and Cheney, 1978a; Homolya and Cheney, 1979).
     Lignite is used where it is plentiful at relatively low cost.   The alkali content (mostly
sodium) of lignite ash has a major effect on the amount of coal sulfur retained in bottom ash.
A  high-sodium  lignite may  retain over  50  percent of the available sulfur,  and a low-sodium
lignite may retain less than 10 percent (U.S. Environmental Protection Agency, 1977).
     Several factors  can affect  the formation  of primary sulfates from coal-fired boilers.
The higher  excess  oxygen levels commonly used in  industrial boilers increase the oxidation of
SOg to  SCL  and HoSO*  (Homolya and  Cheney,  1978a;  Bennett and Knapp, 1978).  Most gaseous SO-
is hydrated  to gaseous or aerosol H-SQ*  before  exiting the boiler stack (Homolya and Cheney,
1979).  Dirty  equipment  also may increase primary sulfate  emissions from coal-fired boilers,
since boiler deposits  can act as catalysts  in  the oxidation of SO,, to sulfates.  Conversely,
the relatively  low flame temperatures used in most coal-fired boilers  lessen the formation of
SO, from S0«.
  w        £.
     After  coal,  oil  combustion  in the  utility and industrial sectors  results  in  the next
largest amount  of  emissions.   In direct contrast  to coal, however, oil is a fast burning, low
ash fuel.   The low ash content results  in  formation of less particulate matter, but the size
of particles formed by oil combustion  is  generally smaller than that of particles  formed by
coal combustion (see Table 4-8).  Also, although coal combustion contributes most of the trace
elements  associated with  particulate  emissions,  oil combustion  is the  source  of 50  to 80
percent of  cadmium, cobalt,  copper, nickel, and vanadium emissions (Surprenant et al., 1976).
     Oil-fired boilers generally  convert over 90  percent of  available fuel sulfur to gaseous
SO, emissions.   However,  high flame temperatures  used in the combustion of oil exacerbate the
formation of primary  sulfates.   Tests have  shown that about  7  percent  by  weight  of sulfur
oxide  emissions from  oil combustion  is  emitted  as  primary  sulfates  (Homolya  and Cheney,
1978b).   Increasing  excess oxygen  and increasing  the oil vanadium  content will  increase the
formation of primary  sulfates in the gas (Homolya and Cheney, 1978b; Bennett and Knapp, 1978;
Oietz  et  al.,  1978).   As the emissions disperse  and cool to  ambient  temperature, vanadium's
catalytic action becomes insignificant (see Chapter 2).
     Low  sulfur oil and  natural  gas  are  the fuels  typically used for  space  heating in the
commercial/institutional and residential sectors.  Total emissions are minor compared with the
utility  and  industrial  sectors.   However,  most  commercial/institutional  and  residential
sources are in areas of high population density and release emissions at OP near ground level,
thereby providing for high population exposure (Surprenant et al.,  1979).   Also, emissions are
concentrated primarily during the winter heating season.
     Currently, homeowners, particularly those in  the northern forested areas of the U.S., are
turning to  wood as a  heating fuel.   In 1976, a total of 16 x 10  metric tons was combusted in
wood stoves  and furnaces,  auxiliary heating devices, and fireplaces (deAngelis et al., 1980).
                                              4-26

-------
The emissions of  participate  matter from wood combustion  primarily consist of organic matter
that is rich in phenols, derivatives of benzaldehyde, and furfural, and a variety of compounds
in the  general  category  of polycyclic organic  matter (POM). Measurements  of the filterable
(with collection  taking  place at 116°C) particles emitted  during  residential  wood combustion
typically  show  levels  of approximately 3  g per  kg of fuel.   Condensable (at  0°C)  organic
emissions are on the order of 8 g/kg.   Of critical importance is the fact that measurements of
ROM's show  values of 0.02 to 0.3 g/kg  and  0,03  g/kg for  stoves and fireplaces,  respectively
(deAngelis  et al.,  1980).   On the national  scale, exclusive of aircraft emissions, 35 percent
of all  POM's are  emitted from  residential  wood combustion (Peters, 1981).   In  1980,  the CO
emissions  from  home  wood  combustion  accounted  for  about 3 percent  of the  national  total;
emissions  of sulfur oxides are  at least two orders  of  magnitude  less  than  the particulate
emissions.   Although emissions from residential  wood burning are a small fraction of national
totals,  they are  injected practically at  ground  level  and can  be  expected to  become an
increasingly significant component in urban air pollution.
4.4.2.2   Industrial Processes—Major  industrial  process  sources  of  particulate and  sulfur
oxide emissions  include  the  metals,  mineral products,  petroleum, and  chemicals  industries.
Others are  grain processing and pulp and paper production (See Table 4-6).
     The  most significant  emission sources  in  the  metals industry are iron and  steel  and
primary  smelting  operations.   The iron  and  steel  industry involves  coke,  iron,  and steel
production.   Coking is  the  process  of heating  coal  in   a  low-oxygen  atmosphere  to  remove
volatile  components, which  are  recovered.   Coke is  used in the  production  of  iron.   Both
particles  and sulfur  oxides   result  from  the  charging of  coal  to the  hot  ovens,  door  and
topside leaks, underfiring, pushing (removal of hot coke),  and quenching.  Some fine particles
consist, at least partly, of condensed organic components.
     Particulate  emission sources of  iron  production  include  the combustion  gases,  tapping
operations,  and  blast furnace  slips  (operations that  require bypassing  the control device).
The emitted particles are probably all fine particles that either escape the control  device or
result  from tapping (see Table 4-8).    Blast  furnace flue  dust is composed primarily of iron,
siljcon dioxide,  and aluminum oxide, among others.
     Steel  is  produced  in  several  different ways.   The basic  oxygen  furnace produces steel
from a  furnace  charge  composed of about 70 percent  molten pig  iron and  30 percent  scrap.  A
stream  of  commercially  pure  oxygen  is used to  oxidize  impurities,  principally carbon  and
silicon.   The tremendous agitation produced by the oxygen lancing produces high dust loadings
consisting  mostly of iron and small amounts of fluorides.   Most of the particles are less than
5 urn in size (U.S.  Environmental Protection Agency, 1977).
     From  1960  to 1975,  steel production in  open  hearth  furnaces  declined from 90 percent of
the U.S.  total  to 20 percent  (Desy,  1978).   Open hearth furnaces are being replaced by basic
oxygen furnaces that produce 272 or more metric tons of steel per hour compared with the 27 to
54 metric tons per  hour  typically produced  in open hearth furnaces.  The  composition of
                                              4-27

-------
particulate  emissions  is  similar  to  those  from  the  basic  oxygen furnace.   Most of  the
emissions before  control  are less than 5 urn, and probably 90 percent are fine particles after
control.
     Two  types  of electric furnaces, the  arc  furnace and the  induction  furnace,  are used to
produce steel.  The arc furnace is used to produce high-alloy steel, as well as a considerable
amount of mild  steel.   The emissions, most  of  which are fine particles, consist primarily of
oxides of iron,  manganese, aluminum, calcium, magnesium, and  silicon.   Particulate fluorides
are  also  emitted.  The  induction furnace produces  primarily  specialty  and high-alloy steels
and has no major emission problems.
     The  primary  metals  industry includes the smelting  of  copper,  lead, and zinc, along with
aluminum  production.  Sulfur in unprocessed copper,  lead, and zinc ores is converted to S0~ in
the  smelting processes  (U.S.   Environmental  Protection  Agency,  1977).   A  relatively  small
portion of the sulfur is emitted as particulate sulfate and sulfuric acid.  The bulk of SO- is
formed  in the  roasting,  smelting,  sintering, and  converting processes  (U.S.  Environmental
Protection Agency,  1974).   Particulate  matter emitted from the same  processes is mostly fine
particles, less than 2.5 urn in diameter.
     Aluminum  production  involves  mainly  bauxite  grinding,  calcining,   and  reduction.
Particulate emissions are primarily alumina with about  25  percent  particulate fluoride (U.S.
Environmental Protection  Agency, 1977).   Before  control,  35  to  44  percent  of  the particles
were below 1 urn in diameter.
     Emissions  from the  mineral products  industry  result primarily from the  production of
Portland  cement, asphalt, and crushed rock and to a  lesser extent, lime, glass, gypsum, brick,
fiberglass,   cleaned coal,  phosphate rock,  and  potash.   Emission  points such  as crushing,
screening, conveying,  grinding, drying or calcining,  and loading are common  to most mineral
products  industries.   Fugitive  dust from most of these  processes tends to be larger than 15
urn,  although drying  and calcining  produce relatively  finer  particles.   The composition of
particulate emissions is similar to the mineral being processed.
     The  more than  30  raw materials  used to  make  cement  can be  grouped into  four  basic
categories:    lime   (calcareous),  silica  (siliceous),  alumina  (argillaceous),   and  iron
(ferriferous).  The kiln  and associated clinker cooler are potentially the largest sources of
particulate  and  sulfur oxides  emissions  (U.S. Environmental  Protection  Agency,  1977).   Kiln
emissions also  include  primary sulfates  (Dellinger et  a!.,  1980).   Probable  particle  size
distribution and chemical composition are shown in Table 4-8.
     Asphalt  concrete  is  a mixture  of aggregate,  asphalt  cement,  and  occasionally mineral
filler.   Commonly,  asphalt concrete is  produced  in  conjunction with  crushed and broken stone
production  facilities.    The  rotary  dryer typically  used  to  dry  and  heat the  aggregate is
potentially  the   largest  particulate emission source  (U.S.  Environmental  Protection Agency,
1977).
                                              4-28

-------
     Emissions  from  the  production  of  crushed  rock  result  primarily  from the  processes
mentioned  earlier.    The  chemical  composition of  particulate emissions  is  similar to  the
material processed.   Usually, the particles emitted are relatively coarse.
     The major  sources  of  sulfur oxide emissions in  the petroleum industry are the catalytic
cracking  and  sulfur  recovery processes  and  off-gas flares  (U.S.  Environmental  Protection
Agency,  1977);   Dickerman  et  al.,  1977).   Sulfur  dioxide  is  emitted  during the  catalyst
regeneration step of the catalytic cracking process.
     Major sour gas streams are usually treated in a sulfur plant.   Most sulfur plants utilize
a  modified Claus   process  that  consists  of   multistage  oxidation  of  hydrogen  sulfide  to
elemental sulfur.   The  sulfur recovery efficiency of these sulfur plants ranges from 92 to 97
percent  depending   on  the number of  catalytic  stages.   Sulfur  plant tail  gas  is  usually
incinerated so  that most  of the  remaining  sulfur  species are oxidized to  SOp.   Some plants
have  installed  tail gas cleanup  systems  to  reduce  SOp emissions  further.   These  units  along
with a sulfur plant can achieve up to 99.8 percent sulfur recovery.
     Minor  off-gas  streams and recovered vapors  are  often combusted in flares.   Most of the
sulfur species present in these vapors are oxidized to S0?.
     A  variety  of   processes  are  used by the chemical production  industry.   Chemical process
industries  that contribute significant  amounts of sulfur oxide   emissions  are sulfuric acid
plants, elemental sulfur plants,  and explosives manufacturing.
     Sulfuric acid  is  manufactured primarily by  the  contact  process.   The three types of raw
materials  charged   to  sulfuric acid  plants are  elemental  sulfur,  spent  acid and  hydrogen
sulfide, and  sulfide  ores  and smelter gases.   The  amount  of S0? emissions in acid plant exit
gases  is  an  inverse  function  of  the  sulfur conversion  efficiency  of  the process  (U.S.
Environmental  Protection  Agency,   1977).  Sulfuric  acid  mist is generated by  the  process S0?
absorbers.  The  quantity  and size distribution of the acid  mist  are dependent on the type of
sulfur  feedstock used,  the  strength of the acid produced, and the conditions in the absorber.
     The  manufacture  of  TNT  and  nitrocellulose explosives  produces  emissions  of  SOp  and
sulfuric  acid  mist.    Sulfuric  acid  is  a  major raw  material  in  the  production   of  these
explosives.   Sulfuric acid  concentrators,   exhaust  from  the preparation of  sodium  sulfite/
sodium  hydrogen sulfite (Sellite),  and incinerators are  the major sulfur  oxides sources in
these  processes.  Sulfur  oxide emissions may vary considerably depending on the efficiency of
the process and the operating conditions (U.S.   Environmental   Protection Agency, 1977).
     Particulate emissions  from  grain processing typically  result from  handling,  cleaning,
drying,  and milling (U.S.  Environmental Protection Agency, 1977).   Grain processing particles
are normally coarse and composed of the parent  organic material.
     Chemical  wood  pulping by the  kraft or sulfite processes  involves cooking wood chips under
pressure  to dissolve  the lignin that binds  the cellulose  fibers,  in addition  to  washing,
milling,  bleaching,  and  drying  (U.S.  Environmental  Protection   Agency,  1977).   Particulate
emissions  occur primarily from  the recovery furnace (used to  recover  cooking chemicals)\and
                                              4-29

-------
the  lime  kiln  (lime   is  used  in  cooking).   Sulfur  dioxide emissions  result  mainly  from
oxidation of reduced sulfur compounds in the recovery furnace.,
4.4.3  Industrial Process Fugitive Particulate Emissions
     Fugitive dust emissions result from wind erosion of storage piles and unpaved plant roads
and  from  vehicular  traffic  over  plant  roads.   Fugitive  process emissions   result  from
industry-related  operations  such  as  materials  handling,   loading,  unloading,  and  transfer
operations.   Point sources  that are  incompletely controlled, such  as  furnace  charging and
tapping,  and equipment  that  is  maintained poorly,  such  as  leaking furnaces and  coke  oven
doors, are also fugitive process emission sources.
     Process fugitive emissions  are not emitted from a definable point such as a stack.  They
are difficult to collect, measure, and control.  A given industry generally has a large number
of  fugitive  particulate  emission   sources.   For  example,  20  separate  sources  have  been
identified for  foundries  (Jutze  et al., 1977).   In  terms  of total emissions, however, one or
two of these sources may predominate.
     Even  though  fugitive  particulate emission  totals may  appear  small  when  compared  with
totals from  large  point sources, they may  take  on importance because of the concentration of
control efforts  on point source emissions.  In  the  integrated iron and steel industry, where
fugitive  particulate  emissions   are  characterized  relatively well,  fugitive emissions  are
estimated  to account  for  about  10  percent of  all  uncontrolled emissions.   However, since
fugitive particulate emissions are poorly controlled, they account for more than 60 percent of
total  controlled  emissions (Spawn,  1979).   Also,  in situations where point  sources  are  we'll
controlled or  use high stacks,  fugitive  particulate emissions exert a major  effect  on local
air quality.   Extremely high  suspended particulate matter  levels  have  been  measured in areas
where process fugitive emissions are predominant (Lynn et al., 1976; Lebowitz, 1975).
     Table 4-11 presents  estimates of uncontrolled  industrial process fugitive particulate
emissions.   Particle size  and  composition characteristics are also presented.   Unfortunately,
many  of  the  emission  factors  used  to   estimate  process  fugitive  emissions  are  based  on
engineering  judgment  or extrapolation  from  similar  processes.    Often,  few  test  data  are
available  to  support   these  estimates  since  process  fugitive  emissions  are  difficult  to
measure.    Therefore,  the  accuracy  of  these  estimates is  questionable.  Also,  some  of the
emissions  presented  in  Table 4-11 may have  already  been accounted for in Table  4-6  (Section
4.4.2).  This overlap results from the use of different references (see Section 4.3).
     As  is  evident  from Table 4-11,  three broad categories account for nearly all  of the
potential  process  fugitive emissions  in  the United  States.  They are mineral  products,  food
and agriculture, and primary metals.  In the mineral  products industries,  fugitive particulate
emissions  tend  to reflect  the composition  of  the parent  materials.   The limited  amount of
particle size data indicates that most particles are relatively coarse.
     Grain elevator operations account for the fugitive particulate emissions in the food and
agriculture  industry.   These emissions  consist almost entirely of grain dust from loading and
                                              4-30

-------
                  TABLE 4-11.  UNCONTROLLED INDUSTRIAL PROCESS
                        FUGITIVE PARTICULATE EMISSIONS3
Source category
                    Annual uncontrolled
                    fugitive particulate
                       ,, emissions
                         Size
(10  metric tons)   characteristics
                    Major components
Mineral products
  Crushed rock
  Extraction,
  Surface coal mining
  Portland cement
  Asphalt concrete



  Lime manufacturing

  Concrete batching

Food and agriculture
  Grain elevators

Primary metals
  Coke/i ron/steel
  Foundries
       730
       700




       100



        50

        30


     1,250


       250
       125
10-50% < 10 pm
 1-2% < 1 pm
10-151 < 10 pm




50-70% < 4 pm



45-70% < 5 pm

10-20% < 5 pm


40% < 10 pm
Coke mfg:  27-80%
< 10 pm, 15-26%
< 2 pm; iron mfg:
1-10% < 5 pm;
Steel mfg:   50%
< 5 pm
50% < 15 pm
Same as parent
material (important
for toxic minerals
such as asbestos,
beryllium, silica)

Limestone, clay,
shale, gypsum,
iron-bearing and
siliceous materials

Sand, crushed stone,
1imestone, hydrated
lime

Limestone, lime

Cement dust
Grain dust
Polycyclic organic
matter, coal dust,
coke dust, iron
oxide dust, kish
(graphite material),
metal fume (pri-
marily iron oxide),
plus trace amounts of
As, Be, Pb, Cr, Cd,
Se, Co, Ni, and
fluorides

Metal oxide fume
(primarily oxides of
silicon and iron),
fine carbonaceous
fume, plus trace
amounts of Pb, Cr,
Ag, Co, and Ni
                                     4-31

-------
                             TABLE 4-11.  (continued)
Source category
Annual uncontrolled
Fugitive particulate
   2 emissions           Size
(10  metric tons)   characteristics
                    Major components
  Aluminum
        60
  Copper
        40
  Lead
        15
Secondary metals
        10
Wood products

     TOTAL
50-90% < 10 urn
10-50% < 5
10-90% < 5 |jm
80-100% < 5
                    40-90% < 10
Particulate fluo-
rides, alumina
(AlpO,), carbon
dust, condensed
hydrocarbons, tars

Cu, Fe, S, SiO,
from ore concen-
trate; metal fume
consisting of
oxides of As, Pb,
In, Cu, Cd, plus
trace amounts of
Se and Ag

Metal fume consist-
ing of oxides of
Pb, Cd, In, Sb, plus
trace amounts of As,
Se, and Ag

Oxides of Al,Cu,Pb,
Sn,Zn; oxides of
alkali metals;
A1CU, NH.Cl.NaCl,
ZnClX; fluorides,
and carbonaceous
materials, plus
trace amounts of
Cr, Cd, and Ni

Sawdust
Note:     Emissions are based on data for the mid-1970's and may differ from
          current levels.                                        «   -

aSources:  Taback et al. (1979)
           Zoller et al. (1978)
           Jutze et al. (1977)
           Norman et al. (1977)
                                     4-32

-------
unloading, drying and cleaning operations, conveyor belts, and transfer points.  Only about 29
percent of these particles are less than 10 pm in size.
     Primary  metals production  encompasses  six separate  industries.   Fugitive  particulate
emissions in this category result from the handling and transporting of raw materials and from
the smelting and refining of these raw materials into their finished metal  products.  Although
emissions of  the first  type  are  not  well  characterized, emissions of the  second type often
consist  of   fine metal  fumes.   This  finding  is  particularly  significant  because  of  the
quantities of  toxic trace  metals  that can  be concentrated and  volatilized  in metal melting
operations.   Some of these  trace toxic components of  particulate emissions  are identified in
Table 4-11.
     The remaining two categories, secondary metals and wood products, account for less than 1
percent of the  national  total  of industrial process fugitive particulate emissions.  Fugitive
metal  fume  particles from  secondary metal melting  operations also  include  toxic components
and are listed in Table 4-11.
4.4.4  Nom'ndustrial Fugitive Particulate Emissions
     Nom'ndustrial  fugitive  particulate emissions,  or fugitive  dusts  are caused  by traffic
entrainment of dust from public paved and unpaved roads, agricultural operations, construction
activities,  surface mining  operations, and fires.  With  the  exception  of  fires, all of these
sources may  be  classified as  open-dust sources; that is, they involve dust entrainment by the
interaction  of  machinery  with  aggregate materials  and  by  the  forces  of  wind  on exposed
materials.
     A  number  of factors  can  affect  emissions  from open  sources  but they  can generally be
classified  under  three  headings:   material,  equipment,  and  climate.    Material  factors
encompass such   influences  as  silt  and  moisture content.  For  example,  increasing  the silt
content and  decreasing  the  moisture content of unpaved road material would probably  result in
more  dust being generated.   Equipment factors  generally refer  to  vehicle  weight and speed.
For example, increasing the speed or weight of a vehicle travelling over an unpaved road would
tend  to  increase  emissions.   Climatic  factors  are  windspeed and  precipitation.    Increased
windspeed and decreased precipitation would both tend to  increase emissions from any  open-dust
source.
     Estimated U.S.  annual  particulate emissions from nonindustrial fugitive dust sources are
difficult to estimate  accurately.    As  shown  in Table  4-12,  fugitive dust  emissions from
unpaved  roads  tend to  be quite  significant.   The two available estimates,  however, vary by
almost  an order of magnitude.   Fugitive  dust from wind  erosion  of  cropland and construction
activities as documented by Cooper  et al. (1979) also appears significant.  However, no other
estimates are  available  for  comparison purposes.   Estimated total   fugitive  emissions range
                             f~                                     /-
from  approximately  112  x  10   metric  tons per year  to  369 x 10  metric  tons per year.  The
lower  figure assumes the  U.S.  Environmental  Protection  Agency  (1980b) estimate of fugitive
emissions from  paved and unpaved  roads;  the  higher figure assumes  the estimate of Cooper et
                                              4-33

-------
           TABLE 4-12.  ESTIMATED ANNUAL PARTICULATE EMISSIONS FROM
                        NONINDUSTRIAL FUGITIVE SOURCES
Source category
                                        Estimated Emissions
                                           c                  _,
                                        (10  metric tons/year)
Cooper et al. (1979)
U.S.  Environmental
 Protection Agency
     (1980b)
Unpaved roads
Paved roads
Wind erosion of cropland
Agricultural tilling
Construction activities
Minerals extraction
Mineral tailing
Prescribed fires
290
7.2
40
2.9
25
3
0.7
0.4
35
4.7
-
-
-
-
-
0.2
 Particles less than 30 pm in diameter.

^Includes prescribed forest burns and agricultural burning.
        TABLE 4-13.  ESTIMATED PARTICLE SIZE DISTRIBUTIONS FOR SEVERAL
          NONINDUSTRIAL FUGITIVE SOURCE CATEGORIES IN CALIFORNIA'S
                             SOUTH COAST AIR BASIN
Source category
Unpaved road dust
Agricultural tillage dust
Road building and construction dust
Agricultural burning
Weight
>10 pin
54
40
36
<1
jpercent
3-10 (jm
16
21
24
2
in size
1-3 urn
12
17
16
8
range
<1 pm
18
22
24
90
Source:  Taback et al. (1979).
                                     4-34

-------
al.  (1979).   Because of  the  disparity between comparable estimates,  the  estimated emissions
presented in Table 4-12 should be considered with a degree of caution.
     Information  on  particle size  distribution  is  also  limited.   Some  limited data  are
presented in  Table 4-13,  but they are  representative only  of California's South  Coast  Air
Basin and  should not  be  extrapolated to  the Nat-ion  as  a whole.   Dust from  unpaved  roads,
agricultural  tilling,  construction,   and  road  building  is  composed  primarily of  silicon,
phosphorus,  aluminum,  iron,  calcium,  and  potassium.  Trace  elements  include  barium,  cobalt,
copper,  lead,  manganese, nickel,  titanium,  vanadium, and zinc (Taback et al., 1979).
     Finally,  it  is  estimated that fugitive dust emissions  exceed  particulate emissions from
stationary point sources in 90 percent of the Air Quality Control Regions that are not meeting
the ambient standards  for total  suspended particulates (Carpenter and Weant, 1978).   However,
the  impact  of fugitive dust emissions on  populated areas may be somewhat  lessened  because a
major portion of these emissions  consists of large particles that settle to the ground a short
distance from the source  and because many  fugitive dust sources,  like unpaved roads, exist
mainly in rural areas (U.S. Environmental Protection Agency, 1980b).
4.4.5  Transportation Source Emissions
     Transportation  source emissions  may  be divided into  two categories:   engine-related
emissions from vehicle exhaust, and other highway vehicle-relate'd particles from tire wear  and
clutch  and  brake  lining  wear  (Bradow  et al.,  1979;  U.S.  Environmental   Protection  Agency,
1978b; Dannis, 1974; and Jacko and DuCharme,  1973).   Total transportation source emissions  for
1978,  including  emissions from  highway  vehicles,  aircraft,  railroads,   and  vessels, were
estimated at  1.3 x  10   metric tons  (particulate matter)  and  0.8 x  10   metric tons  (sulfur
oxides)   (U.S.  Environmental  Protection  Agency,  1980a).   About "75  percent  of  the particulate
emissions and 50 percent  of  the sulfur oxide emissions  in 1978 were  from highway vehicles.
     Engine-related  particulate  emissions  from transportation  sources  are  composed primarily
of  lead  halides, sulfates, and  carbonaceous matter (including absorbed  organics).   Highway
vehicle-related  particles  are emitted at  the rate  of about 0.01  to 0.30  grams per mile by
gasoline engines  and 0.5  to 3.0  grams per mile for diesel engines (Bradow et al., 1979).   The
major components are  lead (except  for vehicles  using  diesel fuel  or  unleaded  gasoline),
carbon,   organics,  and  sulfates.   Vehicles  burning  leaded  gasoline  also  emit  inorganic
compounds of  lead (mostly PbBrCl), bromine, and chlorine (Springer,  1978).   Particulate  matter
from  catalyst-equipped   vehicles   using  unleaded   gasoline  is  dominated  by  sulfate  and
carbonaceous material.
     Diesel  exhaust  particles can  include  traces  of iron,  copper,  calcium,  lead,  and zinc,
along'with  carbon,"organics,  and sulfates  (Lee  and Duffield, 1979).   The particle emission
rate and composition for diesel engines are sensitive to many factors, including vehicle size,
operating conditions (speed,  load),   and  fuel  characteristics.  Normally,  carbon-containing
species  dominate, including  a material  similar  to lubricating oil  (Black and  High,  1978).
                                              4-35

-------
     Engine-related particles  are mostly under 1 pm in diameter.  For vehicles burning leaded
gasoline, the  available data indicate a mass  median  diameter of about 0.25 pn (Moran et a!.,
1971).   Due to  the  predominance of  sulfates,  catalyst-equipped  vehicles burning  unleaded
gasoline  emit  smaller  particles  having a mass  mean diameter  of  about 0.05  jjm (Groblicki,
1976).  The  size  distribution of diesel particulate matter suggests a mass median diameter of
about 0.2 urn (Dolan and Kittelson, 1979).
     Very few  data  exist on nonengine particulate  emissions  from highway vehicles.   About 40
percent of particles from tire wear are less than 10 |jm (about 20 percent are less than 1 |jm);
they  are composed  primarily of  carbon  (Taback  et a!.,  1979).   Particles from  brake lining
attrition are  all  less  than 1 urn and  are  composed mainly of asbestos (80 percent) and carbon
(Taback et a!., 1979).
4.5  SUMMARY
     Participate matter and sulfur  oxides are emitted  into the atmosphere from a  number of
sources,  both  natural  and  manmade.   Natural  source  emissions  include  terrestrial  dust,  sea
spray, biogenic emanations, volcanic emissions, and emissions from wildfires.   The predominant
manmade sources are stationary point sources, industrial  and nonindustrial  fugitive sources,
and transportation  sources.   Annual  U.S. emissions from natural sources are estimated at 84 x
10  metric tons of  particulate matter and 5 x 10  metric tons of sulfur (the equivalent of 10
x 10  metric tons  of sulfur dioxide).   Manmade  sources  emit roughly 125 to 385  x 10  metric
tons of particulate matter per year and 27 x 10  metric tons of sulfur oxides  per year in the
U.S.  Because  of the assumptions and approximations inherent  in emissions  calculations,  the
numbers  quoted above  should  not be  considered  more  than  estimates.   Section  4.2 further
discusses the problem involved with emissions estimates.
     The characteristics of particulate  matter emissions vary  according  to  source type and a
number of other factors.  Particulate emissions from natural sources tend to be rather coarse.
(For the purposes of this chapter, coarse refers to particles with a diameter greater than 2.5
urn.)   Particulate matter generated by nonindustrial fugitive sources (e.g., unpaved roads and
wind erosion of cropland) is quite significant on a mass basis.  However, only  about 50 and 20
percent  is  less  than 10 and 1 urn,  respectively.   Most of the  particulate matter emitted by
stationary sources  and  transportation  sources, on the other hand, is relatively fine, or less
than 2.5 urn in diameter.  Adding control devices further concentrates particulate emissions in
the finer ranges because most control devices are more efficient at removing larger particles.
Therefore, the  estimated 10.5  x  10   metric  tons  of particulate matter generated in 1978 by
stationary point sources probably consists largely of finer particles, since that estimate was
arrived  at  assuming  the  application of  control  devices.   In  addition,  the  finer  particles
emitted by stationary point sources tend to include a greater variety of toxic  substances than
do emissions from natural or manmade fugitive sources.
     Virtually all of the manmade sulfur oxide emissions result from stationary point sources.
Over 90  percent  of  these  manmade sulfur  oxide emissions are in the form  of  sulfur dioxide.
                                              4-36

-------
The balance consists  of  sulfates  in various forms.   Most natural  sulfur is emitted as reduced
sulfur compounds,  but these compounds  are probably  oxidized  in the  atmosphere to  sulfur
dioxide and sulfates.
                                              4-37

-------
4.6  REFERENCES

Adams, D.  F.   Personal communication  [Telephone  conversation  with  E.  Robinson],  October 1980.

Adams, D.  F.,  S.  0. Farwell,  M.  R.  Pack, and W.  L.  Bamesberger.   Preliminary measurements  of
     biogenic  sulfur-containing  gas  emissions  from  soils.   J.  Air Pollut.  Control  Assoc.
     29:380-383, 1979a.

Adams, D.  F.,  S.  0. Farwell,  M.  R. Pack, and  E.  Robinson.   Estimates  of  natural  sulfur source
     strengths.   Presented at  the  Gatlinburg Conference  on  Sulfur,  Gatlinburg, TN,  October
     1979b.

Adams, D.  F.,  S.  0. Farwell,  E.  Robinson, M.  R.  Pack,  and W.  L.  Bamesberger.   Biogenic Sulfur
     Strengths.  Environmental  Science and Technology 15:1493-1498,  1981.

Adams,  F., R,  Dams,  L.   Guzman,  and J.  W.  Winchester,   Background  aerosol composition  on
     Chacaltaya Mountain,  Bolivia.  Atmos Environ,  11:629-634,  1977.

Barringer,  A.  R.   AIRTRACE—an airborne  geochemical  exploration technique.   |n:   Proceedings
     of the First Annual William  T. Pecora Memorial Symposium,  American Mining Congress,  Sioux
     Falls,  South  Dakota,  October  1975.    Geological  Survey  Professional Paper  1015,  U.S.
     Department of  the Interior,  Washington, D.C.,  September 1977.   pp. 231-251.

Baylor, E. R., M. B. Baylor, D. C. Blanchard,  L.  D. Syzdek, and C.  Appel.   Virus  transfer from
     surf to wind.  Science  (Washington, D.C.) 198:575-580, 1977.

Beauford,  W.,  J.  Barber,  and A.  R.  Barringer.   Heavy metal  release  from  plants into the
     atmosphere.  Nature (London) 256:35-37, 1975.

Beauford,  W.,  J.  Barber,  and  A.  R.  Barringer.   Release  of  particles containing metals  from
     vegetation into the atmosphere.  Science  (Washington, D.C.)  195:571-573,  1977.

Bennett,  L.,  and K. T.  Knapp.  Sulfur and  trace  metal particulate emissions  from  combustion
     sources.  In:  Workshop Proceedings on  Primary Sulfate Emissions  from  Combustion  Sources,
     Vol.  2:   Characterization,'  U.S. Environmental  Protection Agency,  Southern Pines,  North
     Carolina,  April  24-26,  1978.   EPA-600/9-78-020b, U.S.  Environmental  Protection  Agency,
     Research Triangle Park, NC,  August 1978.  pp.  165-183.

Berg,  W.  W.  Jr.,   and J.  W.  Winchester.   Aerosol chemistry  of the  marine  atmosphere.  In:
     Chemical Oceanography,  Vol. 7.  2nd Edition.  J.  P. Riley and R. Chester, eds.,  Academic
     Press Inc., London, England, 1978.  pp. 173-231.

Black, F.,  and  L.  High.    Diesel  hydrocarbon emissions, particulate and  gas phase.  Presented
     at  Symposium  on Diesel   Particulate  Emissions  Measurements   and  Characterization, Ann
     Arbor, HI, May 1978.

Blanchard,  D.  C.,  and  B.  C.  Parker.   The  freshwater to air transfer of microorganisms and
     organic  matter.   In:   The  Aquatic  Microbial Community.   J.   Cairns, Jr.,  ed.,   Garland
     Publishing, New York, NY,  1977.  pp. 625-658.

Blanchard, D. C.,  and A.  H. Woodcock.   Bubble formation and  modification  in  the sea  and its
     meteorological significance.  Tellus 9:145-158,  1957.

Bradow, R.  L,,  F.  M. Black, J. N. Braddock, C. T.  Hare, M.  Ingalls, and  J. M.  Kawecki.   Study
     of Particulate Emissions  from Motor Vehicles.  A Report to Congress.   Draft  Report.   U.S.
     Environmental  Protection  Agency, Research Triangle Park,  NC, 1979.
                                              4-38

-------
Carpenter,   B.   H. ,   and   G.   E.   Weant,  III.    Parttculate  Control  for  Fugitive   DUst.
     EPA-600/7-78-071, U.S. Environmental Protection Agency,  Research Triangle Park,  NC,  April
     1978.

Cooper,  D.   W. ,   J.  S.  Sullivan,  M.  Quinn,  R,   C.  Antonelli,  and  H.  Schneider.   Setting
     Priorities   for  Control   of   Fugitive  Particulate   Emissions   from  Open   Sources.
     EPA-600/7-79-186, U.S.  Environmental  Protection  Agency, Washington,  D.C.,  August  1979.

Cooper,  J.   A.   Environmental  Impact  of   Residential  Wood  Combustion  Emissions  and  Its
     Implications.  J. Air Pollut. Control Assoc. 30:855-861,  1980.

Curtin, G.  C.,  H.  D. King, and  E.  L. Moiser.  Movement  of  elements into the atmosphere from
     coniferous  trees  in   subalpine  forests of  Colorado  and  Idaho.   J.  Geochem. Explor.
     3:245-263, 1974.

deAngelis, D. G., D.  S. Ruffin, J. A.  Peters  and R. B. Reznik.   Source  assessment;  residential
     combustion  of  wood.   EPA-600/2-80-Q42b.   U.S.   Environmental  Protection  Agency,   IERL,
     Research Triangle Park, N.C.  1980.

Dannis, M. L.  Rubber dust from the  normal wear of  tires.  Rubber  Chem.  Technol.  47:1011-1037,
     1974.                                                                        ~~

Delany,  A.   C. ,  A.   C.  Delany, D.  W. Parkin,  J.  J.   Griffin, E.  D.  Goldberg,   and  B.  ,E.  F.
     Reimann.   Airborne  dust  collected  at  Barbados.   Geochim.  Cosmochim.  Acta  31:885-909,
     1967.

Dellinger, B. , G.  Grotecloss,  C.  R.  Fortune, J. L. Cheney, and  J. B. Homolya.  Sulfur dioxide
     oxidation and plume formation at cement  kilns.  Environ.  Sci. and  Techno!.,  14:1244-1249,
     1980.                                                                        ~~

Desy,  D.  H.   Iron and Steel.   Bureau of  Mines  Mineral Commodity  Profiles No.  MCP-15,  U.S.
     Department of the Interior, Washington,  D.C.,  July 1978.
                                                                 X
Dickerman, J.  D.,  T. D.  Raye, J.  D. Colley,  and R. H.  Parsons.  Industrial process  profiles
     for  environmental  use.   In:    Petroleum  Refining  Industry.    EPA-600/2-77-023c,   U.S.
     Environmental Protection Agency,  Research Triangle Park,  NC,  January 1977.

Dietz, R. N.,  R.  F.  Wieser, and  L.  Newman.  Operating parameters affecting  sulfate  emissions
     from an oil-fired power  unit.    In:   Workshop Proceedings on  Primary  Sulfate  Emissions
     from  Combustion  Sources,  Vol.  2:   Characterization,   U.  S.   Environmental   Protection
     Agency,  Southern Pines,  North  Carolina,  April  24-26,  1978.   EPA-600/9-78-02Qb.    U.S.
     Environmental Protection  Agency, Research Triangle  Park,  NC, August 1978.   pp.  239-270.

Dolan, D.  F. , and D.  B.  Kittelson.  Roadway  Measurements  of  Diesel  Exhaust Aerosols.   SAE
     technical paper no.  790492, Detroit, MI, March 1979.

Duce,  R.  A. , and E.  J. Hoffman.  Chemical fractionation  at the  air/sea  interface.  Annu.  Rev.
     Earth Planet, Sci. 4:187-228, 1976.

Eriksson,  E.   The  yearly circulation  of   chloride  and sulfur  in  nature; meteorological,
     geochemical  and pedological  implications.   Part I,  Tellus,  11:375-403,  1959;  Part  II,
     Tell us, 12:63-109, 1960.

Fegley,  R.  W., H. T.  Ellis,  and  J.   L. Heffter.  Volcanic  contributions to  the  stratospheric
     sulfate layer.  J. Appl. Heteorol. 19:683-690, 1980.
                                               4-39

-------
Fish,  B.  R.   Electrical generation  of natural  aerosols from vegetation.   Science (Washington,
     D.C.) 175:1239-1240,  1972.

Fruchter,  J.  S., D.  E.  Robertson,  J.  C. Evans,  K.  B.  Olsen, E.  A.  Lepel,  J. C.  Laul,  K.  H.
     Abel, R.  W.  Sanders, P.  0.  Jackson, N. S.  Wogman,  R.  W.  Perkins, H.  H.  Van Tuyl,  R,  H.
     Beauchamp,  J.  W.  Shade,  J. L.  Daniel, R.  L.  Erikson,  G.  A.  Sehmel,  R. N.  Lee,  A.  V.
     Robinson, 0.  R.  Moss, J.  K. Briant, and W.  C.  Cannon.   Mount St.  Helens ash from the 18
     May  1980  eruption:    chemical,   physical,   mineralogical,   and  biological   properties.
     Science (Washington,  D.C.)  209:1116-1125,  1980.

Galloway, J. N., and  D.  M. Whelpdale.   An atmospheric  sulfur budget for eastern North America.
     Atmos. Environ.  14:409-417,  1980.

GEOMET,  Incorporated.   Impact of Forestry  Burning Upon Air  Quality.   A State-of-the-Know!edge
     Characterization  in  Washington   and  Oregon.    EPA-910/9-78-052,  U.S.   Environmental
     Protection Agency,  Seattle,  WA, October 1978.

Gillette, D. A.  On the  production of  soil  wind erosion aerosols  having the  potential for long
     range transport.  J.  Rech,  Atmos.  8:735-744, 1974.

Goldberg, E. D.   Rock volatility and  aerosol composition.   Nature (London)  260:128-129,  1976.

Granat,  L.,  H.  Rodhe, and R.  0.  Hallberg.   The  global  sulfur cycle.   Ecol.  Bull.  (Stockholm)
     22:89-134, 1976.

Groblicki, P.  J.   General Motors Sulfate Dispersion  Experiment:   Aerosol  Sizing Measurements.
     GMR-2127, General Motors  Corporation,  Warren,  MI,  April  1976.

Henry, R.  C.,  and G. M.  Hidy.   Potential for  atmospheric sulfur from microbiological sulfate
     reduction.  Atmos.  Environ.  14:1095-1103,  1980.

Hitchcock, D. R.   Atmospheric  sulfates  from biological  sources.   J.  Air Pollut.  Control Assoc.
     26(3): 210-215,  1976.

Hitchcock, D.  R.,  L.  L. Spiller, and W.  E.  Wilson,   Sulfuric acid aerosols  and HC1 release in
     coastal atmospheres:   Evidence  of rapid formation of sulfuric acid particulates.   Atmos.
     Environ. 14:165-182,  1980.

Hobbs, P.  V.,  L.  F.  Radke, M. W. Eltgroth,  and D.  A. Hegg.   Airborne studies of the emissions
     from the volcanic eruptions  of Mt. St.  Helens.  Science  211:816-818,  1981.
                                                                                     \
Homolya,  J.  B.,  and J.  L. Cheney.   An assessment of sulfuric acid and sulfate emissions from
     the  combustion  of fossil  fuels.   I_n:   Workshop  Proceedings- on  Primary  Sulfate Emissions
     from  Combustion  Sources,   Vol.  2:   Characterization.   U.S.  Environmental   Protection
     Agency,  Southern  Pines,  North  Carolina,   April  24-26, 1978.   EPA-600/9-78-020b,  U.  S.
     Environmental  Protection  Agency,  Research  Triangle Park,  NC,  August 1978a.   pp.3-11.

Homolya,  J.  B.,  and J.  L. Cheney.   An Assessment of Sulfuric Acid and Sulfate Emissions from
     the  Combustion  of Fossil  Fuels.   In:   Workshop  Proceedings on  Primary  Sulfate Emissions
     from   Combustion   Sources,   Vol.   2,   Characterization.     EPA-600/9-78-0206.    U.S.
     Environmental Protection  Agency, Research  Triangle Park, NC,  1978b.

Homolya,  J.  B.,  and  J.  L.  Cheney.   A  study of  primary sulfate emissions  from a coal-fired
     boiler with FGD.  J. Air. Pollut.  Control  Assoc. 29:1000-1004,  1979.

Jacko,  M, G.,  and  R.   T.  DuCharme.    Brake  Emissions—Emission Measurements  from  Brake  and
     Clutch  Linings  from  Selected  Mobile  Sources.   APTD-1557,  U.S.  Environmental  Protection
     Agency, Ann Arbor,  MI. March 1973.
                                              4-40

-------
Junge,  C.  E.   Chemical  analysis  of  aerosol  particles  and of gas  traces on  the Island  of
     Hawaii.   Tell us 9:528-537, 1957.

Jutze,  G.  A., J.  M.  Zoller,  T.  A.  Jansen,  R.  S.  Amide,  C.  E.  Zimmer,  and  R.  W.  Gerstle.
     Technical  Guidance  for  Control  of Industrial  Process  Fugitive Particulate  Emissions.
     EPA-450/3-77-01Q.   U.S.   Environmental  Protection  Agency,  Research  Triangle Park,  NC,
     March 1977.

Lawson, D. R.,  and J. W. Winchester.  A standard crystal  aerosol  as  a reference for elemental
     enrichment factors.  Atmos. Environ. 13:925-930, 1979a.

Lawson,  D.   R. ,  and  J.  W.  Winchester.   Sulfur,  potassium,  and phosphorus  associations  in
     aerosols  from South  American  tropical  rain  forests.   J.   Geophys.  Res.   84:3723-3727,
     1979b.

Lee,  R.  E. ,  Jr., and  F,  V.   Duffield.   Sources of  environmentally  important  metals in  the
     atmosphere.   In:   Ultratrace Metal  Analysis in  Biological   Sciences  and   Environment,  A
     Symposium,  174th  Meeting America!  Chemical  Society,  Chicago,  Illinois,  August 29-30,
     1977.   T.  H. Risby,  ed. , Advances  in  Chemistry Series 172, American Chemical  Society,
     Washington,  D.C.,  1979.  pp. 146-171.

Lebowitz,  M.  F.  Short-term testing for fugitive dust effect.   Presented at the  68th Annual
     Meeting,  Air Pollution  Control  Association,  Boston,  MA,  June  15-20,1975.   Paper  No.
     75-25.4.

Lovelock, J.  E.   CSp and the natural sulfur cycle.  Nature (London)  248:625-626, 1974.

Lovelock,  J.  E.,  R.  J. Maggs,  and  R.  A.  Rasmussen.  Atmospheric  dimethyl sulphide and  the
     natural  sulphur cycle.  Nature  (London) 237:452-453,  1972.

Lynn,  D.  A.,  G.  L.  Deane, R.  C.  Galkiewicz,  and R. M.  Bradway.   National Assessment of  the
     Urban Particulate  Problem.  Vol. 1:  Summary  of National Assessment.   EPA-450/3-76-024,
     JJ. S. Environmental Protection  Agency, Research  Triangle Park,  NC, June 1976.

Maclntyre, F.   Chemical  fractionation and sea-surface microlayer processes.   In:  The  Seas:
     Ideas and Observations on Progress in the  Study of  the Seas,  Vol.  5:   Marine Chemistry.
     John Wiley and Sons, Inc., New  York, NY, 1974.   pp.245-299.

Maroulis,  P.  J.,  and  A.  R.   Bandy.   Estimate  of  the contribution  of biologically  produced
     dimethyl  sulfide to the  global  sulfur  cycle.   Science  (Washington,  D.C.)  196:647-648,
     1977.

McClenny, W.  A. ,  R.  W.  Shaw,  R.  E.  Baumgardner, R.  Paur,  A.  Coleman, R.  S.  Braman, and  J.  M.
     Ammons.    Evaluation  of  Techniques for  Measuring  Biogenic Airborne Sulfur  Compounds:
     Cedar Island Field Study  1977.    EPA-600/2-79-004, U.S.  Environmental  Protection  Agency,
     Research  Triangle  Park, NC, January 1979.

McElroy,  M.   B. ,  S.   C. Wofsy,  and  N.  D.  Sze.   Photochemical sources  for atmospheric  H?S.
     Atmos.  Environ.  14:159-163, 1980.

Miller, M. S., S. K.  Friedlander, and  G. M. Hidy.  A  chemical  element balance for the  Pasadena
     aerosol.  J. Colloid Interface  Sci. 39:165-176,  1972.

Moorby, J. ,  and H. M.  Squire.   The  loss of radioactive  isotopes  from the leaves of plants in
     dry conditions.  Radiat.  Bot. 3:263-167, 1963.

Moran, J. B.,  0.  J. Manary, R.  H. Fay, and M, J. Baldwin.   Development of Particulate  Emission
     Control  Techniques for Spark-ignition Engines.  APTD-0949,  U.S.  Environmental Protection
     Agency, Ann  Arbor, MI, July 1971.


                                              4-41

-------
Meyers, J.  L.,  and R. A. Duce.  Gaseous and participate bromine  in the marine  atmosphere.  J.
     Geophys. Res. 77:5330-5338, 1972a.

Hoyers, J.  L.,  and R. A. Duce.  Gaseous  and particulate  iodine  in the marine  atmosphere.  J.
     Geophys. Res. 77:5229-5238, 1972b.

National   Research  Council,   Subcommittee   on  Airborne   Particles.    Airborne  particles.
     University Park Press, Baltimore, HD, 1979,

Nemeryuk,  G.  E.   Migration  of salts  into the  atmosphere during transpiration.  Sov.  Plant,
     Physio!. 17:560-566, 1970.

Norman, B. M., C.  B. Parnell, and R. V. Avant.   Characterization  of particulate emissions  from
    'grain  sorghum storage and handling  installations.   Paper 77-3516.   American Society of
     Agricultural  Engineers, St. Joseph, MI 1977.

Peters,  J., A.   POM  emissions  from  residential wood  burning;  an  environmental assessment.
     Presented at  International Conference on Residential  Solid Fuels, Portland,  OR, June  1-4,
     1981.  Sponsored by Oregon Graduate Center, Beaverton, OR.,  1981.

Plemons,  0.  S.  and  C.  B.  Parnell.   Comparison of  physical and chemical  characteristics of
     grain  dust  from  five  different  grain  species.    Paper  SWR181-4Q4 American Society of
     Agricultural  Engineers, St. Joseph, MI.  1981.

Radke,  L.  F.,  J.  L.  Stith,  D.  A.  Hegg,  and  P. V.  Hobbs.   Airborne  studies of  particles and
     gases from forest fires.  J. Air Pollut. Control Assoc.  28:30-34, 1978.

Rahn,    K.  A.  Sources of Trace  Elements  in Aerosols—An Approach to Clean Air.   Ph.D. Thesis,
     University of Michigan, Ann Arbor, MI, 1974.

Rahn,  K.  A.  The Chemical Composition of the Marine Aerosol.   Technical Report,  University of
     Rhode Island, Kingston, RI, July 1976.

Rasmussen, R. A.   Emission of biogenic hydrogen  sulfide.  Tellus  26:254-260, 1974.

Rasmussen,  R.   A.,  and  F,  W.  Went.    Volatile organic   material  of  plant  origin  in the
     atmosphere.   Proc.  Natl. Acad. Sci. USA 53:215-220, 1965.

Robinson,  E.,   and  R.  C.  Robbins.   Sources,  Abundance,  and  Fate  of Gaseous Atmospheric
     Pollutants.   Stanford Research Institute, Menlo Park,  CA,  1968.

Robinson,  E.,  and  R.   C.  Robbins.   Emissions,  Concentrations,  and  Fate   of Particulate
     Atmospheric Pollutants.  Stanford Research  Institute,  Menlo  Park, CA, 1971.

Ryan,  P. W., and McMahon.  Some Chemical and Physical Characteristics of Emissions from  Forest
     Fires,  Paper 76-2.3,  presented at the 69th Annual  Meeting of the Air Pollution Control
     Association,  Portland, OR, June 22-July 1,  1976.

Schnell,  R.  C.,  and G.  Vali.  Atmosphere ice nuclei from decomposing  vegetation.    Nature
     (London) 236:163-165, 1972.

Schnell,  R.  D.,  and G.  Vali.  World-wide source  of  leaf-derived freezing  nuclei.    Nature
     (London) 246:212-213, 1973.

Spawn,  P.   Letter  Report  submitted  to  U.  S. Environmental  Protection  Agency, Industrial
     Environmental Research Laboratory, under Contract  No.  68-02-2687, Technical  Directive No.
     007. GCA Corporation, Bedford, MA, September 20, 1979.
                                              4-42

-------
Springer, J.  J,   Current  and Future  Trends  of  Exhaust  Emissions from  HD Highway  Vehicles.'
     Presentation  to  Federally  Coordinated  Program  of  Highway  Research  and Development,
     Federal Highway Administration, College Park, MO, October 1978.

Stith, J. L.,  P.  V.  Hobbs, and  L.  F.  Radke.   Airborne particles  and gas measurements  in the
     emissions from six volcanoes.  J. Geophys.  Res. 83:4009-4017,  1978.

Surprenant,   N,,   R.  R.  Hall, S.  Slater,  T.  Susa,  and  M.  Sussman.   Preliminary  Emissions
     Assessment  of  Conventional  Stationary  Combustion   Systems.   Vol.  II:    Final Report.
     EPA-600/2-76-046b.   U.S. Environmental  Protection  Agency,  Research  Triangle Park, NC,
     March,  1976.

Surprenant,   N. ,   R.  R.  Hall, K.  T.   McGregor,  and A.  S.  Werner.   Emissions  Assessment of
     Conventional  Stationary  Combustion  Systems.   Vol.  1:   Gas  and  Oil-Fired  Residential
     Heating  Sources.    EPA-600/7-79-029b.    U.S.   Environmental  Protection  Agency,   Research
     Triangle Park, NC, May 1979.

Taback,  H.   J. ,   A.  R.  Brienza, J.  Macko,  and  N.  Brunetz.   Fine  Particle  Emissions from
     Stationary  and  Miscellaneous  Sources  in   the  South  Coast  Air  Basin.    Prepared for
     California Air Resources Board by KVB, Inc., Tustin,  CA, February  1979.

U.S.. Department of Agriculture,  Forest Service.    1977 Wildfire Statistics.  U.S.  Department of
     Agriculture, Washington, D.C., April 1979.

U.S.  Department   of  Commerce.   Bureau of  the  Census^.    Statistical  Abstract  of  the  United
     States:   1979.  100th Edition.    U.S.  Department of Commerce,  Washington, D.C.,  September
     1979.

U.S.  Department   of  Energy.   National Energy  Plan II--Appendix:   Environmental  Trends and
     Impacts. U.S. Department of  Energy, Washington, D.C., May 1979.

U.S. Department  of  Energy.  National  Environmental Impact Projection No. 1.   U.S. Department
     of Energy, Washington, D.C., April, 1978.

U.S.  Environmental  Protection   Agency.    Office  of  Air Quality Planning  and  Standards.
     Background  Information  for' New Source Performance Standards:   Primary Copper, Zinc, and
     Lead  Smelters.    Vol.  1:   Proposed  Standards.   EPA-450/2-74-OQ2a.   U.S.   Environmental
     Protection Agency, Research  Triangle Park, NC, October 1974.

U.S.  Environmental  Protection   Agency.    Office  of  Air Quality Planning  and  Standards.
     Compilation of Air Pollutant Emission  Factors  3rd Edition,  (including Supplements 1-7).
     Parts A  and B.   AP-42,  U.S. Environmental Protection Agency,  Research Triangle  Park, NC,
     August 1977.

U.S. Environmental Protection Agency.  Carcinogen Assessment of Coke Oven Emissions.   External
     review draft.  EPA-600/6-82-003.  U.S. Environmental  Protection Agency, Research  Triangle
     Park, NC, April 1978a.

U.S.  Environmental  Protection Agency.  National  Air  Pollutant  Emission Estimates, 1940-1976.
     EPA-450/1-78-003, U.S.  Environmental  Protection  Agency, Research  Triangle  Park,  NC, July
     1978b.

U.S.  Environmental  Protection   Agency.    Office  of  Air Quality Planning  and  Standards.
     Protecting Visibility.   EPA-450/5-79-008.  U.S. Environmental  Protection Agency,  Research
     Triangle  Park, NC, October  1979.

U.S. Environmental Protection Agency.  Office of  Air Quality Planning and Standards.   National
     Air  Pollutant  Emissions  Estimates,   1970-1978.   EPA-450/4-80-002.   U.S.   Environmental
     Protection Agency, Research  Triangle Park, NC, January 1980a.


                                              4-43

-------
U.S.  Environmental  Protection  Agency.   Office of  Air Quality  Planning and Standards.  197)
     National  Emissions  Report.   EPA-450/4-80-OQ5.    U.S.  Environmental  Protection  Agency,
     Research Triangle Park, NC, March  1980b.

U.S. Environmental  Protection  Agency.  Office of Air  Quality  Planning and  Standards.   Control
     techniques  for particulate emissions  from stationary sources (preliminary draft).  U.S.
     Environmental Protection Agency, Research Triangle Park,  NC,  July 1980c.

U.S. Environmental  Protection  Agency.  Industrial  Environmental  Research Laboratory.   Summary
     data  received  from  G.  Johnson, Fine  Particle Emissions  Information  Data System.  U.S.
     Environmental Protection Agency, Research Triangle Park,  NC,  September 1980d.

Vandegrift, A. E., L. J.  Shannon, P.  G. Gorman, E.  W.  Lawless, and E. E. Sallee.    Particulate
     Pollutant  System Study.   Volume  I  -  Mass  Emissions.   APTD-0743.   U.S.  Environmental
     Protection Agency, Research Triangle Park, NC, May 1971.

Vines,  R.  G.,  L.  Gibson,  A.  B.  Hatch,  N.  K.  King, and D.  A. MacArthur.  On  the  Nature,
     Properties,  and Behavior  of  Bush-fire  Smoke.   Commonwealth  Scientific  and  Industrial
     Research Organization, Australia,  1971.

Wallace, J. M.,  and P.  V. Hobbs.  Atmospheric Science.   Chapter  4.  Academic Press,  New York,
     NY, p. 150,  1977.

Went,  F.  W.    Organic  matter   in  the  atmosphere, and  its  possible   relation  to  petroleum
     formation.   Proc. Natl. Acad. Sci. USA. 46:212-221,  1960.

Went,  F.  W.  j  D.  E.  Slernmons,  and  H.   N.  Mozingo.   The  organic  nature of   atmospheric
     condensation nuclei.  Proc. Natl.  Acad. Sci. USA,  58:69-74,  1967.

Wexler,  H.  On the effects of  volcanic dust on  insolation and  weather.  Bull.  Am.  Meteorol,
     Soc. 32:10-15, 1951a.

Wexler,  H.   Spread  of  the  Krakatoa  volcanic   dust cloud  as  related  to the  high-level
     circulation.  Bull.  Am. Meteorol.  Soc. 32:48-51,  1951b.

Winchester, J. W., R. J.  Ferek,  D. R. Lawson, J. 0. Pilotte, M. H.  Thiemens, and L.  E.  Wangen.
     Comparison   of  aerosol  sulfur  and   crustal  element  concentrations  in particle  size
     fractions  from continental U.S.  locations.    Water  Air  Soil  Pollut., 12:431-440, 1979.

World Almanac and Book of Facts  1980.   Grosset and  Dunlap, New York, NY, 1980.

Yamate,  G.  Development  of Emission  Factors for  Estimating Atmospheric Emissions  from Forest
     Fires.   EPA-450/3-73-009,  U.S.  Environmental  Protection  Agency,  Research Triangle Park,
     NC, October  1973.

Zoller, J., T. Bertke, and J. Janzen.   Assessment of Fugitive  Particulate Emission  Factors  for
     Industrial  Processes.   EPA-450/3-78-107.   U.S. Environmental  Protection Agency,  Research
     Triangle Park, NC, September 1978.

Zoller, W. H., E. S. Gladney, and R.  A. Duce.  Atmospheric concentrations and sources of trace
     metals at the South  Pole.   Science (Washington, D.C.) 183:198-200,  1974.
                                               4-44

-------
                         5.  ENVIRONMENTAL CONCENTRATIONS AND EXPOSURE

5.1  INTRODUCTION
     This  chapter  has two  objectives:   (1) to  delineate the concentrations of  S0«  and par-
ticles suspended in  the  air to which human  populations,  other organisms, and manmade objects
are exposed;  and  (2) to show how various  sources  of air pollutants contribute  to  these ex-
posures.
     The first goal is to describe ambient air concentrations of these two pollutants in a way
which  is  relevant  to the  effects  they might  cause.    Measurements  of  S0?,  TSP,  and some
chemical  components  of  PM  in  the ambient  air have  been made for  a long  time,  mostly with
imperfect  methods  and procedures.   In Chapter  3,  the most  current information  relative  to
sources of error in measurement are covered in detail.  Here, only those issues that influence
interpretation are mentioned and then only briefly.  The reader is advised to consult Chapter
3 for more detail.
     Despite imperfections  in  measurement methods,  State and  local  monitoring  data stored by
EPA are  the  largest  available  source of  information  on long-term trends in pollutant concen-
trations and on the geographical distributions of the pollutant levels.   Therefore, the exist-
ing monitoring data are presented first to provide an overall perspective regarding S0« and PM
concentrations encountered in the ambient air.
     Particulate matter,  as a pollutant  class,  is  exceedingly complex both  in  regard to its
physical  properties   and  its chemical  composition.   In  Chapter  2,  those  characteristics  of
particles  generally  observed in  most atmospheres are discussed  in  detail.   Consequently,  in
this  section  only those features of chemical  composition and physical size  are  treated that
influence  data interpretation.   The  reader is directed to  Chapter 2 for more detail on these
subjects.
     Recently, particle-measurements have been collected and analyzed to estimate the relative
contributions of  important sources.   In  this  case, the  elemental and  chemical  complexity of
the particles proves to  be valuable since  many  source  types have identifiable characteristic
chemical  signatures.   Consequently,  it is often possible to make at least approximate assign-
ments  of the relative amounts  of suspended PM derived  from road  dust,  power plants, auto-
mobiles,   and  other   common sources,  provided  that  an  adequate description  of the  source
signature  is available.   Since this technique is still new, only a few studies are available,
some  of  which are  discussed to  show the  approximate  source  contributions in representative
cases.   For  a  more complete description of particle emission factors and inventories, Chapter
4 should be consulted.
     Ultimately, the  importance of the ambient air measurements of pollutant concentrations is
in  identifying and predicting  undesirable effects.   When the  effect considered is visibility
reduction,  the  important  factors  are  concentrations  of  light  scattering  and  absorbing
particles  over the geographical  scale of several miles.   In materials damage, concentrations

                                            5-1

-------
of  SO,  and of  soiling particles  are  important over  time scales of  months or years.   Wher
considering  effects other  than  visibility, the  important matter is  the dose.   Dose  incor-
porates  concentration,  time, uptake  and loss, and  the relationship  among  these  parameters.
Throughout  their lifetime,  people inhale  a  complex mixture  of  gases and  particles.   Other
living  things,  vegetation  and  animals,  are also  exposed to  the same complex  mixtures,  the
composition of  which varies with  time  at any given location  because  of  changing  ataiospheric
conditions and source contributions.  The biological  effects of air pollution are functions of
dose  delivered  to  the  receptor and the ability  of the receptor to  cope with  the resultant
stress.    In  humans,  the  stress  experienced  by  a  critical  organ  or  receptor  tissue  from
particle inhalation  depends on  particle size, composition, morphology, acidity or alkalinity,
and other physicochemical properties of the aerosol.   The delivered dose is also a function of
the anatomical  features of the  receptor  as well  as  the manner of breathing,  breathing rate,
and integrity of bodily defense systems.
     It  is  almost  impossible to measure directly the air pollution  dose to  a  population or
even to  an  individual,  except in  the laboratory.  As  an alternative to direct measurement of
dose,  exposure  can  and often must be used as an approximation of dose for studies on air pol-
lution  risk and effects.   The exposure-response relationship for air pollution is most impor-
tant for establishing  standards.   Unfortunately,  to extrapolate  from  measurements of ambient
levels at a few locations to individual  or population exposure levels  is a very difficult task
at present.   The  contribution of outdoor air to indoor concentrations is still being investi-
gated.   The  additional exposures  to  gases  and particles  from nonoccupational  indoor sources
are not adequately known.
     Indoor air  quality and activity patterns complicate air pollution exposure estimates and
are discussed later in this chapter.   First, the ambient outdoor concentrations of SO, and PM
are examined.
5.2  AMBIENT MEASUREMENTS OF SULFUR DIOXIDE
     Ambient concentrations of $02 are determined by the following factors:
     1.    Density of emission sources.
     2.    Source  characteristics  such as stack height, exit  velocity,  and source strength.
     3.    Local  meteorological conditions.
     4.    Local  topography and surrounding buildings.
     5.    Reaction rate for oxidation of SO^.
     6.    Removal rates by precipitation, deposition at surfaces,  and other reactions.
These  factors  interact  in such  a way  that  in  urban  and  industrialized  areas with  high
densities of  S02  emissions, the S02 concentrations  are  much higher  than in surrounding rural
areas.  It  is quite common to find gradients in S02 concentration within these industrialized
areas,  with  a central  core area  reporting the highest S02 concentrations.   This pattern is
shown diagrammatically in Figure 5-1.
                                            5-2

-------
UJ
o
 CM
O
to

iu
                                       MICRO AND MIDDLE
                                          WITH MAJOR POINT SOURCES
          RURAL AREAS
SUBURBS  |

         I
URBAN CORE

CITY LIMITS
I SUBURBS

I
RURAL AREAS
    Figure 5-1. Distribution of annual mean sulfur dioxide concentrations across an

    urban complex, as a function of various spatial scales,


    Source: Ball and Anderson (1977).
                                         5-3

-------
     Where SCL  emissions are dominated by a single source or a few point sources, the patten
of SOp  concentrations  could be different from the pattern displayed in Figure 5-1.  Depending
on  topography,  meteorology,  and  source  characteristics, the  concentration patterns  may bi
asymmetrical,  and the  temporal  distribution  may be skewed  to low mean  values with   a fev
intermittent  high peaks.   These  differences  in concentration patterns  may be  important ir
relation to the types of effects experienced in exposed human populations.
     Host  urban areas  have  experienced dramatic  improvements  in air quality as  a  result of
restrictions on sulfur  in fuel,  better controls  on  new and existing sources, displacement of
old sources  and building of new sources in less populated regions, and construction of taller
stacks.
     This  section presents  SO, concentration data  for specific  locations and  areas where
levels  are currently high.   The national status of SCL concentrations is reviewed, along with
temporal trend  data.   A  comparison is made  between SCL  levels  in six  cities  in  the early
1960's  and concentrations  in the  late  1970's.   Insights  into important determinants of popu-
lation  exposures  are presented  in the discussion  of diurnal  and  seasonal  SCL  concentration
patterns.  Since  SCL  can  be  measured  by a  variety  of  methods  (see  Chapter  3), a brief
discussion  of  SCL   monitoring  and  instruments  precedes  the  substantive  sections on  con-
centrations.
5.2.1  Monitoring Factors
     The EPA  is now in the  process  of  revising Federal, State, and local  air monitoring net-
works.   By  1981,  States  will  be  operating  a  selected  number  of  sites  in the  National  Air
Monitoring Station  (NAMS)  Network.  These sites  are  to  be  located in densely populated areas
with the highest  pollutant concentrations.   They are designed to serve in assessing pollutant
trends  and progress  in meeting   standards.   By 1983,  State  and  local   agencies are  to be
operating  the  State and   Local  Air Monitoring  Station  (SLAMS)  Network.   This network is
designed to  be  part of each State's  implementation  plan.   It is expected that this will mean
fewer sites than are currently in  operation; however, the Federal coordination of air monitor-
ing should provide  much-needed  quality control.  The trend  toward  reduction in the number of
stations is  already apparent  in   the 1977  SCL data.   There were  117  fewer monitoring sites
reporting data in 1977 than  in 1976 (2365 versus 2482).   Many States terminated all or most of
their 24-hour  West-Gaeke bubbler  sampling in  1978,  and most  remaining bubbler stations are
being fitted with temperature  controls to avoid sample degradation (see Chapter 3).   However,
state and  local agencies  are  relying  primarily  on continuous  monitoring equipment whenever
possible.
     Nationally, SCL monitoring is not as extensive as TSP monitoring.   In 1978 there were 947
sites with continuous monitoring  equipment and 1298 bubbler sites.   Every State conducted SCL
monitoring.  All  reported  sites  produced useful information on short-term (1- to 24-hour) SOg
concentrations.    However,  only those  sites  reporting  a  specified number of  hourly or daily
observations per year are considered valid in terms of their annual mean.   For the EPA,
                                            5-4

-------
minimum criteria  for  a valid annual mean  are  6570 hourly values from a continuous monitor or
five 24-hour values  in each quarter from a bubbler monitor.  It is in the number of S02 sites
with  valid annual  means that  national  coverage appears  inadequate.   Only  99 of  the 1298
bubbler sites  (or 7.6 percent)  had valid  annual  means in 1978;  only 385 of the 947 (or 40.7
percent) continuous sites were considered valid.  There were seven States with no valid annual
SO- data for  1978.   The EPA is currently  taking steps to improve the quality of SO- data and
to increase the number of representative sites reporting valid data.
     For valid bubbler sites,  the average number of 24-hour observations in 1978 was 60.  The
number of  observations per  site ranged from 28  to 322.   For the  valid  continuous  sites, the
mean number of observations was 7806 hourly measurements.  This ranged from a minimum of 6578
hours to a maximum of 8755 hours.
5.2.2  Sulfur Pioxide Concentrations
     Although  there are  natural  sources of SOg  such  as  vofcanoes (see Chapter 4) that can be
important  in proximity to the source,- they are usually unimportant on an urban scale.  Sulfur
dioxide has a rather short half-life in the troposphere (see Chapter 6), and background levels
are often  below  the monitor's detection limit.   Therefore,  it is not surprising that the re-
                                               3
ported annual  mean  S09 concentration is 3 ug/m   in some nonurban locations.  It may be lower;
                                                                   3
most monitoring  techniques  have detection limits  close to  10 ug/m , and apparent zero values
are commonly recorded as half the detection limit.
     Monitoring in urbanized areas near industrial sources that use sulfur-bearing fuels shows
rather high  concentrations  of  SO,.   In 1978  the annual  mean  concentrations  obtained by S0?
bubblers ranged from 3 to 79 yg/m .  The valid continuous monitors registered 1978 annual mean
                                         3
concentrations ranging from 3 to 152 ug/m  .
     The concentration  of SO,, is  affected by  meteorological  variables  influencing transport,
dispersion, and  removal,  as well  as by topography and configuration of sources.  Spatial and
temporal variations  in these parameters are reflected in the range of  maximum  and  90th per-
cent! le  concentrations reported  across the  Nation.   For  bubbler  sites,  the  lowest 24-hour
                                             3                           3
maximum value  reported by a site was 3  ug/m ; the highest was 907 ug/m .   For the valid con-
                                                                                      2
tinuous sites, the  spread of 24-hour maximum values was greater, ranging from 10 ug/m  at one
site to 2512 jjg/m  at another site.  Among all continuous sites reporting in 1978, the extreme
24-hour value was 3931 pg/rn .
     Figure 5-2  presents the distribution of  annual  averages  for  all  valid continuous mopi-
toring sites  in  1978.   On this time scale,  the most commonly measured values fell between 20
            3                                    3
and  30 \ig/m  ,  with  most  values  below 60  ug/m  .   Most monitoring  stations were  situated
specifically to detect higher urban or source-specific levels of SO-, however, and the data in
Figure 5-2 may be judged more nearly representative of populated areas or areas influenced by
specific sources  rather  than the entire U.S.  land area.   The following section discusses the
effect of site location.
                                            5-5

-------
at
ui
CO
1
u.
o

DC
UI
m

3
2
120




110




100




 90




 80




 70




 60




 50




 40




 30




 20




 10




  0
           10   20   30   40   50   60    70   80   90  100  110  120  130  140  150   160



                              ANNUAL AVERAGE CONCENTRATION



      Figure 5-2. Histogram delineating annual average sulfur dioxide concentrations

      for valid continuous sampling sites in .the United States in 1978.



      Source:  SAROAD.
                                         5-6

-------
5.2.B  Sulfur Dioxide Concentration By Siteand Region
5.2.3.1 Analyses by Various Site Classifications—In this section, the distributions of annual
mean and 90th  percentile  S0? concentrations by site descriptors are presented for bubbler and
for  continuous sampling  methods.   A  two-descriptor  code  has  been assigned  to  each  site.
Distributions  for every combination  of Type 1 (population,  source, background) and for Type 2
(central-city, suburban, rural, remote) are not presented.  In some cases the designations are
contradictory,  such  as   "population-remote"  or  "background-central  city."   The  purpose  of
presenting  these  distributions is to  permit  comparison  of these  two  categories of sampling
methods and to examine the SCL concentrations as a function of location.
     In  Table 5-1,  a cross-tabulation  of  mean  concentrations  by method  is  presented for
center-city sites that are primarily either population oriented or source oriented.  The third
and fourth numbers in each cell are the column percentage and total percentage of sites having
concentrations within the designated range.  Examination of each of these numbers reveals that
bubblers are,  on  the average, reporting lower concentrations than the continuous instruments,
an expected result because of the method biases reported in Chapter 3.   In Table 5-1, 14.4 per-
cent  of the  population-oriented  continuous  monitors  reported  mean concentrations  above  62
    •3
|jg/m , whereas only  1.6  percent of the bubblers reported such concentrations.   Of the source-
oriented sites,  a higher  percentage  (7.1  percent)  of the  bubblers were above  62 pg/m , but
this was still less than the 12.5 percent of the continuous monitors in this category.
5.2.3.2  Regional Comparisons—Regional  differences  in  SOn concentrations  are  not striking.
In  part,  this is the result of the location  of  the monitor.   In  Section 5.2.4,  it is shown
that high  SO- levels are found around smelters  in  otherwise clean areas.  In the eastern and
northern States,  most continuous S00 monitoring is in urbanized areas.   In 1978, mean concen-
                                                                                             3
trations across  all  continuous monitors  in Regions I, IIS III, IV, and V ranged from 23 ug/m
to  51  p.g/m  (see Table 5-2).  The maximum  annual mean among the valid sites in these regions
                    •3                         Q
ranged  from  59 |jg/m  in Region  I  to  140 pg/m  in Region III.   In the less industrialized or
less populated regions  (VI through X), the mean annual concentration across all sites in each
region ranged  from 8  ug/m  to 40 M9/m •
     Even  with the  summary of the 1978  continuous  SO- data, it  is  difficult  to speculate on
regional differences in  SOp  concentrations.   Sulfur  dioxide monitors  are  not systematically
sited  for   population exposure monitoring  purposes;  •sometimes  SO^ instruments  are  used for
monitoring  the  local  influence  of  strong  point sources  (e.g.,  the  smelters  noted above).
Therefore,  better  indicators  of  regional  differences  in  SO^ concentrations  and population
exposures  are  sulfur  emission patterns (see Chapter 4).
     The data  base used in compiling Figure 5-3, collected between  1974 and 1976, offers finer
spatial  resolution  of national SO, concentrations on a county scale.   The second  highest 24-
hour average S0? concentration by county is displayed.  Some areas  in the West with extremely
high  concentrations were  still  problem  areas  in the  late  1970's  (see  Table 5-3).  Several
counties and  cities  are  still  reporting  high concentrations;  however,  one should not  infer
that  the  reported  concentration  prevails  throughout  the county.   High readings may exist at

                                            5-7

-------
            TABLE  5-1.  GROSSTABULATIQN OF ANNUAL MEAN S02 CONCENTRATION BY                 OR CONTINUOUS)
                         FOR POPULATION-ORIENTED AND FOR SOURCE-ORIENTED CENTER-CITY SITES
01

00

Purpose of Site
Annual mean
S02 concentration,
Mff/ni3
2-7
Number of sites
Percent of row
Percent of column
Percent of total
7-18
Number of sites
Percent of row
Percent of column
Percent of total
18-33
Number of sites
Percent of row
Percent of column
Percent of total
^33-62
'Number of sites
Percent of row
Percent of column
Percent of total
>62
Number of sites
Percent of row
Percent of column
Percent of total
Column total
Number of sites
Percent of total
Bubbler

139
90.3
27.6
17.7

159
83.2
31.6
20.2

106
55.2
21.1
13.5

91
45.3
18.1
11.6

8
16.3
1.6
1.0

503
63.9
Population
Continuous

15
9.7
5.3
1.9

32
16.8
11.3
4.1

86
44.8
30.3
10.9

no
54.7
38.7
14.0

41
83.7
14.4
5.2

284
36.1
Row Total

154
—
—
19.6

191
__
—
24.3

192
__
—
24.4

201
—
—
25.5

49
__
—
6.2

787
100.0
Bubbler

9
75.0
16.1
8.7

18
64.3
32.1
17.3

12
46.2
21.4
11.5

13
46.4
23.2
12.5

4
40.0
7.1
3.8

56
53.8
Source
Continuous

3
25.0
6.3
2.9

10
35.7
20.8
9.6

14
53.8
29.2
13.5

15
53.6
31.3
14.4

6
60.0
12.5
5.8

48
46.2
Row Total

12
—
—
11.5

28
—
—
26.9

26
--
—
25.0

28
—
—
26.9

10
—
—
9.6

104
100.0

     Note:  1 ppm S02 = 2620 ug/nf
     Source:  SAROAD.

-------
                               TABLE 5-2.   CONTINUOUS S02 MONITOR RESULTS BY REGION,

Region Type
I Valid
All
II Valid
An
III Valid
An
IV Valid
All
V Valid
All
en
r
*° VI Valid
All
VII Valid
All
VIII Valid
All
IX Valid
All
X Valid
All
Number of
sites
22
72
51
87
26
108
100
203
111
254


13
32
13
38
12
49
19
52
18
29
Min.
6665
185
6597
140
657B
94
6678
421
6580
129


6631
1669
6769
334
6741
373
6857
105
6651
625
Number of
observations
per site
Mean Max. s.d.
7519
4582
7540
5815
7562
4381
8305
5754
7640
5512


7443
5461
7540
4676
7739
4694
7952
4973
7854
6158
8416
8416
8697
8697
8638
8638
8755
8755
8715
8715


8452
8452
8325
8325
8624
8624
8638
8638
8677
8677
567
2720
546
2380
670
2534
574
2848
625
2327


607
2072
439
2624
514
2358
507
2525
464
2526
Arithmetic Means
Min. Mean Max.
16
8
15
15
12
7
5
3
7
3


3
3
6
4
3
3
3
3
13
13
33
39
37
41
51
45
23
23
36
37


13
12
31
25
40
34
8
24
34
33
59
138
78
94
140
140
63
77
84
192


31
56
47
82
152
152
29
87
78
78
s.d.
12
23
16
19
23
21
12
13
16
25


1
13
14
20
47
39
6
20
18
17
90th Percenti le
Min. Mean Max.
34
14
35
35
34
14
9
'3
10
5


3
3
13
5
3
3
3
3
35
29
65
77
72
78
97
86
54
49
70
73


31
29
62
52
100
89
16
49
90
72
147
340
159
173
282
282
135
180
167
501


69
160
94
155
488
488
48
213
150
150
s.d.
27
52
30
33
46
40
27
29
30
50


19
38
25
41
146
113
12
49
38
38

Note:   1 ppm S02 = 2620 ug/m .
Source:   SAROAD.

-------
en
                Figure 5-3. Characterization of 1974-76 national SO2 status is shown by second highest 24-hr.
                average concentration. Asterisks denote counties for which this level exceeded 36§ /jg/m3.
                (The current 24-hr, primary standard is 365 ftg/m3, which is not to be exceeded more than once
                per year. Alaska and Hawaii reported no such exceedences.)

                Source: Monitoring and Reports Branch, Monitoring and Data Analysis Division, Office of
                Air Quality Planning and Standards, U.S. Environmental Protection Agency.

-------
       TABLE  5-3.  ELEVEN S02 MONITORING SITES WITH THE HIGHEST ANNUAL MEAN
                    CONCENTRATIONS IN 1978 (VALID CONTINUOUS SITES ONLY)

Annual Means,
Location pg/m
Helena, Deer Lodge
Co., MT
Pittsburgh, PA
Helena, Deer Lodge
County, MT
Magna, Salt Lake Co. ,
UT
Toledo, OH
Pittsburgh, PA
Buffalo, NY
Kellogg, Shoshone Co.,
ID
Shoshone Co. , ID
New York City, NY
Mingo Junction, OH
152
140
95
93
84
79
78
78
77
77
76
Maximum 24- hr,
pg/m Description
2512
602
1450
811
915
376
267
294
493
296
329
Rural mine smelter
Center-city industrial
Rural industrial,
1.6 miles east of
smelter
Suburban industrial
Center-city industrial
Suburban industrial
Suburban industrial
Suburban residential
Suburban industrial
Center-city residential
Center-city industrial

Note:   1 ppm S02 = 2620 ug/m~
Source:  SAROAD.
                                       5-11

-------
one or more monitoring sites (for example, Deer Lodge County, MT), but it is likely that there
are substantial gradients across the county and almost certainly across an air quality control
region (AQCR).
5.2.4  Peak Localized Sulfur Dioxide Concentrations
5.2.4.1   1978 Highest Annual  Average Concentrations—The  reasons  for variability  in ambient
air S02  levels  have been mentioned  in  earlier  sections.   This section examines United States
locations with  highest annual  averages and  highest  maximum concentrations  and analyzes the
distribution  of SCL  concentrations  nationally  by site descriptors.   However,  because of the
differences in S02 concentrations between bubbler and continuous monitoring, the distributions
by site descriptor will be treated separately for each method.
     Table  5-3  lists annual  mean and  maximum  24-hour concentrations  for the  11  valid con-
tinuous  monitoring  sites in  the United- States  with the  highest annual  means  in  1978.   The
highest annual  mean  concentration was 152 (jg/m  , at  a site in Montana 2 miles northeast of a
smelter.    That site  also had  the  highest  24-hour concentration  (2512  (jg/m ) of  any valid
                                                             o
continuous  monitor.   The maximum hourly  level  was  7205 (jg/m , and  the  second highest hourly
level was 6026  (jg/m  at this site.   Of the highest 11 sites, 5 were associated with smelters,
5 were associated with industrialized areas  or  towns,  and one (New York  City)  was  a densely
populated  city.   In  New York  City,  S02 emissions  from  space heating,  power plants,  and  a
variety of  industrial  sources  resulted, in a  high  annual  mean concentration.  These were peak
reporting  sites;  the  urban  sites  do  not  .typify  an  entire  city.    Conversely, even higher
concentrations may exist in unmonitored neighborhoods.
5.2.4.2   1978 Highest Daily  Average  Concentrations—About 50 monitoring  sites in  the United
States have consistently reported maximum 24-hour average  S0? levels in excess of 300 (jg/m  in
recent years.   Almost all  of these  have  had  very  high second and  third  highest values also.
Many  of  the  sites  having high  daily  averages  were  located  near  specific industrial sources
such  as  smelters,  steel  plants, and paper  mills.   Monitors around smelters  have  frequently
                                            o
reported 24-hour values of 1000 to 3000 (jg/m  , the highest levels in the United States.
     In 1978,  high  24-hour  S02 values occurred in 17 States encompassing all major regions of
the country.   Ten  of  the highest sites  were .in  Montana, six were in Wisconsin,  and six in
Minnesota.   Most  of  these  were close  to  intense  sources.   However,  several  urban sites,
especially   center-city   sites   in  industrialized,  communities   such  as  Philadelphia  and
Pittsburgh,  PA;  New  York,„  NY;  Toledo,  OH;-  and Hammond,  ID, still  have  high  maximum 24-hour
                             3
values, above 250 to 300 (jg/m .
5.2.4.3  Highest 1-hour Sulfur Dioxide Conce.ntrations-1978 National Aerometric Data Bank (NADB
Data—Single  hourly  S02  values  greater  than  1000 (jg/m   (0.4 ppm) have been measured in about
100 cities and counties in 28 States in recent years..  Such values were very widespread across
the country;  Maine,  Florida,  Montana,  Texas,,  Arizona,  and Washington all  had  sites in this
category.   Of these  top  100 sites,  all except  15  also had second highest values in excess of
1000
                                            5-12

-------
     Hourly measurements this high are comparatively infrequent and for most of these 100 high
sites, less  than 1 percent  of hourly  values  were  in  this category.   But for  a  few sites,
notably those close to metal-smelting operations in a few cities, such values were observed up
to 5  percent  of  the time.   Highest 1-hour  values  were  found in Deer  Lodge County,  MT, where
                                    2
several measurements over  5000 pg/m  were recorded at two sites in 1978.  Anaconda,  MT; Miami
and San Manuel,  AZ;  Newark,  DE; Buffalo County, WI;  and St. Charles  County  and  North Kansas
City, MO,  all reported at least one value in excess of 4000 pg/m .
5.2.5  Temporal Patterns in Sulfur Dioxide Concentrations
5.2.5.1   Diurnal Patterns—In  some  locations,  SO- concentrations   have  distinct  temporal
patterns.   These patterns depend on the variability of meteorological  factors and on the vari-
ability of source emissions.
     Diurnal variations  in SO^ concentrations  reflect the changing dispersion characteristics
of. the lower  atmosphere and  variations in mixing height.  If emissions are predominantly from
low-level  sources such as residential  and institutional  space heating, the highest hourly con-
centrations will  frequently  occur at night  and in the early-morning  hours.   At  these times,
low mixing height and decreased windspeeds lead to higher concentrations.  During the day more
vertical  mixing  usually occurs and windspeeds  increase, diluting low-level emissions.   Figure
5-4  gives  the composite  diurnal  pattern of hourly concentrations for SCL for  the month of
December 1978 in Watertown, MA.  The pattern just described is apparent.
     In locations where SO- emissions from taller stacks are the major SOp source, a different
diurnal pattern  can occur.   In these  situations,  typical  of power plants  and  smelters,  the
highest concentrations  usually occur  in the morning  hours  just after sunrise.  Levels can be
almost zero at  night  if the source is  emitting into a stratified region  above  a lower level
inversion.  Upon inversion breakup,  when  heating  at  the surface causes  vertical  mixing, an
elevated plume can be mixed to the ground.   Fumigation conditions lasting from several  minutes
to several hours can occur.  Montgomery and Coleman (1975) analyzed the effects of tall stacks
on the peak-to-mean ratios for different averaging times and discussed the influence of inver-
sion breakup.  In essence, even with tall stacks, inversion breakup that catches the plume and
brings it  to  the surface can occur.  So the peak-to-mean ratio is almost independent of stack
height.  The  frequency  of occurrences of fumigation, on  the other hand, would most likely be
less with taller stacks.
     Some  similarity  can be  found in  comparing the diurnal  pattern of  hourly  averages  for
Watertown, MA  (Figure  5-4) and St. Louis,  MO  (Figure 5-5).   In February  1977,  a major local
source of  S0? was still in operation in St. Louis.  The midmorning and  late night maxima were
again  associated with  diurnal  variations  of  meteorological  factors.    By  February  1978,  the
source had shut down,  and the reported SO,, levels at  the monitoring  stations reflected this
fact.  The absence of low-level stacks emitting  into a  stable layer of air near the surface at
night  was  noticeable.   Concentrations  did  not build  up  at night as in  the  previous year.
                                            5-13

-------
     0.040
    0.030
I
cc


Ul
o

o
u
 CM
O
0,020
    0.010
                      1      I      I      I       I      I       I      I
                I	I
                                            lilt
                                                                      I       I
I
I	I
                                   8     10     12    14     16    18     20    22



                                                  HOUR
                                                                                  24
Figure 5-4. Composite diurnal pattern of hourly sulfur dioxide concentrations are shown for

Watertown, MA, for December 1978.



Source: Spengler (1980).
                                                5-14

-------
   0.200
   0.100
   0.090
   0.080
   0.070
a 0.060
z*
2 0.050
   0.040
ui
o
§  0.030
 CM
O
Ui
   0.020
   0.010
            I   I    I   I    I   I    I   I   I    I   I    I   I    I   I    I
I    I   I    I   I
                                                 FEBRUARY 1977
                                                 FEBRUARY 1978
            I    I   I   I   I    I   I    I   I    I   I    I   I    I    I   I   I   I    I   I    I
                                     8   9  10  11  12  13 14  15 16  17  18  19  20 21  22 23 24
                                                 HOUR
Figure 5-5.  Monthly means of hourly sulfur dioxide concentrations are shown for St. Louis (city site
no. 26-4280-007, "Broadway & Hurck") for February 1977 and 1978.
Source: Spengler (1980).
                                                5-15

-------
     Diurnal  patterns  of  hourly SO,  values  for the  industrialized river  valley  town  of
Steubenville, OH, are shown in Figure 5-6.  In June 1976, a distinctive maximum in the diurnal
pattern appeared.  In July 1977, there was no apparent variation across the hours.
     In conclusion, it can be said that the variations in hourly concentrations are influenced
by source configuration and meteorological dispersion.  Therefore, it is difficult to general-
ize  about  diurnal  patterns  of  hourly  S0~  concentrations.   Although  there  may  be  some
similarities, the  daily  patterns in SCL  concentrations  are  different for different locations
and can change in time for a given location.
5.2.5,2   Seasonal  Patterns—Concentrations  of  SO, display  seasonal variability.  The  vari-
ability  is  most  pronounced  in  areas  in which  there  is  strong  seasonal  variation  in the
emission-source  strength  or  in  meteorological  conditions.    Obviously,  in urban areas  where
space heating is  the  major source of S02,  the  levels will  be  much  higher during the heating
season.    Figure  5-7 illustrates  just such  situations in Watertown,  MA  and  Steubenville, OH.
The highest monthly mean concentrations occur in the winter months.
     Figure 5-7 also shows the data for St.  Louis, MO where the seasonal pattern  is different.
Here a  local  industrial  s'ource dominates SQ~ concentration  patterns around  the monitor.   The
higher monthly mean concentrations occur in the months with the higher frequency of southerly
winds.   The  source is  to the south of  the monitoring station.  Any  increase  in SO^ concen-
trations as a result of the winter heating season is not apparent.
5.2.5.3  Yearly Trends—The SQ2  levels  in most urban areas in the United States  have improved
steadily  since  the mid-I9601s.   The trend of  decreasing S02  concentrations  can be  resolved
into three distinct periods.   From 1964  to 1969,  the improvement was gradual.   In the middle
period,  between 1969 and 1972, the improvement in most urban areas was more pronounced.  Since
1973 the  improvement  has again become slower.   The 1977 EPA trends  report  states:   "In  most
urban areas,  this  is  consistent with the switch in emphasis from attainment  of  standards  to
maintenance of air quality;  that is, the initial effort was to reduce pollution to acceptable
levels followed  by efforts to maintain air quality at these lower levels."  From 1972 through
1977 annual averaged  SQy  levels dropped by 17 percent, or an annual improvement rate of about
4 percent per year.   Figure 5-8 summarizes the annual average SQy concentrations for 32 urban
National  Air  Sampling  Network  (NASN)  stations  for  the years  1964  through  1971.   In  this
figure,  the  first-two  periods are  apparent.   In  Figure 5-9,  the  national trends  in annual
average SO,  concentrations from* 1972 through 1977 at 1233 sampling  sites are displayed.  In
this figure, the  diamond  symbolizes the  composite  annual average concentration;  the triangle
is the median value,  while the circles are extreme values  and the thick band covers the 25th
to 75th percent!"le range.
     Over the  period  of 1970  through 1977,  SQ2 emissions have decreased  only slightly (U.S.
Environmental Protection  Agency, 1978).  In  1970 the estimated  annual manmade S02 emissions
were 29.8 million metric  tons.   By 1977 this  was reduced  to 27.4  million  metric tons.   The
improvement in  the ambient air  quality  levels  for S02 reflected  the  displacement of sources
from urban  areas to  rural  areas, restriction  of  sulfur content of  fuels  used  by low-level

                                            5-16

-------
    0.080
    0.070
    0.060

Q.
Q.



O   0.050
UJ
u
    0.040
o   0.030
 CM

8

    0.020
    0.010
                                                                     I       I
I       I      I      I
                                                 I
                   _L
                                          10
             12     14



              HOURS
16
18
20
                                                                                  22
24
 Figure 5-6.   Monthly means of hourly sulfur dioxide concentrations are shown for Steubenville, OH

 (NOVAA site 36-6420-012) for June 1976 and July 1977.


 Source:  Spengler (1980).
                                               5-17

-------
   0.040
Q.
Q.
 %


o


St
EC
   0.030
   0.02O
U]

u



O

u

 CM

O
U)
   0.010
                                    CITY       SITE


                                STEUBiNVtLLE 36 6420 012


                                ST. LOUIS     26 4280 007 Q—-—Q


                                WATERTQWN   22 2380 003
                             I
I
I
I
I
              JAN    FEB   MAR   APR   MAY   JUNE   JULY   AUG  SEPT  OCT   NOV   DEC


                                                  MONTH




    Figure 5-7.  Seasonal variations in sulfur dioxide levels are shown for Steubenviile, St. Louis, and

    Watertown.



    Source: Spengler (1980).
                                                    5-18

-------
         1964
1972
Figure 5-8. Annual average sulfur dioxide concentrations
are shown for 32 urban IMASN stations.

Source: National Academy of Sciences (197i).
                         5-19

-------
                1972
                                                                        90TH PERCENTILE
c_ O> •* 	 COMPOSITE AVERAGE
56

48

to
-I
OI
a, 40
g
cc
z 32
UJ
o
z
o
o
111
S 24
X
o
o
cc
3
^3 16
3
CO
8
n

C




-


—






_





_



I I I A •+ 	 MEDIAN
5
0 I














O





A













T
(>« 	 10TH PERCENTILE
0
0
0 n -


—




•»> . * , '

i 	 I mil a 1 	 | —

^^ >V yv
o o


—
A A A A
A
1
—
* T T T I
1973
1974        1975

      YEAR
                                                                   1976
1977
Figure 5-9.  Nationwide trends in annual average sulfur dioxide concentrations from 1972 to 1977 are
shown for 1233 sampling sites.

Source: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards (1978a).
                                        5-20

-------
area sources, building of sources with taller stacks, and source emission controls.
     The first  air quality criteria  document  for S0x,  published in  1969,  presented the fre-
quency distributions for SOp levels in selected American cities for 1962 through 1967 (National
Air pollution Control Administration, 1970).  The 1960's data came from the Continuous Air Mon-
itoring Project (CAMP), which operated continuous monitors in a few of the largest U.S.  cities:
Chicago, Philadelphia, St.  Louis, Cincinnati, Los Angeles, and San Francisco.   Improvements in
SOg levels in each of the six cities are demonstrated by comparing 1962 to 1967 data with data
for 1977 (Table  5-4).   In  each city  there  is  more than one continuous monitor now operating.
The station  reporting  the  highest levels in 1977 was  used in order  not  to overemphasize any
improvement.   The  comparison is  only an approximation  because the locations  of the monitors
and the instrumental  methods  used were  not  the  same- as those reported in  the 1969 document.
     In each city the  peak concentration  decreased.   In most cities -the  1977  peak was less
than one-half the earlier  values.  The  only exception  was St. Louis, where  the earlier peak
was 0.72 ppm and the 1977 peak was 0.67 ppm. The result was not unexpected in that the earlier
summary was a composite frequency distribution of 5 years of monitoring.
      TABLE 5-4.  COMPARISON OF FREQUENCY DISTRIBUTION OF S02 CONCENTRATION (P£M)
                           DURING 1962-673 AND DURING 1977°

Frequency Distribution
City
Chicago
Philadelphia
St. Louis
Cincinnati
Los Angeles
San Francisco
Year
1962-67
1977
1962-67
1977
1962-67
1977
1962-67
1977
1962-67
1977
1962-67
1977
Percenti le 30
0.03
0.01
0.03
0.01
0.02
0.01
0.01
0.01
0.01
0.01
::
50
0.08
0.02
0.05
0.02
0.03
0.01
0.02
0.02
0.01
0.02
0.001
70
0.17
0.03
0.09
0.04
0.05
0.03
0.03
0.02
0.02
0.02
0.01
0.01
9,0
0/32
0.06
0.-.21
0.08
0.11
0.10
0.07
0.04
0.04
0.03
0.03
0.01
of SO,, (ppm)
99
0.65
0.12
0.45
0.23
0.26
0.37
0.18
0.08
0.08
0.05
0.07
0.03
Maximum
0.95
0.25
0.85
0.44
0,72
0.67
0.53
0.29
0.25
0.09
0.17
0.03

    Concentrations from CAMP stations as reported in Air Quality Criteria for
    Sulfur Oxides, National Air Pollution Control Admini strati on (1970).

    Concentrations from National Aerodynamic Data Base (1977).
                               3
   Note:  1 ppm S02 = 2620 ug/m .
                                            5-21

-------
     More stable  indicators  of improved air quality are  provided by the 50th, 70th, and 90th
percentile  concentrations.   In Chicago,  the 50th percentile concentration  dropped  from 0.08
ppm to  0.022 ppm.   In Philadelphia,  the  levels improved substantially; the  50th,  70th,  and
90th  percentile concentrations were  less than  one-half  the earlier  values.   St.  Louis  and
Cincinnati  showed modest  improvement;  their 50th percentile concentrations were lower in 1977
than  they were in the mid-1960's.   The highest concentrations occurred as  frequently  in  St.
Louis as they did earlier, but in Cincinnati they occurred less frequently.   Review of the St.
Louis SO, data showed  improved air quality for most  of the city.   The  high concentrations
reported  in 1977 were  typical  of only a  small  section of the city.  Los Angeles showed  im-
provement  in reducing  the  high  concentrations,  but  the  50th  percentile  concentration  was
actually slightly higher in 1977 than it was in the previous decade.   Similarly, San Francisco
trimmed the peaks but had a very low median value.
     In  summary,  the  frequency of peak  levels has been reduced in  most urban  areas.   The
steady improvement of  S09 ambient air quality has been slowed somewhat in recent years.   Only
                                                                 3
1 percent of the S02 monitoring sites  show  levels above 80 ug/m  ,  the  current annual  NAAQS.
In 1974, the annual  mean S0_ standard was exceeded in 3 percent of the monitoring stations (31
of 1030), compared with 16 percent in 1970.  In 1977 and 1978, 2 percent of the sites reported
violations of the 24-hour standard.  In 1974, this standard was exceeded in 4.4 percent of the
reporting stations (99  of 2241),  compared with  11 percent in 1970.   Many of  these  sites  re-
porting  violations  of  the  24-hour  standard are  in  remote areas  near  large  point sources.
5.3  AMBIENT MEASUREMENTS OF SUSPENDED PARTICULATE MASS
     The  general  character  of  airborne matter designated as  atmospheric  suspended particles
                                                                                  — Q
has been  described  in  Chapter 2.   These particles range in size from about 5 x 10  m, roughly
                                                                                          -4
corresponding  to agglomerates  of  a few  tens or  hundreds of molecules,  up to  about  10  m,
specks of material discernable to the human eye.  A useful division of these particles by size
into fine and coarse fractions occurs in the range of 1 to 2 x 10  m or 1 to 2 urn, as was dis-
cussed in Chapter 2.
     The mass of  suspended particles,  generally concentrated in particles above about 0.3  urn,
is  usually   estimated  by  filtration of known volumes  of air.   The  goal  of  this filtration
process is  the  separation of the gas  phase  from liquid and solid condensed  phases  of  atmos-
pheric aerosol.   Thus,  the mass of material accumulated on a filter is taken to represent the
volume of aerosol filtered,  and  results  are  presented as  |jg  of PM/cubic  meter of aerosol,
abbreviated ug/m .
     Chapter  3  discusses some  of the complications  of commonly  used filtration  methods,
including retention of  reactive gases  such  as  SO-  and HN03 and loss  by evaporation of water
and other  moderately  volatile substances.   While commonly used  filtering  media are  highly
efficient for  collection  of  fine  particles, samplers  of  coarse-particle  concentrations have
had major design defects.   Despite these  complications,  which were  discussed in  detail  in
Chapter 3,  the largest body of information  on  the distribution of suspended  PM  in time  and
                                            5-22

-------
space has  been obtained with  the hi-vol,  or TSP method.  Routine  monitoring  information is
available  from the  NADB maintained  by  EPA  for many  sites.   Some are  EPA sites;  many  are
operated by State and local  agencies.
     The following discussion  relies  mainly on NADB data; additional analyses are to be found
in the  1976 and  1977  National Air Quality and  Emissions  Trends reports,  in  Trends  in  the
Quality of  the Nation's Air  reports  in  1980,  and  in  the document Deputy  Assistant Admini-
strator' s  Report cm Ambient  Monitoring  Activities  -  Air  Portion.  (U.S. Environmental  Pro-
tection Agency, 1977; 1978;  1980a; 1980b).
5.3.1  Monitoring Factors
     The accuracy and  precision  of PM monitoring are limited by three general considerations:
sampling methods, including instrumentation,  analytical methods, and quality assurance; samp-
ling frequency;  and  location  of  monitors.  Chapter  3 discusses  the first of these considera-
tions;  the  second and  third  are  discussed  in this chapter.  Sampling  frequency affects  the
confidence limits on mean TSP concentrations and annual  or seasonal trends.  It is appropriate
to discuss  this  limitation  at the beginning of this section before the 1978 national TSP data
base  is  presented.   The  siting  of  particle  monitors significantly  influences  the  levels
measured  and,  hence,  the*' interpretation  of  data.   These  considerations are  presented with
examples in several  sections of this chapter.
5.3,1.1   Samp1ing Frequency--1n  1978,  there  were  4105  TSP  monitoring  sites  in  the  United
States  and  its territorie*s that  reported data to the  NADB  of the  U.S.  EPA.   Of these, only
2882 had  enough  observations  per quarter and per year for the data to be considered valid for
estimating  annual  averages.   The  number  of  sites  reporting valid  data ranged  from zero in
Delaware  and American  Samoa  to  318  in  Ohio.  The  most populous states, California and  New
York, had 60 and 236, respectively.
     The U.S.  EPA has  established a uniform sampling schedule to be followed by all State and
local agencies.   It  requires  a 24-hour sample (midnight to midnight) every sixth day.  Hence,
in 1 year there  are 60  or 61  possible  sampling days from which  to  derive the mean value and
distribution,  and  to determine attainment  of current  standards.   In 1978,  14  percent  ofxall
reporting sites had 60 or more observations.
     Sampling days are missed and samples must be voided for a variety of reasons.  Therefore,
a minimum  requirement  has  been established for considering the data from any site as valid in
determining an annual  average:   There must be  at least five observations during each quarter
of a  calender  year.   Of the  Federal, State, and local TSP sites reporting  data to NADB, 70
percent met this requirement.
     The  distribution  of observations for the 2882 valid monitoring sites in 1978 is shown in
Figure 5-10.  Of these sites,  10 percent had less than 47 observations and 50 percent had more
than  56 observations.    However,  90  percent  of the monitoring sites collected fewer than 60
samples.  Three  percent of  the monitors sampled at an equivalent frequency of 1 day  in 3,  and
fewer than 2 percent collected samples at a frequency of 1 day in 2.
                                            5-23

-------
   1000
   900
   800
   700
   600

§  SOO
ui
S  400
D
Z
Z
K
Ul
VI
CO
O
cc
X
u.
O
cc
Ul
CD


z
   300
g  200
u.
CO

O
   100
    90
    80
    70
    60

    50

    40

    30
    20
    10
                               I    I
                                        I   I   I   I  I   I   I
                                       1   I
                                            IT
         I  I  I  I   III
                               I
I
I    I   I   I   I   I
                                                            I
_L
_L
I	L
l
     0.01  0.1   0.5  1   2  5   10  20   30  40 SO  60 70  80 90    95  98   99 99.8   99.9    99.99

                             % WITH NUMBER OF OBSERVATIONS LESS THAN


           Figure 5-10. Distribution shows the number of TSP observations per valid site in 1978;
           total is 2882 sites.

           Source: SAROAD.
                                              5-24

-------
     The current NAAQS  for TSP consists of an annual geometric mean and a once-per-year daily
value.   Frequency  of  monitoring is a  fundamental  parameter of the aiir  quality  data used fov
comparison with the standards.   The period of determining an annual average for comparison is
a calendar year.   If  the number of 24-hour  observations  is less than 365, then the true mean
concentration for the year can only be expressed as residing within a range of values.  On the
assumption that the actual  distribution of values  is  log normal, confidence intervals can be
calculated from  the geometric mean and the geometric standard deviation.   Figure 5-11 shows
the effect of sample size on the 95-percent confidence intervals for a hypothetical site whose
true annual  geometric  mean is equivalent to 75 pg/m , the current annual standard.  From this
example, we  can  only  conclude that with  95  percent probability the annual  geometric mean is
                       3
between 67 and 85  |jg/m , if the mean  for that year was  determined  from a sample size of 61.
Increasing the sampling frequency  to  1 day  in  2 reduces  the level of uncertainty so that the
annual mean  is known  to lie between 70 and  81 ug/m .  Thus, by  increasing the sampling fre-
quency by  a factor of three, the  width of  the 95-percent confidence  interval  for the true
annual geometric mean has been cut by a factor of 1.7 (square root of 3).
     A critical  factor  in  evaluating  compliance with once-a-year standards  is  the  effect of
sampling frequency.   Figure 5-10 shows that  in  1978  the  majority of valid sites (80 percent)
had fewer  than 60  sampling days.  The  sites  with more frequent sampling had a greater chance
of  sampling  the  higher concentrations (Figure  5-11).    Assuming that  there are  a  number of
days on which the observations are above the standards, the probability of selecting 2 or more
days  on  which standards  are exceeded  is a  function of sampling frequency.  If  there are 10
days above the standards,  there is only  a  slightly better than 50-percent chance of actually
monitoring on 2 of those days given a sampling frequency of 61 out of 365 days.   When the sam-
pling frequency is doubled to 122 sampling days, the probability of capturing 2 days out of 10
that exceed  the  standards  increases to 80 percent.   In actuality, samples are not taken ran-
domly; they are taken systematically,  usually at a rate of once every 6 days.  The probability
of  capturing the  highest period is further complicated in that the log normal distribution of
TSP concentrations does not apply uniformly to all sites.
     An additional  complication  in determining compliance or trends  occurs when the meteoro-
logical regimes affecting  the TSP concentrations are considered.  Attainment of standards may
depend  on  the number  of "clean"  sampling  days  versus the number  of  "dirty"  sampling days.
Watson (1979) exemplifies this problem with Portland, OR,  TSP data.  The annual  geometric mean
TSP data  show  a  decreasing trend from  1973 through 1975,  with a significant increase in 1976.
If  these  data are  reexamined  and  weighted  by the meteorological  regimes  actually sampled in
each year,  the  conclusions are changed.  The  stratified  mean TSP values show a large drop in
concentrations occurring between 1974  and 1975;  the  levels  are constant for the years before
and after.  Since these means are determined from a sample set varying from 49 samples in 1974
to  79  samples  in 1976, a  statistical  test  is required to determine whether the means of any
year  are significantly different  from those  of any  other year.  The  95-percent confidence
                                            5-25

-------
   95
   90
OJ
* 85
Z
o
a  80
z
UJ
u

1  75
a.
   70
    65
                                    D
ABOVE THE STANDARD

AREA OF<
UNCERTAINTY

BELOW THE STANDARD
                   61    91  122     183                  365

                  NUMBER OF SAMPLING DAYS PER YEAR
 Figure 5-11.  The 95 percent confidence intervals about an annual
 mean TSP concentration of 75 fig/m? is shown for various sampling
 frequencies {assume the geometric standard deviation equals 1.6).

 Source: Curran and Hunt (1975).
                                  5-26

-------
Intervals for all  stratified  means do overlap.  Watson  concludes,  then,  that there is a rea-
sonable chance that the true means do not really vary from year to year.
5.3.1.2  Monitor Location—The  choice of  sampling  location can obviously affect  the concen-
trations measured.   Remotely located monitors typically record low concentrations;  urban moni-
tors  characteristically  record  higher concentrations.   The  positioning  of  a monitor at  a
chosen location can also affect measured concentrations.   For example, at a specific location,
the height of the  monitor above the  ground  influences  sample concentrations.  If the monitor
is  elevated  above  surface  sources,  lower  concentrations  of coarser  particles  might  be
measured.   Some  studies  clearly  indicate  that TSP  concentrations decrease  with increasing
monitor elevation  (Record  and Bradway, 1978; Record et al., 1979;  Lioy et al., 1980a; Pace et
a!., 1977),  with distance from a roadway, and with distance from other nearby sources.
     The inferences  drawn  about air quality levels, trends, and population exposures from the
TSP data presented in  this chapter are made in full knowledge of the following limitations of
TSP monitoring:
     1.   Sampling sites are not standardized.
     2.   Frequency of sampling is quite varied.
     3.   The majority of sites reporting have fewer than 60 sampling days per year.
     4.   The frequency of sampling is not weighted with respect to meteorological  conditions.
     5.   No spatial averaging is used in analyzing or reporting data
     6.   Though the ambient  air monitor  -is  stationary,  the population  it is  intended to
          represent is highly mobile and spends some time indoors.
5.3.2  Ambient Air TSP Values
     The distribution  of  1978 annual  arithmetic  means for  valid TSP  monitoring  sites is
plotted in Figure 5-12.  Half of all the nation's sites had annual  arithmetic mean values less
             3                                               3
than 60  ug/m .  Annual  mean values range from 9 to 256 ug/m .  Only 14 valid sites had annual
                                                     3
mean  concentrations equal  to or  less  than 16  ug/m .   These  lower  values  were  recorded in
remote  monitoring   sites.   Two  background  sites,   Glacier  National  Park,  MT,   and  Acadia
                                                               3
National Park, ME,  had 1977 "annual averages of 11  and 21 ug/m , respectively.  At  the other
                                                                                     3
end of  the  distribution,  25 percent  of sites  had  annual means greater than  76  ug/m ,  and 10
percent were greater than 96 ug/m .  Higher annual  concentrations were found in many populated
                                                                                            3
and  industrialized areas.   About 30  sites  reported annual  averages in 1978  above 150 ug/m .
Topping  the  list  were  four  central-city  sites  in commercial,  residential, or industrial
                                                                              3
settings.  A  Phoenix,  AZ,  site (0136) had the highest annual mean of 256 ug/m , followed by a
                                  3                                                         3
site  in  Calexico,  CA,  at 201 ng/m  and  an industrial  site in Granite City,  IL, at 197 H9/m -
     These  extremely high  annual  TSP concentrations were found in commercial  and industrial
locations.    Of  the 30 highest  sites,  15  were  industrial.  Many of the  higher concentrations
(19  of 30)  were found at  central-city locations.   Only four were  classified as  rural  sites,
most  of which were  also residential  areas.   It is also likely that  arid climates  and dusty
conditions  in the  vicinity of  some monitoring sites might have led  to  suspension of surface
material.  However,  it is impossible to ascertain the contribution of fugitive or resuspended

                                            5-27

-------
   1000
   900
   800
   700
   600
   500

   400

   300


"H 200

g

g
EC
21 100
g  90
1  80
0  70
2  60
P  50

    40

    30


    20
      IT  I   I   I   I     I    I

     O    O 90TH PERCENT! LE

     D 	Q MEAN
                          T   I   I   I   T
                                                                      1   T
                        I	I
                          I    I   I   I   i   i   I
                                                             J	I
                                                                       J	I
I  I
0.01
0.1
                 0.5 1  2    5  10   20  30 40 50 60  70  SO   90  95   98 99   99.8 99.9 99.99

                               % OF SITES REPORTING ANNUAL MEAN

Figure 5-12, Distribution of mean and 90th percentile TSP concentrations is shown for valid 1978 sites.

Source: SAROAD.
                                            5-28

-------
20
40
60
                   80
100   120   140   160   180   200   220  240  260  280

 TSP CONCENTRATION,,
Figure 5-13. Histogram of number of sites against concentration shows that
over one-third of the sites had annual mean concentrations between 40 and
60 uglm3 in 1978.

Source: SAROAD.
                              5-29

-------
dust  to the  concentrations measured  at  these 30  sites  without more detailed analysis.  The
histogram  of  sites  against  concentrations  (Figure 5-13)  shows  that • over  a third  of all
monitoring  sites had annual mean  values  between  40 and 60 ug/m .  Slightly  less than another
                                                3
third had annual averages  between  60 and  80 ug/m  .
     The distribution of 90th percentile  values  is also plotted in  Figure 5-12.  Half of all
                                                                Q
valid monitors had  a  90th  percentile value in  excess of 97 ug/m .   For 10 percent of the moni-
tors, 10 percent of the observations  exceeded 160 (jg/m ,  For  one monitor,  10 percent of the
                              3
observations exceeded 600  ug/m .
     Daily,  or 24-hour, TSP  concentrations have  a wide range.  In  remote  areas  such as the
Pacific  Islands, daily  values  may be as  low  as  a few micrograms  per  cubic meter.   Over the
                                                           2
continental United  States,  concentrations  from 5 to 20 pg/m  are routinely reported.  In other
locations,  daily TSP values can exceed 10 times the levels found in  remote areas, on occasion
                    3                             3
exceeding 3000 ug/m .  Values exceeding 1000 ug/m  are observed  in  remote arid regions as well
as  in  populated urban areas.  Daily  TSP   levels approaching these  higher values,  500 to 1500
    o
fjg/m , are frequently associated with  adverse  meteorological conditions:  low-level inversion,
stagnation, or high winds  resuspending surface material.
                                                                                    o
     Thirty valid TSP monitoring sites report highest 24-hour values above 600 |jg/m .  Only a
few of these sites  are in  the top  30 in annual  average.   In cities  like Topeka, KS, and Libby,
HT, which  are not  densely populated or  industrialized,  these  high concentrations  may result
from chance occurrences, such as  fires or dust storms.   In other cities like El Paso, TX, and
Granite City,  IL, which are industrialized, the  maximum  concentrations are more likely to be
related to persistent sources of pollution.
5.3.3  TSPConcentrations  by Site andRegion
     Airborne particles  measured  by the  hi-vol TSP  method  arise from many sources, and these
are described in Chapter 4.  Important source  categories include:
     1.    Fugitive  dust emissions  stirred up by mechanical  action  or the  wind as  in dust
          storms.
     2.    Large point sources such as  smokestacks.
     3.    Chemical  reactions  in  the  atmosphere  that transform gaseous substances,  such  as
          sulfur and nitrogen oxides, ammonia,  and volatile organic species, to PM substances.
     4.    Low-level   widely  dispersed combustion  sources  such  as  automobiles  and  trucks,
          residential furnaces, fireplaces, and wood stoves.
     5.    Occasional  or sporadic  sources that' can  in  some  instances cause  very  high  TSP
          values.   Examples are agricultural tilling and burning, wildfires in forests, grass-
          lands or  even cities, and roadgrading.
The relative contributions of these sources can vary widely and  are influenced by location of
TSP monitors  relative to  potential  sources  and meteorological  conditions.    Hence,  it  is  not
surprising to  see a wide variation in daily measurements at a  single site, variation over an
urban area,  or variation  among regions of the country.   This  section  will,  by illustration,
examine these differences  in TSP concentrations by location.

                                            5-30

-------
5.3.3.1  TSP by Site Classifications—Evidence that fugitive dust contributes significantly to
both western  arid  sites and many urban  sites  is quite extensive.  Discussion  of  the general
influence of fugitive dust may be found in section 5,6.4, below.
     The general  differences in annual  TSP  concentration among locations are  seen  in Figure
5-14,  These  differences reflect the character of  the neighborhoods where  the monitors are
located.   This figure  summarizes  the mean concentrations from 154  sites in 14 cities.  Resi-
                                                                                3
dential neighborhoods  in  and near cities have TSP levels between 50 and 70 |jg/m .   Commercial
                                                           3
sites  have a  wider range of concentrations (60 to 110 yg/m ).   Industrial locations generally
range between 80 to 150 ug/m .
     The 1978 data base also has been analyzed on the basis of two additional sets  of descrip-
tors.  One description  scheme  classifies monitoring sites by their purpose:   population expo-
sures, source receptors, or background sites.  The other scheme identifies sites by the amount
of  development:    central-city,  suburban, rural,  or  remote.   These  classifications  are not
mutually exclusive.
     When sites are  grouped by descriptors,  a distinct weighting becomes apparent.  Almost 80
percent of the sites are population oriented; approximately 15 percent are source related, and
less  than  6 percent are  background monitors.   The distribution by development also reflects
its  population  emphasis.   Of  the  total  monitors,  83 percent are  at either  central-city or
suburban sites, 15 percent  are at  rural  sites,  and  2 percent are at  remote sites.   In these
data,  38  percent  of the background sites had median  values  less than  or equal to  21 ug/m ,
whereas only 4.4 percent of all sites had these low values.   Only 30 percent of the background
                                      3                               3
sites  had median  values above 44 ug/m  (none had values above 97 ug/m ), whereas 75.5 percent
                                                 3
of  source sites  had median  values  above  44  M9/m •   The pattern is consistent for the distri-
bution of  the 90th  percentiles cross-tabulated by  site purpose.  Cross-tabulations of site
median values and site 90th percentile values with the development-related site descriptors is
further confirmation of the influence of location on  measured  TSP concentrations.  Rural and
remote sites  have lower  median values  and lower 90th-percentile values.  The  suburban sites
reflect the overall national distribution.  The central-city category has proportionately more
sites  in the higher concentration ranges.
5.3.3.2   Intracity Comparisons—Because  of the  strong neighborhood influence  on  TSP concen-
trations, it  is not unusual to  find considerable variation in peak  and mean  concentrations
across a  community.  Examination of intracity differences illustrates the difficulty in esti-
mating population exposures to TSP.
     Data on the nine cities having the highest annual TSP concentrations in 1977 are given in
Table  5-5.   Only  sites  having enough  observations  per quarter  to report an annual  mean are
used.   Although TSP  concentrations  in  these cities  were  generally  high,  in  1977  the less
developed or  less  industrialized areas in each  city  had annual  geometric mean concentrations
              3
below  75  ug/m ,  (currently  the annual primary NAAQS), with the exception of Granite City, IL.
                                            5-31

-------
   150
en
a.

H
   100
a:

z
Ui
u

o
u
Q.
CO
50
            RESIDENTIAL
                         COMMERCIAL
INDUSTRIAL
 Figure 5-14.  Histogram of mean TSP levels by neighborhood shows

 lowest levels in residential areas, higher levels in commercial areas,

 and highest levels in industrial areas.


Source: Office of Air Quality Planning and Standards (1976).
                                 5-32

-------
                   TABLE 5-5.   RANGE OF ANNUAL GEOMETRIC MEAN CONCENTRATIONS IN
                            AREAS WITH HIGH TSP CONCENTRATIONS IN 1977

Number of sites Annual
Number with annual average range,
City of sites >75 pg/m3 yg/m3
Tucson, AZ
Pocatello, ID
Chicago, IL
Granite City, IL
Taos County, NM
Middletown, OH
Cleveland, OH
Youngstown, OH
El Paso, TX
7
4
25
8
1
3
23
5
14
3
3
12
8
—
2
13
4
10
67-156
65-218
50-170
85-185
168
64-192
48-152
66-172
60-158
Range of
maximum 24- hr
value, jjg/m3
178-591
344-1371
152-1106
227-485
577
157-707
128-705
163-602
205-691

il mean concentration
:leanest sites within
Regional Differences
for
the
in
the dirtiest sites
same city.
can be two to four

Background Concentrations — It has been
times higher tha

demonstrated th
concentrations  can  vary  across an  urban area  and  among cities  with different  sources  and
meteorology.   In  addition,  there  may be regional  differences in the  natural  or transported
fraction  of TSP  concentrations.     Figure  5-15  shows  the contribution  of these  sources  to
nonurban levels.  It was assumed that the global and local  contributions in the average would
be similar.  The  greatest difference among regions is the contribution from "continental"  and
transported  emissions.   These  two  categories  of particles  contribute  in such  a way  that
nonurban  sites in  the West  typically  report  annual  geometric means  of  15  ug/m ;  in  the
Midwest, 25  ug/m  ;  and in the  East,  35  pg/m  .   Except for the Acadia  National  Park site (18
ug/m ) and Millinocket (23 pg/m ), all sites in Maine had 1977 annual geometric means above 30
|jg/m .  Nonurban sites in Wisconsin had mean TSP levels less than 25 pg/m .   Nonurban sites in
Montana had  levels  less  than 20 ug/m  in 1977; the individual means were  Big Horn County, 17
    3                         3                                  3
ug/m ; Custer  County,  15  ug/m  ; and Powder River County, 14 ug/m .  Such an analysis does not
exist for coastal areas in the  far West, where densely populated areas cluster thickly and the
city-to-city transport component is large.
5.3.3.4  PeakTSP Concentrations—To  indicate  the severity of TSP ambient exposures, the 90th
percentile  concentration  of the 24-hour  measurements  was examined for all  4008  sites  in  the
                                            5-33

-------
     50
     40
     30
 <
 oc
 UJ
 Z   20
 O
 O
     10
LOCAL


TRANSPORTED PRIMARY


TRANSPORTED SECONDARY


CONTINENTAL


GLOBAL
                            MIDWEST
                                            EAST
Figure 5-1S. Ave*fiiQe estimated contributions to nonurban
levels in the East, Midwest, and West are most variable for
transported secondary and continental sources.

Source:  Lynn et al'. (1976).
                         5-34

-------
1977 NADB.  The  concentrations  of TSP and other air pollutants had been widely reported to be
log  normally  distributed  (Larsen,  1971).  This  statistical  relationship,  however,  appeared
inappropriate at the  high  and low ends of the distribution (Mage and Ott, 1978).   Because the
extreme values at  the high end are subject to  wide scatter,  the 95th  or 99th percentile was
found to  be  less representative of the  severity  of high TSP levels.  The 90th percentile was
therefore chosen as  being  a more stable indicator, the TSP value which is exceeded 36 days of
the year.
     Figure 5-16 shows  the number of AQCR's where  at  least one of  the  monitors  had its 90th
percentile TSP concentration within the various categories.  Of the country's 247 AQCR's, only
                                                                           q
21  reported  90th percentile  values  for  all  data collected below 100 |jg/m  .   There  were 142
                                                                                     2
AQCR's that had  90th-percentile values at one or more sites,  between 100 to 200 ug/m .   These
data  suggest  that most  of the  U.S.  population  might  experience ambient TSP concentrations
exceeding 100 (jg/m  for at least 36 days of the year.
5.3.4  Temporal Patterns in TSP Concentrations
5.3.4.1   Diurnal Patterns--TSP  concentrations  vary with local emission strength, meteorologi-
cal  conditions,  and  changes  in  contributions  from  background   particles.   The  particle
emissions loadings to the  atmosphere generally increase during the day and decrease at night.
The atmosphere undergoes greater vertical mixing during the day, and windspeeds near the sur-
face  increase  as a  result.   Greater  vertical  mixing coupled with  increased source emissions
cause  particle mass  loadings  to increase. '   At  night,  decreased  mixing  and  the resultant
decreased surface winds  permit  settling of larger particles.   With increased atmospheric sta-
bility, local elevated sources are not as likely to mix to the ground.  Unfortunately, diurnal
cycles are not  well  established because  the  standard  sampling  procedure for TSP measurements
yields a 24-hour sample, midnight to' midnight.
     Trijonis et al.  (1980) found no clear diurnal trend in sub-15 urn particle mass in 6-hour
samples from  the St. Louis Regional  Air Pollution Study (RAPS).  Stevens et  al.  (1980) have
found slightly higher daytime levels of sub-15 urn particle mass in a remote site in the Smoky
Mountains; however,  Pierson et  al.  (1980) noted  no significant diurnal pattern in a forested
region in Pennsylvania.
     It is likely  that day-night patterns are  somewhat  obscured by averaging times.  Heisler
et al. (1980) found peaks  in light scattering and in particle mass corresponding to rush hours
in Denver in the winter of 1978; minimum values were found in mid afternoon corresponding with
mixing height maxima.
5.3.4.2   Weekly  Patterns—Since  human activity follows distinct weekly  cycles,  it is reason-
able  that anthropogenic  sources of particles  will  also  have weekly  patterns.   The most
distinct  weekly  patterns  are weekdays  versus  weekends.   Trijonis et al.  (1980)  examined the
St.  Louis TSP  and  dichotomous data base for weekend-weekday differences in particle loadings.
They concluded that there  was only a  slight (-9 percent) difference between weekend TSP values
and  weekday  values  for  the average  of five  urban  sites in St.  Louis.  For  three suburban
sites, the difference was  -5 percent and  for  two rural  sites the difference was -12 percent.

                                            5-35

-------
150
140
130
120
£ ii°
o
< 100
It
0 90
cc
LU
1 BO
z 70
60
SO
40
30
20
10
0






21






142







32 48
            <100
101-200
201-260
           90TH PERCENT1LE TSP CONCENTRATIONS, jug/m3

Figure 5-16. Severity of TSP peak exposures is shown on the
basis of the 90th percentile concentration. Four ACQR's did
not report.

Source: Environmental Protection Agency (1978).
                           5-36

-------
The  urban  difference  was dominated by  readings  from one monitor in  a  heavily industrial  and
commercial  area.
5.3.4.3  Seasonal Patterns—Analyzing  temporal  patterns  can frequently  provide  insight into
the  nature and  source of PM.   Meteorological parameters  affect the generation and dispersion
of particles.  These parameters include (among others) degree-days, mixing height, ventilation
factors, frequency  of  calms   and  stagnations,   and precipitation.   There are  also seasonal
patterns in some source emissions.
     Because meteorological parameters  are so important, it  is  likely  that seasonal patterns
in one  area  cannot  be generalized to other areas.  Trijonis et al.  (1980) found a modest sea-
sonal pattern of  higher TSP concentrations in the  summer months in St. Louis.  In support of
this observation, Figure  5-17  compares the TSP monthly mean values and the data from dichoto-
mous samplers.   The really distinct seasonal pattern was in the fine aerosol fraction.  Summer
fine-particle concentrations are twice as great as winter values.  As discussed later, aerosol
sulfates make  up most  of the  fine-fraction  particles and  show  a  distinct seasonal pattern.
     To  illustrate  the geographic  specificity  of  these  seasonal cycles,  3  years of monthly
averaged TSP data  are presented  in  Figure 5-18.   The data  came  from  Steubenville,  OH,  an
industrialized  site in  the  upper Ohio River Valley.  Each monthly mean was derived from 20 or
more  sampling   days.   The  TSP concentrations were considerably  higher  than the  St.-  Louis
values.  The months with the   highest TSP  in  Steubenville were March, April, and May in 1977;
July,  August,   September,  and  November in  1978;  and  February and  June in  1979.   No clear
seasonal pattern emerges from  this 3-year period.
5.3.4.4  Yearly Trends—In  1957,  a National Air  Sampling Network (NASN) began to operate rou-
tinely  on  a  national  basis.   The U.S.  Public   Health  Service,  with cooperation  from "State
health  departments,   operated  231  urban  and 37 nonurban  stations.   Some of  these stations
operated every  other  year,  so in a given year there were 143 urban and 37 nonurban TSP hi-vol
monitoring sites  in operation.  These sites collected one 24-hour sample every other week for
a  total  of 26  samples per year.  In 1977, over 4000 stations, most of them in State and local
networks,   reported TSP values to the  National  Aerometric  Data Bank (NADB)  of  the U.S. EPA.
Not  only has the number of sites greatly increased, but the sampling frequency has been 1 day
in  6 since 1971.   In some cities,  TSP monitoring data has  been  recorded  for  more than 20
years.  Although  the  sites  may not be  in  exactly the same  locations for every city, general
trends  in TSP concentrations can be obtained.  Figure 5-19 plots the annual geometric mean TSP
concentrations  for  three  groups of cities.  In 1958, the five cities classified as industrial
had  annual  mean TSP concentrations between 140 and  170 (jg/m  .  By 1974 the annual mean concen-
trations had dropped  to between 80 and 110 jjg/m  .  Similarly, three of four cities classified
as moderately  industrialized  showed substantial   decreases.   Only the Denver station recorded
an  increase,  and  that  is  only  for  a single year.   The  four cities classified  as lightly
industrialized  showed less overall change.
                                            5-37

-------
                                                            TSP (HI VOL)


                                                            •••	DURBAN
                                                                      SUBURBAN
                                                                      RURAL
                                                          DICHOT COARSE
                                                      N      (2515pm)
                                                                      URBAN
                                                                    4J SUBURBAN

                                                                      RURAL
JAN   FEB  ,,MAR U-»APR,,  MAY   JUN   JUL   AUG   SEPT   OCT   MOV   DEC
                                MONTH
   Figure 5-17. Seasonal variations in urban, suburban, and rural areas for four
   size ranges of particles. The data were obtained from a relatively small num-
   ber of monitoring sites.

   Source: After Trijonis et al. (1980).
                                    5-38

-------
 en
 a
 o
 §
 cc
 o
 o
 o
 a.
 Ui
 >
 o
    200
150
100
     SO
         I  I  I  I   I  I   I  I  I   I  I  I  I  I   I  I  i   I  I  I  I  I  [  I  IT  I  I  I  I  I  I   I  I  I  I
         I  I  I  i  i  I  I  i   I  I  I   I  I  f  I  i   i  I  I  I  i  i  i   i  i  I  i  i  i  i   I  i  i   i  i  i
          J  F M A  M  J  JASONDJFMAMJ  JASONDJ  FMAMJ  J  A S O N D
                    1977                         1978                        1979
                                           YEAR AND MONTH

Figure 5-18. Monthly mean TSP concentrations are shown for the Northern Ohio Valley Air .Monitoring
Headquarters, Steubenville, OH. No clear seasonal pattern is apparent.
Source:  Spengler (1980).
                                              5-39

-------
to
      220

      200

      180

      160

      140

  M  120

      100

       80

       60

       40

       20
        0
                                                  I    T
                                                                       O BALTIMORE
                                                                       D CINCINNATI
                                                                       A CLEVELAND
                                                                       O PHILADELPHIA
                                                                       || ST LOUIS
             HEAVILY INDUSTRIALIZED CITIES

            I   I   I     I    I    I   I
                                                  I	I
I    I    I   I
                j	I
            1957
                        1960
                                                                   1970
                                                                                    1974
.6
"o>
3,
                                                      CHATTANOOGA
                                                   A DENVER
                                                      PROVIDENCE
                                                   • SEATTLE
                                          I    I    I    I    I    I   I    I    I    I    I    I
           1957
                        1960
                                            19G5
                                                                  1970
                                                                                   1974
                                                 YEAR
160

140

120
100

 80

 60

 40

 20
  0
             I    I
                                1    I    I    I    I    I    I    I    I
            T   1    f
                                                               O  MIAMI
                                                               Q OKLAHOMA CITY
                                                               A SAN FRANCISCO
                                                               • WASHINGTON, D C
               LIGHTLY INDUSTRIALIZED CITIES
              I   t   I    I    I    I    I    I    j	I    1   I   J    I    I   I   I	L
            1967
                         1B60
                                              1U65
                                                YEAR
                                                                   1970
                                                                                    1974
             Figura 6-19. Annual geometric mean TSP trends are shown for selected NASN sites

             S
-------
     Examination of  the  expanded TSP data set from hi-vol samplers shows that for 2707 sites
                                                                o
the composite  median concentration  has  remained about  60 M9/m  between 1972  and 1977.   The
geometric mean over this period has decreased by approximately 8 percent.  The decrease in the
90th percentile of the  annual  average concentrations is most pronounced over this period (see
Figure 5-20).   Lowering  the TSP concentrations in  locations  with very high levels has been a
target of  State air pollution  control  strategies.   In addition, relocating  sources  to rural
regions,  building  new  sources with  taller  stacks,  converting  to  cleaner fuels, paving  of
streets  and  roads,  and  restricting  open burning have  decreased  the  number  of  locations
                                                   3
experiencing annual concentrations of over 100 ug/m •
     For the period  1970  to 1977, EPA reported  an  almost 50 percent reduction in particulate
emissions.   Most of this reduction occurred in the early 1970's as State air pollution control
programs started  many major emitters on  compliance  schedules.   The  rather  modest composite
overall reduction  of  8  percent in annual  TSP  levels  may be explained by the fact that direct
emissions from stationary sources contribute only a fraction of the TSP loadings in the atmos-
phere.
     Another perspective on regional differences is gained from observations of the 1978 data.
Table 5-6 provides a statistical summary for the 50th and the 90th percentiles for valid moni-
tors.   Region  IX  ranked  highest for the mean  and maximum 50th and 90th percentiles,  followed
by  Region  VII, Region  VI,  and Region  V.   Regions  I  and  II  had consistently  lower values.
     The column presenting  the standard deviations of  the mean values for the  50th  and 90th
percentiles is  also  of  interest.   Smaller standard deviations  suggest more uniformity in re-
ported concentrations among monitoring sites.   Because Regions I, II, and IV had less  variance
among  sites  than   other  regions,  it  could be interpreted that these  regions  had either more
uniform distribution of  pollution levels or more uniform  placement  of monitoring sites.   The
larger  standard deviations  in  other regions,  particularly in the  West,  probably mean  that
there is greater variation  in pollution levels.
     There are distinct regional differences in the trends of TSP concentrations.  The distri-
bution of site means and the actual rate of change in TSP levels differed among regions of the
country.   These trends  are  shown in Figures  5-21 and  5-22; the differences between years and
even over the  entire period were not tested for significance.  Therefore, intraregion and in-
terregion comparisons are presented qualitatively.
     In the Eastern United  States, in EPA Regions I and II, the composite average across sites
                        3                           3
decreased  from 60 ug/m  to approximately 55  ug/m .   The range of concentrations  was  much
narrower in  Regions  I  and  II than  it was  in the more  industrialized  Regions  III, IV, and V.
     In Region  III,  the  composite average decreased from  78  to 60 ug/m  ,  with  the 90th per-
centile  in  the distribution of annual  mean concentrations decreasing  from slightly  over 100
    33
ug/m   to  about 90 jjg/m .   However,  it  has  remained relatively  stable  or  has  even increased
                                                                                        3
slightly since  1975.   In Region V, the  composite average  decreased from 80 to  70 ug/m , and
                                                    3
the 90th  percentile decreased  from 100 to  85 ug/m ,  reflecting the  effectiveness  of point
source control.

                                            5-41

-------
160
140
M
"S 120
§ 10°
< 80
w 60
g 40
a.
{2 20
n
I I
9 9 (
i i
I I
*i ?
X X
6 0
i I •
I
1
6
i
I
?-
6 -
i
               1972     1973    1974    1975    1976    1977

                                 YEAR
                       I
                       O
                       A
                        A.
90TH PERCENTILE



75TH PERCENTILE


COMPOSITE AVERAGE

MEDIAN
                                   25TH PERCENTILE
10TH PERCENTILE
Figure 5-20. (Top) Nationwide trends in annual mean total suspended
particulate concentrations from 1972 to 1977 are shown for 2707
sampling sites. (Bottom) Conventions for box plots.

Source:  U.S. Environmental Protection Agency (1978).
                               5-42

-------
                               TABLE 5-6.   REGIONAL SUMMARIES  OF  TSP VALUES FROM VALID MONITORS
en
i

Number
of sites
128
315
300
534
781
294
136
152
89
113
15
9
10
2882
Median
Region
I
II
III
IV
V
VI
VII
VIII
IX
X
Alaska
Hawaii
Puerto.Rico
Total
Minimum
14.0
10.0
27.0
22.0
13.0
12.0
34.0
7.0
16.0
11.0
11.0
25.0
32.0
7.0
Mean
49.2
43.2
59.8
55.3
64.1
65.0
69.7
54.8
76.5
60.3
48.1
39.7
54,3
58.9
Maximum
100.0
114.0
171.0
137.0
189.0
166.0
154.0
164.0
226.0
129.0
94.0
70.0
85.0
226.0
soa
14.6
15.3
20.3
17.2
22.0
20.9
20.2
32.8
38.0
24.5
22.7
13.7
13.9
22.8
Minimum
32.0
29.0
52.0
41.0
26.0
37.0
58.0
18.0
37.0
23.0
35.0
40.0
66.0
18.0
90th percent! le
Mean
87.3
85,0
105.7
93.5
122.4
110.4
123.6
107.8
133.4
123.6
137.1
63.6
90.7
107.9
Maximum
181.0
286.0
296.0
256.0
383.0
436.0
359.0
412.0
381.1
381.1
250.0
99.0
134.0
436.0
SDa
27.7
30.6
42.1
30.9
42.8
45.8
44.2
64.0
66.0
52.5
68.7
18.9
18.9
44.9

    .SO, Standard deviation of the median and 90th percentile  values.

     Including American Samoa and Guam.

    Source:   SAROAD.

-------
                             U S EPA AIR QUALITY CON1ROl REGIONS, EASTERN STATES
   160
   148
   120
lioo
 S ID
Si 60
*" 4D
   20
    0
                 REGION 1
I    I    I     I    I     I
        1972 1973 1974 1975 1976 1977
                     160
                     140
                     120
                  "e 100
                                                    REGION 2
                                  I     I    I    I    I     I
                                          1372 1973 '1974 1975 1976  1977
                                                                                    REGION 3
                                                                    i    ii    i     I
                                                                           197Z 1973 1974 197S 1976 1977
                                  REGION a _
                                         t.    i    I
                                                                      REGION S
                                                             I     Ii    r    I    I
                          1S72  1973 1974  197S 1976 1i77
                                                            1972 1173 1974 1976 1978 1977
  Fifluro 5-21. Regional trends of annual mean total suspended participate concentrations, 1972-1977, Eastern states.

  Source: U.S. Environmental Protection Agency (1978).
                                                     5'44

-------
                        U S EPA AIR ttUALITV CONTROL REGIONS, WESTERN STATES
   160
   140
   120
1-100
 "I 80
 85'BO
 *-  40
    20
    0
          REGIONS
     I     I    1
        1972 1973  1974 1975 1976 1977
                                           REGION? -
                          I	
                                   I	I
                                                 I
                                                                           REGIONS -
                                                                    II     I    I
                                          1872 1973  1974  1975 1976 1977
                                                                           1972 1973  1974 1975 1978 1977
  160
  140
  120
I 100

2  60
P  40
   20
    0
                                            REGION 9
                           I    i     i     i    I
                                                                               REGION 10 -
                         1972 1973 1974 1975 197S 1977
                                                             1972 1973 1974 1975 1976 1977
                                                   YEAR
Figure 5-22.  Regional trends of annual mean total suspended paniculate concentrations, 1972-1977, Western states.

Source:  U.S. Environmental Protection Agency (1978).
                                                      5-45

-------
     The Western States make up Regions VI through X.  In Region VI, the composite average re-
                                 3                                                            '
mained  at  approximately 75  pg/m ,  and  the  90th percentile increased  slightly  from the 1973
                          2
reading to  about  100 ug/m .   Industrial, utility, and related growth in this area, as well as
in  Region  IV, was  probably  responsible  for  keeping TSP concentrations  from decreasing.   In
Region VII, the composite average was almost constant, varying only slightly between 80 and 75
    3                                                      3
pg/m .  The 90th percentile varied between 110 and 100 pg/m .  Region VIII showed wide distri-
                                                                      3
bution  in  the concentrations.   The 10th  percentile,  at about 20 jjg/m  , was  the lowest among
                                                          o
all regions.   The  90th percentile,  approximately 100 jjg/m  , was roughly equal to the highest
concentrations in  most regions.   The composite average  varied  over the 6-year record but re-
                                                     o
mained  essentially the same,  approximately  80 |jg/m  ,  in  1977  as it was  in  1972.   The back-
ground  air  quality in the upper  States  of this  region  (Montana, North and  South Dakota,  and
                                                                                  2
Wyoming) was among the best in the country.   Thus, some  of the low levels (20 pg/m  and below)
represented some of  the lowest background concentrations measured  in the United States.  The
high  composite average  and high 90th  percentile   levels  reflected the  impact  of locating
monitors near  industrial  sources such as smelters and  the  fugitive dust emissions from wind-
                                                             3                            3
blown soils.   Region IX had a composite  average  of  100 pg/m , which was  up from 90 pg/m  in
                                                                  3
the early 1970's.   The 90th percentile was also high, at 120 ug/m .  Thus, Region IX had some
of the  highest levels in the country.   Region X  had a  composite average  of approximately 70
    3                                    3
ug/rn ,  up slightly from a low of 60 |jg/m  in 1975.   The 90th percentile varied between 90 and
100 ug/m3.
     The overall  trend in  improvement from  1972 through  1975 was followed  by  a reversal in
some  regions   in  1976.   Despite  this  short-term reversal  in  1976,  60  percent of  the sites
showed  long-term improvement  from 1972 to 1977.  For  those  sites at which TSP concentrations
violated the current annual standard, 77 percent showed  long-term improvements.  Approximately
25 percent  of these sites reported their lowest  annual  values in 1977.  Possibly, the short-
term reversal  in  1976 was due to unusually  dry  weather, resulting in windblown dust that may
have contributed  to  elevated TSP levels throughout the Central  Plains,  Far West, Southwest,
and Southeast.
5.4  SIZE OF ATMOSPHERIC PARTICLES
5.4.1  Introduction
     In Chapter  2,  the general  features of  size distributions  of  atmospheric particles were
discussed in  some  detail.   In recapitulation, atmospheric particles tend to be more prevalent
in certain  particle  size  bands or modes  than  in  others.  Particles  that  have  grown from  the
gas phase,  either  because of condensation or atmospheric  transformation  or combustion, occur
initially as  very  fine nuclei  0.05 (jm or smaller.  These tend to grow rapidly to accumulation
mode particles around 0.5  pm  in size,  which are  relatively  stable in the  air.   Because of
their initially gaseous  origin,  this range of particle  sizes  includes inorganic ions such as
  O—    w.        -i.
S0£ , NO-, and NH., combustion-formed carbon, organic aerosols from photochemical conversions,
and a variety of trace elements associated with combustion sources.
                                            5-46

-------
     Airborne particles  of  soil  or dust mostly  result from entrainment by the  motion  of the
air or from  other  mechanical  action, and most of the mass of these materials is in particles
larger than  5 urn.   While the relative amounts of these two particle types are highly variable
in both time and  place,  there is  almost always  a clearly observable minimum or gap in atmos-
pheric mass  distribution  occurring in the particle size  range  of roughly 1 to 3 micrometers.
In this range,  there  are only minor  percentages  of  the total mass, and this material appears
to be  overlap  from the  two major categories.   The  larger particles  frequently contain clay
minerals,  bits of  local  rocks,  limestone aggregate from  roadways,  fly ash from power plants,
and other  substances  ranging  from insect parts,  pollen, and  sawdust to  liquid  globules of
acidic smut  blown from  boiler  tubes (Draftz  and Severin, 1980).   The  elemental  analysis of
these  larger particles  is  usually  dominated  by  silicon,  aluminum, magnesium,  calcium,  and.
iron, all  components of soil and of fly ash (see Chapter 4).
     In the  last  several years,  a general  perception has been growing  that  not all  PM is
equally damaging  to  the  environment  (see Chapters 8  & li).  For this  reason,  information on
the mass  of PM in  various size categories  has  been  gradually accumulating  and  is  here sum-
marized.   Furthermore,  a national  network  of sampling stations  equipped with size-selective
sampling devices is currently being set up.   While some tabulated data are available from this
network and  are summarized  here,  no  detailed interpretative  analysis  has been published yet,
nor are any  chemical  analytical  data available.    The analysis  of monitoring results from the
national  inhalable particle  (IP)  network must wait for subsequent revisions of this document.
     In the  following discussion,  the  major chemical components of atmospheric aerosols are
organized by the  size mode or particle  category in  which they  are  most frequently  observed.
                           2-
This is not  to  say that SCL  ion,  for  example,  is exclusively a component of fine particles,
Sometimes, e.g.  in  the  vicinity of a cement manufacturing facility, there can be substantial
                      Own
amounts of coarse CaSO, .  However, the relationship between size and composition of particles
is so general that more reason is served by this organization than by any other.
     Since the finer particles seem to have less diversity and since measurements of the major
anion  components  of this  fraction have been  made for a long time,  this  group is  discussed
first.   The  more  complicated  coarse fraction has  not been very well defined and, indeed, may
not be definable  by  chemical  analysis alone.   There is  considerable interest  in  this size
range currently, though,  and studies of these materials are cited in Section 5.6.
5.4.2  Size  Distribution of Particle Mass
     Evidence from  chemical analysis and  physical theory (Chapter  2)  strongly  suggests that
atmospheric  aerosols  commonly occur  in two  distinct  modes.  The fine  or  accumulation mode is
attributed to growth of particles from the gas phase and subsequent agglomeration.  The coarse
mode is made up of mechanically abraded or ground particles.   Therefore, it is not surprising
to find atmospheric  PM  distributed among fine and coarse particles with a rather clear inter-
val of demarcation in between.
                                            5-47

-------
     Unfortunately,  gravimetric  data by size fraction were sparse until comparatively recent-
ly.  Furthermore, most were obtained with impactors, which are influenced by particle "bounce"
(see Chapter 3).  Several works suggested the existence of a distinct minimum in the mass-size
distribution in  the 1968 to 1970  time  period.   Lee et al. (1968) observed only 14-percent of
the aerosol mass between 2 and 4 urn in three samples from Fairfax, OH.  Lundgren (1970) found
only 10-percent of aerosol mass  in  this  range in 10 Riverside,  CA,  aerosols samples ranging
                    o
from 47 to 144 vg/m ,  O'Donnell et al, (1970) found only 10-percent in the 2 to 4 pi range in
one  Pittsburgh,  PA,  sample.   Lee  and  Goranson  (1972)  and Lee  et  al. (197-2)  reported many
impactor size  distributions  for six cities obtained in 1970, all indicating 12- to 15-percent
of  aerosol  mass  between  2 to 5 pm.  However, many of these data were clouded  by bounce and
entry losses and  probably were biased toward low coarse-mode distributions.
     More  recently, evidence from  electrostatic sizing equipment has  confirmed this general
trend.   Figures  5-23 through  5-26 show  the  distribution of particle  volume by size.   These
data differ from mass distributions because particle density (mass/volume) was not measured as
a function of  size.  Figures 5-23 and 5-24 present distributions in and around St. Louis, MO,
for a  variety of conditions.   Generally  these distributions show distinct  minimum values in
the vicinity of 1 to 2 pm.
     However,  the combined influence  of  nearby  sources  and aerosol aging  can  produce major
shifts in  volume and,  presumably, mass distribution.  For example, Figure 5-24 shows a third,
very fine "nuclei"  mode of particles centering around 0.05 urn.  This mode can be attributed to
the presence  of  nearby  automotive traffic.   Also  shown  in Figure 5-24 is  the  rather narrow
size distribution from  a coal-fired power plant  adding  to the fine aerosol burden.  However,
the mass-size distribution from large sources can vary dramatically among sources depending on
the type and efficiency of control equipment (see Chapter 4).
     There can be major shifts in the relative proportions of fine and coarse particle mass as
an aerosol ages  (i.e.,  moves with the  wind).  Figures  5-25 and 5-26 show dramatic examples of
this  phenomenon  obtained during  the  1972  California  Aerosol  Characterization  Experiment
(ACHEX).   In the first case, aged aerosol  was transported in the wind to the site from the Los
Angeles area;  during this  process coarse particles  settled out.  In  the  second  case, local
winds stirred up dust, shifting the distribution toward larger sizes (Whitby, 1980).
     A summary of  mass  data  calculated  from electrostatic  size distributions  for several
environments is shown in Table 5-7.  Here, the dramatic variations in coarse and fine particle
fractions found in  practice are clear.
     More recently,  a  number of studies  have  been  done  with dichotomous samplers designed to
obtain mass samples of the 0 to 2.5 |jm fine fraction and a 2.5 to 15 urn coarse fraction.  Re-
searchers from  EPA's Environmental  Science Research Laboratory have  measured coarse and fine
aerosol mass concentrations  in several locations:  Dzubay  et al.  (1977) report  on  18 days of
                                            5-48

-------
      60
      SO
      40
 I
      30
      20
      10
BACKGROUND AEROSOLS NEAR ST. LOUIS, MO.
——  URBAN PLUME INFLUENCED
___ BACKGROUND AVERAGE
_._ AUTO INFLUENCED
__.._ CLEAN BACKGROUND
Figure 5-23.  Linear-log plot of Die volume distributions for the four background distributions.
Notice how much the urban plume adds to the accumulation mode of the background.
Source:  Cantrell and Whitby (1978).
                                           5-49

-------
       70
       50
    S»  30
       20
        10
URBAN AEROSOLS

	" LABADIE POWER PLANT

          PLUMB

___ URBAN

__ . __ URBAN AUTO

          INFLUENCED
        0

        0.003
   .01
0.1
                                                       1.0
10
                                                                       50
Figure 5-24,   Linear-log plot of the volume distributions for two urban aerosols and a typical
distribution measured in the Labadie coal-fired  power plant plume near St. Louis. Size distri-
butions measured above a  few hundred meters  above the ground generally have a rather small
coarse particle mode.

Source: Cantrell and Whitby (1978).
                                                   5-50

-------
 o
M~
 O

^

<3
                           0.1
             1.0


PARTICLE DIAMETER, i
                                                                       10
 Figure 5-25.  Incursion of aged smog from Los Angeles at the Goldstone tracking station in the
 Mojave Desert in California. Note the buiidup in the accumulation mode.

 Source: Whitby (1980).

-------
"
     10 —
                                                   1.0
                                       PARTICLE DIAMETER, jrni
10
 Figure 5-26,  Sudden growth of the coarse particle mode due to local dust sources measured at the
 Hunter-Liggett Military Reservation in California. This shove the independence of the accumulation
 and coarse particle mode.

 Source; Whitby (1980).
                                            5-52

-------
                    TABLE 5-7.   FINE AND COARSE AEROSOL CONCENTRATIONS FROM
                         SOME URBAN MEASUREMENTS COMPARED TO CLEAN AREAS

Concentration (pg/m )a
Location
St. Louis
Los Angeles
Los Angeles
freeway
Denver
Go Ids tone
Mil ford, Mich.
Pt. Arguello
(seaside)
Condition
Very polluted
Grand average
Wind from
freeway
Grand average
Clean
Very clean
Marine air
Fine
particles
296.0
37.0
77.0
16.6
1.5
1.03
1.1
Coarse
particles
94.0
30.0
59.0
23.2
3.0
0.82
53.0

                 Calculated from volume distribution using assumed particle
                 density, pp = 1 gm/cm3.
                 Source;   National Research Council (1979).

summer sampling in  St.  Louis;  Stevens et al.  (1979)  report on 2 months of summer sampling in
Houston^ TX; Stevens  et  al.  (1980) discuss results of an extensive sampling for a week in the
Great Smoky Mountains.   Courtney et al. (1980) discuss the early results from winter sampling
at two locations in Denver, CO.  Table 5-8 summarizes their reported findings.
     In another short-term  study,  Lewis and Macias (1980) sampled atmospheric aerosols for 21
                                                                   3
days  in  Charleston, WV.   The  fine-fraction average  was 33.4 pg/m , and  the coarse fraction
                     3
average was 27.1 pg/m .
     Because of the influence  of particle size on adverse effects such as health, visibility,
and soiling (see  later  effects chapters), EPA  is  establishing a network of size-selective PM
monitors.  Ultimately  this grid  will  include  250  stations to  be established over  a 3-year
period.   During the period from April 1, 1979, to June 30, 1980, 94 stations were established.
A map showing current sampler locations is shown in Figure 5-27 (U.S.  Environmental Protection
Agency, 1981).   Since dichotomous samplers are used in this network, together with hi-vols, it
is possible to  obtain a general conception of the relationship between TSP (0   •*• 60 (j<")> di-
chotomous total or  "inhalable" particle mass (0-15 pm), and the fine and coarse fractions de-
fined above.
     A total of 1960 dichotomous fine- and coarse-mass measurements and 2675 TSP measurements
are now  in  this data base; hi-vol measurements with a size-selective inlet are now also being
                                                                 3
made.  In this  data base, daily TSP  values  range from 33.2 |jg/m   in  Litchfield, CT to 474.4
                                            5-53'

-------
                 2*—.
(
                                                                               BOSTON 82^

                                                                    BU FF ALOB 4 HARTFOB0 • 2 i
ARTFOR0*2^»'


|RKCITY«4*
                ,           J 	  . ™^_ .   ]
                \.*NFOS OUSTER ft i         (
                  I        I     MINNEAPOLIS* 2




     l^^fSK'-l         i        -f      ^^^^i^^'ssi^
     /  «IT LAKE CITY* 2	•	.    I       DESMOINES«2 %'n't'm"J    (AKHON"2 »3P|TTS  PpHILADELPHI/
     /  SALT LAKE CITY w*  .   	J__               ,     \ STEU^ENVILLE l( 1 jy«SH"iBALTIMORE 2

2^!»NFRANSI5CO   /       .'         I        *i	\     '     WASHINGTON, D C f 2 ' *
Z»S»NI=BANtISCO   /		  l	i.     ^  ciNCINNATlkl >   „"  ~

                                             »z  y x   X  ^
                                       | ST LOUIST.   \.^.~*     - -
                                           I 	^     	1       • «* _  	
                                                                                       10
                                                   NFOS » NATIONAL FOREST OZONE STUDY


                                                   'NPS = NATIONAL PARK SERVICE


                                                  SMALL NUMBERS REPRESENT NUMBER OF SAMPLING SITES AT EACH LOCATION
                Figure 5-27. Inhalable-particle network sites established as of March 19,1980.

                Source: U.S. Environmental Protection Agency (1981).
                                                      5-54

-------
               TABLE 5-8.   FINE FRACTION AND COARSE FRACTION DICHOTOMOUS SAMPLING
                 BY ENVIRONMENTAL SCIENCE RESEARCH LAB, USEPA IN FOUR LOCATIONS
         Location    Period   Days   Comments
      Concentration
Fi-ne (ug/m3)  Coarse ((jg/m3)
St. Louis

Houston
Denver





Smoky Mtns

Summer

Summer
Winter



Winter
Spring
Fall

18

28
19



19
28
7

Urban
Rural
Urban
Urban Site D
Urban Site N
Urban Site D
Urban Site N
Urban
Urban
Remote Day
Remote Night
29
26
52.2
18.1
25.4
23.2
26.4
26.5
16.1
26.4
22.0
22
15
39.8
22.5
23.4
33.0
26.5
27.1
9.8
6.2
4.9

         Source:  Courtney et al. (1980)

             3
to 474.4 (jg/m   in Dallas, TX.  Maximum dichotomous sampler totals (fine + coarse) ranged from
28.7 ug/m   in  Pearl  City, HI,  to  267.5 ug/m  in Rubidoux, CA  (U.S.  Environmental Protection
Agency, 1981).
     Because of  the  limited  time period available  for  analysis (April 1979 to June 1980), it
would be unwise  to consider analysis of these  data as indicative of geographical or seasonal
trends  in  particle size.  But  some additional  general factors  associated  with particle size
can be  seen  from inspection of  the  data  summary in Table 5-9.   (The  ratios in the table are
averages of  ratios  of individual sample pairs  and  thus will  not equal  ratios  of the average
concentrations given.)
     On the average the dichotomous sampler total mass was about 67 percent  of the TSP (Pace,
1980),  but this  ratio varied widely across  the country,  from about 0.4  to almost 1 in Port-
land, OR and Litchfield, CT (five samples).  However, most Portland and all  Litchfield samples
were collected  during the  winter months  when  rainfall or  snow cover  could  have materially
reduced dust levels.
     The  fraction of  fine  and coarse components  was even  more variable.  The  coarse mass
fraction  of the total  sub-15  urn  mass ranged  from  about  one-fifth  to  two-thirds  in this
selected set and was even higher for individual days.  Particularly striking were the average
values  for  Dallas and El Paso, TX.  At both sites, the sub-15 urn mass was  only about half the
TSP mass.   However,  in Dallas  only  27  percent  of this was in the 2.5 to 15 urn range while in
El Paso, 64 percent was "coarse."
                                            5-55

-------
                   TABLE 5-9.  RECENT DICHOTOMOUS SAMPLER AND TSP DATA
                         FROM SELECTED SITES—ARITHMETIC AVERAGES
Location
No. of TSP DTOTAL/TSP Coafse
3 3
observations pg/m (# pairs) MS/m
Fine Coarse,
Mg/m3 DTOTAL

Northeast
Buffalo, NY
Erie Co. , NY
Litchfield, CN
Philadelphia, PA
Southeast
Birmingham, AL
Midwest
Minneapolis, MN
Cincinnati, OH
Southwest
Dallas, TX
El Paso, TX
Far West
Los Angeles, CA
Portland, OR
Pearl City, HA
TSP
DTotal
TSP
DTotal
TSP
DTotal
TSP
DTotal
TSP
DTotal
TSP
DTotal
TSP
DTotal
TSP
DTotal
TSP
DTotal
TSP
DTotal
TSP
DTotal
TSP
DTotal
28
40
41
44
5
5
102
109
38
40
44
41
51
48
22
24
29
26
43
50
37
36
27
25
93.7 0.70 (21) 25.2
32.8 0.64 (25) 5.1
18.9 0.86 (1) 6.6
45.1 0.83 (40) 13.3
60.8 0.68 (23) 15.0
50.1 0.61 (26) 15.6
53.6 0.77 (26) 14.4
94.9 0.47 (21) 9.8
86.5 0.51 (7) 46.3
68.4 0.53 (18) 21.3
66.7 0.90 (19) 42.3
33.0 0.43 (11) 7.9
25.9 0.50
16.2 0.24
13.3 0.33
22. 5 0,38
24.4 0,38
16.4 0.46
25.2 0.35
24.1 0.27
11.7 0.64
24.6 0.47
22.0 0.60
8.4 0.47

Source:  U.S. Environmental Protection Agency (1981).
                                         5-56

-------
     Pace et al,  (1981)  analyzed some of the general features of these preliminary IP network
data.  These  authors concluded  that regional scale  fine particle  mass  ranged from  6  to 13
    3                                    "           3
      in the Western States and from 15 to 23 (J§/m  in  the  eastern United States, comparing
only  four  sites  in each category.  They  also  found strong influence of  local  sources around
                                                                    o
monitors in urban  sites.   "Concentration changes averaging 16  ug/m  or 31 percent were found
in  sites  separated  by as  little  as  1  km distance,"  according  to  this study.   No strong
seasonal trends were found in various regions of the United States, although there appeared to
be  slightly  higher  fine-particle mass  in  the summer  than  in spring  in the  eastern United
States.
     It appears that  chemical  analysis and several years  more  data from the IP network could
considerably increase  understanding of temporal  and spatial  distribution of fine- and coarse-
particle fractions.
5.5  FINE PARTICLES IN AIR
     Sulfate,   ammonium ions,  organics,  carbon, and  combustion-associated metals  are widely
recognized to  be   the  major  components  of  fine  PM.  Few  studies  of  aerosol  composition have
attempted material balance, and fewer still have done so with size fractionation.
     Nevertheless, a  great deal  has been learned about the chemical and elemental composition
of  airborne particles since  the  early experiments  in the 1950' s by Junge  in Germany, Massa-
chusetts, Hawaii,  and various sites in Florida (Junge 1952).   Junge's observation that sulfate
and  ammonium  ions appear  predominantly  in the  fine-particle  fraction has  been confirmed in
independent field  observations,  both  in urban and rural areas  (Lewis and Macias, 1980; Dzubay
and  Stevens,   1975).   In  analyzing the  St.   Louis,  MO,  dichotomous  sampler  data by  x~ray
fluorescence,   Dzubay  found 75  percent of  the  zinc,  sulfur, .bromine,  arsenic,  selenium,  and
lead occurred  in the fine particles and at  least 75 percent of, the silicon, calcium, titanium,
and iron in the coarse fraction (Dzubay, 1980).
     In  studies  of  Charleston,  WV,   particles,  Lewis  and  Macias  (1980)  reported material
balances of fine  and coarse  particles accounting for 69 percent and 60  percent of the mass,
respectively.   Eighty-five percent of the sulfate and ammonium  ions were in the fine particles
where  they accounted for  30  and  12,8 percent  of  the mass,  respectively.    Carbon,  both
elemental and organic, was mainly in the fine aerosol  (61 percent) where it accounted for 18.2
percent of the mass.
     Stevens et al.  (1978)  reviewed size-fraction  analyses for St.  Louis, MO and Charleston,
WV  and for  four   other  sites  including New  York, NY;  Portland,  OR;  Philadelphia,  PA;  and
Glendora, CA.   They  conclude  that sulfate  ion  is predominantly a fine component (70 percent)
                                                                     *
that  usually accounts  for 40 percent  of  the mass  of that fraction and  occasionally up to 50
                2-
percent.  The  S04  must  be present as ammonium salts or as I^SO* since metallic sulfate could
be only 10 to  30 percent of the total at maximum (Stevens et al., 1978).
     In  one  site  in  the Great  Smoky Mountains,   89  percent, of  the  fine PM  was  identified
(Stevens et al., 1980).  Sulfate accounted  for 61 percent; ammonium ion, 12 percent; elemental
carbon, 5  percent; and organic carbon, 10  percent.   Trace elements,  mainly lead, made up the

                                            5-57

-------
balance.   In  this study, only organic  carbon  was also a significant  component  of the coarse
particles.
     Studies  in  a number of  sites  in California produced similar  results.   Flocchini  et al.
(1978)  and Cahill et  al,  (1977) reported size-fraction  distributions for sulfur (presumably
  o«
SO, ) in  three  districts of California.  In all  areas,  sulfur was present almost exclusively
in the  sub-3.6  pm fraction.   In dry weather, sulfur was found in sub-0.65 (jm fractions, while
under humid conditions it appeared in the 0.65 to 3.6 pm cut.
              2-     +
     Since  SO-  ,  NhL,  elemental  carbon,  and  organics  are  the  major  components  of  the fine
aerosol,  analytical  data  relating   thereto,  whether  size  fractionated or  total,  will  be
discussed together in this section.
5.5.1  Sulfates
     The  term "atmospheric  sulfates"  describes a several sulfur compounds, including ammonium
sulfate,  NH.HSO.,  H2S04,  letovicite,  calcium sulfate, and a variety of metal salts.   Most of
the historic  data on  atmospheric concentrations of  sulfates are  based  on  the water-soluble
extract of TSP filters  and  measurements of the sulfate ion.  These  samples  were subject to
artifact  formation on  the glass fiber filters  used in  the early  NASN  measurements.   For a
complete  discussion  of these  issues, see Chapter  3.   It is now generally  accepted  that TSP
  2-
S0a  measurements taken before 1974 or 1975 using the traditional glass fiber filters may have
                                           3
overestimated sulfates by as much as 2 pg/m  or more in areas where ambient S0« concentrations
were high.
                           Ow.                                              O
     Annual average  TSP SOA   concentrations  range from  less than 1  pg/m   in  some  States to
               3
almost  20 (jg/m   in urban  industrial  areas  of  the Northeast.  For  24-hour  average concen-
            2_                                                         3
trations,  S0|  concentrations range from near zero to more than 80 pg/m .
     Sulfate,  particularly ammonium sulfate,  appears to account for the majority of fine PM in
many  locations  (Dzubay,  1980;  Stevens  et al.,  1980; Watson,  1979;   Flocchini  et al., 1978;
Stevens et  al.,  1978;  Pierson et al.,  1980).  Although  some of this  material  may be emitted
directly  from sources, the majority  appears  to  be secondary  (i.e.,  formed by chemical re-
actions in the atmosphere) (Friedlander, 1973;  Grosjean and Friedlander, 1975).
                                                                                   2-
5.5.1.1   Spatial  and Temporal  Variations—The  spatial distribution of measured  SO,   concen-
trations  for  1974 is  displayed  in Figure  5-28.   Figure 5-28(a) presents the  annual average
                                                                           3
concentrations.    An  area having  an  annual  average  of more  than  15  pg/m  extended  from the
lower Ohio  Valley through the upper  Ohio Valley, including  major  portions  of Kentucky, West
                                                                                             3
Virginia,  Ohio, and  western  Pennsylvania.   The areas with annual  averages exceeding 10 pg/m
included  almost  all  of the  United  States  east  of the  Mississippi, except   for the South
Atlantic States and upper New England.
                                                               3
     Through  the  Central Midwest area,  values of 4  to  9 pg/m  were  reported.   The  Far West
                                                      2-                         3
States and the Pacific Northwest experienced annual S0|  levels below 2 to 3 ug/m  , except for
the Los Angeles area. The Los Angeles levels are not shown in this figure, but a 1975 National
Academy of sciences  report on  air  quality  and stationary source  emission controls  indicated
                                    o
that they were between 7 and 13 pg/m  (National Academy of Sciences, 1975).

                                            5-58

-------
Figure 5-28. Contour maps of sulfate concentrations for 1974 are
shown for: (a) annual average; (b) winter average; (c) summer
average.

Source: National Research Council  (1978a).
                               5-59

-------
                               O—
     Seasonal variations  in  SO,   concentrations are shown  in  Figures  5-28(b) and 5-28(c) for
                                                                                2-
the winter months  and the summer months  respectively.   The area of elevated SO,  greatly ex-
pands  during the  summer  months.   As demonstrated by several  regional  studies  on atmospheric
  2-               2~
SQ*   transport,  SO.   concentrations can  be elevated over large geographical  regions  under
certain  meteorological  conditions  (Eliassen,  1978;  Lyons  and et  al.,  1978;  Perhac,  1978;
Whelpdale, 1978).   This  supports  the  idea  of transport  and  conversion  beyond  the  source
regions of S09  emissions.   As these contour maps clearly show, a sizeable portion of the U.S.
                                   2-                                    3
population is exposed to  annual  SO.  concentrations of  more than 10 u/m  in the ambient air.
In view  of the  rising S00 emissions  from increased use of coal throughout the United States,
                                                                   2-
particularly  in  the  South  Central  States,  the area of maximum S0|   levels  might expand and
shift to the lower Ohio Valley and the Southeast.
                                              2-
     In a large-scale study of atmospheric  SOt   in eastern Canada,  Whelpdale (1978) reported
                       3
mean levels  of  10  pg/m  over southern Ontario.   The mean levels of sulfates dropped to less
than 2.5 (jg/m   above  the  49th parallel.   Figure  5-29  displays these values for the period of
study.   During episodic conditions  primarily affecting the  lower Great Lakes region, 24-hour
                                                       3
concentrations were reported  as  high as 40  to  50 ug/m .   Such episodic conditions are assoc-
iated with a high-pressure  cell  over eastern Canada with southwest flow occurring on the back
side of the  high pressure.   This synoptic situation favors transport of SQ? and sulfates from
the high SO- source regions of the industrialized northeastern United States.
     Recently, new information on  the  interrelationship  of SOp, N0p»  0,,  TSP,  sulfates, and
nitrates has become available from a large-scale regional study.  The Electric Power Research
Institute (EPRI) Sulfate Regional Experiment (SURE) involves intensive monitoring from some 54
rural stations and an aircraft sampling program.  The area being studied is 2400 by 1840 kilo-
meters;  it  extends from  Kansas  to  the  Atlantic coast  and from mid-Alabama  to southeastern
Canada (Hidy et al., 1979) (see figure 5-30).
     Mueller et  al. (1979)  reported on the  earlier SURE  data collected in 1974  and 1975 and
presented the  preliminary results of an  intensive  field study made during July  1977 through
February 1978.   Using the limited historical data base, they indicated that the rural stations
                                                             2~
experienced  a  frequency of occurrence  of 24-hour average SO,   concentration similar to that
observed around  large metropolitan  areas such  as New  York City.  As seen in Figure 5-31, 24-
                                 3
hour values  greater than  10 ug/m  occurred  in approximately half the data, and the occurrence
             2~                         3
of 24-hour SO.  levels exceeding 20 ug/m  was about 10 to 12 percent.
                                                3
     Based on  concentrations  of 10  to  20  yg/m   as an  indicator  of elevated  exposure,  the
average concentrations over  the  entire SURE network area  were estimated  by a linear interpo-
lation procedure with a  resolution of 80  by 80-km  grids.   Episodes of elevated sulfates were
                                                                        2-
extensive; during  an  episode  in early August  1977, the  area where  SO,   levels exceeded 20
    3                                  2
ug/m  expanded to  more than 500,000 km.  Two regional episodes occurred in January and early
                                                  ?-•                        3
February 1977.   In August,  39 percent of  the  SO,   values exceeded 10 yg/m  ;  in January the
                                                                    3
figure was 30 percent.   Five percent of the values exceeded 20 jjg/m .   In October, 20 percent
                                3                                                           3
of the  values exceeded 10  pg/m ,  and  less than  1 percent of  the  values  exceeded 20 ug/m .

                                             5-60

-------
            5 If I MOUNT FOREST f »

                  \
            MICHIGAN
Figure 5-29. Intensive Suifate Study area in Eastern Canada shows the geometric mean
of the concentration of soluble particulate sulfate during the study period. Units are
micrograms of sulfate per cubic meter.

Source: Whelpdale (1978).
                                        5-61

-------
Figure 5-30.  Map of SURE region shows locations of ground
measurement stations.

Source: Hidy et ai. (1379).
                            5-62

-------
    100
     50
E

Z
UJ
u
3
(a
     20
10
        0.01
         O HIVERHEAD, NY (champs!


         A BRONX, NY (champs)


         O ROCKPORT. (sure I)



         • SCRANTON ( sure I)
                                                                              I      I
                                                                                              O
                                                                                             4
                                                                          RANGE OF OCCURRENCE

                                                                          FOR SURE I AND NYC

                                                                          CHAMP STATIONS
                                    10
                                       30
eo
80   90  95
                                                                         99
99.8
                                                                                             99,9
  Figure 5-31. Cumulative plots show the frequency of sulfate concentrations in the SURE region on

  the basis of the 1974-75 historical data.


  Source: Mueller et al. (1979).
                                             5-63

-------
Figure 5-32  shows the estimated number of days exceeding 10 ug/m  for August 1977 and January
                                                                                       2~
to February  1978.   In August, almost the entire Northeast had at least 10 days with SO.  con-
                                 3                                                           3
centrations  greater  than  10 |jg/m .   The area  having  20 or more days with  more than 10 ug/m
involved  Ohio,  West  Virginia,  Maryland,  Pennsylvania,  and New York.   By contrast,  in  the
                                                 O™.
winter months the area of prolonged elevated  SO.   concentrations  shifted toward the West and
Southeast.   The  upper Ohio Valley remained high, and an increase  in the number  of days with
more than 10 pg/m  also occurred over Tennessee, Alabama, and Georgia,
     Studies of seasonal variations have reported elevated concentrations in the summer months
                                                                                    P—
(Hitchcock,  1976;  Hidy et al., 1978).   The summer monthly mean concentrations of SO,,  in some
                                                                                    Z""
regions  can  be twice those for  the  winter months.   The seasonal   variation  in  SO.  concen-
trations in Southeastern and Midwestern cities is less distinct than the variation in New York
                                                                         2—
City or  Los  Angeles  (see Figures 5-33 and 5-34).  Elevated summertime SO.  concentrations are
generally  reported to be  the  result  of increased homogeneous and  heterogeneous  oxidation of
anthropogenically  produced  S09.   However,  oxidation of biologically produced hydrogen sulfide
                                                         2-
has  been  offered as  an  explanation  for  some  high  SO.  concentrations  in  isolated  areas
(Hitchcock,  1976,  1977;  Hitchcock,  Spiller, and Wilson,  1980).  (See also Chapter 2 relative
to H,S oxidation.)
     Lavery  et  al.  (1979)  postulated  the  existence of  two meteorological  conditions that
                                                     2-                              3
result  in regional  accumulation of  particulate SO*  concentrations above  20 ug/m   in  the
northeastern United States:
              The  first  regime  consists  of cases  where widespread stagnation
              occurs  with  a large high pressure area slowly moving eastward over
              the  midwestern and eastern  United States.  Zones of polluted air
              collect over areas within 100-300 kilometers of high sulfur dioxide
              emissions sources.   These zones maintain themselves over periods of
              one  to  four days in warm, moist air,  with light winds, around the
              southern and  western  parts  of the high  pressure area.   The second
              regime  appears  to  be  conducive to long-range (greater than 500 km)
              sulfate transport and involves a channeling of air flow between the
              west side of the Appalachian Mountains and weak cold fronts approxi-
              mately  oriented  west-southwest  to  east-northeast  and  traveling
              south-eastward.  The  channeling  appears to be combined with capped
              vertical  mixing  associated  with  subsidence   around   the  frontal
              system.  These episodes can last up to four days.
                                                                                            2-
5.5.1.2  Urban Variations—The preceding discussion of spatial and temporal variations of SO.
was derived for the most part from widely spaced rural monitoring stations.   It is of interest
                                                                                            O™.
to  note  spatial  variations  on  the much  smaller  scale of  a  metropolitan  area.   The SO,
measured on  this  scale may consist of  a  natural  background component, a long-distance trans-
ported component,  a  component formed locally  in the  atmosphere,  and/or an artifact formed on
                                                 2-
the filter.  Hidy et al. (1978) compared urban SO.  distributions from the previously reported
works of  Lynn  et al. (1975) for  the New  York City area, and Kurosaka (1976) did the same for
the Los  Angeles  area.  These areas differ  in  meteorology and climate, but the population and
total S02 emissions are similar.
                                            5-64

-------
 B
                                             250
Figure 5-32. Maps show the spatial distribution of number of days
per month that the sulfate concentration equaled or exceeded
10 jug/m3. (A) January-February 1978 (31 days); (B) August 1977
(31 days).

Source:  Mueller et at. (1979).
                               5-65

-------
    200


    190


    180


    170


    160


    150


uj   140
D
I
Z

<
u.
o
130


120


110


100


90


 80


 70


 60


50


40


30


20


10


 0
           1    I    I    I    I   I     T
                                         I    I    I    I
                        •SOZ AMBIENT LEVEL
            I    I    I
                            I   I     I    I    I    I    I    I
                            567



                               MONTH
                                            10  11  12
   Figure 5-33. 1977 seasonal patterns of SO2 emissions and 24-hr
   average SO2 and 804 ambient levels in the New York area are
   normalized to the annual average values.


   Source: Lynn et al. (1975).
                               5-66

-------
 300
JAN   MAR
MAY
                            JUL
                                    SEP
                     NOV
Figure 5-34. Monthly variation in monthly mean of 24-hr average
sulfate concentration at downtown Los Angeles is compared with
monthly mean 1973 Los Angeles County power plant SO2
emissions.

Source:.Hidy et al. (1978).
                            5-67

-------
     The population  density of New York City  is  greater than that of  Los  Angeles (see Table
5-10).  In  addition,  the emission patterns are dissimilar.  As seen in Figure 5-35, there was
                                 2-
a  significant difference  in  SO,   concentrations across  the New  York urban area,  with the
highest values observed  in a strip from Staten  Island northeast into Brooklyn.   This finding
                                                                                    2~
may  have  been  biased by  the wintertime  emissions;  in  summer,  fairly uniform  SO.   concen-
trations  have been  found  in the  New York  metropolitan area.  The  highest density  of SCL
emissions were   in  eastern  New  Jersey,  Staten   Island,  Brooklyn,  and Manhattan.   Within a
                                                                             2-
distance of 10 to 50 km from the sources of highest S0? concentration, the SO*  concentrations
decreased by 30 to 40 percent from their maximum values.
     As shown in Figure  5-36,  the mean  annual   average  concentrations derived  from 24-hour
values  in  Los Angeles showed  a  relatively uniform distribution across the  Los  Angeles basin
area.   A  weak maximum was found near Burbank and  another in the San Bernardino area.   The
areas  of  major S02  emissions were El  Segundo,   Long  Beach,  and Fontana.   One  similarity to
New  York was  found  in the Los Angeles area; at distances exceeding 50 km from highest concen-
                     2-
tration areas, the SO.  levels dropped off significantly.
     Spengler and Dockery  (1979) measured sulfates in  particles  less  than 3.5 urn in diameter
using  a network  of 10 to  12  sites  in  each of six  cities  for periods of  up to  2  years.
Analysis of variance  showed no  significant variation among sites within the cities of Topeka,
KS;  Portage,  WI;  Kingston, TN; and Watertown, MA.   Some slight variations occurred among the
sites  in St.  Louis,  and  significant variations occurred  among the  sites in Steubenville, OH.
Only the Carondolet area of southeast St.  Louis was monitored, not the entire city.  There are
a coke plant and a lead pigment plant nearby, which cause large SO, gradients and perhaps also
  ?-•
SO.   gradients.   In Steubenville,  the TSP and  SO,, values near the  river  were  approximately
twice  the concentrations 5 km to the west of the river.  For sulfates in this size range, the
pattern was similar  and  the gradient  was  not as  pronounced, but the  differences  among sites
were significant.
                                                         2~
     An attempt was made to'explain the variability in SO,  data for both the Los Angeles area
and  the New York City area by means  of stepwise  linear regression.  Table  5-11  displays the
three  principal  independent  variables  and the r values associated with them in explaining the
                         2-
variance of the  daily SQ4  concentrations.   The results were very consistent  in  both areas
except for Vista, CA, a community about 100 miles southeast of Los Angeles.
     The results  indicated that the most  important  variables were the 24-hour 0,,  level, the
                                                                      2-
midday  RH,  and the  total  particulate mass concentration, minus the SO.  and nitrate fraction.
Hidy et al. (1978)  also  suggested that these  three  factors were important in determining the
                      2-
daily variations of SO,  concentrations.  The 0,,  or oxidant levels are an indication of photo-
chemical oxidation, the  RH is an indication of water  content of the air mass,  and TSP is an
indication of reactions involving PM.
                                            5-68

-------
              TABLE 5-10.   SOME CHARACTERISTICS OF POLLUTION IN THE
                         NEW YORK AND LOS ANGELES AREAS

Parameter
Surface area considered, km
Population estimate (1970)
2
Population density, no. /km
S02 emissions, tons/yr
2
SOp emission density, kg/km /yr
Maximum temperature, °CC
Minimum temperature, °CC
Relative humidity, %e
Normal precipitation, cm
Mean wind speed, m/sec
Mixing height, m^
2
Ventilation, m /sec
so2, ug/m3 h
Water-soluble sulfate (S042"), ug/m3 h
N0£, ug/m3 h
Water-soluble nitrate (NO, ), ug/m
3 h
Qg, Mg/m
Total PM concentration (TSP) 3
less SO* and nitrate, TSPM, |jg/m
Los Angeles
21,000
9,000,000
430
238,000
10,300
22.8 (5.5)d
10.8 (4.6)
50.2 (17.0)
36
3.3 (1.4)
849 (472)
2690 (2160^)
12.5 (19.9,)
10.1 (7.9)
83.9 (44.3)
9.1 (7.7)
52 (34)
64.5 (27.4)
New York
17,000
12,000,000
710
266,000
14,200
15.0 (7.4)
9.3 (8.4)
59.6 (16.5)
106
5.8 (2.3)
1290 (906)
7460 (6200)
42.9 (45.0)
8.9 (5.7)
67.6 (36.0)
2.6 (2.1)
20 (22)
40.4 (19.9)
aGreater metropolitan areas; Los Angeles, South Coast Air Basin; New York, tri-
.state metropolitan area.
 Based on EPA Air Quality Control Regions.
 •Annual mean of daily maximum or minimum hourly temperature.
 Numbers in parentheses are standard deviations.
-rAnnual mean of daily minimum humidity.
 Annual mean of noon wind speed at surface.
^Defined by annual mean of daily midday radiosonde sounding.
 Annual mean of 24-hr averaged values, 1974-75; Los Angeles, seven stations,
 New York, four stations.

Source:  Hidy et al. (1978).
                                   5-69

-------
    NEW JERSEY
      JERSEY
       CITY
   13
                ATLANTIC OCEAN
                                          km
Figure 5-35. Map shows annual mean 24-hr average
sulfate levels in micrograms per cubic meter in the
New York area, based on 1972 data from Lynn et al.
(1975). Squares are locations of three CHAMP site
stations. The fourth station is at the tip of Long Island
about 160 km from Manhattan.

Source: Hidy et al. (1978).
                         5-70

-------
                               LOS ANGELES CO.
                                                      zusa
                                                  A
                                                West Covina
                                                  "1*1 «
                                                  •12.5
                                                  **-/
West Los Anelas
        'MB

        1
         Santa Monica
                                               .foil     \

                                                    Anaheim
                                                                             RIVERSIDE CO.
              .CIRCLED NUMBERS; STATION DATA
              A/ CHAMP STATIONS
                  UNCERTAIN
                                       DISCREPANCIES L
                  BETWEEN AGENCY ANALYTICAL METHODS
Figure 5-38. Distribution of annual average sulfate concentration in micrograms per cubic meter in the
greater Los Angeles area is based on 1972-74 data.


Source: Kurosaka (1976).
                                                5-71

-------
        TABLE 5-11.  PRIMARY RANKING OF VARIABLES FOR CORRELATING AIRBORNE

              IN TWO CITIES BASED ON A STEPWISE LINEAR REGRESSION OF 15
                 VARIABLES FROM CHAMP AND RELATED MONITORING STATIONS
A.  Los Angeles area
                               Garderr  West
Santa   Thousand
                     Anaheim   Grove   Covina  Glendora  Monica    Oaks    Vista
Variable
1
2
3
Correlation
coefficient (R)
B. New York
Variable
1
2
3
Correlation
coefficent (R)

°3 °3
TSPM TSPM
RH RH
0.71 0.77
Brooklyn Queens

TSPM TSPM
RH RH
°3 °3
0.60 0.63

03 TSPM 03 TSPM Tm1n
TSPM RH RH RH QX
RH 03 TSPM 03 RH
0.79 0.79 0.79 0.72 0.56
Bronx Riverhead, L.I.

TSPM TSPM
°3 RH
RH 03
0.54 0.62
 Located 50 km north of San Diego and 16 km inland from the coast,
RH = Relative humidity.
0- = 1-hr daily maximum ozone value.
TSPM - Mean TSP.
T -  = Minimum temperature.

Source:  Hidy et al. (1978).
                                     5-72

-------
     The  local  SOp  concentrations did not enter  into the correlation sequence as  one of the
three principal variables.   The  findings of Spengler  et  al.  (1979)  are not inconsistent with
these results,  since  the only city with a significant spatial  variation among sites for sul-
fates also had a variation among sites for the respirable particles and TSP,
5.5.2  Nitrates
     Nitrate  aerosols  make up a varying amount  of the TSP.  Although widely  reported to be
                               2-
significantly less  than  the  SO,   fraction,  nitrates  nevertheless  represent an important con-
stituent.  Most nitrates in the atmosphere are formed in gas-to-aerosol reactions, principally
involving nitrogen dioxide and nitric oxide.   These reactions may yield HNCU (gas or aerosols),
ammonium  nitrate, sodium nitrate,  and lesser amounts of other compounds.   A minor fraction of
the NO-  aerosols  measured in the atmosphere can be attributed to wind erosion of soil and re-
suspension of fertilizers  (National  Research Council, 1978b).  These  sources  may be more im-
portant  locally near  fertilizer  plants,  transfer facilities, or  munition  factories (National
Academy of Sciences, 1977).
     Measurement of participate  NO,,  has  proved especially difficult  (see  Chapter 3).  It has
been clearly shown that gaseous HNQ- is absorbed by glass-fiber filters (Pierson et al., 1980a;
Spicer et al.,  1978;  Appel  et  al.,  1979)  and,  consequently,  NO., values  measured, with such
filters might be erroneously high when interpreted as reflecting particle composition.  On the
other hand,  ammonium  nitrate has significant vapor pressure and could be Tost from some media
by evaporation  or by  acid attack (Pierson et  al.,  1980a).   There exists considerable-contro-
versy over the  interpretation of available analytical data for NO,, but the weight of current
opinion  appea'rs to  favor the idea that  glass-fiber filter  N03 may be the  sum of particulate
nitrate  and  gaseous HNOg.   This  issue is not  yet settled,  nor could any sort of consensus be
said to  exist.   Still, the available  data do  indicate those regions with  elevated levels of
some NO,  species, either gaseous or particulate, previously measured as TSP mass.
       *5
     Mean  NO- aerosol  concentrations  from  urban  and nonurban NASN  sites are  summarized in
             3                                                                       ™,_~™,
Figures  5-37 and 5-38,  respectively.  The  annual average concentrations  shown  are in micro-
grams per cubic meter, as measured from hi-vol samples.  Concentrations in urban air were sub-
stantially higher than those in nonurban air.  A zone of high urban concentrations exceeding 4
    o
ug/m  extended eastward from Chicago through the industrialized Northeast through Pennsylvania
to the Philadelphia area.  Other high NO- zones were in southern Louisiana, around Birmingham,
                                                                                  ,3
AL, and  near Little Rock, AK.  In general, a zone of high urban nitrate concentrations 3 jjg/m
and  higher extended from  southeastern Texas  through the Midwest and  into  the Northeast.  Of
course,  a major emission  source may  cause  high NO- gradients in  the  surrounding  area.   For
example,  a  study  in Chattanooga, TN, (Helms et al., 1970; National Academy of Sciences, 1977)
showed an average  NO-  concentration of 48.9  M9/m   'for  a  site  close  to  the  Volunteer Army
Ammunition  Plant.   This  is  more than three times the NASN maximum  station  average for 1965
(13.5  ug/m ).   This  station  average was 15 to  20 times higher than those of the  four other
                                            5-73

-------
Figure 5-37. Map shows U.S. mean annual ambient nitrate levels in micrograms per cubic meter.



Source: Akland (1977).
                                           5-74

-------
Figure 5-38.  Mean nitrate concentrations in micrograms per cubic meter at nonurban sites in the
U.S., based on valid annual averages from 1971 through 1974.

Source: U.S. Environmental Protection Agency (1977).

-------
Chattanooga sites  presumably  not influenced directly by  the  munitions  plant.   Their averages
ranged  from 2.4  to  3.8  ug/m .   While the  artifact phenomenon  may discredit  the absolute
values, the ratios among sites have more credibility.
     It  is  obvious from  these figures  that  the data  base is quite incomplete  for the West
Coast.  No  data were  reported for the  Los  Angeles  area,  nor for the large metropolitan areas
of San Francisco,  Seattle, and Portland.
     A few  studies  have sought information on  NCL  concentrations  by composition and particle
size.  Orel and Seinfeld  (1977) compared the  formation,  sizes,  and concentrations of ambient
  2-
SO.  and NO,  particles.   Unlike HLSO., the HMO, that is formed tends to remain in the gaseous
phase, although it may be an important component of acid precipitation.   The EPRI SURE Project
                                _                                                     O-
(Kneip et al., 1979) reported NO., ion concentrations one-tenth the concentration of SO^  ions.
                                                                            3
The monthly mean  values for August and October 1977 were less than 0.6 (jg/m  ammonium nitrate
at three  locations across  the Northeast.   On  a few occasions the  daily  levels  exceeded 1.5
ug/m .
                      2-
     Data from the  SO.   and N03 data  base  of the California Aerosol Characterization Experi-
ment  are  reported by Appel  et al.  (1978).   Summer measurements for 1972  and  1973  from five
fixed and one  mobile  site indicate that the mass median diameter for nitrates was between 0.3
and 1.6 urn.   Twenty-four-hour averaged concentrations of NO, ion varied across the Los Angeles
                           3                                        3
basin, from a low of 4 ug/m  in Dominguez Hills to a high of 31 ug/m  in the eastern community
                                2-
of Rubidoux.   In  contrast to SO.  the  diurnal  pattern  for NO, often had a maximum during the
morning  close  to  the maximum  for  gas-phase nitrogen oxides.   The  authors  concluded that the
ratios of  ionic  constituents  and  ambient  NHQ  levels suggested that ammonium  salts were the
                    2-       -
principal form of SO.  and NO,.
     Until   recently such  high NO,  levels  were not  suspected in  other  parts  of  the  United
States.  However,  the EPA early analyses from a sampling program in Denver, CO, found 24-hour
NO,  levels,  primarily  in  the  fine  fraction,  that  often exceeded  10 ug/m   (Courtney et al.,
1980).
     Japanese workers  have been  investigating atmospheric nitrates  for  some  time.   Kadawaki
(1977) found  a bimodal  distribution of nitrates in the  Nagoya  area of Japan.   Submicrometer
particles (0.4 to  0.6 urn in diameter)  were ammonium nitrate; the coarse particles (3 to 5 urn
in diameter) were  sodium  nitrate.   Background  nonurban  levels  as  low as 0.8 to  0.9 ug/m'  on
the outer islands of Japan have been reported (Kito, 1977).  Maximum average concentrations in
the city of Kawasaki were reported to be as high as nearly 7 ug/m  (Terabe, 1977).
     In summary,  our  knowledge of nitrates in  the  atmosphere is rather limited.   No compre -
hensive  data  set exists.   The NASN measures  NO,  ion every 12 days  at  relatively few sites;
spatial  and short-term  temporal  variations  cannot  be  discerned.   In  fact,  there  are many
cities for  which   no  measured  values  of nitrates  have been  reported.   Furthermore, historic
data before 1977 are in doubt because of the artifact formation on the filters.
                                            5-76

-------
There  are  spatial  patterns  In  NO,  concentrations.   Cities  tend to  have higher  levels  of
nitrates than do  rural  regions.   Some studies indicate that localized areas may have substan-
tially higher NO,  levels.   This  raises the concern  that  available data on NO, concentrations
may  underestimate  the  actual population  exposure.    In  the  near future,  new  sampling  and
analysis techniques  should expand  our knowledge  of NO, aerosols,  HNO,,  and  other nitrogen
compounds.
5.5.3  Carbon and Organics
     A variety  of carbon-containing  compounds  often account for a large  but  highly variable
portion of fine  PM.   Elemental  carbon, which is emitted from a variety of combustion sources,
is a significant component also,  usually accounting  for 10  to  20 percent of urban aerosol
mass.
     Organic substances  of biological  origin occur in particles.  Evidence from microscopic
examination of particles,  cited  later in this chapter, has demonstrated the presence of wood,
paper, insect  parts, pollen,  bits of leaf,  and textile fibers in  coarse  particle fractions.
For  the most part,  these materials are made up of cellulose and protein, biologically derived
substances that  are insoluble in  common • organic  solvents and that decompose  to form carbon
residues under  thermal  treatment.    Analysis  techniques capable  of distinguishing  this  bio-
logically  fixed  form of  carbon  from  combustion-formed  soot are only  now becoming available
(Huntzicker et al., 1982; Novakov,  1982; Wolff et al., 1982).
     Organics in particles  usually  have been determined by  extraction  in organic solvents or
by thermal evolution of organic  vapors, procedures  that  concentrate  relatively low molecular
weight  organic  components.   Consequently,  there  is  little chemical  analytical  information
relative to the  organic high-polymer composition of particles, either  those  of biological  or
manmade origin.   In  the following  discussion, the reader  is cautioned  that most studies have
concentrated on  these  simpler  organic  molecules and do  not entirely  account  for organic
material in particles.
     Particles emitted from combustion sources frequently have fuel-derived organic substances
sorbed  on their  sources  (see  Chapter 4),  and  such materials  are  commonly  found  in  the
atmosphere.   Combustion  processes   often  alter  fuel  molecules  considerably   and combustion
product mixtures often  contain  substances not in the fuel.  Some examples include products of
coal   and  wood pyrolysis,  oxidized  or  nitrated hydrocarbons  in  motor  vehicle  exhausts,  and
synthesis of polynuclear aromatic hydrocarbons in rich flames.  Chapter 4 covers many of these
processes.
     Photochemical reactions  are also capable of generating substantial quantities of organic
particles, and high concentrations of solvent-extractable PM are found associated with high 0^
levels.  It is  now believed that atmospheric oxidation of some volatile organic species leads
to formation  of  bifunctional molecules,  especially dicarboxylic  acids,  of  very  much lower
vapor pressure than their precursors.  These reactions are discussed in considerable detail  in
Chapters 2 and 6.
                                            5-77

-------
     In addition to the organic species accounting for aerosol mass, there are also present in
the particles  very much  smaller amounts of polynuclear aromatic  hydrocarbons,  components of
special  interest because  several  are  known  to be  carcinogenic.   One  of  these compounds,
benzo(a)pyrene  was  conventionally  measured in  NASN  TSP samples.   A complete  discussion of
these  substances is  contained  in  the  EPA Health Assessment Document for Polycycljc Organic
Hatter (Santodonato et al., 1979) to which the reader is referred.
     Several  comprehensive reviews  of  airborne-organic PM  have appeared  recently  (National
Academy  of  Sciences,  1972;  Duce, 1978; Daisey,  1980;  Hahn,  1980;  Lamb et  al.,  1980).   The
subject  has also been discussed in other  reviews  (National  Academy of Sciences, 1976; Perera
and Ahmed, 1978; Grosjean, 1977).
5.5.3.1   Physical Properties  of Particulate Organics—Many  atmospheric organic  compounds  are
distributed between  the vapor  and  particulate  phases  of the  aerosol  (De Wiest  and Rondia,
1976;  Krstulovic et  al.,  1977;  Cautreels  and Van Cauwenberghe, 1978),  and,  presumably,  this
distribution can  vary with temperature.  Because of this volatility, there can be substantial
losses  of low  molecular weight  compounds during sampling  (Cautreels and Van  Cauwenberghe,
1978;  Krstulovic et  al.,  1977; De Wiest  and  Rondia,  1976;  Katz and Chan, 1980;  Schwartz et
al., 1981).  At the  high temperatures  found in  combustion  sources, larger proportions of the
emitted  organic  compounds  will  be present  in  the  vapor phase.  These compounds will condense
on  the surface  of  PM as the emissions  cool  and,  thus, be enriched at  the surface.   Natusch
(1976)  found  that this  occurs  when  PAH (polynuclear  aromatic  hydrocarbons)  is emitted  from
powerplant stacks.   Such surface enrichment  can  affect  the biological  impact  of polycyclic
organic  matter  (POM).   While there  is  the  possibility  that  POM may exist as particles formed
by self-condensation, most POM is probably adsorbed on the surface of other particles, much of
it  presumably  associated  with  soot particles  (Thomas et  al.,  1968).   The  effect of  the
substrate upon  which POM  is adsorbed  upon the chemical and biological reactivity  of these
compounds is almost  entirely unknown.  Korfmacher and coworkers (1980) recently reported that
photodegradation  of  some  PAH  compounds proceeds much more  slowly when  the  compounds  are
adsorbed on coal fly ash than when adsorbed on other substrates such as silica gel.
     The  distribution  of organic  particles  between  vapor  and particulate  phases  is  also
profoundly  influenced by  chemical   reactions  in  the  atmosphere.   Grosjean  and  Friedlander
(1975) found that during photochemical oxidant incidents organic substances are converted from
volatile to relatively  nonvolatile  species.  In this process, the fraction of organic mass in
the particle phase  (relative to the  gas)  can  grow from very  low  values,  1 percent or so, to
about  6  percent of the vapor and particle total.
     While both mass and size  distribution of organic substances  in particles  is clouded by
their  volatility, there have been some attempts to establish the fine/coarse ratios.  Some of
the  heavier polycyclic  components   are known  to be  predominantly fine-particle components
(Mueller et al., 1964; De Maio and Corn, 1966; Kertesz-Saringer et al., 1971; Pierce and Katz,
                                            5-78

-------
1975; De  Wiest,  1978;  Van  Vaeck and  Van  Cauwenberghe,  1978, 1980).   In  Los  Angeles oxidant
incidents, virtually all the organic particles were found to be smaller than 2.5 urn (Schuetzle
et al.,  1975;  Mueller  et al., Hidy  et al.,  1975).   Van Vaeck  and Van  Cauwenberghe  (1978)
reported that aliphatic  hydrocarbons  and carboxylic acids were predominantly  (90 percent) in
fine  particles   in  European  samples.   Since  organic  compounds are  generally  distributed
disproportionately in the fine fraction aerosols, it is not surprising that they represent an
important  fraction  of  the  mass.  Steigerwald  (1975)  estimated  that organic  substances  are
between 25 and 47  percent of the fine particle fraction in the United States.   However, there
are also  reports of  significant fractions of  organic  substances  in  coarse particles in rural
samples (Stevens et al., 1982).  No consistent study of organic species or classes by particle
size exists in the literature.   Therefore,  it is currently impossible to trace the origins and
fate of organic  particle components or area or timewise distributions on the basis of ambient
air measurements.
5.5.3.2  Carbon and Total Organic Mass — There are limited historical  data on the mass fraction
of elemental carbon  in  atmospheric  aerosols, but very recent work is contributing information
in this area  (Rosen  et  al., 1982; Novakov, 1982; Wolff et al., 1982; Huntzicker et al., 1982;
Stevens et  al.,  1982;  Lewis  and Macias,  1980; Stevens et al.,  1980).   Techniques currently
employed  detect both organics  and carbon  by  optical  absorption  and selective  combustion
techniques.
     Novakov (1982) found elemental  and organic carbon  in over 1000 samples collected from a
variety  of  urban  sites.   In  New  York  City,  the principal  chemical  species present  was
elemental   carbon,  accounting  for  two- thirds  or  more of  the carbon  mass; the balance  was
organic.    In  Denver,  about 60 percent was  elemental  carbon, while in Los Angeles  about 70
percent was organic and the balance, elemental carbon.
     Wolff et al. (1982), using a somewhat different technique, reported carbon concentrations
                                                                                        a
in 10 U.S.  locations.   "Apparent" elemental carbon was  reported  to  range from 1.1 ug/m  in a
remote South Dakota  location  to 13.3 M9/m  in New York City.   These  values covered a range of
4 to 11 percent of the TSP.
     Stevens et  al.  (1980,  1982)  and Lewis  and Macias (1980) reported total carbon values of
8.4 ug/m   (14 percent of total dichotomous sample mass) in Charleston, WV, while in the Great
                                '                   3             3
Smoky Mountains and Shenandoah Valley only 1.2 MS/m  and 1-5 ug/m  » respectively (4 percent of
                                                                                    o
total  dichotomous  sample  mass  in both  cases),  was  elemental  carbon and  3.4 ug/m  and  0.9
ug/m , respectively  (12  percent and  3 percent of  total  dichotomous sample mass), was organic
carbon.
     The  mass  of  organic  substances  present  in  atmospheric  aerosols was   at  one  time
approximated  by solvent-extraction  with  benzene  in  routine  NASN  hi-vol  samples.   Other
solvents have  also  been  used in such determinations.   Unfortunately,  such determinations were
terminated  in  1970  and,  except for  a few  intensive  studies  (Daisey,  1980;   Grosjean  and
Friedlander, 1975; Wesolowski et al.,  1980),  there has been no extensive data base on organic
                                            5-79

-------
extractables covering the past 10 years.
     Some typical values for particle number and mass concentrations of organic substances are
listed  in Table  5-12.   The organic  fraction of  the mass concentration  as measured  by the
                         *
benzene-soluble  component   is  also  listed,  with the benzo(a)pyrene  fraction  for comparison.
In the  organic  fraction,  a variety of  organic  compounds  have been identified, including some
materials classified  as PAH (Corn,  1968).   However, the  identified  fraction  represents only
10%  of the  organic  components  of  the urban  aerosol.   Although the  total  aerosol  number
concentration is  often very large in  cities,  the mass concentration varies  less  and rarely
exceeds about 200  ug/m  in the United  States.   The benzene-soluble fraction of this is about
10-20 percent of the  total  mass, and the concentration of benzo(a)pyrene is far lower.  Even
in remote areas, there is a contribution of organic material.
     A  limited  number  of  samples have  been collected in unpolluted atmospheres.   Levels  in
remote  areas  and in marine air  for  the ether-soluble  fraction of organic particulate matter
have been as  low as 0.51 (0.18-0.84) ug/m   STP.  Marine air with  a continental  influence had
averages  of  0.93  (0.48-1.38)  ug/m3  and  continental  air 1.2  (0.69-1.71) ug/m3.   Similar
concentrations have been observed at Barrow, AK, a remote' site in the Arctic, for cyclohexane-
and dichloromethane-soluble POM (Daisey et al., 1981).
     Variations in the concentration of organic particulate matter by location, meteorological
conditions,  season and  by time  of  day have  been observed repeatedly  (Hidy et  al.,  3975;
Gordon,  1976;  Calvert,  1976).   By  way of illustration, Figure 5-39  shows the  differing
contributions of the organic fraction in samples obtained in two cities in southern California
(National  Academy  of Sciences,  1972).  In both  instances,  however,  the organic  fraction
represents a sizeable portion of total suspended particulate material.
     A  pattern  of elevated wintertime  concentrations of organic  particulate  matter has been
observed  in  New  York  City  and  in  Mainz,  West  Germany.   Winter  samples   in  Mainz  of
ether-extractable  organic  material  averaged 27  ug/m  for TSP concentrations averaging 150
    o
{jg/m .  Winter  samples collected in February,  1977 in  New York City  had  a total extractable
                           3                             3
organic fraction of 22 ug/nr for a TSP average of 96 ug/m .  The August,  1976 levels were 13.3
pg/ra3 for a TSP average of 86 ug/m3 (Kneip et al., 1979).
     A  1971  study of  Colucci  and Begeman  is  an example  of a more detailed short-term urban
survey of PAH than is available from NASN data.  From 1964 to 1965, Colucci and Begeman (1971)
found that the concentrations of benzo(a)pyrene and benz(a)anthracene were 4 1/2 times greater
in central Los Angeles than at two suburban sites.  However, the suburban site downwind of the
downtown  area  (on  the  average)  appeared  to  have  systematically  higher  benzo(a)pyrene
concentrations  than  the upwind  site.   Daily concentrations reported  in the Los  Angeles area
                       3                    ^
ranged  from  0.1  ng/ra   to  over  10  ng/m ,  depending  on  the  season.   Benz(a)anthracene
concentrations were 1 1/2 times larger than the benzo(a)pyrene concentrations.  Annual average
benzo(a)pyrene concentrations were similar to the NASN data for downtown Los Angeles.  The PAH
The  benzene-soluble extract  is  not necessarily  equivalent to the  tota.l amount  of organic
material in the sample, but it is taken to be representative of such a fraction.
                                            5-80

-------
      TABLE  5-12.   TYPICAL VALUES  OF  AEROSOL CONCENTRATION  FOR DIFFERENT
                     GEOGRAPHIC  AREAS (ANNUAL AVERAGES)3

Number of ,
Location particles/cm^
Nonurban
Continental
General
California
Oregon
Colorado
Indiana
Maine
New York
So. Carolina
Maritime
General
Pacific offshore
Oahu, Hawaii
Urban
Continental
General
Los Angeles
Portland
Denver
Minneapolis
Chattanooga
New York «
Greenville, SQ
Maritime
Honolulu, Hawaii
San Juan, Puerto
Rico


103-
103 -
3
HT -
-
-
-
-

iof-
10; -
io3-

f\
io3-
103 -
*\
103-
103 -
-
-
-

103-

"


wl
IO4
A
IO4




fl
10}
IO4
IO4


10}
IO4
f)
103
IO3




IO4


Mass
concen-
tration,
MS/m^


20 - 80
39
47
14
39
18
29
40

-
19 - 146
10 - 49°


MOO
93
72
110
70
105
105
76

40

77
Benzene
soluble
fraction,
|jg/ms


1.1 - 2.2
2.8
0.9
1.1
2.1
1.2
1.8
2.7

— A
1.5 - e.rj
0.7 - 6.3°


7
12.5
6.6
9.0
6.1
6.9
3.9
7.4

2.3

6.9
BaP
fraction,
ng/m^


-
0.48
0.09
0.11
0.25
0.12
0.25
0.43

-
-
-


-
1.87
2.60
2.52
1.18
4.18
3.63
7.49

0.59

1.42

 Data based on 1969 NASN observations,  except for
 maritime data.
 Aitken nuclei
"Geometric means.
-i
 Short-term data.
                                     5-81

-------
I
    J  I   I  1  I -1  1  I  I  1   }  I  I   I  I  I   I  I   II  I   I  I  I  I  I   I  I  I
   NH/'(3%1
       S04!M4W


        NO  (i
                     (WATER (16%)
                                                      ORGANICS (43%)
    Mill  IN  M  I  I  I   I  I  I   I  II  I  I  I  I  IN  I  I  I
             10
                   20
                                    30
                                                40
                                                            50
                                                                        60
 PASADENA, 9/20/72; TWO HOUR SAMPLE OVER 1200-1400 PST; LOW OXIDANT, TOTAL
 MASS CONCENTRATION, 79 /ng/m3.
      I  MITT IT!  I  I  I  M  i   M M

               NH4+  (10%)

               WATER (10%)

                  S04" (1354!

                              i ORGANICS (24%)

                                    «• C26%!
                                  INOjf (26%!
                                   I  I  M  I  I  I  M  I  I  I  I  I   I  I.
            10
                        20
                               30
                                           40
                                                       50
60
 POMONA, 10/24/72; SAMPLES FROM 1200-1400 PST; MODERATE OXIDANT, TOTAL MASS
 CONCENTRATION, 178 pg/m3.

Figure 5-39. Calculated distribution of aerosol constituents for two aerosol samples
taken in the Los Angeles basin.

Source: National Academy of Sciences (1972).
                                      5-82

-------
concentrations  increased  substantially in winter.  Benzo(a)pyrene  concentrations  were higher
at night,  in  contrast with those of other pollutants.   All pollutants were higher on weekdays
than  on weekends.   Benzo(a)pyrene concentrations  were  found to  be correlated  with carbon
monoxide and  lead  concentrations,  with coefficients ranging  from  0.6 to 0.9.  Benzo(a)pyrene
concentrations  were  also  significantly  related  to those of  hydrocarbon vapors,  oxides  of
nitrogen, and vanadium  (a nonautomotive pollutant).  Despite the strong relation to lead, the
statistics  in  the  study  failed  to  reveal  a  clear  identification of  BaP emissions  with
automotive or stationary combustion sources.
     Trends of  BaP concentrations  as  measured at 34 NASN urban sites are displayed in Figure
5-40.   It is indeed encouraging  to  see the steady  decline  in  BaP concentrations  that has
occurred  since  the  mid-19601 s.    The  90th  percent! le  of   quarterly   measurements  fell
                               3                     3
dramatically  from  near 7  ng/m  to less  than 2  ng/m  .   These changes  reflect  both emission
controls and shifting of sources. Incomplete combustion of fossil fuels, especially coal, is a
primary source  of  BaP.   Major point and area sources include residential coal-fired furnaces,
coal-fired  utilities   and   industrial   boilers,   coke   ovens,   petroleum  refineries,  and
incinerators  (see  Chapter 4).   Shifts  away  from coal  for residential,  commercial,  and light
industrial  use  have   made   a  substantial   contribution  to  the  reduction  of  urban  BaP
concentrations.   To  a  lesser extent,  the control of particulate  emissions has also helped to
lower concentrations.
     The  national  trends  in  benzene-soluble  particulate  matter and BaP as  reported fay Faoro
(1975)  may not  be  true  everywhere.    Indeed,  specific  organic  fractions may  show opposite
trends.   Daisey (1980) discussed  benzene soluble  organic trends  for  New York City; annual
averages  for  the  New  York University  station,  normalized  to  account  for  year-to-year
meteorological variations, are reported in Table 5-13.

                  TABLE 5-13.  ANNUAL AVERAGES OF ORGANIC FRACTIONS IN TSP
                            NEW YORK CITY3, DISPERSION NORMALIZED

Year
1968
1969
1977-78
TSP,
3
jjg/m
95.7
129
59.8
Organic fraction
pg/m
10.2
10.8
8.8C
Percent organics
in TSP
10.6
8.4
14.7




       aNYU Medical Center Station.
       b
        total of nonpolar (benzene-soluble) and polar (acetone-soluble) organics.
       €Respirable (£3.5 u) organics only.
       Source:  Daisey (1980).
                                            5-83

-------
    10
     •
 DC

 Z
 ui
 u
 z

 8   4
 CQ
                                                            . 90PERCENTILE OF

                                                             QUARTERLY
                                                             MEASUREMENTS


                                                            •50PERCENTILE OF

                                                             OUARTERI.Y
                                                             MEASUREMENTS
                                            1
            1966     67      68       69     70      71


                                         TIME, year
72
73
74
75
Figure 5-40,  Benzo{a)pyrene seasonality and trends (1966 to 1975) in the 50th and 90th

percontiles for 34 NASN urban sites.



Source: U.S. Environmental Protection Agency, 1979.
                                            5-84

-------
Although TSP had  decreased by 40% between 1968  and 1978,  the POM fraction  decreased by only
10%.
5.5.3.3   ChemicalComposition of Particulate Organic Matter—Participate  organic  matter  has
often  been  fractionated by  means of acid-base extractions followed  by column chromatography
(Hueper et a!.,  1962;  Tabor et al., 1958; Hoffmann and Wynder, 1977;  Asahina et  al., 1972).
The composition  data for  Detroit PM  found  in Table 5-14 are fairly typical  and  present the
general proportions  of  various broad classes of compounds present.  Hoffmann and Wynder (1977)
found  that  the fraction containing  PAH was  principally responsible  for  the tumorigenic pro-
perties of POM in mice.

                         TABLE 5-14.   COMPOSITION OF THE ORGANIC FRACTION
                               OF AIRBORNE PM COLLECTED IN DETROIT3

         Fraction                                     Percent of total extractable
                                                            organic matter

         Aliphatic hydrocarbons                                 48.3
         Aromatic hydrocarbons                                   3.6
         Neutral oxidized hydrocarbons                          20.8
         Acidic compounds                                       14.8
         Basic compounds                                         0.55
         Insolubles                                             10,8
         a3 jjg/m  annual average benzene-soluble organics.
         blncluding PAH.
          TSP not reported.
         Source:  Hoffman and Wynder (1977).
     Specific classes  of compounds identified in the  organic  fraction of airborne PM include
PAH,aromatic  and aliphatic  hydrocarbons,  aza-arenes,  aliphatic  and aromatic  aldehydes and
ketones,  quinones,  phenols,  polyols,  phthalic  acid  esters,  sulfur  heterocyclics,  aryl and
alky!  halides,  chlorophenols,  nitro  compounds,  and  alkylating  agents  (Hoffmann  and Wynder,
1977;  Daisey,  1980;  Lamb et  al., 1980).  Of  all  the  airborne organic compounds,  the most
information exists  for  the  classes of  POM.   The greatest attention  has  been focused on the
subclasses  of PAH  and the polycyclic heterocyclic compounds  such  as the aza-arenes, because
many of the compounds  in this class are potent carcinogens in  animals.   Some of the polycyclic

                                            5-8$

-------
hydrocarbons   identified  are  pyrene,   BaP,   benzo(e)pyrene,   benz(a)anthracene,  perylene,
chrylene,  chrysene,  coronene, fluoranthene, benzo(ghi)perylene, and  alkyl  derivates of these
compounds (Sawicki et al., 1962; Sawicki et a"!., 1965).
     Benzo(a)pyrene was  one  of the earliest compounds in this mixture of organic matter to be
identified and routinely measured.  Some measurements for BaP in the United States date to the
                                                      »  •*.«   '^
early  1950's.   Sawicki  and  coworkers  in  the 1960's extracted  and identified  many organic
compounds.   Today  there  is  a  renewed  effort,  using  more  sophisticated  techniques  and
attempting  to  answer the  many  questions  still  remaining on  the  biological  significance,
variations and  concentrations,  specific source contributions, and  the  reactivity of airborne
organic matter.
     In  1972,  the National Academy of  Science published an extensive  report  on the biologic
effects of airborne matter entitled Particulate Polycyclic Organic Matter (National Academy of
Sciences,  1972).    According  to  the  report,  emission  source   data   for  airborne  organic
substances are  generally expressed  in  terms  of  estimated BaP emissions.   Benzo(a)pyrene is
used  as  a  surrogate for  detecting the  presence of  airborne  organic  pollutants  because it
appears  to  be a prominent  constituent  of  POM.    Benzo(a)pyrene  is  'also  a  known  animal
carcinogen and  the  best  documented of all  the polycyclic organic compounds (National Academy
of  Sciences,  1972).   It cannot  be regarded  as  a  perfect indicator of  polycyclic aromatic
hydrocarbons in the air nor of their carcinogenic properties; however, because better data are
generally  not available,  BaP  is  presently used  as an  indicator  of the  potential catcino-
genicity of general air pollution (Bridbord, et al., 1976).
     Despite  much  work  on   certain  subfractions  of POM, such   as  the polycyclic  organic
fraction,  other  compound  classes  such  as  the  oxidized  hydrocarbons  remain  relatively
unexplored.   Sawicki  (1976)  has estimated  that  "over 99% (sic) of the  organic  pollutants in
the air have never been determined."
     In photochemical incidents,  volatile hydrocarbons are converted to very large quantities
of  5  to  7  carbon  bifunctional  carboxylic  acids  (Schuetzle  et  al.,  1975;   Grosjean  and
Friedlander,  1975;  Cronn  et al.,  1977).   Schuetzle et al. (1975), in a  report  of  a  1972
incident, state that alkanes and alkyl naphthalenes accounted for 1.5 to 3 percent of the fine
particle  mass,  and  bifunctional  compounds  amounted  to about  11 percent.   In  addition to
glutaric, adipic,   and pimelic acids,  the corresponding hydroxy carboxylic acids and a variety
of their nitrate and nitrite ester derivatives were reported.
     Cronn et al. (1977)  confirmed those  findings  in a  series  of  sub-3.5 pm  samples taken
during the 1973  California  Air Characterization  Experiment.   These authors  found  levels of
                                         3                                           3
organic particulate matter  up to 65 jjg/m  out of a fine-particle loading of 230 pg/m .   These
substances included  small  amounts  of alkanes, alkyl naphthalenes, and  piperidines  (up to 12
ug/m ) and much larger  quantities of CK  to  C-, dicarboxylic acids, hydroxy-acids, and amides.
                                                                                   3
     Grosjean  and Friedlander  (1975) have  found  organic extractables of 141  jjg/m  during an
incident in 1973;  one-half to one-third of this mass was polar organics.   These organic
                                            5-86

-------
substances  together  with  ammonium  sulfate  and  nitrate  accounted for  95  percent of  the
secondary aerosol  during photochemical  incidents.   Therefore, there is  substantial  evidence
that organic particles can be influenced in a very major way by photochemistry.
     Classes of biologically  active  compounds other than  PAH  and related polycyclic organics
have been identified  in  airborne PM.   This includes alkylating agents, N-nitrosamines (Kneip
et al.  , 1979),  nitro  derivatives of PAH  (Jager,  1978),  and the compounds responsible for the
mutagenic activity  of POM  in the Ames assay (Kneip et al.} 1979;  Talcott and Wei, 1977;  Pitts
et al., 1977; Daisey et al.s  1979).   In addition, there may also be unstable compounds present
in the aerosol, such  as epoxides and  lactones  (Van Duuren, 1972), that  are  significant for
human  health  but decompose  when collected by  conventional sampling techniques.   There  is a
need to  identify specific compounds such  as  these,  to evaluate  their  significance for  human
health, and  to  determine their sources and concentrations in the ambient atmosphere.  There is
also a need  to  identify  sampling  artifacts and  develop  improved sampling  techniques  for
organic compounds in the aerosol.
5.5.4  Metallic Components of Fine Particles
     It  is  useful  to study  not only  the  chemical  but  also the elemental  composition of
airborne  particles.   Many trace  elements are known to  be toxic  and can  act  as catalysts in
atmospheric  reactions.   Table 5-15  indicates the mean and  maximum concentrations  of several
elements  found  in  urban  and  nonurban areas in  the  United States from  1970  to 1974.   Certain
trace  elements tend to be enriched in urban airborne particles relative to their concentration
in fuels.   For example,  Table 5-16 lists the  ratios  of  urban to  suburban  concentrations of
trace  elements  in  three  groups.   Since  the  ratio of  urban to  suburban  TSP was  2.8,  those
elements  with  higher ratios (antimony  through  bismuth)  were concluded to be  mainly  of  urban
origin.   The middle group  (mercury  through calcium) had ratios near that of TSP and therefore
occur  equally abundantly in city and suburban particles.  The remaining group (silicon through
bromine)  appeared mainly in  suburban suspended  particulate  matter.  As Table 5-17 indicates,
they were not  homogeneously  distributed among the various particle size fractions.   There was
also spatial variability in the composition of the aerosol, as indicated by Table 5-18.  These
intercity differences reflect the  difference in  industries  and  types of fuel  used  in  these
urban  areas.
     Hi-vol  filter  samples from the NASN have been routinely analyzed for certain metals since
the  early 1960's.   The  data for certain  metals for  the years  from  1965 to  1974 have been
summarized  in  an EPA  report  (Faoro  and McMullen, 1977).   This report  presents the composite
national  and regional  trends for  nine  trace   metals,  fuel-related metals  (lead, vanadium,
nickel,  and titanium)  and  industry-related  metals   (cadmium,   chromium,  copper,  iron,  and
manganese).   These  trends were  derived from  samples collected from 92 urban  and 16 nonurban
NASN hi-vol  stations.
     The  instrumental techniques for detecting metals changed in 1970, significantly improving
the  lower limits of detection.  This report,  in  addition to Akland  (1976) describes the
                                            5-87

-------
                   TABLE 5-15.  COMPARISON OF URBAN AND NONURBAH ANNUAL AVERAGE CONCENTRATIONS FOR SELECTED HETALS, 1970-74

                                  Cd           Cr          Co          Fe         Pb            Hn           N1       	TJ_
                 IT    NUb     U    NU      U   NU     U    NU     U    NU    U      NU     U     NU     U     NU    U     NU     U       NU
    1970
    Tiximuiii     2.9    .24   ,099  .0001  .130  .07S   .014  LD   14.2  1.62  5,83  1.471   2.10  .068   .277  .097   .26   .093   1.222   .112
     ArithiM-
     tic mean     L0q    ID   .003  .0001  .008  .003     LD  ID    1.7   .38  1.19   .088    .07  .015   .015  .005   .05   .013    .052   .008
     Std.  devi-
     ation      0.2    .03   .007   --    .011  .009   .001  ~    1.3   .27   .80   .190    .12  .013   .028  .014   .03   .011    .116   .019
    1971
    TixidtttS     0.7    .24   .215  .0001  .171  .061   .085  LO   16.0  2.80  6,31  1,134   1.95  .102   .347  .083   .51   .069   1.325   .209
     Arithme-
     tic mean    0.1    .01   .004  .0001  .009  .004   .001  LD    2.1   .51  1.23   .047    .08  .018   .015  .003   .04   .017    .041   .007
     Std.  devi-
     ation      0.1    .03   .016   --    .014  .007   .003  --    1.6   .38   .87   .155    .11  .015   .028  .011   .05   .020    .108   .024
tn
s>  19.72
»  TiaxiiMi       LO    LO   .112  ,0001  .143  .039   .042  LD    6.4  1.15  6.88  1.048    .86  .046   .268  .082   .48   .092    .858   .205
     Arithme-
     tic mean      -    LO   .002  .0001  .006  .002    LD   LD    1.2   .25  1.13   .139    .04  .007   .011  .004   .04   .027    .022   .004
     Std.  devi-
     ation        —    —   .007   --    .010  .004   ,002  --     .8   .22   .78   .169    .06  .009   .023  .012   .03   .022    .056   .019
    1973
    "Maximum       LD    LO   .032  .0001  .228  .066   .027  LD    6.9  1.19  5.83   .939    .56  .030   .439  .280   .23   .084    .393   .035
     Arithme-
     tic mean      LD    LD   .001  .0001  .007  .003   .001  LD    1.1   .19   .92   .110    .04  .004   .014  .011   .04   .028    .016   .002
     Std.  devi-
     ation        LD   ---   .003   --    .015  .009   .002  ~     .8   .18   .64   .149    .05  .005   .037  .037   .03   .021    .034   .005
    1974
    "Maximum     0.2     LD   .077  ,0002  .073  .009   .029  LD    6.2   .69  4.09   .534    .35  .033   .639  .026   .22   .066    .248   .023
     Arithme-
     tic mean      LD    LD   .002  .0002  .006  .002    LD   LD    1.1   .24   .89   .111    .04  .006   .009  .002   .04   .020    .019   .002
     Std.  devi-
     ation        -     —   .001   —    .006  .002   .001  —     .7   .17   .57   .111    .04  .007   .029  .004   .03   .017    .037   .004
                       __,                                                            ,
    Expressed in ng/m?;  U  = urban; NU = nonurban;  CLD = less than detectable.   Source:  G. Akland (1976).

-------
    TABLE 5-16.   RATIOS OF URBAN (U) TO SUBURBAN (S) CONCENTRATIONS IN AIR,
                             CLEVELAND, OH, AREA
Enriched in cities
                 Ratio similar to TSP
                          Enriched in suburban par-
                                   ticles
  Element  U/S
 Antimony
 Chloride
Beryllium
 Chromi urn
   Cobalt
  Bismuth
  6.9
  6.5
  6.1
  5.6
  3.4
  3.3
                   Element  U/S
  Mercury
     Iron
  Cadmium
   Sodium
Magnesium
Manganese
  Calcium
3.0
2.8
2.5
2.4
2.4
2.2
2.0
     Element      U/S

    Silicon, tin  1.8
Copper, vanadium  1.8
        Aluminum  1.7
            Zinc  1.6
         Arsenic  1.4
        Selenium  1.3
         Bromine  1.2
Note:
Mean TSP ratio = 2.8
Source:  Economic Commission for Europe's Working Party on Air Pollution Problems
         (1977).
                                   5-89

-------
       TABLE 5-17.  CORRELATIONS OF CHEMICAL CONTENT WITH PARTICLE SIZE

               a) Predominant particlesize for various substances
Normally fine       (formally coarse       Normally bimodal       Variable
                    Iron, calcium
Sulfates
Organic (con-
densed vapors)
Lead
Arsenic
Selenium
Hydrogen ion
Ammonium salts
Soot
                    Titanium

                    Magnesium
                    Potassium
                    Phosphate
                    Silicon
                    Aluminum
Chloride
Nitrate
Nickel
Tin

Vanadium
Antimony
Manganese
Zinc
Copper
       b) Ratios of element distribution between fineandcoarse particles
           (St. Louis urban aerosol, 18-day average, Aug. to Sept. 1975)
       Predominantly fine
                                         Predominantly coarse
Element Fine/coarse
Sulfur 8.90
Lead 3.67



Element
Calcium
Silicon
Iron
Potassium
Titanium
Fine/coarse
0.09
0.13
0.29
0.33
0.55

Source: (a) Miller et al. (1979).
        (b) Dzubay et al. (1977).
                                   5-90

-------
                        TABLE 5-18.   PARTICULATE ANALYSES FROM SELECTED URBAN LOCATIONS
i
tc


Suspended particles
Antimony
Beryl 1 i urn
Bismuth
Cadmium
Chromium
Cobalt
Copper
Iron
Manganese
Nickel
Tin
Titanium
Vanadium
Zinc
Atlanta, 6A
97.
0.000
0.000
0.000
0.017
0.002
0.000
0.04
1.2
0,06
0.007
0.02
0.03
0.001
0.52
Birmingham, AL Baltimore, MD
142.
0.000
0.000
0.000
0.008
0.005
0.000
0.06
1.7
0.15
0.004
0.01
0.03
0.003
1.09
146.
0.000
0.000
0.000
0.003
0.018
0.000
0.06
0.8
0.08
0.034
0.01
0.01
0.071
0.34
Albuquerque, NM
120.
0.000
0.000
0.000
0.000
0.001
0.000
0.07
a
0.03
0.000
0.01
0.01
0.001
0.00

        Not analyzed.
       Arithmetic mean values for 1966 expressed as micrograms per cubic meter.
       Copyright by the American Association for the Advancement of Science, 1970.
       Source:   Corn (1976).

-------
methods  used  and  the  implications  for  trends  analysis.   In general,  the  data  presented
indicate  changes  in  atmospheric concentrations  of  various  metals  occurring  in  different
regions of the country.  For the most part, the reported trends are consistent with changes in
emission patterns (due to industrial source control) and in fuel use.
     Similar  to  trends  in  urban TSP  concentrations,  metals concentrations  declined in most
urban  areas,  with  the  exceptions  of  copper,  titanium,  and possibly chromium.  Table 5-19
summarizes metal  trends  and possible causes for these trends.   Both beryllium and cobalt had
such very  low concentrations that trends  could not  be identified with any certainty.  Trends
in  other  metals, such as  vanadium and nickel, parallelled  air pollution control regulations
mandating the  use of low-sulfur fuels.  There was a drop  in vanadium and nickel, particularly
in  the  Northeast,  because the desulfurization  process  of  petroleum  also  removes  these
impurities.  Titanium, on  the  other hand, may have  increased because of the rise in coal use
by  utilities.    Decreases  in iron,  manganese,  and  cadmium  concentrations  were  probably the
result  of  reduced  particulate  emissions   From steel  plants  and related industries.   Improved
incineration  methods  and  use   of sanitary  landfills  instead  of  incinerators may  also have
contributed to the decrease.  No trends were apparent for copper, but it is felt that the high
concentrations were  caused  by  contamination from the commutators of the high-volume samplers.
5.5.4.1   Lead—The  seasonal  patterns  and  trends  in  the  quarterly  averaged urban  lead
concentrations are  displayed in Figure 5-41.  The national  composite 50th  percentile of lead
                                           3                       3
concentrations  fell from  about  1.1  ug/m   in  1971 to 0.84 ug/m   in 1974,  approximately  a
24-percent decline.  This  is attributed to decreased lead in gasoline and the drop in premium
gasoline sales since 1970.   Premium gasoline has, on  the  average,  a higher lead content, than
regular  gasoline.   Lead concentrations  are expected  to continue to  decrease in the  future
because of increased use of unleaded gasoline in new cars equipped with catalytic converters.
This  national  decrease  is  not  equally evident  throughout  the country,  however, because  of
differences in  growth  rates and  vehicle  miles  travelled.   These results should  be  used with
caution  because  of  the  small   number  of  stations used in  determining  the trends.   An NASN
location  in   downtown  Los  Angeles  experienced the highest  lead   concentrations,  averaging
                     o                                                            o
between 4  and 5  ug/m   until 1971;  the concentrations  decreased  to about  2 ug/m   in  1974.
Again,  while  the reduction  of  lead content  in  gasoline  and the increased  use  of  lead-free
gasoline may  have  contributed  to this decline,  the decrease  in  vehicle miles  travelled  in
downtown urban areas also contributed to the decline.  There is some evidence that rural  sites
have shown either stable or slightly  increasing patterns  in ambient air lead  content  in the
United States.
5.5.4.2   Vanadium,Nickel.and Other Metals—Figure  5-42  shows,   for the   five broad  geo-
graphical areas  of  the United -States,  the 90th percent!les  of the annually averaged vanadium
concentration.  The  Northeast had substantially higher vanadium concentrations than any other
area of the United  States.   Over this 10-year period the concentrations of vanadium decreased
                                            5-92

-------
           TABLE 5-19.   TRENDS IN REPORTED URBAN METAL CONCENTRATIONS
                           AND THEIR POSSIBLE CAUSES
    Metal
Observed trends
     Possible causes
Fuel  combustion-related
 metals:
 Beryllium
 Lead

 Nickel

 Titanium

 Vanadium
Industry-related
 metals:
 Cadmium
 Chromium
 Cobalt
 Copper
 Iron

 Manganese
Unknown
Down; expected to
decrease further
Down
Up
Down
Down

No trend
Unknown
No trend
Down

Down
Lower lead content in
gasolines after 1969
Reduction of Ni in residual
oils
Increasing use of coal in
electric utilities
Reduction of V in residual
oils
Controls in metal industry
and improved incineration
practices
Unknown
Contamination from hi-vol
commutator
Improved incineration or waste
burning practices, fuel switch-
ing, controls in steel industry
Controls in metals industry
Source:  Faoro and McMullen (1977).
                                   5-93

-------
     3.0
M
     2.0
 en

 z
 01
 u

 O
 u
 01
     1.0
      0
1965      66      67     68
                                                    I	I
                                                                72
73      74
                                   69       70      71


                                        TIME, year



Figure 5-41.  Seasonal patterns and trends in quarterly average urban lead concentrations.



Source:  Faoro and WlcWIullen, 1977.
75
                                          5-94

-------
    0.10
"I

o
LU
o

O
o

5
I
    0.01
   0.001
                                        /       X        .
                                       /          <*     s
                                       /    .V        N,  S
   -r   .•  —\->


'\      /      x-
   *      4

                                 \
                        N.
                            NORTHEAST (29 SITES)


                            SOUTH (IS SITES)


                            WEST (15 SITES)


                            NORTH CENTRAL (14 SITES)


                            MIDWEST {19 SITES)
                    I     I      I      I     I     I      I      I      I      I
                  1965   66    67    68    6S   70    71    72    73    74


                                           TIME, year




 Figure 5-42.  Regional trends in the 90th percentile of the annual averages for vanadium. (A indi-

cates value below lower discrimination limit.)




Note:  1971-1974 90th percentile below lower discrimination limit of 0.003 = jug/m3in the midwest.


Source: Faoro and McMullen (1977).
                                           5-95

-------
                          3                     3
74 percent, from 0.35 ug/m  in 1969 to 0.09 ug/m  in 1974.  Most of this drop occurred between
1971 and  1972.   The slight increase apparent  in  the South was caused  mainly  by two or three
stations  showing relatively high  readings in  1972  to 1974;  this was  not  characteristic of
other sites  in  the region.   For both  vanadium and nickel, pronounced and regular high winter
and  low  summer  seasonal  variations in  both  the  50th and  90th  percentiles occurred  in  the
Northeast.   This  is  shown  in  Figure  5-43.    These  variations  are  attributed to  the metal
contents  in  the  fuels used for space  heating.   The  decrease in the 50th and 90th percentiles
of these two metals was caused by the decrease-in the sulfur content in petroleum used in this
area.   This  decrease is  exemplified by  the  approximately 70-percent  decrease  in  the sulfur
content of residual oil in the New York City-Westchester County area since 1979.   The vanadium
concentrations decreased  between  70 and 80 percent  over  the same time period at the New York
City NASN site.
5.5.5  Acidityof Atmospheric Aerosols
     Along with  size and chemical composition,  acidity  of  fine  atmospheric  aerosols  is an
important  property.   Measurement  of  acidity   by  titration  is  preferred to  pH measurements
(Junge  and  Scheich, 1971).   To date,  measurements of the  strong  acid  content of atmospheric
aerosols  have indicated that it is quite variable.  Around certain sources such as cement and
lime kilns,  the  airborne  particles may be  basic,  whereas around  other  sources  such as HpSO^
plants and coke plants the particulate emissions may be very acid.
     Weak  and strong  acids exist in  the  atmosphere in  both  gaseous and  particulate form.
Organic  acids  were  reported  by  Ketseridis   et  al.  (1976)  in  rural,  urban,  and  marine
                                       3             3
atmospheres  ranging  between 0.3  ug/m   to 10 ug/m .    Nitric  acid  has  been  found  in  the
                                                                2
atmosphere in concentrations ranging from < 2 to nearly 100 (jg/m   (Spicer, 1977).  Vapor-phase
HC1 may be present (Rahn  et al., 1979), although quantitative data are sparse.  These organic
and inorganic gases  can  condense or be adsorbed  on  particles either jn situ or during filter
sampling.   As pointed out previously, filters are capable of absorbing HNOg directly.
     Host of the strong acid found in aerosol particles is chemically associated with the fine
              O«*
particulate SOI  aerosol  mass.   Charlson  et al. (1978) and others cited numerous measurements
of approximate chemical balance between ammonium cations  and sulfate  anions.   Thus the major
           ?—
form of  SOT   is  ammonium  sulfate  (that  is,  HgS04  fully neutralized by  ambient  ammonia).
However,  on  occasion  urban  and rural aerosols  can be acidic.,  Brosset et al.  (1975), Brosset
(1978), Hitchcock  et al.   (1980),  Cobourn  et  al.  (1978),  Pierson et al.  (1980),  Lioy et al.
(1980), Leaderer et al.   (1982),  Stevens  et  al.  (1979),  and Tanner et al. (1977,  1979)  all
demonstrated that strong acid in the form of NH*HSO*  and,  less frequently, H-SO^, may exist at
significant levels  in  the ambient atmosphere.   Strong acid levels equivalent to > 15 ug/m  of
HgSO, have been observed for periods :>  6 hours.
     In urban  atmospheres,  sulfate  anion  usually appears  primarily in the form of ammonium
                                    +   ?-
sulfate or  ammonium bisulfate  (NH^ /S0|   molar ratios between 1 and 2)  as reported by Lioy
                                            5-96

-------
      0.800
                                     I      I      I      I      I      I

                                                           VANADIUM
                                                   90th PERCENTILE     _
                                                                  73   74
Figure 5-43. Seasonal variation in quarterly averages for nickel and vanadium it
urban sites in the northeast.

Source: Faoro and McMullen (1977).
                                     5-97

-------
et al,  (1980)  for New York City,  Leaderer  et al.  (1982) for New Haven and New York City, and
Cobourn  et  al. (1978)  for St.  Louis.   Presumably this  greater extent of  neutralization of
        2~
urban SO.  aerosols is due to additional NH- sources in urban areas, although it may be due in
part to  analytical  interferences from coarse, basic particles such as resuspended cement dust
present in large amounts in urban aerosols.
     On  the  basis  of  recent  experimental  and  theoretical  work,  Huntzicker et al.  (1980)
                 2~
indicate  that  SO,  aerosol  more acidic than  (NH-)HSO. should  occur only when  SO- is being
oxidized rapidly and where the ratios of [SO,] to [NH,]'are high or when the equilibrium vapor
                                            €,        3
pressure of NH3  over  the partially neutralized  H^SO.  droplet exceeds the ambient NH, partial
pressure.  The situation is more complicated in ambient aerosols in which partially ammoniated
  p~
SO,  is  present  in  mixtures with NO,,  carbonaceous,  and other aerosol components in solid or
liquid  form which  may  affect  its  neutralization rate.   In particular,  some data (Tanner,
1980a) suggest tha.t the degree of mixing in  the "well-mixed" boundary layer is inadequate to
prevent vertical  stratification of strong acid  levels  since  NH3 is largely emitted (and HNO-
largely  removed  by  dry deposition)  at the  earth's  surface.   Further  information  on  the
vertical  distribution  of  strong  acid  and  related  species  is needed  before  emission  and
neutralization rates may be used to predict acid levels in urban areas.
     Cobourn et al.  (1978) demonstrated that acid aerosol episodes could occur in urban areas
as suggested by Tanner et al. (1977, 1979) and Brosset (1978).  Cobourn et al. (1978), using a
             2-
continuous SO.  monitor to distinguish H^SO* species from ammonium sulfate, recorded two acid
aerosol episodes  lasting 3 days or longer  in  St.  Louis, MO.   Both episodes  were reported to
have occurred  simultaneously with  a regional  haze,  one in July 1977, the  other in February
1978.   Cobourn ascribed  the  city  occurrence  of H?SO. to  conditions where  atmospheric  NI-L
concentrations were exceptionally  low  (NH,  was  temporarily depleted from  the  atmosphere).
                                                             ?—
     The temporal variations  of the acid fraction of  the  SOf  aerosol in St. Louis displayed
patterns  similar  to those  reported by  Cunningham and Johnson  (1976) in Chicago  and Tanner
(1980b)  for  the New  York area.  The  aerosol  acidity often  changed  drastically  within a few
hours.   Cobourn et al. (1978) noted a diurnal pattern with highest acid levels in midafternoon
and  lowest  at night.    Leaderer et  al.  (1982),  taking 6-hour samples,  reported  increased
aerosol acidity in  the  noon to 6 p.m.  samples taken at High Point, NJ and Upton, Long Island
(west  and  east of  New  York City,  respectively),  compared to other  sampling  times.   At this
time the  relative contribution of SO, oxidation chemistry, temporal variations in NH, and SO,
emission rates, diurnal  variations in turbulent mixing rates,  and varying height of the mixing
layer  to  the  diurnal  patterns were  not  known.   Dzubay (1980) observed that  while  sulfur and
lead dominated the  fine particles,  they  were insignificant in  the coarse  particles.   The
similar  composition of the rural and  urban aerosols indicated  that  the  urban materials were
transported to the rural areas.
                                            5-98

-------
     Measurements of  acidity in  eastern  United States  aerosol  samples  indicate  that strong
acids are present more frequently and in larger amounts in rural  and semirural samples than in
the urban samples.  Pierson et al. (1980)  reported 12-hour concentrations of H  ions expressed
as HpSCL  as high as  17 ug/m  in the Allegheny Mountains of Pennsylvania  in July and August
1977.  It is likely that strong acid concentrations were substantially higher for periods less
than 12 hours.
     Lioy et al.  (1980) and Leaderer et  al.  (1982) characterized the aerosol  acidity in the
region surrounding New  York City during the summer of 1977.   The samples taken at High Point,
NJ (west-northwest of the  city), were more acidic (17.8 jjg/m ,  6-hour average maximum H  ions
expressed as  HpSO*)  than  samples taken  in  New Haven,  CT,  and Brookhaven,  Long  Island, NY.
     Part of  the period studied by Lioy  et al.  (1980)  was coincident with the  research of
Pierson et  al.  (1980).   Using  the combined  data  set in conjunction with  air  parcel  trajec-
tories and haze analyses, Lioy et al.  (1980) suggested the presence of a regional acidic aero-
sol  distribution  encompassing  an area at least 200 miles in diameter during the period August
3 to 9, 1977.
     As pointed  out  earlier,  atmospheric  particles are nearly neutral for the most part, pre-
sumably because  of the  presence of NHg in  the atmosphere.   However, instances of particulate
acidity have been reported in several credible  studies  done in the East,  in both winter and
summer and in rural and urban sites.
5.6  COARSE PARTICLES IN AIR
5.6.1  Introduction
     In Chapter  2 and  in  earlier  sections  of this  chapter it was shown  that  air in cities
usually contains  large  amounts  of particles  larger than  1  to 3 pm in size.  Particles larger
than 10 to 20 pm tend to settle out of air suspension under the force of gravity.  Yet in many
areas these very large dust particles are also present in substantial quantity.   This material
is quite  commonly deposited  as  dust on  window ledges  and  silt on  roadways.   For potential
effects of  the  ambient  concentrations described below,  the  reader  is directed to Chapter 10,
Section 10.3.2,  for  descriptions of soiling  and to Chapter  11,  Section 11.2, for information
on   deposition  of  particles  in  human  airways.    Coarse-particle  mass  contributions  are
substantial and important in the context of air pollution effects.
     The composition and sources of coarse particles are not as thoroughly studied as those of
fine  particles.   One reason  is  that they  are  more complex.  For example,  it  is  possible to
recognize dozens of particle types in  ambient air samples;  these  range  from soil particles,
limestone road  aggregate,  fly ash  and  oil  soot to cooking  oil  droplets,  pollen,  wood ashes,
and  even  instant coffee   (McCrone,  1968;  Draftz and  Seven"n, 1980).   Man's industry and
activity  stirs  up dust, quite a  lot  of  dust in  arid climates.  Unfortunately, the chemical
composition of  many  kinds  of coarse particles  can be very similar, at least as determined by
elemental analysis.  Consequently, touch of the evidence on large-particle composition has been
obtained from deductions based on microscopical examination.
                                            5-99

-------
5.6.2  Elemental Analysis of Coarse Particles
     Measurable elements  constituting  the major portion of coarse-particle mass in cities are
silicon, aluminum,  calcium,  and iron (Akselsson et a!., 1975; Lewis and Macias, 1980; Camp et
a!,, 1978;  Stevens et  al.,  1980;  Dzubay,  1980; Stevens  et al., 1978;  Cahill  et a!.,  1977;
Hardy, 1979; Trijonis et al., 1980),  Although these elements do exist in the fine fraction to
a minor degree,-they are everywhere substantially enriched in the coarse fraction.
     Occurrence of  some elements in coarse particles is time and place dependent,  though, and
Table  5-20 shows  some  data illustrative  of this  point.   There  appear  to  be  substantial
differences across  the country  in the  fraction of these  elements  occurring  in  coarse par-
ticles.  The presence  of local  sources dominates both the mass and composition of coarse par-
ticles.  However,  Cooper and Watson (1980)  have  graphically demonstrated  the similarity in
elemental distribution  for  a variety of coarse-particle sources  as  Figure 5-44 shows.   Here,
the  most  that can be  said  is  that  rock-grinding  operations  produce  remarkably similar
coarse-particle elemental compositions,  whether  the mechanical action is intentional or inci-
dental to  other activities.   Cement dust and  limestones  (not surprisingly) also have similar
elemental composition (Draftz, 1979).
     Even  greater  evidence of  localized influence on  coarse-particle concentrations can be
seen with  other elements.   In the Smoky Mountains, titanium and chlorine are greatly enriched
in  the coarse  particles (Stevens  et al.,  1980) while  in St.   Louis,  titanium is  mainly  a
fine-particle  component and  chlorine is  about  evenly distributed  between coarse  and fine
particles  (Stevens  et  al.,   1978;  Akselsson  et al.,  1975).   In the  case of  titanium,  the
emissions from a paint plant greatly influence the fine-particle titanium (Rheingrover,   L981).
Chlorine appears to originate in fine automotive particles at inland sites (Winchester et al.,
1967) but in coarse sea-salt particles near the coast (Hardy, 1979;  Draftz, 1979).
     A variety  of  carbon-containing species, particularly  organics  and  carbonates,  sometimes
can  be found in substantial  quantities  in  coarse particles.   For  example, De Wiest  (1978)
found  30  to 50 percent of extractable  organics in 2  to 10 \*m  particles.   Lewis  and  Macias
(1980) found 40 percent of the carbon (presumably mostly in organic compounds) in dichotomous
sampler coarse  fractions  in  Charleston,  WV, while Stevens et al.  (1980) found about one-third
of the organic  compound mass in the coarse fraction in the Great Smoky Mountains.   Mueller et
al.  (1970) were  able  to  differentiate  between  carbonates  and  elemental  carbon by acid
evolution of CO,,  but  this  technique, unfortunately, has  not  been  applied to coarse-particle
analysis.   Considering  that calcium  carbonate  has often  been found as  a  major component of
urban  coarse-particle samples (Graf et al., 1977; Draftz,  1979), it is surprising that direct
analyses for  carbonate have  not been reported.   Elemental carbon appears  to  be  uncommon in
coarse particles^   As  mentioned  previously,  most water-soluble  inorganic  ions are found in
fine fractions.
                                            5-100

-------
TABLE 5-20.   COARSE PARTICLE SILICON,  ALUMINUM,  CALCIUM, AND IRON

Location
Charleston, WV

Smoky Mountains,
TN
New York, NY
Y1 Philadelphia, PA
1— j
2
St Louis, MO
Portland, OR
Glendora, CA
St. Louis, MO
Dates
08/25-
9/14/76
05/11-
05/19/77
09/21-
09/26/78
02/77
03/77
12/75
02/77
03/77
08/-
09/76

Coarse
mass
27.1
43
5.6
42.6
17.5
NA
27.6
NA
28.0
Mfl/m
Si
2.8
7.7
0.50
2.0
1.8
4.3
2.8
1.0
4.5
3
Al Ca
1.1 0.96
NA 2.2
0.20 0.32
0.84 1.15
0.64 0.94
8.7 1.9
1.2 0.76
0.3 0.44
1.2 2.8
Coarse/fine
Fe
0.59
1.4
0.12
0.96
0.69
1.0
0.95
0.36
1.2
Si
6.8
7
15
5.6
7
10
31
5.3
10
Al
15
NA
10
6.5
13
4.4
5.6
>6
6
mass ratios
Ca
9.7
7.4
20
3.2
6
16
11
4.5
21
Fe
4
4
4
2.5
3.2
3
5.0
3
4.4
Reference
Lewis and Mac i as,
1980
Camp et al . ,
1978
Stevens
et al . , 1980
Stevens
et al . , 1978
Stevens
et al . , 1978
Stevens
et al., 1978
Stevens
et al., 1978
Stevens
et al . , 1978
Dzubay, 1980

-------
ui
IU
o-
IUU

10.0
1.0



0.10
ft f|<{
U.U i
100


10.0

1.0

0.10
0 01


-
-Ni




mat



SOIL
Si







K C.T.









Mn
F*
•Ml




i UT
b





















—




Zn







ROCK CRUSHER






d











Si
r-







Cl
fl



Fa


S Mr
JCI v_Crl
inn
IUU

10.0
z
111
g 1.0
Ul
Q.
0.10
n ni
















^^

—

Cu
I'd*


Si
_ f
Naj
-


-










S












Cl
IUU

10.0
1.0



0.10
0.01
100


10.0

1.0

0.10
n ni
ROAD DUST
Si
~~ r™^ Fo
_pl ,iiflTI




S "•"•


Cl



Mn
/
—


Pb
^m
Zn
Al "
Nil


ASPHALT PRODUCTION
Si
— AJJ Fe _
I /*„



• j«*
S
Cl

Mn
/

—

^^ "

COAL PLY ASH

C. fL
n ''







—


Zn ~
ut, Mjj n pb
RI »J lB/n








                  Figure 5-44.  Elemental composition of some coarse particle components.


                  Source:  Cooper and Watson (1980).
                                                  5-102

-------
5.6,3  Evidence from Microscopical Evaluation of Coarse Particles
     Efforts to understand the importance of coarse particles in air have been hampered by the
inability of simple  chemical  analyses (so very successful with fine particles) to reveal  much
of their  nature.   However, estimates of mass  balances have  suggested  for a  long  time  that
locally generated coarse  particles  must constitute a substantial part of the suspended parti-
cle burden.
     As an  example,  the  most recent data  from the  EPA  network of  dichotomous  samplers and
hi-vols could  be  interpreted  as demonstrating significant amounts of particles larger than 15
pm in  the air,  if  the difference between the  dichotomous  sampler  total and TSP  is  taken as
representing supercoarse particles.   Table 5-21 displays some of these data for selected sites
for  illustrative  purposes.   Most  sites  have  two-thirds  or  more  of the  TSP in coarse and
supercoarse particles.  For the more arid and  dusty  parts  of the country, rough estimates of
this kind and  common  sense  have suggested  to pollution  control  officials that  TSP must be
dominated by locally generated coarse particles.
     Since these  larger objects can  be  readily inspected with an optical  microscope,  a  sub-
stantial  body  of information  has been accumulated  by visually  inspecting particle samples,
such as hi-vol  filters or impactor stages.  The  largest  of these studies involved evaluation
of 300  filters  from  14 U.S.  cities (Bradway and Record, 1976).  Table 5-22 presents  composite
analyses  for all filters from each of the cities.   A wide variation was found in these filters
ranging from virtually all  dust in Denver and  Oklahoma City to mostly dust with considerable
fly  ash and soot  in most of the industrialized  cities.   Chattanooga was  anomalous,  in  that
extremely large amounts of plant materials were found, including pollen, fibers, fragments of
leaves, and other tissues.
     Similar investigations  were  combined  with  emissions  inventory,  modeling,  and control
studies  in  Phoenix,  AZ,   in  1977  (Richard, 1977;  Richard  and  Tan, 1977;  Richard  et  a!.,
1977a,b; Graf et al., 1977).   In that city, 90 percent of the TSP was found to be mineral  dust
apparently entrained in air by automotive traffic over 1100 miles of unpaved roads in the area
and  by intense construction  activities.   Suck et al. (1978,  1979)  found,  through  meteoro-
logical modeling, that very  little motion of this  coarse PM occurs  in  the wind.  Since  wind
velocities are  characteristically low,  agricultural  influences are minor and coarse  particles
stay more or less where they are generated.
     Microscopic  evaluations  of  Miami  and  St. Louis particles have been conducted  both on
total  filter  samples  and  on  impactor plates  (Oraftz, 1979;  Draftz  and  Severin, 1980).  In
Miami, calcite (calcium carbonate) was the principal component of the coarse particles.  There
was  evidence that  a  small part of  the  calcite was recrystallized from ocean spray.   However,
most of the  calcite  appeared  to  be  roadway  aggregate suspended in the  air.   There  were  also
significant quantities  of halite (NaCl) and other trace  elements  characteristic of  sea salt.
                                            5-103

-------
     TABLE 5-21.  RELATIVE AMOUNTS OF FINE, COARSE, AND SUPERCOARSE
                       PARTICLES AT SELECTED SITES


Phoenix, AZ
El Paso, TX
Dallas, TX
Portland, OR
Los Angeles, CA
Akron OH
Philadelphia, PA
Hartford, CN

Fine
<2.5 pm
34
16
26
32
36"
49
51
34
Weight percent
Coarse
2. 515pm
6
51
10
64
31
26
32
32

Supercoarse
60
33
64
4
33
25
17
34

Note:  The term supercoarse refers to the difference between the
       hi-vol TSP concentration and total dicnotomous sampler concen-
       tration.
Source:  U.S. Environmental Protection Agency (1981).
                                5-104

-------
TABLE 5-22,  FQURTEEN-CITY STUDY - MICROSCOPICAL IDENTIFICATION OF COARSE PARTICLES
                              COLLECTED IN URBAN ATMOSPHERES

Wt. % of Component
Location
Oklahoma City
Denver
Miami
St. Louis
Washington, DC
Baltimore
Birmingham
Philadelphia
Providence
Seattle
San Francisco
Cincinnati
Cleveland
Chattanooga
Minerals
88
8}
79
75
70
69
66
64
64
60
52
51
51
36
Combustion
products
8
7
9
21
23
25
22
33
22
27
29
44
40
35
Biological
material
1
1
1
>1
5
3
2
1
1
3
3
1
1
16
Miscellaneous
(rubber tire
debris)
4
11
12
4
2
3
10
2
13
10
16
4
8
13

Source:  Bradway and Record (1976).
                                     5-105

-------
     The  general  picture  from these studies  is that  coarse  particles are  contributed from
numerous  local  sources  and vary dramatically from place to place.  It is likely that dust and
roadway aggregate  suspended  by traffic are major sources of coarse particles.  However, in in
industrialized  cities   especially,   there  is  still  some  evidence  of  combustion  source
contributions.
5.6.4  Fugitive Dust
     In the  discussion  of coarse particles, evidence was presented that a substantial portion
of TSP,  usually more  than half, is  accounted  for by coa'rse'-br  s'upercoarse  particles,  and a
great  portion of this  mass  is  mineral  dust, also  called "crustal material"  in  many of the
papers reporting chemical analytical data.  There is growing opinion that this major component
of TSP  is contributed almost entirely by agitation of soil in some way, and this component is
commonly  called fugitive  dust.  One  of the  major sources for  fugitive dust  generation  is
vehicle traffic; the motion of vehicles can reentrain silt (fine soil particles) that has been
deposited by settling from the air, washing from nearby areas in rain, or falling from vehicle
tires.  All   these  mechanisms  were  significant  in the Phoenix,  AZ, study  cited previously
(Richard,  1977; Richard and  Tan, 1977; Richard  et  al.,  1977a,b).   These emission sources and
related ones are discussed in some detail in Chapter 4.
     In an assessment  of particle source influence on TSP in western States, reentrained dust
from paved and unpaved roads accounted for 10 to 75 percent of TSP emissions and 16 to 49 per-
cent  of  TSP concentrations  at  critical  receptor  locations  in  20  inland western  cities
(Axetell,  1980).   In most of these  locations,  40 to  60  percent of TSP  emissions  were from
roadways.    One  critical  feature   of  this  observation,  apparently  mainly  derived  From
microinventory studies  around  TSP monitors, was the traffic volume on unpaved roads.  Cowherd
et al.  (1979)  reported  emission  factors  from unpaved roads  of about  300  grams/vehicle-km
traveled.    Consequently,  even  rather  small  traffic  volumes  can  generate substantial  TSP
contributions.  There  are very  large State-to-State variations in both  the  remaining number
and length of unpaved roads  and the  traffic  volume such roadways carry.   Carpenter and Weant
(1978)  analyzed  the available  information  on  roadway use  and found  that  while  unpaved roads
were  fairly  common in  the western mountains,  deserts,  and Great Plains, other areas of the
country,  the  Southeast  and New England especially,  also had many miles of  unpaved  roads and
substantial  roadway contributions  to TSP.   In industrial  areas,  truck traffic  over access
roads can be a major source of TSP emissions (Cowherd et al., 1979).
     The wind alone can be a major source when velocities are high and the soil  aggregates are
small.  For   example,  Gillette  (1978)  estimated  soil  fluxes  for six test  soils in  a wind
tunnel.   He  concluded  that windspeed and crust  play  major roles  in wind entrainment.  A sur-
face  crust  effectively  eliminated  fine-particle  entrainment  and  greatly  reduced  coarse
particle entrainment.
     Wilson et al.  (1979) found that car  and  truck traffic produced large amounts of dust on
unpaved mining roads  in northeastern Minnesota.  Particle sizes were mainly in the 6 to 30 urn
                                            5-106

-------
range near the  roadway,  but large particles were  found at about one-fifth the roadside level
400  or  500  meters downwind.  There was  visual  evidence of dust coating  roadside foliage and
gusts of wind  caused major short-term pulses in particle concentrations downwind of the road.
Davidson and Friedlander  (1978)  measured deposition of  coarse  particles  on Avena, the common
wild-oat grass of the Far West.   Dry deposition on the stems of such plants was reported to be
a significant removal mechanism for particles larger than about 7 urn.
     Reentrainment of  road  dust  has been found a  major source of coarse particles in central
business districts.   In a  study of  several  sites in Philadelphia,  Record  and Bradway (1978)
concluded that entrainment of dust  from roadways contributed  the majority  of street-level
coarse  particles   and very significant  levels  at  rooftops,   11  meters  above  the  street.
Rainfall, if  there was  enough  of  it,  reduced  the dust levels  significantly  (e.g.,  about 20
percent).  However,  attempts  to  flush the street  with  water  redistributed the fine particles
and  increased the observed coarse-particle level.
     In  a  study  of  one  site in  Massachusetts,  Record et  al.  (1979)  found  coarse-particle
levels  highly  correlated with traffic volume as  shown in Figure 5-45.   In this study,  very
large contributions  of roadway salt,  used for winter snow control, were  found  in the coarse
particles.
     Yocom et  al.  (1981)  estimated, by analysis of  a variety of records, the contribution of
fugitive  dust  to  areawide  particle  burdens  in  Allegheny  County,   PA.   They found  both
industrial  sources and roadways  to be significant contributors, though in widely varying pro-
portions.  In  the 12 study sites, roadway dust contributed  from a^low of 4 percent to a high
of  45  percent  of annual   geometric  mean TSP.   In  most  sites,  15 to  20 percent  of these
particles came  from  traffic.   Industrial fugitive particle emissions were more significant in
this area, although  the general  range was similar,  5 to 40 percent of the annual level.  In
most sites,  industrial  fugitive  dust contributed  20  to 30 percent of the TSP and was greater
than  roadway  dust.  These  two  sources,  together  with  the general  area background,  accounted
for  80  to 90 percent of the TSP burden in Allegheny County.
     Clearly,  fugitive dust  is  a major contributor  to TSP  in most  U.S.  monitoring sites.
There is  evidence that unpaved  roadways  and commercial streets can be major  sources.  Since
reentrained  dust  appears to  be  mainly coarse mode  particles,  monitor-siting  considerations,
and  especially  monitor height or slant distance  relative  to  roadways,  can markedly alter the
observed TSP level (Pace et al., 1977; Record et al., 1979).
     How important this fugitive  dust might be  in  assessing  exposure relative to potential
effects  is  not clear.   Of those cited  in this section, only the Record  et al.  (1979) study
reported particle size data;  further, the whole  issue  of  the relationship of outdoor concen-
trations to exposures  is under serious question.    (Vide  infra, §5.8)
     It is  also interesting to note  in  Table  5-22 that there is always some PM of biological
origin  in atmospheric particles.   Occasionally, especially during pollen season, this material
can  account  for  a significant fraction  of  the coarse-particle mass   (Draftz  et al., 1980).
                                            5-107

-------
CO
Ul
o
u.
o

DC
Ul
a
2

z
2000




1800





1600





1400





1200




1000





 800





 600




 400





 200




   0
              I      111      I     I      I     I      I
 I     I
                                                       TRAFFIC VOLUME
                              8     10    12    14    16



                              TIME OF DAY (START HOUR)
                                                      18
20
       100




       90





       80





       70





       60





       50





       40





       30





       20





       10





       0
22   24
                  Ul
                  o

                  I
Figure 5-46. Diurnal variation of particle concentrations and Plymouth Avenue traffic

volume in Fall River, Mass,, during March through June 1979 (weekdays only), shows

contribution from reentrained particles.,

Source:  Record et al. (1979).
                                      5-108

-------
Draftz et  al.  (1980)  also  identified cornstarch  as  a major contributor to  particle  mass in
some  nonattainment  areas.   These  authors  also  found  a  few  other  reproductive stages  of
organisms  such  as conidia  spores.   It has  been  hypothesized  that bacteria occur  in coarse
aerosols,  perhaps even   legionella  pneumophilia  (Eraser  and  McDade,  1979).   However,  the
presence of airborne infectious organisms has usually been deduced by disease incidence rather
than by  direct  measurement,  and quantitative techniques are largely lacking.  Considering the
importance of viable  particles in allergies, for  example,  this area could be a desirable one
for future research.
5.6.5  Summary
     The wind,  traffic,  construction, mining,  and general industrial activity  are  the major
causes of coarse particles suspended in the air.  Dry climates,  intense construction activity,
lack of  paving, and  salt from icy streets  or  the sea can all  be  contributing  factors.   The
quantitative assignment of particular sources to the coarse and  fine particle burdens has been
addressed  in a  cursory  and introductory  fashion in  this section.   In  the next part,  more
formal systems for this source apportionment will be addressed.
5.7  SOURCE-APPORTIONMENT OR SOURCE-RECEPTOR MODELS
     For  quite  a  long  time,  the  goal  of  quantitatively determining  the  contributions of
particle sources  to  ambient PM concentrations  has been  pursued.   Recently, Cooper and Watson
(1980) and Gordon (1980)  reviewed  the current status of calculational  systems or  models to
estimate  source  contributions to  atmospheric  particle  concentrations.   Cooper  and  Watson
described  several  methods  in  a  systematic  sense,  and  Figure 5-46  shows  their  analysis of
models  capable  of  yielding at least semi-quantitative  information.   Many microscope-based
conclusions were  discussed  in  the previous  section and  the work of Yocom (1981) is basically
an  example of series  analysis.   However,  chemical mass balance  and  multivariate  models have
been used  quite  effectively recently and a  few examples of these approaches are cited below.
The  results  from three  cities (St. Louis,  MO,  Denver,  CO, and  Portland,  OR)  illustrate the
contrast in fractional contributions of PM from different sources.
     In  analyzing  the  St.  Louis Regional Air Pollution  Study  (RAPS) dichotomous sampler data
by  x-ray  fluorescence,  Duubay and Stevens  (1975)   found 75  percent of  the  zinc,  sulfur,
bromine, arsenic, selenium, and lead occurred in the fine particles and at least 75 percent of
the  silicon, calcium,  titanium, and iron in the coarse fraction.   Using groupings of elements
to  represent sources,  Dzubay (1980) postulated the sources making fractional contributions to
2-month  summer  mean  concentrations  at several sampling locations in St.  Louis.   Approximately
50 to 70 percent of the concentration of fine particles was made up of ammonium sulfates.  The
next  largest identifiable  source  was motor vehicle  emissions,  followed by shale  and other
sources.
                                            5-109

-------
OPTICAL
    MICROSCOPIC
      METHODS
SEM
     AUTOMATED
        SEM
ENRICHMENT
  FACTORS
TIME SERIES
 ANALYSIS
                    CHEMICAL
                      MASS
                    BALANCE
SPECIAL SERIES
  ANALYSIS
                         ADVANCED
                     MULTIVAR1ATE DATA
                     ANALYSIS METHODS
            Figure 5-46. Types of receptor source apportionment models.

            Source: Cooper and Watson (1980).
                                 5-110

-------
     Pace and Meyer (1979) resolved the constituents found in the St.Louis dichotomous sampler
data to demonstrate the relative contributions of sources to the urban and nonurban concentra-
tions  (see  Figure 5-47).   As  might  be  expected,  the  vehicle emission  component  was  much
                                                                           2-
smaller  in  the rural  samples than it  was in  the  urban  samples.   The SO-   fraction in the
nonurban  sites  made up  a larger proportion  of  the total  fine particulate  mass  than did the
  2-
SO^  fraction  in the  urban  samples.    The  crustal  component  and  the  "nondescribed components
remained  about  the  same in  both sets  of  data.    Looking  at the coarse  fraction, it  is
interesting that  the crustal  component was much  larger  in the urban sites than it was in the
nonurban sites.
     The monthly averages of size fractionated Denver aerosol mass were compared for 2 months,
                                                2-
January  and May  1979,  by Dzubay (1980).  The SO^  component was smaller than in St. Louis and
the  motor vehicle  component was  larger  (see  Figure  5-48).   The  winter  concentrations  were
higher for both  the  fine and coarse  fraction.   Much  of this difference appeared to be in the
excess carbon  and NO.,  component.   The coarse  fraction contained road salt particles in the
winter.   Presumably,  accompanying sand could account for part of the  unidentified mass.  In
                                                              2-
Denver samples,  unlike  east.ern  aerosol  samples, the summer SO*  concentrations appeared to be
lower  than in  winter.   The asterisks  (*)  indicate  that some or all  of the component could be
due  to excess  carbon.    In this case "excess carbon"  was determined by the amount of carbon
observed  less  the amount accounted for by  the  identified  sources.   In the winter, the carbon
                                                  3         •
concentration  in the  fine  fraction  was  17  ug/m  .   It  was  not determined for  the coarse
                                                          3
fraction.  In  May,  the unfractionated carbon was 7.2 ug/m .   Wood burning and vehicle exhaust
are believed to be the important carbon sources in Denver.
     Chemical element  balance techniques  were applied to  TSP  and  fine-fraction aerosols col-
lected  in Portland,  OR,  in a year-long study.   Cooper et al.  (1979)  describe the experiment
and  results.   Figure 5-49 summarized  the resulting  source  allocation.   As  in  several  other
findings,  soils   and  road dust were  important  components of  TSP.   The  study  revealed  that
burning vegetation was the second most  important source, contributing almost-15 percent of the
TSP  mass  and  20 percent  of  the .fine-particle mass.   Sulfate  was  not  the most  abundant
component of the fine-particle mass.   In fact, it was only 8 percent of the mass, fourth after
auto exhaust,  volatilizable  carbon,  and  aerosols from vegetation burning.   The  study showed
the contribution of residential wood burning  to ambient aerosol concentrations.
     Studies  resolving  the   source  components  are great  aids in  resolving the  fractional
contribution of   local  versus  distant  sources.   Resolution of this  question may  await the
application of receptor modeling to other cities and other regions of the country.
     Source-receptor techniques,  which  rely on measurements of  PM and composition, can often
be combined with  studies  of wind direction or of synoptic air mass motion to further elucidate
source effects on ground pollutant levels.   For example, Rodhe et al.  (1972) studied the soot
       2-
and  SO,   concentrations  at  four coastal  sites  in Sweden.  By  combining  wind direction data
                                            5-111

-------
             •FINE
COARSE
(27 tig/m3)
                           RAPS URBAN SITES
                          (103,10S, 106,108,,112)


                   *2-MONTH AVERAGE CONCENTRATIONS
             FINE
                         RAPS NON-URBAN SITES
                          (115,118,120,122,124)
COARSE
(21 jiB/m3)
Figure 5-47. Source contributions at RAPS sites for July-August 1976
estimated by chemical element balance.

Source: Pace and Meyer (1979).
                                 •5-112

-------
      FINE FRACTION
COARSE FRACTION
MOTOR VEHICLES I (NH4»2SO4
                              T OM%
                               EFUSE 0,2%
                                LT1%
                              LIMESTONE
                                105%
                                ALE 22%
                       __(NH4)2S04  2.6%

                          NO3 OJ%
                          T 1%
                          RE FUSE 0.8%
                                                                    LIMESTONE 2.6%
        JANUARY
   JANUARY
    27 MD/m3
                  
-------
                                                 (VEGETATIVE BURN
                                                      (14.6%)
                           SOIL AND ROAD DUST
                                 [39.0%7
                                                     VOLATILIZABLE
                                                     CARBON (8.1%)
                                 RESIDUAL OIL
                                    (0.8%)
                                              MARINE
                                               (3.8%)
      SOIL AND ROAD DUST
            (43%)
  NONVOLATtZABLE CARBON
          (4.0%)
PRIMARY INDUSTRIAL (3.0%)

  • STEEL PRODUCTION (1.0%)
  • ALUMINUM PRODUCTION (0.72%)
  • HOG FUEL BOILERS (0.48%)
  « SULFIT1 PROCESS (0.39%)
                                       .UNIDENTIFIED (21.3%)
                                          {NH4, H2O, ate)
                                                                              TOTAL
                                                                        INONVOLATILIZABLE CARBON
                                                                                 <2.2%T
                                                                   (NH4,H2O),8tc)
RIMARY INDUSTRIAL (4.9%)
 * CARBIDE FURNACE, PROCESS EMISSIONS (2.0%)
 * ALUMINUM PRODUCTION (1.35%)
 • STEEL PRODUCTION (0 34%)
 * HOG FUEL BOILERS (0.22%)
 • SULFIDE PROCESS (0 18%)
 » SULFITE PROCESS (0.18%)
   FERROMANGANESE PRODUCTION
   CO.18%)
                                                                              FINE
           ESIDUAL OIL
             (1.4%)
                                      MARINE (3.2%)
        CARBIDE FURNACE, PROCESS EMISSIONS (0.6%)
Figure 5-49.   Aerosol source in downtown Portland, annual stratified arithmetic average. Does not
include the 17%, on the average, of material collected with the standard hi-vol sampler which was not
collected and characterized with the ERT-TSP sampler. Volatilizable and non-volatilizabie carbon are
operational definitions which approximately correspond to organic and elemental carbon, respectively.

Source: Cooper at al. (1979a).
                                                5-114

-------
                              2~
with  observations  of  high SO,   and soot  loadings,  these authors  deduced the  existence  of
long-range transport of air pollutants from industrialized areas of northern Europe,   Trijonis
(1980)  used  similar  analysis  in  detailing  local  source contributions  to  PM at  specific
monitoring sites in St.  Louis.
     Samson  (1981)  developed  a more  elaborate model  based mainly oh  the  trajectory analysis
procedure of  Heffter  et  al.  (1975).  This model was used to plot area source contributions to
               O—
ground-level  SO,  values  reported in the literature for seven rural  sites in the northeastern
United States.   Samson   concluded   that  high  concentrations  of  sulfates   appear   to  be
consistently   associated   with   upstream  (upwind)  stagnation.    Further  he  found  that
incorporation of the  magnitude  of S09 emissions and the upstream mixing height into the model
                                                     2~
did  not  improve the correlation  between measured  SO,  and upwind stagnation  as  estimated by
the inverse of windspeed.
     This class of  models appears to offer some promise for future applications, but like the
somewhat  more  complex  Lagrangian  trajectory  models discussed  in Chapter 6,  it has  some
problems.  Samson (1981)  pointed  out that a complete investigation  of sulfur budget (or that
for any  other pollutant,  for  that matter) would require measurements of the vertical profiles
of pollutants over  a  very wide area. "Clearly,  these data requirements are not  likely to be
met by experiments in the near future.  The  reader  is   directed  to  Chapter   6  for  further
discussion of deterministic models  and  their  data  requirements,  successes,  and shortcomings.
5.8  FACTORS  INFLUENCING EXPOSURE
5.8.1  Introduction
     To  this point only  outdoor concentrations  of S0?  and  PM  have  been considered  in the
present  discussion.   Outdoor  concentrations  are of major  concern  in estimating air pollution
effects  on  visibility,   and  ecological  and  materials  damage.   However,  people   spend the
majority  of their time inside  buildings or  other enclosures; they  breathe  mostly  indoor air
and,  therefore,  indoor concentrations  dominate average exposure.   To  the  extent that indoor
concentrations are  different  from the outdoors, population exposures are different from those
estimated by outdoor monitors.
     In  the United States the population is highly mobile.  Many persons in their daily travel
pass through  areas  of both high  and  low pollution levels within a  city.  Others work or play
outdoors  to  a greater degree than the general population.  Therefore, individual exposures to
SQy  and  PM  vary  more  widely than  measurements  from  stationary  outdoor monitors  suggest
(Spengler et  al.., 1979).
     Furthermore, individual variations  in respiratory anatomy, illness, or smoking habits can
exert  important  influences on  the  dose  of a  pollutant retained  by  individuals receiving the
same  exposure.   For  example,  Cohen et  al.  (1979) found that smokers  retain  test particles
longer  than  nonsmokers..   Figure  5-50 presents  the results of  a study of  12  subjects,  3 of
                                            5-115

-------
100
                   468

                   POST INHALATION, months
10
12
   Figure 5-50. Retention of Au1   labeled fe^O^ particles from
   human lungs; comparison of 9 non-smoking subjects with
   three smoker subjects.

   Source:  Cohen  et al. (19791.
                          5-116

-------
whom were  smokers.   Ten months following exposures  to  a known quantity of metallic dust, the
nonsmokers  had  cleared 85  to 95 percent  of the  dust  from their  lungs.   At  the  same time,
smokers  had retained about half  of  their  original dose   Unfortunately, there  are few other
studies that help in understanding these individual variations.
     In  this  section the two major  factors  influencing human  exposure to SO,  and  PM,  indoor
exposures  and activity  variations,  are presented  because they are  important in understanding
health effects.   First the systematic differences between indoor and outdoor concentrations of
SO- and  fine  and coarse particles are  discussed.   Then the limited evidence is presented for
varying exposures of individuals depending on their activities.
5-8-2  Indoor Concentrations of SulfurDioxide
     Indoor concentrations  of SO, are invariably  lower  than outdoors,  usually by a factor of
two (Spengler et a!.,  1979).   Since indoor sources of SO- (such as matches,  natural gas odor-
ants)  are  usually negligible,  virtually all SO,  indoors originated  outdoors.   Lower indoor
concentrations  are  commonly  attributed  to  SO-  removal   by  contact with  wall  coatings,
furniture,  flooring and carpets, air conditioning filters, and the like.
     Removal of  SO,  inside chambers and rooms has been shown to be a function of the material
present and the relative humidity (RH).  Cox and Penkett (1972) measured the decay rate of SO-
inside  containers.   Reaction  rates were  found to  be  first  order  in SO,,,  and irreversible
absorption  occurred  on  the  walls.   The  removal  rates  were  very sensitive  to the  RH.   As
relative humidity increased, so did SO, removal, approaching a maximum value at slightly above
80  percent RH  (Cox and  Penkett, 1972).   Spedding studied SO,  sorption by  indoor surfaces
(Spedding  and Rowlands,  1970; Spedding, 1970; Spedding  et  al.,  1971).   The  surface finish on
wallpaper  affected  sorption  rates  of SO-.   Conventional wallpaper showed better uptake than
polyvinyl  chloride (PVC)  wallpaper.   Hard woods sorbed SO, better and to a greater depth than
did softer woods.   Sulfur dioxide sorption was  also measured  for leather surfaces.  The rate
of absorption seemed to be influenced by the tanning process and the dyes used.
     Walsh  et al.  (1977) measured  sorption  of  SO, by typical  indoor  surfaces including wool
carpets, wallpaper,  and paint.   Absorption rates, as measured by deposition velocities, were
lower  for  carpets   with  an  acidic pH  than those  which  were either  neutral  or alkaline.
Sorption  of  SO, appeared  to  be  irreversible.   When  carpets  were  preexposed  to  an  SO,
                                                              3
concentration equivalent  to  27  years of exposure  at 30 yg/m , the amount  of  S02  uptake was
reduced  by a factor of three.   Fresh  emulsion  paints had the  highest  deposition velocity or
SO, absorption  rates,   and  vinyl wallpaper  had  the  lowest.   It was concluded that the most
effective  sorbing  materials   likely  to  be  present  inside   homes  are cellulose  wallpaper,
furniture  fabrics,   and  wool   carpets.   Therefore,  most studies report  lower  levels  of SOg
indoors than outdoors.
     Anderson (1972) reports that indoor SO- concentrations averaged 51 percent of the outdoor
concentrations over  a  7.5-month period of paired 24-hour sampling inside and outside a single
room.  The correlation  coefficient was only 0.52  (Anderson,  1972).   Biersteker et al. (1965)
                                            5-117

-------
analyzed  over 800 paired samples  from  the living rooms and exteriors  of 60 Rotterdam homes.
Indoor  SO*  levels averaged  20 percent of the outdoor levels and were  lower  for newer homes
than for  older homes.   This may  imply  longer air turnover times in  the newer homes and/or a
"fresher"  surface  area  for  S02  absorption  (Biersteker  etal.,  1965).   Weatherly  (1966)
measured  S02  and smoke levels inside and  outside a building in central  London in early 1960.
Indoor  SO-  levels were  always lower than the corresponding ambient  conditions,  averaging 40
percent.  Spengler et al. (1979) reported  on paired SO, monitoring inside and outside at least
                                                                                             3
10  homes  in  each  of 6  cities.   Figure 5-51  displays the annual S02  concentration in pg/m
averaged  across  each community's indoor/outdoor  network of monitors (May 1977 to April 1978).
The  cities  were:   Portage,  WI;  Topeka, KS;   Kingston  and  Harriman, TN;  Watertown,  MA;
St. Louis,  MO;   and  Steubenville,   OH.    Where  ambient  levels   were  high,   the  indoor
concentrations were  30  to 50 percent of the  ambient levels.   In Kingston, many of the indoor
levels were less than  the minimal detectable  level  and were averaged in as zeros.   This  was
not done  for  Portage  and Topeka, where  ambient  levels  were  very low;  hence  the indoor  SOj,
levels  in these  cities  appear to be reduced by only 20 percent of the outdoor concentrations.
     The  seasonal indoor/outdoor pattern for each city depends on the S02 sources in each city
and the use of air conditioning.  These differences can be seen by comparing the monthly mean
indoor and outdoor S02 concentrations for Watertown and Steubenville,  as shown in Figures 5-52
and 5-53.  In  Watertown, SO, was  primarily derived  from  sulfur-laden fuels.   The ambient
levels rose in the winter as more residual and distillate oil was used for space heating.  The
indoor/ outdoor  ratios became small because  homes were sealed more  tightly.   In the summer,
ambient levels decreased,  but the indoor/outdoor  ratio approached unity because of increased
ventilation.   In Steubenville,  the  summer  SO™  levels were  not substantially  reduced from
winter  values,  since residential and commercial  space heating was not  the  primary  source of
SO, in  this  area.   Yet the  reduced indoor levels continued since more air conditioning  was
used in Steubenville (50 percent of homes  sampled).   Even  in homes  without air conditioning,
summer levels were reduced by 30 to 40 percent.
     While most  information  supports the  idea of lower indoor S02 concentrations, exceptions
are known.  Yocom  et al. (1971) found  one home,  heated by a leaky coal furnace, in which  the
indoor SO,  level  was periodically 10 times the outdoor level.   Bierstecker et al. (1965) also
found leaking flue gas contributions  indoors.
     However,   the principal  body of  evidence suggests that  indoor  exposures  are  generally
about half  that  found  outdoors.   Consequently,  highest exposure  levels  are  likely  to be
incurred by people who spend time outdoors near local S02 sources.
5.8.3  Particle Exposures Indoors
5.8.3.1   Introduction—Available data  on  indoor particle  levels  were  collected by  a wide
variety of  measurement  procedures  ranging  from  dustfall  to  condensation nuclei  counting
Earlier in this  chapter and in Chapter 3,  it was noted that the various particle measurement
                                            5-118

-------
    60
    SO
    40
 I,
 1,30
 O
    20
    10;
|§%f OUTDOOR
j [ INDOOR
»p <0.05
~ m
WA I ifM^-. m
vm I ^h~i w "~i
1






1
M^HMI




i
i
I


=v. ._

MMPMI

         PORTAGE*  TOP1KA'    KING-
                                STON*
WATER-   ST. LOUIS*  STEUBEN-
TOWN*               VILLE*
Figure 5-51.  Annual sulfur dioxide concentrations averaged across each community's
indoor and outdoor network (May 1977 — April 1978).

Source: Spengler et al., 1979.
                                    5-119

-------
"i
    10b

     96

     84'

     72

     60


     48

     36

     24


     12

      0
             WATERTOWN
                                                               O INDOOR

                                                               O OUTDOOR
             NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP QCT NOV DEC JAN FEB MAR APR

                1976                         1977                            1978


Figure 5-52.   Monthly mean SO? concentrations averaged across Water-town's indoor and outdoor
figure o-o*:.   mommy mean auo concei
network (November 1976 - Aprir1978).

Source; Spongier et al. (1980).
                                            5-120

-------
 i
  &
 i
108




 96



 84



 72




 60



 48




 36



 24



 12



  0
                I    I    i    I


             STEUBENVILLE
              NOV DEC JAN FEBMAR APR MAY JUN JUL AUG SEP OCT NOV DEC JAN FEB MAR APR



                1976                        1977                            1978




Figure 5-53.   Monthly mean SOo concentrations averaged across Steubenville's indoor "and outdoor

network (November 1976 —April! 978).



Source: Spengler (1980).
                                             5-121

-------
procedures  have definite  particle  size biases.   For example, dustfall  and  TSP mass measure-
ments  are dominated by  coarse  particles,  while light scattering  and  nuclei  counting,  on the
one  hand, and  smokeshade  and coefficient  of  haze  on the other,  are  better  measures of fine
particles and participate  carbon mass (one  fine particle component), respectively.
     Evidence  from these  techniques  has   produced  a consistent  view of indoor  and outdoor
particle  concentrations  and  sources,  and this view  is  presented  below  by  separate  con-
sideration  of  coarse-  and fine-particle studies.  Table  5-23 summarizes available studies by
particle measurement technique.
5.8.3.2   Coarse-Particle Concentrations Indoors—Particles  larger than  5  or  10 urn  tend to
settle from the air,  and  two studies, using dustfall collection techniques suggest that these
particles  are greatly  reduced  indoors.  Whitby  et al.  (1957)  studied  dustfall  in offices,
laboratories,  and  homes.  Average  indoqr  dustfall   was  only 15 to  20 percent  of the outdoor
level.   No  significant  differences  were  found  among  residential  or business  locations.
Schaefer  et al.  (1972) found indoor dustfall about  one-eighth outdoor values in a study of 30
residential  sites  in  four cities.   There  was  little correlation between  indoor and outdoor
levels.   Dustfall was higher in homes where windows  were open.
     Yocom  et  al.   (1971) studied  TSP  in  public  buildings,  offices,  and  homes using  a
scaled-down  version of the hi-vol sampler.  As mentioned previously,  the mass of such filter
samples contains both coarse- and fine-particle fractions.   Indoor levels were about half out-
door levels  on  the average.   In  summer and fall,  private homes had almost the same interior
daytime TSP  values  as  those found outside  although  night interior levels were much lower than
outside.    In  the   same  study,  indoor/outdoor  ratios   in  air-conditioned  office  buildings
differed  seasonally.   In summer and winter, indoor  TSP was  about  half the outdoor values, but
in  fall,  when  increased  volumes of  outdoor air were  used  in  air conditioning,  indoor and
outdoor TSP values were about equal.
     Yocom  et  al.  (1970) also obtained some cascade impactor size distributions of indoor and
outdoor particles  in  six structures in the Hartford, Conn., area  in the fall  and winter.   The
mass of particles  larger than 2,5 pm was  always  greater outdoors than  indoors.  However, the
mass of sub-2.5 fjm particles was mostly greater indoors than outdoors, and the indoor/outdoor
ratio varied  from  0.63 to 2.6.   In a subsequent report, Yocom et al.  (1971) reported substan-
tially increased organic particle levels  indoors, confirming similar  findings by Goldwater et
al.  (1961).   This  result was attributed to smoking  and  cooking,  indoor activities that could
also increase fine-particle mass.
     Alzona  et «-al.  (1979)  reported  elemental  analyses  for  calcium and   iron,  normally
coarse-particle components, and for zinc,  lead, and  bromine, components of fine particles.  In
these studies,  an  experimental  room  was cleared of airborne PM and  then allowed  to come to
equilibrium under controlled penetration of outdoor  ambient air.   Experiments were carried out
with windows  "cracked"  open  and wide open  and  with windows artd/or  room surfaces covered with
plastic sheets.  Filter samples drawn throughout the experiment were analyzed by X-ray
                                            5-122

-------
TABLE 5-23.   SUMMARY OF INDOOR/OUTDOOR  (I/O)  PM MONITORING STUDIES BY METHOD

Method
Dustfall3






Total sus-.
pended PMD

n
-t
o
o



Smoke







Author
Whitby'et al.
(1957).
Schaefer et
al. (1972)



Yocom et al.
(1971)







Whitby et al.
(1957)


Goldwater
et al.
(1961)

Location
Minneapolis

Chicago,
Washington,
Atlanta,
Austin-San
Antonio
Hartford, CT








Minneapolis

Louisville

New York



Building
type
Residential
Lab & office
Residential




Public


Office


Residential


Lab & office
Residential
Lab & office
Residential
Lab & office

Residential

Number
of
sites
12
30





2


2


2


__
_-
__
--
12

18

Month
or
season
Annual
Annual
Mar.-
Aug.



Summer
Fall
Winter
Summer
Fall
Wi nter
Summer
Fall
Wi nter
__
__
__
--
Feb.-
Mar.
Feb,-
Mar,
Sampling
period
_«
— ~
—




12 hr
12 hr
12 hr
12 hr
12 hr
12 hr
12 hr
12 hr
12 hr
__
--
__
--
__

__

Mean
Indoor
0.54
0.47
0.44




59
58
63
53
31
42
62
53
43
42
46
101
121
149

164

Outdoor
2.86
2.86
3.48




111
119
277
108
50
122
67
77
97
74
74
124
124
263

179

Number
of I/O
samples ratio
0.19
0.16
26 0.12




0.53
0.49
0.23
0.49
0.60
0.35
0.94
0.69
0.44
0.57
0.62
0.81
0.98
12 0.57

18 0. 91


-------
TABLE 5-23 (continued)

Method Author
Jacobs et al.
(1962}
Weatherly
(1966)
Biersteker
et al.
(1965)
Berdyev
et al.
(1967)

1 Anderson
£ (1972)
Respirable . Binder et al.
particless0 (1976)



Particle Parvis
counts0 (1952)
Ishido
et al.
(1956)

Location
New York

London

Rotterdam


Dushambee

U.S.S.R.

Arhus,
Denmark
Ansonia, CT




Italy

Osaka



Building
type
Residential

Office

Residential


Residential
1st floor
Residential
2nd floor
Classroom

Smoking
homes
Non-
smoking
homes
Residential

Apartment



Number
of
sites
17
— «.
1

60


1

1

1

11

9


5

1



Month
or
season
Apr.-
May
Jan.-
Mar.
Winter


Summer

Summer

Sept. -
Apr.
Sept. -
Dec.
Sept.-
Dec.

—
— •*
March
May
June
Nov.
Sampling
period
-_

1 hr

24 hr




—

24 hr

24 hr

24 hr


—
~—
24 hr
24 hr
24 hr
24 hr
Mean
Indoor
239

195

153


1270

660

27

132

93


45.7
1000
1287
978
738
--
Outdoor
226

205

184


960

960

34

58

58


97.6
1036
1528
1047
752
1897
Number
of
samples
17



800


8

9

150

11

9


__
"""*
—
—
—
-—
I/O
ratio
1.06

0.95

0.84


1.32

0.60

0.81

2.28

1.60


0.47 CN
0.97 PC
0.84
0.91
0.98
—

-------
                                                      TABLE  5-23 (continued)
 1
I—«
r\>
cji

Method Author Location
Ishido Osaka
(1959)


Jacobs New York
et al.
(1962)
Megaw England
(1962)
Lefcoe and
Inculet (1975)
Coefficient Whitby et al. Minneapolis
of haze9 fl9^7)
Louisville

Carey et al. Cincinnati
(1958)
Shephard Cincinnati
(1959)










Number
Building of
type sites
Apartment
Residential
Hospital
School
Office & lab
Homes

Test building

Homes
(AC & ESP)
Residential
Lab & office
Residential
Lab & office
Residential

Residential











1
1
1
I
12
18

1

2

—
—
--
— -
__

9
9
9
9
—
—
—
—
9
9
9
9
Month
or Sampling
season period
	 __
—
—
— — ~
—
—

—

Annual
—
—
—
--
--
Oct.-
Dec.
Jan.
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Mean
Indoor
706
662
1611
2382
424
705

—

510d
156
0.6
1.0
2.4
2.8
2.1

2.2
2.0
1.6
1.6
--
0.8
0.8
1.3
0.8
1.5
1.7
1.8
Outdoor
619
678
1595
2346
481
472

—

227f
•59e
1.05
1.05
2.6
2.6
3.8

2.7
2.3
1.8
1.7
1.0
0.9
0.8
0.8
0.5
0.9
1.1
1.5
Number
of I/O
samples ratio
1. 14
0.98
1.01
1.02
0.88
1.49

5 0.66

1.46
2.60
0.57
0.95
0.92
1. 06
0.55

0.86
0.89
0.85
0.94
-— -~
0.88
1.00
1.63
0.80
1.15
0.89
1.06

-------
                                                   TABLE 5-23 (continued)
ro
en

Number
Building of
Method Author Location type sites
Yocom et al. Hartford, Public
(1971) CT

Office


Homes


Piezoelectric Repace and Metropolitan Public
microbalance Lowrey (1980) Washington, (no smoking)
DC
Public
(smoking)

2


2


2


3


13


Month
or
season
Summer
Fall
Winter
Summer
Fall
Winter
Summer
Fall
Winter
March-
June

March-
June

Sampling Mean
period Indoor
0.32
0,33
0.36
0.29
0.20
0.37
0.41
0.28
0. 32
20 min. 29
to,:
55r
2 min (I) 86
5 min (0) to-
697T
Outdoor
0.36
0.34
0.51
0.41
0.26
0.54
0.42
0.30
0.39
40
to.
55T
22
to.
63T
Number
of
samples
__.
—
-—
-_
--
-_
-_
—
--
3


4-25 (I)
13 (0)

I/O
ratio
0.90
0.97
0.69
0.70
0.78
0.88
0.98
0.93
0.82
0.66
to
1.38
1.56
to
11.62

   ?Measured  as  g/mVmonth.
    Measured  as  pg/m3.
^Measured as number/cm3.
 Particles >0.3 urn.
fParticles >0.5 pm.
 Particles >1.0 pm.
Measured as COH/1000 linear ft.
   Note:   1 ppm  S02 = 2620  ug/m3.
          1 ppm  N02 = 1885  ug/m3.

-------
excitation for elements  of  known outdoor origin (Fe, Zn,  Pb, -Br,  Ca).   Within several  hours,
equilibrium was reached  in  which the indoor/outdoor ratio was  typically 0.3 (see Tables 5-24
anti 5-25).   On the basis of  the indoor/outdoor element ratios, they conclude  that remaining
indoors with doors  and  windows partially closed reduces outdoor dust  exposure by two-thirds.
The indoor/outdoor  ratios for  the  coarse-particle components,  calcium and  iron,  were lower
than  for  the  fine-particle  components  zinc,  lead  and bromine.   Therefore,  it  appears  that
tracer components of  coarse  particles do not penetrate any  of  these structures as readily as
the fine components.
5.8.3.3  Fine Particles Indoors—In addition to the cascade impactor studies mentioned earlier
in conjunction with the  coarse-particle discussion, there have been several recent reports of
sub-3.5 |jm particle mass measurements indoors and outdoors.
     Repace  and  Lowrey   (1980)  reported  even  larger  indoor/outdoor  contrasts, which  they
attribute to smokers.   Using a piezoelectric microbalance,  they sampled fine particles (0.01
to 3  jjm)  inside  and  outside  a  variety of  public  places,  many of  them restaurants.   In the
absence of smokers, the indoor/outdoor ratios were in the range of 1:1,  comparable with ratios
reported  by  investigators using  other  methods.  With smokers  present,  indoor/outdoor ratios
ranged to over 11:1 (Table 5-23).
     Binder  et al.  (1976)  used hi-vol  air  samplers outdoors  and  personal  samplers equipped
with a 3.5 urn  cutoff device.  The personal monitors were carried by school children who spent
60  to 80  percent  of their  time indoors.   In homes  where there  were smokers,  the  indoor
fine-particle mass was almost twice the outdoor TSP.
     Spengler  and Dockery have  measured  indoor and outdoor levels  of  sub-3.5  urn particulate
mass using cyclone-equipped  membrane filter samplers in the same  six  cities noted in Section
5.8.2  (Spengler  et al.,  1981; Dockery and  Spengler,  1981;  and Dockery,  1979).   Figure 5-54
presents  annual   average values for all  sites  in the  six cities.    In  all  cities  except
Steubenville, Ohio, the indoor fine-particle level  was higher than outdoors.  Steubenville, an
industrialized  community, also  had  the  highest  annual  average  outdoor level.   Table  5-26
presents  arithmetic averages  for all homes  in  this  study  stratified by numbers of smokers in
the household (Dockery, 1979).  It is apparent that in the absence of smoking, indoor and out-
door  levels  of fine-particle mass  are  almost the  same.   However, smoking  contributes  very
significantly  to  indoor  levels.   Dockery (1979) calculated that  a one pack/day-smoker con-
tributes  about 18 ug/m  to inside fine-particle mass,  and  this level  is increased by the use
of air conditioning, presumably because of recirculation, to 43 pg/m .
                                            5-127

-------
                  TABLE 5-24.  MEASUREMENTS IN PRINCIPAL ROOM OF STUDY3
Case
J
K
L
M
N
P
Number
of runs
3
2
1
2
6
3
Conditions
Normal
Plastic over windows
Window wide open
Window cracked open
All surfaces plastic covered
All but windows plastic covered

Ca
0.10
0.10
0.52
0.20
0.02
0.10

Fe
0.17
0.15
0.81
0.16
0.12
0.15
I/O
Zn
0.52
0.71
0.93
0.69
0.24
0.58

Pb
0.49
0.17
1.2
0.55
0.15
0.57

Br
0.33
0.17
1.0
0.53
0.20
0.32
    Source:  Alzona et al. (1979).



     The study room in all these c;

     listed below in Table 5-25 as case J.
a                                                   2
 The study room in all these cases was the same 12 m , old university building room
                TABLE 5-25.  MEASUREMENTS IN VARIOUS CLOSED ROOMS

Number
Number of
Case of runs Type of room windows Ca
A
B
C
D
E
F
G
H
I
J
Average
1
1
1
1
1
1
1
2
2
3

10 m2,
50 m2,
30 m2,
20 m2,
20 m2,
new univ.
old univ.
bedroom,
bldg.
bldg.
tight home
attic, tight home
bedroom,
leaky home
Chevrolet Vega
Datsun
20 m2,
20 m2,
»| *} m*"*
(except A)
440
old chem.
new univ.
old univ.


lab
bldg.
bldg.

0
2
8
2
2 0.05
6
6 ""*""
(3) 0.08
3 0.15
(3) 0.10
0.10
Fe
<0
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
.1
33
27
10
33
27
09
13
54
17
24
I/O

Zn Pb

-------
           TABLE 5-26.   RESPIRABLE PARTICULATE CONCENTRATIONS OUTDOORS AND INDOORS
                                    BY AMOUNT OF SMOKING3

Location
Outdoor
Indoor, no smokers
Indoor, 1 smoker
Indoor, 2+ smokers
Number of
homes
74
38
22
9
Number of
samples
2S98
1328
712
323
Mean concentra-
tion, ug/m3
22.3
24.0
42.8
74.5
Standard
deviation of
home means
12.7
11.4
22.2
37.9

       Data averaged across network of samples in six communities for 1977.
      Source:  Dockery (1979).
                                                         2-
     Spengler et al,  (1981)  also reported that indoor SO.  (a fine-particle component) levels
were significantly  lower  than  outdoor levels, but that gas cooking stoves can increase levels
by  about  1 pg/ra   (Dockery  and  Spengler,   1981).   Yocom  et  al.   (1971)  found  that  lead
indoor/outdoor  ratios were  greater  than  TSP  ratios.   Taken  with  the  previously  mentioned
Alzona et  al.  (1979)  study,  it appears that  most fine  particle components analyzed  are found
in high proportion indoors.
     A  number  of   studies  have  reported  indoor  measurements  of  smokeshade  or  COM,  both
estimates  of  fine-particle  carbon.   Whitby  et  al.   (1957),  Shephard  et  al.  (1958),  and
Weatherly  (1966) all  found  smokeshade values inside and outside buildings to be nearly equal.
Goldwater  et al.  (1961)  found  indoor smoke about 75  percent of the outside  levels  in 30 New
York sites,  but that difference  was  not significant.   Jacobs (1962)  found no indoor/outdoor
differences  in a  followup  New  York study.   Anderson  (1972) found  nearly equal and highly
correlated  smoke values  in  a classroom and outdoors in Denmark.   Biersteker et al.  (1965), on
the other  hand,  found no correlation between indoor and outdoor smoke in a winter study of 60
homes in Rotterdam.   Whitby  et al. (1957) and  Yocom  et al.  (1971) both  reported that indoor
COH values were much  closer  to  outdoors  than either dustfall or TSP.   Apparently,  the fine
carbon particles measured by these techniques effectively penetrate buildings.
     A  similar  conclusion  is  reached  in  indoor/outdoor  light-scattering  studies.   Since
scattering  of  visible light is  caused by  the range of particles from about 0.2 to  1.0 pm,
measurements  using  this  technique  provide  another index  of fine-particle  mass.  Indoor and
outdoor  light-scattering values  were found  to  be the same and  highly correlated  in Japan
(Ishido, 1959;  Ishido et al.,  1956), in Italy (Parvis,  1952; Romagnol, 1961), and in New York
(Jacobs et  al., 1962).

                                            5-129

-------
\
    50
    40
    30
    20
     10


—
|||{| OUTDOOR
\ \ INDOOR
«p 
-------
     Therefore, fine particles  readily  penetrate buildings and occur inside to about the same
extent as  outdoors.   Indoor  activity  adds incrementally  to outdoor  levels  and,  frequently,
somewhat higher levels  of  fine particles are observed  indoors.   Smoking adds very materially
to indoor levels.
5,8.4  Monitoring and Estimation of PersonalExposures
     In  previous   sections  of  this  chapter,  the  spatial  and  temporal  variations in  the
concentrations of  S0?  and of fine  and  coarse particles and their  components  were summarized
for both outdoor and indoor exposures.   However, looking forward to health effects summarized
in Chapters  11 to  14,  there is still  one element of exposure remaining  for  discussion.   In
addition to  the particle  concentrations  measured  by  long integrating-time  monitors  (i.e.,
long-term  doses  of  pollutants),   people  are  exposed to short-term  high  concentrations.
Unfortunately,  sufficient  data  do  not  exist  to  establish  the  relative  importance  of
concentration  and  time of  exposure.   There  is,  however,  evidence (cited  in  Chapter 12)  for
gaseous S0« and particles that long-term exposures can cause adverse health effects.   There is
also  evidence that  a   short  burst of  pollutant exposure  can  cause  adverse  health  effects.
Therefore,  these peaks in exposure are likely to be important,  and there is some evidence that
peaks occur both indoors and outdoors.   For example, earlier in this chapter it was noted that
high  levels  of S0?  occur  periodically close  to intense  sources.  Obviously,  people passing
through such an area, even though they are not resident there,  receive a high short-term dose.
On roadways,  particle  concentrations tend to be  very high because  of resuspension  of road
dust.  Clearly,  travelers  experience such concentrations  at  least for the time  they  are in
traffic.   As pointed out earlier in this section, an individual's daily activities, the places
visited, and activities in the home all  play a role in that person's exposure.
     For example,  Repace et al. (1980) followed an individual's daily exposure with a portable
particle monitor and correlated these  measurements with activities.   Figure  5-55 shows time
series  plots  of  PM concentrations  to which James  Repace,  the principal  investigator,  was
exposed on October  16,  1979.   Sharp peaks  were  evident in traffic, indoor smoking areas,  and
his  own  home,  particularly  in  the  kitchen.   Obviously,  controlling outdoor  air  pollutant
levels would have little influence on his exposure to short-term doses of particles except for
those  incurred on  roadways.  There have  been other recent reports of  statistical studies of
the relationships among personal, indoor, and outdoor particle concentrations.
     In a personal  monitoring study designed to test the relationships between outdoor concen-
trations and  personal  exposures  and  to  estimate activity  concentrations,  Spengler et  al.
(1980) collected 12-hour  respirable particle samples for 15 days on 45 individuals in Topeka,
Kans.   Particle  concentrations experienced  by  monitored  individuals  were 2.5  times greater
than average outdoor levels for this time interval.  Further, there was no correlation between
the outdoor  level and the personal exposure of individuals.  Variation in outdoor concentrations.
                                            5-131

-------
Ul

M
OJ
o
z
Hi
O
o
u
280

260

240

220

200

180

160

140

120

100

 80

 60

 40

 20
             I    I    I    I    I   1    I    I    I    I    I    I    !    I   I   II    1    1    I    !   !    !
                    • INDOORS

                    • IN TRANSIT

                    O OUTDOORS
                                  CAFiTERIA, SMOKING SECTOR
                              BEHIND SMOKY DIESEL TRUCK
                                COMMUTING I

                                BEDROOMjqf

                                        STREET SUBURBS, OUTDOOR ,
                                                                          WELL-VENTILATED KITCHEN
                                     OUTSIDE CIGAR
                                     SMOKER'S OFFICE

                                     CAFETERIA, NONSMOKING
                                     SECTION
                                                                              SIDEWALK
                                                                  SUBURBS IN   BUS EXHAUST
                                                                  VEHICLE CITY
                                                                           OFFICE
                                           LIBRARY, UNOCCUPIED CAFETERIA
                                                                     CITY, OUTDOOR
                                                         COMMUTING
                                                                SUBURBS

                                                               JOGGING
                                                            LIVING
                                                            ROOM
                                                                                                             LIVING'
                                                                                                             ROOM
                                 I    I    I    I    I    I    I    I    I   I    I    I    I    I    |    |   I    I
        12  1
    MIDNIGHT
            234
567
   A.M.
                                                     9  10   11   12  1
                                                              NOON

                                                           TIME OF DAY
3   4
567

1   P.M.
8    9  10  11  12
                             Figure 5-55. An example of personal exposure to respirable particles.

                             Source:  Repace et al. (1980).

-------
explained only  4 percent of  the  variability in personal concentrations  experienced.   On the
other  hand,  indoor respirable particle  levels  explained a  fair amount of  the  variability in
personal exposure; for men,  indoor levels accounted for 25 percent of their exposure variation
while  for women  50 percent  was explained by this variation.   It appears that somewhere in the
daily  activities of  these  individuals,  their  exposure was  substantially greater  than  that
measured  by  outdoor  monitors and,  further,  ttie  variation  in  exposure  is  not  measured
adequately  by  fixed  indoor  monitors  either.   Passive  smoke  exposure  accounted  for  a
significant portion of  the  increment above outdoor levels.   Figure 5-56  plots the histograms
of  concentrations  for both volunteers who  reported  no passive smoke exposure  during  the day
and  those  who   reported  some  exposure  to  passive   smoke.   The  means   were  20  ug/m   for
                                       3
nonsmoke-exposed samples versus 40 ug/m  for smoke-exposed samples.
     In  Figure   5-57,  the daily  mean  concentrations  for all  outdoor,  indoor, and personal
samples are presented.  There is  the suggestion that  the  variation in outdoor concentrations
causes  variations  in  indoor  and personal concentrations.  However,  variations  in  indoor con-
centration cause considerable variance in individual  exposures.
     As an  alternative  to direct  measurement (monitoring), typical  personal  exposures may be
estimated on  the basis  of information on indoor and outdoor concentrations and human activity
patterns.
     The exposure to particles and gases that one experiences will be ultimately determined by
location  and  activity.    Certainly,  locational  and activity patterns are  very  complex in our
society.  They  are  functions  of age, sex, social, economic, and educational factors.  While a
limited data base exists on activity patterns within our population and on the distribution of
smokers,  housing stock,  and  various  other building  factors, an  exhaustive discussion is not
appropriate for this document.
     Time budget studies of the U.S. population indicate that on the average, 90 percent of an
individual's  time  is  spent  indoors.  Between 5 and 10 percent of the time  is spent in transit
in  a  vehicle.    Considering   these  figures,  the indoor  environment  is  very important in
determining the time-weighted average exposure.
     However, the time-weighted average is only one measure of pollution exposure.   Time spent
outdoors  is variable.   The  time outdoors varies by the time of day and year, among regions of
the country and among different categories of people.   Therefore, in regard to the concern for
indoor  pollution,  the  fact  that  short-term  peak  ambient concentrations  may be an important
component of exposure should be remembered.
     Much work  remains to  be done  on  personal exposures  to gases and  particles.   Based on
current understanding, the following qualitative statements can be made:
     1.   Depending on  spatial  gradients in ambient air, personal  exposures to SO- should be
less than the outdoor concentrations.
                                            5-133

-------
  20

  18
„  16
e
§
I  14
Z
2  12
§
         1
                                                               NONSMOKING EXPOSED
                 10   15  20  25  30  35  40  45  50   55   60   65  70  75  80  85   90  95

                         NORMALIZED MEAN FINE PARTICLE CONCENTRATION
                                          (<3.5 /urn)
14
I 12
fc*
1 10
o
P 8
5
3
0. 6
O
D.

1
g 2
o
—
—
*"""""* ••••»

::::::=K


::::::::1

lmm=|
nfll 1
















MM*
















































^mm^ .
















F^—>


















•"^
^^"
fln[
SMOKING EXPOSED









nlln,-,^
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90
—
—
—




„,„,,„



95
                        NORMALIZED MEAN FINE PARTICLE CONCENTRATION
                                          (<3.5fJm)

Figure 5-56* -Normalized distribution of personal (12-hour) exposure samples
exposed and smoke exposed samples.
 Source:  Spengler and Tosteson
                                                                           ) for non-smoke
                                          5-134

-------
Z
o
111
o

o
o
LU

O
p
DC
<
a.
LU

m
<
cc
Z
sa
ui
CC
   40
   35
   30
25
20
15
10
   n     m
         O PERSONAL
         Q INDOOR
         A OUTDOOR
                                                               I   I	I
      Th  Sa   Tu  Th Sa   Tu Th Sa   Tu Th Sa   Tu Th  Sa   Tu Th  Sa    Tu


      WEEK1     WEEK 2     WEEKS     WEEK 4      WEEK 5     WEEK 6   WEEK?


        DAILY AVERAGE CONCENTRATIONS FOR THE ENTIRE GROUP OF 46 SUBJECTS

                               IN THE TOPEKA STUDY
    Figure 5-57. Daily mean indoor/outdoor and personal concentrations
    of respirable particles. Daily means averaged over 24 homes and outdoor
    locations and up to 46 personal samples. Samples collected during May and
    June 1979.


    Source: Spengleretal. (1980).
                                     5-135

-------
     2.   Depending on activity times and building characteristics, longer term exposure could
be less than half the ambient concentrations.
     3.   For  estimates  of  personal exposure  to  particle  mass  concentration,  the  ambient
measurement appears  to be  a poor predictor.   While ambient concentrations exert  an  effect,
personal activities and  indoor concentrations cause personal exposures to vary substantially.
     4.   Tobacco smoke is an important contributor to indoor and personal exposures,
     5.   Personal exposures to the components of suspended PM of outdoor origin and contained
in the micrometer and submicrometer size  fraction may be estimated  by ambient measurements.
The smaller size  particles  of toxic trace elements (V, Cd, Ni,  Br, Se, etc.) and some organic
                            o—    -
and inorganic  compounds  (SOt  , NO,), which  are exclusively of outdoor  origin,  penetrate the
indoor  environment  in a  predictable way.   Outdoor  measurements of  primary and  secondary
fine-fraction  aerosols  in   nonindustrialized   communities   may  prove  to  be  adequate  to
characterize population exposures and trends.  This last statement assumes no important indoor
sources  for  this typical  outdoor component.  This question  certainly  needs  verification and
quantification in field studies.
5.9  SUMMARY OF ENVIRONMENTAL CONCENTRATIONS AND EXPOSURE
     The purpose  of  this  chapter is to document  the existing concentrations of SO  and PM in
the environment.   Since the damage caused by these pollutants to man, other living things, and
valuable objects  varies with time,  place, and other circumstances, a wide variety of exposure
conditions are relevant for these pollutants.
     Sulfur oxide concentrations  in  the air have been markedly reduced over the past 15 years
because  of fuel  sulfur  restrictions,  control   technology  implementation  on major  sources,
redistribution of power plants to regions outside cities, and the use of taller stacks.  There
are still  some areas  with very high  S09  concentrations,  though,  and hourly values of 4000 to
         •i                             £•
6000 ug/m  are rather common near large smelters.  In about 100 U.S. locations, maximum hourly
                       o
values  above 1000 ug/m  are occasionally found,  but  much of the  nation  is  basically  in com-
pliance with the current NAAQS for SQ2.
     After a downward  trend from 1970 to 1974,  total suspended PM concentrations have changed
very   little   in  recent  years  despite  major  reductions  in  stationary  source  emissions
inventories.   Dusty  arid  regions of  the country  still  have  high TSP as  do industrialized
cities  in  the  East and Far West.  Ninetieth-percentile values of 24-hour TSP (the values that
                                                        o
are exceeded 10  percent  of the time) are  above 85 ug/m   in  every part of the country except
                                                          Q                           O
Alaska.  Regional mean TSP values range from about 50 ug/m  in EPA Region I to 77 ug/m  in EPA
Region  IX.
     Ambient airborne  particles  exist in two distinct size ranges, fine particles below about
1  urn,  and coarse  particles above about 3 urn.   Rather little mass  is  in intermediate sizes.
Except  that  both sizes  are captured on  filters, the  two kinds  have  very  little  in common.
                                            5-136

-------
Fine and  coarse particles  differ in origin, chemical  composition,  geographical  distribution
and physical behavior.
     Fine  particles  are  composed mainly  of  (1)  sulfate,  nitrate, and  ammonium  ions,  (2)
organic substances from atmospheric photochemical conversions, and (3) carbon, organic matter,
and metallic components  directly emitted from combustion sources.  Sulfate, most often in the
form of  neutral ammonium  sulfate,  but  sometimes  in association  with  acidity as  NH»HSO» or
hLSQ*,   is the  principal  component,  often  accounting  for 40  percent of  fine-particle  mass.
Sulfate and NO, ions are present in high concentrations during both summer and winter episodes
                                                                                          3
over very large sections  of the eastern  United States.  This  area experiences  10 |jg/m  or
           p-
greater SO,   levels  for one  or two  periods 'up  to  a month or more  every  year.   The affected
region is so large in scope that no real background levels of fine particles are available for
measurement east of  the  Mississippi.   Sulfat'e and fine-particle levels are nearly the same in
                                             i
cities  and in  rural areas.   Southern  California  experiences  high levels  of  sulfates  and
nitrates,  particularly  during  photochemical incidents.   In that area,  high levels  of fine
                                                         o
organic aerosols are also found, often exceeding 100 jjg/m .
     Toxic  organic  particulate matter  and imetals  are  mainly  emitted  from combustion  and
industrial sources,  and  their concentrations are highest  in cities.   Trends in fine-particle
components  have  been  mostly  downward because of  control  measures   taken,  such as  lead
reductions in gasoline.
     Coarse particles in  air are stirred up by the wind and by machinery.   Since these parti-
cles settle fairly  rapidly,  they tend to be high  close to sources.   In most cases the coarse
particles account for two-thirds of TSP in  dry  regions like Phoenix, Oklahoma City, El  Paso,
or  Denver  in   the   summer.   The  overwhelming  cause   of  high  TSP  is  local  dust,  but  in
industrialized  cities  there  is  evidence that contributions  of soot, fly  ash,  and industrial
fugitive emissions are also present.
     Coarse particles  are mainly  composed  of silica,  calcium carbonate,  clay  minerals,  and
soot.   Chemical  constituents found  in  this fraction include  the  elements silicon, aluminum,
potassium, calcium,  and  iron together with other alkaline earth and transition elements.  Or-
ganic substances are also found  in coarse particles, although their source is unknown.
     Much of this  coarse material is road dust suspended by traffic action, and street levels
of resuspended  dust  can  be very high.   Traffic  on  unpaved roadways can generate huge amounts
of dust,  which  deposits  on vegetation and  can be  resuspended by wind  action.   Rain and snow
cover  can reduce these  emissions,  but  one  study  suggests that salting of  roadways can be a
major source of winter TSP.  Probably, associated sand is also important.  Industrial fugitive
emissions  can  be even  greater   local sources  of coarse  particles,  particularly  from unpaved
access roads, construction activity,  rock crushing, and cement manufacturing.
     The problem of tracing existing levels of particles to sources is being solved in part by
a  number  of   calculational   methods   generally  categorized  as   source  apportionment  or
source-receptor  models.   The  results  from chemical element  balance calculations  or factor
                                            5-137

-------
analysis are available  now for several cities.  Apportionments for these cities are presented
as examples of  results  to be expected in the future by application of these powerful methods.
     Although outdoor  concentrations of  pollutants  can be measured at  particular  sites,  our,
highly mobile population  can be exposed to either higher or lower values than community moni-
tors show.  Indoor  values of S02 tend to  be  lower than outdoor levels because walls, floors,
and furniture  absorb SOp.   Indoor  particle  levels  can be high because  of smoking, cleaning
operations, or  normal  activities.   Exposure of  individuals to SO  and PM  can  vary more than
community monitors show.
                                            5-138

-------
5.10  REFERENCES

Akland, G. G.   Air Quality Data  for  Metals  1970 through 1974  from the National Air  Surveil-
     lance  Networks.   EPA-600/4-76-041,  U.  S.  Environmental  Protection  Agency,   Research
     Triangle Park, NC, August 1976.

Akland, G.  G.   Air  Quality Data  for Nonmetallic  Inorganic  Ions  1971 through 1974  from  the
     National  Air Surveillance  Networks.   EPA-600/4-77-003,  U.  S.  Environmental Protection
     Agency, Research Triangle Park,  NC, January 1977.

Akselsson, R., C. Orsinc, D. L. Meinert, T. B. Johansson, R. E.  Van Grieken,  H.  C.  Kaufman,  K.
     R. Chapman,  J. W.  Nelson,  and  J.  W.  Winchester.   Application  of proton-induced  x-ray
     emission  analysis to  the  St.   Louis  regional  air  pollution  study.    Adv.  X-ray  Anal.
     18:588-597, 1975.

Alzona, J., B.  L. Cohen, H. Rudolph,  H. N. Jow, and J. 0. Frohliger.   Indoor-outdoor  relation-
     ships for airborne particulate matter of outdoor  origin.  Atmos.  Environ.  13:55-60,  1979.

Anderson,  I.   Relationships between outdoor and indoor air pollution.   Atmos.  Environ. 6:275-
     278,  1972.

Appel, B.  R.,  E.  L.  Kothny,  E.  M.  Hoffer,  G.  M.  Hidy,  and  J.  J.  Wesolowski.  Sulfate  and
     nitrate  data  from  the  California  Aerosol  Characterization Experiment.  Environ.  Sci.
     Techno!. 12:418-425, 1978.

Appel, B.  R., S. M. Wall, Y. Tokiwa,  and M. Ha'ik.   Interference  efects  in sampling  particulate
     nitrate in ambient air.  Atmos.  Environ. 13:319-325, 1979.

Asahina,  S.  J.  Andrea,  A.  Carmel,  E.   Arnold,  Y.  Bishop,  S. Joshi,  0.  Coffin, and S.   S.
     Epstein.  Carcinogenicity of organic fractions of particulate pollutants collected in  New
     York  City and  administered  subcutaneously  to  infant mice.   Cancer  Res. 32:2263-2268,
     1972.

Axetell,  K.   Assessment  of non-traditional  source impacts in  the Western  United States.  In:
     Specialty  conference on  the Technical  Basis  for a Size  Specific Particulate  Standard.
     Proceedings:  Danvers,  Massachusetts and Seattle, Washington,  March  24-26 and April  23-
     25, 1980.   E. R. Frederick,  ed., Air Pollution Control Association, Pittsburgh,  PA,  1980,
     pp.  223-229.

Ball,  R.  J.,  and  G.  E.  Anderson.   Optimum Site  Exposure  Criteria for S02  Monitoring.  EPA-
     450/3-77-013,  U.S.  Environmental Protection  Agency,  Research Triangle Park, NC, April,
     1977.

Berdyev,  Kh. B. ,  N.  V. Pavlovich,  and A.A.  Tu'zhilina.  Effect  of motor vehicle exhaust  gases
     on atmospheric pollution  in  dwelling and in a main street.  Gig.  Sanit.  ^2:424-426,  1967.

Biersteker, K.  , H. de Graaf, and  C. A. G. Nass.  Indoor air pollution  in Rotterdam  homes.   Air
     Water Pollut. 9:343-350,  1965.

Binder, R. E. , C. A. Mitchell, H. R.  Hosein, and A. Bouhuys.  Importance of the indoor environ-
     ment in air pollution  exposure.  Arch. Environ.  Health 31:279,  1976.

Bradway, R. M., and F.  A. Record.  National Assessment of the Urban Particulate Problem.  Vol.
     II - Particle Characterization.  EPA-450/3-76-025, U. S. Environmental Protection Agency,
     Research Triangle Park, NC,  July 1976.
                                             5-139

-------
BHdbord, K., J, F. Finklea, J. K. Wagoner, J. B. Moran,  and  P.  Caplan.  Human  exposure  to  poly-
     nuclear aromatic hydrocarbons.   In:  Carcinogenesis:   vol.  I:  Polynuclear Aromatic Hydro-
     carbons:   Chemistry,  Metabolism and Carcinogenesis,  a Symposium,  U.S.  Environmental  Pro-
     tection Agency and  Battelle  Columbus  laboratories,  Columbus,  Ohio,  October 1975.    R.
     Freudenthal and P. W. Jones, eds.,  Raven Press, New  York,  NY,  1976.  pp. 319-324.

Brosset,  C., K. Andreasspn,  and M.  Perm.   The nature and possible  origin of  acid particles
     observed at the Swedish west coast.  Atmos. Environ.  9:631-642,  1975.

Brosset, C.  Water-soluble sulphur compounds  in aerosols.   In:   Sulfur  in the  Atmosphere,  Pro-
     ceedings of the International Symposium, United Nations  Environmental  Program, Dubrovnik,
     Yugoslavia, September 7-14, 1977.   Atmos. Environ. 12:25-38, 1978.

Cahill,  T.  A., L,  L.  Ashbaugh, J.  B.  Barone,  R.  A.  Eldred, P.  Feeeney, R. G. Flocchine,  C.
     Goodart, D. J.  Shadoan and G.  W.  Wolfe. Analysis of  respirable fractions in  atmospheric
     particulates  via  sequential  filtration.   J. Air Pollut. Control Assoc. 27:675-678, 1977.

Calvert, J.  G.   Hydrocarbon involvement  in photochemical  smog  formation in Los Angeles atmos-
     phere.  Environ. Sci. Techno!.  10:256-262, 1976.

Camp, D.  C., A. L. Van Lehn, and B.  W.  Loo.  Intercomparison of Samplers Used in the Determi-
     nation  of  Aerosol  Composition.   EPA-600/7-78-118, U.  S. Environmental  Protection  Agency,
     Research Triangle Park, NC, July 1978.

Cantrell,  B.,   and R.   T.  Whitby.    Aerosol  size distributions  from  the  Labodie  power plant
     plume.  Atmos. Environ. 12:324,  1978.

Carey, G. C. R., J. J.  Phair, R. J.  Shephard, and M. L. Thomson:  The effects  of air pollution
     on human health.  Am. Ind. Hyg.  Assoc. J. 19:363-370,  1958.

Carpenter,   B.   H.,  and   G.   E.   Weant,  III.    Particulate   Control  for   Fugitive  Dust.
     EPA-600/7-78-071, U.S. Environmental Protection Agency,  Research Triangle Park, NC, April
     1978.

Cautreels, W.,  and K. Van Cauwenberghe.   Experiments on the distribution of organic pollutants
     between airborne  particulate matter and the  corresponding gas phase.   Atmos.  Environ.
     12:1133-1141, 1978.

Charlson,  R.  J.,   D. S.  Covert, T.  V.  Larson,  and A.  P.  Waggoner.    Chemical properties  of
     tropospheric  sulfur  aerosols.   In:  Sulfur in the  Atmosphere, Proceedings of the Inter-
     national Symposim, United Nations  Enviornmental Program, Dubrovnik,  Yugoslavia, September
     7-14, 1977. Atmos. Environ. 12:39-53, 1978.

Cobourn, W. G,, R.  B. Husar, and J.  D.  Hysar.  Continuous jji  situ monitoring of ambient parti-
     culate  sulfur using flame  photometry and  thermal  analysis.  In:   Sulfur in the Atmos-
     phere,  Proceedings of  the International Symposium,  United  Nations  Environmental Program,
     Dubrovnik, Yugoslavia, September 7-14, 1977.  Atmos.  Environ.  12:89-98, 1978.

Cohen, D.,  S.   F.  Arai,  and J. D.  Brain.  Smoking impairs long-term dust clearance from  the
     lung.  Science (Washington, DC)  204:514-517, 1979.

Colucci, J. M., a nd C. R. Begeman.   Polynuclear aromatic  hydrocarbons  and other pollutants in
     Los Angeles air.  In:  Proceedings  of the Second International Clean Air  Congress, Inter-
     national Union  of  Air Pollution Prevention Associations,  Washington, DC,  December 6-11,
     1976.  H.  M.  Englund and W. T.  Berry, eds., Academic  Press,  Inc.,  New York, NY, 1971.  pp.
     28-35.
                                            5-140

-------
Cooper, D. W.,  J.  S.  Evans, M.  Quinn,  R.  C, AntoneTIi, and M. Schneider.  Setting Priorities
     for Control of Fugitive Participate Emissions from Open Sources.  EPA-600/7-79-186, U. S.
     Environmental Protection Agency, Research Triangle Park, NC, August 1979.

Cooper, J. A. ,  J.  G.  Watson, and  J.  J.  Huntzicker.   Summary of  the Portland Aerosol Charac-
     terization Study.   Presented at  the  72nd  Annual  Meeting,  Air Pollution Control Assoc-
     iation,  Cincinnati, OH, June 24-29, 1979a.  Paper 79-24,4.

Cooper, J. A.,  and J.  G.  Watson,  Jr.   Receptor  oriented methods  of air particulate source
     apportionment. J. Air Pollut, Control  Assoc. 30:1116-1125, 1980.

Corn,  M,   Aerosols and  the primary  air pollution  -  nonviable  particles:  their occurrence,
     properties, and  effects.   In:   Air Pollution. Vol. I:  Air  Pollutants, Their Transforma-
     tion and Transport.   A.  C. Stern,  ed. ,  Academic  Press,  New York, NY, 1976.  pp. 78-168.

Corn, M.  Nonviable particles in the air.  In:  Air Pollution.  Vol. I:  Air Pollution and its
     Effects.  A.  C.  Stern, ed., Academic Press, 2nd Edition.  New  York, NY, 1968, pp. 47-94.

Courtney, W.  J., J. W. Tesch, R. K. Stevens, and T. G, Dzubay.  Characterization of the Denver
     aerosol  between  December  1978 and  December 1979.   Presented at the 73rd Annual Meeting,
     Air Pollution Control Association,  Montreal, Quebec, Canada, June 22-27, 1980.   Paper no.
     80-58.1.

Cowherd,  C.,  Jr., R.   Bonn, and  T.  Cuscino, Jr.   Iron and  Steel  Plant Open Source Fugitive
     Emission Evaluation.   EPA-600/2-79-103, U. S.  Environmental Protection Agency, Research
     Triangle Park, NC, May 1979.

Cox, R. A.,  and S. A. Penkett.  Effect  of relative humidity on the  disappearance of  ozone and
     sulphur dioxide in contained systems.   Atmos.  Environ. 6:365-368, 1972.

Cronn,  D.  R., R.  J.  Charlson,  R.  L.  Knights, A. L. Crittenden,  and B. R. Appel.  A  survey of
     the  molecular nature  of  primary and  secondary components  of  particles  in urban air by
     high-resolution mass spectrometry.  Atmos. Environ. 11:929-937, 1977.

Cunningham,  P.  T., and  S.  A.   Johnson.   Spectroscopic observation  of acid sulfate  in atmos-
     pheric particulate samples.  Science (Washington, D.C.) 191:77-79, 1976.

Curran, T.  C.,  and W.  F. Hunt, Jr.   Interpretation of air quality data  with respect to the
     national ambient  air quality standards.  J, Air Pollut, Control  Assoc. 25:711-714, 1975.

Daisey,  J.  M.    Organic  compounds  in  urban  aerosols.    In:   Aerosols:   Anthropogenic  and
     Natural, Sources and Transport, Proceedings of a Conference, New  York Academy of Sciences
     and Air Pollution Control  Association, New York, NY, January 9-12, 1979.  Ann, N.Y. Acad.
     Sci.  338:50-69, 1980.

Daisey,  J.  M.,   I.  Hawryluk,  T.  J.  Kneip,  and  F.  Mukai.    Mutagenic  activity  in organic
     fractions  in airborne particulate  matter.  In:   Proceedings:   (Conference)  Carbonaceous
     Particles  in the Atmosphere,  National  Science Foundation and Lawrence Berkeley Labora-
     tory, University of California,  Berkeley, California,  March  20-22,  1978.   LBL-9037, T.
     Novakov, ed. ,   University of  California,  -Lawrence  Berkeley  Laboratory,  Berkeley,  CA,
     1979.  pp.  187-192.

Daisey, J.  M.,  R.  J. McCaffrey,  and  R. A.  Gallagher.   Polycyclic aromatic hydrocarbons and
     total extractable  particulate organic matter in the Arctic  aerosol.  Atmos. Environ. 15:
     1352-1363, 1981.
                                            5-141

-------
Davidson,  C.   I.,  and  S.  K.  Frledlander.    A  filtration  model  for  aerosol  dry deposition:
     application   to  trace  metal   deposition  from   the  atmosphere.    J.   Geophys.   Res.
     83:2343-2352, 1978.

De Maio,  L,  and M.  Corn.   Polynuclear  aromatic  hydrocarbons associated with participates  in
     Pittsburgh air.   J. Air Pollut. Control  Assoc. 16:67-71, 1966.

De Wiest,  F.   Any factors influencing the  dispersion and the transport of heavy  hydrocarbons
     associated with airborne particles.  Atmos. Environ. 12:1705-1711, 1978.

De Wiest, F., and D.  Rondia.  Sur la validite des determinations du benzo(a)pyrene atmospheric
     pendant   les  mois  d'ete,   [on  the   validity   of  the  determination  of  atmospheric
     benzo(a)pyrene  in the summer.]  Atmos.  Environ.  10:487-489, 1976.

Dockery, 0, W.,  and  J.  D. Spengler.   Indoor-outdoor  relationships of respirable  sulfates and
     particles,  Atmos.  Environ. 15:335.-343,  1981.

Dockery,  D.  W.  Personal  Exposure to Respirable Particulates  and Sulfates:   Measurement and
     prediction.   Ph.D.  Dissertation,  Harvard  School   of  Public Health,  Boston,  MA, 1979.

Draftz, R.  G.   Aerosol  Source Characterization Study in Miami, Florida:  Microscopical Analy-
     sis.   EPA-600/3-79-097,  U.  S.  Environmental  Protection  Agency,  Research Triangle Park,
     NC, September 1979.

Draffz,  R.  6., and  K.  Severin.   Microscopical  Analysis of  Aerosols  Collected  in St. Louis,
     Missouri.   EPA-600/3-80-027,   U,  S.  Environmental  Protection Agency,  Research  Triangle
     Park, NC, February 1980.

Draftz,  R.  G.,  I.  Graf,  and  E.   Arnold.    Preliminary  assessments  of traditional  and  non-
     traditional source  impacts.   In:   Specialty Conference on the Technical Basis  for a  Size
     Specific  Particulate  Standard.   Proceedings,  Danvers, Massachusetts  and  Seattle, Wash-
     ington, March 24-26,  and April 23-25,  1980.  E. R. Frederick, ed. , Air Pollution  Control
     Association, Pittsburgh, PA, 1980.  pp.  94-106.

Duce, R.  A.   Speculations  on the budget of  particulate  and in vapor phase non-methane  organic
     carbon in the global troposphere.  Pure Appl. Geophys. 116(23):244-273, 1978.

Dzubay, T.  G.   Chemical-elemental  balance  method  applied to dichotomous  sampler data.  In;
     Aerosol;  Anthropogenic  and Natural  Sources and Transport,  Proceedings  of a Conference,
     New York  Academy of Sciences and Air  Pollution  Control  Association, New York, New York.
     January 9-12, 1979.  Ann., N.Y.  Acad.  Sci. 338:126-144, 1980.

Dzubay, T.  G. , and  R., K.  Stevens.   Ambient air analysis  with  dichotomous  sampler and x~ray
     fluorescence spectrometer.  Environ. Sci. Techno!.  9:663-668, 1975.

Dzubay, T.  G.,  R.  K. Stevens, and  C.  M.  Peterson.    Application of the dichotomous  sampler  to
     the characterization  of ambient aerosols.   In:  X-Ray Fluorescence  Analysis of  Environ-
     mental Samples.   T. G. Dzubay,  ed.,  Ann Arbor  Science  Publishers,  Inc.,  Ann Arbor, MI,"
     1977, pp. 95-106.

Dzubay, T.  G., R. K.  Stevens, W.  J. Courtney,  and E.  A.  Drane.  Chemical  element  balance
     analysis of Denver aerosol.   In:  Electron Microscopy and X-ray Applications to  Environ-
     mental Health Analysis.  Volume 2.  P.  A. Russell,  ed. , Ann Arbor Science, Ann Arbor, MI,
     1981.  pp. 23-42.

Economic Commission  for Europe's  Working Party on  Air  Pollution Probfems.   ECE Task  Force  on
     Fine  Particulate   Pollution.   World  Health  Organization, Geneva,  Switzerland,  December
     1977.
                                            5-142

-------
Eliassen, A.  The  OECD study of long range transport of air pollutants:   Long range transport
     modelling.   In:   Sulfur in  the Atmosphere,  Proceedings  of the International Symposium,
     United  Nations  Environmental  Program,  Dubrovnik,  Yugoslavia,  September  7-14,  1977.
     Atmos.  Environ. 12:479-487, 1978.

Faoro, R. B.  Trends in concentrations of  benzene-soluble suspended particulate  fraction and
     benzo(a)pyrene.  J. Air Pollut. Control Assoc. 25:638-640,  1975.

Faoro, R. D. , and  T. B. McMullen.  National Trends in Trace Metals  in Ambient Air, 1965-1974.
     EPA-450/1-77-003,  U.S.  Environmental   Protection   Agency,  Research  Triangle  Park,  NC,
     February 1977.

Flocchini, R. G. ,  T. A. Cahill, R.  A.  Eldred, L. L. Ashbaugh,  and  J. B.  Barone.  Sulfur size
     distribution by season and site in California.  Presented at the 71st Annual  Meeting, Air
     Pollution Control Association, Houston, TX, June 25-30, 1978.   Paper  no. 78-39.6.

Fraser, D. W., and J. E. McDade.  Legionellosis.   Sci. Am. 241:82-99, 1979.

Friedlander, S.  K.   Chemical  element balances  and  identification  of  air pollution sources.
     Environ. Sci.  Tech. 7:235-240, 1973.

GEOMET,  Incorporated.   Indoor-Outdoor  Pollution  Levels:   A  Bibliography.    EPRI  EA-1025,
     Electric Power  Research Institute, Palo Alto, CA, March 1979.

Gillette, D.   A  wind  tunnel  simulation  of  the  erosion of  soil:   effect of  soil  texture,
     sandblasting,  wind speed,  and  soil  consolidation  on dust production.   Atmos.  Environ.
     12:1735-1743, 1978.

Goldwater,  L.  J. ,  A.   Manoharan,  and M.  B.  Jacobs.   Suspended particulate matter,  dust  in
     "domestic"  atmospheres.  Arch. Environ. Health 2:511-515, 1961.

Gordon, G. E.  Receptor models.   Environ. Sci. Technol. 14:792-800,  1980.

Gordon,  R.   J.   Distribution  of  Airborne  Polycyclic  Aromatic Hydrocarbons  throughout  Los
     Angeles.  Environ. Sci. Technol. 10:370-373,  1976.

Graf,  J.,  R. H.  Snow,  and R. G.  Draftz.   Aerosol sampling and analysis  - Phoenix, Arizona.
     EPA-600/2-77-015, U.S. Environmental Protection Agency, Washington, DC, 1977.

Grosjean, D.   Aerosols.   In:   Ozone  and Other  Photochemical  Oxidants.  National Academy  of
     Sciences, Washington, DC, 1977. pp. 45-125.

Grosjean, D., and  S. K. Friedlander.  Gas-particle distribution factors for organic and other
     pollutants  in  the Los  Angeles atmosphere.   J. Air Pollut.  Control  Assoc. 25:1038-1044,
     1975.

Hahn, J.  Organic constituents of natural aerosols.  In:   Aerosols:  Anthropogenic and Natural
     Sources and Transport,  Proceedings of a Conference,  New York Academy of Sciences and Air
     Pollution Control  Association,  New York, New York,  January 9-12,  1979.  Ann. N.Y. Acad.
     Sci.338:359-376. 1980.

Hardy,  K.  A.    Aerosol  Source  Characterization  Study   in  Miami,  Florida:    Trace  Element
     Analysis.   EPA-600/7-79-197,  U.S.   Environmental   Protection   Agency,  Research  Triangle
     Park, NC, September 1979.

Heffter,  J.  L. ,  A.  0.  Taylor, and  G.  J.  Ferber.   A  Regional-Continental  Scale Transport,
     Diffusion,  and Deposition  Model.   NOAA Technical Memorandum ERL-ARL-50, U. S. Department
     of  Commerce,  National  Oceanic and  Atmospheric  Administration,  Washington,  DC,  1975.
                                            5-143

-------
Heisler, S. L., R. C. Henry, J. G. Watson,  and  G.  M.  Hidy.   The  1978 Denver Winter Haze Study,
     vol.  II - Final Report.   ERT  Document No.  T-5417-1,  Environmental  Research and  Tech-
     nology, Inc., Westlake Village,  CA,  March  1980.

Helms,  G.  T., J.  H. Southerland,  K. R.  Woodard, I. J.  Hindawi,  and D.  H.  Coventry,   Chat-
     tanooga,  Tennessee  —   Rossville,   Georgia   Interstate Air   Quality  Study  1967-1968.
     APTD-0583,  U.S. Department  of  Health, Education,  and Welfare,  National Air  Pollution
     Control Administration, Durham,  NC,  October 1970.

Hidy, G. M.,  B.  R.  Appel, R.  J. Charlson,  W. E. Clark, S.  K. Friedlander,  D.  H.  Hutchison,  T.
     B.  Smith, J.   Suder,  J.  J.  Wesolowski,  and K. T.  Whitby.   Summary of the  California
     Aerosol  Characterization  Experiment.  J.  Air Pollut.  Control  Assoc.  25:1106-1114,  1975.

Hidy, G. H.,  P.  K.  Mueller, and  E.  Y. Tong.   Spatial  and temporal distributions  of  airborne
     sulfates  in  parts of the United States.  In:   Sulfur in  the Atmosphere, Proceedings  of
     the International Symposium, United  Nations Environmental Program,  Dubrovnik,  Yugoslavia,
     September 7-14, 1977.  Atmos. Environ.   12:735-752,  1978.

Hidy, G. M.,  P.  K.  Mueller, V.  Dio,  and K.  C. Detor. Study and implementation of the Sulfate
     Regional  Experiment  (SURE).   Proc., 4th  Symposium  on  Turbulence  and Diffusion  and Air
     Pollution, Aerican Meteorological Society, Reno, Nevada, January 1518, 1979.

Hitchcock, D. R.  Atmospheric  s'ulfates from biological sources.   J.  Air  Pollut. Control  Assoc.
     26:210-215, 1976.

Hitchcock, D. R.  Biogenic Sulfur Sources and Air  Quality in the United  States.   NSF/RA-770276,
     National Science Foundation, Washington, DC,  August  1977.

Hitchcock, D.  R.,  L. L.   Spiller, and W.  E.  Wilson.   Sulfuric acid  aerosols and HC1 release  in
     coastal atmospheres:   evidence  of rapid formation of sulfuric acid particulates.   Atmos.
     Environ. 14:165-182, 1980.

Hoffmann,  D.,  and  E.  L.  Wynder.   Organic  particulate  pollutants—chemical  analyses and
     bioassays for  carcinogenicity.   In:   Air  Pollution.   A. C. Stern,  ed.,  3d  ed.,  Academic
     Press, Inc., New York, NY, 1977.  pp.  361-455.

Hueper,  W.  C., P.  Kotin, E.  C.  Tabor,  W.  W.  Payne,  H.  Falk,  and E.  Sawicki.  Carcinogenic
     bioassays on air pollutants.  Arch.  Pathol. 74:89-116,  1962.

Huntzicker,  J.  J.,  R. A. Cary,  and  C-S.  Ling.   Neutralization of sulfuric acid  aerosol  by
     ammonia.  Environ.  Sci. Technol. 14:819-824,  1980.

Huntzicker, J. J.,  R.  L.  Johnson, J.  J.  Shah,  and R.  A.  Cary.   Analysis  of  organic  and ele-
     mental carbon  in  ambient aerosols by  a thermal-optical  method.   In:   Particulate Carbon:
     Atmospheric  Life  Cycle,  Proceedings   of an  International  Symposium,  General  Motors
     Research  Laboratories,  Warren,  MI,   October 13-14, 1980.  G. T.  Wolff  and R.  L.  Klimisch,
     eds., Plenum Prss, New York, NY, 1982.   pp. 79-88.

Ishido,  S.   Variations  in indoor and outdoor dust intensities.   Bull.  Dept.  Home Econ.,  Osaka
     City University, 6:53-59, 1959.

Ishido,  S.,  K. Kamada,  and   Nakagawa:   Free  dust particles and airborne  microflora.   Bull.
     Dept.  Home Econ., Osaka City Univ.,  21:31-37,  1956.

Jacobs, M.  B., L. J. Goldwater, and A. Fergany:  Comparison of suspended particulate matter  of
     indoor and outdoor air.   Air Water Pollut. 6:377-380,  1962.
                                             5-144

-------
Jager,  J.  Detection  and characterization  of nitro  derivatives  of  some  polycyclic aromatic
     hydrocarbons by  fluorescence quenching  after  thin-layer chromatography:   application to
     air pollution analysis.  J, Chromatogr. 152:575, 1978.

Junge,  C.   Die (Constitution  des  Atmospharischen Aerosols.   Ann.  Meteorol.  5 (Beiheft) 1-55,
     1952.

Junge,  C.  and G. Scheich.   Determination  of the acid content of aerosol particles.  Atmos.
     Environ. 5:165-175, 1971.

Kadawaki, S.   Size distribution and chemical composition of atmospheric participate  nitrate in
     the Nagoya area.   Atmos. Environ. 11:671-675, 1977.

Katz, M.,  and  C,  Chan.   Comparative distribution of eight polycyclic aromatic hydrocarbons in
     airborne particulates collected by conventional high-volume sampling  and by size fraction-
     ation.  Environ.  Sci. Technol. 14:838-843, 1980.

Kertesz-Saringer, M.,  E.  Meszaros,  and  T.  Varkovni.   On  the  size distribution  of carbon-
     containing particles in urban air.  Atmos. Environ. 5_: 429-431, 1971.

Ketseridis,  G., J.  Hahn, R. Jaenicke, and  C.  Junge.   The organic constituents of atmospheric
     particulate matter.  Atmos. Environ.  10:603-610, 1976.

Kito, T.   Survey  of air pollution  on  outlying islands:   An  attempt  to collect the  background
     data as to urban air pollutions.  Kuki Seijo 15:10-21, 1977.

Kneip,  T.  J.,  M.  A. Leyko, M.  T.  Kleinman, M. Lippmann, and J. M. Daisey.  Organic matter in
     New  York  City TSP:   comparisons of  seasonal  variations and  relationships  to  source
     tracers.   In:   Proceedings:   Conference  on Carbonaceous  Particles  in  the Atmosphere,
     National  Science  Foundation  and  Lawrence  Berkeley  Laboratory,  University of California,
     Berkeley,  California,  March  20-22,   1978.   LBL-9037,  T.  Novakov,  ed.,  University of
     California, Lawrence Berkeley  Laboratory, Berkeley, CA,  June  1979.  pp. 79-83.

Korfmacher,  W.  A. ,  E.  L.  Wehry,  G.  Mamantov, and  D.  F.  S. Natusch.    Resistance  to photo-
     chemical  decomposition of polycyclic  aromatic hydrocarbons  vapor—adsorbed on coal fly
     ash.  Environ.  Sci. Technol.  14:1094-1099, 1980.

Krstulovic, A. M., D.  M. Rosie, and P. R.  Brown.  Distribution of  some atmospheric polynuclear
     aromatic hydrocarbons.  Am. Lab.  9:11-12,14,16-18, 1977.

Kurosaka,  D.   Sulfate Concentrations  in  the South Coast  Air Basin.  DTS-76-1, California Air
     Resources Board, Sacramento, CA,  January 1976.

Lamb, S.  I., C.  Petrowski, I.  R.  Kaplan,  and B.  R.  T.  Simoneit.   Organic compounds in urban
     atmospheres:  a review of distribution,  collection and analysis.  J.  Air. Pollut. Control
     Assoc.  30:1098-1115, 1980.

Larsen,  R.  I.   A  Mathematical  Model for  Relating Air  Quality Measurements to  Air Quality
     Standards.   AP-89,   U.S.  Environmental  Protection Agency,  Research  Triangle  Park, NC,
     November 1971.

Lavery, T.  F. ,  G.  M.  Hidy, R.  L. Basket, and J.  Thrasher.  Occurrence of  long-range transport
     of sulfur  oxides  in the northeastern  United States.   In: Proceedings:  4th Symposium on
     Turbulence  Diffusion  and  Air Pollution  American  Meteorological  Society,  Reno, Nevada,
     Jan.  15-18, 1979.
                                            5-145

-------
Leaderer,  B.  P.,  R.  L.  Tanner, and T.  H.  Holford.   Diurnal variations,  chemical  composition
     and  relation  to meteorological  variables of  the summer  aerosol  in  the  New York  sub-
     region.  Atmos Environ.  16:2075-2087,  1982.

Lee, R.  E.,  Jr., and S.  S. Goranson.   National  Air  Surveillance Cascade Impactor Network.   I.
     Size  distribution  measurements  of suspended particulate  matter in  air.    Environ.  Sci.
     Techno!. (5:1019-1024, 1972.

Lee, R.  E.,  Jr., S.  S.  Goranson,  R.  E. Enrione,  and  G. B.  Morgan.   National Air Surveillence
     Cascade Impactor Network.   II.   Size distribution measurements  of trace metal  components.
     Environ. Sci. Techno!. 6:1025-1030, 1972.

Lee,  R.  E.,  Jr., R.   K,  Patterson,  and J.  Wagman.   Particle-size  distribution  of  metal
     components  in urban  air.   Environ.  Sci. Techno!.  2:288-290,  1968.

Lefcoe, N. H.,  and I. I. Inculet.  Particulates in  domestic premises:   II.  Ambient levels  and
     indoor-outdoor relationships.  Arch. Environ. Health  30:565-570,  1975.

Lewis,  C.  W.,  and E.  S.  Macias.  Composition  of size-fractionated aerosol  in Charleston, West
     Virginia.   Atmos. Environ.  14:185-194, 1980.

Lioy, P. J., P.  J. Samson, R.  L. Tanner, B. P.  Leaderer, T.  Minnich,  and W.  Lyons.  The distri-
     bution  and transport of sulfate "species" in  the New York metropolitan area during  the
     1977 summer aerosol  study.  Atmos.  Environ.  14:1391-1407, 1980.

Lioy, P. J., R.  P. Mallon, and T. J.  Kneip.  Long-term trends in total suspended participates,
     vanadium,  manganese, and lead at  near street level  and elevated sites in  New York City.
     J. Air Pollut. Control Assoc. 30:153-156,  1980a.

Lundgren,  D.  A.   Atmospheric  aerosol composition  and  concentration as  a  function  of  particle
     size and of time.  J. Air Pollut.  Control  Assoc.  20:603-608,  1970.

Lynn, D. A.,  B. S. Epstein,  and C. K.  Wilcox.   Analysis of New  York City Sulfate Data.   Final
     Report.  EPA Contract No.  68-02-1337,  GCA Corp.,  Bedford, MA, 1975.

Lynn, D. A., G.   L. Deane, R.  C.  Galkiewicz, R.  M.  Bradway,  and F.  R.  Record.  National  Assess-
     ment  of  the  Urban  Particulate  Problem.   Volume  I.   Summary of the National  Assessment.
     EPA-450/3-76-024, U. S.  Environmental  Protection  Agency, Research Triangle  Park,  NC, July
     1976.   p. 53.

Lyons,  W.  A.,  J.  C.  Dooley,  Jr., and  K.  T.  Whitby.   Satellite detection of long range pol-
     lution transport and sulfate aerosol haze.  In:   Sulfur in  the  Atmosphere,  Proceedings of
     the International Symposium, United Nations Environmental Program,  Dubrovnik,  Yugoslavia,
     September 7-14,  1977.  Atmos. Environ. 12:621-631, 1978.

Mage, D. T.,  and W.  R.   Ott.   Refinements  of the  lognormal  probability  model for  analysis of
     aerometric  data.   J. Air Pollut. Control  Assoc. 28:796-798,  1978.

Maunsell,  K.  Concentration  of airborne spores  in dwellings under normal  conditions and under
     repair.  Int. Arch.  Allergy Appl.  Immunol.  5:373-376,  1954.

HcCrone, W.  C.   Morphological analysis of particulate pollutants.   In:  Air Pollution.   2nd
     Ed.,  Vol.   II:   Analysis, Monitoring,  and Surveying.   A. C.  Stern, ed., Academic  Press,
     Inc.,  New York, NY,  1968.   pp. 281-301.

McFadden,  J.  E.,  J.  H.   Beard,  JI !.,,.„  and  D.  J.  Moschandreas.   Survey  of Indoor  Air  Quality
     Health Criteria  and Standards.  EPA*-600/7-78-027, 'U.S. Environmental  Protection  Agency,
     Research Triangle Park,  NC, March  1978.
                                            5-146

-------
Megaw, W.  S.   The penetration  of  iodine into buildings.   Air Water Pollut. 5:121-128, 1962.

Miller, F. J.,  D.  E.  Gardner,  J.  A.  Graham, R.  E. Lee, Jr.,  W. E. Wilson,  and J. D. Bachman.
     Particle Size Considerations  for Establishing a  Standard for Inhaled  Particles,  Draft,
     U.S.  Environmental Protection Agency, Research Triangle Park, NC, January 1979.

Montgomery, T.  L.  and J.  H. Coleman,  Empirical  relations between timeaveraged sulfur dioxide
     concentrations.   Environ.  Sci. Techno!. 9:9537, 1975.

Moschandreas, D. J.,  J.  W.  Winchester,  J.  W.  Nelson,  and  R.  M.  Burton.   Fine particle resi-
     dential indoor air pollution.  Atmos. Environ. 13:1413-1418,  1979.

Mueller, P. K.,  H.  L. Helwig,  A.  E.  Alcocer,  W.  K. Gong,  and E.  E. Hones.  Concentration of
     fine particles  and Pb  in car  exhause.  Am.  Soc. Testing  Mater.,  Spec.  Tech.  Publ.  No.
     352:6073, 1962.

Mueller, P. K.,  R. W. Mosely,  and L.  B. Pierce.  Carbonate and noncarbonate carbon  in atmos-
     pheric particles,   jri:   Proceedings  of  the  Second  International  Clean  Air Congress,
     International  Union of Air Pollution Prevention Associations, Washington, DC, December 6-
     11, 1970.  H.  M.  Englund and W. T.  Beery, eds., Academic  Press,  Inc., New York, NY, 1971.
     pp. 532-539.

Mueller, P. K.,  R. W. Mosley,  and L.  B. Pierce.  Chemical composition of Pasadena aerosol by
     particle  size and  time  of   day.   IV.   Carbonate  and noncarbonate  carbon  content.   J.
     Colloid Interface Sci.  39:235-239,  1972.

Mueller, P. K.,  G. M. Hidy, T. F.  Lavery,   K.  Warren, R. L, Bassett.  Some  early  results  from
     the Sulfate Regional  Experiment (SURE).  Proceedings: 4th  Symposium on Turbulence Diffu-
     sion and Air  Pollution,  American Meteorological  Society, Reno,  Nevada, Jan.  15-18, 1979,
     pp. 322-329.

National Academy of   Sciences.   Air Quality and Stationary Source Emissions Control.  Serial
     No. 94-4, U.S. Senate, Committee on Public Works, Washington, DC, March 1975.

National Academy of Sciences.   Nitrogen  Oxides.  National Academy  of  Sciences, Washington, DC,
     1977.

National Academy of  Sciences.   Vapor-Phase  Organic Pollutants.  National Academy  of  Sciences,
     Washington, DC,   1976.

National Air Pollution Control  Administration.  Air Quality Criteria  for  Sulfur Oxides.  NAPCA
     pub. no. AP-50.   U.S. Government Printing Office, Washington, DC, 1970.      /
                                                                                 /
National Aerometric Data Base.  U. S. Environmental Protection Agency, Research Triangle Park,
     NC, 1977.

National  Academy of  Sciences.   Particulate Polycyclic  Organic Matter.   National Academy of
     Sciences, Washington, DC,  1972.

National  Research  Council.    Nitrates:   An Environmental  Assessment.   National  Academy of
     Sciences, Washington, DC,  1978b.                                  ^

National  Research Council.   Sulfur  Oxides.  National  Academy of  Sciences,  Washington, DC,
     1978a.

National  Research  Council.   Airborne Particles.   University  Park Press, Baltimore, MD, 1979.
                                             5-147

-------
Natusch,  D.  F.   A.   Characterizations  of  atmospheric  pollutants  from power  plants.   |n:
     Proceedings of the Second Federal Conference on the Great  Lakes,  1975.  Great  Lakes Basin
     Commission, Ann Arbor, MI, 1976.  pp. 114.

Novakov,  T.   Soot in  the  atmosphere.    In:   Particulate  Carbon;   Atmospheric  Life Cycle,
     Proceedings of an  International Symposium, General Motors  Research  Laboratories, Warren,
     MI, October  13-14,  1980.   G. T. Wolff  and R.  L.  Klimisch, eds., Plenum  Press, New York,
     NY, 1982.  pp. 19-41.

O'Donnell,  H., T.  L.  Montgomery, and M. Corn.  Routine assessment of  the particle  size-weight
     distribution of urban aerosols.  Atmos. Environ. 4:1-7, 1970.

Office of Air Quality Planning and Standards.   U. S. Environmental Protection  Agency,  Research
     Triangle Park, NC, 1976.

Office of Air Quality Planning and Standards.   U. S. Environmental Protection  Agency,  Research
     Triangle Park, NC, 1979.

Orel, A., and J,  Seinfeld.  Nitrate  formation  in atmospheric aerosols.  Environ. Sci.  Techno!.
     11:1000-1007, 1977.

Pace,  T.  G.   Ambient  particulate  baseline  conditions  -  sources  and  concentrations.   In:
     Specialty Conference on the Technical  Basis  for  a  Size  Specific  Particulate Standard.
     Proceedings, Danvers, Massachusetts and Seattle, Washington, March 24-26  and April 23-25,
     1980.  E.  K.  Frederick,  ed. ,   Air  Pollution  Control  Association,  Pittsburgh, PA, 1980.
     pp. 26-39.

Pace, T. G.,  and  E.  L.  Meyer, Jr.   Preliminary characterization of  inhalable particulates  in
     urban  areas.   Presented at  the 72nd Annual Meeting,  Air  Pollution Control Association,
     Cincinnati,  OH, June 24-29,  1979.  paper  no. 79-47.2.

Pace, T. G., W. P. Freas, and E,   M.  Afify.  Quantification  of the relationship between monitor
     height and  measured particulate levels  in six U.S.  urban  areas.   Presented at the 70th
     Annual Meeting, Air  Pollution Control Association, Toronto,  Ontario,  Canada,  June 1977.
     Paper 77-13.4.

Pace, T. G.,  J.  G. Watson, and C. E. Rodes.   Preliminary Interpretation  of Inhalable  Particu-
     late  Network  Data.   Presented  at  the 74th Annual Meeting,  Air Pollution Control Assoc-
     iation, APCA Paper no. 81-5.2.  Philadelphia, PA, June 21-26, 1981.

Parvis,  D.  Condensation  nuclei  in  the air  of artificially heated environments.   Ann. Sanita
     Pub!ica 13:1569-1581, 1952.

Perera,  F.  P., and A.  K. Ahmed.   Respirab,le  Particulates:   Impact  of  Airborne  Fine Parti-
     culates  on  Health  and the  Environment.   Natural Resources  Defense Council, Inc.,  New
     York, NY, October 1978.

Pierce,  R.  C., and M. Katz.   Dependency of polynuclear aromatic  hydrocarbon  content on size
     distribution of atmospheric  aerosols.  Environ. Sci.  Techno!. 9:347353, 1975.

Pierson, W.  R.,  W. W.  Brachaczek,   T. J.  Korniski,  T.  J.   Truex, and  J.  W.  Butler.  Artifact
     formation of  sulfate, nitrate,  and hydrogen ion on  backup filters:  Allegheny Mountain
     Experiment.   J.  Air Pollut.   Control Assoc. 30:30-34, 1980a.
                                            5-148

-------
Pierson, W.  R.,  W. W.  Brachaczek, T.  J.  Truex,  J.  W.  Butler, and  T.  J.  Korniski.  Ambient
     sulfate measurements  on Allegheny  Mountain and the question  of atmospheric sulfates in
     the northeastern  United States.   In:   Aerosols:  Anthropogenic  and  Natural Sources and
     Transport,  Proceedings  of a  Conference,  New York Academy of  Sciences  and Air Pollution
     Control Association, New  York,  New York, January 9-12,  1979.   Ann. N.Y. Acad. Sci. 338:
     145-147, 1980b.

Pitts, J.  N., Jr., D.  Grosjean, T. M.  Mischke, V. F. Simmon, and D. Poole.  Mutagenic activity
     of airborne particulate organic pollutants.   Toxicol. Lett. 1:65-70, 1977.

Rahn, K. A., R.  D.  Borys, E.  L.  Butler, and R.  A. Duce.  Gaseous and particulate halogens in
     the New York City atmosphere.  Ann. N.Y. Acad. Sci.  322:143-151, 1979.

Record, F.   A.,  and  R.  M. Bradway.  Philadelphia Particulate Study.  EPA-903/ 9-78-003, U. S.
     Environmental Protection Agency, Philadelphia, PA, June 1978.

Record, F.  A.,  K. W.  Whiltsee,  Jr.,  and R.  M. Bradway.   Program to  measure,  analyze,  and
     evaluate total suspended  particulates (TSP)  in  Massachusetts.   GCA-TR-79-44-G, GCA Cor-
     poration,  Bedford, MA, September 1979.

Repace, J.  L.,  and A.  H. Lowrey.   Indoor air  pollution,  tobacco  smoke,  and public health.
     Science 208:464-472, 1980.

Repace,  J.   L. ,   W.  R.  Ott,  and L.  A.  Wallace.   Total  human  exposure  to  air  pollution.
     Presented  at the  73rd  Annual  Meeting, Air Pollution  Control  Association,  Montreal,
     Quebec, Canada,  June 2.2.-21, 1980.   Paper no. 80-61.6.

Rheingrover, S. W.  A  statistical model of titanium pollution transport and  dispersion  in the
     atmosphere of St.  Louis.  J. Appl.  Meterol.  20:693-705, 1981.

Richard, G.   An  Implementation Plan for  Suspended Particulate  Matter in the  Pheonix Area.
     Vol.  IV -  Control  Strategy Formation.  EPA-450/3-77-021d, U. S. Environmental  Protection
     Agency, Research Triangle Park, NC, June 1977.

Richard, G., and  R.  Tan.  An Implementation Plan for Suspended Particle Matter in the Phoenix
     Area.   Vol.  I -  Air Quality Analysis.  EPA-450/3-77-021a, U. S. Environmental  Protection
     Agency, Research Triangle Park, NC, November 1977.

Richard, G. ,  J.  Avery,  and L.  Baboolal.   An Implementation  Plan  for  Suspended Particulate
     Matter in  the  Phoenix Area.  Vol,  III  -  Model Simulation of Total Suspended Particulate
     Levels.   EPA-450/3-77-021c,  U.  S.  Environmental   Protection  Agency,  Research Triangle
     Park,  NC,  August 1977a.

Richard, G., R.  Tan, and J. Avery.  An Implementation Plan for Suspended Particulate Matter in
     the Phoenix  Area.   Vol.  II - Emission Inventory.  EPA-450/3-77-021b, U.  S.  Environmental
     Protection Agency,  Research Triangle  Park,  NC, December 1977b.

Rodhe,  H. ,  C.  Persson,  and  Akesson.   An  investigation  into  regional  transport of soot and
     sulfate aerosols.    Atmos.  Environ.  6:675-693,  1972.

Romagnol,  G.   Studies  on the  climatic  conditions in  some elementary  classrooms of Novara.
     Ital.  Rev. Hyg.  21:410-419, 1961.

Rosen,  H. ,  A.  D.  A.  Hansen, R.  L.  Dod, and T.  Novakov.   Soot  in urban atmospheres:   deter-
     mination  by an  optical  absorption technique.  Science  (Washington,  D.C.)  208:741-744,
     1980.
                                            5-149

-------
Samson,  P.  0.  Trajectory  analysis of summertime  sulfate concentrations in the  northeastern
     United States.  0. Appl. Meteorol. 19:1382-94, 1981.

Santodonato,  J.,  P.  Howard, D.  Basu,  S.  Lande,  J.  K, Selkirk, and P. Sheehe.   Health Assess-
     ment Document for Polycyclic Organic Matter.   EPA-6QO/9-79-008, U.S. Environmental  Pro-
     tection Agency, Research Triangle Park, NC, December  1979.

SAROAD  (Storage  and  Retrieval   of Aerometric  Data),   U.S.  Environmental Protection Agency,
     Office of Air Quality  Planning and Standards,  Research Triangle  Park, NC.

Sawicki,  E,   Analysis  of  atmospheric carcinogens  and  their  cofactors.   In:   Environmental
     Pollution and  Carcinogenic Risks.  Proceedings of  a  Symposium,  International Agency  for
  ,   Research on  Cancer and  French National Institute of Health  and Medical Research,  Lyon,
     France,  November 3-5,  1975.   C.\Rosenfeld  and W.  Davis, eds.,  IARC  Scientific  Publi-
     cations  No.   13,  INSERM Scientific  Publications  No. 13,  French  National Institute  of
     Health and Medical Research (INSERML Paris, France,  1976.  pp.  297-354.
Sawicki,  E.   The  genotoxic environmentalXpollutants,  Paper  presentedXat  the Symposium  in
     Management of Residuals from Synthetic Fuel9 Production, Denver, CO,\May 25, 1976.

Sawicki, E., T.  R.  Hauser, W.  C.  Elbert,  F.  T. Fox,  and  J.  E. Meeker.   Polynuclear  aromatic
     hydrocarbons:   Composition of  the  atmosphere in  some large  American  Vities.   Am.  Ind.
     Hyg. Assoc. 0, 23:137-144, 1962.

Sawicki, E., S.  P.  McPherson, T. W.  Stanley,  J.  Meeker,  and W. C. Albert.   Quantitative  com-
     position of  the urban atmosphere in  terms of polynuclear aza-heterocyclic compounds and
     aliphatic  and  polynuclear  aromatic  hydrocarbons.    Air  Water  Pollut.  9:515-/524, 1965.

Schaefer, V.  S., V.  A.  Mohnen,  and V.  R.  Veirs:   Air  quality of  American homes.   Science
     (Washington, D.C.) 175:133-175, 1972.                                        ,

Schuetzle, D.,  D.  Cronn,  A.  L.  Crittenden,  and R.  J.   Charlson.   Molecular  composition  of
     secondary aerosol and  its possible origin.  Environ.  Sci.  Techno!. 9:838, 1975.

Schwartz, G,, J.  M.  Daisey, and P. J. Lioy.  Effect of sampling duration  on  the concentration
     of  particulate  organics collected on glass  fiber filters.  Am.  Ind.  Hyg.  Assoc. J. 42:
     258-263, 1981.

Shephard,  B.  J.   Topographic  and  meteorological  factors   influencing air  pollution  in
     Cincinnati.  AMA Arch. Ind. Health 19:44-54, 1959,

Shepherd, M.  B., G. C. R. Carey and, J. J. Phair:  Critical evaluation  of  a filter-strip smoke
     sampler used in domestic premises.  AMA Arch. Ind. Health  17:236-252, 1958.

Spedding, D.  J.   Sorption of sulphur dioxide  by  indoor suraces — II.  Wood.  J,  Appl. Chem.
     20:226-28,  1970.

Spedding, D.  J., and  R.  P. Rowlands.  Sorption  of sulphur dioxide  by indoor  surfaces —  L
     Wallpapers.  J. Appl.  Chem. 20:143-146, 1970.
              s,
Spedding, D.  J., R.  P.  Rowlands,  and  J.   E.  Taylor.   Sorption of sulphur dioxide by  indoor
     surfaces — III.  Leather.  J. Appl.   Chem. 21:68-70,  1971.

Spengler, J.  D.  [Unpublished  data from  the  ''six-cities  study."]   Harvard School of  Public
     Health,  Department of  Environmental, Health Science(s,  Boston^ MA, 1980.
                                            5-150

-------
Spengler,  J.  D.,  and  0.  W.  Dockery,   Long term  measurements  of  respirable  sulfates and
     particles  inside  and outside  homes.  1979  Electro Professional  Program,  New York, NY,
     April 25, 1979.  Paper no. 2413.

Spengler, J.  D.,  and T. Tosteson.   Personal  exposures to respirable particles.   Presented at
     Environmetrics  81, SIAM  Institute  for  Mathematics  and Society,  Washington,  DC,  April
     8-10, 1981.

Spengler, J.  D., B.  G. Ferris,  Jr.,  D. W.  Dockery, and F.  E.  Speizer.   Sulfur dioxide and
     nitrogen dioxide  levels  inside and  outside  homes and the implications on health  effects
     research.  Environ. Sci. Techno!. 13: 1276-1280,  1979.

Spengler, J.  D.,  D.  W.  Dockery,  M.  P.  Reed,  T.  Tosteson, and P.  Quinlan.  Personal exposures
     to  respirable  particles.   Presented at  the  73rd Annual Meeting,  Air Pollution  Control
     Association, Montreal, Quebec, .Canada, June 22-27,  1980.  Paper  no. 80-61.5b.

Spengler, J.  D.,  D.  W.  Dockery, W. A. Turner, J. M. Wolfson, and  B.  G.  Ferris, Jr.  Long-term
     measurements  of respirable  sulfates  and  particles  inside  and outside  homes.   Atmos.
     Environ.  15:2330,  1981.

Spicer,  C.  W.   Photochemical   atmospheric  pollutants  derived from  nitrogen  oxides.   Atmos.
     Environ.   11:1089-1095, 1977.

Spicer, C. W., P. M. Schumacher, J. A. Kouyoumjian, and  D. W. Joseph.  Sampling and Analytical
     Methodology for  Atmospheric  Particulate  Nitrates.  EPA-600/2-78-067, U. S.  Environmental
     Protection Agency, Research Triangle Park, NC, April 1978.

Steigerwald,  B.  J.    Particulate  control  strategy.    Office  of  Air  Quality  Planning and
     Standards,  EPA,  memorandum  to  Roger Strelow,   Asst.  Administrator  for  Air  and  Waste
     Management,  July  2,  1975;  submitted  by  EPA to the  Subcommittee,  House  Committee on
     Interstate and Foreign Commerce.

Sterling, T.  D. , and  D.  M.  Kobayashi.   Exposure to  pollutants  in  enclosed "living spaces."
     Environ.  Res. 13:1-35, 1977.

Stevens,  R.  K.,  T.  G.  Dzubay,   G.  Russwurm,   and   D.  Rickel.   Sampling  and  analysis of
     atmospheric sulfates  and  related species.    In:   Sulfur  in the Atmosphere, Proceedings of
     the International  Symposium,  United  Nations  Environmental Program,  Dubrovink, Yugoslavia,
     September 7-14, 1977.  Atmos.  Environ. 12:55-68,  1978.

Stevens, R. K.,  T.  G.  Dzubay, R.  W. Shaw, Jr., W. A.  McClenny, C. W. Lewis, and  W. E.  Wilson.
     Characterization  of  the  aerosol  in the Great  Smoky Mountains.   Environ.   Sci. Technol.
     14:1491-1498, 1980.

Stevens,  R.  K.» R.  J.  Paur,  C. W.  Lewis,  T.  G.  Dzubay, and W.  J.  Courtney.  Houston  aerosol
     characterization:   EPA  measurements.   In:   Proceedings of  a Specialty  Conference, Air
     Pollution  Control  Association, Houstonj,  Texas,   October 14-17,  1979.  E.   R. Frederick,
     ed., Air Pollution Control Association,  Pittsburgh, PA,  1979.  pp.  212-220.

Stevens,  R.  K., W.  A.  McClenny,  T. G.  Dzubay,  M.  A. Mason,  and W.  J. Courtney.  Analytical
     methods  to  measure   the   carbonaceous   content  of aerosols.   In:   Particulate   Carbon:
     Atmospheric  Life Cycle,  Proceedings of  an  International  Symposium,  General Motors Re-
     search Laboratories,  Warren,  MI, October  13-14, 1980.   G.   T. Wolff  and  R.   L. Klimisch,
     eds. , Plenum Press, New York,  NY, 1982.  Pp. 111-129.

Suck,  S.  H.,  E.  C.  Upchurch,  and  J.  R. Brock.   Dust transport in  Maricopa County, Arizona.
     Atmos. Environ. 12:2265-2271,  1978.
                                            5-151

-------
Suck,  S.  H.,  E.  C.  Upchurch,  and J. R. Brock.   Dust Transport in Maricopa  County,  Arizona.
     EPA-600/3-79-082,  U.  S.   Environmental  Protection  Agency,  Research  Triangle Park,  NC,
     September  1979.

Tabor,  E.  C.,  T.  E. Hauser,  J.  P.  Lodge,  and  R.  H.  Burttschell.   Characteristics of  the
     organic particulate matter  in  the atmosphere of certain cities.   AMA Arch.  Ind.  Health
     17:58-63,  1958.

Talcott,  R., and  E.  Wei.  Airborne  mutagens bioassayed in Salmonella typhimurjym.   J.  Natl.
     Cancer  Inst. 58:449-451, 1977.

Tanner, R.  L.   An ambient .experimental  study of phase  equilibrium  in  the  atmospheric system:
     aerosol  H+,  NH.+,  SO,  ,  NO,—NH,(q), HN03(g),   Presented at the  180th National  Meeting,
     American Chemical Society, Las Vegas,  NV, August 24-29,  1980a.

Tanner, R. L.   Sulfur and nitrogen compounds  in  urban aerosols.   In:  Aerosols:   Anthropogenic
     and  Naturan Sources and Transport,  Proceedings of  a Conference, New York, NY  January
     9-12, 1979.  Ann. N.Y. Acad. Sci. 338:39-49,  1980b.

Tanner, R.  L.,  R. Cederwall, R.  Garber, D. Leahy, W.  Marlow, R. Meyers,  M.  Phillips,  and L.
     Newman.  Separation  and analysis of airborne sulfate at ambient  concentrations.   Atmos.
     Environ. 11:955-966, 1977.

Tanner, R.  L.,  W. H. Marlow, and  L.  Newman.   Chemical  composition  correlations  of size frac-
     tionated sulfate in  New York City aerosol.   Environ.  Sci.  Techno!. 13:75-78, 1979.

Terabe, M.   Air  pollution  in  Japan:   present  state area trends.   Kuki  Seijo 15:1-10,  1977.

Thomas, J.  F.,  M. Mukai, and B.  D.  Tebbens.   Fate of airborne benzo(a)pyrene.   Environ.  Sci.
     Technol, 2:33-39, 1968.

Trijonis, J.,  J.  Eldon,  J.  Gins,  and  G.  Berglund.   Analysis  of  the  St.   Louis  RAMS Ambient
     Particulate  Data.   Vol. I - Final  Report.  EPA-450/4-80-006a,  U.S.   Environmental  Pro-
     tection Agency,  Research Triangle Park, NC,  February 1980.

U.S.  Environmental  Protection  Agency.   Deputy  Assistant Administrator's  Report on  Ambient
     Monitoring  Activities—Air Portion.   U.S.   Environmental  Protection  Agency,  Washington,
     DC, 1980a.

U.S. Environmental  Protection  Agency.   Inhalable Particulate Network Annual Report:   Opera-
     tion and Data Summary.   (Mass concentrations only), April  1979 - June 1980.   EPA-600/4-
     81-037, U.S. Environmental Protection  Agency, Environmental  Monitoring  Systems Laboratory
     Research Triangle Park, NC, May  1981.

U.S. Environmental  Protection  Agency.   National  Air  Quality, Monitoring  and  Emissions  Trends
     Report, 1977.   EPA-450/2-78-052, U.S.  Environmental  Protection Agency, Research  Triangle
     Park, NC, December 1978.

U.S. Environmental Protection Agency.  National  Air Quality and Emissions Trends  Report,  1976.
     EPA-450/1-77-002,  U.S.   Environmental   Protection  Agency,  Research  Triangle  Park,  NC,
     December, 1977.

U.S. Environmental  Protection  Agency.  Trends in  the Quality of  the Nation's  Air—A Report to
     the People.   OPA16/9,  U.S. Environmental Protection Agency, Office  of Public  Awareness,
     Washington, DC,  October 1980b.

Van Duuren, B. L.  Epoxides, hydroperoxides, and peroxides  in air pollution.   Int.  J.  Environ.
     Anal. Chem. 1:233-241, 1972.
                                            5-152

-------
Van Vaeck,  L.,  and  K.  Van  Cauwenberghe.   Cascade impacted measurements  of the size distri-
     bution  of  the  major classes  of organic  pollutants  in  atmospheric  particulate matter.
     Atmos. Environ. 12:2229-2239, 1978.

Van Vaeck,  L. ,  and K.  Van Cauwenberghe.   Differences  in particle size distributions on major
     organic pollutants in ambient aerosols in  urban, rural and seashore areas.  Environ. Sci.
     Techno!. 13:1494-1502, 1980.

Walsh, M.,  A. Black,  and A.  Morgan.  Sorption  of SOp by typical indoor sources  including wool
     carpets, wallpaper, and paint.  Atmos. Environ. 11:1107-1111, 1977.

Watson, J. G., Jr.  Chemical  Element Balance-Receptor Methodology for Assessing  the Sources of
     Fine and Total  Suspended Particulate Matter in Portland, Oregon.   Ph.D.,  Thesis, Oregon
     Graduate Center, Beaverton, OR, 1979.

Weatherly, M. L.  [Untitled.]  Air Water Pollut. 10:404-409, 1966.

Wesolowski, J.  J. ,  C.  P.  Flessel, S. Twiss,  J. Cheng, R. Chan,  L.  Garcia, J.  Ondo, F. Fong,
     and  S.  Lum.   The  chemical and biochemical characterization of particulate  matter as part
     of  an  epidemiological   cancer   study.   Proceedings:   1980  Conference  on Aerosols  in
     Science,  Medicine  and  Technology — Physical  and  Chemical  Properties  of  Aerosols.
     Gesellschaft fgr Aerosolforschung, Schmallenberg, West Germany, 1980.

Whelpdale,  D.  M.   Large  scale  atmospheric  sulfur  studies  in  Canjida.   Atmos.  Environ.
     12:661-670, 1978.

Whitby,  K.  T.   Aerosol formation  in  urban plumes.   Ann. N. Y.  Acad.  Sci. 338:258-275, 1980.

Whitby,  K.  T.,  A.  B.  Algren, R. C. Jordan, and J. C. Annis.   The ASHAE air-borne dust survey.
     Heat. Piping Air Cond. _29:185-192, 1957.

Wilson,  J.  C.,  K.  T.  Whitby, V.  A.  Marple,  and J. E.  McCormack.  Characteristics and origins
     of coarse particles in the air of northeastern Minnesota.  Particle Technology Laboratory
     Publication No. 390, University of Minnesota, Minneapolis, MN, April  1979.

Winchester, J. W., W. H. Zoller, R. A. Duce, and C. S.  Benson.  Lead and halogens in pollution
     aerosols and sndw from Fairbanks, Alaska.  Atmos.  Environ. 1:105-119,  1967.

Wolff, G. T., P. J. Groblicki, S. H. Cadle, and R. J. Countess.  Particulate carbon at various
     locations  in  the  United  States.    In:   Particulate Carbon:   Atmospheric  Life  Cycle,
     Proceedings of an Internation Symposium, General Motors Research Laboratories, Warren MI,
     October 13-14,  1980.  G.  T. Wolff and  R.  L.  Klimisch, eds., Plenum  Press, New York, NY,
     1982.  pp.  297-315.

Yocom, J.  E. ,  E.  T. Brookman,  R.  C.  Westman,   and 0. Ambardar.  Determining the contributions
     of traditional and non-traditional sources of particulate matter.  J.  Air Pollut. Control
     Assoc., 31:17-23, 1981.

Yocom,  J.  E. ,  W.   A.  Cote,  and W.  L.  Clink.   Summary Report—Study of Indoor-Outdoor Air
     Pollution  Relationships.   National   Air  Pollution Control  Administration  Contract No.
     CPA-22-69-14,  TRC—The   Research Corporation  of  New  England,   Hartford,  CT,  May 1970.

Yocom, J.  E., W.  L. Clink, and W. A. Cote.  Indoor/outdoor air quality relationships.  J. Air
     Pollut. Control Assoc. 21:251-259, 1971.
                                             5-153

-------

-------
                   6.  ATMOSPHERIC TRANSPORT, TRANSFORMATION, AND DEPOSITION

6.1  INTRODUCTION
     The preceding  chapters discussed  the  physical  and chemical properties  of  sulfur oxides
and  PM  (Chapter  2);  methods  of  measuring them  (Chapter  3);  their  sources and  emissions
(Chapter 4); and  measurements  of ambient levels in  urban  and rural environments (Chapter 5).
These  chapters  also discussed  information  relevant to  parts of this  chapter;  whenever pos-
sible, reference  is  made to resource material  in  previous  chapters.   Emissions, which can be
viewed as a  model  input, are discussed only  in the  context of their relevancy to air quality
simulation modeling in the final section.
     This chapter reviews our knowledge of the physical and chemical processes that contribute
to the transport  and diffusion, transformation and  deposition  of  PM and sulfur oxides in the
atmosphere  and  discusses  the  theoretical  approaches  for  integrating  these processes  with
source  emission  contributions   through  the  use  of mathematical  models.    Such  integrating
approaches  help  improve understanding  of  the complex processes  that  operate in  polluted
atmospheres.  These source-receptor  relationships  provide  a  credable scientific basis  for
determining  the nature  and  extent of emission  control  required to meet specified ambient air
quality  levels.
     The concentration  of  a pollutant species at a fixed point in time and space after it has
been  emitted from a source at  another  given point depends on  four fundamental  factors:   (1)
emission—the rate  of  pollutant emitted and the configuration  of  its source; (2) transforma-
tion—the  chemical   and  physical  reaction  processes  that  convert  one pollutant  species  to
another; (3) transport and diffusion—the movement and dilution of a pollutant species through
time  and space  as  a  result of various meteorological variables; and  (4)  deposition—the re-
moval of pollutant species through their interaction with land and water surfaces (dry deposi-
tion) and through interaction with precipitation or cloud droplets (wet deposition).
     Figure  6-1 schematically  illustrates  the principal process pathways  of airborne pollut-
ants.  Ideally, each of these processes should be treated explicitly in any air quality simu-
lation model, but generally this is not the case.
     The modeling  approaches  discussed here include explicit treatments of the dynamic physi-
cal and  chemical atmospheric processes that simulate relationships between pollutant emissions
and ambient  air quality.  More  implicit statistical-empirical approaches, which deduce source
contribution  through analysis  of  empirical information only,  are not  within the  purview of
this chapter.  A brief discussion of source-apportionment techniques that have shown consider-
able  promise in developing source-receptor relationships for  particulate  matter is presented
in Section 5.7 of Chapter 5.
6.2  CHEMICAL TRANSFORMATION PROCESSES
     Chapter 2  presents a  detailed discussion  of the chemistry of S0«  and other  gases that
react  to form PM in the atmosphere.   Section  6.2.1 summarizes the results of the atmospheric
                                             6-1

-------
FREETROPOSPHERIC
    EXCHANGE
                                         VERTICAL
                                         DIFFUSION
                         AEROSOL
           "T BY WIND*1  COMPENSATION

                       COAGULATION

                    CHEMICAL REACTIONS
        SEDIMENTATION
          AS AEROSOL
                            ABSORPTION IN
                           CLOUD ELEMENTS
    NATURAL
    SOURCES
                     DRY DEPOSITION ON   »//////, 777/7,,77///7
                        THE GROUND     ////^/.V'/////'///.'//

                               O
ANTHROPOGENIC
   SOURCES
ABSORPTION IN
PRECIPITATION
                                   WASHOUT IN PRECIPITATION
          Figure 6-1.  Pathway processes of airborne pollutants.

          Source: Adapted from Drake and Barrager(1979).
                              6-2

-------
chemical transformation processes of SCL and PM presented in Chapter 2.   Section 6.2,2 reviews
the status of  field  measurements of the rate of S0? oxidation in industrial and urban plumes,
and their contribution  to  elucidating the transformation pathway processes of SO- oxidation.
6.2.1  Chemical Transformation of Sulfur Dioxide and Particulate Hatter
     Present understanding  of the  homogeneous  gas  phase reactions  of S0?  indicates  that the
rate of S0?  oxidation  in the atmosphere is dominated by free radical  reaction processes.   The
free radical  species important  to  the SOp  oxidation process  are  HO,  HOp,  CHgO-,  and other
organic peroxyl species (RCL, R'0?,  etc.).   The concentrations of these radicals in the atmos-
phere depend on  many factors, the more important of which  are the  concentrations of volatile
organic compounds  and nitrogen oxides  (NO  and  NCL) in the  atmosphere, temperature  and solar
intensity.   Theoretical estimates have shown that maximum S0? oxidation rates of 4.0 percent/h
are possible  in polluted  atmospheres.   Recent experimental  rate constant  determinations for
the H0?  and CH,0, reactions with S0?, however,  indicate that these processes may  not be as
important as previously  thought,  and that the maximum,possible homogeneous S0? oxidation rate
under optimum  atmospheric  conditions may  only be of the order of 1.5 percent/h.  This rate is
a result of S0? reaction with HO radical only.
     Present knowledge of  heterogeneous pathways to S02 oxidation in the atmosphere indicates
                                                         2+    3+
that the  liquid phase  catalyzed oxidation  of  S0?  by Mn  ,  Fe  , and carbon are potentially
important processes,  as is  oxidation by  hydrogen  peroxide.  Theoretical  estimates  of atmo-
spheric S02  oxidation  rates by these processes  are of the   order of  10  percent/h.   Unfortun-
ately,  the actual availability of these catalyzing substances in ambient fine PM is uncertain.
The quantitative  determination of  rates  of  S0?  oxidation  by these  processes  has  never been
demonstrated under actual atmospheric conditions.
     Processes  that  form organic and nitrate particles  are thought  to be  dominated  by homo-
geneous gas  phase reactions.   In the case of atmospheric  nitrates,  a  significant production
pathway is  through reaction between HO free radical and N0_, resulting in nitric acid (MONO,)
formation.   The fate of nitric acid  in the atmosphere is not well understood, though a portion
of  gaseous  nitric acid  is  known to enter  into  an equilibrium with NH^  to form particulate
NH.NO_.   Information on  the production  rates  and  mechanism details of  organic particulate
matter is limited.  Available product information indicates   that oxidation reactions involving
the interaction  of ozone,  nitrogen   oxides,  and  HO  free  radicals with higher molecular weight
organics represent a major pathway to organic particle production.
6.2.2  Field Measurements on the Rate of Sulfur Dioxide Oxidation
     The majority of atmospheric S0? oxidation studies  have been carried  out  only  in recent
years,  and  most have involved power  plant  plumes.   One  reason for the late start in this re-
search  was  the  lack of adequate measurement technology for particulate  sulfur,  but recent
developments  (e.g.,  Huntzicker  et  al., 1978;  Cobourn et al. , 1978) seem  to have alleviated
this  problem  (see  Chapter  3).   Table 6-1  summarizes  S0?  oxidation  rates,  based  on field
measurements  in  power  plant, smelter,  and  urban  plume studies carried out from 1975 to 1980.
                                             6-3

-------
        TABLE 6-1.  FIELD MEASUREMENTS ON THE RATES OF S02 OXIDATION IN PLUMES"
Plume type/

 location
 SO,, oxidation

rate (percent/h)
          Method
     Reference
Power plants
KeybLone
  (Pennsylvania)

L abaci ie
  (Missouri)
    0-10
 0.41-4.9
Four Corners     0.27-0.84
  (New Mexico)
Labadie and
Portage des Sioux
  (Missouri)
Muscle Shoals
  (Alabama)
Kycjer Creek
  (Ohio)
LabaiUe
  (Missouri)

Four Corners
  (New Mexico)

Labadie
  (Missouri)

Cumberland
  (Tennessee)

Great Canadian
Oil Sands
  (Alberta, Canada)
Keystone
  (Pennsylvania)

Centralia
  (Washington)
Four Corners
  (New Mexico)

Four Corners
  (New Mexico)
   0-5
   0-3
   2-8
   0-4
   0-7
   0-3
   0-5
   0-6
0.15-0.5
32  34
  S/  S,ratio, change with
 oxidation

Total change in particle
 volume

Submicron sulfate and
S02 - change of ratio
with time
Newman et al.  (1975)
Cantrell and Whitby (1978)
                                                Ursenbach et al.  (1977)
Particulate sulfur to total   Forr%st and Newman (1977)
 sulfur ratio
Particulate sulfur to total   Gillani et al.  (1978)
 sulfur ratio

CCNb production (CCN to S02   Pueschel and Van Valin (1978)
 ratios)

Particulate sulfur to total   Husar et al. (1978)
 sulfur ratio

Particulate sulfur to total   Meagher et al.  (1978)
 sulfur ratio

Particulate sulfur to total   Lusis et al. (1978)
sulfur ratio
Particulate sulfur to total   Dittenhoefer and dePena (1978)
sulfur ratio
Total change in particle
 volume
Hobbs et al. (1979)
CCN  production (CCN to S02   Mamane and Pueschel (1980)
 ratios)
                                           6-4

-------
                                  TABLE 6-1.   (continued)
Plume type/
 location
 S0« oxidation
rate (percent/h)
Method
Reference
Leiand-Qids
  (North Dakota)
Sherburne County      0-5.7
  (Minnesota)
Big Brown
  (Texas)

Smelters

INCO Nickel           0-7
  (Copper Cliff,
   Canada)

INCO Nickel         1.2-5.2
  (Copper Cliff,
   Canada)

MT ISA Mines           0.25C
  (MT ISA, Australia)

Urban
                  Total  change in particle
                   volume
                  Particulate sulfur to total
                   sulfur ratio
                  Particulate sulfur to total
                   sulfur ratio
                  Particulate sulfur to
                   lead ratio
                    Hegg and Hobbs (1980)
                    Lusis and Wiebe (1976)
                    Forrest and Newman (1977)
                    Roberts and Williams (1979)
Los Angeles
  (California)

St. Louis
  (Missouri)

St. Louis
  (Missouri)
   1 2~13         Particulate sulfur to total
                   sulfur ratio

     7-12.5       Particulate sulfur to total
                   sulfur ratio

   3.6-4.2        Particulate sulfur to total
                   sulfur ratio
                    Roberts and Friedlander (1975)
                    Alkezweeny and Powell (1977)
                    Chang (1979)
 Adapted in part from Hegg and Hobbs (1980)
 Cloud condensation nuclei
cDiurnal average rate
                                           6-5

-------
The  rates  of SO,  oxidation in  industrial  plant  plumes consistently range from  0  to 10 per-
cent/h, with  urban  plumes showing only a  slightly  greater maximum rate of 13 percent/h.  The
pre-1975 studies  (Gartrell  et al., 1963;  Dennis  et al., 1969; Weber,  1970;  and Stephens and
McCaldin, 1971), which observed conversion rates an order of magnitude larger than more recent
observations, are suspect due to possible artifact formation in the sulfate analysis technique
and limitations in the analytical methods in general.
     Newman  (1980)' reviewed the  majority of  the  power plant and  smelter  plume studies pre-
sented in Table 6-1 and arrived at the following conclusions:
     1.   The diurnal  average  oxidation rate  of  S09  to sulfate is probably  less than 1 per-
          cent/h.                                    £
     2.   Little or no oxidation of SCL occurs from early evening to early morning.
     3,   Maximum oxidation rates of S0? to sulfate of 3 percent/h can occur under midday con-
          ditions.
     4.   The separate contribution  of homogeneous  and heterogeneous mechanisms to S0? oxida-
          tion in plumes cannot be deduced from the present studies.
     The reported SO,, oxidation rates are estimates based on analyses of measured physical and
chemical parameters  and  in  many instances have  incorporated  certain  simplifying assumptions
that are not  totally substantiated.   Also, present understanding  of  S0? chemical transforma-
tion processes indicate that SCL oxidation rates can vary significantly as a result of differ-
ences  in  the composition  of source plumes  and the  air masses into which the  plumes enter.
Typical experimental  uncertainties  in  measured S0? oxidation  rates  reported  in Table 6-1 are
50 percent, but may be greater  if  inappropriate  assumptions have  been used.   Even with these
uncertainties, the  overall  consistency in the observed range of S0? oxidation rates is grati-
fying.
6.3  PHYSICAL REMOVAL PROCESSES
     The removal  of PM  and gases  from  the atmosphere generally occurs  through two physical
processes:   (1) dry deposition—the removal of chemical species from the atmosphere at the air-
surface interface;  and  (2)  precipitation scavenging—the removal of chemical  species from the
atmosphere by interaction  with  various types of precipitation such as rain, snow, etc.  These
processes have both  a positive  and negative impact on environmental air pollution issues.  On
one hand, they  constitute the  major mechanisms by  which the polluted atmosphere cleanses it-
self,  lowering  ambient air  concentrations of pollutant species and thereby  reducing health-
related risks.  On  the  other hand, the deposited pollutant materials may constitute increased
risks to our terrestrial and aquatic ecosystems.
     Since wet  and  dry removal  processes  significantly  affect the lifetime of  SO,  and  PM in
the atmosphere and thereby affect the distance traveled and the concentration of these species,
understanding  these processes  is essential  for  a proper  assessment  of their  environmental
significance.   The   removal  of   pollutant  species  by dewfall  has not  been  studied, and it
remains  for  future  research  to  determine  whether  this   process  is  an  important  removal
mechanism for atmospheric contaminants.
                                             6-6

-------
     In the sections to follow, dry deposition and precipitation scavenging are discussed with
emphasis on experimental data bases and theoretical treatment.
6.3.1  Dry Deposition
     Sehmel  (1980),  HcMahon and  Dem'son  (1979),  Chamberlain  (1980),  and  Garland  (1978) re-
viewed particle and  gas dry deposition.   The dry  deposition of S0? and PM, like other atmos-
pheric species, is governed by three major components: meteorological variables, properties of
the depositing pollutant,  and  surface variables.  These components are influenced by specific
parameters that  interact in complex ways,  which in many instances  are  not completely under-
stood.
     The  most  important meteorological  processes  affecting  dry  deposition  are  transport-
related phenomena,  which are  governed by the wind and  temperature  profiles;  eddy  diffusion;
and sedimentation  across the  boundary  layer to the vegetation  canopy  or  the  surface.   Two
meteorological parameters strongly  influence these processes:   the friction velocity (u*) and
the aerodynamic  surface roughness  (z ).   Both  of these parameters are used  to describe the
windspeed  profile  above  a  given  surface  under  given  conditions of atmospheric  stability.
Typically, these two variables are determined empirically by  fitting  optimal  curves to wind-
speed data as a function of height.  The strong diurnal dependence of dry deposition is linked
to the  formation  of a stable  layer of air at the earth's surface  at  night (nocturnal inver-
sion) that effectively inhibits the vertical transport  of pollutant species to the canopy or
the surface.   The formation  of the  nocturnal  inversion and  its  effect on other atmospheric
processes is discussed in the section on transport and diffusion.
     Solubility in water is an important property influencing the dry deposition of a pollut-
ant.   Other  important  characteristics of PM are  size  distribution, density,  morphology, and
composition.   Among  the  important surface properties are: (1) the moisture content of the sur-
face  which,  in conjunction with  the  solubility  of the  pollutant  species,  govern the overall
sticking  efficiency  of the deposited material; and (2) the physiological state of the vegeta-
tion  surface,  especially the opening and closing of stomatal pores, where the rate of pollut-
ant uptake is thought to be strongly governed.
     Chamberlain and Chadwick  (1953) introduced  a convenient  way to express  the rate of dry
deposition of both  gases  and  particles  in terms of velocity.   Dry deposition velocity (V )
(defined  as  the  downward flux  (F) of the species, divided by its ambient concentration (x) at
some  specified  height [typically 1 to 2  meters  above the surface]), is the standard form in
which  all measured  deposition  rates  are  reported.   Dry deposition velocities typically are
reported  in  units of cm  sec

                            v    =   "F
                            g       *

      Dry  deposition  velocity is positive by convention  and therefore requires a minus sign on
F, the  downward flux, which  is  defined as negative.
                                             6-7

-------
     Hicks et al. (1980) reviewed and evaluated the measurement techniques for the dry deposi-
tion of pollutant  species.   They sorted measurement  methods  into  three major categories:  (1)
estimates of accumulation, (2) flux monitoring, and (3) flux parameterization.
     Though none of  the experimental techniques has proven to be useful in all dry deposition
measurements, a  general  consensus  has been reached on the overall  accuracy of the methods  and
their suitability  for specific applications.   Based  on Hicks et al.  (1980),  the  three cate-
gories are described briefly, with general comments on their limitations.
     Estimates  of  accumulation  may  be  considered  using atmospheric  radioactivity or  mass
balance methods.   Radioactive techniques compare ambient  concentrations  of  selected radioac-
tive species  with  concentrations in  water  bodies,  vegetation, etc.,  to  evaluate  the rate at
which material  enters  the ecosystem over long periods.  The technique generally is limited to
small particle  uptake  of long-lived species and has difficulty distinguishing between dry and
wet  removal and resolving short-term variations.  Mass balance studies attempt to measure the
various inflow  and  outflow processes in the  ecosystem,  with  the exception of dry deposition,
which is  then  determined through a budget calculation.  The method's major limitation is that
dry deposition  is inferred by indirect measurements, which in themselves are difficult to make
accurately.
     Flux monitoring  considers  the  direct measurement of total deposition over a well-defined
surface  for set  periods.   Several  types  of  deposition  surfaces  have  been used  with  this
general category,  including  open  pots,  flat  filters,  flat  plates  and shallow  pans,  fiber
filters, and sticky films.  Overall, the methods are limited due to their lack of standardiza-
tion,  dissimilarity  to  natural  surfaces,   and potential  for contamination  by  locally  re-
suspended particles.
     Flux  parameterization  includes  a variety  of  methods,  one  of which,  eddy correlation,
shows promise  as a  measurement standard for dry deposition  of gases.   Eddy correlation  re-
quires the simultaneous measurement of the concentration of pollutant species and the vertical
component of the wind velocity at a sufficiently fast rate to determine the turbulent flux of
the  pollutant.   The  lack of adequate  fast-response instruments  for many of  the  pollutant
species of  interest significantly  limits  the  technique.   In addition,  particle  flux  due to
gravitational  settling  is not  detected,  and this can produce  invalid results if significant
particle  resuspension occurs.   Laboratory methods,  including  chamber  and wind tunnel studies
are  also  considered tinder the  flux  parameterization  category.    In  these  controlled experi-
mental studies,  plants,  leaves,  simulated canopy surfaces, etc., are exposed to known pollut-
ant concentrations.  Measurements of the change in concentration, which can be accomplished by
a variety of methods, are then used to determine pollutant uptake.
6.3.1.1   Sulfur  Dioxide Dry Deposition—The  dry deposition of SO, onto grass, crops, forests,
soil, and building surfaces  has been reviewed by Sehmel  (1980),  McMahon and Denison (1979),
Chamberlain (1980),  and Garland (1978).   McMahon and Denison (1979) presented compilations of
dry  deposition  laboratory and field measurements of S0?.  A review of these results  indicates
                                             6-8

-------
measured dry  deposition  velocities ranging from 0.04  to  3.7 cm sec  , but with  the majority
in the  range  from  0,3  to  1.6  cm sec  .   The  apparent wide  range of dry  deposition values
is not particularly disconcerting when the variety of surfaces, meteorological conditions, and
experimental methods are considered.   Table 6-2 summarizes the average dry deposition velocity
by surface type.
             TABLE 6-2.   AVERAGE DRY DEPOSITION VELOCITY OF S02 BY SURFACE TYPE


Surface
Alfalfa
Grass
Wheat
Forest
Sandy Soil
Clayey Soil
Soil
Land
Water (Fresh)
Ocean
Snow
Laboratory measurement,
v (cm sec )
1.2 (2)
__
—
—
0.6 (2)
0.8 (2)
--
—
--
—
..
Field measurement,
v (cm sec )
1.6 ( 2)
1.1 (14)
0.4 ( 3)
1.4 ( 5)
--
—
1-2 ( 4)
1.2 ( 4)
1.1 ( 6)
0.5 ( 2)
0.3 ( 2)

  Note;  Values in parentheses indicate the number of separate studies used to obtain the
         average deposition velocity.
  Source:  McMahon and Denison (1979).

     After reviewing  the  same set of data, Garland  (1978) concluded that the mean deposition
velocities for  S0? over  surfaces ranging  from  water and soil through  short  grass  to forest
                  ^                                                -^
were  very  similar and  suggested that a value of  about 0.8 cm sec   was  applicable to large
areas of Europe.
     In  a  more detailed  effort  to  estimate dry deposition velocities  of  S0?  and particulate
sulfate  over the  eastern  half  of  the  United States,  southern  Ontario,  and  nearby oceanic
regions,  Sheih  et al.   (1979)  computed  deposition  velocities  as  a   function  of  land  use
characteristics, surface  roughness scale lengths, and,surface resistances to pollutant uptake.
Gridded  dry  deposition  velocity maps  of  sulfur  dioxide  and  sulfate  corresponding  to half
degree  increments  of  longitude and  latitude were computed for a range of atmospheric stabili-
ties.  The results indicate that deposition velocity distributions for SO, are not uniform for
the  less stable atmospheric  conditions.   For very  unstable  atmospheric conditions (Pasquill
category A)  dry deposition velocities over the eastern  United States ranged from 0.4 to 0.9 cm
sec    (excluding  water surfaces)  for SO,, with a  mean areawide dry  deposition  velocity of
                         -1
approximately 0.6  cm  sec   .   Under  the same  conditions, the deposition velocity for sulfates
ranged  from  0.7 to 0  9 cm sec   with a mean value of approximately 0.8 cm sec  .  Sheih et al.
(1979)  noted that,  under nearly calm conditions at  night, stability classification  schemes do
                                             6-9

-------
not  adequately  represent the nocturnal  inversion  formed  at the surface and  recommend  that a
dry deposition velocity of 0.07 cm sec   be assumed for both SCL and sulfate particles.
6.3.1,2  Particle Dry Deposition—As  a  measurement for pollutant species of interest, the dry
deposition of particulate matter is less understood.  In Sehmel (1980) and McMahon and Denison
(1979),  deposition  velocities  for  particle  species  are  compiled for  both artificial  and
natural  surfaces.   Unfortunately,  with  the exception of  lead particles  from automotive ex-
haust, virtually no  data exist for other  particulate  pollutants,  such as sulfates, nitrates,
and  carbon-containing  particles.   The  interpretation  of the  relationship between deposition
measurements with  fallout  collectors  and dry  deposition rates  on  natural  surfaces creates an
additional  problem.   Fallout  collectors,   which  were  used in a  significant portion  of the
measurements  reported,  generally do  not  have the characteristics  of the surfaces  they are
attempting to simulate.
     Tables 6-3 and 6-4 from the study by McMahon and Denison (1979) compile literature values
for  the  deposition velocities  of  particles  measured  under laboratory  and field conditions,
respectively.   The  data cover  a  range  of surface  variables,  particle sizes  and composition,
and meteorological conditions.  A review of the data indicates:
     1.   Deposition  in nature  varies   considerably  through  processes  that are  not totally
          understood.
     2.   The minimum  deposition velocity  for particles  occurs at  diameters  from  0.1  to 1.0
          urn.
     3.   Deposition velocities  are  often  reported for particle diameters  and size distribu-
          tions that do not reflect typical atmospheric PM characteristics.
     The experimental uncertainties  associated with particle dry deposition velocity measure-
ment have  stimulated  the development of theoretical models  for simulating the dry deposition
process  and  predicting  dry deposition velocities  given specific meteorological  data (Sehmel,
1980; Slinn, 1978, 1977; Davidson and Friedlander,  1978).   The models describe only the physi-
cal  processes of  bringing  the particle to the depositing surface.    Particle shapes other than
spherical, particle composition, or surface properties have not been considered with regard to
particle retention.  Particle size, an important property in the aerodynamic flow of particles
to  surfaces,  is considered in  these  studies.   The typical model  result  shows that predicted
deposition velocities increase as surface roughness and/or friction  velocity increases and are
nearly  independent  of  atmospheric  stability.   The deposition  velocity for  particles  passes
through  a  minimum  in the 0.1 to  1  ^m diameter particle range. Figure 6-2 (Sehmel, 1980) pre-
sents  deposition  velocities predicted  by  one model at 1 m from  the surface, for  U* = 30 cm
sec   and particle densities of 1, 4, and 11.5 g/cm .
     Table 6-5  shows, the range of predicted  deposition  velocities at a height of 1 m for two
particle  size  regions  and  for a  range of aerodynamic surface roughness  lengths,  mean wind-
speeds,  and  calculated friction velocities.   These results are based on  the  model  of Sehmel
(1980) and should be representative of most meteorological and surface conditions.
                                             6-10

-------
                        TABLE  6-3.   LABORATORY MEASUREMENTS OF DEPOSITION VELOCITIES OF PARTICLES

Author
(date)
Chamberlain (1967)

Moller and Schumann
(1970)
Chamberlain and
Chadwick (1972)
Clough (1973)
Sehmel (1973)
Sehrael and Sutter (1974)
Belot and Gauthier
(1975)
Klepper and Craig (1975)
Craig et al. (1976)
Reference
v height
(cm sec ) (m)
0.03
0.03
0.1
0.8
2/3
V « D '
g
v =0.06 u*
V9=0.12 u*
0.005
0.003 0.1
0.3
2
2 x 10"3-10 0,01
5 x 10~3-29 0.01
vga"4
vgd
0.0035
0.01
Particle
diameter
((jm) Surface
0.1
i Grass
5

20-30 Cereal crops
0.08
0. 5 Filter paper
5
20
0.1-28 Smooth brass
0.2-30 Water
1-10 Shoots of pine
and oak trees
0.8 Bean leaves
0.1 1 Smooth
Comment
~—

D = diffusion coefficient ~ ,
2 x 10"^ > D > 2 x 10 cnTsec
Dry Includes wind tunnel
Wet and field data
v to copper also measured
v^ found to be a function of
* windspeed
—
—
u = wind speed
d = particle diameter
—
Wind tunnel
Wedding et al.  (1976)
Deposition rate on pubescent
leaves of sunflower was nearly
7 times that of the non-
pubescent leaves of tulip
poplar.

-------
                                                     TABLE 6-3.  (continued)

Author
(date)
Little and Wiffen (1977)

Little (1977)








vg_
(cm sec )
0.11
0.02
0.5
0.04
0.3
0.9
0.1
0.3
1.5
0.3
0.8
Reference Particle
height diameter
(m) (pro)
5 x 10"2
0.2
o 71:


5


8.5


Surface
Short grass

Nettle
Beech
White Poplar
Nettle
Beech
White Poplar
Nettle
Beech
White Poplar
Comment
__

These data are for whole
shoots for..wind speeds of
2.5 m sec . Data for other
wind speeds and separate
plant surfaces are given in
reference.




   Source:  McMahon  and  Denison (1979).
I
t—»
rsa

-------
                           TABLE 6-4.   FIELD  MEASUREMENTS  OF  DEPOSITION VELOCITIES OF PARTICLES
cr>
i

Author
(date)
Chamberlain (1953)


Eriksson (1959)

Small (1960)

Neuberger et al. (1967)

White and Turner (1970)





Esmen and Corn (1971)

_#
Chamberlain and
Chadwick (1972)
Pierson et al. (1973)


vg_
(cm sec )
2.1
1.1
0.5
0.7
1.6
0,5
(0.2-3.4)


5.6
4.7
3.0
7.1
0.8


v = 0.50
9
v = 0.06 M*
vj= 0.12 M*

0.1-0.6

Reference
height Particle
(m) composition
0.3-0.9 16 pm*
0.3-0.9 16 Mm*
0.3-0.9 16 Mm*




Ragweed

Na
K
Ca
Mg
P


0.1-10 pro*

20-30 M"i*




Surface

Grass

Ocean
Land
1 anrl
L-u 1 114
Coniferous
forest

Mixed
deciduous
woodland


Filter paper
Millipore filter
Glass slide
Cereal crops


Land

Comment
u = 9.2 m/sec
u = 3.2 m/sec
u = 1.1 m/sec
Chloride over Scandinavia
— —
Radioactive particles over
Norway
80 percent ragweed pollen re-
moved from air by forest
1. Probable overestimation of
aerosol income, hence v .
2. Standard deviation varied
between 65 and 95 percent of
mean v .
g
—


Dry, Includes wind tunnel and
Wet* field data.
v estimated for 23 trace
g
elements based on several
years of data

-------
                                                       TABLE 6-4.   (continued)
en
i

Author
(date)
Cawse (1974)











Hart and Parent (1974)


Clough (1975)







vg_
(cm sec )
1.3
0.22
(0.45)
0.50
(0.50)
1.1
0.56
(0.45)
0.30
(1.0)
0.29
0.62



3.4
7.3
11
61
100
0.74
1.1
0.75
12.7
Reference
height Particle
(ra) composition
Al
As
Cd
Cr
Cu
Fe
Mn
Ni
Pb
Ti
V
Zn
Na, Ca,
Mg, K,
P, N03


30 jjm*


4 pi"*
3jb
|jm*
Surface












Douglas fir
and junipers

Grass
Grass
Grass
Dry moss
Wet moss
Grass
Grass
Dry moss
Comment
Extracted from Gatz (1975).
Values in parentheses were
estimated by Gatz from a
relationship between particle
size and v .
g






Deposition ratio:
beneath trees _ ~ ,,.
open terrain
Dry u* = 37 cm sec ,
Dry u* = 87 cm sec ,
Wet u* = 87 cm sec


Dry u,< = 37 cm sec

Dry u* = 37 cm sec
   *particle diameter

-------
                                                   TABLE 6-4.  (continued)

Author
(date)
Abrahamsen et al. (1976)
Dovland and Eliassen
(1976)
Fritschen and Edmonds (1976)
Prahm et al. (1976)
Krey and Toonkel (1977)
Wesley et al. (1977)
vg_
(cm sec )

0.16
0.68
0.07
0.46
0.4
0.5
0.6
Reference
height Particle
(m) composition
S°4?
Atmospheric
aerosol
3 (jm*
Atmospheric
aerosol
5 0.05-0.1 |jm*
Surface
Spruce and
pines
Snow
Douglas fir
Atlantic
ocean
Bare soil
and grass
Comment
Deposition ratio:
beneath trees _ 0
open terrain
Lead
SO. : upper bound value

so;2
90S : HASL wet-dry collector
u < 2 m sec : Eddy correlatio
method.

Source:   McMahon and Denison  (1979).

-------
     =  111 HIM)    I  I I IMII|    I  I  I IIIIIj    I   I I II
                           UPPER LIMIT
                     NO RESISTANCE BELOW AND
                   ATMOSPHERIC DIFFUSION FROM
                            1 cm TO 1 m
                                \
                               STABLE ATMOSPHERE
                                WITH ROUGHNESS
                                  HEIGHT, cm
                                                  p  = PARTICLE DENSITY  —
                                                  0  = ROUGHNESS HEIGHT ~
                                                     = FRICTION VELOCITY
        I   I  I I Hill  S\  I  1 Hill    1  I  I Hill     I   I 1 I Hill    I  I  I 111II
10
                                10"'            1

                              PARTICLE DIAMETER,
    Figure 6-2. Predicted deposition velocities at 1 m for /n. = 30 cm s  and particle
                                 "3
    densities of 1,4, and 11.5 g cm""1.

    Source:  Sehmel (1980).
                                 6-16

-------
                      TABLE 6-5.   PREDICTED PARTICLE DEPOSITION VELOCITIES3

Deposition Velocity Range
Z U/UK
cm


0.1
10
0.1
10
m sec


2.
1.
11.
5.
/cm sec


3/10
2/10
5/50
8/50
0.


1.5x1-
9 0x10
2 0x10
1 0x10
1 to

cm
-2 _
-2
~2 _
-1
Particle Diameter
1

sec
5
1.
5,
2.
pro
-1

0x10
5x10
5x10
0x10



-2
-1
-2
-1
1 M


5.
1,
5.
2.
to 10 ur
-1
cm sec
0x10"? -
5xlO~2 -

Oxio"1 -
n


4
4
4
4

          .Based on model predictions in.,Sehmel (1980)
           Particle density of 11.5 g/cm

6.3.2  Precipitation Scavenging ,
     As with dry  deposition,  precipitation scavenging or wet removal results from a series of
complex physical  and chemical  interactions  involving properties of the  scavenging media and
the species removed.  For the past 30 years,  research in the area has focused on removing from
the  atmosphere  radioactive  debris  introduced  by  nuclear  weapons  testing  (Bowen,  1960;
Engelmann, 1968;  Volchok et  a!.,  1971) and  in conjunction  with  material  balance  or budget
studies on the removal of various elemental species from the atmosphere (Robinson and Robbins,
1970; Rasmussen et al. ,  1975; Junge, 1972,  1974).
     Engelmann (1968), Postma  (1970),  Hales  (1972), SI inn et al. (1978) and SI inn (1981) have
researched and reviewed  the theory of precipitation  scavenging.   Though  our understanding of
the details of the complex processes operating in precipitation scavenging is incomplete," sig-
nificant progress  has been  made in deducing the general scavenging pathways and in developing
appropriate parameters for  their quantitative treatment.  As  pointed  out by SI inn (1981) and
others, the  removal  of  trace  constituents  from  the  atmosphere by precipitation scavenging
depends on: (1)  the position of the trace consitituent relative to the scavenging media; (2)
the  physical  form  of the  scavenging media;  (3)  the chemical and  physical  properties of the
trace  constituent;  and   (4)  the specific physical/chemical process  that  is  operative   These
basic factors are schematically illustrated in Figure 6-3.
     A convenient practice in evaluation of precipitation scavenging is to distinguish between
below-cloud and  in-cloud scavenging processes.  Unfortunately, use of the terms "rainout" for
in-cloud and "washout" for  below-cloud scavenging, has  led  to confusion.  It is difficult to
clarify the  contribution  of these  processes to  the  total   scavenging during  precipitation.
Washout,  which  is easier to  study,  has received more  scientific  attention.   In  many experi-
mental studies,  the distinction between the two processes has been ignored and only the total
precipitation  scavenging has been  considered.  The  theoretical  approaches discussed  in the

                                             6-17

-------
CTl
I
                                                       \. >  1
                                                IN-CLOUD SCAVENGING
                                                     (RAIN OUTI '
                                                                            PRECIPITATION MEDIA: e.g., RAIN, SNOW
                                                                         POLLUTANT PARAMETERS: e.g., GAS, PARTICLE, SOLUBILITY
                                                                          SCAVENGING PROCESSES: e.g., GAS-LIQUID MASS TRANSFER
                                                                                                  DIFFUSION, IMPACTION
                                                    "\XV^>\^
                       n
                                                       \ \\ BELOW-CLOUD SCAVENGING
                                 Figure 6-3. Basic factors influencing precipitation scavenging.
                                 Source: Adapted from Slinn (1981).

-------
following section  are  for washout processes only, while  the empirical parameterizations con-
sider total  gas and particle scavenging.
     The parameterization of the precipitation scavenging process has generally taken the form
of a  loss rate per unit volume and has evolved from various assumptions applied to the conti-
nuity equation  (SI inn,  1977)    Parameterizations for the removal of gases by rain and the re-
moval of particles  by  rain,and by snow are considered in the following sections.  The formal-
ism and  technical  rigor used in their development are discussed elsewhere (Hales, 1972, 1978;
Slinn, 1977) and are beyond the scope of the present discussion.
6.3.2.1 Sulfur Dioxide Wet Removal—The removal of SO- from the atmosphere by rain is governed
by basic physical  processes of absorption and desorption of the SO- molecules from the hydro-
meteor  (Hales, 1972,  1978) and  by  a  series  of chemical  reactions  (Postma, 1970;  Hill  and
Adamowicz,   1977;  Barrie,  1978)  that account  for  the  liquid  phase  oxidation  of SO-.  As  a
physical process,  the  rate of scavenging of a  gas  by rain  is a function of the size spectrum
of the  rain  droplets,  the fall path of the rain droplet to the ground, the rainfall rate, and
the solubility of the gas.  Hales (1978) and Slinn et al.  (1978) developed general expressions
for  computing  the  scavenging  of  SO,,  and  other gases,  given simplifying assumptions  on the
character of the  precipitation and solubility of the  gas.   As  pointed out  in  Chapter  2, the
liquid phase oxidation  of SO- is complex and not thoroughly understood.  The uptake of SO- by
rain  droplets  proceeds  through the dissolution of SO,, and a subsequent series of dissociation
reactions.    The  chemical   equilibria  and  associated  reaction  rates   for  the  dilute  sulfur
dioxide-pure water system  are  well   known  and the. reactions  sufficiently fast  that thermo-
dynamic equilibrium between gas phase and liquid droplet phase can be assumed.  Therefore, the
physical dissolution of SO- in rain droplets follows a Henry's Law relationship that predicts
reversible absorption  to  a degree that depends on the physical  and chemical  state of the rain
droplet and the ambient atmosphere through which it travels  (Hales, 1972).  Treatment of these
processes  is  relatively  straightforward,   but  there  is  uncertainty  in  the SO-  wet removal
process  because of  the effect of trace  substances  on the SO- dissociation equilibria and the
effective oxidation rate of SO-  in  solution.   In  this light,  several  recent studies by Hill
and Adamowicz  (1977),  Barrie (1978), Garland  (1978),  and"Gravenhorst  et al.  (1978) have con-
sidered  the 50,,-bisulfite  oxidation  process  in predicting wet  removal  rates of  SO,, under
various atmospheric conditions.
      Barrie  (1978)  modeled the washout  of  SO- from a plume  under  varying meteorological and
SO-  concentration  conditions.    He   assumed  that  SO- oxidation  in   the  raindrops  could  be
neglected due  to  the limited time for reaction (0 to 5 min), the low pH's encountered (Beilke
et al.  1975),  and the  dissolution of the SO- governing pH of the raindrop.  He concluded that
the  fractional  plume washout rate (percent/mm rain) is inversely related to the plume concen-
tration  and thickness.   That  is,  for  a   given  precipitation   rate,  the plume  washout rate
(percent/h)  increases  with decreasing plume concentration or decreasing plume thickness.  For
heavy rain  (25 mm/h),  washout from a 1000 ppb(v) SO- plume of 20 m thickness occurs at a rate
                                             6-19

-------
of 56 .percent/h,  while under drizzle conditions (0.5 mm/h) for a 300 ppb(v) S0? plume of 50 m
thickness the rate was 2 percent/h.
     In a more  explicit treatment of S0? washout, Hill and Adamowicz (1977) accounted for the
effects of S02  oxidation within rain droplets and of the pH of precipitation on the SO,, wash-
out process.  They  indicated that pH can  be  quite variable (over six  orders  of magnitude in
hydronium ion concentration,  equivalent to six pH  units)  at  SO,, ambient levels of 10 ppb and
less.   As SO-  levels increase,  the variability in background pH decreases.   The SO,, oxidation
rate of 3.6  percent/h used in the calculations is based on the catalytic oxidation studies of
Brimblecombe and  Spedding  (1974).   In a typical  calculation  of the rate of SO,, washout, Hill
and Adamowicz (1977)  assumed various ambient SO,, concentrations, well mixed through a layer 1
km in  depth,  and  a rainfall rate of 1 mm/h, with a predominant drop radius of 0.5 mm and a pH
of 7.   Calculated washout  rates of SO,, under these conditions were 2.6 percent/h and 0.8 per-
cent/h for ambient S02 levels of 10 ppb and 100 ppb (30 and 260 (jg/m3). respectively.
     A convenient empirical expression for the wet removal of gases takes the form of an expo-
nential decay process, where the time constant for decay (scavenging coefficient for the gas,),
determined in field  and laboratory studies,  is a function of the rainfall  intensity. The ex-
pression takes the form
                                   (-At)
                           xt = V
where xt and  x_ are the atmospheric concentrations of the gas at time t and zero, respective-
       v      O                                                                           j
ly, and  A  is  the scavenging  coefficient for the  gas.   Chamberlain  (1953),  Beilke (1970),
Hales et al.  (1971),  Dana et al.  (1975)  and  others have reported estimates of the-scavenging
coefficient for sulfur dioxide.   Calculated scavenging rates  of  S02 using  these coefficients
can range typically from 2 percent/h to 22 percent/h.
6.3.2.2  Particle Wet  Removal—The study  of precipitation scavenging of particles has focused
on theoretical  studies (Slinn,  1977; Grover et al., 1977; Wang et al., 1978), but emphasis in
experimental work has  taken hold in  recent years (Dana and Hales,  1976; Radke  et al. , 1980;
Gatz,   1977).  The wet removal  of sulfate  PM  in the ambient environment has been of particular
interest (Scott,  1978; Hales,  1978;  Dana, 1980) due to the acidity of many of these particles
and the increased concern for the phenomenon termed "acid rain."  Acidic precipitation and its
associated scientific  issues are discussed in Chapter 7.
     As with  gases,  the size spectrum of  the rain droplets, the fall path of the-rain droplet
to the ground,  and the rainfall rate, as  well as the size distribution and composition of the
particulate  matter  affect  the  particle  scavenging  rate  by  precipitation.   Slinn  (1977)
developed general  expressions  for computing the scavenging of particles given certain simpli- .>
fying assumptions, as  has been done with gases.
     A practical  approach  in predicting the wet removal of particles, as mentioned previously
for gases, has been through the measurement of empirical scavenging coeFficients.  McMahon and
Denison  (1979)   compiled  a  comprehensive   list  of  field  measurements   of  wet  scavenging
coefficients  of particles, which is  presented  in Table 6-6.   A cautionary  note is  in order.

                                           6-20

-------
TABLE 6-6.   FIELD MEASUREMENTS OF SCAVENGING COEFFICIENTS OF PARTICLES

Author
(date)
Kalkstein et al . (1959)
Georgii (1963)
Banerji and Chatter jee (1964)
Makhon'ko (1964)
Shirvaikar et al. (1960)
Makhon'ko and Dmitrieva
(1966)
Makhon'ko (1967)
Wolf and Dana (1969)
Bakulin et al. (1970)
Burtsev et al . (1970)
Dana (1970)
Perkins et al. (1970)
Peterson and Crawford
(1970)
Esmen (1972)
AP
(sec"1)
2 x 10~|?
2 x 10"5
4 x 10~|?
22 x 10";?
4 x 10
0.4 x 10"5
2 x 10~j?
< 1 x 10"3
7 x 10"5
20 x 10"5
7 x 10"5
0.5 x 10"5
3 x 10"5
15 x 10~|?
20 x 10
13 x 10"5
300 x 10"5
16 x 10"5
0.4 x 10~5
Particle
composition
Cl^NO™-'
Dissolved
inorganic
contaminant
Radon
Fission
products
Atmospheric
dust
Fission
products
Atmospheric
dust
J 0.5 urn*

J0'l ^0.2 urn*
JU'D -0.2 urn*
J 7.5,3 urn*
Atmospheric
aerosol
,0.5 ,- *
J 5 urrr
Atmospheric
aerosol
Comment
Rainout As-. calculated by
Washout Makhon'ko (1967)
Rainout
Washout
Rainout As calculated by
Makhon'ko (1967)
Rainout
Washout
Rainout As calculated by
Makhon'ko (1967)
Rainout
Rainout plus washout
Snowout, see also Knutson
and Stockham (1977)
Pb; washout from
thunderstorm
Washout
Rainout
Uranine and rhodamine
particles respectively
Rainout
Based on Engelmann's
data (1965)
Includes rainout
                                    6-21

-------
                                    TABLE 6-6.   (continued)

Author
(date)
Rodhe and Grandell (1972)
Acres-AESC (1974)
Graedel and Franey (1975)
Hicks (1976)
Graedel and Franey (1977)

Radke et al. (1977)

AP
(sec"1)
Particle
composition
Comment

Suggest A proportional
to rainfall intensity
0.7
snow
50
19
18
28
43
65
92

x 10
= 25-50Ar
x 10"5
x 10 R
x 10"^
x 10"?
x 10 ^
x 10~?
x 10 D

Atmospheric
aerosol
•ain O-4"1 Mm*
< 1 M"1*
0.3-0.5 urn*
0.5-0.7 |jm*
0.7-0.9 Mm*
0.9"1.5 M"1*
1.5-3 Mm*

Includes rai
nout
See SI inn (1976)
Rainout
Condensation
Snow

See Fig. 6.4

nuclei



0 = rainfall intensity in mm/h
^particle diameter
Source:  HcHahon and Denison (1977).
                                             6-22

-------
The scavenging coefficients  are  dependent on the rainfall rate, the mean raindrop radius, and
the particle size.   When these factors are considered, the scavenging coefficients reported in
Table 6-6 show reasonable consistency, as demonstrated by Figure 6-4.
     Airborne measurements by  Radke  et al.  (1980) on precipitation scavenging of aerosol  par-
ticles greater  than 0.01  urn diameter  in  aged air  masses,  coal-fired power  plant  plumes,  a
kraft paper mill,  and  a plume from a volcanic eruption supported theoretical estimates of wet
removal   for  aerosol particles greater  than  1.0  jjm.   Marked differences were  observed in the
submicrometer particle  region, where measured scavenging efficiencies for submicrometer aero-
sol particles were  typically an  order of magnitude greater than theoretical predictions.   The
scavenging gap,  that portion  of  the aerosol particle size range  where  scavenging collection
efficiencies are at a minimum, was narrower than theoretically predicted,   Radke et al. (1980)
offer some  explanations for  the  discrepancies,  including deliquescent growth and nucleation
scavenging of the  submicrometer  particles in convective clouds.  Considering the varied aero-
sol particle sources and precipitation studied, the measurements showed marked continuity (see
Figure 6-5).
6.4  TRANSPORT AND DIFFUSION
     Pollutant substances emitted into the atmosphere are transported and diffused as a result
of  a  series  of  complex physical  interactions that  result  in  the mean motion  of  air  and its
fluctuating  components.   Transport  and  diffusion  are  associated with  spatial  and temporal
scales.   The spatial or temporal  domain directly  influences what specific physical  phenomena
will most affect transport and diffusion.
     Studies of  air pollution transport and diffusion fall broadly  into  two  categories, de-
pending  on  the extent  of  the horizontal  scale  studied.  High pollutant  concentrations  that
occur in the vicinity  of  a major  emission source  are  dominated  by physical  processes  that
operate  on  a local  horizontal  scale  of the order  of 1 to  5  km, or  approximately  1  hour of
transport.  Since the majority of criteria pollutants are emitted directly into the atmosphere
by  major sources,   this  has  been the primary area  of interest  in  air  pollution regulation.
However,  as  air pollution  issues are  raised  with regard to pollutants of  a  more ubiquitous
nature that  have appreciably longer  lifetimes and,  in  some  cases, form through secondary re-
action processes,  the  horizontal  scale of  interest  expands considerably.  Sulfur dioxide and
particulate  matter span a horizontal  scale  ranging  from local  to global.   A  brief  review of
the physical processes  contributing to transport and diffusion is presented in Section 6.4.1,
while Section 6.4.2 considers residence times of pollutants and their long distance transport.
6.4.1  The Planetary Boundary Layer
     The mean wind within  approximately the first 1000  meters  above the earth's surface car-
ries most  of the pollutants within  the atmosphere.   The mean wind  is determined primarily by
the  interaction  of  three  forces  governed by thermodynamic and  mechanical  processes:  (1) the
force due to the  horizontal  pressure gradient produced by differential solar heating of the
                                             6-23

-------
    100
 X
s
UJ

tr
o
z
5
Ul
u
V)
X
     10
      0.01
                                      I
                                          —I RADKE ET AL. (1977)
                                          1 BUHTSEV ET AL. (1970S
                                          2 HICKS (1976)
                                          3 DANA (1970)
                                          4 PETERSON & CRAWFORD (1970)

                                                     I	
                      0.1              1              10

                     EQUIVALENT PARTICLE DIAMETER, p
     Figure 6-4. Relationship between rain scavenging rates and particle size.

     Source: McMahon and Denison (1979).
                                6-24

-------
tu
K
u
I-
cc
a.
ui
O
o
a-
Ul
Q-
    100
o
1    50
z
tu

1
m   ,
Q
tu
            DATE      SOURCE OF AEROSOL PARTICLES
       -MAY 13, 1974  PT. TOWNSEND PAPER MILL, WA
       -MAR. 25,1976  "NATURAL," NEAR CENTRAL1A, WA
       -MAY 10,1976  CENTBALIA POWER PLANT, WA
                     (a) 400 s AND (b) 265 s SOAVEMGING--*-
                     TIME                 y     x
                                         /'
          (a)
                   tijii
                                               .....I
   10
        ,-2
                 10
                   ,-1
iou
10'
              DRY AEROSOL PARTICLE DIAMETER, j
100
     50
        I   I I I I III!   I  I  1 I I llll|    I  I I I I III]

            PATE      SOURCE OF AEROSOL PARTICLES
        -JUL. 1,1976    "NATURAL" (a) 3 km MSL AND
                      (b) 2,5 km MSL AT MILES CITY, MT
        -APR. 21, 1977   VOLCANIC MAAB, AL
        -JUN. 29,1977   FOUR CORNERS POWER PLANT, NM
                   i ml
                                       i *i 1 1 1 1 ii
       10
         t-2
                 10
                   ,-1
                                10"
              101
              DRY AEROSOL PARTICLE DIAMETER, jum
                              (b)

   Figure 6-5, Percentages of aerosol particles of various sizes
   removed by precipitation scavenging.


   Source:  Radkeetal. (1980).
                           6-25

-------
earth's  surface;  (2) the  Con'olis  force  due  to the  earth's  rotation; and  (3)  the friction
force due to the texture of the earth's surface.  The planetary boundary layer is that portion
of the  atmosphere within  which surface  frietional  effects have a  substantial  impact  on the
mean wind.  Typically, this layer is hundreds of meters deep and varies diurnally.
     Diffusion in the planetary boundary layer, which governs the spreading of pollutants per-
pendicular to  the  transport flow,  is regulated  by  turbulence.   Turbulence,  which comprises a
complex spectrum of fluctuating motion superimposed on the mean wind, is generated through the
interaction  of directional and  speed  differences  (shear)  in  large-scale  atmospheric motions
and  perturbations  introduced  into  the mean  flow by the roughness of the  earth's surface,  as
well as by solar heating.
     The  theory of  the  mean vertical structure of the planetary boundary layer is fairly well
understood (Haugen, 1973) and can be characterized by measuring the basic meteorological para-
meters.  The description of the turbulent properties within the mean motion,  which govern dif-
fusion, is more elusive.  Detailed theoretical approaches to turbulence are difficult to solve
because  they have more  unknown parameters  than equations.  Higher  order closure techniques
apply assumptions that permit new unknowns to be expressed  in terms of others in such a way as
to allow  solution  of the equation set.   However,  practical application is limited because of
their  intensive computer  and  data  requirements.   Consequently,  the practical  treatment  of
atmospheric  diffusion to  air  pollution-related processes  is  based on highly parameterized
theories that depend strongly on basic experimental data sets.
     Practical  approaches  to  the  treatment of  atmospheric diffusion have  been  derived from
statistical theory, similarity theory, and gradient transport or K-theory (Pasquill, 1974).   A
brief description of  these approaches and their  usefulness in  air pollution-related problems
is  presented  below.  More detailed discussions of the  theories can be  found  in  the cited
references.
     Statistical theory considers the time history of the motion of a single  fluid "particle,"
relative  to  a  fixed coordinate axis (Taylor,  1921),  and of groups or clusters of such parti-
cles  relative  to their  centroid (Batchelor,  1953).   The  theory  provides the basis  for the
development  of  the  Gaussian diffusion formula  and  provides an  effective means of correlating
empirical dispersion data.   As a result, diffusion equations for various emission source types
have  been developed  (Gifford,  1968,  1975; Turner, 1970; Pasquill,  1974,  1975).   A practical
limitation  of  this approach is  that  it makes the  fundamental  assumption  of turbulence homo-
geneity,  whereas  boundary-1ayer turbulence  is  inhomogeneous,  especially   in  the  vertical
dimension.
     The  similarity theory of diffusion relates the mean position and other properties of dif-
fusing clouds and plumes to the characteristic parameters of the surface layer, by dimensional
reasoning.   Results  (Monin and  Yaglom,  1971) are reasonably complete  for the surface layer,
but  extension  to  the entire boundary  layer  introduces further parameters, which limits their
practical use.
                                             6-26

-------
     The gradient-transport,  or K-theory,  of  diffusion is the oldest,  originating  with Pick
(1855) and  Boussinesq (1877).   Atmospheric applications  have  been most  successful  at large
scales, including global  diffusion.   At boundary-layer scales the behavior of K is quite com-
plicated.   Useful results  can be obtained  (Pasquill,  1974; Yaglom," 1975;  Csanady, 1973), but
the mathematics  tend to  be  elaborate.   The  essential problem is  to account  for the strong
space-time scale  dependence  of  eddy-diffusivity,  which was  first  demonstrated by Richardson
(1926).  Berlyand (1974)  based a comprehensive system of air pollution analysis entirely on a
form of K-theory.
     Obukhov (1941)  showed that the parameterization of atmospheric  diffusion  follows a form
of Richardson's law  of  diffusion,  where the total  amount  of  turbulent energy dissipation and
the pollutant spreading  is proportional to the diffusion  time  to the 3/2 power.  Data on the
instantaneous values  of  the  spreading of plumes and puffs (i.e.,  on relative diffusion) shows
that this law describes  diffusion  up to t  on  the order of an  hour (Gifford,  1976a).   On the
other hand,  the time-average  spreading of plumes was shown by Taylor (1921) to obey the asymp-
totic  laws  a «  t  (where a =  the  standard  deviation of  the  horizontal  spreading  of the
                                                           1/2
pollutant cloud and  t = time) for small t values and a 
-------
mean wind;  and (3)  the time  of  day and  height at which  the pollutant  is  emitted  into the
atmosphere.
     Residence  times  for pollutants  are governed  by  the extent  of wet  and  dry removal and
chemical transformation  the  pollutant species undergoes  in  the  atmosphere.   Figure 6-6 esti-
mates  residence times for  typical  pollutants and  their  associated characteristic horizontal
meteorological  scale.   Average windspeeds  of 5  m/sec were  assumed  in  approximating  the dis-
tance scale.  Residence time estimates are based on the work of Junge (1972, 1974).
     Transport scales for pollutants such as S0?, which have appreciable dry deposition veloc-
ities, are sensitive to the relative height at which the pollutant is emitted.  Pollutant dis-
tributions are  also  sensitive  to the stability of the atmosphere, which governs the extent of
vertical mixing to the surfaces.
     Industrial facilities  emitting  large  quantities  of  S0~ have tried to  take  advantage of
these  natural  meteorological  phenomena  to reduce  ambient  levels of  SO,  in  the vicinity of
their  stacks.   IJy  building  taller effluent stacks, emitting facilities injected SCL at higher
levels  in  the  atmosphere,  allowing the pollutant  more time to  be  dispersed and transported
before reaching ground level,  thereby effectively reducing ground-level concentrations of S0?.
A  great  deal of  controversy  arose  over whether this approach circumvents  the  intent of the
Clean  Air  Act.   As  a  result of the  Clean  Air Act  Amendments of  1977, EPA  requires  that the
degree of  emission limitation  necessary for  control  of any air  pollutant cannot be achieved
through  the construction of stacks  higher  than would be  considered  appropriate  using good
engineering  practice  design standards.  An  interesting corollary of the  tall  stack  issue is
the potential  for  such sources to enhance the production of particulate sulfate.   When S0? is
emitted at  higher  levels in the atmosphere,  the probability of its removal by dry deposition
is lowered, thus extending its  lifetime in the atmosphere and subsequently enhancing the prob-
ability of its being transformed to particulate sulfate through chemical reaction.
     Of particular importance  to horizontal transport is the strength and time of formation of
the nocturnal  inversion,  a  stable layer of  air  formed at the surface due to the differential
cooling of  the  earth's surface relative to the night air.   This stable layer varies in thick-
ness from approximately 50 m to 500 m depending on meteorological conditions.  At the onset of
the  nocturnal  inversion,   all pollutants  present  in  the  well-mixed layer  from the  day's
emission  are cut  off  from  the  surface by  this  stable  layer of  air:   There  are  no  major
mechanisms  for transport  through  this  layer,  so  dry  deposition  processes  virtually  stop,
leaving  the pollutant reservoir  aloft  free to travel  long  distances  with negligible losses.
In addition,  horizontal  mean windspeeds are  typically  higher  in layers aloft, due to the re-
duction  in  frictional  drag  at  the earth's  surface  as a result of  the  presence  o^ Jfhe stable
nocturnal layer.
     Associated with  this  overall  process  is  a  phenomenon  of particular  importance to night-
time transport, the  nocturnal  jet.   According to Blackadar  (1957),  the nocturnal jet forms as
a  result of  decoupling  of  winds previously  restrained by  frictional  forces  at  the surface.
These  winds  are  now free  to  accelerate  in  response to  existing pressure gradients.   As  a

                                             6-28

-------
RESIDENCE
TIME.hr
103
10^ *+
101 -i
10° -I
10'1 *t

m*3 ^

HORIZONTAL
LENGTH
SCALE
10 000 km '"—
2 000 km 	
200 km -- i
20km




CLIMATOLOGICAL
SCALE
CH4







SYNOPTIC AND
PLANETARY
SCALE

0,1— 1.0 jum
PARTICLES





MESO
SCALE


so2
r




MICRO-SCALE



io2
PARTICLES



Figure 6-6.  Estimated residence times for select pollutant species and their associated hori-
zontal transport scale.
                                           6-29

-------
result, some  overshooting  in windspeed occurs as the flow attempts to establish a new balance
with  inertia! forces.   Bonner  (1968)  examined  2  years  of  upper-level  wind data  from  the
National Weather Service's rawinsonde network to determine the frequency and geographical dis-
tribution of  the  low-level  jet.   Recently, high-resolution measurements of wind profiles col-
lected  over  central  Illinois  (Sisterson  and Frenzen, 1978)  showed that  nocturnal,  low-level
wind maxima occur more frequently than indicated in Bonner's analysis.  In these studies, made
during  the  summers of  1975  and  1976,  low-level wind maxima were observed  on 24  out  of 30
nights  for which  meteorological  field experiments were  conducted.  Typical  average  windspeed
profiles observed  under  the  decoupled conditions showed windspeeds of the order of  1 to  2 m
sec    near the  surface,  increasing to maximum  values  of 8 m sec   at 100 to 200 meters  above
the surface.
     In summary, it appears that the nocturnal jet and nighttime flows are significant factors
in the  transport of pollutants over long distances.
     Definitive studies on the long-range transport of atmospheric tracers have been primarily
associated with radioactive debris (Islitzer and Slade, 1968) and in many instances at heights
not of  particular interest for air pollution-related work.  Some analyses of ambient data have
been performed to  provide  qualitative indications of  the  long-range  transport and dispersion
of  certain  pollutant species  (Altshuller,  1976; Lyons  and Husar, 1976;  Rodhe  et al.,  1972;
Brosset and Akerstrom,  1972);  but very few quantitative studies exist, primarily because of a
lack of appropriate experimental data.   Recent monitoring and field studies of long-range air
pollutant transport phenomena should alleviate this problem somewhat (Perhac, 1978; MacCracken,
1978; Schiermeier et  al., 1979).
6.5  AIR QUALITY SIMULATION MODELING
     The air  quality  simulation model (AQSM) primarily describes the quantitative relationship
between the distribution  of  emissions and ambient air quality in time and space.  Air quality
simulation models are intended to improve understanding of the physical and chemical  processes
operating  in  polluted atmospheres  and provide  a basis for  sound, credibly based scientific
decisions on  the  nature and extent of emission control required to meet specified ambient air
quality standards.
     The AQSM was first developed in the early 1930's, when Sutton (1932) introduced his basic
theory  on  diffusion  in  the  atmosphere.  Sutton's  theory established  the  foundation for the
Gaussian equations  used in  describing  the dispersion of  effluents in the atmosphere.   Since
then, the evolution of AQSM's has continued, treating increasingly complex air pollution prob-
lems and using  advanced theoretical approaches to describe the details of physical and chemi-
cal  atmospheric  processes.   The use of  mathematical  models  for  air quality impact analysis
associated with SO,.and total  suspended particulate matter,  both criteria  pollutants,  had
become  a  standard practice,  though discretionary,  prior to the passage of  the  Clean Air Act
Amendments in 1977.
     With  the passage of  the 1977  amendments,  EPA was  required to  take  certain regulatory
steps  related to  the use of air quality  simulation models.   The workhorse of operational air

                                             6-30

-------
quality  simulation  modeling has  been the  single  and multiple  source  Gaussian  plume models.
These models have  been used primarily to predict ground-level concentrations ranging from the
immediate vicinity  to several  kilometers  downwind of  the effluent  source.   Many reviews on
dispersion  modeling  are available  (e.g.,  Gifford,  1968;  Strom,  1976;  and Turner,  1979).
Section 6.5.1 provides  a brief discussion of the  status  of the Gaussian modeling techniques,
while section 6.5.2  discusses  the scientific basis and current status of air quality simula-
tion modeling over  long  distances, and the extent to which available modeling techniques have
furthered our understanding  of the physical and chemical  processes  affecting the fate of S0?
and particulate matter in the air environment.  A discussion of model  evaluation and data bases
is provided in Section 6.5.3.
6.5.1  Gaussian Plume Modeling Techniques
     The Gaussian diffusion  formulation  is  used in a variety of air quality simulation model-
ing approaches.  The formulation is a result of the  Gaussian or normal distribution function
being  a  fundamental  solution  to  the  Fickian  diffusion equation.    Strictly  speaking,  the
Gaussian distribution  applies  only in the  limit of  large diffusion  time and for homogeneous,
stationary conditions.
     The Gaussian  diffusion  formulation  for a continuous  point source emitting  pollutants at
height h and calculated receptor concentrations at ground level is given by
               x (x,y) =
                          noy
-------
     The Gaussian plume  model  formulas, however, have not been without criticism.   Aside from
fundamental disadvantages,  such  as  inability to treat spatially varying meteorological param-
eters, discontinuity  under calm  wind conditions, and  inability to  treat  nonlinear reactive
pollutants, the major  criticisms  relate to the improper application of the models through the
use of  inadequate or  inappropriate input data.   Turner (1979) pointed out that  the  ease of
using the  Pasquill-Gifford  dispersion  estimates  rather  than  collecting  recommended onsite
meteorological data  for developing  inputs to  the  model has  led  to applications  beyond the
scope intended for the dispersion schemes.   Turner (1979) indicates that a significant improve-
ment  in dispersion  modeling estimates would Be achieved through the collection and use of on-
site  measurements,  including:   hourly averaged windspeed  and direction at a  height of  10 m;
standard deviation  of horizontal wind fluctuations; bulk  Richardson number  (a  quantitative
measure of  atmospheric stability) as  determined  from temperature  and 2 and  10  m temperature
differences; height of the top of the boundary layer under unstable conditions; and the top of
the  inversion under  stable atmospheric conditions.   Additional  improvements  will   grow as
existing  theories and  experimental  data  bases are  drawn together  in  a unified  scheme for
estimating  dispersion parameters  as  a  function of  stability,  effluent release  height, and
surface roughness.
     A variety  of operational Gaussian  air  quality  dispersion models, used for  most  S0? and
total suspended particulate matter regulatory applications, are available through EPA's User's
Network for Applied Modeling of Air Pollution (UNAMAP).   Turner (1979) briefly describes these
models.
     A final  aspect of  dispersion  modeling  calculations  is the prediction  of  the effective
height of  the effluent release,  the  so-called  "plume rise," which  strongly  affects  the pre-
dicted ground-level concentration of pollutants.   Research into the processes affecting plume
rise  and its  prediction  has been underway for  the past 20 years.   Briggs (1975) reviewed the
physics of plume  rise and its prediction and presented basic formulations for calculating the
height to which  plumes rise as a function of atmospheric stability and several standard stack
parameters.   He   indicated  that  further investigation  is  needed  in the  area of  plume rise
limited by ambient turbulence under convective atmospheric conditions.
6.5.2  Long-RangeAir Pollution Modeling
     The recent growing  interest in long-range transport  of  air pollutants has resulted from
extensive  reviews on the  subject by  Bass (1980),  Eliassen  (1980), Pack  et  al.  (1978), and
Smith and  Hunt  (1978).   (Long range  is  defined in this document as  horizontal  scales of the
order of 1000 km resulting from transport times of the order of several days.)  Typical model
spatial resolutions on  this scale range from 20  to  100 km.  Bass  (1980)  concluded that most
long-range  transport  models are  Lagrangian  based.   In  the  Lagrangian  approach,  an emitting
source element is represented by a series of discrete pollutant parcels that are advected and
diffused by  a time- and  space-dependent wind field.  In principle,  calculation  on the  indi-
vidual pollutant parcels can treat time-dependent chemical transformation, dry deposition, and
precipitation scavenging  processes.   Fixed space-time averages of pollutants are generated by

                                             6-32

-------
superimposing all  elements that pass a  specified  point over the averaging  time  of interest.
The Lagrangian models, although all based on the same theory, have evolved individual nuances.
     Little consensus of  opinion  exists  for standard treatments of:   (1) wind field analysis;
(2) choice of wind height level for trajectory; (3) mixing height variations; and (4) dry and
wet removal and  chemical  transformation  rates.  Even the basic generation of discrete air par-
cels  is  viewed  from  four different approaches:  puff,. superposition,  segmented  plume,  square
puff, and statistical.  In Figure 6-7,  three of the approaches are contrasted to the idealized
continuous plume they are attempting to represent (Bass, 1980).
     Eulerian or grid-based  approaches  are less prevalent in  long-range transport air pollu-
tion modeling.   This may result from problems associated with the numerical integration of the
advection equation that give rise to pseudodiffusion effects.  A more likely reason is the in-
creased complexity and enhanced data base and computation requirements of the Eulerian models.
     Table 6-7 provides  a representative sampling of long-range transport models discussed in
the available  literature and for  each referenced  model,  presents a brief  description  of the
modeling approach, including characteristic averaging times;  approaches to dry and wet removal
and chemical  transformation; and pollutant species modeled.
     A review of  the  models presented in  Table  6-7 indicates that S0?  and  sulfates have re-
ceived the,most  attention.   This is because many of the models were developed specifically to
study the  acid  rain  phenomenon.  These sulfur species have been identified as major contribu-
tors  to  the  acidification  of  precipitation.   Though  none  of the  models  consider primary
emitted particulate matter, its inclusion would be reasonably straightforward given the avail-
ability  of appropriate  emission  inventories.   Considering  the gas-to-particle  forming pro-
cesses of  nitrates and  organic species,  aerosol dynamic processes and size distributions need
further  research  to  be  understood.  Until these basic  processes,  as well as gas-liquid phase
transfer  and  solution phase chemistry of  rain droplets, are  treated  adequately  within the
models,  significant skepticism  toward  the scientific credability and usefulness of the mo'dels
will remain.
     Similarly,  regional visibility impairment, which results from the physical interaction of
sunlight with light-absorbing gases and particles, and light-scattering aerosols, requires the
consideration of many of  the  processes  described above.   Though  several  empirical analysis
techniques  have been developed  that  provide  a qualitative  understanding  of the  scope and
general meteorological characteristics of the visibility  impairment problem, no adequate quan-
titative relationships are available for emission control strategy assessments.
     Although verification studies of  long-range transport  models  are  limited,  it has been
recognized for  some  time  that errors in  observed wind direction (Pack  et  al.  1978) and the
specification  of  wind  fields  in  general  (Sykes  and  Hatton,  1976;  Smith  and  Hunt,  1978;
Draxler,  1979)  can  result  in  drastic  errors  in  spatial predictions over  long-range travel
distances.
                                             6-33

-------
      CONTINUOUS PLUME MODEL
  SEGMENTED PLUME MODEL
PUFFSUPERIMPOSmON MODEL
'SQUARE PUFF MODEL
                Rgure 6-7. Trajectory modeling approaches are shown.
                Source: Bass (1980).

-------
                                 TABLE 6-7.  SUMMARY OF SELECT LONG RANGE TRANSPORT AIR POLLUTION MODELS
Model
ARL-ATAD1
EURMAP-1,-22
ENAMAP-1
MESOPUFF3
MESOPLUME4
ASTRAP5
l^IRSOX
MESOGRIO6
Type
Segmented plume
trajectory
Puff supenmposition
trajectory
Puff super-imposition
trajectory
Segmented plume
trajectory
Statistical
trajectory
Puff superimposition
trajectory/vertical
finite difference
Squence puff grid
Description
averaging time
daily to yearly
daily to yearly
daily to yearly
daily to yearly
monthly to yearly
daily to yearly
daily to monthly
Removal process
First order wet
and dry removal
First order wet
and dry removal
First order dry
removal
First order dry
removal
Diurnal and sea-
sonal dry removal/
first order wet
removal
First order wet
and dry removal
First order wet
and dry removal
Chemical process
None
SO- first order decay
SO- first order decay
SO. first order decay
Diurnal and seasonal
dependent SO. first
order decay
SO, first order decay
Typically SO, first
order decay
Pollutant species
Inert substances
S02 and SO*"
SOj and SO^
SO. and SO*"
S0? and S0*~
2-
SO^ and SO^
SOZ and SO*"
Reference
Heffter et al. (1975)
Heffter (1980)
Johnson et al. (1978)
Mancuso et al (1979)
Bbumralkar (1980)
Benkley and Bass (1979a)
Benkley and Bass (1979b)
Shannon (1979)
Shieh (1977)
Meyers et al. (1979)
Morris et al (1979)
see also model  by Start and Wendell  (1974)
see also models by Eliassen  (1978),  Nordo (1976)
and Eliassen and Saltnones (1975)

see also models by Draxler (1977,  1979)
see also models by McNaughton (1980), Hales et al  (1977), Pendergast (1979),
and Henmi (1980)

see also models by Bolin and Persson (1975); Fisher (1975, 1978); and
and Scriven and Fisher (1975)

see also models by Lui and Durran (1977), Rao et al  (1976), Lavery et al. (1980);
  and Cantnchael and Peters (1979)

-------
     The  sparse temporal  and  spatial  resolution  of upper-level  wind information  leads  to
uncertainties  in predicting  transport residues.   The  National  Weather Service's  rawinsonde
network  provides vertical  profiles  for  windspeed and  direction,  temperature,  and  moisture
every 12 hours at 70  sites  across the continental United  States.  This provides upper-level
winds at a horizontal resolution  of  the  order of  400 kilometers,  considerably less  resolved
than the  20  to 80 kilometer grid  spacing  required in air quality simulation modeling techni-
ques.  Increased temporal  resolution  in upper-level winds should also diminish transport pre-
diction uncertainties.
6.5.3  Model Evaluation and DataBases
     Almost no long-range transport models have been verified through comparison of model pre-
dictions with  actual  observation.   The major deterrent has been insufficient or inappropriate
monitoring  data for  the  spatial  scales  of  interest.   Long-range transport  models predict
ambient  pollutant  concentrations  that  represent horizontal spatial averages of  the  order  of
10  square kilometers.  Standard monitoring networks, established for local high concentration
measurement within  the vicinity of the emission sources, do  not provide  representative data
for long-range  models.   Some routine  monitoring data  for S02  and S0f~ from EPA's Storage and
Retrieval  of  Aerometric Data  (SAROAD)  system have been  useful  in  model testing (Bhumralkar,
1980} but in  general, the data  have  proven  less than adequate.  The  Electric  Power  Research
Institute  (EPRI)-sponsored Sulfate Regional  Experiment   (SURE)  air quality network,  (Perhac,
1978} which operated from August 1977 to October 1978, has provided the most extensive S00 and
  2~                                                                                     "•
SQ,  data base to date for  long-range  transport model evaluation.   But even these data col-
lected over  this limited  period are  only  sufficient in  providing spatially  resolved S00 and
  2-                                                                                     <;
SO.  regional  concentration fields during the  intensive study  periods—August 1977, October
1977, mid-January  to  mid-February 1978,  April  1978,  July 1978,  and October  1978—when the
extended  54-site  monitoring network  was activated.  During the SURE  study period, data also
exist from the  Department  of Energy-funded Multi-State Atmospheric Power Production Pollution
Study (MAP3S)  precipitation  chemistry network,  (MacCracken,  1978) which  had  at  least four
sites operating during  the program.   No dry  deposition  data are  available  for the  study
period.    Since  the  data have only recently  become available,  they have had  limited  use, .but
future long-range transport model  evaluations are certain to consider them.
     Another limitation  in model  evaluation  studies is the quality of the  emission inventory.
Until recently,  there was  no  national  gridded emission  inventory.  Clark  (1980)  prepared  an
annual  gridded emissions inventory for the United States  and Southern Canada east of the Rocky
Mountains, using data compiled by EPA, the  Ontario Ministry  of the Environment, and Environ-
ment Canada.   In preparing the  gridded inventory,  Clark found  significant errors  in many  of
the United States point source records, which  had  to be corrected.  Models, like chains, are
only as strong as their weakest links.
     Although  long-range  transport  model  evaluations  are  extremely limited,  two  recent
studies should be noted.  Mancuso  et al. (1979) evaluated a trajectory puff model using monthly
                                             6-36

-------
averages from the Organisation for Economic Cooperation and Development (OECD) monitoring pro-
gram.  They  generated  two sets of evaluation results.  The first considered model predictions
versus observations  using parameters  originally specified  for  the model.   This resulted in
root mean square (RMS) differences of predicted versus observed monthly averaged concentration
                                        •*      3               ?—
at all receptors  of  12.9 (5 ppb) and 4.8 pg/m  for S0?* anS' SO. ", respectively.  In the second
evaluation, half of the data were used to optimize model  parameters through a regression anal-
ysis technique  and the other half of the data were used to evaluate the model.  This resulted
                                               o               ^_
in RMS differences of 7.7 (3 ppb) and 2.9 ug/m  for S09 and SO.  and correlation coefficients
                            2-                         £4
of 0,72  both  for S0« and SO,  ,  a  marked  improvement in the model's performance.  None of the
optimized  parameters  assumed values  that were physically unrealistic.   Lavery  et al.  (1980)
evaluated  a  grid  model using data from the SURE monitoring network and based on 24-hour aver-
aged concentrations.   Four days  were selected for parameter  "adjustments"  and  "fine tuning"
                                                                                            2-
and  an  additional three  days were  used  for the model evaluation.   RMS  differences for SO,
                              3                                                   3
ranged from  6.9 to  23.4 ug/m   for  the  3  days,  with a  mean value of 14.1 ug/m .   RMS dif-
                2~                              3                              3
ferences for  SO.   ranged from 5.2 to 14.4 ug/m  with a mean value of 9.3 ug/m .   Mean concen-
                                                       32-
tration  for the 3 days were 26.1 (10 ppb) and 13.9 ug/m  for SO, and SO,  , respectively.  Mean
                                          2-
correlation  coefficients for  S0? and SO,   for the  three  days  tested  were 0.31  and 0.53,
respectively.  Based on results of the 3-day comparison between observations and model predic-
tions, Lavery  et  al.  (1980) concluded that  the  model typically overpredicts observed 24-hour
average  sulfate  by  20 to 30 percent  and  observed 24-hour average S02  by a factor of 2 to 3.
     Overall,  the results from both models  are  encouraging.   Evaluation  of a trajectory puff
model  for the  United States  (Bhumralkar,  1980) using  the  SURE  data  base,  is  also showing
promising  results.
6.5.4  Atmospheric Budgets
     Atmospheric budgets  have proven a convenient technique for quantitatively evaluating the
overall  source  and  sink contributions of specified pollutant species within a selected region
of  interest.   The budget is formed by estimating the various input and output processes asso-
ciated with  the region, such as  anthropogenic  and  natural emissions, pollutant concentration
inflow and outflow,   and wet and  dry removal.   Budget  analyses provide  a general  long-term
indication of  the significant factors contributing to the pollutant burden  in a given region.
Sulfur budgets  have  been of  interest  both in Europe  (Rodhe, 1972, 1978; Garland, 1978) and in
North  America (Galloway  and  Whelpdale,  1980)  because of  the  association of sulfur with acid
precipitation  phenomena.   Conclusions drawn from the  eastern  North American sulfur  budget by
Galloway and Whelpdale (1980) were that  manmade  emissions exceed natural ones by a  factor of
10;  wet  and  dry  deposition over  the region  is  approximately equivalent;  and  at least one-
quarter  of the  emissions leave the  region  via the atmosphere to  the  east.   As with Western
Europe,  the   North  American  budget  showed  that  human  activities  dominate   the  regional
atmospheric  sulfur cycle.
                                             6-37

-------
6.6  SUMMARY
     The processes governing transport and diffusion, chemical transformation, and wet and dry
removal of SO,  and PM in the  atmosphere  are extremely complex and are  not completely under-
stood.  The oxidation  rates  of S0? observed  in  industrial  plumes and urban atmospheres range
from 0 to  15  percent/h and would  seem  to be only partially accounted for through homogeneous
gas-phase reactions.  Liquid-phase catalytic oxidation reactions involving Mn2  and carbon, as
well as solution-phase  oxidation by H202 are possible  contributors  to the observed oxidation
rates, but  further research  is  required to determine the  rates and detailed  mechanisms of
these processes under typical atmospheric conditions.
     Dry deposition  of S0«  is fairly  well  understood as a  result of extensive measurements
over  various  surfaces.   Particle  dry  deposition  studies have focused  more on  the  physical
aspects of  the deposition process, and have generated very  few  supporting data on particles
with compositions typical of those found in the polluted atmosphere.
     Our understanding  of the wet removal of S0? has progressed considerably in recent years,
including  consideration of solution-phase chemistry within  rain  droplets.   Particle removal,
like gas removal,  depends on the  physical  characteristics  of the precipitation events, which
may, in many instances, be the determining factors in accurate wet removal  prediction.
     Characterization of  the dynamics of the planetary boundary  layer is essential to an ade-
quate understanding of pollutant transport and diffusion over all spatial scales.  Though con-
siderable advances have been made  in this area, our ability to predict mean transport and dif-
fusion over long  distances is inadequate.   No doubt this is partly due to the sparse spatial
and temporal resolution of the data from the upper air wind observation network used to gener-
ate models of the transport winds.
     Present generation  long-range air pollutant transport models use simple parameterization
for chemical transformation  and  wet and dry removal, and varying degrees of sophistication in
the treatment of transport and diffusion.  None of the models adequately treat the dynamics of
the  planetary   boundary  layer.   Evaluations of  long-range  transport models,  though limited
because of lack of  data bases,  have shown that with  further  research  and development these
models should  be adequate tools  in  addressing air pollution issues  associated  with the mean
movement of pollutants over long distances.
                                             6-38

-------
6.7  REFERENCES


Acres Consulting Services Ltd. and Applied earth Science Consultants Inc.  Atmospheric loading
     of  the upper  Great Lakes,  vol.  2, Report  for Canada  Centre for  Inland Waters,  1974.

Alkezweeny, A. J.,  and D. C. Powell.   Estimation  of transformation rate  of S0? and SO. from
     atmospheric concentration data.  Atmos. Environ. 11:179-182, 1977.

Altshuller, A. P.  Regional  transport and transformation  of sulfur dioxide  to  sulfates  in the
     United States.  J. Air Pollut. Control Assoc. 26:318-324, 1976.

Bakulin, V. N.,  E.  E.  Senko, B.  G.  Starikov,  and V. A. Trufakin.  Investigation of turbulent
     exchange and washout by measurement of natural radioactivity in surface air.  J. Geophys.
     Res. 75:3669-3674, 1970.

Banerji,   P.,   and  D.   D.   Chatterjee.   Radon  content   of  rainwater.   Nature  (London)
     204:1185-1186, 1964.

Barrie,  L.  A.   An  improved  model  of  reversible,  S0~~washout  by  rain.   Atmos.  Environ.
     12:407-412,  1978.                                   *

Bass, A.  Modeling long range transport and diffusion.  In:  Second Conference  on Meteorology,
     American  Meteorological  Society  and  Air  Pollution  Control  Association,  Second  Joint
     Conference on  Applications  of Air Pollution Meteorology,  New  Orleans,  Louisianna,  March
     24-28, 1980.  American Meteorological Society, Boston, MA, 1980.  pp. 193-215.

Batchelor,  G.  K.   The  Theory  of  Homogeneous  Turbulence.   Cambridge  University   Press,
     Cambridge, England 1953.

Beilke,  S.   Laboratory  investigations  on  washout  of   trace  gases.   In:   Precipitation
     Scavenging (1970), Proceedings of a Symposium.  Battelle Pacific Northwest Laboratory and
     U.S. Atomic Energy Commission, Richland, Washington, June 1-5, 1970.  R. J. Engelmann and
     W.  G.  N.  Slinn,  eds.,  AEC  Symp. Ser.  22,  U.S.  Atomic Energy  Commission,  Oak Ridge, TN,
     December 1970.  pp. 261-269.

Beilke,  S., D. Lamb, and J.  Muller.  On the uncatalyzed oxidation of atmospheric SOp by  oxygen
     in  aqueous system.  Atmos.  Environ. 9:1083-1090, 1975.

Belot,  Y.,  and   D.  Gauthier.   Transport  of  micronic  particles  from  atmosphere  to   foliar
     surfaces.   In:   Heat and  Mass Transfer  in  the Biosphere.  Part  I.  Transfer Process in
     Plant  Environment.   D.  A.  deVries and N.  H.  Afgan,  eds., John Wiley and  Sons, Inc., New
     York, NY, 1975.  pp. 583-591.

Benkley, C. W., and A.  Bass.  Development of mesoscale air  quality  simulation models.  Vol. 3.
     User's guide  to MESOPUFF (mesoscale puff) model, September 1979.  EPA-60Q/7-80-056, U.S.
     Environmental Protection Agency, Research Triangle Park, NCS 1979a.

Benkley,  C. W.,  and A. Bass.  Development of mesoscale air quality simulation  models. Vol. 2.
     User's   guide  to   MESOPLUME   (mesoscale  plume   segment)  model,  September,   1979.
     EPA-600/7-80-057,  U.S.   Environmental   Protection  Agency,  Research  Triangle  Park,  NC,
     1979b.

Berlyand,  M.  E.   Modern Problems in  Atmospheric  Diffusion  and Atmospheric  Pollution.   Hydro-
     meteorological Publishing House, Leningrad, USSR, 1974.

Bhumralkar, C.  M., R.  L. Mancuso,  D.  E. Wolf,  R.  A.  Thuillier,  and  K. C.   Nitz.  ENAMAP-1
     Long-Term  Air Pollution Model:   Adaptation and  Application  to Eastern  North America.
     EPA-6QO/4-80-039,  U.S.  Environmental  Protection Agency, Research Triangle Park, NC, July
     1980.


                                             6-39

-------
Blackadar,  A.   K.    Boundary  layer  wind  maxima  and their  significance  for  the  growth of
     nocturnal inversions.  Bull. Am. Meteorol. Soc.  38:283-290, 1957.

Bolin, B.,  and  C.  Persson.  Regional dispersion and  deposition of atmospheric pollutants with
     particular application to sulfur pollution over  western Europe.  Tellus 27:281-310, 1975.

Bonner, W. D.  Climatology of the low leva! jet.  Mon. Weather Rev. 96:833-850, 1968.

Boussinesq, J.   Essai  sur la theorie des  eaux courantes.  [Essay on the theory of circulating
     currents.]  Mem. Acad. Sci.  Inst.  Fr. 23:1-61,  1877.
                                               90
Bowen, V.  T.,  and  T.  T,  Sugihara.   Strontium   in  the "mixed-layer"  of the Atlantic Ocean.
     Nature (London) 186:71-72, 1960.

Briggs,  G.  A.   Plume  rise  predictions.   In:  Lectures  on  Air Pollution and Environmental
     Impact  Analysis.    D.  Haugen,   ed.,  American  Meteorological  Society, Boston,  MA, 1975.

Brimblecombe, P.,  and D.  J.  Spedding.   The catalytic oxidation  of micromolar aqueous sulfur
     dioxide.  Atmos. Environ 8:937-945, 1974.
                  o  o
Brosset,   C.,  and A. Akerstrom.   Long distance transport of air  pollutants - measurements of
     black airborne  particulate  matter  (soot) and particle-borne sulphur in Sweden during the
     period of September-December 1969.   Atmos. Environ. 6:661-672, 1972.
                                                                                   32
Burtsev,   I.  I.,  L.  V.  Burtseva, and S.  G. Malakhou.  Washout characteristics of a   P aerosol
     injected  into   a  cloud.  In:   Atmospheric Scavenging of  Radio-isotopes, Papers  from a
     conference, Lithuanian  SSR  Academy of Sciences, Palanga,  Lithuania SSR,  June 7-9, 1966.
     B,  Styra,  C.   Garbaliauskas  and  V.  Lujanas,  eds.,  Israel  Program  for  Scientific
     Translations, Ltd.; Jerusalem,  Israel, 1970.  pp. 242-250.

Cantrell, B.  K., and K.  T.  Whitby.   Aerosol   size distributions  and  aerosol  volume formation
     for a coal-fired power plant plume.  Atmos. Environ. 12:323-334, 1978.

Carmichael, G.  R.,  and L. K.  Peters,   Numerical  simulation of the  regional  transport of SO*
     and  sulfate  in  the  eastern  United  States.   In:   Preprints - Fourth  Symposium  on
     Turbulence, Diffusion and Air  Pollution, American  Meteorological  Society,  Reno, Nevada,
     January 15-18,  1979.  American  Meteorological   Society, Boston, MA,  1978,   pp. 337-344.

Cawse, P. A.  A  survey of atmospheric  trace  elements in the U.K. (1972-73).  Environmental &
     Medical Science Division.  AERE Harwell,  U.K. Report R 7669, 1974.

Chamberlain, A.  C.   Aspects  of travel  and  deposition of aerosol  and vapour clouds. AERE HP/R
     1261, 1953.

Chamberlain, A.  C.   Aspects of  the  deposition of radioactive and  other gases and particles.
     Int. J. Air Pollut. 3:63-88, 1960.

Chamberlain,  A.   C.   Transport  of  Lycopodium  spores  and other  small  particles  to rough
     surfaces.   Proc. R. Soc. London Ser. A 296:45-70, 1967.

Chamberlain, A.  C.   Deposition  of  S02  by gaseous diffusion.   Unpublished manuscript, 1973.

Chamberlain, A.  C.   Movement of particles  in plant  communities.  j_n:   Vegetation  in the
     Atmosphere.   J. L. Monteith, ed., Academic Press Inc., New York, NY, 1975.   pp. 155-201.
                                             6-40

-------
Chamberlain, A.  C.   Dry  deposition of  sulfur dioxide.   In:   Atmospheric Sulfur Deposition,
     Environmental Impact  and  Health Effects, Proceedings of the 2nd Life Sciences Symposium,
     U.S. Department  of Energy,  U.S.  Environmental  Protection Agency,  and  Tennessee Valley
     Authority, Gatlinburg,  Tennessee,  October  14-18, 1979.   D. S.  Shriner,  C.  R.  Richmond,
     and S.  E.  Lindberg,  eds., Ann  Arbor Science Publishers Inc.,  Ann Arbor,  MI, 1980.   pp.
     185-197.

Chamberlain, A, C.,  and R. C, Chpdwick.,  Deposition of  airborne radio-iodine vapor.  Nucleonics
     8:22-25, 1953.

Chamberlain, A. C.,  and R. C.  Chadwick.   Deposition  of spores and  other particles on vegeta-
     tion and soil.   Ann.  Appl. Biol. 72:141-158, 1972.

Chang, T.  Y.  Estimate of the conversion rate of  S0« to SOA  from  the Da Vinei  Flight Data.
     Atmos.  Environ.  13:1663-1664, 1979.                ,     4

Clark, T.   L.   Annual   anthropogenic pollutant  emissions in  the United  States  and southern
     Canada east of the Rocky Mountains.   Atmos. Environ. 14:961-970, 1980.

Clough, W.  S.  Transport of particles to a surface.  J. Aerosol Sci. 4:227-234, 1973.

Cobourn,  W.  G.,  R.  B.  Husar, and J.  D.  Husar.   .Continuous  in  situ  monitoring of ambient
     particulate  sulfur using  flame photometry and thermal  analysis.   Atmos.  Environ. 12:89-
     98,  1978.

Craig, D.  K. ,  B. L.  Klepper,  and  R.  L.  Buschbom.   Deposition of various piutoniurn-compound
     aerosols onto plant  foliage  at very  low wfnd velocities.  Atmosphere-Surface Exchange of
     Particulate  and Gaseous Pollutants  (1974), Proceedings of a  symposium,  Battelle Pacific
     Northwest Laboratories and U.S. Atomic Energy Commission,  Richland, Washington, September
     4-6, 1974.   R.  J.  Engelmann and  G,  A.  Sehmel, coord., ERDA  symposium series 38, Energy
     Reserach  and Development Administration, Oak Ridge,  TN, January 1976.    pp.  244-263.

Csanady,  G.  T.   Turbulent Diffusion in the Environment.  D. Reidel  Publishing Co., Dordrecht,
     Holland, 1973.

Dana,  M.  T.   Scavenging  of soluble  dye particles by rain.   In:   Precipitation Scavenging,
     (1970), Proceedings of a Symposium, Battelle Pacific Northwest  Laboratory and U.S. Atomic
     Energy  Commission,  Richland, Washington, June 1-5,  1970.   R.   J.  Engelmann  and  W.  G: N.
     Slinn,  eds., AEC  symposium series 22,  U.S.  Atomic  Energy  Commission, Oak Ridge,  TN,
     December 1970.   pp. 137-147.

Dana, M.  T.   S09 versus sulfate wet deposition in the eastern United States.  J. Geophys.  Res.
     85:4475-4480, 1980.

Dana,  M. T. , and J.  M. Hales.   Statistical  aspects of the  washout  of polydisperse aerosols.
     Atmos.  Environ.  10:45-50,  1976.

Dana, M.  T., J. M. Hales,  and M.  A,  Wolf,  Rain scavenging of S09 and sulfate from power plant
     plumes.  J.  Geophys.  Res.  80:4119-4129, 1975.              £

Davidson,  C.   I.,  and  S.  K.   Friedlander.   A filteration model  for aerosol  dry deposition:
     Application  to  trace metal  deposition from the  atmosphere.  J.  Geophys.  Res.   83:2343,
     1978.

Dennis,  R. ,  C.  E. Billings, J. A.  Record, P.  Warneck, P. Arin, and M.  L. Arin.  Measurements
     of  sulfur dioxide losses from stack plumes.   Presented at the 62nd Annual  Meeting, Air
     Pollution Control Association, New York, NY, June  22-26, 1969.
                                             6-41

-------
Dittenhoefer, A. C., and R. G. de Pena.  A study of production and growth of sulfate particles
     in plumes from a coal-fired power plant,  Atmos. Environ. 12:297-306, 1978.

Drake, R.  L., and S.  M.  Barrager.   Mathematical Models  for Atmospheric Pollutants. EA-1131,
     Electric Power Research Institute, Palo Alto, CA, August 1979.

Oraxler,  R. R.  Determination of atmospheric diffusion parameters.  Atmos. Environ.  10:99-106,
     1976.

Oraxler,  R. R.  _A Mesoscale Transport and Diffusion Model.   NOAA technical memorandum ERL-ARL-
     64,  U.S.  Department of  Commerce,  National  Oceanic  and Atmospheric  Administration, Air
     Resources Laboratories, Silver Spring, MD, June 1977.

Draxler,  R.  R.   Modeling the results of  two recent mesoscale dispersion experiments.  Atmos.
     Environ. 13:1523-1533, 1979.

Eliassen, A.  The  OECD study of long range transport of  air pollutants:  Long  range transport
     modeling.  Atmos. Environ. 12:479-488, 1978.

Eliassen, A.  A review of long-range transport modeling.  J. Appl. Meteorol. 19:231-240,  1980.

Eliassen,  A.,  and J.  Saltbones.   Decay  and transformation rates for  S0?,  as estimated from
     emission data,  trajectories  and measured air concentrations.  Atmos. Environ.  9:425-429,
     1975.

Engelmann,   R.  J.  The  Calculation  of  Precipitation Scavenging.  BNWL-77,  Battelle Memorial
     Institute, Pacific Northwest Laboratory, Richland, WA,  July 1965.

Engelmann,   R.  J.  The  calculation  of precipitation scavenging.   In;   Meteorology and Atomic
     Energy - 1968.   D.  H.  Slade,  ed. , U.S. Atomic Energy Commission, Division of Technical
     Information, Oak Ridge, TN, July 1968.  pp. 208-221.

Eriksson,  E.   The yearly  circulation  of chloride and  sulfur  in nature; meteorological, geo-
     chemical and pedological implications.  Part I.  Tell us 11:375-403, 1959.

Esmen, H.  A.,  and  M.  Corn.   Residence  time  of  particles in  urban  air.   Atmos. Environ.
     5:571-578, 1971.

Esmen, N.  A.   Particle  retention  efficiency  of  scavenging  rainfall.   In:   163rd National
     Meeting, American  Chemical  Society, Division of Water, Air and Waste Chemistry, Boston,
     Massachusetts,  April  9-14,  1972.   Prepr,  Pap. Natl. Meet.  Div. Water Air  Waste Chem. Am.
     Chem.  Soc. 12: 12-18, 1972.

Pick, A,   Uber diffusion [On diffusion.]  Ann. Phys. Chem.  2:59-86, 1855.

Fisher,  B.  E. A.  The  long  range transport of  sulfur  dioxide.   Atmos. Environ. 9:1063-1070,
     1975.

Fisher,  B.  E.  A.  The calculation of  long term sulfur deposition in Europe.   Atmos. Environ.
     12:489-501, 1978.

Forrest, J. ,  and L.  Newman.   Further studies on the oxidation of  sulfur dioxide  in  coal-fired
     power plant plumes.  Atmos. Environ.  11:465-474, 1977.

Forrest, J.,  and L.  Newman.  Oxidation  of  sulfur dioxide  in the  Sudbury smelter plume.  Atmos.
     Environ. 11:517-520, 19775.

Galloway, J.  N., and  D. M. Whelpdale.   An  atmospheric sulfur budget for eastern North America.
     Atmos.  Environ.  14:409-417, 1980.


                                             6-42

-------
Garland,  J.  A.   Dry and wet  removal  of sulfur from  the  atmosphere.   Atmos. Environ.  12:349-
     362, 1978.

Garte11,  F,  E. ,  F.  W.  Thomas, and  S.  B.  Carpenter.  Atmos.  oxidation  of S0« in coal-burning
     power plant plumes.  Am. Ind. Hyg. J. 24:113-120, 1963.

Gatz,  D.  F.   A  review  of  chemical  tracer  experiments  on precipitation  systems.   Atmos.
     Environ. 11:945-954, 1977.

Georgii,  H.   W.   On  the  effect of  rainfall  on the  sulphur  dioxide  concentration  in the
     atmosphere.   Air Water Pollut.  7:1057-1059,  1963.

Gifford,  F.  A.  An  outline of  theories  of  diffusion in the  lower  layers of the  atmosphere.
     In:   Meteorology  and  Atomic  Energy—1968.    0.  H.  Slade,  ed.,   U.S.   Atomic  Energy
     Commission,  Division  of Technical  Information,  Oak  Ridge, TN, July 1968.  pp.  65-116.

Gifford,  F.  A.   Atmospheric dispersion models  for  environmental pollution applications.   In:
     Lectures  on  Air Pollution  and Environmental Impact Analysis.   D.  Haugen,  ed., American
     Meteorological Society, Boston, MA, 1975.

Gifford,  F.  A.  Tropospheric  Relative Diffusion Observations.   Unpublished manuscript,  File
     No.  76/10, National Oceanic and Atmospheric  Administration, Oak  Ridge,  TN,  1976a,

Gifford,  F.  A.  Turbulent  diffusion  typing schemes:   a  review.   Nucl.  Saf. 17:68-86, 1976b.

Gillani,  N.  V.,  R.  Husar,  D.  E.  Patterson,  and W.  E. Wilson.   Project MISTT:   kinetics of
     particulate sulfur formation in a power plant out to 300  km.  Atmos.  Environ.  13:589-598,
     1978.

Colder,  D.   Relations   among stability  parameters  in  the  surface layer.   Boundary  Layer
     Meteorol. 3:47-58, 1972.

Graedel,  T.  E. ,  and J.  P.  Franey.   Field measurements of  submicron  aerosol washout by  snow.
     Geophys. Res.  Lett. 2:325-328, 1975.

Graedel,  T.  E. ,  and J.  P.  Franey.   Field measurements of  submicron  aerosol washout by  rain.
     In:   Precipitation Scavenging (1974),  Proceedings of a Symposium,  Illinois State  Water
     Survey  and  U.S.  Atomic Energy Commission, Champaign,  Illinois,  October 14-18,  1974.  R.
     G.  Semonin  and  R. W.  Beadle,  coord. ,  ERDA  symposium  series  41,  Energy Research and
     Development Administration, Oak Ridge, TN, June  1979.   pp.  503-523.

Gravenhorst,  G. , T.  H.  Janssen-Schmidt, D. H.  Ehalt, and E.  P.  Roth.  The  influence of clouds
     and  rain on the  vertical   distributions  of  sulfur  dioxide in  a one-dimensional  steady-
     state model.  Atmos. Environ. 12:691-698,  1978.

Grover,  S.  N. ,  H.   R.  Pruppacher,  and A.  E. Hamielec.    A  numerical   determination  of the
     efficiency with  which  spherical  aerosol  particles collide  with  spherical water drops due
     to  inertial  impaction  and phoretic and  electrical  forces.   J.  Atmos.  Sci. 34:1655-1663,
     1977.

Hales,  J. M.  Fundamentals of the  theory of  gas scavenging by  rain.  Atmos. Environ. 6:635-
     659,  1972.

Hales,  J. M.  Wet removal  of sulfur compounds from  the  atmosphere.   Atmos. Environ.  12:389-
     399,  1978.

Hales,  J.  M. , D.  C.  Powell, and T. D, Fox.   STRAM -  An air pollution model  incorporating non-
     linear  chemistry,  variable  trajectories,  and plume segment  diffusion.   EPA-450/ 3-77-012,
     U.S.  Environmental Protection Agency, Research Triangle Park, NC, 1977.


                                             6-43

-------
Hales, J. H., J. M. Thorp, and M. A. Wolf.  Field Investigation of Sulfur Dioxide Washout from
     the  Plume  of a  Large Coal-fired Power Plant  by  Natural  Precipitation.   APTD-0739, U.S.
     Environmental Protection Agency, Research Triangle Park, NC, March 1971.

Hart, G.  E,,  and D.  R. Parent.  Chemistry of throughfall under Douglas Fir and Rocky Mountain
     Juniper.  Am. Midi. Nat. 92:191-201, 1974.

Haugen,  D.   A,   ed.   Workshop  on  micrometeorology,  Boston,  1972.    American Meteorological
     Society, Boston, MA, 1973.

Heffter,  J.  L.   Air  Resources  Laboratories  Atmospheric  Transport  and  Dispersion  Model
     (ARL-ATAD).   NOAA  technical   memo   ERL-.ARL-81,   U.S.  Department  of  Commerce,  National
     Oceanic  and Atmospheric Administration,  Air  Resources  Laboratories,  Silver Spring, MD,
     1980.

Heffter,  J.   L.,  A.   D.  Taylor, and  G.  J.  Ferber.  A Regional-Continental  Scale Transport,
     Diffusion  and Deposition  Model.   NOAA  technical  memo ERL-ARL-50,  U.S.  Department  of
     Commerce,  National  Oceanic and  Atmospheric Administration,  Air Resources Laboratories,
     Silver Spring, MD, June 1975.

Hegg, D,  A. ,  and P.  V. Hobbs.   Measurements  of gas-to-particle conversion in the plumes from
     five coal-fired electric power plants.  Atmos.  Environ. 14:99-116, 1980.

Henmi,  T.   Long-range transport model of SO-  and  sulfate and  its  application to the eastern
     United States.  J. Geophys. Res. 85:443^-4442,  1980.

Hicks,  B.  B.   An evaluation of  precipitation  scavenging  rates  of  background  aerosol.  In:
     Argonne  National  Laboratory  Radiological   and  Environmental  Research  Division  Annual
     Report:  Atmospheric Physics:  January through December, 1976.  ANL-76-88 Part IV,  Energy
     Research and  Development  Administration,'Argonne National Laboratory, Argonne, IL, 1976.
     pp. 65-75.

Hicks,  B. B.,  M. L.  Wesely, and J. L. Durham.   Critique of Methods to Measure Dry Deposition.
     EPA-60Q/9-8Q-05Q,  U.S.  Environmental Protection  Agency,  Environmental Sciences Research
     Laboratory, Research Triangle Park, NC, September 1980.

Hill,  F.  B., and  R. ,F.  Adamowicz.   A  model  for  rain composition and  the washout of  sulfur
     dioxide.  Atmos. Environ.  11:917-927, 1977.

Hobbs,  P.  V.,  D.  A.  Hegg,  M.  W.   Eltgroth,  and L. F. Radke.  Evolution  of  particles  in the
     plumes  of  coal-fired  power  plants.   I.  Deductions  from  field  measurements.    Atmos.
     Environ. 13: 935-951, 1979.

Huntzicker, J.  J.,  R.  S.  Hoffman, and C.  S.  Ling.  Continuous measurements  and speciation of
     sulfur-containing aerosols by flame photometry.   Atmos. Environ. 12:83-88, 1978.

Husar,  R. B., D. E. Patterson, J. D. Husar, N. V. Gillani, and W. E, Wilson.   Sulfur budget of
     a power plant plume.  Atmos. Environ. 12:549-568, 1978.

Islitzer, N.  F., and D. H. Slade.  'Diffusion and transport experiments.  In:  Meteorology and
     Atomic  Energy-1968.    D.   H.   Slade,  ed.,  U.S.  Atomic Energy  Commission,  Division  of
     Technical  Information, Oak Ridge, TN, July  1968.  pp. 117-188.

Johnson, W.  B., D.  E. Wolf, and R. L. Mancuso.   Long  term regional patterns  and transfrontier
     exchanges  of airborne  sulfur  pollution  in  Europe.   Atmos. Environ.  12:511-527, 1978.

Junge, C. E.  The cycle of atmospheric trace gases  - natural and manmade.   Q.  J. R.  Meteorol.
     Soc. 98:711-729, 1972.
                                             6-44

-------
Junge, C.  E.   Residence time and variability of tropospheric trace gases.  Tellus  26:477-488,
     1974.                                                                          ~~

Kalkstein, M.  I.,  et al.  Natural  aerosols  and nuclear debris  studies.  AFCRL-TN-59-627, Air
     Force Cambridge Research Laboratory, Hanscom Air Force Base, MA,  1959.

Klepper,  B.,  and D.  K.  Craig.   Deposition  of ainbonne pa-rticulates onto  plant  leaves.   J,
     Environ. Qua!. 4:495-499, 1975.

Knutson,  E.  0. ,  and J.  D. Stockham.   Aerosol  scavenging by snow:  comparison  of single-flake
     and  entire-snowfall  results.   In:   Precipitation Scavenging  (1974),  Proceedings of  a
     Symposium,  Illinois  State  Water  Survey  and  U.S.  Atomic  Energy Commission,   Champaign,
     Illinois, October  14-18,  1974.  R. G.  Semonin and R.  W.  Beadle, Coords.,  ERDA Symposium
     Series 41,  Energy  Research and Development Administration, Oak Ridge, TN,  June 1977. pp.
     195-207.

Lavery,  T.  F. ,  R.  L.  Baskett,  J.  W,   Thrasher,  N. J.  Lordi,  A.  C.  Lloyd, and  G. M.  Hidy.
     Development and  validation of a regional model to  simulate atmospheric  concentrations  of
     sulfur dioxide  and sulfate.   In;   Second  Joint Conference  on Applications of Air Pollu-
     tion  Meteorology,  and Second  Conference on  Meteorology,  American Meteorological  Society
     and  Air  Pollution  Control Association,  New Orleans,  Louisiana,  March  24-28,  1980.
     American Meteorological Society, Boston, MA, 1980.  pp. 236-247.

Little, P.  Deposition of  2.75,  5.0 and 8.5  urn particles to plant and  soil  surfaces.  Environ.
     Pollut.  12:293-305, 1977.

Little,  P.,  and  R.  D. Wiffen.   Emission and deposition  of petrol engine  exhaust Pb—I  Deposi-
     tion of  exhaust Pb to plant and soil surfaces.  Atmos. Environ. 11:437-447, 1977.

Liu, M. K. and D. R. Durran.  The Development of a  Regional Air  Pollution Model  and its Appli-
     cation  to  the  Northern Great Plains.   EPA-908/1-77-001,  U.S.  Environmental   Protection
     Agency,  Denver, CO, July 1977.

Lusis, M.  A., and  H.  A.  Wiebe.   The  rate  of  oxidation of sulfur dioxide  in  the  plume of a
     nickel smelter stack.  Atmos.  Environ.  10:793-798,  1976.

Lusis, M.  A., K.  G.  Anlauf,  L. A. Barrie, and H.  A.  Wiebe.   Plume chemistry studies at a
     northern Alberta power plant.  Atmos. Environ.  12:2429-2437, 1978.

Lyons, W.  A., and R. B. Husar.  SMS/GOES visible images detect  a synoptic-scale air pollution
     episode.  Mon. Weather Rev. 104:1623-1626, 1976.

MacCracken,  M.  C.   MAP3S:   an investigation of atmospheric energy related pollutants  in the
     northeastern United States.  Atmos. Environ. 12:649-659, 1978.

Makhon'ko,  K. P.   Determination of capture coefficient of  radioactive dust  by   rain.   Izv.
     Akad. Nauk  SSSR Ser.  Geofiz. 11:1709, 1964.

Makhon'ko, K.  P.   Simplified theoretical motion  of contaminant removal  by precipitation from
     the  atmosphere.  Tell us 19:467-476, 1967.

Makhon'ko, K. P., and G. V. Dmitrieva.  Capability  of various types of precipitation for wash-
     out  of  fission products from  the  atmosphere  and characteristics of washout.   Izv.  Akad.
     Nauk SSSR Ser. Phys.  Atm.  I. Okeana 3:508, 1966.

Mamane,  Y. ,  and R.  F.   Pueschel.   Formation of  sulfate particles in the plume of the  Four
     Corners  Power Plant.  J. Appl. Meteorol. 19:779-790, 1980.
                                              6-45

-------
Mancuso,  R.  L., C.  M.  Bhumralkar,  D.  E. Wolf,  and W.  B.  Johnson.   The  exchange of sulfur
     pollution  between  the  various countries  of  Europe based  on  the  EURMAP  model.    In:
     Preprints-Fourth  Symposium  on  Atmospheric  Turbulence,  Diffusion  and  Air  Pollution,
     American  Meteorological  Society,  Reno, Nevada,  January 15-18,  1979.  American Meteoro-
     logical Society, Boston, MA, 1978.  pp. 345-354.

McMahon, T. A., and P. J. Oenison.   Review Paper - empirical  atmospheric deposition  parameters
     - a survey.  Atmos. Environ. 13:571-585, 1979.

McNaughton, D.  J.   Initial  comparision of SURE/MAP3S sulfur  oxide observations with long-term
     regional model predictions.  Atmos. Environ. 14:55-63, 1980.

Meagher, J. F., L. Stockburger, E. M. Bailey, and 0. Huff.  The oxidation of  sulfur  dioxide to
     sulfate aerosols in the plume of a coal-fired power  plant.  Atmos. Environ. 12:2197-2204,
     1978.

Meyers, R.  E., R,  T. Cederwall,  and W.  D.  Ohmstede.  Modeling regional atmospheric transport
     and  diffusion:   Some environmental applications.   In:  Advances in Environmental  Science
     and  Engineering.   J.  R.  Pfafflin and  E.  Ziegler,  eds., Gordon  and Breach, New York,  NY,
     1979.

HBller,  U.,  and  G.  Schumann.   Mechanisms  of  transport  from  the atmosphere  to  the  earth's
     surface.  J. Geophys. Res. 75:3013-3019, 1970.

Honin, A.  S., and A. M. Yaglom.   Statistical  Fluid Mechanics, vol.  1.  MIT  Press,  Cambridge,
     MA, 1971.

Neuberger,  H.,  C.  L. Hosier, and W.  C.  Kocmond.  Vegetation as an  aerosols  filter.   In:  Bio-
     meteorology.  Volume 2.  S.W.  Tromp and W.H.  Weihe,  eds.,  Pergamon Press, Ltd.,  Oxford,
     England,  1967.  pp. 393-702.

Newman,  L.   Atmospheric oxidation  of sulfur  dioxide.    In:   Atmospheric Sulfur Deposition,
     Environmental  Impact  and Health Effects, Proceedings of the  2nd Life Sciences  Symposium,
     U.S.  Department of Energy,  U.S. Environmental  Protection Agency,  and  Tennessee Valley
     Authority, Gatlinburg,  Tennessee,  October 14-18, 1979.  D.S.  Shriner, C.R. Richmond,  and
     S.E.  Lindberg,  eds., Ann  Arbor Science Publishers  Inc., Ann Arbor,  MI, 1980.   pp.  131-
     143.

Newman,  L.,  J. Forrest, and B. Manowitz.   The application of  an  isotope  ratio technique  to a
     study of the atmospheric oxidation of  sulfur dioxide in  the plume from a coal-fired power
     plant.   Atmos.  Environ. 9:969-977, 1975.

NordiJ, J.   Long-range transport of  air pollutants  in  Europe  and acid precipitation  in  Norway.
     Water, Air, Soil Pollut. 6:199-217,  1976.

Obukhov,  A.  M.  On the  distribution of energy in the spectrum of turbulent  flow.   Izv. Akad.
     Nauk  SSSR Ser.  Geogr. Geofiz. 5:453, 1941.

Pack,  D.  H«,  G. J.  Ferber,  J.  L. Heffter,  K. Telegades, J.  K.  Angell, W. H.  Hoecker, and L.
     Machta.   Meteorology of  long-term transport.   Atmos.  Environ. 12:425-444,  1978.

Pasquill,  F.   The estimation of  the dispersion  of windborne material.  Meteorol. Mag. 90:33-
     49,  1961.

Pasquill,  F.   Atmospheric Diffusion.  2nd  ed.   John Wiley and  Sons,  Inc.  New York,  NY.  1974.

Pasquill,  F.   The dispersion of  materials  in the atmospheric  boundary layer - the basis  for
     generalization.   In;   Lectures on Air  Pollution  and Environmental Impact Analysis.   D.
     Haugen,  ed.,  American Meteorological Society,  Boston MA, 1975.


                                             6-46

-------
Pasquill, F.  Atmospheric  Dispersion Parameters in Gaussian Plume Modeling.   Part 2:  Possible
     Requirements for Change  in the Turner Workbook Values.  EPA-600/4-76-030b, U.S.  Environ-
     mental   Protection  Agency, Environmental Sciences  Research Laboratory, Research  Triangle
     Park, NC, June 1976.

Pendergast,   M.  M.   Model  evaluation for  travel  distances 30-140  km.  In:   Preprints-Fourth
     Symposium  on  Turbulence,  Diffusion, and Air  Pollution,  American Meteorological  Society,
     Reno,  Nevada,  January 15-18,  1979.   American Meteorological  Society, Boston,  MA, 1978.
     pp. 648-651.

Perhac, R. M.  Sulfate regional experiment in northeastern United States:   The "SURE"  program.
     Atmos.  Environ. 12:641-647, 1978.

Perkins,  R.   W.,  C.  W.  Thomas,  and  J.  A.  Young.  Application of short-lived cosmogenic  radio-
     nuclides as traces  of  in-cloud scavenging  processes.   J.  Geophys.  Res,  75:3076-3087,
     1970.                                                                       ~

Peterson, K.  R., and  T.  Y.  Crawford.   Precipitation  scavenging  in  a  large-cloud diffusion
     code.   In:   Precipitation Scavenging  (1970),  Proceedings  of  a  Symposium,  Richland,
     Washington, June  1-5, 1970.   R.  J. Engelmann and W.  G.  N. Slinn,  eds., Symp.  Ser.  22,
     Battelle Pacific Northwest Laboratory and U.S. 'Atomic Energy  Commission, Oak  Ridge, TN,
     December 1970.   Available as  CONF-700601  from  National  Technical  Information  Service,
     Springfield, VA.   pp.  425-431.

Pierson,  D.   H.,  P.  A.  Cawse, L. Salmon, and R. S. Cambray.  Trace elements in the atmospheric
     environment.  Nature,  (London) 241:252-256, 1973.

Postma,  A.   K.   Effect  of  solubilities  of  gases  on  their   scavenging   by  raindrops.  In:
     Precipitation  Scavenging  (1970),  Proceedings of a Symposium,  Battelle Pacific Northwest
     Laboratory and U.S.  Atomic Energy Commision, Richland, Washington, June 1-5, 1970.  R. J.
     Engelmann  and W.  G.  N,  Slinn,  eds.    AEC  symposium  series   22,   U.S.  Atomic  Energy
     Commission, Oak Ridge, TN, December, 1970.  pp. 247-259.

Pueschel  R.   V.,  and C.  C.  Van Valin.  Cloud nucleus formation in a power plant plume.   Atmos.
     Environ. 12:307-312 1978.

Radke,  L. F., E.  E. Hindman, II, and P. V. Hobbs.  A case study of plume scavenging by  a rain
     shower.  In:  Precipitation Scavenging (1974), Proceedings of a Symposium, Illinois State
     Water Survey and U.S.  Atomic Energy Commission, Champaign, Illinois, October 14-18, 1974.
     R. G.  Semonin  and R.  W.  Beadle,  coord..,  ERDA symposium  series  41,  Energy Research and
     Development Administration, Oak Ridge, TN, June 1977.  pp.  425-436.

Radke,  L. F., P.  V. Hobbs, and M.  W. Eltgroth.  Scavenging of aerosol particles by precipita-
     tion.  J. Appl. Meteorol. 19:715-722, 1980.

Rao, K. S.,  J.  S.  Lague, and  B. A.  Egan.   An  air  trajectory  model  for regional transport of
     atmospheric  sulfates.    Preprint   Volume-Third   Symposium  on   Atmospheric  Turbulence,
     Diffusion  and Air  Quality, American  Meteorological  Society,  Raleigh,  North Carolina,
     October  19-22,  1976.   American  Meteorological  Society, Boston,  MA,  1976.   pp.  325-331.

Rasmussen,  K.  H.,  M.  Taheri,  and  R.  L. Kabel.   Global emissions  and  natural  processes for
     removal of gaseous pollutants.   Water Air Soil Pollut. 4:33-64, 1975.

Richardson,   L.  F.   Atmospheric diffusion shown on  a  distance-neighbour graph.  Proc. R. Soc.
     London Ser. A 104:640-654, 1926.

Roberts,  P.  T., and  S.  K. Friedlander.   Conversion  of S02 to sulfur particulate  in the Los
     Angeles atmosphere.   EHP Environ. Health Perspect.  10:103-108, 1975.
                                             6-47

-------
Roberts, D. B. ,  and D.  J. Williams.   The  kinetics of oxidation of sulphur dioxide within the
     plume from  a  sulphide  smelter in a remote  region.   Atmos.  Environ.  13:1485-1499, 1979.

Robinson, E.,  and  R.  C.  Robbins.  Gaseous nitrogen compound pollutants from urban and natural
     sources.   'J. Air Pollut. Control Assoc. £0:303-306, 1970.

Rodhe,  H.   Budgets  and  turn-over times  of atmospheric  sulfur compounds,   Atmos.  Environ.
     12:671-680, 1978.

Rodhe, H., and J.  Grandell.   On the  removal  time of aerosol particles from the atmosphere by
     precipitation scavenging.  Tellus 24:442, 1972.

Schiermeier,   F.  A.,  W.   E.  Wilson,  F.  Pooler,  J.  K.  S. Ching,  and J.   F.  Clarke.   Sulfur
     Transport and  Transformation  in the Environment (STATE):   A  Major EPA Research Program.
     Bull.   Am. Meteorol. Soc. 60:1303-1312, 1979,

Scott,  B.  C.    Parameterization  of  sulfate removal  by  precipitation.   J.   Appl.  Meteorol.
     17:1375-1389, 1978.

Scriven, R. A.,  and B.  E. A.  Fisher.   The long  range transport of airborne  material and its
     removal  by deposition and washout.  Atmos. Environ. 9:49-68, 1975.

Sehmel, G. A.   Particle eddy diffusivities  and  deposition velocities for  isothermal flow and
     smooth surfaces.  J. Aerosol Sci. 4:125-138,  1973.

Sehmel, G. A., and S. L. Sutter.   Particle deposition rates on a water surface as a function
     of particle diameter and air velocity.  J. Rech. Atmos. 8:911-920, 1974.

Sehmel, G.  A.   Particle and gas dry deposition:   a review.  Atmos.  Environ. 14:983-1011, 1980.

Shannon,  J.   D.   The   Advanced   Statistical   Trajectory   Regional   Air  Pollution  Model.
     ANL/RER-79-1,  U.S.  Department  of  Energy Argonne  National  Laboratory, Argonne,  IL,  May
     1979.

Sheih,  C.  M.   Application  for  a  statistical  trajectory model to the simulation  of sulfur
     pollution over northeastern United States.   Atmos. Environ. 11:173-178, 1977.

Sheih, C. M.,  M. L. Wesely,  and B.  B.  Hicks.  Estimated  dry  deposition  velocities of sulfur
     over the  eastern United States and  surrounding regions.  Atmos.  Environ.  13:1361-1368,
     1979.

Shirvaikar, V,  V.  et al.  A  study of washout of  radioactive fallout and particulate matter in
     individual  rain showers.  AEET/AM/10 (India),  1960.

Sisterson, D.  L.,  and  P.  Trenzen.   Nocturnal  boundary-layer wind maxima and the problem of
     wind power  assessment.   Environ. Sci. Techno!. 12:218-221, 1978.

Slinn, W. G.  N.  Dry deposition and  resuspension  of aerosol particles  —a  new  look at some old
     problems.   In:  Atmosphere-Surface Exchange  of Particulate and Gaseous Pollutants  (1974),
     Proceedings  of  a  Symposium,   Battelle Pacific  Northwest  Laboratories  and  U.S.  Atomic
     Energy Commission,  Richland,  Washington, September 4-6, 1974.   R. J.  Engelmann and G. A.
     Sehmel,  coord., ERDA symposium  series  38, Energy Research and Development Administration,
     Oak Ridge,  TN, January 1976.  pp. 1-40.

Slinn, W. G,  N.   Some approximations  for  the wet  and dry removal of  particles and gases from
     the atmosphere.  Water,  Air, Soil Pollut. 7:513-543,  1977.

Slinn, W. G.  N.   Precipitation scavenging.   In:   Atmospheric Sciences and Power Production—
     1981.  D. Randerson, ed. ,  U.S. Department  of Energy, Washington,  DC.   In press, 1981.


                                             6-48

-------
Slinn, W. G.  N.,  L.  Hasse, B. B. Hicks, A. W. Hogan, D. Lai, P. S.  Liss, K. 0. Munnich, G. A,
     Sehmel, and 0.  Vittorl.   Some aspects of  the transfer of atmospheric  trace  constituents
     past the air-sea interface.  Atmos. Environ.  12:2055-2087, 1978.

Small, S.  H.   Wet  and  dry deposition of fallout materials at  Kjeller.   Tellus  12:308-314,
     1960.                                                                          ~

Smith, F. B.   A scheme for estimating the vertical <5-ispei%ion of  a plume from a  source  near
     ground level.   In:   Proceedings of the Third  Meeting of the Expert  Panel  on Air Pollution
     Modelling, NATO  Committee  on  the Challenges of Modern Society,  Paris, France, October
     2-3, 1972.  NATO-CCMS N.  14, North Atlantic Treaty Organization, Brussels, Belgium, 1972.
     pp.  XVII-1 - XVII-14.

Smith, F. B. ,  and  R.  D. Hunt.   Meteorological aspects of the transport  of pollution over  long
     distances.  Atmos.  Environ. 12:461-477, 1978.

Start,  G.  E. , and  L.   L.  Wendell.   Regional   Effluent  Dispersion  Calculations   Considering
     Spatial  and  Temporal Meteorological  Variations.  NOAA  technical  memo  ERL  ARL-44,  U.S.
     Department of Commerce, National  Oceanic and  Atmospheric Administration,  Idaho Falls,  ID,
     May 1974.

Stephens, W.  T.,  and  R.  0.  McCaldin.  Attenuation of power station plumes as determined, by^
     instrumental aircraft.   Environ. Sci. Techno!.  5:615-621  1971.                '          "

Strom, G. H.  Transport and diffusion  of stack effluents.   In:  Air Pollution.  Volume I:   Air
     Pollutants, Their Transformation and Transport.   A.  C.  Stern, ed. ,  3rd ed.,  Academic
     Press, Inc., New York, NY,  1976.  pp. 401-501.

Sutton,  D.  G.  A  theory of eddy diffusion  in  the atmosphere.   Proc.  R.  Soc. London.  Ser.  A
     135:143, 1932.

Sykes, R.  I.,  and  L.  Hatton.   Computation  of  horizontal  trajectories based on   the surface
     geostrophic wind.  Atmos. Environ. 10:925-934, 1976.

Taylor,  G.  I.   Diffusion by continuous movements.  Proc.  London Math.  Soc. 20:196-212, 1921.

Turner,  D.  B.   Workbook of Atmospheric Dispersion Estimates.   Office of Air Programs -Publica-
     tion No.  AP-26,  U.  S. Environmental  Protection  Agency, Research Triangle Park, NC, 1970.

Turner,  D.  B.   Atmospheric dispersion modeling  -  a critical review.  J. Air.  Pollut. Control
     Assoc. 29:502-519, 1979.

Ursenbach,  W.  0.,  A.  C.  Hill,  W.  H.  Edwards,  and  S.  M.  Kunen.   Conversion rate of S02 to
     submicron  sulfate in the plumes of a  coal-fired  power  plant in the  western United StaT.es.
     Presented  at  70th Annual  Meeting, Air  Pollution Control  Association, Toronto, Canada,
     June 20-24, 1977.

Volchuk,  H.  L. , V.  T.  Bowen,  T.  R.  Folsom, W.  S. Broecker,  E. A.  Schuert,  and  G.  S. Bien.
     Oceanic distribution of radionuclides from  nuclear explosions.   In:   Radioactivity  in the
     Marine Environment.  National Academy of Sciences, Washington, DC,  1971.

Wang,  P. K.,  S.  N.  Grover  and H.  R. Pruppacher.  On the  effect  of electric charges on the
     scavenging  of  aerosol  particles  by   clouds  and  small  raindrops.    J.   Atmos.   Sci.
     35:1735-1743, 1978.

Weber, E.   Contribution to the  residence  time  of  sulfur dioxide  in a  polluted atmosphere. J.
     Geophys.  Res. 75:2909-2915, 1970.
                                             6-49

-------
Weber, A.  H.   Atmospheric dispersion parameters of Gaussian Plume Modelling.  Part I:  Review
     of  Current  Systems  and  Possible Future  Developments.   EPA-600/4-76-030a,  U.S.  Environ-
     mental Protection Agency, Research Triangle Park, NC, 1976.

Wedding, J. B.,  R.  W. Carlson, J.  J.  Stukel,  and F. A.  Bazzaz.   Aerosol deposition on plant
     leaves.   In:   Proceedings  of the First International Symposium on Acid Precipitation and
     the  Forest  Ecosystem, Ohio  State University,  Columbus,  Ohio,  May  12-15,  1975.   L.  S.
     Dochinger  and T.  A. Seliga,  eds.,  NEFES/77-1, U.S.  Department of Agriculture,  Forest
     Service,  Northeastern Forest  Experiment  Station,  Upper  Darby,  PA,   1976.   pp.  897-903.

White, E.,  and'F.  Turner.  A method  of  estimating income of  nutrients in a catch of airborne
     particles by a woodland canopy.  J. Appl. Ecol. 7:441-461, 1970.

Wolf, M.  A.,   and  M.  T.  Dana.   Experimental  studies on  precipitation scavenging.  Battelle-
     Northwest Annual  Report.   Report BNWL-1051 (Pt. 1), U.S. Atomic Energy Commission, 1969.
     pp. 18-25.

Yaglora, A. M.   Equations with time-dependent coefficients describing diffusion in a stationary
     atmospheric  surface layer.   Izv.  Acad.  Sci.  USSR  Atmos. Oceanic  Phys.  (Engl.  Transl.)
     11:704-708, 1975.
                                             6-50

-------
                                     7.  ACIDIC DEPOSITION

7.1  INTRODUCTION
     Acidic  precipitation has become  a major  environmental  concern  in  many regions  of the
world.  Acidic  precipitation (rain  and snow)  in  the Adirondack Mountains of  New York State,
eastern  Precambrian Shield area  of Canada,  southern Norway, and  southwest Sweden  has  been
associated with  the acidification of  waters  in ponds, lakes, and  streams and the subsequent
disappearance  of  animal   and  plant  life.  Acidic  precipitation  also  is  believed to  have the
potential for leaching elements from sensitive soils and causing direct and indirect injury to
forests.   It is also believed to play a role in damaging stone monuments and buildings, corrod-
ing metals and the deterioration of paint.
     The  story  of acidic  precipitation  is  ever-changing;  new  information concerning the
phenomenon is appearing continuously.   This chapter explains how particulate matter and sulfur
                                                    *
oxides are  involved in  acidic deposition phenomena  and associated  ecological  effects.   The
information  as  presented reflects  the understanding of the scientific community at  the  time
this chapter was written.  A critical assessment document on acidic decomposition currently
being written  under the direction of  the  Office  of Research and Development  of the  Environ-
mental  Protection Agency, will  present a more detailed, up-to-date discussion of  the  many
facets of the acidic deposition problem.
     The  sections that  follow emphasize the  effects associated with the wet  deposition  of
sulfur and  nitrogen oxides  and their products on aquatic  and  terrestrial  ecosystems.   Dry
deposition also  plays  an  important role,  but  the relative  contribution of  this  process  is
still unknown.  Because  sulfur and nitrogen oxides are  so  closely linked in the formation of
acidic precipitation,  no  attempt  has  been made to limit the discussion that follows to sulfur
oxides.
7.1.1  Overview of the Problem
     The generally  held  hypothesis  is that sulfur and nitrogen  compounds are largely respon-
sible  for  the acidity of precipitation.   The emissions of the  sulfur and  nitrogen compounds
involved in acidification are attributed chiefly to the combustion of fossil  fuels.  Emissions
may occur at ground level,  as from automobile exhausts, or from stacks 300 meters (1000 feet)
or more  in  height.   Emissions from natural sources  are  also  involved; however, in highly in-
dustrialized areas, emissions  from  manmade sources far exceed those from natural sources.  In
the  eastern  United States, the  highest emissions  of sulfur oxides  are from  electric power
generators burning  coal,  while on the West Coast, particularly around large cities, emissions
of nitrogen oxides, chiefly from automotive sources, predominate.  (See Chapter 4.)
     The fate  of sulfur  and  nitrogen  oxides,  as  well as other pollutants  and  gases emitted
from natural sources  into the atmosphere,  depends on their dispersion, transport, transforma-
tion and deposition.   Sulfur  and nitrogen oxides may be deposited locally or transported long
distances  from the emission  sources.   Therefore,  residence  time  in the atmosphere  will  be

                                            7-1

-------
brief  if the emissions  are deposited  locally  or may  extend to days or weeks  if long range
transport  occurs.   The  chemical  form  in  which emissions  ultimately  reach the  receptor is
determined  by  the complex  chemical transformations  that  take place  between the  emission
sources  and the receptor.   Long range transport over distances of  hundreds  or thousands of
miles  allows  time  for a greater  number of  chemical  transformations to occur (see Chapter 6).
     Sulfates and  nitrates are  among the products of the  chemical  transformations  of sulfur
and  nitrogen  oxides.   Ozone and  other  photochemical  oxidants are believed  to  be  involved in
the  chemical  processes  that  transform  sulfur  dioxide  and  nitrogen  oxides  in  the atmosphere
into  sulfuric and nitric  acids.   When these  acids are  brought to  earth  in rain  and snow,
acidic  precipitation   occurs.   Because  of  long  range  transport,  acidic precipitation  in  a
particular  state or  region can be the  result  of emissions from sources  in  states or regions
many miles  away,  rather than from local sources.  To date, however, the complex nature of the
chemical  transformation processes has  made it  difficult to  demonstrate  a direct  cause and
effect  relationship  between emissions of sulfur and  nitrogen oxides and the  acidity of pre-
cipitation.
     Acidic precipitation  is arbitrarily defined as precipitation  with a  pH  less  than 5.6.
This value  has  been  selected because precipitation formed in an atmosphere relatively free of
natural  or manmade emissions  would  have a pH  of  approximately 5.6 due to  the combining of
carbon dioxide with water  in the air to form carbonic acid.
     Acidity  of  aqueous solutions  is  determined  by  the concentration  of  hydrogen  ions (H )
present  and  is  expressed  in  terms  of pH  units—the  logarithm of  the inverse  activity of
hydrogen  ions.   The  pH  scale  ranges from 0 to  14, with  a value of  7  representing  a neutral
solution.   Solutions   with values less than  7  are acidic,  while values  greater than  7 are
                   4
basic.  Because pH is  a  logarithmic scale, a change of one unit represents a tenfold change in
acidity, hence pH 3 is ten times as acidic as pH 4.  Currently the acidity of precipitation in
the  Northeastern  United States  normally  ranges  from pH  3.9 to 5.0; in  other  regions  of the
United States  precipitation  episodes  with a  pH  as  low  as  3.0  have  been reported.   For
comparison, the pH of some familiar substances are:  cow's milk, 6.6; tomato juice,  4.3; cola
(soft drink) 2.8; and  lemon juice, 2.3.
     The  pH of  precipitation  can vary during  an event,  from event to  event,  from  season to
season,  and from geographical  area  to geographical  area.   Substances  in  the  atmosphere can
cause  the  pH to shift by  making it more acidic  or more  basic.   Dust and debris  swept up in
small  amounts from the ground into the atmosphere may become components of precipitation.  In
the West  and  Midwest  soil  particles tend to  be basic,  but  in the  eastern  United States they
tend  to  be acidic.  Industrial  emissions of limestone particles and similar  oxides  and car-
bonates are basic.  As gaseous ammonia" from decaying  organic matter makes precipitation more
basic, ammonia  influences  the  acidity of precipitation in  areas where  there are large stock-
yards or other sources of  organic matter.
                                            7-2

-------
     In  the  Eastern  United States  sulfur  oxide  emissions are  greater than  nitrogen  oxide
emissions; hence, sulfates  are  greater contributors in this region  to the formation of acids
in precipitation.  The ratio between the two emissions, however, has been decreasing.   Sulfate
concentrations are  greater in  summer  than  in  winter in the  Eastern  United States;  however,
around some  of  the  larger cities in California nitrate^, contribute more to  the  formation of
acidity in rainfall.   In  coastal  areas sea  spray  strongly  influences  precipitation chemistry
by contributing  sodium,  calcium, potassium, chloride, and sulfates.  In the final  analysis,
the pH of precipitation reflects the overall contributions of all  of these components.
     The  impact  of  acidic  precipitation on lakes,  streams, ponds, forests,  fields and manmade
objects,  therefore, is not the  result of one, or even of several  precipitation events, but of
continued  additions of  acids   or  acidifying  substances  over  time.   When  did precipitation
become acidic?   Some  scientists state that  it began with the Industrial  Revolution  and the
burning  of  large amounts  of  coal;  others  say  that,  in the United States,  it  began  with the
introduction  of  tall  stacks   on  power  plants  in  the   1950's;  other  scientists  disagree
completely and state that rain has  always been acidic. In other words,  no definitive answer to
the question exists at the present time, nor  are  there data to indicate with much confidence
trends of pH  in  precipitation  because the  pH  of rain has not  been  continuously  monitored in
the United States for  any  period of time.   In  Scandinavia,  on the other hand, the pH of rain
has been  continuously monitored for  many  years;  it  is  more  nearly  possible,  therefore, to
determine when  precipitation  began to become more acidic.   Data  from the  European Chemistry
Network  in the  late 1950's revealed that the  pH of the precipitation falling on southwestern
Scandinavia was  < 4.7.   At that time  it  was postulated that  the  acidification of freshwater
lakes and streams which  had first been noted  in southern Norway in 1911 and later during the
1920's had been caused by acidic precipitation.
     Though acidic  precipitation (wet  deposition)  is usually  emphasized, it  is  not  the only
process by which acids  or acidifying  substances are  added  to  bodies of water or to the land.
Dry deposition  also occurs.  During  wet deposition,  substances  such as sulfur  and nitrogen
oxides are  scavenged, by  precipitation (rain  and  snow) and deposited on the  surface of the
earth.   Dry  deposition processes include gravitational sedimentation  of particles,  impaction
of aerosols, adsorption  and absorption of  gases by  objects  at the earth's  surface  or by the
soil or  water.   Gases,  solid  and liquid  aerosols  can be  removed by both wet and dry deposi-
tion.   Dew,  fog,  and  frost are also involved  in the deposition processes but do not strictly
fall  into the category of either wet  or  dry deposition.   Dry deposition processes are not as
well  understood  as  wet  deposition  at  the  present time;  however, all of the deposition pro-
cesses contribute  to the  gradual accumulation  of  acidic  or acidifying  substances  in  the en-
vironment.   In  any  event,  precipitation  in  the Eastern United States at the present  time is
acidic and has  been associated with changes  in ponds, lakes,  and streams that are considered
by humans to be detrimental to their welfare.
                                            7-3

-------
     The most  visible changes  associated with  both  wet and dry acidic  deposition  are those
observed  in the  lakes  and  streams  of  the Adirondack  Mountains  in New  York State,  the
Precambrian Shield  areas  of Canada and in  the  Scandinavian countries.   In these regions, the
pH of  the  fresh water bodies has  decreased,  causing  changes in animal  and plant populations.
The most readily observable effect, however, has been the decrease in fish populations.
     The chemistry  of fresh waters is determined primarily by  the  geological structure (soil
system  and bedrock)  of  the lake or stream catchment  basin,  by the ground cover  and  by land
use.   Near coastal  areas,  that is up to about 160 kilometers (100 miles) from the sea,  marine
salts also may  be important in determining the  chemical  composition of the stream,  river, or
lake.
     Sensitivity of a lake to acidification depends on the acidity of both wet and dry deposi-
tion plus  the  same factors—the soil system  of  the  drainage basin, the canopy effects of the
ground  cover, and  the composition of the watershed bedrock—that  determine the chemical com-
position of fresh water bodies.   The capability of a lake and its drainage basin to neutralize
incoming acidic substances, however, is determined largely by the composition of the bedrocks.
     Soft  water lakes,  those most sensitive to  additions of  acidic substances,  are  usually
found in areas  with igneous bedrock which contributes few soluble solids to the surface waters,
whereas hard waters  contain large concentrations of alkaline earths  (chiefly bicarbonates of
calcium and sometimes magnesium and iron) derived from limestones and calcareous sandstones in
the  drainage basin.   Alkalinity  is  associated with  the capacity  of  lakes  to  neutralize or
buffer the incoming acids.   The quantity of acidic precipitation necessary to acidify a sensi-
tive lake  system has yet to be determined.
     The disappearance of fish populations from freshwater lakes and streams is usually one of
the most readily observable signs of lake acidification.  Death of fish in acidified waters has
been attributed to  the  modification of a number of physiological processes by a change in pH.
Two  patterns of pH change have been observed;  the  first involves a sudden short-term drop in
pH and  the second,  a gradual decrease in  pH  with time.   Sudden short-term drops  in pH often
result  from a winter thaw or the  melting  of  the snow pack in early spring and the release of
the  acidic constituents  of  the   snow  into  the  water.    These short-term  changes  in  water
chemistry  may have significant impacts on aquatic biota,  especially if they occur at sensitive
times in the life cycle (e.g., during spawning or early stages of development).
     A gradual   decrease in pH, particularly below 5, can interfere with reproduction and spawn-
ing  of  fish until  elimination of the population occurs.   In some lakes, aluminum mobilization
in fresh waters at a pH below 5 has resulted in fish mortality .
     Although the  disappearance of  and/or reductions in fish  populations  are usually empha-
sized as significant results of lake and stream acidification, changes of equal or greater im-
portance are the effects on other aquatic  organisms  ranging from waterfowl to bacteria.  Or-
ganisms at all  trophic  (feeding) levels in the  food  web appear to be affected.   Reduction in
                                            7-4

-------
number and diversity  of  species  may occur; biomass (total mass of living organisms in a given
volume of water)  may  be  altered and processes,  such  as primary production and decomposition,
impaired.
     Primary production  and decomposition are the  bases  of the two major  food  webs  (grazing
and detrital) within an ecosystem by which energy is passed along from one organism to another
through a  series of  steps  of eating and  being  eaten.   Green plants, through the  process  of
photosynthesis, are the  primary  energy  producers in the  grazing  web,  while bacteria initiate
the detrital food web by feeding on dead  organic  matter.   Disruption of either of these two
food webs results in a decrease in the supply of nutrients, interferes with their cycling, and
also reduces energy flow within  the affected ecosystems.   Acidification  of lakes  and streams
affects both these processes when a slowing down in the rate of microbial  decomposition causes
an alteration of the species composition and structure of the pondweed and algae plant communi-
ties.
     At present,  there are  no documented  observations or measurements of  changes  in natural
terrestrial ecosystems that  can  be directly attributed to acidic precipitation.   The informa-
tion available is an  accumulation of  the results of  a  wide variety of controlled research
approaches largely  in the laboratory,  using in most instances some form of "simulated" acidic
rain,  frequently dilute sulfuric acid.
     Soils may become gradually  acidified from an  influx  of hydrogen (H ) ions.  Leaching of
the mobilizable  forms of mineral nutrients may  occur.  The rate of leaching  is  determined  by
the buffering  capacity of  the  soil as  well  as the amount  and  composition -of precipitation.
Unless the buffering  capacity of the soil is high and/or the salt content of precipitation is
high,  leaching will, in time, result in acidification.   At present, there are no studies show-
ing this process has occurred because of acidic precipitation.
     Damage to stone  monuments  and buildings, corrosion of  metals and deterioration of paint
can result  from  acidic  precipitation.   Because  sulfur compounds  are a dominant component  of
acidic precipitation  and  are deposited  during dry deposition also, the effects resulting from
the two processes cannot be distinguished.   In addition, the deposition of sulfur compounds on
stone surfaces provides  a medium for microbial  growth  that can result in deterioration. (See
Chapter 10 for a more detailed discussion of materials effects.)
     Human health effects due to the acidification of  lakes and rivers have been  postulated.
Fish in acidified water  may contain toxic metals  mobilized due to the acidity  of  the water.
Drinking water may  contain toxic metals or leach  lead from the pipes bringing water into the
homes. Humans  eating  contaminated  fish  or drinking contaminated water could  become  ill.   No
instances of these effects having occurred have been documented.
    , Several  aspects  of  the acidic precipitation problem  remain subject to debate because
existing data are ambiguous or inadequate.   Important unresolved issues include:
     1.  The rate at which rainfall is becoming more acidic and the rate at
         which the phenomenon is becoming geographically widespread.

                                            7-5

-------
     2.   The  relative  extent to which the  acidity  of rainfall in a  region  depends  on local
          emissions  of  nitrogen and  sulfur oxides versus emissions  transported from distant
          sources.
     3.   The  relative  importance  of  changes in total mass emission rates compared to changes
          in the nature of the emission patterns e.g., ground level versus tall stacks in con-
          tributing to regional acidification of precipitation.
     4.   The  relative  contribution of wet  and dry deposition to  the acidification  of lakes
          and streams.
     5.   The  geographic  distribution  of  natural  sources of  NO   and SO  ,  and NH,  and  the
                                                                 XX         i5
          significance and seasonality of their contributions.
     6.   The  existence and  significance  of anthropogenic, non-combustion sources of SO ,  NO
          and HC1.
     7.   The  dry  deposition  rates  for  S02,  NOp,  sulfate,  nitrate and  HC1   over  various
          terrains and seasons of the year.
     8.   The  existence  and  reliability of  long-term pH measurements  of  lakes  and headwater
          streams.
     9.   The  acceptability  of current  models for predicting long  range  transport of SO  and
          NO  and of those for predicting the acid tolerance of lakes.
            X
    10.   The  feasibility and costs  of  using liming or other corrective procedures to prevent
          or reverse damage from acidification.
    11.   The  differential  effects of  SO   and NO   and hydrogen  ion deposition on ecosystem
                                         
-------
fields, lakes, rivers, estuaries, and oceans) within the biosphere obtain energy from the sun,
nutrients from the  earth's  crust (the lithosphere), gases from the atmosphere, and water from
the hydrosphere.   All of the  living systems  are  interdependent.  Energy  and nutrients move
from  one to  another.  The  living  systems,  together with  their physical  environments,  the
lithosphere,  hydrosphere  and  atmosphere  make  up the  ecosystem  that is  the planet  Earth
(Billings, 1978;  Boughey,  1971; Odum, 1971; Smith,, 1980).
     Ecosystems are composed of  biotic  (living) and abiotic (nonliving) components.  The bio-
tic component  consists  of:   (1)  producers, green  plants that capture the  energy  of the sun;
(2) consumers  that  utilize  as  their energy source  the  food stored by  the  producers;  and (3)
decomposers who obtain  their  energy by breaking down and converting  dead^organic matter into
inorganic  compounds.   (See  Table  7-1).   The  abiotic   components  are air,  water,  the soil
matrix,  sediment,  particulate matter,  dissolved  organic  matter,  and  nutrients  in  aquatic
systems, and dead or  inactive organic matter in terrestrial systems (see Table 7-1) (Billings,
1978;  Boughey, 1971; Smith,  1980).
     Ecosystems are  basically energy processing systems  "whose  components have  evolved to-
gether  over  a long period  of time.  The  boundaries  of  the system are determined  by the en-
vironment, that is, by  what forms of life can be sustained by the environmental conditions of
a  particular  region.   Plant and animal populations  within the  system  represent  the objects
through which the system functions." (Smith, 1980).
     Ecosystems are  open  systems.   They both  receive from and contribute  to  the environment
that  surrounds  them.-  The  environment  contributes gases,  nutrients,  and  energy.   Ecosystems
utilize  these substances  and,  in  turn,   make their  own  contributions  to the  environment.
Energy  flows  through  the  system  unidirectionally while water, gases and nutrients are usually
recycled and fed back into the system.  The functioning of ecosystems is greatly influenced by
the extent to  which the gases and  nutrients are fed back into the system. When materials are
not returned  to  an ecosystem  through  recycling,  they  must be obtained in another-way.  The
organismal populations are the structural  elements of the ecosystem through which energy flows
and nutrients are cycled.   (Smith, 1980; Billings, 1978; Odum, 1971).
     Energy from the  sun  is the driving force in  most  ecosystems.   Without it, virtually all
ecosystems would cease to function.   The energy of the sun is captured by green plants through
the process  of photosynthesis  and  is stored  in plant tissues.   This  stored energy is passed
along  through ecosystems   by  a  series of  feeding  steps,  known  as  food  chains,  in  which
organisms  eat  and  are eaten.   Energy flows  through ecosystems in two  major food chains, the
grazing  food  chain  and  the  detrital food chain.  The amount of energy that passes through the
two food chains  varies  from community  to  community.   The detrital  food chain is  dominant in
most terrestrial  and  shallow-water ecosystems.   The grazing food  chain may be dominant in deep-
water  aquatic  ecosystems  (Smith, 1980).  The  two  fundamental  processes involved in these two
food  chains are:   (1) photosynthesis, the capture of energy from the sun by green plants, and
(2) decomposition,  the  final  dissipation  of  energy and the reduction  of  organic matter into
inorganic nutrients.
                                            7-7

-------
                   TABLE 7-1.  COMPOSITION OF ECOSYSTEMS
     Component
     Description
Biotic (biological):
 Individuals

 Producers
 Consumers
 Decomposers

 Populations

 Communities
Plants, animals (man), and microorganisms.
 These are either producers, consumers, or
 decomposers.
Green plants.
Herbivores, carnivores.
Macroorganisms (mites, earthworms, millipedes,
 and slugs) and microorganisms (bacteria
 and fungi).
Groups of interbreeding organisms of the same
 kind, producers, consumers or decomposers,
 occupying a particular habitat.
Interacting populations linked together by
 their responses to a common environment.
Abiotic (physical):
 Atmosphere
Cosmic radiation, temperature, thermal
 radiation, radioactivity including fallout.
Water vapor, cloud and precipitation.
Gases, pressure and wind,
Heat and temperature,
Fire and pollutants
 Lithosphere
Rock and soil particles
Minerals, water,
Radioactivity, heat and temperature, gases
Topography
 Earth mass
Gravity
Adapted from Billings (1978).
                                       7-8

-------
     In addition  to the  flow of energy,  the  existence  of the living world  depends  upon the
circulation of  nutrients  through the ecosystems.  Both  energy  and nutrients move through the
ecosystem as organic matter.   It is not possible to separate one from the other.  Both influ-
ence the  abundance of  organisms,  metabolic rate  at which they live, and  the complexity and
structure of  the  ecosystem  (Smith,  1980).  Nutrients,  unlike energy, after moving  from the
living to the nonliving return to the living components of the ecosystem in a perpetual cycle.
It is  through  the cycling of nutrients  that plants  and animals obtain the minerals necessary
for their existence.
     The gaseous  and sedimentary cycles are the two basic types of nutrient or biogeochemical
cycles.  The gaseous cycles  involve carbon, oxygen,  and nitrogen.   Water,  also, is sometimes
considered as belonging to the gaseous cycle.   In the gaseous cycles, the main nutrient reser-
voir is the  atmosphere  and the ocean.  In the sedimentary cycle, to which phosphorus belongs,
the reservoir  is  the  soil and rocks of the earth's crust.  The sulfur cycle, a combination of
the two cycles, has reservoirs in both the atmosphere and earth's crust.
     Nitrogen,  sulfur and water  cycles are involved-in  acidic  deposition.   Nitrogen, through
the agency of plants (chiefly legumes and blue green algae) and microorganisms, moves from the
atmosphere to  the  soil  and  back (see  Figure  7-1).   Human intrusion  into  the nitrogen cycle *
includes  the  addition  of nitrogen  oxides to the  atmosphere, and  ammonia and  nitrates  to
aquatic ecosystems.   Sulfur  enters  the atmosphere  from volcanic  eruptions,  weathering,  the
surface of the  ocean,  gases  released in the decomposition processes, and combustion of fossil
fuels  (see Chapter  4  for details).   There is  no way to harness or use energy without creating
an environmental  impact.   Both  the  nitrogen   and  sulfur cycles have  been  overloaded in some
parts of the world by the combustion of fossil  fuels by man.   For these cycles to function, an
ecosystem must possess a  number of structured  relationships among its components.  By changing
the amounts of  nitrogen and  sulfur moving  through the  cycles,  humans have perturbed or upset
the structured  relationships  that  have existed for millions of years and altered the movement
of  the elements  through  the  ecosystems.   The  pathways that  the  elements take  through  the
system depend  upon  the  interaction  of the  populations  and their relationships to each other.
     Change is  one  of  the basic characteristics of our environment.   Weather changes from day
to day, temperatures rise and fall,  rains come and go, soils erode, volcanoes erupt, and winds
blow across  the  land.    These are natural phenomena.  Significant environmental  changes also
result when  human beings clear  forests,  build cities and factories, and dam  rivers.   All  of
these  environmental changes  influence  the organisms  that live in  the  ecosystems where  the
changes are occurring (Moran et al., 1980).
     Existing studies indicate that changes occurring within ecosystems,  in response to pollu-
tion or other  disturbances,  follow definite patterns that  are  similar even in different eco-
systems.   It is possible, therefore, to predict the basic biotic responses of an ecosystem to
disturbances caused by  environmental  stress  (Woodwell,  1970;  Woodwell,  1962;  Odum,  1965).
These  responses  to disturbance  are:   (1)  removal  of sensitive organisms  at  the species and

                                            7-9

-------
                                                              	Jf-~*^,  -^     minim
                                                             '   \  f V   sxcntiioN  /
                                                             '•   NMj'  \  06ATH  4 /
                                                             v_j/ I   \     ORGANIC N
                                                             >=ZLj~~~±e—~—p-~~~'
                                                             I*'  N IN SEDIMENTS  r
                                                        •«; "^ SEE-AOi
Figure 7-1.  Schematic representation of the nitrogen cycle, emphasizing human activities that affect fluxes of
nitrogen (National Research Council, 1978).

-------
subspecies level due  to  differential  kill; (2) reduction  in  the number of plants and animals
(standing crop); (3) inhibition of growth or reduction in productivity; (4) disruption of food
chains; (5) return  to a  previous state of  development;  and (6) modification in  the  rates  of
nutrient cycling.
     Ecosystems  can  respond  to  environmental  changes  or  perturbations  only  through  the
response  of  the  populations  of  organisms of which  they are  composed (Smith,  1980).   The
species of organisms  sensitive  to environmental changes are removed.   Therefore, the capacity
of an  ecosystem to maintain  internal stability  is determined  by  the  ability  of individual
organisms  to  adjust  their physiology  or  behavior to  change.  The  success  with which  an
organism copes with environmental  changes is determined by its ability to produce reproducing
offspring.  The size  and success  of a  population  depends  upon  the collective  ability  of
organisms  to  reproduce   and  maintain  their  numbers  in  a  particular  environment.   Those
organisms  that  adjust  best  contribute  most  to  future  generations  because  they have  the
greatest  number  of progeny in  the population (Smith, 1980; Billings,  1978; Woodwell,  1970;
Odum, 1971).
     The capacity of  organisms  to withstand injury from  weather extremes,  pesticides,  acidic
deposition or  polluted  air follows the  principle of  limiting  factors  (Billings,  1978;  Odum,
1971; Moran et a!., 1980; Smith, 1980).   According to this principle,  for each physical  factor
in the  environment  there exists for each  organism  a  minimum  and a maximum limit beyond which
no members  of a  particular  species can survive.   Either  too much  or too  little  of  a  factor
such as heat,  light,  water,  or minerals (even though they are necessary for life) can jeopar-
dize the  survival  of  an  individual and,  in extreme cases, a species  (Billings,  1978;  Smith,
1980; Boughey,  1971;  Odum,  1971).   The range of tolerance (see Figure 7-2) of an organism may
be broad  for one  factor and narrow  for another.  The  tolerance  limit  for each  species  is
determined by  its  genetic makeup and varies from species to species for the same reason."  The
range  of  tolerance also  varies depending on  the age,  stage  of growth, or growth  form  of  an
organism.   Limiting  factors  are,  therefore,  those  which, when  scarce  or overabundant,  limit
the  growth,  reproduction, and/or  distribution of  an  organism  (Billings,  1978;  Smith,  1980;
Boughey, 1971;  Odum,  1971;  Moran et a!., 1980).  The increasing acidity of water in lakes and
streams is such a factor.
     Organisms can exist only within their range of tolerance.  Some populations of organisms,
annual   plants,  insects,  and  mice, for example, respond rapidly to environmental change.   They
increase  in  numbers   under  favorable  conditions  and  decline rapidly  when conditions  are
unfavorable.    Populations of other  organisms, such  as  trees  and  wolves,  fluctuate less  in
response  to  favorable or unfavorable conditions  by showing little variation in  the  rates  of
reproduction  and  death.   Adaptation  is  the  ability  of  an  organism  to  conform  to  its
environment.    Ecosystem  stability ultimately  is based on  the  adaptability  of  the organisms
that compose  it.   Stability  may be associated with the  ability of a system to  return  to  an
equilibrium  state after a temporary disturbance  (Holling,  1973;  May,  1973).   The less  it

                                            7-11

-------
  ZONE OF
INTOLERANCE
               tOWtR UNITS
               OF TOLERANCE
               ZONE OF
            PHYSIOLOGICAL
               STRESS
TOLERANCE RANGE
RANGE OF OPTIMUM
   ZONE OF
PHYSIOLOGICAL
   STRESS
                                          ZONE OF
                                        INTOLERANCE
              ORGANISMS
             INFREQUENT
 ORGANISMS
  ABSENT
                                       GREATEST
                                      ABUNDANCE
                             ORGANISMS
                            INFREQUENT
                                         ORGANISMS
                                          ABSENT
   LOW4-
  -BBAOIiNT-
              -*HtGH
                        Figure 7-2.  Law of tolerance.

                        Source:  Adapted from Smith (1980).
                                         7-12

-------
varies  from and  the faster  it  returns  to its  original  state, the  more stable  the  system
(Smith, 1980).   Stability also involves persistence, the ability of the populations of an eco-
system to persist  through time.   Persistence involves resilience, the ability of an ecosystem
to absorb changes.   Although  individual  populations within a  system may fluctuate greatly in
response to environmental  changes,  the system may  be  highly  resilient (Holling, 1973;  Smith,
1980).  Contrasted  with resilience  is resistance, the  ability of  a system, because  of its
structure, to resist changes from disturbances.   Typically, the most resistant ecosystems have
large  living  components, trees  for example, and store nutrients and energy in the standing
biomass.  Resistant systems such as forests, once highly disturbed,  are very slow in returning
to their original state (Smith, 1980).
     Aquatic ecosystems  that  lack components in which energy  and nutrients may be stored for
long  periods  of time usually  are  not very resistant to environmental changes (Smith,  1980).
For example, an influx  of pollutants, such  as  effluents from sewage disrupts the  system be-
cause  more  nutrients enter the system than  it  can 'handle.  However, since  the  nutrients are
not retained or recycled within the  system  it  returns  to  its original  state in a relatively
short time after the perturbation is removed.
     No barriers  exist  between the  various  environmental  factors,  or between an organism or
biotic community and its environment.  Because an  ecosystem  is  a complex of interacting com-
ponents, if one factor  is changed, almost all  will change eventually.   "The ecosystem reacts
as a  whole.  It is practically impossible to wall  off a single  factor  or organism in nature
and control it  at  will  without affecting the rest of the ecosystem.   Any change no matter how
small   is  reflected  in some way throughout the ecosystem:  no 'walls' have yet been discovered
that prevent these interactions from taking place"  (Billings,  1978).
     Continued  or  severe  perturbation of an ecosystem can  overcome  its  resistance or prevent
its recovery, with  the  result that the original  ecosystem will  be replaced  by  a new system.
In the  Adirondack  Mountains  of New York  State,  eastern  Canada,  and parts of Scandinavia, the
original aquatic ecosystems  have  been and are continuing to be replaced by ecosystems differ-
ent from  the original due to  acidification  of  the aquatic habitat.   Forest ecosystems appear
to be  more  resistant thus far, because changes  due to stress from acidifying substances have
not been detected.
7.2  CAUSES OF ACIDIC PRECIPITATION
7.2.1  Emissions of Sulfur and Nitrogen Oxides
     The  generally  held hypothesis  is that increased  emissions of  sulfur  and  nitrogen com-
pounds  are  largely responsible for  the  acidity of precipitation (Smith,  1872;  Bolin et al.,
1972;  Likens, 1976).  The emissions of the sulfur and nitrogen compounds involved in the acidi-
fication  are  attributed chiefly  to the  combustion of  fossil fuels.  Emissions from natural
sources can also  be involved; however, in  highly industrialized areas emissions from manmade
sources usually exceed those from natural sources (see Chapter 4).  (See Chapter 5 for environ-
mental concentrations of sulfur compounds.)

                                            7-13

-------
     After 1900  there  was a nearly exponential  increase  in the consumption of gas and oil in
the United States  (see Figure 7-3). Although  the  total  consumption of coal has not increased
greatly since  about  1925, the consumption of oil and gas has continued to rise precipitously,
thus overshadowing coal  as the dominant fuel source during the past 50 years (Hubbert, 1976).
Within this overall increase in fossil-fuel use, there have been shifts in the pattern of con-
sumption.   Whereas formerly a considerable proportion of coal was used for transportation and
heating, these functions  have since been taken over by oil and gas.  Coal is now predominantly
devoted to electric  power generation (Figure 7-4).  In fact, electric power generation is the
primary  factor  accounting  for an  absolute increase  in  coal  consumption  over the  past two
decades.  (The decline in the use of coal in the 1930s was due to the general economic depres-
sion, and the  decline  in the 1950s was  due to the availability of relatively inexpensive oil
and gas.)  Approximately 550 MM (million metric) tons (National Research Council, 1978b) were
used per  year during  1918-1928 compared  to  672 MM  tons/year during  1979  (Hamilton, 1980).
There was, however, a seasonal shift in the pattern of coal consumption.  Summer coal consump-
tion has  increased since 1960 because of  increased  usage of electricity in air conditioning,
while winter consumption  has decreased due to the use of alternative fuels.
     These changes in  the pattern of fuel use have been accompanied by changes in the pattern
of pollutant emissions.   Figure 7-5A and 7-5B  illustrate  the rise since 1940 in emissions of
sulfur  and  nitrogen  oxides,  the  primary gaseous pollutants  resulting  from  the combustion of
fossil fuels (See also Chapter 4).  Although there has been a net increase in both categories,
the more  consistent  rise has been  in  emissions of nitrogen  oxides.  Almost  all  (93 percent)
emissions of sulfur  oxides in the United  States  arise from stationary point sources, princi-
pally industrial and power plant stacks.   Nitrogen oxide pollutants, on the other hand, origi-
nate  about  60 percent  from transportation  (mobile)  sources  and  30 percent  from stationary
                   «
sources, which include not only industrial and power plants, but residential and institutional
heating equipment as well (U.S. Environmental  Protection Agency, 1980). (see Chapter 5 of Air
Quality Criteria for Oxides of Nitrogen for a more detailed discussion.)
     The geographic distributions  of sources of the presumed gaseous precursors of acidic pre-
cipitation are depicted in Figures 7-6 and 7-7.  Clearly,  the main sources of sulfur oxides in
the  United  States are  in the eastern  half  of  the  country, particularly  the  northeastern
quadrant.   Major nitrogen oxide sources also  show a tendency to  be  concentrated somewhat in
the Northeast.   Chapter  4 should  be consulted  for  a  more detailed account of the sources and
emissions of sulfur oxides.
7.2.2  Transport of Nitrogen and Sulfur Oxides
     Among the factors  influencing  the formation as well  as the location where acidic deposi-
tion occurs is  the  long-range transport of nitrogen and sulfur oxides.   Neither the gases nor
their transformation  products always remain near the sources from which they have been emitted.
                                            7-14

-------
 I
10

 o
 o


 o
 cc
 LU
     50
     40
     30
20
     10
      1850
                                          fOlL
              1900
                            1950



                           YEAR
2000
                                                           2050
     Figure 7-3. Historical patterns of fossil fuel consumption in the

     United States



     Source: Adapted from  Hubbert (1976).
                        7-15

-------
o
oo

O
O



I
O



ce
800



700



600



BOO



400



300



200



100
            1     I     I     I
                         I     I
I     I
                            TOTAL
RES. & COMM,

    HEATING

  RAILROADS Hi

    I     I     I
                                        - OTHER



                                        -OVENCOKE
                                         ELECTRIC

                                         UTILITIES
                                                  I	I
      1900  10    20   30   40   50    60    70   80   90  2000


                                YEAR



Figure 7-4.  Forms of coal usap in the United States, Electric
power generation is currently the primary user of coal. (Data
from U.'S. Bureau of Mines, Minerals Yearbooks 1933-1974).



Source: U,S. Bureau of Mines (1954, 1976).
                                 7-16

-------
(a
O
S
tu
d*
tn
ta
O
    35
    30
    2F
    20 _
    15
    10
     1940
    35

    30

.S   25
4=
09

§
    20
    15
    10
     0 I	
     1940
                      TRANSPORTATION
                 1950
                             1960

                            YEAR
             1970
            1980
                              T
                 TOTAL
                         TRANSPORTATION

                               I	I
                 1950
 1960

YEAR
1970
                                                     1980
 Figure 7-5a. Trends in emissions of sulfur dioxides.
 Figure 7-5b. Trends in emissions of nitrogen oxides.
 Source:  Office of  Air  Quality  Planning and  Standards
 (1978).
                              7-17

-------
        KEY
EMISSION  DENSITY, tons/mi"2
20-50

>50
    5'10
EH 10-20
              Figure 7-6.  Characterization of U.S. SOX emissions density by state.
             {Roman numerals indicate EPA Regions.)

              Source: U.S. Dept. of Energy (1981).

-------
                                                     0)
                                                     +*
                                                     R
                                                     S
                                                     0)
                                                     o

                                                    Jj


                                                     Q)

                                                     X «-.


                                                    i   1

                                                    °?  "1?
                                                    3  (T
                                                    H«   ^
                                                     O
                                                     o
                                                    i
                                                         a.
                                                         UJ
                                                     S  "O
                                                     tS  .5
                                                     2   «
                                                     co   eg
1
 O

Ul
 e
Q

(0
3*
                                                    w   i

                                                    *   S    »
                                                    ^   c    §

                                                     S   I    i
                                                     o>   o    o
                                                    E  £    co
7-19

-------
They may be transported for long distances downwind (Altshuller and McBean, 1979; Pack el al.,
1978;  Cogbill  and Likens, 1974).  (See  Chapter  6 for a detailed  discussion  of transport and
transformation of sulfur oxides.)
     The  geographic  picture of  the  problem  of  acidic precipitation  in  North  America can be
better understood  in  the  light of some  information on prevailing wind patterns.  Winds trans-
port  the  precursors of acidic precipitation  from  their points of origin  to  areas  where the
acidified rain and  snow eventually fall.  Prevailing  winds  in the Eastern United States tend
to be from the'west and southwest.  Atmospheric pollutants, therefore, are carried in a gener-
ally  northeasterly  direction.   Thus,  pollution  originating  in  the Ohio  River Val.ley can be
carried toward the  New England States.  Seasonal meteorological patterns, however, can modify
the direction  of windflow,  particularly in the summer.  The Maritime Tropical air masses from
the Gulf of Mexico that occur in late summer have the greatest potential for the formation and
transport of high concentrations of sulfate into the Northeastern United States and into east-
ern Canada (Altshuller and McBean, 1979).
     Cogbill and  Likens (1974)  associated acidic rainfall in  central  New York during 1972-73
with high altitude  air masses  transported into the region from the Midwest.  They stated that
the NO  and  S02  that is involved in acidic rain formation may be transported distances of 300
to 1500 km.   Reports  by Miller et al.  (1978),  Wolff et al.   (1979),  and  Galvin et al. (1978)
all support  the  concept that the trajectories of the  air masses which come  from the Midwest
carry sulfur and nitrogen  compounds that acidify precipitation in New York State.
     A significant, though  disputed  factor,  in this transport  picture is the height at which
the pollutants are  emitted.   Industrial  and power plant smokestacks emit their effluents into
the atmosphere  at higher elevations than  do  motor  vehicles  or most  space heating equipment.
In fact,  there has  been a trend  since  the 1960s toward building  higher  stacks as a means of
dispersing pollutants  and thereby reducing pollutant concentrations  in the vicinity of power
plants, smelters,  and similar  sources   (Grennard and  Ross,  1974).  The  result has  been that
sulfur and nitrogen  oxides  are carried  by prevailing  winds  for long distances and allowed to
diffuse over greater areas through the atmosphere.  Concomitantly, long-range transport allows
greater time  for chemical  reactions  to  convert these  pollutant gases into particulate forms.
(Eliassen and Saltbones,  1975;  Smith and Jeffrey,  1975;  Prahm et al., 1976).   Chapter 6 dis-
cusses the chemical transformations,  wet and dry deposition,  transport and diffusion of sulfur
oxides in the  atmosphere.   Sulfur dioxide and nitrogen  oxides are oxidized and hydrolyzed in
the atmosphere to  nitric  (HNO~) and sulfuric (H-SQ,) acids.   These acids are considered to be
the main components of acidic precipitation.
     The mechanisms of these chemical reactions are quite complex and depend on a host of vari-
ables ranging from physical  properties of the pollutants to weather conditions and the presence
of catalytic or interacting agents (Fisher, 1978).  Although the atmospheric chemical processes
are not well  understood,  it does appear that the long-range transport of sulfur compounds can
cover 1000 to 2000 km over three to five days (Pack et al., 1978).  Thus,  the impact of sulfur
                                            [
                                            7-20

-------
 pollutants  in  the  form  of  acidic  precipitation may  be  far  removed  from  their  points of origin.
 It  is not yet clear whether  nitrogen  oxide pollutants may be  transported  distances comparable
 to  that of sulfur compounds  or  are local in origin (Pack,  1978); in the Northeast, however,
 nitrates  are  currently thought to  contribute 15  to 30 percent of  the acidity of  polluted pre-
 cipitation.   This  estimate has increased  over the  past few years  and  is  expected to increase
 still  further  in the future (Robinson  et  al., 1978).
      Evidence  from northern Europe  (Oden,  1968) also supports  the  idea  that acidic rainfall  is
 a   large-scale regional   problem  involving  long  distances   between   emission  sources  and
 deposition  of acidic  precipitation.   The  acid  rains  that have  received  intensive  study  in
 southern  Scandinavia have  been shown to result primarily from  emissions  of nitrogen and  sulfur
 oxides in  Great   Britain  and the  industrial regions of continental   Western  Europe,  e.g.,
 Holland,  Belgium,  and West Germany  (Brosset,  1973).
 7.2.3   Formation
      Precipitation is   that portion of the  global-water  cycle  by which water vapor from the
 atmosphere  is converted to  rain or  snow  and  then deposited on  the  earth  surfaces (Smith,
 1980).  Water moves  into  the atmosphere  by evaporation and transpiration  (water  vapor lost  by
 vegetation).   Once it  reaches  the atmosphere,  the water vapor is cooled,  then condenses  on
 solid particles and soon reaches  equilibrium with atmospheric  gases.  One  of  the  gases is car-
 bon dioxide.   As carbon dioxode dissolves  in water, carbonic acid  (H^GOO  is  formed.  Carbonic
•acid is a weak acid and in  distilled water only dissociates  slightly,  yielding  hydrogen ions
 and bicarbonate  ions   (HCQ.,  ).   When relatively pure water   is  in equilibrium with  normal
 atmospheric concentrations  and  pressures  of carbon   dioxide,  the  pH  of rain  and  snow  is
 approximately  5.6  (Likens  et  al., 1979).
      The  pH of precipitation may vary and become more basic  or more acidic  depending on sub-
 stances  in  the  atmosphere.   Dust and  debris may be  swept  from the  ground   and  into  the
 atmosphere  in small amounts  where it can become  a   component  of rain.  The amounts  of the
 various  substances  in  the  atmosphere   originating   from seawater,  desert   sands,  volcanic
 islands,  or vegetated   land influence  the chemistry of natural  precipitation.  In regions with
 calcareous  soils, calcium   and  carbonate   may  enter  precipitation  as dust,  subsequently
 increasing  the pH of rain or snow to 6.0 or above (Likens et al., 1979).  Soil  particles are
                                                                                          2+
 usually slightly  basic  in distilled water and  release positive ions,  such as calcium (Ca   ),
 magnesium (Mg  ),  potassium (K ),  and sodium (Na+) into solution.   Bicarbonate usually  is the
 corresponding  negative ion.   Decaying organic matter  adds gaseous ammonia to the atmosphere.
 Ammonia gas, in  rain  or .snow forms ammonium  ions  (NH, )  and tends to increase the pH.    In
 coastal areas, sea spray pJ.ayS' a  strong role  in the chemistry  of precipitation.   The  important
 ions  entering into  precipitation—sodium,  magnesium, calcium,   potassium,  and  the  anions
            •»                  O«~
 chloride  (Cl  ) and sulfate (SO.   )—are also  those  most abundant in ocean  water (Likens, 1976;
 Likens et al., 1979).
                                             7-21

-------
     Gases,  in  addition to  CQ^,  which enter precipitation are  ammonia,  sulfur dioxide (S02)
and the nitrogen oxides (NO  ).  Sulfur gases originating from natural sources (e.g., volcanoes
and swamps)  may also enter precipitation.  In the wet atmosphere, S0« can be oxidized to sul-
furic acid  (Likens,  1976;  Likens et a!., 1979).  Strong acids dissociate completely in dilute
aqueous solutions and may lower the pH of precipitation to less than 5.6.  Rain or snow with a
pH below 5.6 has been arbitrarily considered acidic precipitation by many scientists (Galloway
and Cowling, 1978; Wood, 1975; Likens et al., 1979).
     The most  important natural  sources of atmospheric sulfur  are  biologically reduced com-
pounds coming from  lake and sea bottoms, marshes,  swamps,  polluted estuaries, streams, tidal
basins, and decaying vegetation.   Such reduction occurs most readily  under oxygen deficient
conditions  when organic  matter  is  present  (Cullis  and Hirschler,  1980).   It  is  virtually
impossible  to  measure or calculate  the quantities of volatile  sulfur  compounds emitted from
biogenic sources, mainly due to lack of quantitative information concerning their reactions in
the atmosphere  (Cullis and  Hirschler,  1980).   Hydrogen sulfide, assumed  initially  to be the
predominant  reduced  sulfur  compound,   reacts  very  slowly  with oxygen  in  the absence  of
catalysts,  but  is oxidized  photochemically (Eggleton and Cox, 1978).   Estimates for the life
time of H,S near the Earth's surface range from 2 hours (Cadle and Ledford, 1966) to nearly 4
days (Friend,  1973).  Dimethyl  sulfide has been  shown to be present  in  considerably higher
concentrations above  a  pond (Rasmussen, 1974); however, it is much less readily oxidized than
HqS (Cadle, 1976).    Biogenic sulfur  emissions  are  much  higher over  the oceans  than land
(Cullis and Hirschler, 1980). (See also Chapter 4.)
7.2.3,1   Composition and pH of Precipitation—Sulfur  and nitrogen  compounds are  chiefly re-
sponsible  for  the excess acidity of  precipitation.   Continuous  measurement of  pH  in  rain by
Likens et  al.  (197^) for the Hubbard  Brook Experimental  Forest  in New Hampshire from 1964 to
1971 indicated  the  precipitation was acid with an annual weighted average pH range of 4.03 to
4.19.   (A  weighted  average  takes into account the amount of rain as well as its composition.)
Cogbill and Likens   (1974),  analyzing precipitation from  the Ithaca area  and  Hubbard Brook,
which consistently  had  a pH of  less  than 4.4,  reported that their  analysis of precipitation
showed that  65  percent  of  the acidity  was  due  to H,SO_ 30  percent  to  HN03,  and less than 5
percent was due to HC1.
     In 1976,  Likens (1976)  reported that  the  continued monitoring of precipitation at the
Hubbard Brook  Forest  through 1974  indicated the  average  annual weighted  pH for  the  years
1964-1974 ranged from 4.03  to 4.21.  There  was  a downward trend in annual  pH  values  between
1964-65 and 1970-71  followed  by an  upward trend until 1973-74, but no  statistically signi-
ficant trend was noted for the 1964-74 period; however, pH values of 2.1 and 3.0 were observed
for individual  storms  at  various locations.  There was an increase in nitric acid in the pre-
cipitation (rain and snow) falling there.  This change in the composition of acidic precipita-
tion suggests  that  the sources  of  nitrogen oxide emissions  increased while those  for sulfur
oxides remained constant.

                                            7-22

-------
     The acidity of  precipitation  is a reflection of the hydrogen ions in precipitation.   The
contribution of sulfate  and  nitrate anions has changed with time, and analysis indicates  that
the nitrate anion  makes  up an ever-increasing fraction of the total negative ion equivalents.
                                                                                +       2-
Following the reasoning of Granat (1972), Likens et al, (1^,76) found [assuming 2H  per SOt   ion
as in hLSO. or  one H  ion per  NO-  as in (HNO,)] that  the  contribution of sulfate to acidity
declined from 83 to 66 percent of the total acidity between 1964 tp 1974 at Hubbard Brook, and
the contribution of  nitrate  increased from 15 percent  to  30 percent of the  total  during the
same period.  Furthermore, increased annual  input of H  was closely correlated with increased
input of nitrate,  but there was little correlation between H  input and sulfate input.
     Data for nitrate, ammonium,  and sulfate in  rain  at Ithaca and Geneva, New York, consti-
tute the longest record  of precipitation chemistry in the United States (Likens, 1972).   Data
are available from 1915 to the present, but long gaps exist in the measurements, especially at
the Geneva site.  Figures 7-8 (A) to (C) show that marked changes in composition have occurred
at Ithaca:  a  gradual  decline in ammonium, an increase in nitrate beginning around 1945,  and a
marked  decrease in  sulfate   starting  between 1945  and 1950.   Early  data for  Ithaca  showed
higher concentrations of sulfate in winter than in summer, presumably because of greater local
burning of  coal  in winter.   Data for 1971 showed the reverse trend, however, with nearly half
the  annual  sulfate  input occurring during  the months  of  June to  August.   Likens  (1972)
concluded that,  despite  deficiencies  in the historical  data and  questions  concerning  their
reliability, the trends are real and can be explained by changes in fuel consumption patterns,
i.e.,  natural  gas  began to  replace coal for  home heating  near the  time  of the  shifts  in
precipitation chemistry.   On  the  basis  of  United States  Geological  Survey  data  for  nine
stations, Likens (1976)  reported a sharp increase in nitrate concentrations in New York State
during the past decade (Figure 7-8 [D]).
     Data  for eastern  North America  (the  U.S.  east  of  the  Mississippi River)  indicate  a
roughly three-fold increase   in  nitrate in  rainfall  since 1955, whereas  sulfate  in rain has
roughly  doubled  in   this  period.   According  to  Nisbet  (1975),  sulfate/nitrate  ratios  in
rainfall averaged about 4 in the Eastern United States in 1955-1956, but the average ratio had
fallen to about  3  in 1972-1973.  Nisbet  calculated  that the fraction of.H  deposition attri-
butable to  nitrate rose  from 19 percent  in  1955-1956 to 24 percent  in  1972-1973,  while the
deposition attributable to H«SO. decreased from 80 to 73 percent,
                                          2~       +
     Lindberg et al.  (1979)  noted that  SO,   and  H  were by  far  the  dominant constituents of
precipitation at the  Walker  Branch Watershed, Tennessee.   Comparison  with the annual average
concentration of major elements in rain at the Walker Branch Watershed on an equivalent basis
indicated  that  H  constitutes  approximately 50  percent of  the cationic strength  and  trace
elements account  for only 0.2  percent.   Sulfate constituted approximately 65  percent  of the
anionic strength and on an equivalent basis was 3.5 times more concentrated than NQ3, the  next
                                            7-23

-------
Ul
u.
_l
        I   I  I  III  I  I   I  I
       1920 1930  1940  19SO  1960


                 YEAR
                               1970
       1920 1930  1940  1950  1960  1970
                  YEAR
                                        8

                                        z
                                        01
                                        o
                                        D>
1920  1930 1940  1<«0  1960  1970


          YEAR
                                           06

                                           05


                                           04
                                        iu  0.3
                                        t-

                                        K  02

                                        Z
                                           01


                                            0
  1968  1970  1972  1974

       YEAR
      Rgure 7-8. Trends in mean annual concentrations of sulfate, ammonia,
      and nitrate in precipitation. (A), (B), and (C) present long-term data for
      Ithaca, New York; (D) presents data for eight years averaged over eight
      sites in New York and one in Pennsylvania. One point in (A), for 1946-7,
      is believed to be an anomaly (see Likens, 1972, for discussion}.

      Source:  (A), (B), and (C) modified from Likens (1972); (D) modified from
      Likens (1976).
                                     7-24

-------
most abundant  anion.   The  incident  precipitation  for the 2-year  (1976-1977)  period was des-
cribed as  "a dilute  mineral  acid solution,"  primarily H,,S04,  at a pH  approximating 4.2 and
containing  relatively minor  amounts of  various  trace salts  (Lindberg et  a!.,  1979).   In
Florida,  Hendry  (1977)  and Hendry and Brezonik  (1980)  found that the relative proportions of
sulfate,  nitrate and  chloride ions in rainfall  at  Gainesville,  Florida, during 1976, were 69
percent,  23 percent and 8 percent, respectively.
                                        2-
     Based on most  reports, sulfate  (SO. ) appears to be the predominant anion in acidic pre-
cipitation in the  Eastern  United States.   In California, however, nitrate (N03) seems to pre-
dominate.   Liljestrand and Morgan  (1978)  reported  that  their  analyses of  acidic rainfall
collected from February 1976 to September 1977 in the Pasadena,  CA area showed that the volume-
weighted mean  pH was  4.0,  with  nitric  acid being  32  percent  more important as  a source of
                                                              +    +   +    2+       2+
acidity than sulfuric  acid.  The major cations  present were H  ,  NH.,  K , Ca   and Mg   while
                                       2-
the major  amons were  Cl  , NO,, and SO.  .  McColl  and  Bush (1978) also noted the strong in-
fluence of nitrate  on rain in the Berkeley, CA,  region.  However, they note that in bulk pre-
                                               2-
cipitation (wet  plus  dry  fall-out) sulfate (SO.  ) constituted 50 percent of the total anions.
     Nearly all  of  the nitrate in rainfall is formed in the atmosphere from  NO  .   Little is
                                                                                /\
derived from wind  erosion  of nitrate salts in soils.  Similarly,  nearly all of the sulfate in
rainfall  is  formed in  the atmosphere from  S0?  (Galloway,  1978).   Thus,  all  atmospherically
                                                                                          +
derived nitrate  and  sulfate can contribute to the acidification of precipitation, since H  is
associated stoichiometrically with the  formation  of each.   A  second  stoichiometric process
that affects the acidity  of rain is the reaction of nitric and  sulfuric acids with ammonia or
other alkaline substances (e.g., dust particles)  in the atmosphere to form neutral nitrate and
sulfate aerosols.  To the extent that such neutralization occurs,  the acidity of precipitation
will be  reduced (National   Research  Council,  1978).   However,  since much of  the  ammonium ion
reaching soil is converted  to nitrate, these neutral salts can still have an acidifying effect
on the soil.  (See Section 7.3.2.1)
     Seasonal fluctuations  in composition as well as pH of rainfall have been reported by many
workers.   In addition, the  pH and composition of rainfall fluctuate from event to event, with-
in an event, from locality  to locality, and from storm to storm.
                   2-       +
     In general  SO.   and  H  concentrations in precipitation  in the eastern United States are
higher in the  summer than   in the  winter.   Wolff et al. (1979) found  this  to be true for the
New York metropolitan area.  Hornbeck et al.  (1976) and Miller  et al.  (1978) both stated that
a summer maximum- for sulfate was associated with an increase in hydrogen ion concentration in
upstate New  York,  the Hubbard Brook Experimental  Forest in New Hampshire, and in portions of
Pennsylvania.   Pack  (1978),   using  data  (1977)  from  the four  original  MAP3S  (Multistate
Atmospheric  Power  Production Pollution  Study) precipitation chemistry  stations,  plotted the
                                            7-25

-------
     weighted monthly  sulfate ion  concentrations  shown  in  the figure by  Lindberg,  (Figure  7-9).
     Maximum sulfate  concentrations occurred  from June  through  August.   Lindberg et  al.  (1979),
     studying wetfall  deposition  of sulfate  in  the  Walker Branch  Watershed,  also noted  summer
                    2-       +
     maximum for SO*    and  H .   Using the same  MAP3S  data as did Pack, they plotted weighted mean
     concentrations of  sulfate  in  rain  collected from  November 1976  through November 1977.   The
     peak summer  concentrations  at Walker Branch  Watershed,  Tennessee, are lower than  all of the
     stations except remote  Whiteface  Mountain,  New York,  The  regional  nature  of the  wet deposi-
     tion of sulfate  is  apparent.   Reasons for  the  existence of the high summer maxima of sulfate
     for the Eastern United States are discussed in some detail  in Chapter 5, Section 5.5.1.
          Seasonal  variations of  nitrogen  compounds  and  of pH  in  precipitation  have been reported
     by  several  workers,  but no  simple  trends are  apparent  (see U.S.  Environmental  Protection
     Agency Air Quality  Criteria for Oxides of  Nitrogen, 1982).   Hoeft et  al.  (1972)  found  rela-
     tively constant levels of nitrate in rain and snow collected in Wisconsin throughout the  year,
     but deposition of ammonia and organic nitrogen was lowest in winter and highest in  spring, per-
     haps because of the thawing of frozen animal wastes.  Haines (1976) reported large  random vari-
     ations, but  relatively small seasonal variations,  for  the various forms of  nitrogen  in wet-
     only precipitation at  Sapelo  Island,  Georgia; nitrogen  concentrations  were lowest during the
     rainy  months  of July and September.  The highest nitrogen loadings occurred during July and
     were-asosciated with the lowest range in pH (4.2-4.8).   Hendry (1977)  and Hendry and Brezonik
     (1980) found  relatively  smooth  seasonal  trends in  ammonia  and nitrate concentrations in both
     wet-only and  bulk  collections  (wet- and dryfall)  at Gainesville, Florida, with lowest concen-
     trations in winter (Figure  7-10).   In addition,  the pH of the bulk  precipitation showed no
     seasonal trend.   Wet-only  collections, however,  showed  the lowest pH  value  (4.0)  during the
     spring and summer.  The historical  record  suggests*there  has  been an  increase in  the concen-
     tration of inorganic nitrogen in florida over the past 20 years.
          Scavenging by  rainfall  produces  large changes  in atmospheric contaminant concentrations
     during a given  rainfall  event.   The decline  in constituent levels is  usually rapid, at  least
     in localized  convective  showers;  low, steady-state  concentrations.are  usually  reached within
     the  first  half hour of a rain  event due  to cleansing  of the atmosphere  by rain (Brezonik,
     1975).   Major ions  [chloride  (CT )  and sulfate (SQ.~)],  inorganic forms of nitrogen [nitrate
         *•                    -t-
     (NO- )  and  ammonium (NH. )],  total  phosphorus and pH were measured in  rain collected  in 5-
     minute segments within  three  individual  rainstorms.  Initially,  rapid  decreases were observed
     for  nitrate and  ammonium and total  phosphorus.   Therf was  also  a decrease,in pH from 4.65 to
     4.4.  Steady state concentrations were reached in 10 minutes.   Two other storms sampled in the
     same manner showed similar but less  defined patterns (Hendry & Brezonik, 1980).
          Wolff et al. (1979) examined spatial, meteorological and seasonal  factors associated with
     the  pH of precipitation in the  New York metropolitan area.   Seventy-two events were studied
                                                 7-26

\

-------
Figure 7-9. Comparison of weighted mean monthly concentrations of
sulfate in incident precipitation collected in Walker Branch Water-
shed, Tenn. (WBW) and four MAP3S precipitation chemistry monitor-
ing stations in New York, Pennsylvania, and Virginia.

Source:  Lindberg eit al. (1979).
                                   7-27

-------
< s
p
gj 0>
o 5
o
    4.80


    4.60


    4.40


    4.20




    0.40


    0.30


    0.20


    0.10
                                       M
                                               M
                                               I	I
              A   S  O

             	1976	
                                       Wl  A

                                       -1977 —
                                              M
                            MONTH

 Rgure 7-10.  Seasonal variations in pH (A) and ammonium
 and nitrate concentrations (B) in wet-only precipitation at
 Galnaville, Florida. Values are monthly volume-weighted
 averages of levels in rain from individual storms.

 Source: Hendry (1977).
                             7-28

-------
from  1975  through  1977.    There was  some  site-to-site  variability  in  the  hydrogen  ion
concentration expressed  as  pH values among the eight sites they studied in the Manhattan area
(Table 7-2).  The  standard  deviation for individual  sites  ranged  between 0.20 and 0.37. They
also noted  that  the pH varied according to storm type (Table 7-3).  Storms with a continental
origin have  a  slightly lower pH than storms originating over the ocean.  The storms with tra-
jectories from  the south and southwest had the  lowest pH's,  while those from  the north and
east had the highest pH's (Wolff et al., 1979).
     The mean  pH of precipitation falling  on  the New York metropolitan  area  during a 2-year
(1975 to 1977)  study  was 4.28; however, a  pronounced seasonal  variation was observed (Figure
7-11).  The minimum pH at  all  sites  except Manhattan was recorded during July  to September,
while the maximum  occurred  during October to December.  The minimum pH in Manhattan, however,
was measured between January  and March and  then gradually increased  through the year.  The
lowest mean  pH  of  4.12 for the New  York  Metropolitan area occurred during the  summer months
(Wolff et al., 1979).   In general, the pH  of  rain is usually lower in the summer than in the
winter and  is associated with  the high summertime sulfate'concentrations.   In  addition, the
            »
lowest pH's were associated with  cold  fronts  and air mass type precipitation  events.   These
events occur more  frequently  during the  summer months.   The lower  pH's  also  occurred  on
westerly or southwesterly winds (Wolff et al.,  1979).
     Seasonal variations  in pH measured  at several  sites in  New York State 70  km (45 mi.)
apart demonstrated  a  significant  difference between seasons (winter had an average pH of 4.2;
summer, 3.9.)  but  no  significant difference  between  sites.   In  New  Hampshire,  however, six
summer storms sampled  at 4  sites less than 3 km (2 mi.) apart showed a significant difference
(3.8  to  4.2) indicating  considerable variation  in  pH may  occur  in the same  storm (Cogbill
1976).
     Stensland (1978,  1980) compared the  precipitation chemistry  for  1954  and 1977 at a site
in central  Illinois.   The pH for the 1954  samples had not been measured, but were calculated
from  the  data of  Larson and  Hettick (1956)  and  compared with those  measured  in 1977.  The
calculated  pH for  1954 was  6.05;  the pH for 1977 was 4.1.  The more basic pH in 1954, accord-
ing  to the  author, could  have  resulted   from  low  levels  of acidic  ions  (e.g.  sul fate  or
nitrate) or  from high  amounts of basic ions (evg. calcium and magnesium).  Stensland suggests
                                                   I I                   -iil--|r
that the higher  pH in 1954 was due to calcium (Ca  ) and magnesium (Mg  ) ions from the soil.
7.2.3.2  Geographic Extent,  of Acidic Precipitation—Acidic precipitation has been a reality in
New York State  for an undetermined period  of  time.   Data collected by the United States Geo-
logical Survey  (Harr  and Coffey,  1975) are presented in Figure 7-12.   The pH of precipitation
has  remained nearly at the same  average  level during the entire  ten-year  period; therefore,
since data  for  the years prior to 1965 are lacking,  it is difficult to determine when the pH
in precipitation first began to decrease (Harr and Coffey, 1975).
                                            7-29

-------
         TABLE 7-2.  MEAN pH VALUES IN THE NEW YORK METROPOLITAN
                              AREA (1975-1977)

Site
Cal dwell, N.J.
Piscataway, N.J.
Cranford, N.J.
Bronx, N.Y.
Manhattan, N.Y.
High Point, N.J.
Queens, N.Y.
Port Chester, N.Y.
All sites
Mean pH
4.32
4.25
4.34
4.31
4.29
4.25
4.63
4.60
4.28
SD
0.26
0.36
0.34
0.37
0.25
0.30
0.35
0.19
0.32
No. obsd
50
64
48
57
39
25
20
21
72
Range
3.35-5.60
3.57-5.50
3.44-5.95
3.42-5.75
3.80-5.50
3.74-4.90
3.98-5.28
4.00-5.10
3.50-5.16

Source;  Wolff et al. (1979.)
               TABLE 7-3.  STORM TYPE CLASSIFICATION


Type
Description of dominant storm
system
No.
obsd
Mean
PH

I

2

3
4
5
6
7
8
Closed low-pressure system which formed
over continental N. Amer.
Closed low-pressure system which formed in
Gulf of Mexico or over Atlantic Ocean
Closed low which passed to W or N of N.Y.C.
Closed low which passed to S or E of N.Y.C.
Cold front in absence of closed low
Air mass thunderstorm
Hurricane Belle
Unclassified
22

21

26
17
16
5
1
6
4.35

4.43

4.39
4.39
4.17
3.91
5.16
4.31

Source:  Wolff et al. (1979).
                                7-30

-------
    4.6
    4.5
    4.4
i.   43
    4.2
    4.1
    4.0
               JFM         AMJ        JAS         OND

               MONTHS OF THE YEAR (1975 THROUGH 1977)

Figure 7-11. Seasonal variation of precipitation pH in the New
York Metropolitan Area.
Source: Wolff et al. (1979).
                               7-31

-------
70

60

SO

40


30
  «
00


TO

60

50

«0
                                            ALBANY. NEW YORK
  001


  60

  50

  40

  30

  oo"




  60

  SO
                                      ALLEGHENY STATE PARK. NEW YORK
    <        I         I
    p i i i. i . i i ii 11 r i i i i .1.1.1.1 i il
                                        ATHENS. PENNSYLVANIA
  40
                                          CANTON NEW YORK
    it,
  00 1 . i i i t t i
            1965       1966       1967       1MB       %K9      1910      1971      1972
                                                                                     19J3
Figure 7-12. History of acidic precipitation at various sites In and adjacent to State of
New York.
Source: Harr and Coffey (1975).
                                               7-32

-------
70


60


SO


40

00*





70


60


SO


40

oo":
                                       HINCKLEV, NEW YORK
                                       MAYS POINT. NEW YORK
   70


   60


   50


   40

   OO-lim.Mm
   60


   SO


   40


   00
                                        MINEOLA, NEW YORK
                                       ROCK HILL NEW YORK
                                        UPTON Nf W YORK
   "
        1965
                 1966
                                   196S
                                             1969

                                            YEAR
                                                      1970
                                                                                  1973
Figure 7-12 (cont'd).  History of acidic precipatation at various sites in and adjacent to
State of New York.
Source: Harr and Coffey (1975).
                                              7-33

-------
     Reports  indicate that precipitation  is acidic  in parts of  the country  other  than the
northeastern  United  States  (See  Figure  7-13.).   Average pH  values around  4.5 have  been
reported  as  far  south  as northern  Florida  (Likens, 1976;  Hendry and Brezom'k,  1980),  from
Illinois  (Irving,  1978),  the  Denver  area  of  Colorado   (Lewis  and Grant,  1980),   the  San
Francisco Bay area of California  (McColl and Bush, 1978; Williams, 1978), Pasadena, California
(Liljestrand and  Morgan,  1978), the Puget Sound area of Washington (Larson et al., 1975), and
from eastern Canada  (Glass et  al.,  1979;  Dillon  et al., 1978).    Data from  the San Francisco
Bay  area  indicate that  precipitation has become more acidic in  that region since 1957-1958
(McColl and Bush, 1978).  The pH  decreased from 5.9 during  1957-1958 to 5.0 in 1974, and seems
to  be related  to an  increase in  the  NCs  concentration  (McColl  and Bush,  1978).   Another
report,  using  data  from  the  California  Air Resources Board  (Williams, 1978),  states  that
acidic  precipitation has  been reported from  such widespread areas  as  Pasadena,  Palo Alto,
Davis, and Lake Tahoe.
     Studies in  the  Great Smokey Mountain National Park (Herrmann and Baron, 1980) indicate a
downward  trend  in pH  has occurred  there  over the  past twenty years.   Over a  period of 20
years, there has been a drop  in pH from a range of 5.3 to 5.6 in 1955 to 4.3 in 1979.
     The absence of a continuous precipitation monitoring network throughout the United States
in the past  makes determination of  trends in  pH  extremely difficult and controversial.  This
shortcoming has  been  rectified recently through the establishment of the National Atmospheric
Deposition  Program.   Under  the program,  monitoring  stations collect precipitation  samples,
determine their pH and then send the samples to a Central Analytical Laboratory in Illinois to
be analyzed.  This  long-term network plans  to  have  75  to   100 collection sites throughout the
United States; 74 are already operational.
7.2.4  Acidic Deposition
     The previous sections of this chapter have discussed the formation, composition,  and geo-
graphic distribution  of  acidic precipitation.   Usually when  the  effects  of acidic deposition
are discussed,  emphasis  is placed on the  effects resulting from the scavenging of sulfur and
nitrogen compounds by precipitation.  Dry deposition of gaseous and particulate forms of these
compounds  also  occurs and  is  beginning to  receive more  emphasis in research  (Galloway and
Whelpdale,  1980; Schlesinger and  Hasey,  1980;  Stensland, 1980; Sehmel,  1980;  Chamberlain,
1980).   Gaseous  compounds  reach  the  surface  of the  earth by  turbulent  transfer,  whereas
particulate  sulfates  and  nitrates  reach the earth's surface by gravitational sedimentation,
turbulent  transfer,   and  impaction  (Galloway  and  Whelpdale, 1980;  Sehmel, 1980; Hicks  and
Wesely, 1980).   A comparison of the relative significance  of wet and dry deposition is diffi-
cult  (See  Chapter  6).    Dry deposition,  however,  is  always  removing  pollutants from  the
atmosphere,  while removal  by  wet  deposition  is  intermittent  (Sehmel,  1980).   Marenco  and
Fontan  (1974)  suggest  that   dry deposition  is  more   important than  wet  in  removing  air
pollutants.
                                            7-34

-------
-a
 i
                            5.0
                       *  Wltto«(X>«for«m(NMJP)
                       a
                       -O- dWontelMMiltiafncMwKieir
                       A  MlMMUUAMMflMltetaMr
                          Production Mftitfeo tuaxium)
                       A
                       0
                       e
                          IMxraNyglMnM
                           Figure 7-13. pH of rain sample as measured in the laboratory and used in combination with Jhe
                           reported amount of precipitation.
                          Source: Wisniewski and Keitz (1981).

-------
     Lindberg et al.  (1979)  have calculated the annual mass transfer rates of sulfates to the
forest floor  in  Tennessee (Figure 7-14).  Their calculations  for SO/~ suggest wet deposition
by incident precipitation to be 27 percent of the total annual flux compared with a total dry
precipitation of 13  percent.   The dry deposition and foliar absorption of S02, a very import-
ant component, is  missing from this calculation.  The  wet and dry deposition percentages are
indicative only  of the  relative magnitude of the two processes.  The percentages, however, do
point out that the effects of acidic deposition usually attributed to precipitation scavenging
alone are probably "a result of both wet and dry deposition.  At the present time the accuracy
with which dry deposition can be measured is still  under question.
     The studies  of McColl and  Bush  (1978),  Hendry and Brezonik  (1980),  and Schlesinger and
Hasey (1980)  also  point out that both  wet  and dry deposition  are important  when considering
                a.    9—        *-
the effects of H ,  SO, , and N03 ions on aquatic and terrestrial receptors.
     The effects of the dry deposition of SO, and particulate matter on vegetation and terres-
trial ecosystems is discussed in Chapter 8.   The processes of wet and dry deposition of sulfur
oxides are  discussed in  Chapter 6 of  this  document;  such processes for  nitrogen  oxides are
discussed in  Chapter 6  of Air  Quality  Criteria for Oxides of Nitrogen  (U.S.  E.P.A.,  1982).
7.3  EFFECTS OF ACIDIC DEPOSITION
     Acidic precipitation  has  been  implicated in the acidification  of  surface waters leading
to the degradation of aquatic ecosystems, in the erosion of stone buildings and monuments, and
as a potential  source of harm to forests and other terrestrial ecosystems.  The sections that
follow discuss these effects.
7.3.1  Aquatic Ecosystems
     Acidification of surface waters  is a major  problem  in  regions of  southern Scandinavia
(Oden, 1968;  Aimer et  al.,  1974; Gjessing  et al., 1976), Scotland (Wright et al. ,  1980a),
eastern Canada (Beamish  and  Harvey,  1972; Dillon et al., 1978), and the Eastern United States
in the Adirondack Region of New York State (Schofield,  1976 a»b,c,; Pfeiffer and Festa,  1980),
Maine (Davis  et  al.,  1978),  and northern Florida (Crisman et al., 1980).   Damage to fisheries
is the  most obvious  effect  of  acidification  on freshwater life.  The disappearance  of fish
populations from acidified fresh water lakes and streams was first noted in southern Norway in
the  1920's.   In 1959,  Dannevig proposed that  acidic  deposition  was  the probable  cause for
acidification and  thus  for the loss of fish populations (Leivestad et al., 1976).  Subsequent
studies have  attempted  to verify this postulate.  Declines in fish populations have been re-
lated to acidification  of surface waters in southern Norway (Jensen and Snekvik, 1972;  Wright
and Snekvik, 1978), southwestern Sweden (Aimer et al.,  1974), southwestern Scotland (Wright et
al., 1980a),  the Adirondack  Region of New York  State  (Schofield, 1976a,b,),  and the LaCloche
Mountain Region  in southern  Ontario  (Beamish and Harvey,  1972).   Acidification may also have
serious repercussions on other aquatic biota inhabiting these systems.   Changes in the acidity
and chemistry of freshwater  affect the communities of  organisms  living there.  Pertinent de-
tails of these effects are described in the following sections.

                                            7-36

-------
                   IN CLOUD
                PRECIPITATION
                 SCAVENGING
                     25%
IBELOW CLOUD   , '/.'< •
PRECIPITATION  *.':."'.'.'.
 SCAVENGING  <[::^^
     2%         T      TO GROUND
                    (DORMANT PERIOD)
                           2%
          »,,! '«. J|t I,
       INCIDENT'PRECIPITATION
                                         •   .   TOTAL DRY
                                        •..-:.•  DEPOSITION
                                                  13%
                                                         /   \
                                                     TO LEAFY
                                                     CANOPY
                                                       10V.
                          •, ' i ' 27%" 1*1
                           '•  jiWj
                                 j
                                                            TO BRANCHES
                                                          (DORMANT PERIOD)
                                                                 1%
                      INTERNAL   EXTERNAL
                               100%
               RELATIVE ANNUAL MASS TRANSFER  RATES
                   OF  SOj-S  TO THE FOREST FLOOR
Figure 7-14. Annual mass transfer rates of sulfate expressed as a percentage of the estimated
total annual flux of the element to the forest floor beneath a representative chestnut oak stand.
Source: Lindberg et al. (1979).
                                   7-37

-------
7.3.1.1   Acidification of lakes  and streams—Precipitation  enters lakes  directly as rain  or
snow or  indirectly as runoff  or seepage water from the surrounding watershed.   The relative
magnitude of the  influents from these two sources is dependent on the surface area and  volume
of the  lake,  and  the size  of the watershed  and  its  soil  volume and type.   In  general,  the
watershed plays a dominant role  in determining the composition of water entering the lake.   As
a result, the water will be strongly influenced by processes in the edaphic environment  of  the
watershed, such  as weathering,  ion exchange,  uptake  and  release of  ions by  plants,  carbon
dioxide production by vegetation, microbial respiration, and reduction and oxidation reactions
of sulfur and  nitrogen compounds (Seip, 1980).  Precipitation  as  a  direct source of water to
the lake  plays  a  relatively greater role when lake  areas  are large in comparison to the size
of the watershed.
     Acidification of  surface  waters  results when the sources of hydrogen ion exceed the abi-
lity of  an ecosystem  to neutralize the  hydrogen  ion.   In  general, the soils and crust  of  the
earth are composed principally of basic materials with large capacities to buffer acids.   How-
ever,  areas  where  bedrock  is particularly  resistant  to  weathering and  soils are  thin  and
poorly developed  have  much less neutralizing  ability.  This  inability  to neutralize hydrogen
ions does not  usually arise from a  limited  soil  or mineral buffering capacity.   Instead  low
cation exchange capacity  and  slow mineral  dissolution rates  in  relation to  the  relatively
short retention time  of water within the soil system  may  result in incomplete neutralization
of soil  waters  and acidification  of  surface waters  (Driscoll,  1980).   Characteristics  of
regions  sensitive to  surface water  acidification  are discussed in  more detail  in Section
7.4.1.
     Sources  of  hydrogen  ions  to the edaphic-aquatic system include, besides  acidic deposi-
tion, mechanisms  for  internal  generation of hydrogen  ion -  oxidation  reactions (e.g.,  pyrite
oxidation, nitrification), cation uptake by vegetation (e.g., uptake of NH4  or Ca  ), or gene-
ration of organic acids from incomplete organic litter decomposition (Figure 7-15).   The rela-
tive importance of  the hydrogen ion content in  acidic deposition to the overall  hydrogen  ion
budget of an  ecosystem has been discussed by many researchers (Rosenqvist, 1976; SNSF Project,
1977).
     The consensus  is  that changes  in internal hydrogen ion generation related to land  use or
other changes  (e.g.,  Drabl^s and Sevaldrud,  1980) can  not  consistently  account for the wide-
spread acidification  of  surface waters  occurring  in southern  Scandinavia,   the  Adirondack
Region of  New York,  the LaCloche Mountain  Area of Ontario, and  elsewhere.   Driscoll  (1980)
developed a hydrogen  ion budget for the Hubbard Brook Area in New Hampshire.   Based on these
calculations,  atmospheric hydrogen ion sources represent 48 percent of the total Hubbard Brook
ecosystem hydrogen ion sources.
     As noted  above,  fresh water ecosystems  sensitive  to  inputs  of  acids are generally found
in areas of poor soil  development and underlain by highly siliceous types of bedrock resistant
to dissolution  through  weathering (Likens et a!., 1979).   As a result, surface waters in such

                                            7-38

-------
     ALLOCHTHONOUS SOURCES OF HYDROGEN ION
            PRECIPITATION,
            DRY DEPOSITION,
         , r  DRAINAGE WATER
       • ECOSYSTEM BOUNDARY
        HYDROGEN ION
           SOURCES

        OXIDATION RXN
        CATION UPTAKE
           PYR1TE
          OXIDATION
         NH/ UPTAKE
                 HYDROGEN ION
                     SINKS

                 REDUCTION RXN
                 ANION UPTAKE
                     OXIDE
                  WEATHERING
            STREAM EXPORTS
               H*. HCOg, OH-LIGANDS,
               ORGANIC ANIONS
Figure 7-1 i.
cycle.
Schematic representation of the hydrogen ion
Source: Driscol! (1980).
                         7-39

-------
areas typically  contain  very low concentrations of  ions  derived from weathering.   The waters
are diluted with low levels of dissolved  salts  and  inorganic carbon, and low in acid neutra-
lizing capacity.  The  chemical  composition of acid lakes is summarized in Table 7-4 for lakes
in southern Norway (Gjessing, et a!., 1976), the west coast (Hornstrom et al., 1976) and west-
central  regions  of  Sweden  (Grahn,  1976),  the  LaCloche Mountains  of  southeastern  Ontario
(Beamish, 1976),  and  the vicinity of Sudbury, Ontario (Scheider et al., 1976), as well as for
lakes not yet affected by acidification but in regions of similar geological  substrata in west-
central Norway (Gjessing et al., 1976) and the experimental lakes area of northwestern Ontario
(Armstrong and  Schindler,  1971).   Basic cation concentrations  (Ca,  Mg,  Na,  K) are low (e.g.,
calcium  levels  of 18-450 jjeq/liter or  0.4 -  9 mg/liter) relative to  world-wide  averages (15
mg/liter calcium,  Livingstone,  1963).   Bicarbonate  is the  predominant anion  in  most fresh-
waters (Stumm and Morgan, 1970).  However, in acid lakes in regions affected by acidic deposi-
tion, sulfate replaces  bicarbonate as the dominant anion (Wright and Gjessing, 1976; Beamish,
1976).  With a  decreasing pH level in  acid  lakes,  the importance of  the  hydrogen ion to the
total cation content increases.
     Surveys to determine the extent of effects of acidic deposition on the chemistry of lakes
have been conducted  in Norway (Wright and Snekvik,  1978;  Wright and Henriksen, 1978), Sweden
(Aimer et  al.,  1974; Dickson,  1975),  Scotland (Wright et al.,  1980a),  the  LaCloche Mountain
area  of  Ontario  (Beamish  and  Harvey,  1972),  the  Muskoka-Haliburton area  of south-central
Ontario  (Dillon  et  al.,  1978),  and  the Adirondack  Region  of New  York State  (Schofield,
1976a,b), Maine  (Davis  et al.,  1978), and Pennsylvania  (Arnold et al., 1980).  In regions of
similar geological substrata not  receiving acidic deposition,  lake  pH levels average 5.6-6.7
(Armstrong and  Schindler,  1971).   Of 155  lakes  systematically  surveyed in southern Norway in
October 1974, over 70 percent had pH  levels  below  6.0,  56 percent  below  5.5,  and 24 percent
below  5.0  (Wright and  Henriksen,  1978).  Of  700 lakes  in  the SeJrlandet Region  of southern
Norway surveyed in 1974 to 1975 (May-November), 65 percent had pH levels below 5.0 (Wright and
Snekvik,  1978).   On  the  west coast of  Sweden,  of 321 lakes investigated during 1968-1970, 93
percent had a pH level 5.5  or  lower.   Fifty-three  percent had pH levels  between  4.0 and 4.5
(Dickson, 1975).  In  the LaCloche Mountain Region of Ontario, 47 percent of 150 lakes sampled
in 1971  had pH  levels less  than  5.5, and  22 percent had  pH  levels below  4.5  (Beamish and
Harvey, 1972).   In the Adirondacks, 52 percent  of the high elevation (> 610  m)  lakes had pH
values below  5.0 (Schofield, 1976a,b).    In each of  these  studies,  the pH level  of an indi-
vidual lake could be related to, in most cases, the intensity of the acidic deposition and the
geologic  environment  of the  watershed.   Atmospheric  contributions  of sea  salts are  also
important in coastal  regions.
     Several methods  have  been  developed to assess  the  degree of acidification in a lake and
relate it to  inputs  of hydrogen ion or sulfate.  Henriksen (1979) utilized alkalinity-calcium
and pH-calcium relationships in lakes to estimate the degree of acidification experienced by a
surface water.   This technique  is based  on  the premise  that when carbonic  acid weathering

                                            7-40

-------
                           TABLE 7-4   CHEMICAL COMPOSITION (MEAN 1 STANDARD DEVIATION) OF ACID LAKES (pH <5) IN REGIONS RECEIVING HIGHLY
                            ACIDIC PRECIPITATION (pH <4 5), AND OF SOFT-WATER LAKES IN AREAS NOT SUBJECT TO HIGHLY ACIDIC PRECIPITATION
                                                                             (pH M 8)
Region
I LAKES IM MID ARIAS
Scandinavia
Southernmost
Norway
Westcoast
Sweden
West-central
Sweden
North America
La Cloche Mtns,
Ontario
Sudbury,
Ontario
II. LAKES IN UNAFFECTED
Scandinavia
West-central
Norway
North America
Experimental
Lakes Area,
Ontario
No of
lakes


Measured. 26
Less s «*"
Measured 12
Less s **;
Measured! 6
Less s w*

Measured- *
Less % \f
Measured 4
Less s w*.
AREAS

Measured 13
Less s w*

Measured 40
Less s w*'

Specific
conductanceTt H (pH)


27*10 18*11 (4 76)
18
72"« 43" (4 37**)
43««
47*23 22*15 (4 66)
22

38±8 20*9 (4 7}
20
120*40 36*5 (4 5)
36


1313 612 (5 2)
6

19 02-2 (5 6-6 7)
0 2-2


Na


70+40
9
330
-50
1651120
20

26*4
9
100*30
50


50*20
9

40
4


K


5*3
4
20
13
1518
12

10i3
10
40110
40


»1
3

10
10


Ca


56*35
50


-
75*10
70

150*25
150
450*180
450


1849
16

80
80


Mg


41*16
25
*M»n

-
80+40
50

75*8
65
310*120
300


16±5
7

75
65

Meq/1
HC03


11±26
11
0
0
-
-

0
0
842
8


13*8
13

60
60


HC1


71± 45
0
440
0
170*90
0

2216
0
50*20t
0


46±21
0

40
0


so«


100*33
92
ZOO
155
200*70
180

290140
2iO
8004290
800


33*8
30

60
55


N03


442
4
8
8
19*4
19

-
-
-
"


5*2
5

<1 5
<1 5


I cations


189
106
673
-
360
175

280
255
940
880


93
41

200
160


I anions


186
107
648
-
390
200

310
290
850
800


97
48

160
120

Reference


G jess ing
et al , 1976
BornstrSn
et al , 1976
Grahn
et al , 1976

Beaiii sh ,
1976
Arnstrong,
1971


Gjessmg
et a) , 1976

Armstrong,
1971

 *Less s w = Concontrations after subtracting the seawater contribution according to the procedure explained by
   Wright and Gjessino,  (1978)
"Data for 112 lakes
 treasured after past  liming of  the  lakes
ttyS/c«! at 20"C

-------
occurs  one equivalent of  alkalinity (acid neutralizing capacity) is  released  to  the aquatic
environment for  every equivalent of basic cation  (Ca,  Mg,  K, or Na) dissolved.  On the other
hand,  if  mineral  acid weathering is occurring,  for  example as a result of acidic deposition,
one equivalent of hydrogen ion is comsumed for every equivalent of cation solubilized.  There-
fore,  for a given  basic cation  level,  there is  less  aqueous acid  neutralizing  capacity in
lakes  in  systems experiencing  strong acid weathering  than  in systems  experiencing carbonic
acid weathering.  When comparing alkalinity p,lots from two watersheds, one experiencing strong
acid  contributions  and  the  other undergoing largely carbonic acid weathering  (assuming both
watersheds  have  similar edaphic  environments),  the difference in alkalinity between the two
plots for a given calcium level (the dominant basic cation) should be indicative of the amount
of  strong acid the watershed receives and the  degree of acidification of  the  surface water.
For waters  with pH  levels below 5.6,  alkalinity is approximately  numerically equal  to the
hydrogen  ion concentration  with its sign changed.  Therefore, pH level can be substituted for
alkalinity, and  pH-calcium  plots developed (Figure 7-16).   In the figure most of the lakes in
northern  and  northwestern  Norway  fall  below an empirically shown  curve, whereas  lakes  in
southernmost and southeastern  Norway lie above this  curve.   Data of  this type  for Norway in-
dicate that significant  acidification of lakes has  occurred  in areas receiving precipitation
with  volume-weighted average  concentrations of  H   above  20-25  pec/liter (pH 4.7-4,6)  and
sulfate concentrations above 1 mg/liter (20 peq/liter) (Henriksen, 1979).
     Henriksen (1979) also utilized the concentration of excess sulfate in lake  water (sulfate
in excess of that of marine origin) to estimate acidification.  This suggests that bicarbonate
anions lost in  acidified lakes have been  replaced  by an equivalent amount of sulfate.  Aimer
et  al.  (1978)  plot  pH  levels  in  Swedish lakes  as  a function of excess  sulfur  load (excess
sulfur in  lake  water multiplied by the yearly runoff) (Figure 7-17).   Based on  this relation-
ship, they  estimate  that the most  sensitive  lakes  in Sweden may resist a  load of only about
       2                                                2
0.3 g/m   of sulfur in lake water each year.   At 1  g/m of sulfur,  the pH  level  of the lake
will probably decrease below 5.0.
     Elevated metal  concentrations  (e.g.,  aluminum,  zinc,  manganese,  and/or iron)  in surface
waters are often associated with acidification (Schofield and Trojnar, 1980; Hutchinson et al.,
1978; Wright and Gjessing, 1976; Beamish, 1976).   Mobility of all  these metals is increased at
low pH values  (Stumm and Morgan, 1970).   For example, an inverse  correlation between aluminum
concentration and pH  level  has been identified for lakes in the Adirondack Region of New York
State, southern  Norway,  the west coast of Sweden, and Scotland (Wright,  1980b)  (Figure 7-18).
Aluminum  appears  to be  the primary element  mobilized by strong  acid  inputs in precipitation
and dry deposition (Cronan, 1978).                          .               .
     Aluminum is the third most abundant element by weight in the  earth's crust  (Foster, 1971).
In general, aluminum is  extremely insoluble and retained within the edaphic environment.  How-
ever, with  increased hydrogen  ion  inputs (via  acidic  deposition or  other sources) into the
                                            7-42

-------
 I
 a
                                                                     200km
     6.0
     6.5
     7.0
     7.5
             \
                \
<^8o°0§  O  °
  °**  s°°°
 ° SIR o  %o
 *o ^^ y
teo  o  o, -  *
                     s»».
       o  "aJ-*   "o   O
       ooocfgo o
                 o  „
250
                                             tCa]
Figure 7-16.  pH and calcium concentrations in lakes in northern and northwestern Norway
sampled as part of the regional survey of 1975, in lakes in northwestern Norway sampled in
1977  (o)  and in  lakes in southernmost and  southeastern Norway sampled in 1974  (•).
Southern Norway receives highly acid precipitation (pH 4.24.5) and a large number of lakes
have lost their fish populations due to high acidity. Inset shows areas in which these lakes
are located. Areas south of isoline receive precipitation more acid than pH 4.6.

Source:  Henriksen (1979).
                                            7-43

-------
             D
                                             O
                                  I
I
                                                                         CURVE 2
        0123

                               EXCESS S IN LAKE WATER, g/m2/year

Figure 7-17.  The pH value and sulfur loads In lake waters with extremely sensitive surroundings
(curve 1} and  with  slightly  less sensitive surroundings (curve 2). (Load = concentration of
"excess" sulfur multiplied by the yearly runoff.)

Source; Aimer et al. (1978).
                                         7-44

-------
1000



j£ 100


10
L


1000



1 100
<

10
4

— III-
SOUTH NORWAY 1974
— 154 LAKES —
*
* «
-•;•-:-••'•• .-
*,.* * * *
— % '•..** ... —
• *§ « * *
.*.>l **.-. « ~ .

*
* •
I i i
I 5 6 7 8
pH

\— I ! i —
WEST COAST SWEDEN
— * 37 LAKES 	
*
*
» *
- :'. .* .
*
* *
« *>* « * *
* *
I * i ** * I
t 6 7 f
pH
1000



OJ
5 100


10
t


1000



5 100


10
i i

— I 1 1 -
SCOTLAND
	 72 LAKES 	
*
•
	 **»*.. » * _
*'•."• . -' % •
9 •
~~ * » •* * . . "~~
* .
. * . ' * %*.
"" • * »» • '""='
* «
* *
1 I i '
15678
pH
* ..1 1 1
_ * * ADIRONDACKSUSA _
• 134 LAKES
	 ».*«.* * 	
•Vw ;.*
..£* .*.,. ... . . .
• * a. " * »
(••WH ** * ^BBWH
** » '* .*** "
* * » * *
* *
* *
*
I I . I
I 5 6 7 8
pH
Figure 7-18,   Total dissolved  Al as a function of pH level in lakes in acidified areas in Europe and
North America.

Source: Wright et al. (I980b).
                                              7-45

-------
edaphic environment,  aluminum  is rapidly mobilized.  Cronan and Schofield (1979) suggest that
input of  strong  acids may inhibit the historical trend of aluminum accumulation in the B soil
horizon.  Consequently,  aluminum tends  to  be transported  through the soil  profile  and into
streams  and lakes.   Evidence  from field data (Schofield  and  Trojnar,  1980)  and laboratory
experiments  (Driscoll  et al.,  1979;  Muniz and  Leivestad,  1980)  suggest  that  these  elevated
aluminum  levels  may  be toxic to fish.   Concentration  of aluminum may be as important or more
important than pH levels as a factor leading to declining fish populations in acidified lakes.
Aluminum toxicity to aquatic biota other than fish has not been assessed.
     Surface water chemistry,  particularly  in streams and rivers, may be highly variable with
time.  Since many  of the neutralization reactions  in  soils are kinetically slow, the quality
of the leachate from the edaphic system into the aquatic system varies with the retention time
of water  in the  soil  (Johnson  et  al.,  1969).   The  longer the contact period  of water with
lower soil  strata, the greater the neutralization of acid contribution from precipitation and
dry  deposition.  Therefore,  during  periods  of heavy rainfall or snowmelt and rapid water dis-
charge, pH levels in receiving waters may be relatively depressed.
     Many of the regions currently affected by acidification experience freezing temperatures
during  the  winter  and  accumulation  of a  snowpack.   In  the  Adirondack  Region of  New York
approximately  55  percent  of  the  annual   precipitation  occurs  during  the  winter  months
(Schofield,  1976b).  Much of the acid load deposited in winter accumulates in the snowpack and
may  be  released  during  a relatively  short time  period during snowmelt  in the  spring.   In
addition, on melting, 50  to  80  percent of  the pollutant  load  (including hydrogen  ion  and
sulfate) may be  released in the first 30 percent of the meltwater (Johannessen and Henriksen,
1978).  As  a result,  melting  of the snowpack and  ice  cover can result in  a  large influx of
acidic pollutants  into lakes and streams (Figure  7-19)  (Gjessing  et al., 1976; Schofield and
Trojnar,  1980;   Hultberg,  1976).   The  rapid  flux  of this  meltwater  through  the  edaphic
environment, and its interaction with only upper  soil  horizons,  limits  neutralization of the
acid  content.  As  a  result,  surface waters only moderately acidic during most of the year may
experience extreme drops in pH level during the spring thaw.  Basic cation concentrations (Ca,
Mg,  Na, K)  may  also  be lower during this time period (Johannessen et al., 1980).  Similar but
usually  less drastic  pH drops  in  surface  waters  (particularly  streams)  may  occur during
extended  periods of heavy  rainfall  (Driscoll,  1980).   These  short term  changes  in  water
chemistry may have significant impacts on aquatic biota, especially if they occur at sensitive
times in the life cycle (e.g., during spawning or early stages of development).
7.3.1.2  Effects on  decomposition.  The  processing of dead organic matter  (detritus)  plays a
central role  in the energetics  of lake  and stream  ecosystems (Wetzel,   1975).   The organic
matter may have been generated either internally (autochthonous) via photosynthesis within the
aquatic ecosystem  or  produced outside  the  lake or stream  (allochthonous)  and  later  exported
to the  aquatic  system.  Detritus is an important food source for bacteria, fungi, some proto-
zoa, and other animals.  These organisms, through the utilization of detritus, release energy,

                                            7-46

-------
                                            F

                                         1976/77
M
Figure 7-19,  pH levels in Little Moose Lake, Adirondack region of New York State, at a depth of 3
meters and at the lake outlet.

Source:  Adapted from Schofield and Trojnar (1980).
                                                7-47

-------
minerals and  other  compounds stored in the organic matter back into the environment.   Initial
processing of coarse participate detritus is often accomplished by benthic invertebrate fauna.
Among  other  things, the  particles  are physically broken down  into  smaller  units,  increasing
their  surface area.   Biochemical  transformations of particulate and  dissolved  organic matter
occur  via  microbial metabolism  and are  fundamental  to the dynamics of  nutrient  cycl-ing and
energy flux within the aquatic ecosystem.
     In general, the growth and reproduction of microorganisms is greatly affected by hydrogen
ion concentration (Rheinheimer,  1971).   Many bacteria can grow only within  the pH range of 4
to  9;  however,  the  optimum for most aquatic bacteria  is  between  pH 6.5 and 8.5.   There are
more acidophilic fungi than bacteria; consequently, in acid waters and sediments the proportion
of  fungi  in the  microflora is greater  than  in waters  or sediments with neutral  or slightly
alkaline pH  levels.  Most  aquatic  fungi require free oxygen for  growth  (Rheinheimer, 1971).
     Numerous studies have  indicated that acidification of  surface  waters  results in a shift
in microbial species and a reduction in microbial activity and decomposition rates.   It should
be  noted,  however,  that microorganisms in general are highly adaptive.   With sufficient time,
a given species may adapt to acid conditions or an acid-tolerant species may invade and colo-
nize  acidified  surface  waters.   Therefore,  some  caution  is  necessary  in  interpreting
short-term  experiments   on  the   effects   of   acidification   on  microbial   activity  and
decomposition.  On the other hand, increased accumulations of dead organic matter (as a result
of decreased decomposition rates) are commonly noted in acidic lakes and streams.
     Abnormal  accumulations  of  coarse  organic matter have been observed on  the bottoms of six
Swedish lakes.  The pH levels in these lakes in July 1973 were approximately 4.4 to 5.4.  Over
the last three  to four  decades, pH levels appear to have decreased 1.4  to 1.7 pH units (Grahn
et  a!.,  1974),   In  both Sweden  and Canada,  acidified  lakes have been treated with alkaline
substances  to raise  pH levels.  One  result of  this  treatment has  been an acceleration  of
organic decomposition processes and elimination of excess accumulations  of detritus (Andersson
et  al., 1978; Scheider  et al., 1975).   Litterbags containing coarse particulate detrital mat-
ter have been used to monitor decomposition rates in acidified lakes and streams.  In general,
the rates  of  weight loss were reduced in acidic waters when compared with more neutral waters
(Leivestad et  al.,  1976; Traaen,  1980;  Petersen,  1980).   Traaen (1980) found that  after  12
months of  incubation  dried  birch leaves or aspen sticks showed a weight loss of 50-80 percent
in waters with pH levels 6 to 7 as compared to only a 30-50 percent weight loss in waters with
pH  4 to 5.   Petersen (1980) likewise  found  reduced weight loss of leaf packs incubated in an
acidic stream when  compared to leaf packs in either a stream not affected by acidification or
a  stream  neutralized  with addition   of lime.   Petersen,   however,   found no  evidence  of
differences  in  microbial  respiration   between  the  streams.   The  acidic stream  did  show a
reduction  in  the  invertebrate  functional group that specializes in processing large particles
(shredders).   Gahnstrom  et  al.  (1980)  found no  significant  differences  in  oxygen consumption
by sediments from acidified and non-acidified lakes.  Rates of glucose decomposition were also

                                            7-48

-------
studied  in  lake  sediment-water  systems  adapted  to  pH  values  from  3  to  9.    Glucose
transformation  increased  at  pH levels  above 6.   Lime  treatment of  acidic Lake Hogsjon  in
Sweden also increased  rates  of glucose processing.   However in a humic lake, the maximum rate
of glucose transformation occurred at the in situ value pH 5 (Gahnstrom et a!,,  1980)
     Laboratory and  field  experiments involving decomposition rates have  fairly consistently
found decreasing microbial activity  with increasing acidity.   Traaen (1980) found that litter
decomposition at pH  level  5.2 was only  50  percent of that at pH  7.0  and at pH  3.5,  only  30
percent that at pH  7.0.   In addition, increasing  acidity  (pH 7.0 to 3.5) led to a shift from
bacterial  to  fungal  dominance.   Incubations  of profundal  lake  sediments  at  pH 4, 5,  and 6
indicated a significant reduction in community respiration with increasing acidity, as  well  as
a possible inhibition  of  nitrification and a lowering of sediment redox potentials.   Bick and
Drews  (1973)  studied  the  decomposition  of peptone -in the  laboratory.   With  decreasing  pH,
total  bacterial  cell   counts  and  numbers  of  species   of  ciliated  protozoans  decreased,
decomposition and nitrification  were reduced, and oxidation of ammonia ceased below pH 5.   At
pH 4 and lower, the number of fungi increased.
     Disruption of the detrital trophic structure and the resultant interference with nutrient
and energy cycling within the aquatic ecosystem may be one of the major consequences of acidi-
fication.   Investigations  into  the effects  of acidification  on  decomposition  apparently have
produced somewhat inconsistent results.  However, many of these apparent inconsistencies arise
only from a lack of complete understanding of the mechanisms relating acidity and rates of de-
composition.    It  is   fairly   clear  that  in  acidic  lakes  and  streams  unusually  large
accumulations of detritus  occur,  and these accumulations are related,  directly or indirectly,
to the  low pH  level.   The processing  of  organic matter has been  reduced.   In  addition, this
accumulation of organic debris  plus  the development of extensive mats  of filamentous algae on
lake bottoms  (discussed in  Section  7.3.1.3) may  effectively seal off  the mineral  sediments
from interactions with the overlying water.   As a result,  regeneration  of nutrient supplies to
the  water  column  is decreased  both  by reduced processing and mineralization of dead  organic
matter and by  limiting sediment-water interactions.  Primary productivity  within the  aquatic
system may  be substantially decreased as a result of this process  (Section 7.3.1.3).   These
ideas  have  been formulated  into the  hypothesis  of "self-accelerating  oligotrophication"  by
Grahn et al.  (1974).
7.3.1.3   Effect on primary producers and primary productivity.   Organisms  obtain their food
(energy) directly  or  indirectly from solar energy.  Sunlight, carbon dioxide,  and  water are
used by primary producers (phytoplankton, other algae, mosses, and macrophytes) in the  process
of photosynthesis  to  form  sugars  which in turn are  used  by the  plants or stored  as  starch.
The  stored energy may  be used by the plants or pass through the food chain or web.   Energy in
any  food chain  or  web passes through  several  trophic  levels.  Each link in the food chain is
termed  a  trophic  level.   The  major  trophic levels are  the primary producers,  herbivores,
carnivores,  and the  decomposers.    Energy  in  an   ecosystem  moves primarily  along two main
pathways:   the  grazing food chain (primary producers-herbivores-carnivores)  and the detrital
                                            7-49

-------
food  chain  (Smith, 1980;  Billings,  1978;  Odum,  1971).   Interactions between  these  two food
chains are,  however,  extensive.   Green plants convert solar  energy to organic matter and, as
such, are  the base  for both  food  chains.   The grazing food chain involves  primarily living
organic matter;  the detrital  food  chain,  dead  organic  matter.   Any  changes  as  a  result of
acidification  in the  green  plants  and  primary production within  the aquatic  ecosystem  may
therefore have a profound effect on all other organisms  in the aquatic food web.  As noted in
Section 7.3.1.2,  a portion  of the detrital  food chain  is  supported by  dead  organic matter
imported into the aquatic system from external sources.
     Extensive surveys  of acidic lakes in Norway and Sweden (Leivestad et a!.,  1976;  Aimer et
al., 1974) have observed changes in species composition and reduced diversity of phytoplankton
correlated with decreasing lake pH level (Figure 7-20).   Generally at normal pH values of 6 to
8,  lakes  in the  west coast  region  of Sweden  contain  30 to 80  species of phytoplankton  per
100-ml sample  in mid-August.   Lakes  with  pH  below 5 were  found to have  only about a dozen
species.   In some very acid lakes (pH<4), only three species were noted.   The greatest changes
in species composition occurred in the pH interval 5-6.   The most striking change was the dis-
appearance of many diatoms and blue-green algae.  The families Chlorophyceae (green algae)  and
Chrysophyceae (golden-brown algae) also had greatly reduced numbers of species in acidic lakes
(Figure 7-21).   Dinoflagellates  constituted the bulk of the phytoplankton biomass in the most
acidic lakes (Aimer et al., 1978).  Similar phenomena were observed in a regional survey of 55
lakes in  southern Norway  (Leivestad  et  al.,  1976) and  in  a study of nine  lakes in Ontario
(Stokes,  1980).   Changes  in  species composition and reduced diversity have also been noted in
communities  of attached  algae  (periphyton)  (Leivestad  et  al.,  1976;  Aimer et  al.,  1978).
Hougeotia, a green alga, often proliferates on substrates in acidic streams and lakes.
     Shifts in the types and numbers of species present may or may not affect the total levels
of primary productivity and  algal biomass  in acidic  lakes.   Species favored by acidic condi-
tions may  or may  not have comparable  photosynthetic efficiencies or desirability as  a prey
item for herbivores.   On the other hand, decreased  availability of nutrients in acidic water
as a  result  of reduced rates of  decomposition  (Section  7.3.1.2) may decrease primary produc-
tivity regardless of  algal  species  involved.   In field surveys and experiments, relationships
between pH  level and  total  algal biomass  and/or productivity were not as consistent  as  the
relationship between pH and species diversity.
     Kwiatkowski   and  Roff (1976)  identified  a significant linear  relationship of decreasing
chlorophyll  a  concentrations  (indicative  of algal  biomass)  with  declining  pH level  in  six
lakes near Sudbury, Ontario,  with a pH range of 4.05 to 7.15.   In addition, primary productiv-
ity was reduced  in  the two most acid lakes (pH 4-4.6).   Stokes (1980) also reports a decrease
in total phytoplankton  biomass with decreasing pH  level  for  nine lakes in the same region of
Ontario.   Crisman et  al.  (1980) reported a linear decrease in functional chlorophyll  measure-
ments with  declining  pH  for 11  lakes  in  northern Florida,  pH range 4.5  to 6.9.    On  the
other hand, Aimer et  al.  (1978) note that  in 58 nutrient-poor lakes in the Swedish west coast

                                            7-50

-------
    80


    70


tn   60
01
o
£   50
K
ui
CO
S
D
Z
40


30


20


10
                  T    I     I      1     I     I     I     I     1     li     I     I     I
                                                        PHYTOPLANICTON SPECIES IN 60 LAKES

                                                        ON THE SWEDISH WEST COAST

                                                        AUGUST 1976
      pH   4,1   43  4.5  4.7   4.9   5.1   5.3  5.5   5.7   5.9   6.1   6.3  6.5   6.7   6.9   7.1
 NUMBER   1  104324331210331002035054123101  1
OF LAKES

     Figure 7-20,  Numbers of phytoplankton species in 60 lakes having different pH values on the Swedish
     West Coast, August 1976.


     Source: Adapted from Aimer et al. (1978).
                                                 7-51

-------
            pH 4.60-5.45
                                     pH 6.25-7.70
                                  BIOMASS
                                  SPECIES
       D1ATOMEAE


|     | CHLOROPHYCEAE


       CHRYSOPHYCEAE
                                               CYANOPHYCEAE


                                               PYRROPHYTA


                                               SEPTEMBER 1972
Figure 7-21.   Percentage distribution of phytoplankton species and their biomasses.
September 1972, west coast of Sweden.    Biomass = living weight per unit area.

Source:  Adapted from Aimer et al. (1978).
                                     7-52

-------
region, the largest mean phytoplankton biomass occurred in the most acid lakes (pH <4.5).   Van
and  Stokes  (1978)  concluded that  they have  no  evidence that  the phytoplankton biomass  in
Carlyle Lake, with  a  summer pH level  about 5.1, is below that observed in circumneutral  lakes
in  the same  region.    In  a continuing whole-lake acidification  project (Schindler et  a!,,
1980), a  lowering  of  the epilimnion pH level from 6.7-7.0 in 1976 to 5.7-5.9 in 1978 resulted
in no significant change in the chlorophyll concentration or primary production.   Both i_n  situ
and  experimental  acidification  have   resulted  in large  increases in  periphyton  populations
(Muller, 1980; Hendrey,  1976;  Hall  et al.  ,  1980).  Hendrey  (1976) and Muller (1980) observed
carbon uptake by periphyton incubated i_n  vitro.  They  found  that, although the total rate of
photosynthesis increased with decreasing pH level due  to  the larger bioraass at the  lower pH,
the photosynthesis per unit biomass decreased with pH.
     From the above  discussion  it is  obvious that not  only is  there no  clear  correlation
between pH  level  and algal  biomass or productivity,  but the effects  of  acidification  appear
inconsistent between  systems.  Again,  these apparent inconsistencies  probably  reflect  a  lack
of  knowledge  about  exact mechanisms relating acidification and  lake  metabolism,  and also the
complexity of these mechanisms  and interactions.   Changes in the  algal community biomass and
productivity  probably reflect the  balance between a number  of  potentially  opposing factors:
those  that tend  to decrease productivity  and biomass  versus  those that tend to increase  pro-
ductivity and/or biomass when acidity increases.  Factors working to decrease productivity and
biomass with  declining  pH  levels may  include:  (1) a  shift in pH level below that optimal for
algal growth; (2) decreased nutrient availability as  a result of decreased decomposition rates
and  a  sealing-off of  the  mineral  sediments  from the  lake water;  and  (3)  decreased nutrient
availability  as  a result  of changes   in aquatic  chemistry with acidification.   For example,
despite the fact that the optional pH range for growth of label 1 aria flocculosa is between 5.0
to 5.3 (Cholonky, 1968) or higher (Kallqvist et al.,  1975), this species dominated experiment-
ally acidified stream communities at  pH level  4  in  three out of  five  replicates  (Hendrey et
al., 1980a).   As noted in Section 7.3.1.1,  aluminum concentrations increase with decreasing pH
level  in  acidified  lakes  and  streams.   Aluminum is  also a  very effective precipitator  of
phosphorus,  particularly in the  pH range  5  to  6  (Dickson, 1978; Stumm and Morgan, 1970).  In
oligotrophic lakes, phosphorus is most commonly the limiting nutrient for primary productivity
(Wetzel, 1975; Schindler,  1975).   Therefore, chemical  interactions between aluminum  and phos-
phorus may result  in  a decreasing availability of phosphorus with decreasing pH level and, as
a result, decreased primary production.
     Factors working  to increase productivity and/or biomass with acidification  of  a lake or
stream may  include:   (1)  decreased loss   of algal  biomass to herbivores;  (2)  increased  lake
transparency; and  (3)  increased  nutrient  availability resulting  from  nutrient  enrichment of
precipitation.   Decreased  population  of  invertebrates  (as  discussed  in  Section  7.3.1.4),
particularly  herbivorous  invertebrates, may decrease  grazing pressure  on algae and  result
in  unusual  accumulations  of biomass.  Hendrey  (1976)  and  Hall  et  al.  (1980)  include  this

                                            7-53

-------
mechanism as one hypothesis to explain their observation of increased biomass of periphyton at
pH level 4 despite a decreased production rate per unit biomass.
     Increases in  lake  transparency over time have been correlated with lake acidification in
Sweden  (Aimer  et  a!.,  1978)  and the  Adirondack Region  of New York  (Schofield,  1976c).   In
addition, after  the  second year of experimental lake acidification (pH 6.7-7.0 to 5.7-5.9) in
northwestern Ontario  (Schindler et a!.,  1980),  lake transparency  increased  by  1-2 m.   These
increases in  transparency have  not been correlated with decreases  in phytoplankton biomass.
Two mechanisms have  been proposed.   Aluminum acts as  a very efficient precipitator for humic
substances.   Dickson  (1978) found  that humic  substances  are readily precipitated  in  the pH
range 4,0 to 5,0.   Dickson (1978) and  Aimer  et al.  (1978) suggest that increases in aluminum
levels with lake acidification  (Section 7.3.1.1) have  resulted  in  increased precipitation of
humics from the  water column and therefore increased  lake  transparency.   Aimer at al.  (1978)
provide data for one  lake on the west coast of Sweden.  The pH level declined from above 6 to
about 4.5  between 1940  and 1975.  The secchi  disc  reading (Depth at  which 9-inch disc is
visible to  the  naked  eye when  lowered  into  the water) increased  from about 3m  to about 10m
over the  same  period.   Organic matter  in  the water (as estimated  by  KMnO*  demand) decreased
from 24 to 8 mg/liter from 1958 to 1973.  Schindler et al. (1980),  on the other hand, found no
change in levels of  dissolved organic  carbon with  acidification.   Instead,  changes in  hydro-
lysis of organic matter with declining  pH  level may affect the light absorbancy characteris-
tics of the molecules.   Levels  of particulate organic carbon, and  changes with pH level, were
not reported by Schindler et al. (1980).
     Acidification of precipitation  (and dry  deposition) has been  accompanied by increases in
levels of sulfate  and nitrate.   Both of these  are  nutrients required by plants.   However, as
noted above, the primary nutrient limiting primary productivity in most oligotrophic lakes is
phosphorus.   Aimer et  al.  (1978) report that atmospheric  deposition rates of phosphorus have
also  increased  in recent years.   The  world-wide  extent  of  the  correlation between  acidic
deposition and increased atmospheric phosphorus  loading, however, is not known.   It is expect-
ed that changes in atmospheric phosphorus loading would be much more localized than changes in
acidic deposition.  It  is possible  that in some areas increased atmospheric loading of phos-
phorus  has  occurred  in  recent  years  coincidentally  with  increased  acidic  deposition.
Increased phosphorus  nutrient loading  into lakes may  then  increase primary  production  rates.
     The effect  of acidification on  primary productivity  and algal  biomass of  a particular
stream or lake system depends upon the balance of the above forces.   Differences in the  impor-
tance of  these  factors  between systems may  account for  inconsistencies  in the  response of
different aquatic systems  to acidic  deposition.   Acidification does,  however,  result  in  a
definite change  in the  nutrient and energy  flux of the aquatic system,  and this  change may
eventually limit the total system biomass and productivity.
     Acidification of lakes has  also been correlated with changes in the macrophyte community.
Documentation for  these changes  comes mainly from lakes in Sweden.   Grahn (1976) reported that

                                            7-54

-------
in five of six lakes studied during the last three to five decades, the macrophyte communities
dominated  by Lobe!ia  and Isoetes  have  regressed whereas  communities dominated  by  Sphagnum
mosses have  expanded.-  Acidity  levels in these  lakes  apparently have increased, as indicated
by a  drop of approximately  1,3 to  1.7  pH  units  since  the 1930's to 1940's.   In acid lakes
where  conditions  are suitable  the Sphagnum peat  moss may  cpver more than 50  percent of the
bottom above the  4-m depth,  and may also grow at much lower depths (Aimer et al., 1978).   The
Sphagnum  invasion  may start  at lake  pH  levels just  below 6 (Aimer  et  al.,  1978).   Similar
growths  of Sphagnum occur  in Norwegian  lakes  (Galloway, 1978).  Increases in  Sphagnum  as a
benthic  macrophyte  have been documented  from  one lake  in  the Adirondack Region  of, New  York
(Hendrey and Vertucci, 1980).
     Under acid conditions the Sphagnum moss appears to simply outgrow flowering plant aquatic
macrophytes.    In  laboratory  tests,  the growth and  productivity of  the rooted macrophyte
Lobelia was  reduced  by  75 percent at  a pH  of 4, compared with the control (pH 4.3-5.5).   The
period of flowering was  delayed  by ten  days at the  low pH (Laake,  1976).  At  low pH levels
(pH<5),  essentially  all the  available inorganic  carbon  is in the form  of carbon dioxide or
carbonic  acid  (Stumm and  Morgan,  1970).   As  a result,  conditions may be more favorable for
Sphagnum, an acidophile that is  not able to utilize the carbonate ion.
     Besides the  shift in macrophyte species,  the  invasion of Sphagnum  into  acid lakes may
have  four other  impacts   on  the  aquatic ecosystem.   Sphagnum  has  a very high ion-exchange
capacity, withdrawing basic cations such as Ca    from solution  and  releasing H (Anschiltz and
Gessner,  1954; Aimer et al.,  1978).   As  a  result, the presence of Sphagnum may intensify the
acidification of  the system and decrease the  availability  of basic  cations for other biota.
Second,  dense  growths of  Sphagnum form  a  biotype that  is  an unsuitable substratum  for  many
benthic  invertebrates  (Grahn,  1976).  Growths  of  Sphagnum in acidic  lakes  are  also often
associated  with  felts  of white  mosses  (benthic  filamentous  algae)  and accumulations  of
nondecomposed organic matter.   In  combination, these organisms  and  organic  matter may form a
very effective seal.  Interactions between the water column and the mineral sediments, and the
potential  for  recycling  of  nutrients from  the sediments  back  into  the water body,  may be
reduced (Grahn,  1976; Grahn, et al., 1974).   These soft bottoms may also be colonized by other
macrophytes.    In Sweden,  Aimer et al.   (1978) report  that  growths  of  Juncus, Sparagam'um,
ytricularia,  Nuphar, and/or  Nymphaea, in addition  to Sphagnum,  may be extensive  in acidic
lakes.   Thus primary production  by macrophytes in  lakes with  suitable  bottoms  may be  very
large.   Increased lake transparency  may  also  increase  benthic macrophyte and algal primary
productivity.
7.3.1.4   Effects  on  invertebrates—In regional surveys conducted  in  southern  Norway (Hendrey
and Wright,  1976), the west coast  of Sweden (Aimer et al., 1978), the  LaCloche Mountain Region
of Canada (Sprules,  1975), and near  Sudbury,  Ontario  (Roff and Kwiatkowski,  1977) numbers of
species  of  zooplankton  were strongly correlated  with  pH  level  (Figure 7-22).   Changes in
community  structure  were  most noticeable at pH  levels below 5.  Certain  species (e.g., of the

                                            7-55

-------
   25
   20
 CO
 UJ
 o
 UJ
 Q.

 2 15
 O
 m


 lio
 <
 ui
                                                            pH INTERVAL

                                                            NUMBER OF LAKES
Figure 7-22.  The number of species of crustacean zooplankton observed in 57
lakes during a synoptic survey of lakes in southern Norway.
Source: Hendryetal(1976).
                                      7-56

-------
genera  Bosmina,  Cyclops,  Diaptomus,'  as  well  as  rotiferans,   of  the  genera  Polyarthra,
Keratella, and  Kellicottia) apparently  have  a high  tolerance of acidic  conditions  and were
commonly  found  in the  pH interval 4.4  to 7.9.   Others,  such as cladocerans  of  the Daphnia
genus, apparently are more sensitive and were only rarely found at pH <6 (Aimer et al., 1978).
     Similar studies of  the relationship between pH level and biomass or productivity of zoo-
plankton are not  available.   Proposed mechanisms for interactions between lake acidification
and zooplankton populations are therefore largely hypothetical.
     The  species, population  size,  and  productivity  of  zooplankton  are  affected   both  by
changes  in  the quality  and quantity of  the  food supply and  shifts  in  predator populations.
Changes  in zooplankton  species  and production in response to changes in fish populations have
been clearly demonstrated (Brooks  and Dodson, 1965; Walters and Vincent, 1973; Dodson, 1974).
Elimination  of fish  predators often  results  in dominance of  the  zooplankton community  by
large-bodied species.   Absence  of  invertebrate predators (e.g., large-bodied carnivorous zoo-
plankton) as a  result  of fish predation  or  other reasons often results in  the prevalence  of
                                                       >
small-bodied species (Lynch,  1979).   Surveys  of  acidic  lake  waters   often  have shown  the
dominance of small-bodied herbivores in the zooplankton community (Hendrey et al, 1980a).  Fish
also often are  absent at these pH levels (Section 7.3.1.5).   Different zooplankton species may
have different  physiological  tolerances  to depressed pH levels (e.g., Potts and Fryer, 1979).
Food  supplies,  feeding  habits,   and  grazing  of  zooplankton  may  also  be  altered  with
acidification  as  a  consequence  of  changes  in  phytoplankton   species  composition  and/or
decreases in biomass or productivity of phytoplankton.   Zooplankton  also rely on bacteria and
detrital organic  matter  for part  of their food  supply.   Thus  an inhibition of the microbiota
or  a  reduction  in  microbial  decomposition  (Section  7.3.1.2)  may  also  affect  zooplankton
populations.   These alternate mechanisms postulated to underlie changes  in community structure
and/or  production  of zooplankton  communities probably  play an important  role in  zooplankton
responses to acidification.
     Synoptic and  intensive  studies  of lakes and  streams also have  demonstrated that numbers
of  species  of  benthic  invertebrates  are  reduced along a gradient of decreasing  pH  level
(Sutcliffe and Carrick, 1973; Leivestad et al., 1976; Conroy et al.,  1976;  Aimer et al.,  1978;
Roff  and Kwiatkowski,  1977).  In  1500  freshwater localities  in Norway  studied from  1953-73,
snails were generally  present  only in lakes with pH levels  above 6 (Okland,  J. , 1980).  Like-
wise Gammarus  Lacustris,  a freshwater shrimp and an important element in the diet of  trout in
the Norwegian  lakes  where it occurs, was  not found at  pH levels below 6.0 (0kland, K.,  1980,
J. and  K. 0kland,  1980).  Experimental investigations have shown  that  adults of this species
cannot tolerate two  days of exposure to pH 5.0 (Leivestad et al.,  1976).  Eggs were reared at
six different  pH  levels (4.0 to 6.8).  At a  pH of 4.5, a majority of the embryos  died within
24  hours.   Thus,   the  short-term  acidification which  often occurs during the  spring melt  of
snow  could  eliminate this  species from  small  lakes (Leivestad et al.,  1976).   Fiance (1977)
concluded that  ephemeropterans (mayflies) were particularly  sensitive  to low  pH levels  and

                                            7-57

-------
their  populations  were reduced  in  headwater  streams  of the  Hubbard  Brook watershed  in  New
Hampshire.   In  laboratory studies, Bell  (1971),  Bell  and Nebecker (1969),  and  Raddum  (1978)
measured the tolerance of some stream macroinvertebrates to low pH levels.   Tolerance seems to
be  in  the order  caddisflies  >  (more  tolerant than)  stoneflies  > mayflies (Hendrey et a!.,
1980).
                                                                               2         2
     Leivestad  et  al.  (1976)  reported on decreased  standing crops (numbers/m   and g/m ) of
benthic invertebrates  in  two  lakes  with pH  levels  near 4.5 as compared to five lakes with pH
near 6.0.  Chironamids were  the dominant group in  all  lakes.   No fish were found in the acid
lakes.   Lack of predation by fish should favor  increases  in benthic biomass,  the opposite of
that observed.   On  the other hand,  Hendrey  et al.  (1980a),  based on data  from  eight Ontario
lakes  (pH 4.3  -  <  5.7),  reported no  reduction  in  abundance of benthos related  to  pH  level.
     Air-breathing  aquatic  insects  (e.g.,  backswimmers, water  boatmen,  and water  striders)
appear to  be very  tolerant  of acidic environments.   Population densities are often greater in
acidic lakes and in the most acid lakes than in circumneutral lakes.   Abundance of these large
invertebrates may be related to reduced fish predation (Hendrey et al.,  1980a).
     Hall  et al.  (1980)  experimentally acidified a stream  to  pH 4 and monitored reactions of
macroinvertebrate  populations.   Initially  following  acidification,   there  was  a 13-fold
increase  in downstream   drift  of  insect larvae.   Organisms  in  the  collector and scraper
functional groups were affected more than predators.   Benthic samples  from the acidified zone
of Morris  Brook contained 75 percent fewer individuals than those for  reference areas.   There
was also  a 37  percent reduction in insect emergence;  members of the collector group were most
affected.   Insects  seem to  be particularly sensitive at emergence (Bell, 1971).   Many species
of aquatic insects emerge early in the spring through cracks in the ice and snow cover.   These
early-emerging  insects therefore  are  exposed in many cases to the extremely acidic conditions
associated with srfowmelt (Hagen and Langeland, 1973).
     Low  pH  also  appeared  to  prevent  permanent colonization  by a  number  of  invertebrate
species,  primarily  herbivores, in  acidified reaches  of River Dudden,  England  (Sutcliffe  and
Carrick,  1973).  Ephemeroptera,  trichoptera,  Ancylus  (Gastropoda) and  Gammarus were absent in
these reaches.
     Damage  to  invertebrate  communities may  influence other components  of the  food  chain.
Observations that   herbivorous  invertebrates  are especially  reduced  in  acidic streams,  as
reported in Norris Brook and River Dudden, support the hypothesis (Hendrey, 1976; Hall et al.,
1980a)  that  changes in  invertebrate  populations  may be responsible for  increased periphytic
algal  accumulations in acidic  streams and  benthic  regions of acidic  lakes (Hendrey et al.,
1980).    Benthic invertebrates  also assist  with the  essential  function  of  processing dead
organic matter.   Petersen (1980) noted that decomposition of coarse particulate organic  matter
in  leaf packs was  lower in an acidic stream than in two streams with circumneutral pH levels.
The invertebrate community  also showed a reduction in  the  invertebrate functional group that
specializes  in  processing large  particles  (shredders).  In unstressed  aquatic  ecosystems, a
continuous  emergence of  different  insect species  is  available to  predators  from  spring to
                                            7-58

-------
autumn.  In  acid-stressed  lakes or streams, the  variety  and numbers of prey may  be  reduced.
Periods may  be  expected  to occur in which the amount of prey available to fish (or other pre-
dators) is diminished.
7.3.1.5  Effects on fish—Acidification of  surface  waters has had its most  obvious,  and per-
haps the most severe, impact on fish populations.   Increasing acidity has resulted not just in
changes in species  composition or decreases in biomass>"but,'' in many cases,  in total  elimina-
tion of populations  of  fish from a given  lake  or stream.  Extensive depletion of fish stocks
has occurred in large  regions of Norway,  Sweden,  and parts of eastern  North  America.   Both
commercial  and  sport fisheries have  been  affected  in these areas.   However,  precise assess-
                                                                             *.
ments  of  losses—in terms  of  population extinctions,  reductions  in yields, or  economic and
social  impacts—either  have not  been attempted  or  are  still  in  the process  of evaluation.
Potential  damage to fish populations inhabiting other acid-sensitive aquatic  ecosystems in New
England, the Appalachians,  and parts  of Southeastern,  North  central,  and Northwestern United
States have not yet been assessed (Galloway, 1978).
     Declines in fish populations have been related to acidification of surface waters in the
Adirondack Region  of New  York State  (Schofield 1976c),  southern Norway  (Jensen  and  Snekvik,
1972; Wright and Snekvik,  1978),  southwestern Sweden (Aimer et a!., 1974), the LaCloche Moun-
tain Region  in  southern  Ontario (Beamish and Harvey, 1972), and southwestern Scotland (Wright
et  al., 1980a).   Schofield (1976c) estimated that  in 1975 fish populations in 75 percent of
Adirondack lakes  at  high   elevation  (<610  m)  had been  adversely  affected  by  acidification.
Fifty-one percent of the lakes had pH  values  less  than 5, and 90  percent of these lakes were
devoid of fish life (Figure 7-23).  Comparable data for the period  1929 to 1937 indicated that
during that  time only  about 4 percent of these lakes had pH values below 5 and were devoid of
fish (Figure 7-24),  Therefore,  entire fish communities consisting of brook trout (Salvelinus
fontinalis),  lake  trout ( Salvelinus  namaycush),  white sucker  (Catostomus commersoni),  .brown
bullhead (Icaturus  nebulosus), and several cyprinid species were apparently eliminated over a
period of  40 years.   This  decrease in  fish populations was associated with a decline in lake
pH  level.   A survey of more  than  2000 lakes  in southern Norway, begun  in 1971,  found that
about  one  third of these  lakes had  lost  their fish population  (primarily brown  trout,  Sal mo
trutta  L.) since the 1940's (Wright and Snekvik,  1978).  Fish population status was inversely
related to lake  acidity (Leivestad et al., 1976).  Declines in salmon populations in southern
Norwegian  rivers were  reported as early as  the 1920's.   The catch of Atlantic salmon (Salmo
salar,  L.) in  seven acidified southern Norwegian rivers  is now virtually zero (Figure 7-25).
In  northern  and western rivers not  affected by acidification,  no distinct  downward  trend in
catch  has  occurred (Leivestad et al.,  1976;  Wright et al., 1976;  Jensen  and  Snekvik, 1972).
Similar changes  have been  observed in  Sweden  (Aimer et al., 1974) where it is estimated that
10,000  lakes have  been  acidified to  a  pH  less  than 6.0 and 5,000  below a pH of 5.0 (Dickson,
1975).  Populations  of  lake trout, lake herring (Coregonus artedii), white suckers, and other
species disappeared rapidly during  the 1960's from a group of remote  lakes  in  the  LaCloche
Mountain Region of Ontario  (Beamish et al., 1975).
                                            7-59

-------
                                            pH

Figure 7-23.  Frequency distribution of pH and fish population status in Adirondack Mountain
lakes greater than 610 meters elevation. Fish population status determined by survey gill netting
during the summer of 1975.

Source; Schofield (1976b).
                                              7-60

-------
       20
       10
   60
   LU

   <
   E
   UJ
   CQ
       10
                                       I         I
                                          1975
                                NO FISH PRESENT


                                FISH PRESENT
                                          1930s
             r-n
                              6

                             pH
Figure 7-24.   Frequency distribution of pH and fish  pop-
ulation status in 40 Adirondack lakes greater than 610 meters
elevation, surveyed during the period 1929-1937 and again in
1975.

Source: Schofield (1976 b).
                            7-61

-------
 STAVANGER
                           RIVER TOVDAL
                                                             RIVER DALALVEN
        300
        250
        200
        150
            1900
1920
1940
1960
1980
    en
    z
    o
         30
         20
         10
           1900
                         1920
             1940
             1960
             1980
Figure 7-25.  Norwegian salmon fishery statistics for 68 unacidified and 7 acidified

rivers.



Source: Adapted from Aimer et a!, (1978).
                                     7-62

-------
     It is difficult  to  determine at what pH level fish species disappear from lakes.   Disap-

pearance of the  fish  is  usually not due to massive fish kills, but is the result of a gradual

depletion of the  population  following reproductive failures (Leivestad  et  a!.,  1976).   Field

surveys in Scandinavia and eastern North America (Wright and Snekvik, 1978;  Aimer et a!., 1974;

Schofield, 1976c) suggest  that  many species do not occur  in lakes with pH  values  below 5.0.

     However,  large spatial  and temporal  fluctuations,in ,pH,  and  the possibility for "refuge

areas" from acidic  conditions  during critical  periods make  it extremely difficult to genera-

lize about effects  of acidification on fish populations based on  grab samples or annual mean

pH  levels.   The  pH levels  identified  in the  literature  as  critical  for  reproduction  of  a

species or correlated  with the absence of a species  in lake surveys are summarized  in Table

7-5.  Values range from pH 4.4 to over 6.0, and are highly species  dependent.


              TABLE 7-5.   pH LEVELS IDENTIFIED IN FIELD SURVEYS AS  CRITICAL TO
                           LONG-TERM SURVIVAL OF FISH POPULATIONS
       Family
        Species
Critical pH
     Reference
       Salmonidae     Brook trout (Salvelinus               5.0
                        fontinalis)
                      Lake trout (Salvelinus                5.1
                        namaycush)                        5.2-5.5

                      Brown trout (Salmo trutta)            5.0
                      Arctic char (Salvelinus alpinus)      5.2

       Percidae       Perch (Perca fluviatilis)           4.4-4.9
                      Yellow perch (Perca flavescens)     4.5-4.7
                      Walleye (Stigostedion yitreum)      5.5-6.0+

       Catostomidae   White sucker (Catostomus            4.7-5.2
                        commersoni)                         5.1

       Ictajuridae    Brown bullhead (Icaturus            4.7-5.2
                        nebulosus)                          5.0

       Cyprim'dae     Minnow (Phoxinus phoxinus)            5.5
                      Roach (Rutilus rutilus)5.5
                      Lake chub (Couesius plumbeus)       4. 5-4. 7
                      Creekchub (Semotilus atromaculatus)   5.0
                      Commonshiner (Notropis cornutas)      5.5
                      Goldenshiner (Notemigonus             4.9
                        crysoleucas)

       Centrarchidae  Smallmouth bass (Micropterus        5.5-6.0+
                        dolomieui)
                      Rock bass (Ambloplites rupestris)   4.7-5.2
       Esocidae
Pike (Esox jucius)
  4.4-4.9
                                                  Schofield,  1976c

                                                  Schofield,  1976c
                                                  Beamish,  1976

                                                  Aimer et  al.,  1978
                                                  Aimer et  al.,  1978

                                                  Aimer et  al.,  1978
                                                  Beamish,  1976
                                                  Beamish,  1976

                                                  Beamish,  1976
                                                  Schofield,  1976c

                                                  Beamish,  1976
                                                  Schofield,  1976c

                                                  Aimer et  al.,  1978
                                                  Aimer et  al.,  1978
                                                  Beamish,  1976
                                                  Schofield,  1976c
                                                  Schofield,  1976c
                                                  Schofield,  1976c
Beamish, 1976

Beamish, 1976

Aimer et al., 1978
                                            7-63

-------
     Recent field and laboratory studies (Schofield and Trojnar, 1980; Dickson, 1978; Driscoll
et  al.,  1979;  Baker and  Schofield,  1980;  Muniz  and  Leivestad,  1980) have  indicated  that
aluminum levels in acidic surface waters (Section 7.3.1.1, Figure 7-18) may be highly toxic to
fish (and perhaps other biota).  Schofield and Trojnar (1980) analyzed survival of brook trout
stocked into  53  Adirondack lakes as a function of 12 water quality parameters.  Levels of pH,
calcium, magnesium, and aluminum were significantly different between the two groups of lakes,
with and without  trout  survival.  However, after accounting  for the effects of aluminum con-
centrations on differences between the two groups of lakes, differences in calcium, magnesium,
and pH  levels were  no  longer  significant.   Aluminum, therefore,  appears  to be  the primary
chemical factor controlling survival of trout in these lakes.   Likewise, in laboratory experi-
ments with  natural Adirondack  waters  and synthetic  acidified aluminum  solutions,  levels of
aluminum,  and  not the pH level  per  se,  determined  survival  and growth  of  fry of brook trout
and white suckers (Baker and Schofield, 1980).   In addition,  speciation of aluminum had a sub-
stantial effect on  aluminum toxicity.   Complexation of aluminum with organic chelates elimin-
ated  aluminum toxicity  to fry (Baker  and  Schofield,  1980;  Driscoll  et  al.,  1979).   As  a
result, waters high  in  organic carbon, e.g., acidic bog lakes, may be less toxic to fish than
surface waters at similar pH levels but with lower levels of dissolved organic carbon.
     Inorganic aluminum levels, and not low pH levels, may therefore be a primary factor lead-
ing to declining fish populations in acidified lakes and streams.  However, many laboratory or
jjn situ field experiments have been conducted on the effects of pH on fish without taking into
account aluminum  or  other  metal concentrations in naturally acidic waters.   As a result,  many
of the conclusions  based on these experiments regarding  pH  levels critical for fish survival
are suspect.  Therefore these experiments will  not be reviewed here.
     Sensitivity  of  fish and  other biota to low pH  levels  has also been  shown  to  depend on
aqueous calcium levels (Wright and Snekvik, 1978; Trojnar, 1977a,b; 8ua and Snekvik,  1972).  In
southern Norway,  the mean calcium  level  in  lakes studied was  approximately  1.1  mg/liter, as
compared to  about 3  mg/liter in the LaCloche Mountain Region (Table 7-4)  or  2.1 mg/liter in
the Adirondack Region (Schofield, 1976b).  In Norwegian lakes, Wright and Snekvik (1978) iden-
tified  pH  and calcium  levels  as the  two most important chemical  parameters  related to  fish
status.
     Decreased recruitment  of  young fish has been cited  as  the primary factor leading to the
gradual extinction of fish populations (Leivestad et al,, 1976; Rosseland et al., 1980; Wright
and Snekvik,  1978).   Field observations (Jensen and  Snekvik,  1972; Beamish,  1974; Schofield,
1976a; Aimer  et  al.,  1974) indicate changes in population structure over time with acidifica-
tion.  Declining  fish populations  consist primarily of  older  and larger fish with a decrease
in  total  population  density.   Recruitment failure may  result from  inhibition  of adult fish
spawning and/or increased mortality of eggs and larvae.  Effects on spawning and decreased egg
deposition may be associated with disrupted spawning behavior and/or effects of acidification
on reproductive  physiology  in  maturing adults (Lockhart  and  Lutz, 1976).   Field observations

                                            7-64

-------
by Beamish et  al.  (1975) related reproductive failure in white suckers to an inability of fe-
males to  release their  eggs.   On  the  other hand, Amundsen and  Lunder  (1974)  observed total
mortality of naturally spawned trout eggs in an acid brook a few weeks after spawning.  A sum-
mary of Norwegian studies (Leivestad et al., 1976) concluded that egg and fry mortality is the
main cause of  fish  reproduction failure.  Spawning periods  and early life history stages for
many fish  species coincide with  periods of extreme acidity, particularly  during and immedi-
ately after snowmelt in the spring.
     In  some  lakes,  fish  population  decreases  are  associated  with a lack  of older  fish
(Rosseland et  al.,  1980).   In Lake Tveitvatn  on the Tovdal River  in southern Norway,  brown
trout mortality  apparently occurs  primarily after the  first  spawning.   Since 1976, no fish
past spawning age have been found, and population density has decreased steadily (Rosseland et
al., 1980).  Fish kills of adult salmon in rivers in southern Norway were recorded as early as
1911 (Leivestad et al., 1976).
     When  evaluating  the potential  effects  of  acidification on  fish or other  biotic popula-
tions,  it  is  very important to keep in  mind the highly diversified nature of aquatic systems
spatially, seasonally,  anc! year-to-year.  As a  result  of this diversity,  it is necessary to
evaluate each  system  independently  in  assessing  the  reaction  of  the population to acidifica-
tion.  Survival  of  a fish population may depend more on the availability of refuge areas from
acid conditions during spring melt or of one tributary predominantly fed by baseflow (fed from
the  bottom) and  supplying an adequate area  for  spawning,  than on mean annual  pH, calcium, or
inorganic aluminum levels.
7.3.1.6   Effects on vertebrates otherthan  fish.   Certain  species  of  amphibians may be the
vertebrate animals,  other than fish, most immediately and directly affected by acidic deposi-
tion (Rough and  Wilson,  1976).  Their vulnerability  is  due  to their reproductive habits.  In
temperate  regions,  most  species of frogs and toads, and approximately half of the terrestrial
salamanders,   lay eggs  in ponds.  Many of these  species breed  in  temporary  pools formed each
year by  accumulation of  rain and  melted  snow.   Approximately 50 percent  of  the species of
toads and  frogs  in the United  States  regularly  breed in ephemeral  pools;  about one-third of
the  salamander  species  that  have  aquatic   eggs  and  larvae and  terrestrial adults  breed in
temporary  pools.  Most  of these pools are small  and collect drainage from a limited area.  As
a result, the acidity of the eater in these pools is strongly influenced by the  pH of the pre-
cipitation that  fills  them.   Ephemeral  pools are  usually more  acidic than adjacent permanent
bodies of  water.  Rough  and  Wilson (1976)  report that  in  1975, in the vicinity of Ithaca,
N.Y., the average pH of 12 temporary ponds was 4.5 (range 3.5 to 7.0), while the average pK of
six  permanent  ponds  was  6.1 (range 5.5 to 7.0).   Amphibian eggs and larvae in temporary pools
are exposed to these acidic conditions.
     Rough and Wilson (1976)  and Rough (1976)  studied the  effect  of  pH  level  on embryonic
development of two common species of salamanders, the spotted salamander (Ambystgma maculatum)
and  the  Jefferson salamander  (A. Jeffersonianum).  In  laboratory experiments,  embryos of the

                                            7-65

-------
spotted salamander  tolerated  pH levels from 6 to 10 but had greatest hatching success at pH 7
to 9.  The  Jefferson salamander tolerated pH levels 4 to 8 and was most successful at 5 to 6.
Mortality of  embryos rose  abruptly beyond  the  tolerance  limits.   In a four-year  study  of a
large breeding  pond (pH  5.0-6.5),  938 adult  spotted salamanders  produced  486 metamorphosed
juveniles (0.52 juveniles/adult), while 686 adult Jefferson salamanders produced 2157 juvenile
(3.14 juveniles/adult).   Based on  these  findings,  Rough and Wilson  (1976)  predict that  con-
tinued acidic  deposition may  result in  substantial  shift in salamander  and other amphibian
populations.  Gosner and Black (1957) report that only acid-tolerant species of amphibians can
breed in the acid (pH 3.6 to 5.2) sphagnaceous bogs in the New Jersey Pine Barrens,
     Frog populations in Tranevatten, a lake near Gothenberg, Sweden, acidified by acidic pre-
cipitation,   have also  been  investigated  (Hagstrom,  1977;  Hendrey,  1978).   The lake  has  pH
levels ranging from 4.0 to 4.5.  All fish  have disappeared, and frogs belonging to the species
Rana temporaria  and Bufg  bufo are  being eliminated.   At  the  time of the  study (1977)  only
adult frogs eight  to  ten  years old  were  found.  Many egg masses  of Rana  temporaria  were
observed in 1974,  but few were found  in  1977,  and the few larvae (tadpoles) observed at that
time died.
     Frogs  and  salamanders are  important predators on  invertebrates, such  as mosquitoes and
other pest species, in pools,  puddles,  and lakes.  They also are themselves important prey for
higher trophic  levels in an  ecosystem.   In  some habitats, salamanders are  the most abundant
vertebrates.  In a  New  Hampshire  forest,  for example, salamanders  were  found to exceed birds
and mammals  in both numbers and biomass (Hanken et a!., 1980).
     The elimination of fish  and vegetation from  lakes  by acidification  may have an indirect
effect on a  variety of vertebrates:   species of fish-eating birds (e.g., the bald eagle, loon,
and  osprey),  fish-eating mammals (e.g.,  mink and  otter), and  dabbling  ducks which  feed on
aquatic vegetation.   In fact,  any animal  that depends on  aquatic organisms (plant or animal)
for a portion of its food may be affected.
     Increasing acidity  in  freshwater  habitats results in shifts in species, populations, and
communities.  Virtually  all trophic levels are affected.   Summaries  of  the  changes which are
likely to occur in aquatic biota with decreasing pH are listed in Tables 7-6 and 7-7.
7.3.2  Terrestrial  Ecosystems
     Determining the effects  of acidic precipitation on terrestrial ecosystems is not an easy
task.  In aquatic  ecosystems,  it has been  possible  to measure changes in  pH  that occur in
acidified waters  and then observe  the  response  of organisms living  in aquatic ecosystems to
the  shifts  in  pH.   In  the case  of  terrestrial ecosystems, the  situation  is more complicated
since no components  of  terrestrial  ecosystems appear  to be as  sensitive to a change in pH as
organisms living  in poorly buffered aquatic  ecosystems.   In addition,  indirect effects may
only be expressed  after long periods of time.  Nonetheless, the possibility exists that soils
and  vegetation  may be  affected,  directly or  indirectly,  by acidic  precipitation,  albeit in
complex ways.

                                            7-66

-------
  TABLE 7-6.  CHANGES IN AQUATIC BIOTA LIKELY TO OCCUR WITH INCREASING ACIDITY
1.   Fish populations are reduced or eliminated.

2.   Bacterial decomposition is reduced and fungi may dominate saprotrophic
    communities.  Organic debris accumulates rapidly, tying up nutrients and
    limiting nutrient mineralization and cycling.

3.   Species diversity and total numbers of species of aquatic plants and
    animals are reduced.  Acid-tolerant species dominate.

4.   Phytoplankton productivity may be reduced due to changes in nutrient
    cycling and nutrient limitations.

5.   Biomass and total productivity of benthic macrophytes and algae may
    increase due partially to increased lake transparency.

6.   Numbers and biomass of herbivorous invertebrates decline.  Tolerant
    invertebrate species, e.g., air-breathing bugs (water-boatmen, back-
    swimmers, water striders) may become abundant primarily due to reduced
    fish predation.

7.   Changes in community structure occur at all trophic  levels.
                                7-67

-------
                 TABLE 7-7.  SUMMARY OF EFFECTS ON AQUATIC ORGANISMS ASSOCIATED WITH A RANGE  IN pH
CTI
00
      8.0-6.0  •   Long-term changes of less than 0.5 pH units in the range 8.0 to
                 6.0 are  likely to alter the biotic composition of freshwaters to
                 some degree.  The significance of these slight changes, however,  is
                 not great.

              •   A  decrease  of 0.5 to 1.0 pH units in the range 8.0 to 6.0 may cause
                 detectable  alterations in community composition.  Productivity of
                 competing organisms will vary.  Some species will be eliminated.
                 Phytoplankton plentiful and well distributed but numbers of species
                 begin  to  decrease  as pH decreases.
6.0-5.5 •   Decreasing pH from 6.0 to 5.5 will  cause a reduction  in species
           numbers and,  among remaining species,  alterations  in  ability
           to withstand  stress,  and change in  species dominance.
           Reproduction  of some  salamander species  is impaired.


5.5-5.0  •  Below ptf 5.5, numbers and diversity of species will be  reduced.
           Many species  will  be  eliminated.  Crustacean  zooplankton,  phy-
           toplankton, molluscs, amphipods,  most  mayfly  species, and  many
           stone fly species  will begin to drop out.   In contrast, several
           pH-tolerant invertebrates will  become  abundant,  especially the
           air-breathing forms (e.g., Gyrinidae,  Notonectidae, Corixidae),
           those with tough cuticles which prevent  ion losses (i.e.,
           Sialis lutaria). and  some forms which  live within  the sediments
           (Oligochaeta, Chiromomidae,  and Tubificidae).  Overall, inver-
           tebrate biomass may be reduced.

5.0-4.5 •   Below pH 5.0, decomposition  of organic detritus  will  be severely
           impaired.   Organic matter accumulates  rapidly.   Some  fungal
           species increase (Hyphomycetes,  basidomycetes).  Many fish
           species are eliminated,  (see Table  7-5 for fish  species
           eliminated.)
                                                                                   Aimer et al.,  1974;
                                                                                   Leivestad et al.,  1976;
                                                                                   Aimer et al.,  1978
Aimer et al., 1974;
Leivestad et al., 1976;
Conroy et al., 1976;
Aimer et al., 1978

Aimer et al., 1974;
Kwiatkowski and Roff, 1976;
Aimer et al., 1978

Aimer et al., 1974;
Leivestad, 1976;
Kwaitkowski and Roff, 1976;
Aimer et al., 1978
Rough and Wilson, 1976

Aimer et al., 1974;
Leivestad et al., 1976;
Hendrey et al., 1976;
Grahn et al., 1974;
Grahn, 1976;
Kwiatkowski and Roff, 1976;
Hagen and Langeland, 1973;
Henriksen and Wright, 1977
Hultberg, 1976;
Aimer et al., 1978

Leivestad et al., 1976;
Schofield, 1976b;
Aimer et al., 1978
Hall et al., 1980.

-------
                                                 TABLE 7-7 (continued)
      4.5 and
      below
                Macrophytes, such as Lobelia, are replaced by Sphagnum moss.
                Number of algal species decreases.  Acid-tolerant forms remain.
Below pH 4.5,  all  of the above changes  will be  greatly exacerbated,
and all  fish will  be eliminated.   Lower limit for  many algal
species.
I
en
vo
Leivestad et al., 1976
Hendrey et al., 1976;
Grahn et al., 1974;
Aimer et al., 1978

Leivestad et al., 1976;
Hendrey et al., 1976;
Grahn et al., 1974;
Aimer et al., 1978

Aimer et al., 1974
Leivestad et al., 1976;
Schofield, 1976a,b;
Wright et al., 1976
Beamish et al., 1975;
Menedex, 1976;
Trojnar, 1977a,b;
Schofield, 1979
      Source:  Modified from Hendrey (1978).

-------
7.3.2.1   Effects onsoils—Acidity  Is  a critical  factor  in  the  behavior of  natural  or
agricultural soils.  Soil  acidity influences the availability of  plant nutrients  and various
microbiological processes which  are  necessary for the  functioning  of terrestrial  ecosystems;
therefore,  there  is  concern that  acidic precipitation over time  could have  an  acidifying
effect on soils through  the addition of  hydrogen  ions.   As water containing hydrogen cations
(usually from  weak  acids)  moves  through  the  soil,  some of the hydrogen ions replace adsorbed
exchangeable cations, such  as  Ca  ,  Mg   , K  , and Na  (see Figure 7-26).  The removed cations
are  then carried  deep   into  the  soil  pr.ofile  or  into the ground  water.   In native  soils
hydrogen ions are derived from the following sources (Wiklander,  1979):
     1.   nutrient uptake by plants—the roots adsorb cation nutrients and desorb H ;
     2.   COp produced by plant roots and micro-organisms;
     3.   oxidation of NH4+ and S, FeS2, and H2S to HN03 and H2S04;
     4.   very acid  litter  in  coniferous forests, the  main  acidifying source for the A and B
          horizons;
     5.   atmospheric deposition  of  H^SO. and some HNO-,  NO  , HC1  and NH.  (after nitrifica-
          tion to HN03).
In addition to the  acidifying  factors listed above, the use of ammonium fertilizers on culti-
vated  lands  increases the  hydrogen  cations  in the water  solution.   Ammonium fertilizers are
oxidized by  bacteria to  form  nitrate  (NO.,  )  and hydrogen ions (H )  (Donahue et  al., 1977).
Increased  leaching  causes  soils to  become  lower in  basic Ca  , Mg  ,  Na , and  K  cations
(Donahue et al., 1977).   Sensitivity to leaching is according to the following sequence:  Na
» K+ > Mg2+ > Ca2+ (Wiklander, 1979).
     Norton (1977)  cited the potential effects of acidic  deposition  on soils that are listed
in Table 7-8.  Of  those  listed,   only  the increased  mobility of cations and their accelerated
loss has been  obse'rved  in field experiments.  Overrein  (1972) observed an increase in calcium
                                                                                     ++    -!_.{.
leaching under simulated acid  rain conditions, and  increased  loss by leaching of Ca  ,  Mg   ,
       +3
and  Al    were observed  by Cronan (1980) when he treated New Hampshire soils with simulated
acid rain at a pH 4.4.
     Wiklander (1979) notes  that in humid areas leaching leads to a gradual decrease of plant
nutrients in available and mobilizable forms.  The rate of nutrient decrease is  determined by
the  buffering  capacity   of  the  soil  and  the amount and composition  of precipitation (pH and
salt  content). Leaching  sooner   or  later  leads  to soil  acidification unless  the buffering
capacity of  the  soil is great and/or the salt concentration of precipitation is  high.   Soil
acidification  influences the amount  of exchangeable  nutrients  and  is  also likely to affect
various biological  processes in the soil.
                                                      o-        -
     Acidic precipitation increases the amounts of SO.   and N03  entering the soils.  Nitrate
is easily  leached   from  soil;  however,- because it  is  usually deficient in  the  soil  for both
                                            7-70

-------
ACID RAIN
1
SOIL PARTICLES
^flLtf* _^^^^
3INu ^
*

Mg2+
K*
^^
Na*^
NH4*
sof-_
SOIL SOLUTION
Ca2*
Mg2*
11+
K*
Na*
NOJ
so?-
V -rf
CAN BE LEACHEC
Figure 7-26. Showing the exchangeable ions
of a soil with pH  7, the soil solution com-
position, and the replacement of IMa+ by H+
from acid rain.

Source: Wiklander (1979).
                     7-71

-------
             TABLE 7-8.   POTENTIAL EFFECTS OF ACID PRECIPITATION ON SOILS
                    Effect
             Comment
        Increased mobility of
          most elements
        Increased loss of
          existing clay minerals
        A change in cation
          exchange capacity

        A general propor-
          tionate increase in
          the removal of all
          cations from the soil

        An increased flux in
          nutrients through the
          ecosystem below the
          root zone
Mobility changes are essentially
  in the order:   monovalent,
  divalent, trivalent cations.

Under certain circumstances may
  be compensated for by production
  of clay minerals which do not have
  essential (stoichiometric) alkalies
  or alkali earths.

Depending on conditions, this
  may be an increase or a decrease.

In initially impoverished or
  unbuffered soil, the removal
  may be significant on a time
  scale of 10 to 100 years.
     Source:  Norton (1977).



plants and  soil  microorganisms,  it  is rapidly  taken up  and retained within  the soil-plant

system (Gjessing  et al., 1976; Abrahamsen  et a!.,  1976; Abrahamsen and  Dollard,  1979).   The

fate  of  sulfate  is determined by  its mobility.   Retention  of sulfate  in soils  appears  to

depend on  the amount  of hydrous oxides of  iron and aluminum present.   The amounts  of these

compounds present varies  with  the soil type.  Insignificant amounts of the hydrated oxides of

iron  (Fe)  and aluminum (Al) are  found in  organic soils; therefore, sulfate retention  is  low

(Abrahamsen and Dollard, 1979).  The presence of hydrated oxides of iron and aluminum, however,

is only  one of the factors  associated with  the capability of a soil to retain  sulfur.   The

capacity of soils to adsorb  and retain anions increases as the pH decreases and with the salt

concentration. Polyvalent anions  of  soluble salts  added experimentally  to  soils  increase

adsorption and decrease  leaching  of salt cations.  The effectiveness of the anions studied in
                                                           -       -       2        -
preventing leaching  increased  in  the following  order:   Cl  ~ N03   < SO,  < H^PO. (Wiklander,

1980).  Additions of sulfuric acid to a soil will have no effect on cation leaching unless the
sulfate anion  is mobile,  as  cations cannot  leach without associated anions  (Johnson  et al.

1980; Johnson, 1980; Johnson and Cole, 1980).
                                            7-72

-------
     Leaching of soil  nutrients  is efficiently inhibited by  vegetation  growing on it.   Plant
roots take up the  nutrients frequently in larger  amounts  than required by the plants.   Large
amounts of these nutrients  will  later be deposited on the soil surface as litter or as  leach-
ate from the vegetation canopy (Abrahamsen and Dollard, 1979).
     In lysimeter experiments  in  Norway,  plots with vegetation cover were used.  One plot had
a dense layer of the grass, Deschampsia flexuosa (L.) Trin.  and the other a less dense  cover.
                                                      O—
The soil retained approximately  50 percent of the SO,   added to it.   The greatest amount was
retained in the lysimeters  covered with grass; the relative retention increased with increas-
ing additions of  sulfate (Abrahamsen and  Dollard,  1979).   Leaching of  cations  from the soil
                                       ?—                         ?+       ?+
was reduced by the  retention of the SO,  ; however, leaching of Ca   and Mg   increased  signi-
ficantly as the acidity of  the simulated  rain  increased.  In the most acid treatment leaching
of Al  was  highly significant.   The behavior of K  ,  NO,  ,  and NH,  was  different  in the two
lysimeter series.  These ions were retained in the grass-covered lysimeters whereas there was a
net  leaching  of K   and  N03  in  the  other  series.   Statistically significant  effects were
obtained only when  the  pH  of the  simulated rain  was 3.0 or  lower  (Abrahamsen and Dollard,
1979).
     The Scandinavian  lysimeter  experiments  appear to demonstrate that  the  relative rate of
adsorption of sulfate  increases  as the  amounts applied are  increased.   In  the control  lysi-
meters  the  output/input  ratio was approximately  one.   These  results  are in  agreement with
results of watershed studies which frequently appear to demonstrate that, on an annual  basis,
sulfate outflow is  equal  to or greater than  the  amounts  being added  (Gjessing  et  al.,  1976;
Abrahamsen and Dollard,  1979).   Increased outflow may be attributed to dry deposition and the
weathering of sulfur-bearing rocks.  The increased deposition of sulfate via acidic precipita-
tion appears to  have increased the leaching of sulfate from the soil.   Together with the reten-
tion  of hydrogen  ions  in  the soil  this results  in an increased  leaching of  the nutrient
         +    2+     2 +
cations K , Ca  , Mg  , and Mn (Abrahamsen and Dollard, 1979).  Shriner and Henderson (1978),
however, in their study of  sulfur distribution and cycling  in the Walker Branch Watershed in
eastern Tennessee noted the additions of sulfate sulfur by precipitation were greater than the
amount  lost in  stream  flow.  Analysis of  the biomass and soil concentrations of sulfur indi-
cated that sulfur was being retained in the mineral soil horizon.  It is suggested that  leach-
ing  from  organic  soil  horizons  may be  the   mechanism  by which sulfur  is  transferred  to the
mineral horizon.   Indirect evidence suggests that vegetation scavenging of atmospheric sulfate
plays an important role by adding to the amounts of sulfur entering the forest system over wet
and dry deposition.
     Studies of the  nutrient cycling of sulfur in a number of forest ecosystems indicate that
some ecosystems  accumulate  (Johnson et al., 1980; Heinrichs and Mayer, 1977; Shriner and Hender-
son, 1978) while other ecosystems maintain a balance between the additions and losses of sulfur
or show a  net loss (Cole & Johnson, 1977).  Sulfur accumulation appears to be associated with
sulfate  adsorption   in  subsoil  horizons.   Sulfate  adsorption  is  strongly dependent  on pH.
                                            7-73

-------
Little adsorption occurs above pH 6-7 (Harward and Reisenaur, 1966).  The amount of sulfate in
a soil is a function of a soil's adsorption properties and the amount of sulfate that has been
added to the soil, integrated over time.  Soil properties may favor the adsorption of sulfate;
however, the fiet annual accumulation of sulfate at any specific time will be influenced by the
degree of soil saturation (Johnson et al., 1980),
     Lowered  soil  pH also influences the  availability and toxicity of  metals  to plants.   In
general,  potentially  toxic  metals  become more  available  as  pH  decreases.   Ulrich  (1975)
reported that aluminum  released by acidified soils could be phytotoxic if acid rain continued
for a  long  period.  The degree of  ion  leaching increased with decreases in pH, but the amount
of cations  leached  was  far less than the  amount of acid added (Malmer,  1976).   Baker et al.
(1977) found that sulfur dioxide in precipitation increased the extractable acidity and alumi-
num, and decreased  the  exchangeable bases, especially calcium and magnesium.  Although dilute
sulfuric acid in  sandy podsolic  soils caused  a significantly  decreased  pH of  the  leached
material, the  amount of acid  applied  (not  more than  twice the yearly airborne supply over
southern Scandinavia)  did  not  acidify  soil   as  much  as did nitrate fertilizer (Tamm  et al.,
1977).  Highly acidic  rainfall,  frequently with a pH less than 3.0, in combination with heavy
metal particulate fallout from smelters, has caused soils to become toxic to seedling survival
and establishment  according  to  observations  by  Hutchinson and Whitby  (1976).  Very  low soil
pH's are  associated with  mobility of  toxic  aluminum compounds  in the  soils.  High acidity,
High sulfur, and heavy metals in the rainfall have caused fundamental  changes in the structure
of soil organic  matter.  The sulfate and  heavy  metals were borne by air from the smelters in
the Sudbury area  of Ontario  and brought to earth by dry and wet deposition.  Among the metals
deposited in rainfall and dustfall  were nickel, copper, cobalt, iron,  zinc, and lead.   Most of
these metals  are retained in  the  upper layers  of  soil,  except in very  acid  or  sandy soils.
     The accumulation of metals in soils is mainly an exchange phenomenon.  Organic components
of litter,  humus, and soil  may bind heavy metals as stable complexes (Tyler, 1972).  The heavy
metals when  bound may  interfere with  litter decay and nutrient cycling,  and  in this manner
interfere with  ecosystem functioning  (Tyler, 1972).    Acidic  precipitation, by  altering  the
equilibria of the metal complexes through mobilization, may decrease the residence time of the
heavy metals in soil and litter (Tyler, 1972, 1977).
     Biological  processes  in  the   soil  necessary for  plant growth can  be affected  by soil
acidification.   Nitrogen  fixation,  decomposition  of organic  material, and  mineralization,
especially  of nitrogen, phosphorus and  sulfur,  could be  affected (Abrahamsen  and  Do!lard,
1979; Tamm  et  al.,  1977; Malmer, 1976; Alexander, 1980).   Nearly all  of the nitrogen,  most of
the phosphorus and sulfur,  as well  as other nutrient elements in the soil are bound in organic
combination. In this form,  the elements are largely or entirely unavailable for utilization by
higher  plants (Alexander, 1980).   It  is  principally  through  the activity  of heterotrophic
microorganisms that  nitrogen,  phosphorus, and sulfur  are  made available  to  the autotrophic
                                            7.74

-------
higher plants.   Thus, the microbial processes that lead to the conversion of the organic forms
of these elements  to the inorganic state are crucial for maintaining plant life in natural  or
agricultural ecosystems.   The key role  of  these degradative processes  is  illustrated  by the
fact that nitrogen  is  limiting for food production  in  much of the world  and  governs primary
productivity in many terrestrial  habitats (Alexander, 1980).
     Many,   and probably  most, microbial  transformations  in  soil  may be  brought  about  by
several species.  Therefore, the reduction or elimination of one population is not necessarily
detrimental  since  a  second population,  not affected by  the stress,  may fill the partially or
totally vacated niche.   On the other hand,  there are a few processes that are carried out,  so
far as  it is  now known, by only  a single species, and elimination of that species could have
serious consequences.   Examples  of this  are  nitrification in which ammonium  is  converted  to
nitrate, and nodulation  of leguminous plants, for which  the  bacteria are reasonably specific
according to the leguminous host (Alexander, 1980).
     Nitrification  is  one of the  best  indicators of   pH  stress  because  the  responsible
organisms,   presumably largely  autotrophic  bacteria,  are  sensitive  both  in culture  and  in
nature  to  increasing acidity  (Dancer et al.,  1973).   Although  nitrification will  sometimes
occur  at pH values  below  5.0, characteristically the rate decreases with increasing acidity
and often is undetectable much below pH 4.5.  Limited data suggest that the process of sulfate
reduction to sulfide in  soil  is markedly inhibited  below a pH of 6.0.  (Connell  and Patrick,
1968)  and studies  of the presumably responsible organisms in culture attest to the inhibition
linked with the acid conditions (Alexander,  1980).
     It is  difficult to  make  generalizations concerning  the  effects  of soil acidification on
microorganisms.   Many microbial  processes  that  are important for  plant growth  are clearly
suppressed as the pH declines; however, the inhibition noted in one soil at a given pH may not
be  noted  at  the   same  pH  in  another  soil  (Alexander,  1980).    The  capacity  of  some
microorganisms  to  become   acclimated to  changes  in  pH  suggests  the  need  to study this
phenomenon  using  environments  that have been maintained at different pH values for some time.
Typically the   studies  have been  done  with  soils  maintained only for  short periods  at the
greater acidity (Alexander, 1980).   The consequences of increased acidity in the subterranean
ecosystem are   totally  unclear,   however;  the pH  of soil  influences  not only  the  microbial
community at  large, but  also those  specialized populations that colonize  the root surfaces
(Alexander,  1980).
     The addition of nitrate  and other  forms  of nitrogen from  the  atmosphere to ecosystems
through the activity of  microorganisms is an  integral  function  of  the terrestrial nitrogen
cycle.   The contribution  of   inorganic  nitrogen  in wet  precipitation  (rain plus  snow)  is
usually equivalent to  only a   few percent of the total nitrogen assimilated annually by plants
in terrestrial  ecosystems;  however,  total nitrogen contributions, including organic nitrogen,
in bulk precipitation  (rainfall  plus dry fallout)  can  be significant,  especially in unferti-
lized  natural systems.
                                            7-75

-------
      Atmospheric contributions  of nitrate  can  range from less  than-0.1 kg  N/ha/yr  in  the North-
 west (Fredricksen,  1972) to 4.9  kg  N/ha/yr in the Eastern  United States (Likens et al., 1970;
 Henderson  and Harris, 1975).   Inorganic nitrogen  (ammonia-N plus nitrate-N) additions in wet
 precipitation ranged from less than  0.5  kg/ha/yr to more  than  3.5 kg/ha/yr in Junge's (1958)
 study of rainfall over the United States.   On  the  other hand,  total  nitrogen loads in bulk pre-
 cipitation range from  less  than  5  kg/ha/yr  in  desert regions  of the West  to  more than 30
 kg/ha/yr near barnyards in the Midwest.   Total contributions of  nitrogen  from the atmosphere
 commonly range from about 10 to 20 kg N/ha/yr  for  most  of the  United States (National  Research
'Council, 1978).
      In  comparison, rates of annual  uptake by plants in ecosystems  selected from  several bio-
 climatic zones;  (deciduous  forest,  tundra, desert, western coniferous forest, grassland and
 tropical forest) range  from  11 to 125 kg  N/ha/yr.  (National  Research Council, 1978).  Since
 the  lowest  additions are  generally  associated  with  desert  areas  where   rates of  uptake by
 plants are low, and the highest  additions usually occur in moist areas where plant  uptake is
 high,  the  contributions  of  ammonia  and  nitrate  from  rainfall  to terrestrial ecosystems are
 equivalent to about 1 to 10 percent of annual  plant uptake.   In eastern deciduous and western
 coniferous forest ecosystems, contributions from bulk precipitation,  on the other  hand, repre-
 sent from  about 8  to 25 percent  of  the annual  plant requirements.   Although these comparisons
 suggest  that plant growth in terrestrial  ecosystems depends to  a  significant extent on atmo-
 spheric  deposition, it is not  yet possible to estimate the  importance of these contributions
 by comparing them with the biological fixation and mineralization of nitrogen in  the soil. In
 nutrient-impoverished  ecosystems, such as  badly eroded abandoned croplands or soils  subjected
 to prolonged leaching by acidic precipitation,  nitrogen additions from atmospheric depositions
 are  certainly important to biological productivity.  In largely unperturbed forests,  recycled
 nitrogen from the  soil organic pool  is  the chief  source of nitrogen for  plants,  but nitrogen
 to support increased production  must  come either  from  biological  fixation or from atmospheric
 contributions.   It  seems possible,  therefore, that man-generated contributions  could play a
 significant ecological role  in  a  relatively large  portion of the forested  areas near  industria-
 lized regions (Galloway, 1978).
      Sulfur,  like  nitrogen,  is  essential for optimal  plant growth.   Plants  usually obtain
 sulfur from the soil  in  the  form  of  sulfate.   The  amount of mineral  sulfur in soils is usually
 low  and  its release from organic matter  during microbial  decomposition is a major source for
 plants (Donahue et al., 1977).   Another major source is the  wet and dry  deposition  of atmos-
 pheric sulfur (Donahue et al.,  1977;  Brady, 1974;  Jones, 1975).
      In  agricultural  soils  crop residues, manure,  irrigation  water,  fertilizers,  and soil
 amendments are  important sources of sulfur.  The  amounts  of  sulfur entering the soil system
 from atmospheric sources  is dependent on proximity to industrial  areas, the sea coast, and
 marshlands.   The prevailing winds and the amount of precipitation  in  a given region are also
 important  (Halstead and  Rennie, 1977).  Near fossil-fueled  power plants and industrial sources,
                                             7-76

-------
the amount  of sulfur In precipitation  may be as much  as  150 pounds per acre  (168  kg/ha) or
more (Jones,  1975).   By contrast, in rural  areas,  based on the equal  distribution  of sulfur
oxide emissions over  the coterminous states, the amount of sulfur in precipitation is gener-
ally well below the  average 15 pounds per  acre  (17 kg/ha).  Approximately 5 to  7 pounds per
acre (7 to  8  kg/ha)  per year were reported for  Oregon in 1966 (Jones, 1975).   Shinn and Lynn
(1979) have estimated that  in the Northeastern United States, the area where precipitation is
most acidic,  approximately  5  x 10  tons of sulfate per year is removed by rain.  Hoeft et al.
(1972) estimated the  overall  average sulfur as sulfate  deposition at 26 pounds of sulfur per
acre per year (30  kg S/ha/year).   Estimates for rural areas were 14 pounds of sulfur per acre
per year (16  kg/ha/yr).  Approximately  40 to 50 percent of the sulfur additions occurred from
November to  February.  Tabatabai and  Laflen (1976)  found that SQ*-S  deposition  in Iowa was
greatest in fall  and winter when precipitation was low.
     Experimental  data have shown that even though plants are supplied with adequate soil sul-
fate they  can absorb 25  to  35  percent of  their sulfur  from  the atmosphere  (Brady,  1974).
Particularly  if the  soil sulfur is low and atmospheric sulfur is high, most of the sulfur re-
quired by the plant  can come from the  atmosphere (Brady,  1974).  Atmospheric sulfur would be
of  benefit  chiefly to  plants growing on  lands  with  a low  sulfur  content (Brezonik,  1976).
     Tree species  vary  in  their ability to  utilize sulfur.   Nitrogen and sulfur are biochem-
ically associated in plant  proteins,  therefore,  a close relationship exists between the two in
plants.  Apparently,  nitrogen is only taken up at the rate at which sulfur is available.  Pro-
tein formation, therefore,  is limited by the amount  of sulfur available (Turner and Lambert,
1980).   Conifers  accumulate as sulfate any sulfur  beyond the amount  required  to  balance the
available nitrogen.   Protein  formation  proceeds  at the  rate  at  which nitrogen becomes avail-
able.  Trees  are   not injured when sulfur  is applied as sulfate  rather than S0?  (Turner and
Lambert, 1980).
     When  discussing the  effects of  acidic precipitation,  or  the  effects  of  sulfates or
nitrates on  soils, a distinction should  be made between managed  and  unmanaged soils.   There
appears to  be general agreement  that managed agricultural soils  are  less  susceptible  to the
influences of acidic  precipitation  than are unmanaged  forest or rangeland soils.   On managed
soils  more  than  adequate  amounts of lime  are  used  to counteract  the acidifying effects of
fertilizers   in  agricultural  soils.   Ammonium  fertilizers,  usually  as  ammonium  sulfate
                                                                                           2-
[(NH^nSOfl] or ammonium nitrate,  (NH.NCU)  are oxidized  by bacteria  to  form  sulfate (SO.  )
and/or nitrate (NOl)  and hydrogen ions (H  )  (Donahue et al.  1977; Brady, 1974).   The release
of  hydrogen ions  into the  soil causes the  soil  to become  acidified.   Hydrogen ions are also
released  into the soil  when plants  take up mineral  nutrients.  Hence,  substances (notably
various complexes  of ammonium  and  sulfate  ions),  although of neutral pH, or  nearly so, are
acidifying in  their effects when they are taken up by plants or animals.  Thus,  the concept of
"acidifying precipitation"  must be added to the concept of "acid precipitation."
                                            7-77

-------
     The acidifying  effects  of fertilization or acidic precipitation  is  countered in managed
soils through the  use of lime.  Liming tends to raise the pH and thereby eliminate most major
problems associated  with  acidic soils (Donahue et  al., 1977;  Likens et al., 1977).  Costs of
liming all natural  soils  sensitive to acidification would be prohibitive as well as extremely
difficult to carry out.
     Precipitation  adds many chemicals to terrestrial, aquatic,  and agricultural  ecosystems.
In addition to  sulfur and nitrogen, phosphorus and potassium  are biologically most important
because they often  are in limited supply  in  the  soil (Likens et al., 1977).  Other chemicals
of varying biological  importance and varying concentration  found  in precipitation over North
America are the  following:   chlorine, sodium, calcium, magnesium, iron, nickel, copper, zinc,
cadmium, lead,  manganese (Beamish, 1976; Hutchinson and Whitby, 1976; Brezonik, 1976), mercury
(Brezonik,  1976),   and cobalt   (Hutchinson  and Whitby,  1976).   Rain over  Britain and  the
Netherlands, according to Gorham (1976), contained the following elements in addition to those
reported  for  North  American  precipitation:   aluminum, arsenic,  beryllium,  cerium, chromium,
cesium, antimony,  scandium,  selenium,  thorium,  and vanadium.   Again,  it  is obvious that many
of these  elements  will be found in  precipitation  in highly industrialized areas and will  not
be of  biological importance  until  they enter  an  ecosystem where they may  come into contact
with some form  of  life,  as  in  the  case of heavy metals in the waters and soils near Sudbury,
Ontario.  Of the chemical  elements found in precipitation, magnesium, iron, copper, zinc,  and
manganese  are  essential   in  small   amounts  for   the  growth  of  plants;   however,  at  high
concentrations these  elements,  as  well as the  other  heavy metals, can be toxic to plants  and
animals.  Furthermore, the  acidity of precipitation  can affect the solubility, mobility,  and
toxicity of these  elements  to the foliage or roots of plants and to animals or microorganisms
that may ingest oh decompose these plants.
     Wiklander  (1979)  has pointed  out that  based  on the  ion exchange theory,  ion exchange
experiments, and  the  leaching of  soil  samples, the following  conclusions  can  be  drawn about
the acidifying effect on soils through the atmospheric deposition of mineral acids:
     1.   At a  soil  pH >  6.0, acids are fully neutralized by decomposition of CaCOo and other
          unstable minerals and by cation exchange.
     2.   At soil  pH < 5.5, the efficiency of the proton to decompose minerals and to replace
                         2+     2+    +        +
          exchangeable Ca   ,  Mg   ,  K , and Na  decreases with the soil pH.   Consequently,  the
          acidifying effect  of  mineral acids on soils decreases, but the effect on the runoff
          water increases in the very acid soils.
     3,   Salts of  Ca  +,  Mg  ,  K , and NH»  in the precipitation counteract the absorption of
          protons and, in that way,  the decrease of the base saturation.   A proportion of the
          acids percolate through the-soil  and acidify the runoff.
     The  sensitivity  of  various  soils  to  acidic  precipitation  depends  on the  soil  buffer
capacity and on  the soil  pH.  Noncalcareous sandy soils with pH > 5 are the most sensitive to
acidification;  however, acidic soils would be most likely to release aluminum.
                                            7-78

-------
     Very  acid  soils  are less  sensitive  to  further  acidification because  they are already
 adjusted  by soil formation to  acidity and are therefore more  stable.   In these soils easily
 weatherable  minerals  have disappeared, base  saturation  is  low, and the pH of  the  soil may be
 less than  that of precipitation. The  low  nutrient  level  is  a crucial factor  which  limits pro-
 ductivity  in these soils.  Even  a  slight  decrease in nutrient  status  by  leaching may have a
 detrimental  effect on  plant  yield  (Wiklander,  1979).    Fertilization  appears  to be the only
 preventive measure.
     In properly managed  cultivated  soils,  acidic precipitation should  cause  only a slight in-
 crease in the  lime requirement,  with a cost compensated  for  by  the  supply  of  sulfur,  nitrogen,
 magnesium, potassium, and calcium made available to plants (Wiklander,  1979).
 7.3.2.2   Effects on  vegetation—The atmosphere, as well  as  the soil, is a source of nutrients
 for plants.   Chemical  elements  reach  the  plant surface via wet  and dry deposition.  Nitrates
 and sulfates  are  not  the only  components of  precipitation falling onto the plant surface.
 Other  chemical  elements  (cadmium,  lead,  zinc, and manganese),  at  least partially  soluble in
 water,  are  deposited  on  the  surface of  vegetation  and may  be  assimilated  by it, usually
 through  the   leaves.    (See   Chapter  8  for  discussion  of  particulate matter.)   An average
 raindrop  deposited on trees  in  a typical  forest washes  over three tiers  of  foliage before it
 reaches  the  soil.   The  effects of  acidic  precipitation   may  be beneficial  or  deleterious
 depending  on its chemical composition, the species of plant on which it is deposited, and the
 physiological  condition  and  maturity  of  the  plant (Galloway  and  Cowling, 1978).   Substances
 accumulated  on the leaf  surfaces strongly influence the chemical composition of  precipitation
 not only at  the leaf  surface,  but also  when it  reaches the  forest  floor.  The chemistry of
 precipitation  reaching the forest   floor  is  considerably different from that  collected above
 the forest canopy or a ground  level where the canopy has no influence  (Lindberg  et al. 1979).
 Except  for  the  hydrogen   ion (H )  the mean  concentrations  of  all  elements  (lead,  manganese,
 zinc  and  cadmium)  studied in the Walker  Branch Watershed in Tennessee were  found by Lindberg
 et al.  (1979) to  be present in greater  amounts   in the throughfall   than  in incident rain.
'Their  study indicated that the  presence  of trace  elements  was  more variable than that of the
 sulfate  and  hydrogen  ions  and  that  throughfall  appeared  to  be  a more  dilute solution of
 sulfuric  acid  than rain with  a  pH ~ 4.5 not influenced  by the forest canopy.  The solution was
 found  to contain  a  relatively  higher concentration of alkaline  earth salts  of  sulfate and
 nitrate  as well as a  somewhat  higher  concentration of  trace elements  (Lindberg  et al. 1979).
      Lee  and  Weber  (1980) studied  the effects of sulfuric acid  rain on two model hardwood
 forests.   The  experiment, conducted under  controlled field conditions,  consisted  of the appli-
 cation of simulated  sulfuric  acid rain (pH values  of 3.0, 3.5,  and  4.0), and  a  control rain of
 pH 5.6 to the two model  forest ecosystems for a duration of 3  and  1/2  years.   Rainfall appli-
 cations_  approximated   the annual   amounts  of  areas  in which sugar maple  and  red  alder
 communities  normally  occur.
                                             7-79

-------
     In evaluating the results of the study, the authors conclude that a well developed forest
canopy and  litter layer  can increase the  pH and  concentration of bases  (i.e.,  calcium and
magnesium)  in  rainwater.   Such  conditions  would tend  to decrease the  acidification  rate of
forest soils by acid rain.  However, as  bases  are continually  leached  from  the  soil  column,
these  cations  could eventually  be  lost  from  the ecosystem and unavailable  to  influence the
acidification reactions.  Changes in the ionic and pH balance of forest systems may impact the
productivity of  forests  through  acidity-induced changes  in  the  nutrient cycling  process, de-
composition, reproduction,  tree  growth,  and structure of  forest systems  (Alexander,  1980).
     The additions of  hydrogen,  sulfate and nitrate  ions to soil and plant systems have both
positive and negative  effects.   It has generally been assumed that the free hydrogen ion con-
centration in acidic precipitation is the component that is most likely to cause direct, harm-
ful effects  on vegetation  (Jacobson,  198Qa).  Experimental studies  support this assumption;
however,  to date,  there are no confirmed reports  of exposure  to ambient acidic precipitation
causing foliar  symptoms  on field grown vegetation in the continental  United States (Jacobson,
198Qa) or Canada (Linzon personal communication 1980).
7.3.2,2.1   Direct effects on vegetation.    Hydrogen  ion concentrations equivalent  to  that
measured in more  acidic rain events (5 pH  3.0)  have been observed experimentally through the
use of simulated  acid  rain to cause tissue  injury  in the form  of  necrotic lesions to a wide
variety of plant  species  under greenhouse and  laboratory conditions.  This  visible injury has
been reported as occurring between pH 3.0 and 3.6 (Shriner,  1980). The various types of direct
effects which have been reported are shown in Table 7-9. Such effects must be interpreted with
caution because  the  growth and morphology of leaves on plants  grown in greenhouses frequently
are atypical  of  field conditions  (Shriner, 1980).    (See  Chapter 8  for  discussion  of the
vegetational effects of SCL).
     Small necrotic lesions, the most common form of direct injury, appear to be the result of
the collection  and retention  of water on  plant surfaces  and  the subsequent  evaporation of
these water droplets which concentrates the solution's constituents causes  a lesion to occur.
The  depression  formed  by  the lesion  further  enhances  the collection of  water.   A  large
percentage  of  the leaf  area may exhibit lesions after repeated exposures  to  simulated acid
rain at  pH concentrations  of 3.1,  2,7, 2.5 and  2.3 (Evans et  al. 1977a,  1977b).   In leaves
injured by  simulated acidic  rain,  collapse and  distortion of  epidermal  cells on  the upper
surface  is  frequently  followed  by  injury  to  the  palisade cells  and ultimately  both leaf
surfaces are affected (Evans et al., 1977b).  Evans et al. (1978), using six clones of Populus
spp. hybrids,  found  that leaves that had just  reached full expansion were  more  sensitive to
simulated acid  rain at  pH 3.4,  3.1, 2.9,  and  2.7  than those which were unexpanded or fully
expanded.    On   two  of  the clones,  gall  formation  due  to abnormal  cell  proliferation  and
enlargement  occurred.    Other  effects   attributed  to  simulated  acid   rain   include  the
modification of the leaf  surface,  e.g.  epicuticular waxes, and  alteration of physiological
processes such as carbon fixation and allocation.
                                            7-80

-------
           TABLE 7-9.  TYPES OF DIRECT, VISIBLE INJURY REPORTED IN RESPONSE TO SIMULATED ACIDIC WET DESPOSITION
      Injury Type
Species
pH Range
     Reference
Remarks
I
CO
      Pitting, curl
       shortening, death

      1-mm necrotic lesions.
       premature abscission
      Cuticular erosion
      Chlorosis
      (A) small, shallow
       circular depressions:
       slight chlorosis

      (B) larger lesions,
       chlorosis always present
       palisade collapse

      (C) 1-ffim necrotic lesions
       general distortion

      (D) 2-iwii bifacial necrosis
       sue to coalescence of
       smaller lesions, total
       tissue collapse.

      Wrinkled leaves, excessive
       adventitious budding,  pre-
       mature abscission
Yellow birch
Kidney bean,
 soybean,
 loblolly pine,
 E. white pine,
 willow oak

Willow oak
Sunflower,
 bean

Sunflower,
 bean
Sunflower,
 bean
Sunflower,
 bean

Sunflower,
 bean
Bean
2.3-4.7   Wood and Bormann, 1974
3,2
Shriner et al., 1974
3.2
Shriner, 1978a,
 Lang et al., 1978
2.3-5.7   Evans et al., 1977b
2.7       Evans et al., 1977b
2.7       Evans et al., 19J7b
2.7       Evans et al., 19775
2.7       Evans et al., 1977b
l.S-3.0   Ferenbaugh, 1976
                         More frequent near
                               veins.  (A) - (D)
                               represent sequential
                               stages of lesion
                               development, through
                               time, up to 72 hrs (one
                               6-min rain event daily
                               for 3 days)

-------
                                                        TABLE 7-9 (continued)
         Injury Type
                              Species
                    pH Range
     Reference
Remarks
00
         Incipient bronzed spot

         Bifacial necrotic pitting
         Necrotic  lesions,
         premature abscission
        Marginal and tip necrosis
Galls, hypertrophy,
 hyperplasia

Dead leaf cells
                              Bean

                              Bean
E. white pine,
 scotch pine,
 spinach,
 sunflower,
 bean

Bean, poplar,
 soybean, ash
 birch, corn,
 wheat

Hybrid poplar
                                      Soybean
                    2.0-3.0   Hindawi et al., 1980

                    2,0-3.0   Hindawi et al., 1980
                                                  2.6-3.4   Jacobson and
                                                             van Leuken, 1977
                                                  Submicron Lang et al.,  1978
                                                   H2S04
                                                   aerosol
                                                          2.7-3.4   Evans et al., 1978
                              After first few hours

                              After 24 h (reported
                               pooling of drops =
                               more injury)
                              Injury associated with
                               droplet location
                               within 24-48  h.
                    3.1
Irving, 1979
         Source:  Shriner, 1980

-------
     Lee et al.  (1980) studied the effects of simulated acidic precipitation on crops.   Dependa
ing on  the crop  studied,  they reported  beneficial,  detrimental  or no effects  on  yield when
crops were exposed  to sulfuric acid rain at  pH  values of 3.0, 3.5, and 4.0 and were compared
to crops exposed  to a control rain of pH 5.6.   The yield of  tomatoes,  green peppers, straw-
berries, alfalfa,  orchard gra.ss,  and  timothy were stiaulated.  Yields  of  radishes, carrots,
mustard greens, and broccoli  were inhibited.   Potatoes were ambiguously affected except at pH
3.0 where  their yield,  as well as that  of  beets,  was inhibited.    Visible  injury  of tomatoes
might have decreased  their market value.  In sweet corn,  stem and leaf production was stimu-
lated,  but no statistically  significant effects  on  yield were observed for  15 other crops.
Results suggest that the possibility of  acid rain affecting yield depends on  the  portion of
the plant  being commercially  utilized  as well as the species.   Plants were regularly examined
for foliar injury associated  with acid rain.   Of the 35 cultivars examined, the foliage of 31
was injured at pH 3.0; 28 at pH 3.5; and 5 at pH 4,0.   Foliar injury was not generally related
to effects on yield.   However, foliar injury of swiss chard,  mustard greens, and spinach was
severe  enough to adversely  affect marketability.   These  results  are  from a  single  growing
season and therefore considered to be preliminary.
     Studies  indicate that wet deposition of acidic  or  acidifying substances may result in a
range of  direct  or indirect  effects  on vegetation.   Environmental  conditions  before, during
and after  a precipitation event  affect  the  responses of vegetation.  Nutrient  status of the
soil, plant  nutrient requirements, plant sensitivity, growth  stage and the  total  loading or
deposition of critical   ions  (e.g.  H , N0«   and SO.   all  play a  role  in  determining vegeta-
tional response to  acidic precipitation).
     Wettability of leaves appears to be an important factor in the response of plants to acid
deposition.  This has been demonstrated in the work of Evans and Curry (1977), Oacobson and van
Leuken (1977), and  Shriner (1978a), who variously report a threshold of between pH 3.1 and 3.5
for development of foliar lesions on  beans.   The  cultivars of Phaseolus vulgaris  L.  used in
the above  studies are all relatively non-waxy and therefore fairly easily wettable.  By com-
parison, studies  with the very waxy leaves of  citrus (Heagle et al., 1978) reported a thres-
hold for visible  symptoms to be near pH 2.0.  Waxy leaves apparently minimize the contact time
for the acid solutions,  thus accounting  for the  <400X increase  in  H   ion concentration re-
quired  to  induce  visible injury.   Table  7-10 summarizes the thresholds, species sensitivity,
concentration, and  time  for visible injury associated with experimental studies of wet deposi-
tion of acidic substances.
     Leaching  of  chemical elements  from exposed  plant  surfaces   is  an  important effect that
rain, fog, mist,  and dew  have on vegetation.  Substances leached  include a great diversity of
materials.  All of the essential  minerals, ami no  acids, carbohydrate growth regulators, free
sugars, pectic substances, organic acids, vitamins, alkaloids, and  allelopathic substances are
among the  materials which have been detected in  plant  leachates   (Tukey, 1970).  Many factors
                                            7-83

-------
               TABLE 7-10,  THRESHOLDS FOR VISIBLE INJURY AND GROWTH EFFECTS ASSOCIATED WITH EXPERIMENTAL

                        STUDIES OF WET DEPOSITION OF ACIDIC SUBSTANCES (AFTER JACOBSQN, 1980a,b)
00
-fe
Effect
Species
Threshold
Reference
Remarks

Foliar lesions, decrease
in growth
Foliar aberrations,
decrease in growth
Foliar lesions
Foliar lesions
Foliar lesions
Foliar lesions
Foliar symptons, no
reduced growth
Increased growth,
i ncreased/decreased
nutrient content"
Reduced growth
Reduced yield
Reduced growth
Reduced yield
Yellow birch
Bean
Bean, sunflower
Bean
Hybrid poplar
Sunflower
Soybean
Lettuce
Pinto bean
Pinto bean
Soybean
Soybean
pH 3.1
pH 2.5
pH 3.1
pH 3.2
pH 3.4
pH 3.4
pH 3.0
pH 3.0, 3.2
pH 3.1
pH 2.7
pH 3.1
pH 2.5
Wood and Bormann (1974)
Ferenbaugh (1976)
Evans et al. (1977a)
Shriner (1978a)
Evans et al. (1978)
Jacobson and
van Leuken (1977)
Jacobson (1980b)
Jacobson (1980b)
Jacobson (1980b)
1/m
1/m
1/rn
greenhouse
greenhouse
greenhouse
greenhouse
greenhouse
greenhouse
greenhouse
greenhouse (varied
witft S04 & N03")
greenhouse
1/m
1/m
1/m

-------
                                                 TABLE 7-10 (continued)
"•J
I
co
en
Effect
Species
Threshold Reference
Remarks

Increased yield
Foliar symptoms
Reduced growth
Reduced yield
Reduced quality
No foliar symptoms, or
effects on growth
No foliar symptoms, but
a) decreased growth, yield
b) increased yield
No effect on growth, yield
Reduced quality
Soybean
Tomato
Tomato
Tomato
Tomato
Soybean
Soybean
Soybean
Soybean
Tomato
Tomato
pH 3.1
pH 3.0 Jacobson (1980b)
pH 3.0
pH 3.0
pH 3.0
pH 3.1 Irving (1979)
pH 2.8 Jacobson (1980b)
pH 2.8
pH 2.8
pH 3.0 Jacobson (1980b)
pH 3.0

greenhouse



field
field, low ozone
field, high ozone
field, low ozone
field
field

       Highest pH to elicit a negative growth response, or lowest pH to elicit a positive growth response


      Shriner, 1980.

-------
influence the quantity  and quality of the substances leached from foliage.   They include fac-
tors associated directly with the plant as well as those associated with the environment.   Not
only are there  differences among species with respect to leaching, but individual differences
also exist among  individual  leaves of the same crop and even the same plant, depending on the
physiological age of the leaf.  Young, actively growing tissues are relatively immune to leach-
ing of mineral nutrients and carbohydrates, while mature tissue which is approaching senescence
is  very  susceptible.   The stage  of  plant development, temperature, and  rainwater  falling on
foliage  and  running  down plant  stems  or tree  bark  influences leaching.    Rainwater,  which
naturally has a pH  of about 5.6, washing  over vegetation may become enriched with substances
                               o
leached from the tissues (Nihlgard, 1970).
     Leaching of  organic and inorganic materials from  vegetation to the soil is part of the
normal functioning  of terrestrial ecosystems.   The nutrient  flow from  one  component of the
ecosystem to  another is  an  important phase  of nutrient cycling  (Comerford  and White,  1977;
Eaton et al., 1973).  Plant leachates have an effect upon soil  texture, aeration, permeability,
and  exchange capacity.   Leachates,   by  influencing the number  and  behavior of soil  micro-
organisms, affect soil-forming processes,  soil  fertility, and  susceptibility or immunity of
plants to soil pests and plant-chemical interactions (Tukey, 1970).
     It  has  been  demonstrated  under experimental conditions  that  precipitation of increased
acidity can  increase  the leaching of various  cations  and organic carbon from the tree canopy
(ABrahamsen et a!.,  1976; Wood and Bormann, 1975).  Foliar losses of potassium, magnesium, and
calcium from bean plants and maple seedlings were found to increase as the acidity of an arti-
ficial mist  was  increased.   Below a pH  of 3.0 tissue  damage occurred;  however, significant
increases in  leaching were measured  at pH  3.3 and  4.0 with no observable tissue damage (Wood
and Bormann,  1975).   Hindawi  et al.   (1980) also  noted that,  as the acidity  of  sulfuric acid
mist increased, so  did  the foliar leaching of nitrogen,  calcium, phosphorous, and magnesium.
Potassium  concentrations were  not  affected,  while the concentration  of  sulfur  increased.
Abrahamsen and  Dollard  (1979),  in experiments  using  Norway  spruce  (Picea abies  L.  Karst),
observed that despite  increased leaching under the most acid treatment, there was no evidence
of  change  in the foliar cation  content.   Wood and Bormann (1977), using  Eastern  white pine
(Pinus strobus  L.),  also noted no significant changes in calcium, magnesium or potassium con-
tent  of  needles.   Tukey (1970)  states that increased  leaching  of  nutrients  from foliage can
accelerate nutrient uptake by plants.  No injury will occur to the plants as long as roots can
absorb nutrients  to  replace  those being leached; however, injury could occur if nutrients are
in  short supply.  To  date, the effects,  if any,  of the increased leaching of substances from
vegetation by acidic precipitation remain unclear.
     Some experimental  evidence  suggests  that acidic solutions affect the chlorophyll content
of  leaves  and the  rate  of photosynthesis.  Sheridan and  Rosenstreter  (1973)  reported marked
reduction of  photosynthesis  in a moss exposed  to  increasing  H  ion concentrations.  Sheridan
                                            7-86

-------
and Rosenstreter (1973), Ferenbaugh (1976), and Hindawi et al.  (1980) reported reduced chloro-
phyll content  as a  result  of tissue exposure  to  acid solutions.  In the  case  of Ferenbaugh
(1976),  however, the significant reductions in chlorophyll in the leaves of Phaseolus vulgaris
at pH 2.0  were associated with large areas  of  necrosis.   A significant aspect  of this study
was the  loss  of capacity by the plant  to  produce  carbohydrates.  The rate of  respiration in
these plants  showed  only a slight, but significant increase, while the rate of photosynthesis
at pH  2.0 increased nearly  fourfold  as determined by oxygen  evolution.  Ferenbaugh concluded
that due  to  a  reduction  in  biomass  accumulation and sugar  and  starch concentrations,  photo-
phosphorylation in the treated plants was in some way being uncoupled by the acidic solutions.
     Irving  (1979)  reported  a higher  chlorophyll  content and  an  increased rate  of  photo-
synthesis in field-grown soybeans exposed to simulated rain at pH 3.1.  She attributed the in-
creases to improved  nutrition  due  to the sulfur and nitrogen components of the simulated acid
rain overcoming any negative effects.
     Vegetation is commonly exposed to gaseous phytotoxicants such as ozone and sulfur dioxide
at  the  same  time as  acidic precipitation'.   Little information  is  available upon  which to
evaluate the potential  for determining the effects of the interaction of wet-and dry-deposited
pollutants on vegetation.  Preliminary studies by Shriner (1978b), Irving (1979), and Jacobson
et al.  (1980) suggest that  interactions may occur.   Irving (1979) found  that  simulated acid
precipitation  at  pH  of 3.1 tended to limit the decrease in photosynthesis observed when field
                                                                              3
-grown  soybeans were  exposed  17  times during  the  growing  season to 500 ug/m   (0.19 ppm) of
S09.   Shriner  (1978b),  however,  reported no significant interaction between multiple exposure
                                                                 3
to simulated  rain at pH 4.0 and four  SO,,  exposures (17860 M9/m . 3  ppm peak for 1 hr.) upon
                                                                            3
the growth of  bush beans.  Shriner (1978b)  also  exposed  plants to 290 ug/m  (0.15 ppm) ozone
(4 3-hour  exposures)  in between 4 weekly exposures to rainfall  of pH 4.0, and observed a sig-
nificant  growth reduction at  the  time of harvest,   Jacobson et  al.  (1980b),  using ope"n-top
exposure chambers with field-grown soybeans, compared growth and yield between three pH levels
of simulated  rain (pH  2.8, 3.4, and 4.0)  and two levels of ozone (<60  and <240 |jg/m  , <0.03
and £0.12  ppm).   Results demonstrated that  ozone  not  only  depressed both growth and yield of
soybeans  with all three rain treatments,  but that the depression was greatest  with the most
acidic  rain.   Ozone concentrations  equal  to or  greater than  those  used  in  the studies are
common  in most areas  of  the Northeastern  United States  where acidic  deposition  is a problem
(Jacobson et al. , 1980b); therefore, the potential for possible  ozone-acidic deposition inter-
actions is great.
     Shriner (1978a) studied the effect of acidic precipitation  on host-parasite interactions.
Simulated  acid rain  with a pH of 3.2 inhibited the development  of bean rust and production of
telia (a  stage in the rust  life cycle)  by the oak-leaf rust fungus Cronartiurn fusiforme.  It
also inhibited reproduction  of root-knot  nematodes and inhibited or stimulated development of
halo blight  of  bean  seedlings depending  on the time  in the disease cycle during which the
                                            7-87

-------
simulated acid rain was applied.  The effects which inhibited disease development could result
in a net  benefit  to plant health.  Shriner (1976, 1980) also observed that root nodulation by
Rhizobium on common  Beans  and soybeans was inhibited by the simulated acid rain, suggesting a
potential for reduced nitrogen fixation by legumes so affected.
     Plants such  as  mosses and lichens are particularly sensitive to changes in precipitation
chemistry because many of their nutrient requirements are obtained directly through precipita-
tion.   These plant forms are typically absent from regions with high chronic S0? air pollution
and may be  affected by acidic  precipitation (Nieboer  et al.,  1976;  Denison et  a!.,  1977;
Sheridan and Rosenstreter,  1973).  Gorham (1976) and Giddings and Galloway (1976) have written
reviews  concerning this  problem.  Most  investigations  on the  effects  of air  pollution on
epiphytes have dealt with gaseous pollutants.  Very few studies have considered acidic precipi-
tation.  Denison  et  al.  (1977), however, did observe  that the  nitrogen-fixing ability of the
epiphytic lichen Lobaria oregana was reduced when treated with simulated rainfall with a pH of
4.0 and below.    Investigations  concerning the  effects  of acidic  precipitation on epiphytic
microbial populations are very few (Abrahamsen and Do!lard, 1979).
     Limited fertilization could occur  in the bracken fern  Pteridium  aquilinum under condi-
tions of acidic precipitation (pH and sulfate concentrations) that prevail in the northeastern
United States.   Evans and Bozzone (1977), using buffered solutions to simulate acidic precipi-
tation, observed  that  flagellar movement of sperm was reduced at pH levels below 5.8,  Ferti-
lization was reduced after exposure to pH's below 4.2.  Sporophyte production was also reduced
by 50  percent  at  pH levels below 4.2  when compared to 5.8.  Addition  of sulfate as sulfuric
acid (86 mM) to the buffered solutions decreased fertilization at least 50 percent at all pH
values observed.  In another study, Evans and Bozzone (1978) observed that both sperm motility
and fertilization  in  gametophytes of Pteridium aquilinum were reduced when anions of sulfate,
nitrate, and chloride were added to buffered solutions.
     Sulfur and nitrogen in precipitation have been shown to play an important role in vegeta-
tional  response to  acidic deposition.  Jacobson  et al.  (1980b)  investigated the  impact of
simulated acidic  rain on  the  growth  of lettuce  at  acidities of pH 5.7 and 3.2.   At pH 3.2,
solutions with  NO.,:SO. mass ratios of  20:1,  2:1, and 1:7.5 were compared.  For those growth
parameters  (root  dry weight and apical leaf dry weight) that responded to the treatments, the
results of  the  high nitrate concentrations applied  at pH 3.2 could not be distinguished from
the control  treatment at  pH 5.7.  The  effects, however, were  significantly  less  than those
obtained from the low nitrogen, high sulfur treatment.  These observations suggest that sulfur
was possibly a limiting factor in the  nutrition of  these plants, with the  result that the
plant  response  to sulfur overwhelmed the hydrogen  ion effect.   Other studies also have cited
the beneficial  effects of  simulated acidic precipitation.   Irving  and  Miller  (1978) observed
that  an acidic  simulant  had  a  positive  effect on  productivity of field-grown  soybeans as
reflected by seed weight.  Increased growth was attributed to a fertilizing effect from sulfur
                                            7-88

-------
and nitrogen delaying  senescence.   Irving and Miller (1978),  in  the same study, also exposed
soybeans to SO, and acidic precipitation.   No visible injury was apparent in any of the plots;
however, a histological  study  revealed significant increases  in  the number of dead mesophyll
cells in all plots  when compared to the  control.   The  proportion of dead  mesophyll  cells  of
plants  exposed  to acid  rain  and SCL  combined was more  than additive  when compared  to  the
effects of each  taken  singly.   Wood and  Bormann  (1977)  reported  an increase in needle length
and the weight of seedlings of Eastern white pine with increasing acidity of simulated precipi-
tation  where  sulfuric  and  nitric  acid were  used to  acidify the mist.   Increased growth  was
attributed to increased  NO,  application.  Abrahamson and Dollard  (1979) also presented data
suggesting positive growth responses in forest tree species resulting from nitrogen and sulfur
in simulated rain.  Simulated acidic precipitation was observed to increase the growth of Scots
pine saplings in experiments conducted in Norway.   Saplings in plots watered with acid rain of
pH 3.0, 2.5,  and 2.0 grew more than the control plots.  The application of acid rain increased
the nitrogen and sulfur content of the needles.  As the acidity of the artificial rain was  ad-
justed  using sulfuric  acid only,  the increased growth was  probably due to increased nitrogen
mineralization and  uptake.  Turner  and Lambert (1980) reported evidence indicating a positive
growth  response  in Monterey pine  from the  deposition  of sulfur in ambient precipitation  in
Australia.
     Acidifying forest soils that  are already acid by acidic  precipitation or air pollutants
is a  slow  process.   Growth effects probably  could  not be detected for a long time.  To iden-
tify the possible  effects  of  acidification on poor pine  forests,  Tamm et al. (1977) conducted
experiments using 50 kg and 100 kg of sulfur per hectare  as dilute sulfuric acid (0.4 percent)
applied annually with and without NPK (nitrogen, phosphorous, potassium) fertilizer.  Nitrogen
was found  to be  the limiting factor  at both experimental  sites.   Acidification  produced  no
observable influence on  tree  growth.   Lysimeter and  soil  incubation experiments conducted at
the same time as the experiments described above suggest that even moderate additions of sul-
furic acid or sulfur to soil affect soil biological processes, particularly nitrogen turnover.
The soil incubation studies indicated that additions of sulfuric acid increased the amount of
mineral nitrogen but lowered the amount of nitrate.
     Soil  fertility may increase as a  result of acidic precipitation as  nitrate  and sulfate
ions, common  components of chemical  fertilizers,  are deposited; however,  the advantages  of
such additions are  possibly short-lived as depletion of  nutrient cations through accelerated
leaching could eventually  retard  growth (Wood, 1975).  Laboratory  investigations  by Overrein
(1972)  have  demonstrated that  leaching of potassium, magnesium, and  calcium,  all important
plant nutrients,  is accelerated  by increased acidity of rain.  Field studies in Sweden corre-
late decreases in soil  pH with increased additions of acid (Oden et al., 1972).
     Major uncertainty in estimating effects of acid rain on forest productivity is the capac-
ity of  forest  soils to buffer against  leaching by hydrogen ions.   Forest  canopies have been
                                            7-89

-------
found to  filter  90 percent of the  hydrogen  ions  from rain (pH 4.0)  falling on the landscape
during the  growing season (Eaton et  al. ,  1973).   As a result, solutions  reaching  the forest
floor  are less  acidic  (pH  5.0).   Mayer and  Ullrich (1977),  however,  point out  that  their
studies suggest that for most elements the addition by precipitation (wetfall plus dryfall) to
the  soil  beneath the  tree canopy  is  considerably  larger  than that  by  precipitation to the
canopy surface  as measured  by  rain  gauges  on  a non-forested area.   The  leaching  of metabo-
lites, mainly  from leaf  surfaces,  and  the  washing  out  from leaves, branches,  and stems of
airborne  particles  and  atmospheric aerosols  intercepted by  trees  from the  atmosphere,  are
suggested as the main reasons for the mineral increase.
     Forest  ecosystems  are complicated  biological  organizations.   Acidic  precipitation will
cause some components  within the ecosystem to  respond even though it is not possible at pre-
sent to evaluate  the  changes that occur.  The impact of the changes on the ecosystem can only
be determined with certainty after the passage of a long period of time.
7.3.2.3   Effects  on Human  Health—One  effect of acidification that is potentially  of concern
to human health is the possible contamination by toxic metals of edible fish and of water sup-
plies.  Studies  in  Sweden (Landner and Larsson, 1975; Turk and Peters, 1977), Canada (Tomlin-
son, 1979; Brouzes  et  al. , 1977), and  the  United States (Tomlinson,   1979) have revealed high
mercury concentrations  in  fish  from acidified  regions.   Methylation  of  mercury to monomethyl
mercury occurs  at low pH while  dimethyl  mercury  forms.at  higher  pH  (Fagestrb'm and Jernelb'v,
1972).  Monomethyl  mercury in the water  passing  through the gills of fish  reacts  with thiol
groups in the  hemoglobin of the  blood  and  is  then transferred to the muscle.  As methyl mer-
cury is eliminated very slowly from fish, it accumulates with age.
     Tomlinson  (1979)  reports  that  in  the Bell  River area  of  Canada  precipitation  is  the
source of mercury.   Both  methyl  mercury and  inorganic mercury were  found  in precipitation.
                 <                                                         •
The source of mercury in snow and rain was not known at the time of the study.
     Zinc, manganese,  and aluminum  concentrations  also increase  as the acidity  of lakes in-
creases (Schofield, 1976b).   The ingestion of fish contaminated by these metals is a distinct
possibility.
     Another human health aspect  is the possibility that, as drinking-water reservoirs acidify
owing to  acidic  precipitation,  the increased concentrations  of metals may exceed the public-
health limits.  The  increased metal concentrations  in  drinking water are caused by increased
watershed weathering and,  possibly  more importantly, increased leaching of metals from house-
hold plumbing.   Indeed,  in New York State, water from the Hinckley Reservoir has acidified to
such an extent  that "lead concentrations in water  in contact with household plumbing systems
exceed the  maximum  levels  for   human  use  recommended  by  the New York State  Department of
Health" (Turk and Peters, 1977).   The lead and copper concentrations in pipes which have stood
over night (U) and those in which the" water was used (F) are depicted  in Table 7-11.
                                            7-90

-------
   TABLE 7-11.   LEAD AND COPPER  CONCENTRATION  AND pH OF WATER  FROM  PIPES
   CARRYING OUTFLOW  FROM HINCKLEY  BASIN  AND HANNS AND STEELE  CREEK BASIN,
                        NEAR AMSTERDAM, NEW YORK

Collection site
and date
Hinckley Dam
Nov. 21, 1974
Nov. 21, 1974
Nov. 7, 1974
Nov. 7, 1974
Oct. I, 1974
Oct. 1, 1974
Aug. 15, 1974
Aug. 15, 1974
Amsterdam
Jan. 6, 1975
Jan. 6, 1975
Pipe ,
condition

U
F
U
F
U
F
U
F

U
F
Copper
(Mi/D

600
20
460
37
570
30
760
40

2900
80
Lead
(M9/D

66
2
40
6
52
5
88
2

240
3
pH


7.4
6.3
6.3
6.8
7.1
6.3
6.3

4.5
5.0

 U,   unf lushed,   (water   stands
Source:   Turk and Peters (1979).
pipes   all   night);   F,   flushed.
                                7-81

-------
7.3.2.4   Effects ofAcidic Precipitation on Materials—Acidic precipitation  can  damage  the
abiotic as well  as  the biotic components of an ecosystem.   Of particular concern in this sec-
tion are the deteriorative effects of acidic precipitation on materials and cultural artifacts
of manmade  ecosystems.   At  present,  in most  areas, the dominant factor  in the  formation  of
acidic precipitation is sulfur, usually as sulfur dioxide (Likens, 1976; Cowling and Dochinger,
1978).  Because  of  this fact, it is  difficult to isolate the effect  of  acidic precipitation
from changes induced  by sulfur pollution in general.  (The  effects  of sulfur oxides on mate-
rials are discussed in Chapter 10.)  High acidity promotes corrosion  because the hydrogen ions
act as a sink for the electrons liberated during the critical corrosion process (Nriagu, 1978).
Precipitation as  rain  affects corrosion by forming a  layer  of moisture on the surface of the
material and by  adding hydrogen (H+) and sulfate (S02   ) ions as corrosion stimulators.  Rain
also washes out the sulfates deposited during dry deposition and thus serves a useful function
by removing the  sulfate and stopping corrosion (Kucera,  1976).  Rain plays a critical role in
the corrosive process because in areas where dry deposition predominates,  the washing effect is
greatest, while  in areas  where  the  dry and wet deposition processes  are  roughly  equal,  the
corrosive effect is greater  (Kucera, 1976).  The  corrosion effect,  particularly  of certain
metals, in  areas where  the pH of precipitation  is very low may be greatly enhanced by that
precipitation  (Kucera,  1976).   In  a  Swedish study,  the  sulfur content of  precipitation,
                  o
expressed as meq/m  per year, was found to correlate closely with the corrosion rate of steel.
The metals most likely to be corroded by precipitation  with a low pH  are those whose corrosion
resistance may be ascribed to a protective layer of basic carbonates, sulfates, or oxides, such
as those  used  on zinc or copper.  The decrease in pH of rainwater to 4.0 or lower may accele-
rate the dissolution of the protective coatings (Kucera,  1976).
     Materials reported to be  affected by  acidic  precipitation, in  addition  to steel, are:
copper materials, linseed oil, alkyd paints on wood, antirust paints  on steel,  limestone, sand-
stone, concrete, and both cement-lime and lime plaster (Cowling and Dochinger,  1978).
     Stone is one of the oldest building materials used by man and has traditionally been con-
sidered one of the  most durable because structures such as the pyramids, which have survived
since  antiquity,  are made of stone.  What  is  usually  forgotten  is  that  the  structures built
with stone that was not durable have  long since disappeared (Sereda,  1977).
     Atmospheric  sulfur compounds (mainly  sulfur dioxide, with  subsidiary amounts of sulfur
trioxide  and ammonium sulfate) react with the carbonates  in limestone and dolomites, calcar-
eous  sandstone and  mortars to form calcium  sulfate  (gypsum).   The results of these reactions
are  blistering,  scaling,  and loss of surface cohesion which, in turn, induces similar effects
in neighboring materials not  in themselves susceptible to direct attack (Sereda, 1977).
     Sulfates  have  been implicated by Winkler (1966)  as very important in the disintegration
of stone.  The surface flaking on the Egyptian granite obelisk (Cleopatra's Needle) in Central
Park,  New York  is  cited as  an  example.  The  deterioration occurred within two  years of its
erection  in 1880.
                                            7-92

-------
     A classic  example of  the  effects of  the  changing chemical climate on  the  stability of
stone is the deterioration of the Madonna at Herten Castle, near Recklinghausen, Westphalia in
Germany.   The sculpture of  porous Baumberg sandstone was  erected  in 1702.   Pictures taken of
the Madonna in  1908 shows slight-to-moderate damage during the first 206 years.  The features
of the  Madonna—eyes,  nose,  mouth  and hair—are  readily discernable.  In pictures  taken in
1969 after 267 years, no features are visible (Cowling and Dochinger, 1978).
     It is not certain in what form sulfur is absorbed into stone,  as a gas (SO,) forming sul-
furous and/or sulfuric  acid or  whether it is deposited in rain.  Rain and hoarfrost both con-
tain sulfur compounds.   Schaffer (1932) compared the sulfate ion in both rain and hoarfrost at
Heading!ey, Leeds,  England  in 1932  (Table 7-12) and  showed  that the content of hoarfrost was
approximately 7  times  greater than  rain.   Wet stone surfaces unquestionably increase the con-
densation  or  absorption of sulfates.   Stonework kept dry and  shielded  from  rain,  condensing
dew, or hoarfrost  will  be damaged less by S0? pollution than stone surfaces which are exposed
(Sereda, 1977).
           TABLE 7-12.  COMPOSITION OF RAIN AND HOARFROST AT HEADINGLEY, LEEDS
                                 Average rain                     Hoarfrost
                               parts per million              parts per million
Suspended matter
Tar
Ash
Acidity
SO, as sulfur
SO- as sulfur
Total sulphur
Chlorine
Nitrogen as NH,
Nitrogen as N-Og
Nitrogen as albuminoid
115
15
28
1.9
22
5.7
27.7
7.3
1.98
0.196
0.434
4620
158
67
102.9
148
41.0
189.0
94.6
8.57
0.0
1.618

       Source:  Adapted from Schaffer (1932)
                                            7-93

-------
     Acid  rain  may leach ions from  stonework  just as acidic runoff and  ground  water leaches
ions from  soils  or bedrock; however, at  the  present time it is not possible to attribute the
deleterious effects of atmospheric sulfur pollution to specific compounds.
     Hicrobial  action  (an  indirect  result of  sulfur deposition) can also  contribute  to the
deterioration of stone surfaces.  Tiano  et  al.  (1975) isolated large numbers  (250  to  20,000
cells per  gram) of sulfate-reducing  bacteria  from the stones of two  historical  buildings of
Florence, Italy.  The majority of the bacteria belonged to the genus Thiobacillus.  This genus
of  chemosynthetic  aerobic microorganisms oxidize  sulfide, elemental sulfur,  and thiosulphate
to  sulfate to  obtain energy  (Andersson, 1978).   Limestone  buildings,  and  particularly the
mortar used in the construction of brick and stone buildings, are susceptible to deterioration
when  conditions permit  Thiobacillus to  convert  reduced forms  of  sulfur to  sulfuric  acid.
Sulfate  in acidic  precipitation  as well as other sulfur compounds deposited in dry deposition
permit the formation  of  sulfur compounds utilizable by microorganisms.  (For more information
concerning the effects of sulfur oxides on materials, please consult Chapter 10.)
7.4  ASSESSMENT OF SENSITIVE AREAS
7.4.1  Aquatic Ecosystems
     Why do some lakes  become acidified  by  acidic precipitation  and others not?  What  deter-
mines susceptibility?  Are terrestrial ecosystems likely to be susceptible; if so, which ones?
     The sensitivity of  lakes  to acidification is determined by:   (1)  the acidity of both wet
deposition (precipitation)  and dry  deposition;  (2)  the  hydrology  of the lake;  (3)  the soil
system,   geology, and  canopy effects; (4) the  surface water.   Given acidic precipitation, the
soil system and associated canopy effects are most important.  The hydrology of lakes includes
the sources,  amounts,  and pathways of water entering and leaving a lake.   The capability of a
lake and its  drainage  basin to neutralize acidic contributions as well as the mineral content
of  its surface  water is  largely governed  by  the composition of the bedrock of the watershed.
The chemical  weatheripg  of  the watershed strongly influences the salinity (ionic composition)
and the  alkalinity of  the surface water  of  a lake (Wetzel, 1975; Wright  and Gjessing,  1976;
Wright and Henriksen,  1978).   The cation exchange  capacity  and  weathering rate of the  water-
shed and the alkalinity of the surface water determine the ability of the system to neutralize
the acidity of precipitation.
     Lakes  vulnerable to acidic  precipitation  have  been shown  to  have watersheds  the geo-
logical   compositions  of which  are highly  resistant  to  chemical   weathering  (Wright  and
Gjessing,  1976;  Galloway and  Cowling,  1978;  Wright and Henriksen, 1978).   In  addition, the
watersheds of the vulnerable  lakes  usually  have  thin, poor soils  and  are  poorly vegetated.
The cation exchange  capacity of  such soils is low and, therefore, their buffering capacity is
low (Schofield, 1979; Wright and Henriksen, 1978).
     Wright and Henriksen  (1978)  point out  that  the chemistry  of Norwegian lakes  could be
accounted for primarily on the basis of bedrock geology.  They examined 155 lakes and observed
                                            7-94

-------
that 59 of  them  lay in granite or felsic gneiss basins.   Water in these lakes was low in most
major ions and had low electrical conductivity.   The fewer the minerals in water the lower its
conductivity (Wetzel, 1975).   The  waters in the lakes surveyed were "among the softest waters
in  the  world"  (Wright and  Henriksen,  1978).   Sedimentary  rocks generally  weather  readily,
whereas  igneous  rocks are  highly  resistant.   The  Adirondacks,  as  pointed out  by  Schofield
(1976a; 1979),  have granite  bedrock with much of  the-araa covered with  a mantle  of mixed
gneisses.    Shallow soils  predominate  in  the  area.   Thus,  these  areas  are  susceptible  to
acidification.
     Limestone terrains,  on the other hand, are capable of buffering intense concentrations of
acids.   Glacially derived sediment has been found to be more important than bedrock in assimi-
lating acidic precipitation in the Canadian Shield area (Kramer, 1976).   The detailed mineral-
ology of  the unconsolidated  post-glacial  cover is  the most  important  parameter  in assessing
the  H   ion assimilation  of acid  precipitation in  non-calcareous terrain.   Knowledge of the
complete  surface  and subsurface hydrol,ogy is  required  as lower  horizons  may  be calcareous,
whereas surface  deposits may  be non-calcareous  (Kramer  1976).  Generally,  however,  bedrock
geology is the best predictor of the sensitivity of aquatic ecosystems to acidic precipitation
(Hendrey et al., 1980b).
     Areas with aquatic  ecosystems that have the potential for being sensitive to acidic pre-
cipitation  are  shown in  Figure  7-27.   In Figure 7-27,  the shaded  areas on  the  map  indicate
that the  bedrock  is composed of igneous or metamorphic rock while in the unshaded areas it is
of  calcareous  or  sedimentary  rock.   Metamorphic and igneous  bedrock weathers  slowly; there-
fore,  lakes  in  areas  with this type  of  bedrock would  be  expected to be  dilute  and of low
alkalinity  [<0.5  meq  HC03/liter  (Galloway and Cowling, 1978)].  Galloway and Cowling verified
this  hypothesis  by compiling alkalinity  data.   The  lakes  having low alkalinity  existed  in
regions  having  igneous  and metamorphic rock  (Galloway  and  Cowling,  1978).  Hendrey  et al.
(1980b) have  developed new bedrock  geology  maps  of the eastern  United States  for predicting
areas  which might be impacted  by  acidic  precipitation.   The  new  maps permit  much greater
resolution for detecting sensitivity than has been previously available for the region.
     Henriksen  (1979)  has  developed a  lake acidification  "indicator  model" using pH-calcium
and  calcium-alkalinity relationships as an indicator for  determining  increased surface water
acidification.   The  indicator is based  on the  observation that in pristine lake environments
(e.g., Northwest  Norway  or the Experimental Lakes  area  in northwest Ontario, Canada) calcium
is  accompanied  by a proportional  amount of bicarbonate because  carbonic acid  is the primary
chemical weathering agent.  "The pH-calcium relationship found for such  regions is thus defined
as the reference  level for  unacidified  lakes.  Acidified lakes  (e.g., Southeast Norway and the
Adirondack region) will exhibit lower pH or lower alkalinity than the reference lakes, at com-
parable  calcium  levels,   due  to  the  replacement  of bicarbonate  by  strong  acid anions.  (See
Section 7.3.1.1).
                                            7-95

-------
Figure 7-27.  Regions in North America with lakes that are sensitive to acidification
by acid precipitation by virtue of their-underlying bedrock characteristics.

Source: Galloway and Cowling (1978).
                                    7-96

-------
     A report by  Hendrey  et al.  (1980b) compared pre-1970 data with post-1975 data.   A marked
decline in both alkalinity and pH was noted in waters of North Carolina and New Hampshire sen-
sitive to acidification.   In the former, pH and alkalinity have decreased in 80 percent of the
streams tested and,  in the latter, pH has decreased in 90 percent of the streams tested since
1949.  These areas  are predicted to be sensitive  by geological  mapping on the basis of their
earlier alkalinity  values.   Detailed  county  by county  maps of  other states in  the eastern
United States suggest the sensitivity of these regions to acidic precipitation.
     Though bedrock  geology  generally  is  a good predictor of the susceptibility of an area to
acidification due to acidic precipitation, other factors also have an influence.   Florida, for
example,  is underlaid by  highly calcareous and phosphate  rock,  suggesting that acidification
of  lakes  and  streams is  highly  unlikely.   Many  of  the soils,  however,   (particularly  in
northern  Florida)  are very  mature and have  been  highly  leached of  calcium carbonate;  as  a
result, some lakes  in which groundwater  inflow is minimal  have become acidified (Hendrey and
Brezonik,  1980).  Conversely,  there are areas in Maine with granitic bedrock where lakes have
not become acidified, despite receiving precipitation with an average pH of approximately 4.3,
because the  drainage basins contain lime-bearing  till  and marine clay  (Davis et  al.,  1978).
Small amounts of  limestone in a drainage  basin exert a strong influence  on  water quality in
terrain which  would otherwise  be  vulnerable to  acidification.   Soils  in Maine in  the areas
where the  pH of  lakes  has  decreased  due to acidic precipitation  are  immature,  coarse, and
shallow,  are  derived largely  from granitic material,  and commonly  have a  low capacity for
assimilating hydrogen ions  from  leachate  and surface runoff in lake watersheds (Davis et al.,
1978).  The occurrence of limestone outcroppings in the Adirondack Mountains of New York State
is highly correlated with  lake pH levels.  The occurrence of limestone apparently  counteracts
any effects of acidic precipitation.  Consequently, when predicting vulnerability of a partic-
ular  region to  acidification,  a careful classification of rock mixtures should be  made.  Rock
formations should be classified  according to their potential buffering capacity, and the type
of soil overlying  the  formations should be noted.   Local  variations in bedrock and soils are
very  important in explaining variations in acidification among lakes within an area.
7.4.2  Terrestrial Ecosystems
      Predicting the sensitivity of terrestrial ecosystems to acidic precipitation is much more
difficult than  for  aquatic  ecosystems.   With aquatic  ecosystems,   it is  possible to compare
affected and  unaffected  ecosystems  and to note  where the changes  have occurred.   With ter-
restrial  ecosystems, comparisons are difficult  to make because the effects of acidic precipi-
tation have  been difficult  to detect.   Therefore,  predictions regarding  the sensitivity of
terrestrial ecosystems must,  as  much as possible,  use the data which link the two  ecosystems,
i.e., data on  bedrock geology.  Since, in most regions  of the world, bedrock is  not exposed
but  is  covered with  soil,  it  is  the sensitivity of different  types  of soil which must be
assessed.   Therefore, the first step is to define "sensitivity" as it is used here  in relation
to soils and acidic precipitation.   Sensitivity of soils to acidification alone,  though it may
                                            7.97

-------
be the most  important long-term effect, is too narrow a concept.  Soils influence the quality
of waters in associated streams and lakes and may be-changed in ways other than simple pH-base
saturation relationships, e.g.,  microbiological  populations of  the  surface  layers  or accele-
rated loss  of aluminum  by  leaching.    Therefore,  criteria need to  be  used  that would relate
soil "sensitivity"  to any  important  change brought  about in  the  local ecosystem  by acidic
precipitation (McFee, 1980).
     All soils are  not equally susceptible to acidification.   Sensitivity  to leaching and to
loss of buffering capacity  varies according to the  type of parent material from which a soil
is derived.   Buffering  capacity is greatest in soils  derived from  sedimentary rocks, especi-
ally those containing carbonates,  and least in soils derived from hard crystalline rocks such
as granites and quartzites (Gorham, 1958).   Soil  buffering capacity varies widely in different
regions of the country.  Unfortunately, many of the areas now receiving the most acidic preci-
pitation are also those with relatively low natural buffering capacities.
     The buffering capacity of soil depends on mineralogy, texture, structure, organic matter,
pH, base saturation,  salt content, and soil permeability.  Above a pH of 5.5 virtually all of
the H   ions, irrespective  of source,  are  retained by  ion exchange and chemical weathering.
Below pH 5.5, the retention of the H  ion decreases with the soil pH in a manner determined by
the composition  of  the  soil  (Donahue et a!., 1977).   With a successive drop in the soil pH
below 5.0,  an increasing proportion  of hydrogen  ions  (H ) and deposited  sulfuric  acid will
pass through the soil  and  acidify runoff  water  (Donahue et a!., 1977).   The sensitivity of
different soils based on pH, texture,  and calcite content is summarized in Table 7-13.
     Soils are the  most stable component of a terrestrial ecosystem.  Any changes which occur
in  this  component  would  probably have  far-reaching effects.   McFee  (1980)  has  listed four
           TABLE 7-1-3.  THE SENSITIVITY TO ACID PRECIPITATION BASED ON:  BUFFERING
          CAPACITY AGAINST pH-CHANGE, RETENTION OF H , AND ADVERSE EFFECTS ON SOILS

Noncalcareous

Buffering
H retention
Adverse
effects
Calcareous
soils
Very high
Maximal
None
clays
pH > 6
High
Great
Moderate
sandy soils
pH > 6
Low
Great
Considerable
Cultivated
soils
pH > 5
High
Great
None -
slight
Acid
soils
pH < 5
Moderate
Slight
Slight

     Source:  Wiklander (1979).
                                            7-98

-------
parameters which are of importance in estimating the sensitivity of soils to acidic precipita-
tion.  They are:
     1.   the total  buffering  or cation exchange capacity,  which  is provided
          primarily by clay and soil organic matter.
     2.   the base  saturation  of that  exchange capacity, which can be esti-
          mated from the pH of  the soil.
     3.   the management  system  'imposed' on  the soil;»is«.>it  eultivated  and
          amended with fertilizers  or  lime  or renewed by flooding or by other
          additions?
     4.   the presence or absence of carbonates in the soil profile.
     McFee  (1980)  mapped  soils  of the  Eastern United  States taking  into account  several
factors, e.g.,  the  sensitivity to acidic precipitation, effects of cultivation (Figure 7-28).
The  areas  containing most of  the soils potentially sensitive  to  acidic precipitation  are in
the  upper  Coastal  Plain  and  Piedmont  regions  of  the  southeast,   along  the  Appalachian
Highlands, through the east central  and northeastern areas, and in the Adirondack Mountains of
New York  (McFee,  1980).   The  current  limited  state of  knowledge  regarding the  effects of
acidic precipitation  on soils  makes a  more definitive  judgment of the location of areas with
the most sensitive soils difficult at the present time.
     The capacity  of  soils  to  absorb and retain am'ons, also important in determining whether
soils will become acidified,  was not discussed by McFee (1980).  The capacity for anion absorp-
tion is  great  in  soils rich in hydrated oxides of aluminum (A!) and iron (Fe).  Reduced leach-
ing of salt cations is of great significance not only in helping to prevent soil acidification
but in geochemical circulation  of nutrients, fertilization in agriculture and preventing water
pollution (Wiklander, 1980; Johnson et al.  1980; Johnson, 1980).  (See Section 7.3.2.1.)  This
parameter, as well  as those  listed by Me Fee (1980), should be used in determining the sensi-
tivity of soils to acidification by both wet and dry deposition.
7,5  SUMMARY
     Occurrence of  acidic  precipitation (rain and snow) in many regions of the United States,
Canada,  and  Scandanavia has  been implicated in  the disappearance  or reduction of fish, other
animals,  and plant  life   in  ponds,  lakes,  and  streams.   In  addition, acidic precipitation
appears  to possess  the potential for impoverishing  sensitive  soils, degrading natural areas,
injuring forests, and damaging  stone monuments and buildings.
     Sulfur  and nitrogen  oxides, emitted  through  the  combustion  of  fossil  fuels,  have been
implicated as the chief contributors to the acidification of precipitation.  The fate of sul-
fur  and  nitrogen  oxides,  as  well as other  pollutants emitted  into the atmosphere, depends on
their dispersion, transport,  transformation, and deposition.  Emissions from automobiles occur
at ground level, those from electric power generators from smoke stacks 300 meters (1000 feet)
or more  in height.  Transport and transformation of the sulfur  and nitrogen oxides are in part
associated with the height at which they are emitted.  The greater the height, the greater the
likelihood of  a  longer residence time  in  the  atmosphere  and a  greater  opportunity  for the
                                            7-99

-------
                                                REGIONS WITH SIGNIFICANT
                                                AREAS OF SOILS THAT ARE
                                                [~| NON SENSITIVE
                                                    SLIGHTLY SENSITIVE
                                                    SENSITIVE
                                                WITHIN THE EASTERN U.S.
Figure 7-28. Soils of the eastern united states sensitive to acid rainfall.
Source: McFee (1980).
                                    7-100

-------
chemical transformation  of the  oxides  to sulfates,  nitrates or other  compounds.   Ozone and
other  photochemical  oxidants  are believed  to  be involved  in the  chemical  transformations.
Because of long  range  transport, acidic precipitation  in  a  particular state or region can be
the result of  emissions  from sources in states  or regions hundreds of miles away rather than
local  sources.   To date the complex  nature  of  the chemical  transformation  processes  has not
made possible the demonstration of a direct cause-and-effect relationship between emissions of
sulfur and nitrogen oxides and the acidity of precipitation.
     Natural  emissions of  sulfur and nitrogen compounds are also involved in the formation of
acidic  precipitation;   however,  in  industrialized  regions  anthropogenic  emissions  exceed
natural emissions.
     Precipitation  is   arbitrarily  defined  as  being  acidic if  its pH is  less  than  5.6.
Currently  the  acidity of  precipitation in  the  Northeastern  United States, the  region  most
severely impacted, ranges  from pH 3.0 to 5.0.  Precipitation episodes with a pH as low as 3.0
have been  reported for other regions of the  United  States.   The pH of precipitation can vary
from event to  event,  from season to  season  and  from geographical area  to  geographical  area.
     The  impact  of  acidic  precipitation  on aquatic and terristrial  ecosystems is  not the
result of  a  single or several precipitation  events,  but the result of continued additions of
acids  or acidifying  substances over time.   Wet  deposition of acidic substances on freshwater
lakes, streams,  and  natural  land areas is only  part  of the problem.  Acidic substances exist
in  gases  and particulate  matter transferred into the  lakes, streams, and  land  areas by dry
deposition.  Therefore all the observed biological effects should not be attributed to acidic
precipitation alone.
     Sensitivity of a lake to acidification depends on the acidity of both wet and dry deposi-
tion,  the  soil  system  of the drainage  basin, canopy  effects of ground cover and the composi-
tion of the watershed bedrock.
     An extremely  close  mutual  relationship exists between the chemistries of the environment
and of living  organisms.   There is a continuing exchange of nutrients and of energy.  The two
are  closely  intertwined responses.   There is no  action  without a  reaction.   Ecosystems can
respond to environmental changes or perturbations only through the response of the populations
of  organisms of  which  they are composed.   Species of organisms sensitive to specific environ-
mental  changes are removed.   Therefore,   the capacity  of an ecosystem  to  maintain internal
stability  is determined  by the ability of individual organisms  to adjust their physiology or
behavior to environmental change.  The success with which an organism copes with environmental
changes is determined  by its ability to yield reproducing offspring.  The size and success of
a population depends  upon the collective ability of organisms to reproduce and maintain their
numbers  in a  particular environment.  Those organisms that  adjust best contribute  most to
future generations because they have the greatest number of progeny in the population.
                                            7-101

-------
     The capacity of  organisms  to withstand injury from  weather  extremes,  pesticides, acidic
deposition,  or  polluted air  follows the  principle  of limiting  factors.   According  to  this
principle,  for  each  physical   factor  in  the  environment  there  exists  for each  organism  a
minimum  and maximum  limit beyond which  no members  of  a particular  species can  survive.
Either too much or too little of a factor such as heat, light, water, or minerals (even though
they are necessary for life) can jeopardize the survival of an individual and in extreme cases
a species.  When  one  limiting factor is removed another takes its place.  The range of toler-
ance of an  organism  may be broad for one factor, narrow for another.  The tolerance limit for
each species is determined by its genetic makeup  and  varies from species to  species  for the
same reason.   The range of tolerance also  varies depending  on  the age, stage of growth or
growth form of an organism.  Limiting factors  are,  therefore, factors which,  when scarce or
overabundant, limit the growth,  reproduction and/or distribution of an organism.  The increas-
ing acidity of water  in lakes and streams appears to  be such a  factor.   Significant  changes
that have occurred in aquatic ecosystems with increasing acidity include the following:
        1.   Fish populations are reduced or eliminated.
        2.   Bacterial  decomposition  is  reduced and fungi  may dominate saprotrophic communi-
             ties.   Organic  debris  accumulates  rapidly,  tying  up  nutrients, and  limiting
             nutrient mineralization and cycling.
        3.   Species diversity  and total  numbers of species of aquatic plants and animals are
             reduced.   Acid-tolerant species dominate.
        4.   Phytoplankton productivity may  be  reduced due to changes in nutrient cycling and
             nutrient limitations.
        5.   Biomass and total  productivity  of benthic macrophytes and algae may increase due
             partially to increased lake transparency.
        6.   Number^ and biomass  of  herbivorous invertebrates decline.   Tolerant invertebrate
             species, e.g., air-breathing  bugs  (water-boatmen,  back-swimmers, water striders)
             may become abundant primarily due to reduced fish predation.
        7.   Changes in community structure occur at all trophic levels.

     Studies indicate that pH levels between 6.0 and 5.0 inhibit reproduction of many species
of  aquatic  organisms.   Fish  populations  become seriously  affected at  a  pH  lower than  5.0.
     Disappearance of fish from lakes and streams follows  two  general  patterns.  One results
from sudden  short-term  shifts  in pH, the  other  arises from a long-term decrease in the pH of
the water.   A  major  injection of acids and other soluble substances occurs when polluted snow
melts during warm periods in winter or early spring.   Fish  kills are a dramatic consequence of
such episodic injections.
     Long-term  increases   in  acidity interfere  with  reproduction  and spawning,  producing  a
decrease in population density and a shift in size and age of the population to one consisting
                                            7-102

-------
primarily of  larger and  older  fish.   Effects  on  yield often are not  recognizable  until  the
population is  close to extinction; this  is particularly true for late-maturing  species  with
long lives.   Even  relatively  small  increases (5 to  50  percent)  in mortality of fish eggs and
fry can decrease yield and bring about extinction.
     Aluminum is mobilized  at low pH values.   Aluminum  may  be as important or more important
than pH  levels  as  factors leading to declining fish populations in acidified lakes.  Certain
aluminum  compounds in  the  water  upset  the osmoregulatory  function  of  the blood  in  fish.
Aluminum toxicity to aquatic biota other than fish has not been assessed.
     An  indirect  effect oF acidification  potentially of concern to  human  health  is possible
heavy metal  contamination of edible  fish  and  water supplies.  Studies in  Canada  and Sweden
reveal high mercury concentrations  in fish from acidified regions.   Lead and copper have been
found in plumbing systems with acidified  water,  and persons drinking  the  water  could suffer
from lead poisoning.
     Acidic  precipitation may indirectly  influence terrestrial plant productivity by altering
the  supply  and  availability  of  soil  nutrients.  Acidification  increases leaching  of  plant
nutrients (such as calcium, magnesium,  potassium,  iron, and manganese), increases the rate of
weathering of most minerals,  and  also makes phosphorus  less  available  to plants.   Acidifica-
tion also decreases the rate  of many soil microbiological processes such as nitrogen fixation
by  Rhizobium  bacteria on  legumes  and  by the free-living  Azotobacter, mineralization  of
nitrogen from forest  litter,  nitrification of  ammonium  compounds, and  overall  decay rates of
forest floor litter.
     Plants   usually  take up  sulfur  in the  form of  sulfate  from the soil;  however,  they can
also take up  S0«  from the atmosphere through  their  leaves and utilize it  as a sulfur source
for plant nutrition.   If soil sulfur is  low,  plants may obtain most of their required sulfur
from  the atmosphere.   Though  small amounts  of  S0? may be beneficial,  large  amounts and  high
frequency of uncontrolled applications can be detrimental in the long term.
     At  present,  there are no  documented observations or measurements of  changes  in natural
terrestrial  ecosystems that can be directly attributed to acidic precipitation.   This does not
necessarily  indicate   that  none  are  occurring.   The information  available on  vegetational
effects  is an  accumulation  of the results of a wide variety of controlled research approaches
largely  in the  laboratory,  using  in most instances some form of "simulated" acidic rain,  fre-
quently  dilute  sulfuric acid and/or  nitric acid.   The  simulated "acid rains"  have deposited
hydrogen  (H  ),  sulfate (SQ/~) and nitrate (NO,) ions on vegetation  and  have caused necrotic
lesions  in a wide variety of plants species under greenhouse and laboratory conditions.   Such
results  must be interpreted with caution, however, because the growth and morphology of leaves
under  greenhouse  conditions  are   often  atypical  of field  conditions.  Based  on  laboratory
studies,  sensitivity  of plants  to acidic depositions seems to be associated with the wettabi-
lity of  leaf surfaces.  The shorter the time of contact, the lower the resulting dose, and the
less likelihood of injury.
                                            7T103

-------
     Erosion of stone  monuments  and buildings and corrosion  of metals can result from acidic
precipitation.   Because sulfur compounds  are a dominant component of acidic precipitation and
are deposited during  dry  deposition also, the effects resulting from the two processes cannot
be distinguished.  In  addition,  the deposition of sulfur compounds on stone surfaces provides
a medium for microbial growth that can result in deterioration.
     Certain aspects of the  acidic deposition issue remain subject to debate because existing
data are ambiguous or  inadequate.   A comprehensive evaluation of  scientific evidence bearing
on these issues is being prepared as part of a forthcoming EPA critical assessment document on
acidic deposition.
                                            7-3.04

-------
7.6   REFERENCES

Abrahamsen, G. ,  and G.  J,  Dollard.   Effects  of acid precipitation  on forest vegetation and
     soil.   In:   Ecological  Effects  of  Acid  Precipitation,  Report  of  a  Workshop, Electric
     Power  Research  Institute, Gatehouse-of-Fleet,  Galloway, Scotland,  September 4-7, 1978.
     M.  J. Wood,  ed. ,  EPRI  SQA77-403, Electric Power  Research Institute, Palo Alto, CA, July
     1979.  section 4.2.  17 pp.

Abrahamsen, G.,  K.  Bjor, R.  Horntvedt, and B.  Tveite.  Effects of acid precipitation on coni-
     ferous forest.  In:   Impact of Acid Precipitation on Forest and Freshwater Ecosystems  in
     Norway:   Summary   Report  on  the Research Results  from  the  Phase I  (1972-1975)  of the
     SNSF-Project.   F.  H.  Braekke,  ed. ,  Research  Report  FR 6/76,  SNSF  Project, Oslo-As,
     Norway, March, 1976.  pp. 37-63.

Alexander,  M.   Effects  of  acidity  on microorganisms and microbial  processes  in soil.   In:
     Effects of  Acid Precipitation on Terrestrial Ecosystems, North Atlantic Treaty Organiza-
     tion,  Toronto,  Ontario,  Canada,  May 21-27,  1978.   T. C. Hutchinson  and M.  Havas, eds. ,
     Plenum Press, New York, NY, 1980.  pp.  363-374.
Aimer, B., W.  Dickson,  C.  Ekstrom, E.  Hornstrom,  and U. Miller.
     Swedish lakes.  Ambio 3:30-36, 1974.
                                                              Effects of acidification on
Aimer,  B.,  W.  Dickson,  C.  Ekstrom,  and  E.  Hornstrom. Sulfur pollution  and the aquatic eco-
     system.   In:   Sulfur  in  the Environment,  Part II:    Ecological  Impacts.   J.  0. Nriagu,
     ed,, John Wiley & Sons, Inc., New York, NY, 1978.  pp. 271-311.

Altshuller, A.  P., and G. A. McBean.  The LRTAP Problem in North America:  a preliminary over-
     view.  United  States-Canada Research  Consultation Group on  the  Long-Range transport of
     Air Pollutants, Research Triangle Park, NC, and Downsview, Ontario, Canada, 1979.  p. 48.

Amundsen, T.,  and K. Lunder.  Report on fishery-biological surveys  in Tjagevatn and Sonstevatn
     in Grandsherod,  Notodden in Telemark.  Fiskerikonsulenten  i  Ost-Norge, Direktoratet for
     vitt og ferskvannsfisk, Oslo, 1974.
Anderson, J.  W.   Oxidation of sulphur by sulphur by chemautotropic bacteria.
     Biology.  University Park Press, Baltimore, MD, 1978.  pp. 17-19.
                                                                          In:  Sulphur in
Andersson,  I.,  0.0Grahn,  H.  Hultberg,  and  L.  Landner.   Jamforande undersokning  av olika
     tekniker fSr  aterstallande  av forsurade  sjoar.   [Research  for developing different tech
     niques  for  removing acid  from lakes.]   STU Report  73-3651.   Stockholm:  Institute for
     Water and Air Research,  1975.  Cited  in:   National  Research Council, 1978b.  pp. 63-75.

Anschiitz, J., and  F. Gessner.   Der lonehaustausch bei torfmoosen  (Sphagnum).   [Ion exchange by
     peat moss.]   Flora 141:178-236, 1954.  Cited uj:  Aimer et a!., 1978.

Armstrong, F. A.  J.,  and D. W. Schindler.  Preliminary chemical  characterization of waters in
     the experimental lakes area, Northwestern Ontario.  J. Fish. Res. Board Can. 28: 171-187,
     1971.

Arnold,  D.  E.,  R.  W.  Light,  and V. J.  Dymond.   Probable  Effects of Acid  Precipitation on
     Pennsylvania  Waters.   EPA 600/3-80-012, U.S. Environmental  Protection Agency, Corvallis,
     OR, January 1980.
Baker, J. P., and C. L. Schofield.  Aluminum toxicity to fish as related to acid precipitation
     and  Adirondack surface  water  quality.   In;   Ecological  Impact  of  Acid Precipitation,
                      an  International  Conference,  SNSF  Project,  Sandefjord,  Norway,  March
                      Drablrfs and A. Tollan eds.  SNSF Project, Oslo-As, Norway, October 1980.
Proceedings  of
11-14, 1980.  D.
pp. 292-293.
                                            7-105

-------
Baker,  J.,  D. Hocking, H.  Nyborg.   Acidity of open  and  intercepted precipitation in forests
     and  effects  on forest soils in  Alberta,  Canada.   In:  First International Symposium on
     Acid Precipitation and the Forest Ecosystem,  U.S.  Department of. Agriculture and Others,
     Columbus, Ohio, May 12-15, 1975.  Water Air Soil  Pollut.  7:449-460, 1977.

Beamish,  R.  J.   Growth and survival of white  suckers (Catostomus commersoni)  in an acidified
     lake.  J. Fish Res. Board Can. 31:49-54, 1974.

Beamish,  R.   J.   Acidification  of lakes  in  Canada  by  acid precipitation  and the resulting
     effects  on fishes.   In;   First  International  Symposium  on Acid  Precipitation  and the
     Forest  Ecosystem,  U.S.  Department of Agriculture and  Others, Columbus, Ohio, May 12-15,
     1975,  Water Air Soil  Pollut. 6:501-514, 1976.

Beamish, R.  J., and H. H.  Harvey.  Acidification of the La  Cloche  Mountain  Lakes, Ontario, and
     resulting fish mortalities.  J. Fish. Res. Board Can.   29:1131-1143, 1972.

Beamish,  R.  J.,  W.  L.  Lockhart, J. C. Van Loon, and  H. H.  Harvey.   Long-term acidification of
     a lake and resulting effects on fishes.  Ambio 4:98-102, 1975.

Bell, H.  L.   Effect of low pH  on the survival and emergence of aquatic insects.  Water Res.
     5:13-319, 1971.

Bell, H.  L.,  and A. V. Nebecker.   Preliminary studies on  the tolerance of aquatic insects to
     low pH.  J. Kansas Entomological Soc. 42:230-236, 1969.

Bick, H.,  and E.  F.  Drews.  Autopurification and ciliated protozoa in  an acid environment
     (model  study.)  Hydrobiologia 42:393-402, 1973.

Billings, W.  D.   Plants and the  Ecosystem.  3rd  Edition.  Wadsworth Publishing Company, Inc.
     Belmont, CA., 1978.  pp. 1-62.

Bolin, B.,  L.  Granat,  L.  Ingelstam, M. Johannesson,  E. Mattsson,  S. Oden,  H. Rodhe, and C. 0.
     Tanim.  Air Pollution Across National Boundaries:  The  Impact  on the Environment of Sulfur
     in Air  and  Precipitation.   Sweden's Case Study  for the United Nations Conference on the
     Human Environment, P. A. Norstedt and Sons, Stockholm,  Sweden,  1972.

Boughey, A.  S.   Fundamental Ecology.   Intext Educational Publishers, Scranton, PA, 1971.  pp.
     11-50.

Brady, N. C.   The  Nature  and Property of Soils.  8th Edition.  McMillan Publishing Co., Inc.,
     New York, NY, 1974.  pp. 462-472.

Braekke,  F.   H.,  Ed.    Impact of Acid  Precipitation  on  Forest and  Freshwater Ecosystems in
     Norway:   Summary  Report  on  the  Research Results  From the  Phase  I  (1972-1975)  of the
     SNSF-Project.   Research Report 6/76, SNSF Project, Oslo-fis, Norway, March, 1976.

Brezonik, P.  L.  Nutrients  and other biologically active substances  in atmospheric precipita-
     tion.   In:   Atmospheric Contribution  to  the  Chemistry of  Lake  Waters,   First Specialty
     Conference of  the International  Association for Great Lakes  Research,  Longford Mills,
     Ontario,  Canada,  September  28 - October 1,  1975.   J. Great  Lakes  .Res-.   2  (suppl.  1):
     166-186, 1976.

Brock,  T.  D.   Lower  pH   limit  for  the  existence   of  blue-green  algae:  evolutionary  and
     ecological implications.  Science (Washington, D.C.) 179:480-483, 1973.

Brooks, J.  L.,  and  S.  I.  Dodson.   Predation,  body size  and composition of plankton.  Science
     150:28-35, 1965.
                                            7-106

-------
Brosset, C.  Air-borne acid.  Ambio 2:2-9, 1973.

Brouzes, R.  J.  P., R. A.  N.  McLean, and G.  H.  Tomlinson.   Mercury - the  link between pH of
     natural  waters  and  the  mercury  content  of fish.   Presented  at  a  meeting  of  the
     Panel  on Mercury  of the Coordinating Committee for Scientific and Technical Assessments
     of  Environmental  Pollutants,  National  Academy of Sciences,  National  Research Council,
     Washington,  DC,  May  3,  1977.   Domtar  Research  Center, Montreal,  .Quebec, Canada, 1977.

Bua, B., and  E.  Snekvik.  Hatching  experiments  with  roe of salmonid fish 1966-1971.   Effects
     of acidity and salt content of hatchery water.  Vann 7:86-93,  1972.  Cited  in:  Wright and
     Snekvik, 1978.

Cadle,  R.  0.   The  photo-oxidation of  hydrogen sulfide and dimethyl  sulfide in air.  Atmos.
     Environ. 10:417, 1976.

Cadle,  R.   D.  and  M.  Ledford.   Reaction  of  ozone  with  hydrogen  sulfide.   Air Water
     Pollut. 10:25-30, 1966.

Chamberlain,  A.  C.  Dry deposition of  sulfur  dioxide.  In:   Atmospheric Sulfur Deposition:
     Environmental  Impact  and  Health  Effects,   Proceedings  of  the  Second  Life  Sciences
     Symposium, Oak Ridge National Laboratory and Others, Gatlinburg, Tennessee, October 14-18,
     1979.   D.  E.  Shriner, C. R.  Richmond,  and S.  E.   Lindberg,  eds., Ann  Arbor Science Pub-
     lishers, Inc., Ann  Arbor, MI, 1980.  pp. 185-197.

Cholonky,  B.  J.   Die  Okologie  der  diatomeen  in  Binnengewasser.   [Diatom  ecology  in inland
     waters.]  Cramer, Weinheim, 1968.  Cited in:   Hendrey et a!.,  1980a.

Clapham, W. B., Jr.  Natural  Ecosystems.  The Macmillan Co., New York, NY, 1973.

Cogbill, C.  V.    The  history and  character  of  acid  precipitation in  eastern North America.
     In:   First  International   Symposium on  Acid Precipitation  and  the  Forest  Ecosystem,
     U.S.  Department  of Agriculture  and Others, Columbus,  Ohio,  May  12-15,  1975.   Water Air
     Soil Pollut. 6:407-413,  1976.

Cogbill, C.  V.    The  effect  of  acid precipitation on  tree growth in  eastern North America.
     In:   First  International   Symposium on   Precipitation  and  the  Forest  Ecosystem,  U.S.
     Department  of Agriculture  and  Others,  Columbus,  Ohio, May  12-15,  1975.   Water Air Soil
     Pollut. 8:89-93, 1977.

Cogbill, C.   V.,  and  G. E.  Likens.   Acid precipitation in the  northeastern United  States.
     Water Resour. Res.  10:1133-1137, 1974.

Cole,  D. W. ,  and D. W.  Johnson.  Atmospheric sulfate additions and cation leaching in  a Doug-
     las fir ecosystem.  Water Resour. Res. 13:313-317, 1977.

Comerford,  N. B, , and E. H. , White.   Nutrient  content of throughfall in  paper birch  and red
     pine stands  in northern  Minnesota.  Can. J.  For. Res. 7:556-561, 1977.

Connell, W. E., and W. H. Patrick, Jr.  Sulfate  reduction in soil:  effects of redox potential
     and pH.  Science (Washington, D.C.)  159:86-87, 1968.

Conroy,  N.  ,  K,  Hawley, W. Keller,  and  C,  Lafrance.  Influences of the atmosphere on lakes in
     the Sudbury  area.   In:   Atmospheric Contribution  to the  Chemistry of Lake Waters, First
     Specialty  Conference  of the International   Association for Great Lakes Research, Langford
     Mills,  Ontario,  Canada,   September  28  -  October 1,  1975.   J.   Great  Lakes  Res.   2
     (Suppl.  1):146-165, 1976.
                                            7-107

-------
Cowling,  E.  B.,  and L. S. Dochinger.  The  changing  chemistry  of precipitation and its  effects
     on vegetation and materials.   In:  Control  and  Dispersion of Air  Pollutants:   Emphasis  on
     NOV and Participate Emissions.  AICE Symp.  Ser.  74(175):134-142,  1978.
       )\                                              ™T""™

Crisman,  T.  L.,  R.  L.  Schulze,  P. L. Brezonik,  and S.  A,  Bloom.   Acid precipitation:  the
     biotic  response  in  Florida   lakes.   _In:   Ecological  Impact  of  Acid  Precipitation,
     Proceedings  of  an  International Conference,  SNSF  Project, Sandefjprd,  Norway,* March
     11-14,  1980, D.  Drab1*5s and  A.  Tollan, eds.,  SNSF Project,  Oslso-As,  Norway,  October
     1980.  pp.  296-298.

Cronan,  C.  S.    Solution  chemistry  of  a  New  Hampshire subalpine  ecosystem:   biochemical
     patterns and processes.  Ph.D.  Thesis, Dartmouth College,  Hanover,  NH,  1978.

Cronan, C. S.  Consequences of sulfuric acid  inputs  to a  forest soil.   In:   Atmospheric Sulfur
     Deposition:   Environmental   Impact  and  Health  Effects,  Proceedings of  the Second  Life
     Sciences Symposium,  Oak Ridge  National  Laboratory and  Others,  Gatlinburg,  Tennessee,
     October 14-18,  1979.   D. S. Shriner,  C. R. Richmond, and S.  E. Lindberg,  eds.,  Ann Arbor
     Science Publishers, Inc., Ann  Arbor, MI, 1980.   pp.  335-343.

Cronan,  C.  S.,   and C.  L.   Schofield.   Aluminum  leeching  response  to  acid  precipitation:
     effects  on  high-elevation  watersheds  in  the  Northeast.   Science  (Washington, D.C.)
     204:304-305, 1979.

Cullis,  C. F.  and  M.  M.  Hirschler.   Atmospheric   sulphur:   natural  and  man-made sources.
     Atmos. Environ. 14:1263-1278,  1980.

Dancer, W.  S.,  L.  A.  Peterson,  and G.  Chesters.   Ammonification and  nitrification of  N  as
     influenced by  soil pH  and  previous  N treatments.    Soil Sci.  Soc.  Am.  Proc.  37:67-69,
     1973.

Dannevig,  6.  Influence  of  precipitation on  the acidity of water courses and on fish  stocks.
     Jeger  og  Fisker.   3:116-118,  1959.    Cited in:   Leivestad  et  al., 1976; Wright  et  a!.,
     1976; Wright and Snekvik, 1978.

Davis,  R.  B.,  M.  0. Smith,  J.  H.  Baily,  and S. A.  Norton.   Acidification  of Maine (U.S.A.)
     lakes  by  acidic precipitation,   Verh. Int. Ver. Theor.  Angew  Limnol.  20:532-537, 1978.

Denison, R., B.  Caldwell, B. Bormann,  L. Eldred, C.  Swanberg,  and S. Anderson.   The  effects  of
     acid  rain  on  nitrogen  fixation  in  western  Washington coniferous  forests.   In:   First
     International Symposium  on  Acid Precipitation  and  the  Forest Ecosystem,  U.S.  Department
     of  Agriculture and  Others,  Columbus, Ohio,  May 12-15,  1975.   Water Air Soil  Pollut.
     8:21-34, 1977.

Dickson,  W.   The acidification  of  Swedish lakes.    Fish.  Board Swed.  Inst.  Freshwater  Res.
     Drottingham Rep. 54:8-20, 1975.

Dickson,  W,   Some effects  of the  acidification of  Swedish  lakes.   Verh.  Int. Ver.   Theor.
     Angew. Limnol. 20:851-856,  1978.

Dillon, P. 0.,  D.  S. Jefferies,  W.  Snyder, R.   Reid, N.  D. Van,  D. Evans, J.  Moss,  and W. A.
     Scheider.   Acidic precipitation in south-central Ontario:   recent observations.  J. Fish.
     Res. Board Can. 35:809-815,  1978.

Dodson,  S. I.   Zooplankton  competition  and  predation:  an  experimental  test  of the size-
     efficiency hypothesis.   Ecology 55:605-613, 1974.

Donahue, R. L.,  R. W. Miller, and J. C. Shickluna.   Soils:  An Introduction  to Soils  and Plant
     Growth.  4th Edition.    Prentice-Hall,  Inc., Englewood  Cliffs,  NJ,  1977.   pp. 244-248.


                                            7-108

-------
Drablj^s, D., and I. Sevaldrud.  Lake acidification, fish damage, and  utilization  of outfields.
     A comparative  survey  of six highland areas, southeastern Norway.   In:   Ecological  Impact
     of  Acid  Precipitation,  Proceedings  of  an  International  Conference,  SNSF  Project,
     Sandefjord, Norway,  March 11-14,  1980,   D.  Drablsfe  and A.  Tollan, eds.,  SNSF  Project,
     Oslo-As, Norway, October, 1980.  pp. 354-355.

Driscoll,  C. T. , Jr.   Chemical characterization of Some Dilute Acidified Lakes and Streams  in
     the Adirondack  region of New York  State.   Ph.D.  Thesis, Cornell  University,  Ithaca, NY,
     1980.

Driscoll,  C.  T. ,  Jr., J.  P.  Baker,  J.  J.  Bisogni,  Jr.,  and  C.   L.  Schofield.   Effect  of
     aluminum speciation  on  fish  in dilute  acidified waters.  Nature  (London) 284:161-163,
     1979.

Dvorak, A. J.,  B.  G. Lewis,  P. C.  Ghee, J.  D. Jastrow, J. C. Prioleau,  E.  H. Dettmann,  F.  C.
     Kornegay, L. F. Soholt,  R. F. Freeman, III, D. L. Mabes, W. S. Vinikour,  R.  M.  Goldstein,
     P. A.  Merry, E. W. Walbridge, R. R. Hinchman, and E, D.  Pentecost.   Impacts  of Coal-Fired
     Power Plants  on Fish,  Wildlife,  and Their Habitats.   FWS/OBS-78/29, U.S.   Department  of
     the Interior,  Fish and Wildlife Service, Ann Arbor, MI,  March 1978.   pp.  64-70.

Eaggleton,  A. E. J.  and R. A.  Cox.   Homogeneous oxidation of sulfur compounds  in  the  atmos-
     phere.  In:  Sulfur in the Atmosphere, Proceedings of'the International  Symposium,  United
     Nations  Environment   Program  and  Others,  Dubrovnik,  Yugoslavia,  September 7-14,  1977.
     Atmos. Environ. 12:227-230, 1978.

Eaton,  J.  S.,  G.   E.  Likens,  and  F.  H.  Bormann.   Throughfall  and  Stemflow  Chemistry in  a
     Northern Hardwood Forest.  J. Ecol.  61:495-508, 1973.

Eliassen,  A.,  and  J.   Saltbones.   Decay and  transformation rates  of SO,,  as estimated from
     emission data,  trajectories  and measured air  concentrations.   Atmos Environ.  9_:425-429,
     1975.

European Inland  Fisheries  Advisory Committee.  Water quality criteria  for European freshwater
     fish - extreme pH values and inland fisheries.  Water  Res. 3:593-611, 1969.

Evans, L. S., N. F. Gmur, and J. J. Kelsch.  Perturbations  of upper leaf  surface  structures  by
     simulated acid rain.   Environ. Exp. Bot. 17:145-149s 1977a.

Evans, L.  S. , and  0. M. Bozzone.   Effect  of buffered solutions and  sulfate  on vegetative and
     sexual development in gametophytes of Pteridium aquilinum.  Am.  J.  Bot.  64:897-902,  1977.

Evans,  L.  S. ,  N.  F.  Gmur, and F.  Da Costa.   Leaf surface and  histological perturbations  of
     leaves of Phaseolus vulgaris and He!ianthus annuus after exposure  to simulated acid rain.
     Am. J. Bot. 64:903-913,  1977b.

Evans, L.  S., and  D. M. Bozzone.   Effect  of buffered solutions and  various  anions on vegeta-
     tive  and  sexual   development   in  gametophytes  of  Pteridium aquilinum.    Can.  J. Bot.
     56:779-785, 1978.

Evans, L.  S., N.  F. Gmur, and F.  Da Costa.  Foliar response  of six clones of hybrid poplar  to
     simulated acid rain.   Phytopathology 68:847-856, 1978.

Evans,  L.  S. ,  and  T.  M.   Curry.   Differential  responses  of plant  foilage to simulated acid
     rain.   Am.  J.  Bot. 66:953-962, 1979.

Fagestrom, T.,  and A.  Jernelov.  Some aspects of  the quantitative ecology  of mercury.   Water
     Res. 6:1193-1202, 3972.
                                            7-109

-------
Ferenbaugh, R. W.  Effects of  simulated acid  rain  on  Phaseolus  vulgaris  L.  (Fabaceae).   Am.  J.
     Bot. 63:283-288, 1976.

Fiance,  S.  'B.   Distribution  and biology  of mayflies  and  stoneflies  of  Hubbard Brook,  New
     Hampshire.  M.S. Thesis,  Cornell University,  Ithaca,  NY, 1977.

Fisher,  B.  E. A,  Long-range  transport  and deposition  of sulfur  oxides.   In;  Sulfur  in  the
     Environment,  Part I:   The  Atmospheric  Cycle.   J.  0.  Nriagu,  ed.  , John Wiley and Sons,
     Inc.,  New York, NY,  1978.  pp. 243-295.

Foster,  R.  J.   Geology.   Charles E. Merrill  Publishing Co.,  Columbus,  OH,  1971.   162  pp.

Fredricksen,  R.  L.  Nutrient  budget  of  a Douglas-fir forest on an experimental watershed  in
     western  Oregon.   In;  Research on Coniferous  Forest  Ecosystems:    First  Year  Progress  in
     the  Coniferous  Forest  Biome,  US/IBP,  Proceedings of  a   Symposium,  Forty-Fifth  Annual
     Meeting, Northwest Scientific Association, Bellingham,  Washington,  March  23-24, 1972.   J.
     F.  Franklin,  L. J.  Dempster, and  R.  H.  Waring,  eds., U.S.  Department of  Agriculture,
     Forest Service, Portland, OR, 1972.  pp. 115-131.

Friend,  J.  P.   The global  sulfur  cycle.    In:   Chemistry of the lower  atmosphere.   S.  I.
     Rasool,  ed.  Plenum  Press, NY, 1973.

Gahnstrom,  G. ,  G.   Andersson,  and  S.  Fleischer.   Decomposition  and  exchange  processes  in
     acidified lake  sediment.  In:  Ecological  Impact of Acid Precipitation, Proceedings of an
     International  Conference,  SNSF  Project,  Sandefj,ord,  Norway,   March  11-14,  1980.   D.
     Drabltfs  and A.  Tollan,  eds., SNSF Project,  Oslo-As,  Norway, October  1980.   pp.  306-307.

Gallo./ay, J.  N.  Effects  of  sulfur oxides on  aquatic  ecosystems.   In;  Sulfur  oxides,  National
     Academy  of  Sciences, Washington, DC, 1978.

Galloway, J.  N., and E.   B.  Cowling.   The effects of precipitation on aquatic and  terrestrial
     ecosystems:  a  proposed precipitation  network.  J.  Air  Pollut. Control Assoc.  28:229-235,
     1978.

Galloway, J.  N., and D. M. Whelpdale.  An atmospheric sulfur budget for  eastern  North  America.
     Atmos. Environ.  14:409-417, 1980,

Galvin,  P.  J. ,  P.  J. Samson,  P.  E.  Coffey, and D.  Romano.   Transport of sulfate  to  New York
     State.   Environ. Sci. Techno!. 12:580-584, 1978.

Galvin,  P.  J.,  and  J.  A. Cline.  Measurement  of anions in the snow  cover of the Adirondack
     Mountains.  Atmos. Environ. 12:1163-1167,  1978.

Giddings,   J.  ,   and  J.  N.  Galloway.   The  effects  of acid  precipitation  on  aquatic  and
     terrestrial  ecosystems.    In;    Literature   Reviews   on   Acid   Precipitation.   Cornell
     University,  Centers  for  Environmental  Quality  Management  Water  Resources  and  Marine
     Sciences, Ithaca, NY, March 1976.  40 pp.

Gjessing, E.  T., A.  Henriksen,  M.  Johannessen,  and R.  F.  Wright.   Effects  of acid precip-
     itation  on  freshwater  chemistry.   In:   Impact   of  Acid  Precipitation on  Forest  and
     Freshwater  Ecosystems in Norway:  Summary  Report on the Research  Results  From  the Phase I
     (1972-1975)  o£ the  SNSF-Project.   F.   H.  Braekke, ed. ,  Research  Report  FR 6/76,  SNSF
     Project, Oslo-As, Norway, March 1976.  pp. 65-85.

Glass, N. R. , G. E.  Glass,  and  P.  J.  Rennie.  Effects  of acid precipitation.  Environ.  Sci.
     Techno!.  13:1350-1355,  1979.
                                            7-110

-------
Gorham,  E.    The   influence  and  importance  of  daily weather  conditions  in  the  supply of
     chloride, sulfate and other ions to fresh waters  from atmospheric precipitation.  Philos.
     Trans.  R. Soc. London Ser. 8.241:147-178, 1958.

Gorham,  E.   Acid precipitation  and its influence  upon  aquatic  ecosystems—an overview.  In;
     First  International  Symposium  ,on  Acid  Precipitation _and  the  Forest  Ecosystem,  U. S.
     Department of Agriculture and  Others,  Columbus,  tfhio," May  12-15,  1975.   Water Air  Soil
     Pollut. 6:457-481, 1976.

Gosner, K. L., and I.  H. Black,  The effects of acidity on the development and hatching  of New
     Jersey frogs.  Ecology 38:256-262, 1957.

Grahn, 0.   Macrophyte  succession in Swedish lakes  caused by deposition of  airborne acid  sub-
     stances.  In:  Proceedings of the First International Symposium on Acid  Precipitation and
     the  Forest  Ecosystem,  May  12-15,  1975,  Columbus,  Ohio.   L.  S.  Dochinger  and  T. A.
     Seliga,  eds.,  U.  S.  Forest  Service General  Technical Report  NE-23,  U.  S.  Department of
     Agriculture,  Forest  Service,  Northeastern Forest  Experiment  Station,  Upper  Darby, PA,
     1976.  pp. 425-458.

Grahn, 0.,  H.  Hultberg,  and L.  Landner.   Oligotrophication—a  self-accelerating  process in
     lakes subjected to excessive supply of acid substances.  Ambio 3_: 93-94,  1974.

Granat,  L.   On the relation between  pH and the  chemical composition  in atmospheric  preci-
     pitation.  Tell us 24:550-560, 1972.

Grennard, A.,  and  F.   Ross.   Progress  report  on  sulfur dioxide.   Combustion  45:4-9,  1974.

Hagen, A., and A.  Langeland.   Polluted snow in southern Norway and the effect of the meltwater
     on freshwater and aquatic organisms.  Environ. Pollut. 5:45-57, 1973.

Hagstrom, T.   Grodornas forsvinnande i an forsurad sjo.   [Disappearence of  frogs from an  acid
     lake.]   Sver. Nat. 11/77:367-369, 1977.  Cited in:   Hendrey, 1978.

Haines, E. B.  Nitrogen content and acidity of rain on the Georgia coast.  Water Resour. Bull.
     12:1223-1231, 1976.

Hall,  R.  J.,  G.  E. Likens,  S.  B.  Fiance,  and G. R. Hendrey.  Experimental  acidification  of a
     stream  in the  Hubbard Brook  Experimental  Forest,  New Hampshire.   Ecology 61:976-989,
     1980.

Halstead, R.  L.,  and  P. J.  Rennie.   Effects  of sulphur  on soils in Canada.  In:  Sulphur and
     its  Inorganic Derivatives  in the  Canadian  Environment.    NRCC  N.C.  15015,  National
     Research  Council  of   Canada,  NRC  Associate   Committee   on  Scientific  Criteria  for
     Environmental Quality,  Ottawa, Canada, 1977.  pp. 181-219.

Hamilton,  W.  C.    Personal  communication  to  Dr.  J.   H.  B.   Garner,  U.  S.  Environmental
     Protection Agency, Research Triangle Park, NC, January 31, 1980.

Hanken,  I. ,  J. F.  Lynch, and  D.  B. Wake.   Salamander  invasion of  the tropics.   Nat. Hist.
     89:(12) December 47-53, 1980.

Harr, T.  E., and  P. E. Coffey.  Acid Precipitation  in  New York State.  Technical Paper No. 43,
     New  York State  Department  of Environmental  Conservation,  Albany,  NY, July  1975.  pp,
     21-29.

Harward,  M.  E.,  and H. H. Reisenaur.   Movement and reactions of inorganic  soil sulfur.   Soil
     Sci. 101:326-335, 1966.
                                            7-111

-------
Heagle, A.  S., W.  W,  Heck,  W.  M.  Knott,  J.  V.  Johnston,  E. P.  Stahel,  and E. B. Cowling.
     Responses of  citrus  to acidic  rain  from  simulated SRM fuel exhaust mixtures and exhaust
     components.    Interim  report to NASA John  F.  Kennedy Space Center, Titusville,  FL, 1978.
     11 pp.

Heinrichs,  H.,  and R.  Mayer.   Distribution and cycling  of  major and  trace elements in two
     central forest ecosystems.  J.  Environ. Qua!. 6:402-407,  1977.

Henderson, 6. S., and W. F. Harris.  An ecosystem approach to  characterization of the nitrogen
     cycle  in  a  deciduous forest watershed.   Iji Forest Soils and  Forest Soils Conference. B.
     Bernier and C.  H.  Winget,  eds., Les Presses de  TUniversite Laval, Quebec,  Canada, 1975.
     pp. 179-193.

Hendrey, G. R.  Effects of pH on the Growth of  Periphytic Algae in  Artificial Stream  Channels.
     Report IR 25/76, SNSF Project,  Oslo-Ss, Norway,  October 1976.

Hendrey,  G.  R.,  and R. F.  Wright.   Acid precipitation  in Norway:   Effects on aquatic fauna.
     In:  Atmospheric Contribution to the Chemistry of  Lake Waters, First Specialty Conference
     of the International  Association  for Great  Lakes  Research, Longford  Mills,  Ontario,
     Canada, September  28-October 1,  1975.  J.  Great Lakes  Res.  2(Suppl. 1):192-207, 1976.

Hendrey, G.  R.,  and F.  Vertucci.  Benthic  plant communities  in acidic  Lake Golden,  New York:
     sphagnum and  the  algal mat.  In:  Ecological  Impact of Acid  Precipitation,  Proceedings
     of an  International  Conference, SNSF  Project, San^lefjord,  Norway,  March 11-14, 1980, D.
     Drablife and  A. Tollan,  eds., SNSF-Project, Oslo-As, Norway,  October  1980,  pp. 314-315.

Hendrey, G.  R.   Aquatics  Task Force on  Environmental Assessment of the Atikokan Power Plant:
     Effects on Aquatic  Organisms.  BNL 50932,  U. S.  Department of  Energy,  Brookhaven National
     Laboratory,  Upton, NY, November 1978.

Hendrey, G. R., K.  Baalstrud, T. S.  Traaen, M.   Laake, and G. Raddum.  Acid  precipitation: some
     hydrobiological changes.  Ambio 5:224-227,  1976.

Hendrey, G.  R. ,  J.  N.  Galloway, S.  A. Norton,  C. L.  Schofield, P.  W. Shaffer and D.  A. Burns.
     Geological and Hydrochemical  Sensitivity  of the Eastern  United States to Acid Precipita-
     tion.   EPS-600/3-80-024, U.S.  Environmental  Protection  Agency, Corvallis, OR, January
     1980b.

Hendrey, G. R., N.  D.  Yan, and K. J. Baumgartner.  Responses of freshwater  plants and inverte-
     brates  to  acidification.   Presented  at the  EPA/OECD  International Symposium for Inland
     Waters and Lake Restoration, Portland, ME,  September 8-12, 1980a.

Hendry, C.  D., Jr.  The Chemistry  of Precipitation in North-Central Florida, M.S. Thesis M.S.
     Thesis,  University   of  Florida,   Department   of  Environmental   Engineering   Sciences,
     Gainesville, FL, 1977.

Hendry, C.  D. ,  and P.  L.  Brezonik.   Chemistry  of  precipitation at  Gainesville, Florida.
     Environ. Sci.  14:843-849, 1980.

Henriksen,  A.   A  simple  approach for identifying and  measuring acidification of freshwater.
      Nature (London) 278:542-545, 1979.

Henricksen,  A.  and  R.  F.  Wright.   Effects  of acid precipitation on  a small   acid lake in
     Southern Norway.   Nord. Hydro!. 8:1-10, 1977.

Hermann,  R.,  and J.  Baron.   Aluminum  mobilization   in  acid  stream environments, Great Smoky
     Mountains National Park, U.S.A. In:  Ecological  Impact of Acid Precipitation, Proceedings
     of an  International  Conference, SNSF  Project, Sangiefjord, Norway,  March 11-14,  1980.  D,
     Drablrfs and  A Tollan,  eds., SNSF  Project, Oslo-As, Norway,  October  1980,  pp. 218-219.

                                            7-112

-------
Hicks, B.  B.,  and  M.  L. Wesely.   Turbulent transfer processes  to  a surface and  interaction
     with  vegetation.   In:   Atmospheric  Sulfur Deposition:   Environmental  Impact and Health
     Effects, Proceedings of the Second Life Sciences Symposium,  Oak  Ridge National Laboratory
     and Others, Gatlinburg,  Tennessee,  October 14-18,  1979,   D. S.  Shriner, C. R, Richmond,
     and S,  E.  Lindberg, eds., Ann  Arbor Science Publishers,  Inc.,  Ann  Arbor,  MI, 1980, pp.
     199-207.

Hindawi, I.  J.,  J.  A.  Rea,  and W.  L. Griffis.   Response of bush bean  exposed  to acid mist.
     Am.  J. Bot.  67:168-172, 1980.

Hoeft, R.  G., D.  R. Keeney, and L. M. Walsh.  Nitrogen  and sulfur in precipitation and sulfur
     dioxide in the atmosphere in Wisconsin. J. Environ. Qua!.  1:203-208, 1972.
Hornstrom,   E., C,  Ekstrom,  U. Miller, W.  Dickson.   Effects of the  Acidification on Lakes  in
     the Swedish  West  Coast  Region.   Statens  Naturvardsverk,  Slona,  Sweden,   Publikationer
     1973:  7.  Cited in:  Wright and Gjessing, 1976'.

Hoi ling, C.  C.    Resilience  and  stability  of ecological  systems.    Annu.   Rev.  Ecol.  Syst.
     4:1-23, 1973.

Hornbeck, J. W.,  G. E.  Likens, J.  S. Eaton.  Seasonal patterns  in acidity of  precipitation and
     their  implications  for forest  stream ecosystems.   'In;   Proceedings  of the First Inter-
     national Symposium on  Acid  Precipitation  and  the Forest Ecosystem,  May 12-15,  1975,
     Columbus,  Ohio.   L. S.  Dochitiger and  T.  A.  Seliga, eds.   U.  S.  Forest Service, North-
     eastern Forest Experimental Station, Upper Darby, PA, 1976.  pp. 597-609.

Hubbert, M.  K.   Outlook for  fuel  reserves.   In:   McGraw-Hill  Encyclopedia  of Energy.   D.  N.
     Lapedes, ed., McGraw-Hill Inc., New York, NY, 1976.  pp 11-23.

Hultberg,  H.  Thermally stratified  acid water  in  late winter  - a  key  factor  inducing self-
     accelerating  processes  which  increase acidification.   In:   Proceedings  of  the  First
     International  Symposium  on Acid precipitation and  the Forest Ecosystem, May 12-15, 1975,
     Columbus,  Ohio.  L.  S.  Dochinger and  T.  A.  Seliga,  eds.    U.  S.  Forest Service, North-
     eastern Forest Experimental Station, Upper Darby, PA, 1976.  pp. 503-517.

Hutchinson,  T.  C. ,  and  L.  M. Whitby.   The effects  of acid  rainfall and heavy metal parti-
     culates on  a  boreal  forest  ecosystem  near  the Sudbury  smelting region of Canada.   In:
     First  International  Symposium  on   Acid  Precipitation  and the Forest  Ecosystem,  U.S.
     Department of Agriculture and  Others,  Columbus, Ohio, May  12-15,  1975.  Water Air Soil
     Pollut. 7:421-438, 1977.

Hutchinson, T.  C. ,  W.  Gizym, M. Havas, and V. Zoberns.  Effects  of long-term lignite burns  on
     Arctic  ecosystems at  the  Smoking Hills,  N.W.T.    In:   Proceedings  of  the First Inter-
     national Symposfum on  Acid  Precipitation  and  the Forest Ecosystem,  May 12-15,  1975,
     Columbus,  Ohio.   L. S.  Dochinger and  T.  A.  Seliga, eds.   U.  S.  Forest Service, North-
     eastern Forest Experimental Station, Upper Darby, PA, 1976.  pp. 745-746.

Irving,  P.  Rainfall   acidity at  argonne.   In:   Radiological  and Environmental  Research
     Division Annual Report:   Ecology, January through  December 1978.  ANL-78-65 (Pt. 3} U.S.
     Department of Energy, Argonne National  Laboratory,  Argonne,  IL,  1978, pp. 21-33.

Irving,  P.  M. ,  and J.  E. Miller.   The effects of acid  precipitation alone and  in  combination
     with  sulfur  dioxide  on  field-grown  soybeans.   In:    Radiological   and   Environmental
     Research Division Annual Report:  Ecology, January  through December 1978.   ANL-78-65 (Pt.
     3), U.S. Department of Energy, Argonne National Laboratory,  Argonne, II, 1978, pp. 17-20.

Irving,  P.  M.   Response of Field-Grown Soybeans  to Acid  Precipitation Alone and  in Combina-
     tion  with  Sulfur Dioxide.   PhD.  Thesis,  University of Wisconsin,  Milwaukee, WI, 1979.

-------
Jacobson, J.S.  Experimental studies on the phytotoxiclty of acidic precipitation:   the United
     States experience.   In:   Effects of Acid  Precipitation  on Terrestrial Ecosystems, North
     Atlantic  Treaty  Organization,  Toronto,  Ontario,  Canada,  May   21_27,   1978.   T,  C.
     Hutchinson. and M. Havas, eds., Plenum Press, New York, NY,  1980a, pp.  151-160.

Jacobson, J.S.  The influence of  rainfall composition on the yield and quality  of agricultural
     crops.   In:   Ecological Impact  of Acid  Precipitation,  Proceedings  of, an International
     Conference,  SNSF Project,  Sande|jord,  Norway,  March  11-14, 1980.   D.  DrabljJs  and A.
     Tollan, eds., SNSF Project, Oslo-As, Norway, October 1980b, pp. 41-46.

Jacobson,  J.S. ,   and  P.   Van Leuken.   Effects of  acidic  precipitation  on  vegetation.    In:
     Proceedings  of  the  Fourth  International Clean Air Congress, International Union of  Air
     Pollution Prevention Associations,  Tokyo, Japan, May 16-20,  1977.  S.,Kasuga,  N. Suzuki,
     T. Yamada, 6.  Kimura, K.  Inagaki, and K. Onoe, ets., The Japanese  Union of Air Pollution
     Prevention Associations, Tokyo Japan, 1977, pp. 124-127.

Jacobson, J.A., J. Troiano, L. J.  Colavito, L. I. Heller, and D. C. McCune.  Polluted rain  and
     plant growth.  In:   Polluted  Rain, Twelfth International Rochester  Conference on Environ-
     mental Toxicity,  University  of  Rochester,  Rochester,  New  York, May  21-23, 1979.   T. Y.
     Toribara, M.  W.  Miller, and P. E. Morrow, eds., Environmental Science  Research  Volume  17,
     Plenum Press, New York, NY, 1980, pp. 291-305.

Jensen, K. W.,  and E.  Snekvik.   Low pH levels wipe out salmon and trout populations in south-
     ernmost Norway.   Ambio I:223-225, 1972.

Johannessen, M.,  and A.  Henriksen.   Chemistry  of  snow meltwater:  .changes in concentration
     during melting.   Water Resour. Res. 14:615-619, 1978.

Johannessen, M.,  A.  Skartveit,  and R.  F.  Wright.   Streamwater chemistry before,   during  and
     after  snowmelt.    In:    Proceedings  of  an  International  Conference,   SNSF Project,
     Sandegjord,  Norway,  March 11-14,  1980.   D. Drab¥s  and A.  Tollan,  eds,, SNSF Project,
     Oslo-As, Norway, October 1980, pp. 224-225.

Johnson,  D. W.   Site susceptibility to  leaching  by H^SC), in acid rainfall.   In:    Effects of
     Acid  Precipitation   on  Terrestrial  Ecosystems,  North  Atlantic  Treaty Organization,
     Toronto, Ontarip, Canada,  May 21-27, 1978.  T.  C.  Hutchinson and  M.  Hawas, eds., Plenum
     Press, New York, NY,  1980, pp. 525-535.

Johnson, D. W., and D. W.  Cole.   Anion mobility in sols:  Relevance to nutrient transport from
     forest ecosystems.   Environ.  Int. 3:79-90, 1980.

Johnson,  D. W., J.  W.  Hornbeck, J. M.  Kelly, W.  T. Swank, and  D. E. Todd.  Regional patterns
     of soil  sulfate  accumulation:  relevance to ecosystem  sulfur budgets.  In:  Atmospheric
     Sulfur Deposition:  Enfironmental Impact  and Health  Effects, Proceedings  of  the,Second
     Life Sciences Symposium, Oak  Ridge National Laboratory and  Others,  Gatlinburg,  Tennessee,
     October 14-18,  1979.   D.  S,   Shriner, C. R. Richmond, and S.  E. Lindberg,  eds.,  Ann Arbor
     Science Publishers,  Inc., Ann Arbor, MI, 1980, pp. 507-520.

Johnson,  N. M., G.  E.  Likens, F.  H. Bormann, D. W. Fisher, and  R. S. Pierce.   A working model
     for the variation in  stream water chemistry at the Hubbard  Brook Experimental Forest,  New
     Hampshire.  Water Resources Res. 5:1353-1363, 1969.

Jones. M.  B.   Sulfur in  agricultural lands.  In: Sulfur in the  Environment.  Revised Edition.
     Missouri Botanical Garden, St. Louis, Mo., July 1975, pp. 146-158.

Junge,  C.  E.   The  distribution of ammonia and nitrate in  rainwater  over the  United States.
     Trans. Am. Geophys.  Union 39:241-248, 1958.
                                            7-114

-------
Kallqvist, T.,  R.  Romstad, and J.  Kotai.  Preliminary study of pH-effects on algal communities
     in experimental  channels.   SNSF-Project  IR 7/75,  1975.  Cited in:  Hendry  et al., 1980.

Kramer, J. R. , Geochemical  and lithological factors  in  acid precipitation.   In:  Proceedings
     of the First International Symposium on Acid Precipitation and the Forest Ecosystem, U.S.
     Department of Agriculture  and  Others,  Columbus, Ohio,  May  12-15,  1975.   L. S. Dochinger
     and T. A.  Seliga, eds. , NEFES/77-1", U.S. Department T»T Agriculture, Forest Service, Upper
     Darby, PA, August 1976, pp. 611-618.

Kucera, V.  Effects  of sulfur dioxide and  acid  precipitation on metals and anti-rust painted
     steel.  Ambio 5:243-248, 1976.

Kwiatkowski,  R. E.,  and J.  C. Roff.  Effects of acidity on the phytoplankton and primary pro-
     ductivity of selected northern Ontario lakes.  Can.  J. Bot.  54:2546-2561, 1976.

Laake,  M.   Effekter  av  lav pH pa  produksjon,  nedbryting og stoffkretsltfp  i  littoralsonen.
     [Effects  of  low pH on the production,  breaking down of harmful substances  in a coastal
     region.]  Research  Project IR  29/76,  1432  Aas-NLH,  Norway.   SNSF-Projects Secretariat,
     1976.  75 pp.  Cited in: Galloway, 1978.

Landner,  L., and  P.  0. Larsson.  Biological  Effects of Mercury Fall-Out  Into  Lakes from the
     Atmosphere,   IUL Report BUS.   Institute for Water and  Air Research, Stockholm, Sweden,
     1972.  18 pp.   (in Swedish,  translated by H. Altosaar,  Domtar Research Centre, December
     23, 1975)

Lang, D.  S., S.  V.  Krupa, and  D. S.  Shriner.   Injury to  vegetation  incited by sulfuric acid
     aerosols  and acidic rain.   Presented  at the 71st Annual Meeting,  Air Pollution Control
     Association,  Houston, TX, June 25-30, 1978.   Paper No. 78-7.3.

Larson, T. V.,  R.  J.  Charlson, E.  J, Knudson, G.  D. Christian, and H. Harrison.   The influence
     of a sulfur dioxide  point source on  the rain  chemistry of a single  storm in the Puget
     Sound region.  Water Air Soil Pollut. 4:319-328, 1975.

Larson, T. E.  and J.  Hettrick, Mineral composition of rainwater.   Tellus 8: 196, 1956.

Lee,  J.  J. ,  and  D.  E. Weber.  Effects  of  Sulfuric  Acid  Rain on  Two  Model  Hardwood Forests:
     Throughfall,  Litter  Leachate,  and  Soil Solution.    EPA-600/3-80-014,  U.S.  Environmental
     Protection Agency, Corvallis, OR, January 1980.

Lee,  J.  J.,  G. E. Neely,  and S.  C.  Perrigan.   Sulfuric Acid Rain Effects  on  Crop Yield and
     Foliar  Injury.   EPA-600/3-80-016,  U.S.  Environmental  Protection  Agency,  Corvallis, OR,
     January 1980.

Leivestad, H., G. Hendrey,  I.  P.  Muniz,  and E.  Snekvik.   Effects of acid precipitation on
     freshwater organisms.   In:  Impact  of Acid Precipitation  of  Forest and Freshwater Eco-
     systems in Norway:   Summary Report on the Research Results from the Phase I (1972-75)0of
     the  SNSF-Project.   F.  H.  Braekke,  ed., Reserach Report FR 6/76,  SNSF Project, Oslo-As,
     Norway, March 1976, pp. 87-111.

Lewis, W. M., Jr., and M. C. Grant.   Acid precipitation in the western United States.  Science
     (Washington, D.C.) 207:176-177, 1980.

Likens, G. E.  The Chemistry of  Precipitation in the Central Finger Lakes Region.  Technical
     Report 50,  , Cornell  University, Water  Resources and Marine  Science Center,  Ithaca, NY,
     October 1972.
                                            7-115

-------
Likens, G. E.  Acid precipitation.  Chem. Eng. News 54:29-44, 1976.

Likens,  G.   E.,  and  F.  H. Bormann.   Acid rain:   a  serious regional  environmental problem.
     Science (Washington, D.C.) 184:1176-1179, 1974.

Likens,  G.  E., G,  H.  Bormann,  N.  M.  Johnson,  D.  W,  Fisher,  and  R.  S.  Pierce.  Effects of
     cutting  and  herbicide treatment on nutrient  budgets  in  the Hubbard  Brook Watershed -
     ecosystem.  Ecol. Monogr. 40:23-47, 1970.

Likens,  G.  E., F.  H.  Bormann,  J.  S.  Eaton,  R.  S.  Pierce, and  N.  M.  Johnson.  Hydrogen  ion
     input to  the Hubbard  Brook Experimental Forest, New  Hampshire,  during the last decade.
     Im  First International  Symposium on Acid  Precipitation  and the Forest  Ecosystem, U.S.
     Department of  Agriculture and  Others, Columbus,  Ohio, May  12-15,  1975.  Water Air Soil
     Pollut.  6:435-445, 1976.

Likens, G. E.,  J.  S.  Eaton, and  J.  N.  Galloway,  precipitation  as  a source of  nutrients  for
     terrestrial and  aquatic ecosystems.   In:   Precipitation  Scavenging (1974),  Proceedings
     of  a Symposium,  Illinois State Water Survey and Others,  Champaign,  Illinois, October
     14-18,  1974,  EROA Symposium Series 41,  Energy  Research  and Development  Administration,
     Oak Ridge, TN, June 1977, pp. 552-570.

Likens,  G.   E.,  R.  F.  Wright,   J.  N.   Galloway,  and  T.  J.  Butler.   Acid rain.   Sci.   Am.
     241:43-51, 1979.

Liljestrand,  H. M., and J. J.  Morgan.  Chemical composition  of acid precipitation in  Pasadena,
     Calif. Environ. Sci. Techno!. 12:1271-1273, 1978.

Lindberg, S.  E., R.  C.  Harriss,  R.  R.  Turner, D. S.  Shriner, and D. D. Huff.  Mechanisms  and
     Rates of  Atmospheric Deposition of Selected  Trace Elements and  Sulfate  to a  Deciduous
     Forest  Watershed.   ORNL/TM  6674,  Environmental  Sciences  Division Pub. No.  1299,   U.S.
     Department of Energy, Oak Ridge  National  Laboratory, Oak Ridge, TN, June 1979.

Linzon, S. N.   Statement made to workshop  reviewers.   Research Triangle Park, NC.,  November,
     1980.

Livingstone,   D.  A.   Chemical  composition  of  rivers  and  lakes.   In:  Data of  Geochemistry.
     Sixth Edition - Chapter G. M. Fleischer,  ed., Geological Survey Professional Paper 440-G,
     U.S. Government Printing Office, Washington, DC, 1963.

Lockhart, W.  L.,  and A.  Lutz.    Preliminary biochemical observations  of  fishes  inhabiting an
     acidified lake in Ontario, Canada.  In:   Proceedings of the  First  International  Symposium
     on Acid Precipitation and the Forest  Ecosystem, May  12-15,  1975, Columbus, Ohio.  L. S.
     Dochinger and T.  A. Seliga,  eds.  U. S. Forest Service  General Technical Report  NE-23, U.
     S.  Department of  Agriculture,   Forest  Service,  Northeastern Forest  Experiment Station,
     Upper Darby, PA,  1976.  pp.  725-735.

Lynch, H.  Predation,  competition and zooplankton community structure: an experimental study.
     Limnol.  Oceanogr. 24:253-272, 1979.

Malraer, N.  Acid precipitation:   chemical changes in the soil., Ambio 5:231-234,  1976.

Marenco, A,  and J.  Fontan.  Influence  of dry deposition on the  residence time of particulate
     pollutants  in the  troposphere.   In:   Atmosphere-Surface  Exchange  of  Particulate   and
     Gaseous  Pollutants  (1974),  Proceedings  of  a  Symposium,  Bate'lle   Pacific   Northwest
     Laboratories and Others,  Richland, Washington, September 4-6, 1974.  ERDA  Symposium Seris
     38,  Energy Research  and Development  Administration,  Oak Ridge,  TN,  January  1976,   Pp.
     54-61,
                                            7-116

-------
May, R., Stability and Complexity in Model Ecosystems.  Princeton University  Press, Princeton,
     N.J., 1973.

Mayer, R. ,  and B.  Ulrich.   Acidity  of precipitation  as  influenced by  the  filtering of at-
     mospheric  sulphur  and nitrogen  compounds-its role in the  element balance and effect on
     soil.  In:  First International Symposium on  Acid  Precipitation and  the  Forest Ecosystem,
     U.S. Department  of  Agriculture and Others, Columbus, Ohio, May 12-15, 1975.  Water, Air,
     and Soil Pollut. 7:409-416, 1977.

McColl, J. G.,  and D. S. Bush.   Precipitation  and throughfall  chemistry in  the San Francisco
     Bay Area.  J. Environ. Qua! 7:352-357, 1978.

McFee, W.  W.,  J.  M.  Kelly,  and R. H.  Beck.   Acid precipitation effects on  soil  pH and base
     saturation of  exchange  sites.   In;  First  International  Symposium on Acid Precipitation
     and the  Forest Ecosystem, U.S. Department  of Agriculture and Others, Columbus, Ohio, May
     12-15, 1975.  Water Air Soil Pollut. 7:401-408, 1977.

McFee, W. W.   Sensitivity of  Soil  Regions To Long Term Acid Precipitation. U.S. Environmental
     Protection Agency, EPA-600/3-80-013, Corvallis, OR.  January 1980,

Menendex, R.   Chronic effects  of reduced pH on  brook trout- (salvelinus fontinalis).  J. Fish.
     Res. Board Can. 33:118-123, 1976.

Miller, J. M.,  J.  N. Galloway, G.  E.  Likens.   Origin  of air masses producing  acid precipita-
     tion at  Ithaca, New  York, a  preliminary  report.   Geophys.  Res.  Lett.  5:757-760, 1978.

Moran, J. M.,  M.  D. Morgan, and J. H. Wiersma.  Introduction to Environmental  Science.  W. H.
     Freeman and Co., San Francisco, CA, 1980.   pp. 7-75.

Muller,  P.   Effects of artificial  acidification on the growth  of  periphyton.   Can.  J. Fish.
     Aquat.  Sci. 37:355-363, 1980.

Muniz, I. P.,  and H. Leivestad.  Toxic  effects  of aluminum on  brown  trout,  Salmo trutta, L.
     In:  Ecological Impact of Acid Precipitation, Proceedings of an International Conference,
     SNSF Project,  Sand%fjord, Norway,  March 11-14, 1980.   D.  Drablrfs  and  A. Tollan, eds.,
     SNSF Project, Oslo-As, Norway, October 1980,  pp. 320-321.

National  Research  Council.    Nitrates:   An  Environmental  Assessment.   National  Academy of
     Sciences, Washington, DC, 1978.  pp. 275-317.

National  Research  Council, Sulfur  Oxides.   National Academy  of Sciences,  Washington, D.C.,
     1978b.

Nieboer,  E.,  D. H.  S.  Richardson, K.  J.  Puckett and  F. D. Tomassini.   The  phytotoxicity of
     sulfur  dioxide in  relation  to measureable responses  in  lichens.   In:    Effects  of Air
     Pollutants  on Plants, Society for Experimental Biology,  Liverpool, England,  April  10,
     1975.   T.  A.  Mansfield,   ed. , Cambridge University Press, London,  England,  1976*.   pp.
     61-85.
     o
Nihlgard, B.   Precipitation,  its  chemical composition  and effect on soil water in a beech and
     a spruce forest in south  Sweden.  Oikos, 21:208-217, 1970.

Nisbet,  I. C.  T.   Sulfates and acidity  in precipitation:  their relationship to emissions and
     regional  transport  of sulfur  oxides.   In;   Air  Quality and  Stationary  Source Emission
     Control.   A  report  by the Commission on Natural Resources, National Academy of Sciences,
     National Academy of Engineering, National Research Council.  Serial  No.  94-4, U.S. Senate
     Committee  on Public Works, Washington, DC,  March 1975.   pp. 276-312.
                                            7-117

-------
Norton, S. A.  Changes in chemical processes in soils caused by acid precipitation.  In: First
     Inaternational Symposium  on  Acid Precipitation and the Forest Ecosystem, U.S. Department
     of  Agriculture and  Others,   Columbus,  Ohio,  May 12-15,  1975.   Water Air  Soil  Pollut,
     7:389-400,.1977.

NHagu,  J.   Deteriorative  effects  of  sulfur  pollution on  materials.   In:   Sulfur  in the
     Environment.  Part  II:   Ecological  Impacts.  J. 0.  Nriagu,  ed.,  John Wiley & Sons, Inc.,
     New York, NY, 1978.  pp. 1-59.

0kland,  J.   Environment  and  snails (Gastropoda):   studies of  1000  lakes in  Norway.   In:
     Ecological Impact of Acid Precipitation., Proceedings of an International Conference, SNSF
     Project, Sandefjord,  Norway, March  11-14, 1980.   D.  Drablrfs and A.  Tollan, eds., SNSF
     Project, Oslo-As, Norway, October 1980.  pp. 322-323.

0kland, J., and K. A.  0kland.  pH level and food organisms for fish:  studies of 1000 Lakes in
     Norway.  In;   Ecological  Impact  of Acid  Precipitation,  Proceedings of an International
     Conference,  SNSF Project,  Sandegjord, Norway,  March  11-14, 1980.   D.  Drablj^s  and A.
     Tollan, eds., SNSF Project, Oslo-As. Norway, October 1980.  pp. 326-327.

0kland,  K.  A.   Mussels and  crustaceans:   studies of 1,000  lakes  in  Norway.  In:   Ecological
     Impact of Acid Precipitation,  Proceedings of  an  International  Conference,  SNSF Project,
     Sandeijord,  Norway,  March 11-14,  1980.  D.  Drabl^s and A.  Tollan, eds.,  SNSF Project,
     Oslo-As, Norway,  October 1980.  pp. 324-325.

Odin,  S.   The  Acidification  of  Air and  Precipitation and  Its  Consequences on  the Natural
     Environment.   Ecology  Committee  Bulletin  No.  1,  State  National  Science  Research Com-
     mittee, Stockholm, Sweden, May 1968.

Odin,  S.   Acid   precipitation:   A  world  concern.   In:   Proceedings  of  the  Conference on
     Emerging Environmental  Problems, Acid  Precipitation,  May 19-20,  1975, The Institute on
     Man and Science,  Rensselaerville,  N.Y.  U.S.  Environmental Protection  Agency, Region II.
     Water  Resources  and Marine  Sciences Center,  Cornell University  and Center for Environ-
     mental  Quality Management,  Cornell  University,   EPA-902/9-75-001,  November  1975.   pp.
     5-44.

Odin, S.,  R.  Anderssan,  and M. Barting.   The  long-term changes in the chemistry  of soils in
     Scandanavia  due to  acid precipitation.  In:   Supporting  Studies  to Air Pollution Across
     National Boundaries.  The  impact on the environment of sulfur in air  and precipitation,
     B.  Bolin,  ed., Sweden's  case study  for  the United Nations  conference on  the human en-
     vironment, Royal Ministry  of Foreign Affairs., Royal Ministry of  Agriculture, Stockholm,
     Sweden, 1972.  p. 20.

Odum,  E.  P.  Summary.   In:   Ecological  Effects  of Nuclear War,  Proceedings  of a Symposium,
     Ecological Society of America, Amherst, Massachusetts, August 1963.  G. M. Woodwell, ed.,
     BNL 917(C-43), Brookhaven National Laboratory, Upton, NY, August 1965.  pp. 69-72.

Odura, E.  P.   Fundamentals of Ecology.  Third  Edition.   W.  B.  Saunders Co., Philadelphia, PA,
     1971.  p. 5, 8-139.

Office of Air Quality Planning and Standards.   National Air Quality, Monitoring and Emissions
     Trends  Report,  1977.   EPA-450/2-78-052,  U.S.  Environmental  Protection Agency, Research
     Triangle Park, NC, December 1978.

Overrein,  L.  N.   Sulphur pollution patterns  observed; leaching  of  calcium in  forest soil
     determined.   Ambio. 1:145-147, 1972.
                                            7-118

-------
Pack, D. H., G.  J.  Ferber, J. L.  Heffter,  K.  Telegadas, J.  K.  Angell,  W. H. Hoecker, and L.
     Machta.  Meterology of  long-range  transport.   In:  Sulfur in the Atmosphere, Proceedings
     of  the  International  Symposium,   United  Nations  Environmental  Program  and  Others,
     Dubrovnik, Yugoslavia, September 7-14, 1977.  Atmos. Environ. 12:425-444, 1978.

Pack, D. W.  Sulfate  betiavior in eastern U.S.  precipitation.,.. Geophys.  Res. Lett. 5:673-674,
     1978.

Peterson, R. C.   Dynamics  of coarse particulate organic material processing  in diversified pH
     regimes.   Presented at  yie  Internal.  Conf.  The Ecological Impact of Acid Precipitation,
     March 11-14, 1980, Oslo-As,  Norway.

Pfeiffer, M.,  and  P.  Festa.   Acidity status of lakes  in the Adirondack region of New York in
     relation to  fish  resources.   Publication  FW-P168, New York State, Department of Environ-
     mental Conservation, Albany, NY, August 1980.

Potts, D. T. W., and C. Fryer.  The effect of pH and salt content on sodium balance in Daphnia
     magnet and Acantholeberis curvirostris (Crustacea; Cladocera).  J. Comp.  Physio!. Series B
     129:289-294, 1979.

Rough,  F.  H.   Acid precipitation  and  embryonic mortality of  spotted salamanders, Ambystoma
     maculatum.  Science (Washington, D.C.) 192:68-70, 1976.

Pough,  F.  H.,  and  R.  E.  Wilson.   Acid precipitation  and reproductive  success  of Ambystoma
     salamanders.  In:  Proceedings of the First International Symposium on Acid Precipitation
     and the  Forest Ecosystem,  May 12-15, 1975,  Columbus,  Ohio.   L. S.  Dochinger and T. A.
     Seliga, eds.   L).  S.  Forest Service  General Technical Report NE-23,  U.  S.  Department of
     Agriculture, Forest Service, Northeastern  Forest Experiment Station, Upper Darby, PA, 1976.

Prahm, L. P. , U. Torp, and R. M.  Stern.   Deposition and transformation rates  of sulphur oxides
     during atmospheric transport over the Atlantic.  Tellus 18:355-372, 1976.

Raddum, G.  Invertebrates:   quality and quantity as fish  food.   In:   Limnological Aspects of
     Acid  Precipitation,  Proceedings of  the  International Workshop,  U.S.  Environmental  Pro-
     tection Agency and Others, Sagamore, New York, September 25-28, 1978.  G. R. Hendry,  ed.,
     BNL 51074,  U.S.   Department of Energy, Brookhaven  National Laboratory  Upton, NY, 1978.
     pp. 17-24.

Rasmussen, R.  A.  Emission of biogenic hydrogen  sulfide.  Tellus 26:254-260,  1974.

Rheinheimer, G.  Aquatic Microbiology.  John Wiley & Sons, New York, 1971.  184 pp.

Robinson, E.,  R.  B.  Husar, and J.  N. Galloway.  Sulfur oxides in the atmosphere,  jn:  Sulfur
     Oxides.  National Academy of Sciences, Washington, DC, 1978.  pp. 18-62.

Roff, J. R.,  and R.  E.  Kvnatkowski.  Zooplankton  and zoobenthos communities of selected  nor-
     therr, Ontario lakes of different acidities.  Can. J. Zool. 55:899-911, 1977.

Rosenqvist, I.  T.  A contribution towards analysis of buffer properties of geological materials
     against strong acids  in precipitation water, report written for the Council for Research
     in Natural Sciences,  Norwegian General Research Council, 1976.

Rosseland,  B.  0., I.  Sevaldrud, D.  Svalastog, and I. P. Muniz.   Studies on freshwater  fish
     populations -  effects of acidification or reproduction, population structure, growth and
     food  selection.   In:   Ecological  Impact of Acid  Precipitation,  Proceedings of an Inter-
     national  Conference,  SNSF  Project, §andefjord, Norway, March 11-14, 1980, D. Drablris and
     A. Tollan, eds.,  SNSF Project, Oslo-As, Norway, October 1980.  pp. 336-337.
                                            7-119

-------
Schaffer, R. J.   The  Weathering of Natural Building Stones,  Spec. Ref. No. 18, London, HMSO,
     1932«  Cited in:   Sereda, 1977.

Scheider, W., J. Adamski, and M. Paylor.  Reclamation of Acidified Lakes Near Sudbury, Ontario
     Ministry of the Environment, Rexdale, Ontario, Canada, June 1975.

Scheider,  W.,   J.  Adamski,  and  M.  Paylor.   Reclamation  of Acidified  Lakes  Near Sudbury,
     Ontario. Ontario Ministry  of the Environment, Rexdale, Ontario, Canada, 1975.  Cited in:
     Wright and Gjessing, 1976.

Schindler, D. W.   Whole lake eutrophication experiments with phosphorus, nitrogen and carbon.
     Verh. Internet.  Verein. Limnol. 19:577-582, 1975.

Schindler, D. W.,  R.  Wagemann,  R.  B. Cook,  T.  Ruszczynski, and J. Prokopowich.  Experimental
     acidification of Lake  223,  experimental lakes area:  background data and the first three
     years of acidification.  Can. J. Fish. Aquat. Sci. 37:342-354, 1980.

Schlesinger, W.  H., and M.  M. Hasey.  The nutrient content of precipitation, dry fallout, and
     intercepted  aerosols   in   the  chaparral   of  southern  California.    Am.  Midi.   Nat.
     103:114-122, 1980.

Schofield, C. L.  Acid precipitation:  effects on fish.  Ambio. 5:228-230, 1976a.

Schofield, C. L.   Dynamics  and  Management of Adirondack Fish Populations.  I.   Natural Waters
     with Lethal  Conditions  for Fish.  I-a.  Acidification of Adirondack lakes  by atmospheric
     precipitation:  Extent  and  Magnitude of the Problem,  Final Report.  Project No, F-28-R,
     State of New York,  Department of Environmental Conservation, Albany, NY, 1976b.

Schofield, C. L.   Dynamics  and  Management of Adirondack Fish Populations:   I.  Natural Waters
     with Lethal  Conditions  for Fish.  I-b.  Acidification of Adirondack Lakes  by Atmospheric
     Precipitation: Long  Term and  Seasonal  Trends.Project  No.  F-28-R-4, State of New York,
     Department of Environmental Conservation, Albany, NY, 1976c.

Schofield, C. L.  The acid precipitation phenomenon and its impact in the Adirondack Mountains
     of New York  State.   In:   Scientific Paper from the Public Meeting on Acid  Precipitation,
     New York State Assembly Committee on Environmental Conservation and Others, Lake Placid,
     New York,  May 4-5, 1978.  H. H. Icord and J. S. Jacobson, eds., New York State Assembly,
     Science and Technology Staff, Albany, NY, March 1979. p. 86-91.

Schofield,  C.  L.  Effects  of acid  rain  on lakes.   Presented at  the  Session  on  Acid Rain,
     National Convention of the American Society of Civil Engineers, Boston, MA, April 2, 1979.

Schofield, C. L.  and J. R.  Trojnar.  Aluminum toxicity to brook trout (Salvelinus fontinali§)
     in acidified waters.   In:   Polluted Rain, Twelfth  International  Rochester Conference on
     Environmental Toxicity,  University of  Rochester, Rochester, New  York,  May 21-23, 1979.
     T. Y.  Toribara,  M.  W.  Miller,  and  P.  E. Morrow,  eds.,  Environmental  Science Research,
     Volume 17, Plenum Press, New York, NY, 1980.  pp. 341-366.

Sehmel, G. A.  Particle and gas dry deposition:  a review.  Atmos. Environ.  14:983-1011, 1980.

Seip,  H.  M.  Acidification  of  freshwater-sources and mechanisms.   In;   Ecological  Impact of
     Acid Precipitation, Proceedings of an International Conference, SNSF Project, Sandefjojjd,
     Norway, March 11-14,  1980,  D.  DrablizJs  and A.  Tollan,  eds.,  SNSF  Project,   Report-As,
     Norway, October 1980.  pp.  358-366.

Sereda,  P.  J.   Effects  of  sulphur  on  building materials.   lYi:    Sulphur  and  its  Organic
     Derivatives  in  the Canadian  Environment.   NRC  No.  15015,  National  Research  Council of
     Canada.   NRC Associate Committee on  Scientific  Criteria  For  Environmental Quality,
     Ottawa, Canada, 1977.  pp.  359-426.

                                            7-120

-------
Sheridan, R. P.,  and R.  Rosenstreter.  The  effect of hydrogen ion concentration in simulated
     rain on the moss fortyla rural is (Hedw) Sm,  Bryologist 76:168-173, 1973.

Shinn,  J.  H.,  and  S.  Lynn.   Do man-made  sources affect  the sulfur  cycle  of northeastern
     states?  Environ. Sci. Techno!. 13:1062-1067, 1979.

Shriner, D.  S.   Effects of simulated rain acidified with sulfuric acid on host-parasite  inter-
     actions.  In;  Proceedings of the First International Symposium on  Acid Precipitation and
     the Forest Ecosystem, May 12-15, 1975, Columbus, Ohio.  L. S. Dochinger and T. A. Seliga,
     eds.  U.  S.  Forest  Service  General  Technical  Report  NE-23,  U.  S.  Department of Agri-
     culture, Forest  Service,  Northeastern Forest Experiment  Station,  Upper Darby, PA, 1976.
     Pp. 919-925.

Shriner,  D.  S.   Effects  of simulated  acidic  rain  on  host-parasite   interactions  in plant
     diseases.   Phytopathology 68:213-218, 1978a.

Shriner,  D.  S.    Interactions  between acidic  precipitation  and S0? or  0.,:   Effects on plant
     response.   Phytopathol. News, 12:153, 1978b.

Shriner, D.  S.  Terrestrial  vegetation - air  pollutant interactions:    nongaseous  pollutants,
     w,et  deposition.   Presented  at  the International Conference on Air Pollutants and Their
     Effects on Terrestrial Ecosystems, Banff, Alberta, Canada, May 10-17, 1980.

Shriner,  D.  S.,  and G. S.  Henderson.   Sulfur distribution and cycling  in a deciduous  forest
     watershed.  J. Environ. Qua!. 7:392-397, 1978.

Shriner,  D.  S.,  M.  DeCot, and E.  B. Cowling.   Simulated acid rain caused direct injury to
     vegetation.  Proc. Am. Phytopath. Soc. 1:112, 1974.

Smith,  F. B.,  and G. H.  Jeffrey.   Airborne  transport of sulfur dioxide from the U.K.   Atmos.
     Environ. 9:643-659,  1975.

Smith, R. A.   Air and Rain:  The  Beginnings of Chemical Climatology.  Longmans, Green, London,
     1872.  600 pp.

Smith,  R.  L.  Ecology and Field Biology. 3rd Edition.   Harper and Row,  New York, NY, 1980.
     pp. 11-199.

SNSF Project.   Acid precipitation and some alternative sources as the cause of the acidifying
     water  courses.   Norway's Agrarian  Science  Research  Board,  Norway's  Technical   Nature
     Science Research Board, Dept. of Environmental Protection, ed., A.   Tollan, 1977.

Sprules, G. W.   Midsummer crustacean zooplankton communities in acid-stressed lakes.  J. Fish.
     Res. Board Can. 32:389-395,  1975.

Stensland, G.  J.   A comparison of  precipitation  chemistry data at a central Illinois site in
     1954 and  in  1977.   Presented at 71st Annual Meeting, Air  Pollution Control. Association,
     Houston, TX, June 25-26, 1978.  Paper No. 78-48.7.

Stensland,  G.  J. ,   Precipitation  chemistry  trends in  the  northeastern  United  States.  In:
     Polluted  Rain,  Twelfth International  Rochester  Conference  on Environmental Toxicity,
     University of Rochester,  Rochester,  New York,  May 21-23, 1979.   T.  Y.  Toribara, M. W.
     Miller, and  P.  E.  Morrow, eds., Environmental Science Research Volume 17,  Plenum  Press,
     New York, NY, 1980.   pp. 87-108.

Stokes,  P.   Phytoplankton  of acidic  lakes in  Killarney Ontario:   community structure related
     to  water  chemistry  abstracts of voluntary  contributions.   In:   Proc.  International Con-
     ference on the Ecological  Impact of Acid Precipitation, Sandefjord, Norway, March  11-14,
     1980, SNSF-project report, Oslo, Norway, 1980.

                                            7-121

-------
Stumm, W.,  and J.  J. Morgan.  Aquatic Chemistry, Wiley-Interscience, New York, NY, 1970.  583
     pp.

Sutcliffe, D. W., and T. R. Carrick.  Studies on mountain streams in the English  lake district
     I.  pH,  calcium, and  the  distribution of  invertebrates  in  the River Duddon.  Freshwater
     Biol. 3:437-462, 1973.

Tabatabai, M.  A.,  and J. M. Laflen.   Nitrogen  and sulfur content  and  pH  of precipitation in
     Iowa.  J. Environ.' Qual.  5:108-112, 1976.

Tamm, C.  0.,  G.  Wiklander, and B. Popovie.  Effects of applications of sulphuric acid to poor
     pine  forests.   In:   First  International  Symposium on Acid  Precipitation and the Forest
     Ecosystem, U.  S.  Department of Agriculture and  Others,  Columbus,  Ohio, May 12-15, 1975.
     Water Air Soil Pollut. 8:75-87, 1977.

Tiano, P.,  R.  Bianchi,  G.  Gargani, and S. Vannuci.  Research on the presence of  sulphur-cycle
     bacteria  in the stone of some historical  buildings  in Florence.   Plant Soil 43:211-217,
     1975.

Tomlinson,  G.  H.    Acidic  precipitation  and  mercury  In  Canadian  lakes  and  fish.   In:
     Scientific Paper  From the Public Meeting  on  Acid Precipitation,  New York State Assembly
     Committee on Environmental Conservation and Others, Lake Placid, New York, May 4-5, 1978.
     H.  H.  Izard and  J.  S. Jacobson,  eds.,  New York State  Assembly,  Science and Technology
     Staff, Albany, NY, March 1979.  pp. 104-118.

Traaen, T.  Effects of acidity on decomposition or organic matter in aquatic environments. In:
     Ecological Impact of Acid Precipitation, Proceedings of an International Conference, SNSF
     Project,  Sande/jord,  Norway,  March  11-14, 1980,  D.  Drablrfs  and  A.  Tollan,  eds.,  SNSF
     project, Oslo-As, Norway.   October 1980.  pp.  340-341.

Trojnar, J.  R.   Egg and larval survival  of  white  suckers (Catostomus  commersoni) at low pH.
     J. Fish. Res.  Board Can. 34:262-266, 1977a.

Trojnar, J.  R.   Egg hatchability and tolerance  of brook trout (Solvelinus fontinalis) fry at
     low pH.  J. Fish.  Res. Board Can. 34:574-579, 1977b.

Tukey,  H.  B., Jr.  «The  leaching  of  substances  from  plants.    Annu.  Rev.   Plant  Physiol.
     21:305-324, 1970.

Turk, J. T., and N. E. Peters.   Acid-Rain Weathering of a Metasedimentary Rock Basin, Herkimer
     County, New York.   Open file Report 77-538, U. S. Department  of the Interior, Geological
     Survey, Washington, DC, July 1977.

Turk,  J.  T., and N.  Peters.   Acid-Rain Weathering of a  Metasedimentary  Rock Basin,  Herkimer
     County, New York.   In:   Scientific Paper  From  the  Public Meeting on Acid Precipitation,
     New York  State Assembly Committee on Environmental Conservation and Others, Lake Placid,
     New York, May  4-5,  1978.   H. H.  Izard and J.  S. Jacobson, eds., New York State Assembly,
     Science and Technology Staff, Albany, NY, March 1979.  pp. 136-145.

Turner, J.,  and  M.  Lambert.   Sulfur nutrition of forests.  Jn:  Atmospheric Sulfur Deposition
     Environmental   Impact  and  Health  Effects.   Proceedings  of  the Second  Life Science Sym-
     posium, Oak Ridge  National  Laboratory and Others, Gatlinburg, Tennessee, October 14-18,
     1979.  D. S.  Shriner, C.  R. Richmond,  and S.  E. Lindberg, eds. ,  Ann  Arbor Science Pub-
     lishers, Inc., Ann Arbor,  MI, 1980.  pp. 321-333.
                                            7-122

-------
Tyler, G.  Heavy metals pollute nature, may reduce productivity, Ambio, 1:52-59, 1972.

Tyler, G.   Leaching rates  of heavy metal  ions  in  forest soil.   Water  Air  Soil  Pollut.  9_:
     137-148, 1977.

Ulrich,  B.   Environmental  influences ,on the  nutrient cycle pf a  beech"  forest on acid soil.
     Forstwiss.  Centralbl. 94:280-287, 1975.

U.S.  Bureau  of Mines.   Minerals  Yearbook  1952.   Volume  II.   Area Reports:   Domestic.  U.S.
     Department of the Interior, Washington, DC, 1954.  p. 117.

U.S.  Bureau  of Mines.   Minerals  Yearbook  1974.   Volume  II.   Area Reports:  ^Domestic.  U.S.
     Department of the Interior, Washington, DC, 1976.  p. 392.

U.S. Department of Commerce.   Climatic Atlas of the United States.  National Climatic  Center,
     Asheville, NC, June 1968.

U.S.  Department  of  Energy.   Acid  Rain  Information  Book.   Final  Report  DOE/EP-0018, U.  S.
     Department of Energy, Washington, DC, May 1981. pp. 2-12;  2-13.

U.S.  Environmental  Protection Agency.   Air Quality  Criteria  for  Oxides of Nitrogen.  Draft
     Final.   EPA-600/8-82-026,  U.S. Environmental  Protection Agency,.Research Triangle Park,
     NC, September, 1982.

Walters, C.  J.,  and R, E.  Vincent.   Potential  productivity of an  alpine lake  as  indicated by
     removal and reintroduction of  fish.  Trans. Amer. Fish Soc. 102:675-697, 1973.

Wetzel,  R.  G.  Limnology,  p. 287-621,  W.  B.  Saunders Co.,  Philadelphia,  PA,  1975.    743 pp.

Whelpdale,  D.  M.   Acidic  Deposition.   Ecological Effects of  Acid  Precipitation,  Report of  a
     Workshop, Electric Power Research Institute, Gatehouse-of-Fleet, Galloway, Scotland, Sep-
     tember  4-7,  1978.   M. J.  Wood,  ed., EPRI SDA77-403,  Electric Power Research Institute,
     Palo Alto, CA, July 1979.  Section 4.1.  22 pp.

Wiklander,  L.   Interaction between cations and  anions  influencing adsorption and  leaching.
     In:   Effects  of  Acid Precipitation  on  Terrestrial  Ecosystems,  North  Atlantic Treaty
     Organization, Toronto, Ontario,  Canada, May 21-27, 1978.  T.  C. Hutchinson and  M. Havas,
     eds., Plenum  Press, New York,  NY, 1980.  pp. 239-254.

Wiklander,  L.   Leaching  and Acidification of Soils.   In:   Ecological Effects of Acid Precipi-
     tation.   Report  of  a workshop,  Electric Power  Research Institute, Gatehouse-of-Fleet,
     Galloway, Scotland, September  4-7, 1978.  M. J. Wood, ed., EPRI SOA77-403, Electric Power
     Research  Institute,  Palo Alto, CA, July 1979.  Section 4.3.  24 pp.

Williams,  T.   Acid rain:   the California context.  Citizens  for Better Environ.:  6-8,  10, May
     1978.

Winkler, E.  M.  Important agents of weathering  or building and monumental stone.  Eng. Geol.
     1:381-400, 1966.

Wisniewski,  J. and  E.  L.  Keitz.  Acid rain deposition  patterns  in  the  continental United
     States.   Submitted to  the Bull. Amer. Meteoro, Soc.   1981.

Wolff,  G.  T., P.  J.  Lioy, H. Golub,  and  J.  S.  Hawkins.  Acid precipitation in the New York
     Metropolitan  Area:    Its  relationship to  meteorological  factors.   Env.  Sci.   Techno!.
     13:209-212, 1979.
                                            7-123

-------
Wood,  T.   Acid  Precipitation.   In;   Sulfur  in  the Environment.  Revised Edition.   Missouri
     Botanical Garden, St. Louis, MO, July 1975.  pp. 39-50.

Wood, T., and F. H. Bormann.  The effects of  an artificial  acid  mist  upon the growth of Betula
     alleghaniensis Britt. Environ. Pollut. 7=259-267,  1974.

Wood, T., and F.  H. Bormann.   Increases  in  foliar  leaching  by  acidification of an artificial
     mist.  Ambio 4:169-171, 1975.

Wood, T., and F.  H. Bormann.   Short-term effects of a  simulated acid rain upon the growth and
     nutrient relations  of Pinus strobus L.   In:   First International Symposium on Acid Pre-
     cipitation  and  the  Forest  Ecosystem,   U.  S.  Department of  Agriculture  and  Others,
     Columbus, Ohio, May 12-15,  1975.  Water,  Air Soil  Pollut. 7:479-488,  1977.

Woodwell, G, M.  Effects of  ionizing radiation on terrestrial  ecosystems.  Science (Washington,
     DC) 138:572-577, 1962.

Woodwell, G.  M.   Effects of pollution on the structure and physiology of ecosystems.   Science
     (Washington, DC) 168:429-433, 1970.

Wright, R.  F., and E,  T.  Gjessing.   Changes  in  the  chemical  composition  of  lakes.   Ambio.
     5:219-223, 1976.

Wright, R.  F.,  and A. Henriksen.  Chemistry  of small  Norwegian lakes, with  special reference
     to acid precipitation.  Limnol, Oceanogr. 23:487-498,  1978.

Wright, R.  F., and E.  Snekvik.   Acid precipitation:   chemistry and fish populations  in 700
     lakes  in southernmost Norway.   Verh. Int. Ver. Theor. Angew.   Limnol.  20:765-775, 1978.

Wright, R.  F., N.  Conroy, W.  T.  Dickson, R.  Harriman, A.  Henriksen,  and  C.   L.  Schofield.
     Acidified  lake  districts   of  the world:  a  comparison  of water chemistry of  lakes  in
     southern Norway,  southern Sweden, southwestern Scotland, the  Adirondack Mountains of New
     York,  and southeastern Ontario.   In:    Ecological  Impact  of  Acid Precipitation,  Pro-
     ceedings of  an International Conference, SNSF  Project,  Sajidefjord,   Norway,  March 11-14,
     1980,  D.  Drablrfs and A. Tollan, eds., SNSF project, Oslo-As,  Norway, October 1980b.  pp.
     377-379.

Wright, R.  F.,  T.  Dale, E. T.  Gjessing, G. R. Hendry,  A. Henriksen,  M. Johannessen, and I.  P.
     Huniz,   Impact  of  acid  precipitation  on freshwater  ecosystems  in Norway.   In:   First
     International  Symposium on Acid Precipitation  and the Forest  Ecosystem, U.  S.  Department
     of  Agriculture  and  Others,  Columbus,   Ohio,   May  12-15,  1975.   Water  Air  Soil  Pollut.
     6:483-499, 1976.

Wright,  R.  F., R.  Harriman, A.  Henriksen,   B. Morrison,   and L.  A.   Caines.   Acid  lakes and
     streams  in  the  Galloway  area,  southwestern Scotland.   In:  Ecological  Impact  of Acid
     Precipitation,  Proceedings  of an  International  Conference,  SNSF   Project,  Sandefjord,
     Norway,  March  11-14,  1980,  D. DrabljJs and A. Tollan, eds.,  SNSF  project, Oslo-As, Norway,
     October  1980a.   pp. 248-249.

Van, N. D., and P.  Stokes.   Phytoplankton of  an acidic  lake,  and its  responses to experimental
     alterations of pH.  Environ. Conserv. 5:93-100, 1978.  '
                                             7-124  
-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing]
1 REPORT NO
 EPA-600/8-82-029b
                                                           3 RECIPIENT'S ACCESSION NO
4 TITLE AND SUBTITLE
 Air Quality Criteria  for Participate Matter
 and Sulfur Oxides.  Volume II.
             5 REPORT DATE
                December 1982
             6 PERFORMING ORGANIZATION CODE
7 AUTHOR(S)

 See list of Authors,  Contributors, and Reviewers
                                                           8 PERFORMING ORGANIZATION REPORT NO
9 PERFORMING ORGANIZATION NAME AND ADDRESS
 U.S. Environmental  Protection Agency
 Environmental  Criteria and Assessment Office
 MD-52
 Research Triangle  Park, NC  27711	
                                                           10 PROGRAM ELEMENT NO
             11 CONTRACT/GRANT NO
12 SPONSORING AGENCY NAME AND ADDRESS
 U.S. Environmental  Protection Agency
 Office of Research  and Development
 Office of Health and Environmental Assessment
 401 M Street,  SH, Washington, DC  20460  	
                                                           13 TYPE OF REPORT AND PERIOD COVERED
                FINAL
             14 SPONSORING AGENCY CODE
                EPA/600/00
15. SUPPLEMENTARY NOTES
16 ABSTRACT

 The document  evaluates and assesses scientific  information on the health and welfare
 effects  associated with exposure to various concentrations of sulfur oxides and
 particulate matter in ambient air.  The literature  through 1980-81 has been reviewed
 thoroughly for information relevant to air quality  criteria, although the document
  is not  intended as a complete and detailed review of all  literature pertaining to
 sulfur  oxides and particulate matter.  An attempt has been made to identify the major
 discrepancies in our current knowledge and understanding  of the effects of these
 pol1utants.

 Although this document is principally concerned with the  health and welfare effects of
 sulfur  oxides and particulate matter, other scientific data are presented and evalu-
 ated  in  order to provide a better understanding of  these  pollutants in the environment
 To this  end,  the document includes chapters that discuss  the chemistry and physics
 of the  pollutants; analytical techniques; sources;  and types of emissions; environ-
 mental  concentrations and exposure levels; atmospheric chemistry and dispersion
 modeling; acidic deposition; effects on vegetation; effects on visibility, climate,
 and materials; and respiratory, physiological,  toxicological, clinical and
  epidemiological aspects of human exposure.
                               KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b IDENTIFIERS/OPEN ENDED TERMS
                           c  cos ATI field/Group
18 DISTRIBUTION STATEMENT

       RELEASE UNLIMITED
19 SECURITY CLASS /This Report)
  UNCLASSIFIED
21 NO OF PAGES
      625
                                              20 'SFCUJRITY CLASS fThlSO&tei
                                               UNCLASSIFIED
                                                                        22. PRICE
iPA Farm 2220-1 (Rsv, 4-77)   PREVIOUS EDITION is OBSOLETE

-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
    Office of Research and Development
 Center for Environmental Research Information
          Cincinnati, Ohio 45268

          OFFICIAL BUSINESS
   PENALTY FOR PRIVATE USE. S3OO
  AN EQUAL OPPORTUNITY EMPLOYER
                                     If your address is incorrect, please change on the above label
                                     tear off; and return to the above address.
                                     If you do not desire to continue receiving these technical
                                     reports, CHECK HEREd; fear off label, and return it to the
                                     above address.
                                    Publication No. EPA-600/8-82-029b

-------