Draft
Do Not Quote or Cite
                                External Review Draft No. 2
                                           December 1980
             Air Quality Criteria
           for Particulate  Matter
              and Sulfur Oxides

                                  t
                      Volume IV
                           NOTICE

This document is a preliminary draft. It has not been formally released by EPA and should not at this stage be
construed to represent Agency policy. It is being circulated for comment on its technical accuracy and
policy implications.
            Environmental Criteria and Assessment Office
           Office of Health and Environmental Assessment
               Office of Research and Development
              U.S. Environmental Protection Agency
               Research Triangle Park, N.C. 27711

-------
                              NOTE TO READER

     The  Environmental  Protection Agency  is  revising the existing criteria
documents for particulate matter and sulfur oxides (PM/SO ) under Sections 108
and 109  of  the Clean Air Act,  42  U.S.C.  §§ 7408, 7409.   The first external
review draft of a revised combined PM/SO  criteria document was made available
                                        i\
for public comment earlier this year.
     The Environmental Criteria and  Assessment Office (ECAO)  filled more  than
4,000 public requests for copies of  the first external review draft.   Because
all those who received copies of the first draft from ECAO will  be sent copies
of the second  external  review draft, there is no need to resubmit a request.
     To facilitate public review,  the  second external review draft  will  be
released in five volumes on  a staggered schedule as the volumes  are completed.
Volume I (containing Chapter 1) Volume II  (containing Chapters 2, 3,  4, and 5)
Volume III (containing Chapters 6, 7, and 8) Volume IV (containing Chapters  9
and 10) and Volume V (containing Chapters  11, 12, 13,  and 14) will be released
during January-February, 1981.  As  noted  earlier, they will  be released  as
volumes are completed, not in numerical  order by volume.
     The first external review draft was announced in the Federal Register of
April 11, 1980 (45 FR 24913).  ECAO received and reviewed 89 comments from the
public, many of which were quite extensive.  The Clean Air Scientific Advisory
Committee (CASAC)  of the Science  Advisory Board also provided advice and
comments on  the  first external review draft at  a public meeting of August
20-22, 1980 (45 FR 51644, August 4, 1980).
     As with the first external review draft, the second external review draft
will  be  submitted to  CASAC for its advice and  comments.   ECAO is also
soliciting written  comments from the public  on  the  second external review
draft  and requests  that  an original and  three  copies of all  comments be
submitted  to:    Project  Officer for PM/SO ,  Environmental  Criteria  and
                                           J\
Assessment  Office,   MD-52,  U.S. Environmental Protection  Agency, Research
Triangle Park, N. C. 27711.   To facilitate ECAO's consideration of comments on
this lengthy and complex document, commentators with  extensive comments should
index the major  points  which they  intend  ECAO to address,  by providing a  list
of  the major  points and a  cross-reference  to the pages in  the  document.
Comments should be submitted during  the forthcoming comment period, which will
be announced in  the Federal Register once all volumes of the second external
review draft'are available.
SOX9A/C                                                                    12-23-80

-------
Draft
Do Not Quote or Cite
                                External Review Draft No. 2
                                          December 1980
                ir Quality  Criteria
           for  Particulate Matter
              and  Sulfur  Oxides
                      Volume IV
                          NOTICE

This document is a preliminary draft. It has not been formally released by EPA and should not at this stage be
construed to represent Agency policy. It is being circulated for comment on its technical accuracy and
policy implications.
          Environmental Criteria and Assessment Office
          Office of Health and Environmental Assessment
              Office of Research and Development
             U.S. Environmental Protection Agency
              Research Triangle Park, N.C. 27711

-------
                                PREFACE

     This document is a revision of External Review Draft No. 1, Air
Quality Criteria for Particulate Hatter and Sulfur Oxides, released in
April 1980.   Comments received during a public comment period from April
15, 1980 through July 31, 1980, and recommendations made by the Clean Air
Scientific Advisory Committee in August have been addressed here.
     Volume IV contains Chapters 9 and 10 which cover effects on visibility
and climate and effects on materials.   A Table of Contents for Volumes I,
II, III, and V follows.
                                    11

-------
                                  CONTENTS

                        VOLUMES I, II, III, IV, AND V

                                                                      Page

Volume I.
     Chapter 1.     Executive Summary	       1-1

Volume II.
     Chapter 2.     Physical and Chemical Properties of Sulfur
                    Oxides and Particulate Matter	       2-1
     Chapter 3.     Techniques for the Collection and Analysis of
                    Sulfur Oxides, Particulate Matter, and Acidic
                    Precipitation	       3-1
     Chapter 4.     Sources and Emissions	       4-1
     Chapter 5.     Environmental Concentrations and Exposure	       5-1

Volume III.                                      '
     Chapter 6.     Atmospheric Transport, Transformation and
                    Deposition	       6-1
     Chapter 7.     Acidic Deposition	       7-1
     Chapter 8.     Effects on Vegetation	       8-1

Volume IV.
     Chapter 9.     Effects on Visibility and Climate	       9-1
     Chapter 10.    Effects on Materials	      10-1

Volume V.
     Chapter 11.    Respiratory Deposition and Biological Fate
                    of Inhaled Aerosols and SO-	      11-1
     Chapter 12.    Toxicological Studies	      12-1
     Chapter 13.    Control led Human Studies	      13-1
     Chapter 14.    Epidemiological Studies of the Effects of
                    Atmospheric Concentrations of Sulfur Dioxide
                    and Particulate Matter on Human Health	      14-1
                                    111

-------
                              TABLE OF CONTENTS
9.   EFFECTS ON VISIBILITY AND CLIMATE	     9-1
     9.1  INTRODUCTION	     9-1
     9.2  THE VALUE OF VISIBILITY	     9-1
     9.3  PERCEIVED EFFECTS ON VISIBILITY	     9-3
     9.4  FUNDAMENTALS OF ATMOSPHERIC VISIBILITY	     9.7
          9.4.1  Measurement Methods	    9-12
                 9.4.1.1  Human Observer (Total  Extinction)	    9-14
                 9.4.1.2  Photography (Total  Extinction)	    9-14
                 9.4.1.3  Telephotometry (Total  Extinction)		    9-14
                 9.4.1.4  Long-path Extinction (Total  Extinction)	    9-15
                 9.4.1.5  Nephelometer (Scattering)	    9-15
                 9.4.1.6  Light Absorption Coefficient	    9-16
          9.4.2  Role of Particulate Matter in Visibility Impairment	    9-16
                 9.4.2.1  Rayleigh Scattering	    9-16
                 9.4.2.2  Nitrogen Dioxide Absorption	    9-16
                 9.4.2.3  Particle Scattering	    9-17
                 9.4.2.4  Particle Absorption	    9-27
                 9.4.2.5  Summary	    9-28
          9.4.3  Chemical Composition of Atmospheric Particles	    9-28
                 9.4.3.1  Summary	    9-32
     9.5  HISTORICAL PATTERNS OF VISIBILITY	    9-32
          9.5.1  Natural Versus Manmade Causes	    9-41
          9.5.2  Summary	    9-45
     9.6  SOLAR RADIATION	    9-45
          9.6.1  Spectral and Directional  Quality of Solar Radiation	    9-48
          9.6.2  Total Solar Radiation:   Local to Regional Scale	    9-51
          9.6.3  Radiative Climate:   Global Scale	    9-53
     9.7  CLOUDINESS AND PRECIPITATION	".	    9-55
     9.8  SUMMARY	•>.	    9-58
     9.9  REFERENCES	-.	    9-60

10.   EFFECTS ON MATERIALS	'	     10-1
     10.1 INTRODUCTION	     10-1
     10.2 SULFUR OXIDES	     10-4
          10.2.1 Corrosion of Exposed Metals	     10-4
                 10.2.1.1 Physical and Chemical  Considerations	     10-4
                          10.2.1.1.1 Relative Humidity and
                                     Corrosion Rate	      10-5
                          10.2.1.1.2 Influence of Rainfall on
                                     Corrosion	      10-7
                          10.2.1.1.3 Influence of Temperature
                                     on Corrosion	      10-9
                          10.2.1.1.4 Hygroscopicity of Metal
                                     Sulfates	      10-9
                          10.2.1.1.5 Electronic Conductivity of Rust	     10-11
                          10.2.1.1.6 Cathodic Reduction of Rust	     10-11
                          10.2.1.1.7 Corrosion-Protective Properties
                                     of Sulfate in Rust	     10-11
                                   iv

-------
            10.2.1.2 Effects of Sulfur Oxide Concentrations
                     on the Corrosion of Exposed Metals ............     10-12
                     10.2.1.2.1 Ferrous Metals .....................     10-12
                     10.2.1.2.2 Laboratory and Field Studies
                                Emphasizing Ferrous Metals .........     10-13
                                10.2.1.2.2.1  Laboratory Studies       10-13
                                10.2.1.2.2.2  Field Studies ........     10-15
                     10.2.1.2.3 Comparison of Ferrous and
                                Nonferrous Metals ..................     10-20
     10.2.2 Protective Coatings ....................................     10-23
            10.2.2.1 Zinc-Coated Materials .........................     10-23
            10.2.2.2 Paint Technology and Mechanisms of Damage .....     10-27
     10. 2. 3 Fabrics ................................................     10-30
     10.2.4 Building Materials .....................................     10-34
            10. 2.4. 1 Stone .........................................     10-34
            10. 2. 4. 2 Cement and Concrete ...........................     10-35
     10.2.5 Electrical Equipment and Components .....................    10-35
     10. 2. 6 Paper ......................... , .........................    10-35
     10. 2. 7 Leather .................................................    10-36
     10.2.8 Elastomers and Plastics ........ : ........................    10-36
     10.2.9 Works of Art ............................................    10-37
    10.2.10 Review of Damage Functions Relating S09 to Material
            Damage ................................ r .................    10-37
10. 3 PARTICIPATE MATTER .............................................    10-39
     10.3.1 Corrosion and Erosion ...................................    10-39
     10.3.2 Soiling and Discoloration ............. -. .................    10-41
            10.3.2.1 Building Materials .............................    10-41
            10. 3. 2. 2 Fabrics ........................................    10-41
            10. 3. 2. 3 Household and Industrial Paints ................    10-41
10.4 SUMMARY, PHYSICAL EFFECTS OF SULFUR OXIDES AND PARTICULATE
     MATTER ON MATERIALS ............ : ..............................     10-43
10.5 ECONOMIC DAMAGE OF AIR POLLUTION TO MATERIALS - S0y AND
     PARTICULATE MATTER ................................ . ...........     10-45
     10. 5.1 Introduction ............. : .............................     10-45
     10. 5. 2 Economic Damage to Materials ...........................     10-46
            10.5.2.1 Metals ........................................     10-46
            10.5.2.2 Paints ........................................     10-48
            10.5.2.3 Economic Cost of Soiling ......................     10-49
            10.5.2.4 Combined Studies ..............................     10-52
     10.5.3 Summary of Economic Damage of Air Pollution
            to Materials. . .........................................     10-53
10. 6 SUMMARY AND CONCLUSlbNS , EFFECTS ON MATERIALS .................     10-54
10. 7 REFERENCES .......... . .................. ......... J- .............     1°~56
GLOSSARY

-------
                               LIST OF FIGURES

 figure                                                                    Page
9-1.   Map shows median yearly visual range (miles) and isopleths
       for suburban/nonurban areas, 1974-76	   9-6

9-2.   Median summer visual range (miles) and isopleths for suburban/
       non-urban areas, 1974-76	   9-6

9-3.   (A) A schematic representation of atmospheric extinction,
       illustrates (i) transmitted, (ii) scattered, and (iii) absorbed
       light.  (B) A schematic representation of daytime visibility
       illustrates: (i) residual light from target reaching observer,
       (ii) light from target scattered out of observer's Tine of
       sight, (iii) air light from intervening atmosphere and (iv) air
       light constituting horizon sky	   9-9

9-4    The apparent contrast between object and horizon sky decreases
       with increasing distance from the target	  9-10

9-5.   Measured apparent contrast of farthest visibility marker was
       identified in 1000 determinations of visual range by 10
       observers	  9-11

9-6.   Inverse proportionality between visual range and the scattering
       coefficient, bscat, was measured at the point of observation	  9-13

9-7.   For a light-scattering and absorbing particle, the scattering per
       volume has a strong peak at particle diameter of 0.5 urn (m=1.5-
       0.05i; wavelength = 0.55um)	  9-18

9-8.   Calculated scattering cross-section per unit mass at a wave-
       length of 55 Mm for absorbing and nonabsorbing materials is shown
       a function of diameter for single-sized particles.  (B) Calculated
       absorption cross-section per unit mass at 0.55 urn for single-
       sized particles of carbon and iron (C) Calculated extinction
       cross-section per unit mass at 0.55 pro for single-sized particles
       of carbon, iron, silica and water	  9-20

9-9.   For a typical aerosol volume (mass) distribution, the calculated
       light-scattering coefficient is contributed almost entirely by
       the size range 0.1-1.0 urn	  9-21

9-10.  Scattering-to-volume ratios are given for various size
       distributions	  9-22

9-11.  Simultaneous monitoring of b   .  and fine-particle mass in
       St. Louis in April 1973 showed a high correlation coefficient of
       0.96, indicating that b   t depends primarily on the fine
       particle concentration	  9-23
                                   v1

-------
9-12.  Aerosol mass distributions, normalized by the total mass, for
       New York aerosol at different levels of light-scattering
       coefficient show that at high background visibility, the fine-
       particle mass mode is small compared with the coarse-particle
       mode	   9-25

9-13.  Humidograms for a number of sites show the increase in b   which
       can be expected at elevated humidities for specific sitelpor
       aerosol types	   9-31

9-14.  Historical trends in hours of reduced visibility at Phoenix
       and Tucson are compared with trends in SO  emissions from
       Arizona copper smelters	   9-33
9-15.  Seasonally adjusted changes in sulfate during the copper strike
       are compared with the geographical distribution of smelter SO
       emi ssions	   9-35
9-16.  Seasonally adjusted percent changes in visibility during the
       copper strike compared with the geographical distribution of
       smelter SO  emissions		.	   9-36
                 A

9-17   Compared here are summer trends of U.S. coal consumption and
       Eastern United States average extinction coefficient	   9-38

9-18.  In the 1950's the seasonal coal consumption peaked in the winter
       primarily because of increased residential and railroad use.
       By 1974, the seasonal pattern of coal usage was determined by the
       winter and summer peak of utility coal usage	   9-38

9-19.  In 1974, the U.S. winter coal consumption was well below, while
       the summer consumption was above the 1943 peak.  Since 1960, the
       average rate of summer consumption was 5.8 percent per year, while
       winter consumption'increased at only 2.8 percent per year	   9-39

9-20.  Trends in coal consumption in the continental United States
       are shown by region	   9-40

9-21.  The spatial distribution of 5-year average extinction
       coefficients shows the substantial increases of third-quarter
       extinction coefficients in the Carolinas, Ohio River Valley and
       Tennessee-Kentucky area.1	   9-42

9-22.  Average annual number of days with occurrence of dense fog	 9-44

9-23.  Annual percent frequency of occurrence of wind-blown dust when
       prevailing visibility was 7 miles or less, 1940-1970	 9-44

9-24.  Solar radiation intensity spectrum at sea level in cloudless sky
       peaks in the visibile window, 0.4-0.7 urn wavelength range, shows
       that in clean remote locations, direct solar radiation
       contributes 90 percent and the skylight 10 percent of the
       incident radiation on a horizontal surface	   9~46
                                   vii

-------
9-25.  Extinction of direct solar radiation by aerosols is depicted	   9-47

9-26.  On a cloudless but hazy day in Texas, the direct solar radiation
       intensity was measured to be half that on a.clear day, but
       most of the lost direct radiation has reappeared as skylight	   9-49

9-27.  To interpret these monthly average turbidity data in terms of
       aerosol effects on transmission of direct sunlight, use the
       expression l/ln = 10  , where B is turbidity and l/ln is the
       fraction transmitted	   9-50

9-28.  Seasonal turbidity patterns for 1961-66 and 1972-75 are shown for
       selected regions in the Eastern United States.	   9-52

9-29.  Analysis of the hours of solar radiation since the 1950's shows
       a decrease of summer solar radiation over the Eastern United
       States	   9-54

9-30.  Numbers of smoke, haze days are plotted per 5 years at Chicago,
       with values plotted at end of 5-year period	   9-57

10-1   Relationship among emissions, air quality, damages and benefits,
       and policy decisions	    10-2
10-2   Steel corrosion behavior is shown as a function of average
       relative humidity at three average concentration levels of
       sulfur dioxide	    10-6
10-3   Steel corrosion behavior is shown as a function of average
       sulfur dioxide concentration and average relative humidity 	    10-8
10-4   Empirical  relationship between average relative humidity and
       fraction of time relative humidity exceeded 90 percent is
       shown for data from St.  Louis International Airport 	   10-10
10-5   Relationship between corrosion of mild steel and corresponding
       mean S02 concentration is shown for seven Chicago sites 	   10-16
10-6   Adsorption of sulfur dioxide on polished metal  surfaces is shown
       at 90 percent relative humidity	   10-21
10-7   Relationship between retained breaking strength of cotton
       print cloth samples and S0? and sulfation rate at selected
       sites in St. Louis and Chicago.	   10-33
10-8   Representation of.soiling of acrylic emulsion house paint
       as a function of exposure time and particle concentrations	   10-44
10-9   Summary of economic damage estimates of air pollution to
       materials	   10-55
                                   viii

-------
                              TABLES
Number                                                                Page

9-1  Light scattering per unit mass of fine aerosol	     9-24
9-2  Correlation/regression analysis between airport
     extinction and copper smelter SO	     9-34
9-3  Some solar radiation measurements in the Los Angeles area	     9-53

10-1 Some empirical expressions for corrosion of exposed
     ferroalloys 	    10-19
10-2 Critical humidities for various metals 	    10-22
10-3 Experimental regression coefficients with estimated standard
     deviations for small zinc and galvanized steel speciments
     obtained from six exposure sites 	    10-25
10-4 Corrosion rates of zinc on galvanized steel products exposed
     to various environments prior to 1954	    10-26
10-5 Paint erosion rates and T-test probabilify data 	    10-31
10-6 Selected physical damage functions related to S02 exposure	    10-38
10-7 Results of regression for soiling of building materials as a
     function of TSP dose 	    10-42
10-8 Summation of annual extra losses due to corrosion damage by air
     pollution to external metal structures for 1970 	    10-47
10-9 Estimates of total costs from air pollution damage to
     materials in 1970 	    10-53
                                  ix

-------
                      9.  EFFECTS ON VISIBILITY AND CLIMATE

9.1  INTRODUCTION
     Airborne particles  are  among the most conspicuous  trace  constituents  of  the atmosphere;
their presence can be detected from afar by the unaided eye.  Particles scatter and absorb  visi-
ble radiation, redirecting a portion of it to form "air light," which obscures  distant objects.
As such, degradation  of visual air quality is unique in that it is an effect that can be seen
by everyone.  Section 9.2 outlines attempts to quantify the value of visibility to individuals
and to society.   Section 9.3 discusses the subjectively perceived effects of particles on visi-
bility.  Section  9.4  covers  the objective methods for measuring aerosols' influences on light
transmission through the atmosphere, and the fundamental physics involved.  Historical patterns
are discussed briefly in Section 9.5.
     Particles also indirectly affect the properties of  the  atmosphere.   For  example, essen-
tially all  atmospheric  water vapor condensation occurs  by' nucleation, that is, water deposi-
tion onto  cloud  condensation nuclei or ice  nuclei.   Addition of such particles to the atmos-
phere  therefore  affects cloudiness, which in turn influences insolation.  Section 9.6 of this
chapter discusses the effects of particles on climate.
     The incorporation  of  aerosols into cloud and fog droplets can change the quality as well
as the quantity  of precipitation as discussed  in  Section 9.7.  The mechanisms and effects of
acid precipitation are discussed in Chapter 8.
     In 1977,  the National  Science Foundation  Workshop on  Inadvertent  Weather Modification
(Robinson, 1977) identified five major areas in which effects on the atmosphere can be attrib-
uted  to  presence  of  particles:   visibility,  solar  radiation,  cloudiness,  precipitation
quantity,  and  precipitation quality.   On the basis  of  current scientific knowledge, weather
phenomena were assessed to determine the likelihood and qualitative scale of their inadvertent
modification by  human  activities.   Evidence of major  inadvertent  modification is considered
scientifically established for three areas: changed precipitation quality, increased haze,  and
increased  cloudiness.   Evidence of  moderate inadvertent modification is established for  de-
creases in solar  radiation  and sunshine  and  for urban-scale  increases in quantity of pre-
cipitation  and  thunderstorms.   There  is  probable  but not yet  firmly  established evidence of
inadvertent modifications in mesoscale increases or decreases in cloudiness and thunderstorms.
9.2  THE VALUE OF VISIBILITY
     Reductions  in visibility  can adversely affect transportation safety, property value,  and
aesthetics.   Impaired  visibility has  triggered automobile accidents  when  smoke stack plumes
smothered  a  roadway.  When visibility drops below 3 miles, FAA regulations restrict flight in
controlled  air  to aircraft  equipped with  IFR instrumentation  (14  CFR  91.105).   Aside froa
these  hazards,'  reduced visibility carries  social   and economic  costs.  The  value  of  the
"wilderness  experience" and  the  importance of  preserving  natural  heritage have  long been

SOX9A/A                                      9-1                                       12-23-80

-------
recognized  in  the United  States (Environmental Protection  Agency,  1979).   Accordingly,  the
unique charm of  the  expansive, scenic vistas of the  Southwest suggests that long-range visi-
bility may  be  what economists call a public  good—an asset  held in common  by  the  population
(Rowe et  al.,  1980).   The key issue concerning visibility  is the amount of its value nation-
wide.
                                                           i
     To establish  visibility values, it  is  necessary to develop ways  to quantify  visibility
impairment as  perceived  by the human eye.  The Park  Service, in a brochure entitled "My Eyes
Need a  Good Stretching"  (NPS, 1978), attempted to summarize this relationship  between visi-
bility and  experience  by relating the views  of artists, humanists, and scientists on the sub-
ject of visibility degration.  According to the Environmental Protection Agency (1979), estab-
lishing visibility values must, therefore, involve  relating the whole visual experience to in-
dices such  as  visual  range,  contrast transmittance,  and color alteration.   Although the value
of visibility  may  prove  intangible,  it  is conjectured  that  it is to some extent quantifiable
(can be identified as  a  discrete point  on a  scale) and, hence, can be generalized for display
to the public.   Given  the lack of consistent units  for the  evaluation of aesthetic qualities
such as visibility, values can be categorized according to:   (1) economic criteria,  the dollar
cost/ benefit  associated with  visibility;  (2) psychological  criteria,  the individual need and
benefits resulting from  visibility;  and  (3)  social/political criteria, community opinions and
attitudes held in  common  with regard to visibility  (Environmental  Protection  Agency, 1979).
     Economists  have  made  some  progress toward quantifying the  values of visibility using
dollars as  the measure.   While market values and  prices reflect the marginal  value of visi-
bility, total  value must be estimated from revealed consumer preference, either in a market or
market-like similation.   According to Rowe et  al.  (1980),  one approach that  has appeared in
the literature is to estimate directly an individual's economic values for a unit reduction in
visibility, usually  drawn from  the  work of  Bohm  (1971), Bradford  (1970),   and Davis (1963).
Other studies have used iterative bidding techniques such as  those conducted by Randall et al.
(1974) and  Brookshire  et al. (1976).   Brookshire  (1979)  explains  the iterative bidding tech-
nique as  "a direct  determination of economic  values from data which  represent responses of
individuals to contingencies posited  to them via a survey instrument."  The iterative bidding
technique in its current form was first developed  and applied by Randall et al. (1974) in the
Four  Corners   region  of  the Southwest.   Three contingencies  were  considered:   (a) limited
visibility  reductions and a  view of  a power  plant  with   limited  visibile   emissions;  (b)
moderate  emissions  from the plant,  moderate visibility  reductions  and moderate existence of
unreclaimed soil banks  and transmission lines; and (c) extensive emissions, visibility reduc-
tions  and unreclaimed  soil banks  and  transmission  lines.   Unfortunately   this  selection of
scenarios prohibits disaggregation of results into component values for visibility, power plant
location, and  unreclaimed  soil banks and transmission lines.   A mean reduction in sales tax of
$85 per household was required to make bidders accept scenario C instead of  scenario A;

SOX9A/A                                      9-2                                       12-23-80

-------
and a  reduction of  $50 to accept  C in  place  of B (Environmental Protection Agency,  197S).
Biases are  associated with these  techniques; however,  no bias tests were conducted in  this
experiment.   Because bids are  requested  without actual  payment,  there may be  incentives  to
misrepresent one's bids or the questions may distort an accurate response.
     According to  Rowe  et al.  (1980), the current state of methodological  and  empirical know-
ledge about  obtaining the economic value of atmospheric visibility suggests that much work  is
left to  be  done.   While Randall, et  al.  (1974)  and Brookshire et al. (1976) used a series  of
pictures in  depicting the varying levels of the environmental good under question, no linkage
was defined between the  pictures  and  actual  physical  parameters of  visibility.   Clearly,
differing levels  of  visibility  and other  aesthetic  effects were  depicted, but the question
remains  as  to  the relationship of economic  value  to  the  level of visibility and to the level
of emissions from power plants.  For benefit-cost  type  of evaluations for policy purposes  in
satisfying  the U.S.  Environmental  Protection Agency  guidelines,  a decision  regarding a re-
duction  in   the  aesthetics (visibility)  of  a region wguld  necessitate  all   such linkages.
According to Environmental Protection Agency (1979), certain psychological benefits, actual  or
perceived, are associated with good visibility.  These benefits include those associated with
a person's  wish to preserve the option  for a clear view  of a scenic area and those from just
knowing pristine areas exist, regardless  of any intent of  visiting them.
9.3  PERCEIVED EFFECTS ON  VISIBILITY
     The  term  "visibility" is  used colloquially  to   refer  to  various characteristics  of the
optical  environment, such as the  clarity with which  distant details  can  be resolved and the
trueness of  their apparent coloration.
     Traditionally,  visibility  has been  defined in terms  of visual range the distance  from an
object that  corresponds to a minimum or threshold contrast between that object and  some appro-
priate background.   Threshold  contrast refers to the  smallest difference between  two  stimuli
that the human eye  can distinguish.  The  measurement of these quantities depends on the nature
of the observer,  his or  her physical health, and mental attitudes of  attention  or  distraction
such as  effects of  boredom and  fatigue.
     Although  visibility  defined in terms of  visual  range is a resonably precise  definition,
visibility  is  really more than being able to see a black  target, or any target, at a distance
for which the  contrast is reduced  to the threshold  value.   Visibility also  includes seeing
vistas at  shorter distances and being able to appreciate  the details  of line, texture, color,
and form.  The definition  of visibility and the selection  of methods for monitoring visibility
impairment  should  relate  to  these different  aspects of  perceiving  distant objects.  Visual
impairment  resulting from air pollution  can  manifest  itself in two distinct ways; as  layered
or as  uniform  haze  (Malm  et al. 1980a and b).  Layered haze can be thought of as any confined
layer  of pollutants  that  results  ip a visible spectral  discontinuity between that layer and
its background (sky or landscape) while  uniform  haze exhibits ftself as an overall reduction

SOX9A/A                                      9-3                                        1^23-80

-------
in air clarity.  The classic example of a layered haze is a tight vertically constrained coher-
ent plume (plume blight).  However, as an atmosphere moves from a stable to unstable condition
and a plume  mixes with the surrounding atmosphere, the plume impact on visual  air quality may
manifest itself in an overall reduction in air clarity rather than as a layer of haze.
     Quantifications  of the perceptible  effect of  these  two types  of impairment  is  really
quite different.  The eye is much more sensitive to a sharp demarcation in color or brightness
than it  is  to a gradual change  in  brightness  or color over a period  of time (Green 1965 and
Patel 1966).
     Since a change in uniform haze takes place over time periods of hours or days, an evalua-
tion of visual air quality change resulting from a uniform haze requires a person to "remember"
what the scene  looked like before a given change in air pollution took place.   A layered haze
is evaluated  at only  one point in time  in  that the layer of haze is directly compared to its
background.   Judgments of  visual  air quality as a function of air pollution, whether the pol-
lution manifests  itself  as  layered or uniform haze, might  be  made different by variations in
sun angle,  cloud  cover, and  landscape features.   These and other aspects  of  visibility are
addressed in an EPA report to Congress (U.S.  Environmental  Protection Agency 1979).
     The selection of a good parameter (or parameters)  to characterize visibility is dependent
on whether one's  interest  is  in human perception  of visual  air  quality or the cause of visi-
bility degradation,  that is, the  air pollution  itself.   The variable, apparent target con-
trast, relates well to how a person perceives visual air quality  and serves as the fundamental
measure of  visibility impairment  (Mulu  et al.,  1980a).   Color  change also meets  the first
criteria.   Fine particulate concentration and  scattering coefficient  are  good measures when
relating visual impact  to  pollutant sources.   Visual range  cannot be directly measured by an
instrument.   However, when site intercomparisons are required (such as for establishing regional
trends) it will be necessary to use visual range as a normalizing parameter.  Also, because of
its historical  popularity,  it remains a  useful  concept to  indicate  atmospheric "clarity" to
the lay person.
     As assessment of  visibility impairment produced by air pollution must consider the pro-
cess of human visual  perception as well as changes in the optical characteristics of the atmos-
phere.  The perception of brightness, contrast, and color is not determined simply by the pat-
tern and intensity of incoming radiation; rather, it is a dynamic searching for the best inter-
pretation of  the  visible  scene.   For example,  notice the  difference between  a candle in a
brightly lit  room compared with one in a dimly  lit room,  or how  sunlit  treetops may appear
dark against  the  horizon sky but bright  when  viewed against the  shadowed  forest floo(r (EPA,
1979).  The  detection of  contrast between an  object  and its surroundings  is  fundamental) to
visibility.   Without contrast, as for example in a thick fog, objects cannot be perceived.  As
the contrast  between  object and background is  reduced  (for example, by increased pollution).,
the object  becomes less  distinct.   When  the  contrast becomes  very low,  the  object, will no

SOX9A/A                                      9-4                                       12-23-80

-------
longer be  visible.   This  liminal  or threshold  contrast has been the object of  considerate
study.  The threshold  contrast is of particular interest for atmospheric visibility,  since  it
influences the maximum distances  at which various components of a scene  can be  discerned.   Of
equivalent importance  to  threshold contrast is the smallest perceptible  change  in contrast  of
a viewed scene caused by a small increment in pollution haze (EPA, 1979).
     For small objects,  the  size of the  visual  image on the retina of  the eye also  plays  an
important role in  the  perception of contrast.  We all know from experience that,  as an ojbect
recedes from  us  and apparently becomes smaller, details with low contrast become difficult  to
perceive.   The reason for this loss of contrast perception is not only that the  relative bright-
ness of adjacent areas changes but also  that the visual system is less  sensitive to  contrast
when  the  spacing of contrasting areas decreases.   If the contrast spacing  is  a  regular pat-
tern of light and dark bands, e. g., a picket fence, a "spatial frequency" can  be readily de-
scribed in terms of the number of pattern repetitions or "cycles'1 per degree of  viewing angle.
The human visual  system is much more sensitive to contrast at certain spatial frequencies than
to contrast of other spatial frequencies  (EPA, 1979).
     Data reported by Van Nes and Bouman  (1967) can be used to estimate the reduction  of view-
ing  distance  which  is needed  to  compensate  for  loss of  contrast.   For  example, a regular
pattern of 40 cycles per degree can normally be resolved with about 5.2 percent contrast.  If
the contrast  is  reduced to 0.26 percent  the pattern would remain visible only if the viewing
distance is reduced to 1/5 its original value.
     The  relationship  between perceived  contrast threshold  and target  characteristics (size
and pattern)  is  important for visibility, because a  scenic vista usually contains a number  of
targets of varying sizes and arrangement.   The  calculation of the perceptibility of all tar-
gets  would  require specification  of their angular  size distribution.   The  perception  of
"texture," consisting  of contours of small  angular size and high spatial frequency, is parti-
cularly affected by this loss of threshold sensitivity.
     At the   practical  level, measurements  of pattern  preception thresholds  by  several  re-
searchers  (Van Nes  and  Bouman, 1956; Schober and Hilz,  1965;  Cornsweet,  1970;  Henry, 1977)
make  it particularly clear that the scattering and absorption of light by particles and gases
added to  the  atmosphere can lead to a dramatic loss  of visibility through contrast reduction.
The operator  of  a motor vehicle or the pilot of an airplane who must react quickly to minimal
visual cues  may be  greatly disadvantaged  by an increase  in  atmospheric  pollutants.   In the
above illustration  of  visibility loss by contrast reduction a target that  is normally visible
at 100 m  would only become visible  at  20 m.  For the  operator of a high-speed vehicle, this
loss of visibility may be fatal given the limits of human reaction time.
     An insight  into  general  visibility conditions  in the United States  can  be obtained  by
examining  available regional  airport visibility data.   Figures 9"!  and 9-2  (Trijonis and
Shapland,  1979)  show  isopleths  of median  yearly and summer visibility,  a statistic that  is
insensitive  to  the  site-specific  availability  of  markers as  long  as  the  farthest Barkers
SOX9A/A                                      9-5                                       12-23-80

-------
    r BASED ON PHOTOGRAPHIC
      PHOTOMETRY DATA

    N BASED ON NEPHELOMETRY DATA
    • BASED ON UNCERTAIN EXTRAPOLATION
      OF VISIBILITY FREQUENCY DISTRIBUTION
                                                                     15
         Figure 9-1. Map shows median yearly visual range (miles) and isopleths
         for suburban/nonurban areas, 1974-76.

         Source:  Trijonis and Shapland (1978).
  photometry data

N: based o<~ nepnelometfy data

• : basf-G on uncertain extrapolation
  vis ib 111ty frequercy distfibu
                                                                                  ^  /
                                                                                  10^
     Figure 9-2. Median summer visual range (miles) and isopleths for suburan/
     nonurban areas, 1974-76.
     Source: Trijonis and Shapland (1978).

                                                9-6


-------
consistently  reported lie  beyond the  median  visibility.   The  data represent midday  visual
ranges for  1974-76  from  100 suburban/nonurban locations.  Visibilities at 93 of the locations
were determined  from  airport observations, usually by the  National  Weather Service.   Instru-
mental visibility measurements  from seven sites in the Southwest are also included.   Although
some uncertainties  arise  from the use of airport visibility observations, there is reasonably
good consistency among airport  observations within regions and between airport and instrumen-
tal results in the Southwest.
     The best visibility (70+ miles, 110+ km) occurs in the mountainous Southwest.   Visibility
is  also  quite good  (45-70 miles, 70-110  km)  north  and  south of that  region,  but sharp gra-
dients occur  to  the east and west.  Most of the area east of the Mississippi and south of the
Great Lakes has a median visibility of less than 15 miles (24 km).  Within the East, the lowest
visibilities occur along the heavily industrialized Ohio River Valley.
     Although natural sources of visibility impairment are undoubtedly an important factor in
producing  these  geographical and seasonal patterns, analysis of visibility trends and other
information discussed in  later sections suggests that manmade air pollution has a significant
impact.   It  is  also important  to note that  the regions  with  the  best existing visibility
levels  are the  most  sensitive  to additional  impairment  and most  responsive  to  incremental
pollution reductions.  The  reasons for this are discussed in  the next section.
9.4.  FUNDAMENTALS OF ATMOSPHERIC VISIBILITY
     The  effect of the  intervening atmosphere  on the  visual  perception  of  distant objects
(e.g.,  structures,  mountains,  and vistas) is determined by  the concentration and character-
istics of air molecules, particles, and nitrogen dioxide along the line of site along with the
illumination  conditions.   The  rigorous  treatment of visibility requires a mathematical des-
cription  of  the  interaction  of  light with  the atmosphere  known  as  the  radiative transfer
equation.   The  description presented here  is  intended  to  provide a  qualitative understanding
of  the  underlying  theory.   Detailed  treatments  are available  in  a  number  of  publications
(Middleton, 1952; Chandrasekhar,  1960).
     Figure 9-3(A)  shows  the simple case  of  a beam of  light (e.g.,  from the sun  or a search-
light)  transmitted horizontally  through  the  atmosphere.   The  intensity of the  beam  in the
direction  of  the observer,  I(x),  decreases with distance from the source as light is absorbed
or  scattered  out of  the beam.   Over  a short interval, this decrease  is proportional  to the
length of the interval and  to the intensity of the beam  at  that  point:
•-    >•-•                     -dl = bextldx
where dl is the decrease in  intensity, bgxt is the extinction coefficient (the proportionality
constant),  I  is the original  intensity of the beam, and  dx  is the path  length.  The extinction
coefficient b  ,.  has  units of  inverse  distance.   The extinction coefficient is determined by
              cX w
the  scattering  and  absorption  of  particles  and  gases   and  varies  with particle and gas
concentration and wavelength of  light.

SOX9A/A                                      9-7                                       12-23-80

-------
     Consider  an observer  looking  at a  distant  target in the daytime as depicted  in  Figure
9-3.  Just as a  light beam  is attenuated by the atmosphere, the light from the target reaching
the observer  is  diminished by absorption and  scattering.   In  addition,  the observer receives
extraneous  light (often called  "air light")  scattered into  the  line of sight by  the  inter-
vening atmosphere.   The net effect is that, as  shown  in Figure 9-4, a target darker than the
horizon  appears  brighter than it actually  is  and a target brighter than the  horizon appears
darker than  it  actually  is.   At  increasing  distances, the apparent  brightness  of  dark and
bright targets  approaches  the horizon  brightness.   At sufficient  distance,  the target and
horizon  are so  close in brightness that the target is indistinguishable;  this distance is the
"visual  range," or "visibility."
     The example just  discussed can be restated in terms of a target's "contrast,"  defined as
target  brightness  minus  horizon  brightness,  divided  by horizon  brightness.   A  target's
apparent contrast  begins with  some  inherent  value  at  the target and approaches zero  as one
backs away from  the target.  When the observer's  "contrast threshold" is reached,  the target
cannot be distinguished and the visual range is known.
     In  a  uniform  atmosphere,  the  apparent contrast  between  a  target  and the horizon sky
decays exponentially with observer-target distance x (Middleton, 1952):
                              C = CQ exp(-bext x)                               (9-1)
where CQ is the initial contrast at x = 0.  The maximum distance 'at which a given  large target
can be  distinguished from  the  horizon  is therefore inversely proportional  to  the  extinction
coefficient:

                                 x    - 1oge |Co| " 1p9e  £                    (9-2)
                                  max ~       K
                                              bext
where e is the observer's contrast threshold.
     The proportionality factor, logg |cJ - logg  e  in equation 9-2, depends on  the target's
intrinsic contrast  with the  horizon  and on  the observer's contrast  threshold.  For a black
target,  CQ =  -1, so that log   C-  = 0 and the proportionality  factor  reduces to -log   e .
The visual range (V) in a uniform atmosphere is thus given by the Koschmieder formula:
                                         V = "Io9e  £                          (9~3^
     The observer's contrast threshold, e, is of course no universal constant.  It varies with
apparent  target size  and  overall  illumination,  and it  varies with  observer  (Figure 9-5;
Middleton,  1952).    However,  it  enters  equation  9-3  only  logarithmically,   so  that  the
relationship  of visual range to extinction  is  not unduly sensitive to  the  psychological  and
physiological nature of  the observer.   A sevenfold increase in e, for example, does not quite


SOX9A/A                                      9-8                                       12-23-80

-------
Figure 9-3. (A) A schematic representation of atmospheric extinction,
illustrates (i) transmitted, (ii) scattered, and (iii) absorbed light. (B) A
schematic representation of daytime visibility illustrates: (i) residual
light from target reaching observer, (ii)  light from target scattered
out of observer's line of sight, (iii) air light from intervening atmos-
phere, and (iv) air light constituting horizon sky. (For simplicity,
"diffuse" illumination from sky and surface is not shown.) The
extinction of transmitted light attenuates the "signal" from the tar-
get at the same time as the scattering of air light is increasing the
background "noise."
                         9-9

-------
in
V)
2
E
GO
        LIGHT INTENSITY OF HORIZON
                        BLACK OBJECT
                     OBJECT-OBSERVER DISTANCE



Figure 9-4. The apparent contrast between object and horizon sky
decreases with increasing distance from the target. This is true for
both bright and dark objects.


Source: Charlson et al. (1978).
                        9-10

-------
  o
 o
 UJ
 cc
220

200

180

160


140


120

100


 80


 60

 40


 20

  0
                               N • 1000
         00
                05
10
                                              15
20
                             CONTRAST, |e|

                     J	I	I
           4.6        3.0           2.3          1.9

                           LOG CONTRAST. K
                                                      1.6
Figure 9-5. Measured apparent contrast of farthest visibility marker
was identified in 1000 determinations of visual range by 10 observ-
     Scatter is due to both the variability of observer thresholds and
ers.
the discrete nature of the marker set. The corresponding value for
the Koschmieder constant, K = -log e, is 3.2 * 0.6.

Source: From Middleton (1952).
                             9-11

-------
double  -log   e  .   The conventional  choice of  c   for a  "standard"  observer is  0.02,  which
yields V = 3.9/bext.
     Little  error  is introduced  into the determination  of visual range by  using as  targets
such nonblack features as dark forests and deep shadows.  For a target whose intrinsic  bright-
ness  is  as  much  as 30 percent  that of  the  horizon sky, for example,  the  limiting distance
given by equation 9-2 is within 10 percent of the visual range.  On the other hand, the intrin-
sic brightness of artificial lights or sunlit objects can approach that of the horizon  sky,  in
which case  their use as  targets can lead to  large underestimates of visual  range.   It must
also be  noted that  nonblack targets can lead to serious underestimates of visual  range if one
uses a  target at a  distance significantly less that the visual  range and then calculates  or
estimates visual range  from the  apparent contrast of the target.   For example, for an  intrin-
sic target brightness of  30 percent of the horizon brightness, an error of 18 percent  results
from using a target at  half the true visual  range and an error  of  91 percent results if the
target is at one-tenth the true visual range.
     The Koschmieder  formula's neglect of pollution gradients and the  earth's curvature and
topography  limit its applicability  near  sources  of primary particulate  matter  and  in very
clean air.   Where  visibility  is  restricted  by diffuse haze, however,  equation  9-3 performs
well.    Comparisons  of daytime visual range, as measured by a human  observer,  and extinction
from scattering, as  measured instrumentally  at a  single point, -show visual range  to correlate
with  the reciprocal  of extinction,  as illustrated  in Figure 9-6  (Horvath and  Noll,  1969;
Samuels et al.,  1973).  The correlation coefficients are commonly in the neighborhood  of 0.9,
which is quite good considering that the point measurement of extinction is being  extrapolated
along a sight path several tens of kilometers long.
     In  summary,  visibility (or visual  range) is  inversely proportional to the atmospheric
extinction coefficient.
9.4.1  Measurement Methods
     The term "visibility"  is  often used to mean  visual  air quality which takes  into  account
not only how far one can see, but also how well one can see nearby objects (i.e.,  their appar-
ent  contrast or  discoloration).   These  multiple  meanings  have  confounded  the  choice  of a
particular method for measuring and ultimately regulating visual  air quality.
     The extinction  coefficient  introduced  in Section 9.4 represents a summation of  contri-
butions  from scattering and absorption by gases and by particles:  b  .  = bR  + b   +  b   t +
babs' wnere bpa ^s scattering by air molecules (Rayleigh scatter); b   is absorption by nitro-
gen dioxide  gas; b   .  is scattering  by  particles; and b .   is  absorption  by particles.   A
number of methods  are in use which  measure  or allow estimation of the  various terms  of this
equation.  A full  discussion of  the more common methods follows.   For more detailed accounts,
see Middleton (1952), Charlson et  al.  (1978), Tombach and Allard (1980),  Ellestad and Speer
(1980), and Waggoner et al.  (1980).

SOX9A/A                                      9-12                                      12-23-80

-------
     o.so

     0.40

     0.30

  E
  *•  0.20
 u
 u
 UJ
 8  0.10
 a
 cc
 E
     0.05
     0.03
                  I  I I I
                                   I
I
                                              I
                                                     I  I  I  i
                       10         20             50

                           VISUAL RANGE, km
                  100
Figure 9-6. Inverse proportionality between visual range and the scat-
tering coefficient, bscat, was measured at the point of observation.
The straight line shows the Koschmieder formula for nonabsorbing
(bext= ''scat) med'a- v = 3-9fascat- Tne linear correlation coefficient
for V and 1/bscat is 0.89.

Source: Horvath and Noll  (1969).
                             9-13

-------
9.4.1.1  Human Observer (Total Extinction)—The human observer can make qualitative statements
about overall  visual quality, unusual  coloration,  and the presence of plumes.  However,  the
observer has been  most  often used to determine the visual range.   In standard practice,  a set
of dark targets  at known distance is selected  and  an observer then records whether or not he
can see each target.   Different  days or studies  are  difficult to compare because  even minute
changes of scenes  from day  to  day may  affect  human perception.  This shortcoming can  be
resolved by  photography.   The observer method has  been  widely  used  for many years  and its
qualitative  judgments  of visual  air quality are the ultimate reference for any  regulatory
effort.   However,  the method  is  labor intensive,  subjective,  and  often must use  ill-placed or
non-ideal   targets.   The intercomparison of qualitative  judgments from the  visual  range data
may  be  converted  to  b  .,   if  one  accepts  certain  assumptions  (Middleton,   1952),  by
Koschmieder's formula:
          K    -      3.9
          b	
                  visual range
9.4.1.2  Photography (Total Extinction)--Photographs can  be used  to document scenes for later
qualitative analysis by  humans or for later analysis  of  a target's apparent contrast by film
densitometry.  Photographs provide for  more accurate long-term retention of a scene than does
the human  mind and  enable  large numbers  of people  to evaluate a given  scene for perception
studies.    However, photography may introduce significant  errors in reproducing a scene due to
varying film characteristics  and  exposure, aging, storage, and reproducibility  of the image.
The faithfulness  of  rendition  may,  therefore, not  be  true  and  densitometry and  even some
qualitative applications may be  seriously affected.
9.4.1.3  Telephotometry (Total Extinction)---A telephotometer  is a telescope which can measure
the apparent brightness  of  a far away object.  By  measuring  the brightness  of  an object and
the horizon sky around it,  one can compute the object's apparent contrast.  This  number may be
used directly or, if one is willing to assume Koschmeder's restrictions (Middleton, 1952), may
be converted to b  .  by the equation:

           bext =
Telephotometry is  attractive  because it is  a path  measurement (thus atmospheric nonuniformi-
ties are averaged),  the instrument's absolute  calibration  is  unimportant (only  its linearity
and  short-term stability matter),  it  requires  no  sample  aspiration  (and  thus  avoids  large
particle  losses  and  sample  heating or  cooling),  and  is perhaps  the  closest instrumental
approximation  to  human observation.  The  method has  limitations  when  the target's intrinsic
contrast is unknown or assumed (a very common situation), when measuring dark objects'at'close
range  (due to internal  stray light  errors), and when clouds  cause  uneven illumination.  The
cloud  problem  makes telephotometry  difficult  to automate and thus the  method  remains  labor
intensive.
SOX9A/A                                      9-14                                     12-23-80

-------
9.4.1.4  Long-path Extinction (Total Extinction)--The  most  direct way  to  measure b    is to
measure the decrease in intensity of a light beam over a known distance  x,
          bext = 4  In  T-
                  *      Xo
where I and  I   are the final and  initial  intensities, respectively.   The  method is  appealing
in that  no assumptions  are involved,  it  measures average  extinction  over the path,  and  it
requires no sample aspiration.  Unfortunately, even for values of b  .  about 0.2 km   (typical
at the eastern  U.S.)  decrease over short  paths  (about 1 m) is extremely small (200  parts per
million) and  cannot be measured accurately.   An alternative is to increase the  path length,
but  source intensity  fluctuation  (or mirror  reflectivity  changes  for  single-ended  systems),
alignment, thermally-induced scintillation, and the large background light of daytime make the
measurement difficult  again.
     Hall  and Riley  (1976) have measured  extinction  by observing an uncontaminated source at
two  ranges.  Any decrease in intensity with range in  excess of the inverse-square decrease is
due to extinction.  This method avoids any need for absolut'e calibration, since only the ratio
of intensities at two  ranges need be measured.  This method as used by Hall and Riley is labor
intensive  but  has been  demonstrated  in clean and urban  environments  during night operation.
9.4.1.5  Nephelometer  (Scattering)--The  integrating  nephelometer measures only the scattering
coefficient, bscat,  of an aerosol.  By  simple adjustment,  Rayleigh scatter, bR  , can be sub-
tracted or included.   The  instrument consists  of an enclosed volume painted black, a sensitive
light detector  looking through the volume, and a  light source at one side of the volume.  The
only light reaching the detector  is  that  scattered  by gas molecules and particles within the
volume.   The nephelometer  is  sensitive,  is easily calibrated,   is easily automated, enables
one to modify the sample  if desired,  and provides  a point measurement for correlation analysis
with  point measurements  of mass loading  and  chemical composition.   Errors of application of
the  nephelometer include sites where the  atmosphere  is nonuniform  (e.g., near  aerosol  sources
or  in mountainous  terrain) or  sites which may  have  significant  absorption  occurring  (e.g.
urban  sites),  unless  a  measurement  of  absorption is being  made as well.   Nephelometers can
also give  suspect data if  the operator  is not careful to avoid  inadvertent  heating or  cooling
of the  sample and resulting modification  of  the aerosol size distribution.   The nephelometer
has  two sources  of  inherent errors,  that of  angular truncation (which results in underesti-
mates  of scattering,   especially when large particles are present) and  that of sample  aspira-
tion (which  results in large particles  being  impacted on the  ductwork and thus  never  measured).
These  inherent  errors may  result  in  depressed scattering/total  mass correlations and elevated
scat/fine  mass  correlations when  significant large  particle  concentrations  occur.   Despite
these  limitations, the nephelometer remains one  of the most widely  used visibility measurement
methods.   Most  of the  data relating  cause  and  effect (i.e.,  particulate concentration or
composition  and optical  effect) have been acquired  with the nepbelometer.  Agreement  between

SOX9A/A                                       9-15

-------
the nephelometer and  the  long-path transmission has been demonstrated  in  several  cases  (Wag-
goner and Charlson, 1976;  Weiss et al., 1978).
9.4,1.6  Light Absorption Coefficient—Without doubt, the absorption  component  of  extinction,
b .  , is the  most  difficult to measure.  As yet,  no  single method has proven  to  be  the most
effective.   Were it not for the fact  that graphitic  carbon as soot is  a prominent species  in
cities and  industrial regions, b  .   would be  inconsequentially  small.  However,  even  a few
percent of the submicrometer  mass  as soot produces a  significant  effect on b  .   or bgx^.  The
methods that  have  so  far  been used to  evaluate  b  .   include:   (1) determining  the difference
between b  .  and b   .by  combining long-path transmissometer with nephelometer  (Weiss et al.,
1978); (2)  determining absorption  on Nuclepore filters with scattered light  removed  by  an
integrating plate  of  opal  glass  (Lin et  al,  1973);  (3) determining absorption on Millipore
filters (Rosen et  al., 1980);  (4)  determining the  reflectivity of a white  powder with aerosol
mixed  into  it, called the  Kubelka-Monk method  (Lindberg  and  Laude,  1974);  (5)  determining
absorption of light  by  a  sample   of particles inside  a  white  sphere (integrating  sphere)
(Fischer, 1975);  (6)   estimating  an  imaginary  refractive  index   from  regular  scattering  or
polarization and size  distribution  (Eiden,  1971); (7)  measuring the amount  of  graphitic carbon
and its size  distribution and  them calculating b . ;  (8) detecting the acoustical pulse pro-
duced when energy  is   absorbed  by  particles as  light and  transformed  to  heat  (spectrophone)
(Truex and Anderson, 1979).
9.4.2  Role of Particulate Matter  in Visibility Impairment
     As noted in  section 9.2.2, the  extinction coefficient comprises  contributions  from gas
and particle scattering and  absorption:
                        bext = bRg * bag * bscat * babs
This section discusses the relative magnitudes of these contributions.
9.4.2.1  Rayleigh Scattering—A particle-free atmosphere  at  sea level  has an extinction coeffi-
cient of  about 0.012  km    for green light  (wavelength 0.55  urn)  (Penndorf, 1957),  limiting
visual range  to about 325 km.   The coefficient b_   decreases with altitude.   In some  areas  of
                                                 Kg
the western United States,  the extinction of the atmosphere is at times essentially that of a
particle-free  atmosphere  (Charlson  et al.,  1978).  Rayleigh scattering  thus amounts to a
simply definable  and  measurable background  level   of extinction  with  which  other extinction
components (such as those caused  by manmade pollutants)  can be compared.   Rayleigh scattering
decreases with the fourth  power of wavelength, and contributes a strongly wavelength-dependent
component to extinction.   When Rayleigh scattering dominates, dark objects  viewed at distances
over  several  kilometers appear behind a blue  haze of scattered light, and bright objects  on
the horizon (such as snow, clouds,  or the sun) appear reddened at  distances greater than about
30 km.
9.4.2.2   Nitrogen Dioxide Absorption—Of  all  common gaseous  air  pollutants,  only  nitrogen
dioxide  has   a  significant absorption band in  the visible part  of the  spectrum.   Nitrogen

SOX9A/A                                      9-16                                      12-23-80

-------
dioxide strongly absorbs  blue light and can  color  plumes  or urban atmospheres red,  bro«r.,  ^
yellow if  significant  concentrations and path lengths are  involved.   The  effects of nitroc--
dioxide on  visibility  are discussed more fully  in  the criteria document for oxides  of nitro-
gen, and  in Hodkinson  (1966),  White  and  Patterson (1980), and Charlson et  al.  (1972j.   Its
contribution to total extinction is  in general minor.
9.4.2.3   Particle  Scattering--As the  particle concentration increases  from  very low levels,
where  Rayleigh scattering  dominates,  the  particle scattering coefficient,  b   .,  increases
                                                                               SCoTr
until  eventually  bscat  is  greater  than  bR .   At  40-km  visual range,  a  better-than-average
value for the eastern U.S., Rayleigh scattering contributes only about one-eighth of  the total
extinction (Trijonis and Snap!and, 1979).
     Two principal problems in understanding the degradation of the visual  quality of air have
been:  (1) defining the size  range and other physical characteristics of particles most effec-
tive in causing scatter;  and (2) defining  the chemical  composition of particles in  this size
range  (Charlson  et  al., 1978).   Size, refractive index,, and  shape  are  the  most  important
parameters  in relating  particle  concentration  to  particle-related  extinction coefficients,
b   t and  b .  .   If these properties are established and the shapes are geometrically simple,
the  light  scattering and absorption can be calculated.  Alternatively, the extinction coeffi-
cient  associated  with  an  aerosol  can  be  measured directly  as  discussed in  Section 9.4.1.
     From  the  point of view  of  aerosol  optics,  a  key question is whether an aerosol particle
is  spherical.   For  such  particles, rigorous  Mie  theory  (Mie,  1908) is  applicable,  and the
optical properties can be readily calculated  from their size and refractive index (Dave, 1969;
Wiscombe,  1980).   Measurements  in St.  Louis by Allen et al. (1979) show that in the fine mode
less than 5 percent of  the aerosol population  is  nonspherical.   Pueschel  and Wollman (1978)
found that spherical  particles also dominate  fine-mode aerosols near Cedar Mountain, in east-
central Utah.
     Charlson  et  al.  (1978)  used Mie  theory  to  calculate  the  light-scattering and absorption
efficiency  per unit volume of particles for a typical aerosol  containing some light-absorbing
soot (Figure 9-7).   As illustrated  in the  figure,  particles of 0.1  to  2  \an diameter are the
most efficient light scatterers.  The remarkably high scattering efficiencies of these parti-
cles are  illustrated by the  following examples:   a given  mass of  aerosol  of 0.5-um diameter
scatters  about a  billion  times more  light than the same  mass of  air;  a  1-mm-thick sheet of
transparent  material,  if  dispersed as  0..5-ym particles,   would  be sufficient to scatter 99
percent of the incident light,  that is, to completely obscure vision across such an aerosol
cloud.
     A  more  revealing  explanation  of the  usual  dominance of scattering by  fine  particles
(i.e.,  those particles (yf  diameter less  than 1-3  um)  is possible:  particles  smaller than
0.1 um, though sometimes- present in  high numbers, are very  inefficient at scattering light and
thus contribute  very  little  to visibility  loss;  particles larger than about 1-3 um, though

SOX9A/A                                      9-17                                      12-Z3-80

-------
      10
 ==  e
"u T 5
»E  ?
 §*§>
  si
  M mO
                        j  i
SCATTERING
        10
         ,-2
 10'1
10U
                                                               4.00
                               DIAMETER,
   Figure 9-7.  For a light-scattering and absorbing particle, the scatter-
   ing per volume has a strong peak at particle diameter of 0.5 /im (m =
   1.5 — O.OSi; wavelength = 0.55 fim). However, the absorption per
   aerosol volume is only weakly dependent on particle size. Thus the
   light extinction by particles with diameter less than 0.1, pm is primar-
   ily due to absorption. Scattering for such particles is very low. A
   black plume of soot from an oil burner is a practical example.

   Source: Charlson et al. (1978).
                                  9-18

-------
efficient  per  particle  at  scattering  light,  usually  exist  in  quite  small  numbers  and
contribute only a  small  fraction of visibility loss.   Coarse particles (i.e.,  larger  than  1-2
urn) are occasionally  important  in determining visibility, particularly near roadways  and some
industrial sites and during natural occurrences of fog and wind blown  dust.
     Coarse particles are  almost exclusively nonspherical; using the  Mie theory  to calculate
their optical properties will  therefore give only a  crude  approximation.   There  is,  however,
an  extensive  body  of data  on  the  optical  properties of  the nonspherical particles  (e.g.,
Pinnick et al.,  1976; Fowler and Sung, 1979; Mugnai and Wiscombe, 1980).
     Atmospheric particles are  made  up  of  a number  of  chemical compounds.   All  of  these
compounds exhibit  a peak  scattering efficiency  in  the  same  diameter range (0.1  to 1.0  um)
calculated to be optically important for the typical aerosol in Figure 9-7.  However, because
of differences  in  refractive  index, the values of the peak efficiency and the exact  particle
size at  which  it  occurs vary  considerably  among the  compounds (Figure 9-8;  Faxvog,  1975).
     It  is  instructive to  note in  Figure  9-8 the high  extinction efficiences of  carbon  and
water.   As will  be discussed in  Section 9.4.3, these compounds are often significant fine mass
components and are therefore often responsible for significant amounts of extinction.
     Measured  particle  size  distributions  can  be  used  in  conjunction with  Mie  theory
calculations  to determine  the  contribution of  different  size  classes to  extinction.   The
results of this kind of calculation are shown in  Figure  9-9.  The peak in scattering  per unit
volume is at 0.3 urn, so that the fine particles dominate  extinction in most cases.
     Because the peak and  shape of the bimodal particle  mass distribution curve can vary,  the
light-scattering characteristics of  a given  particle  mass might  also  be expected  to  vary.
However,  as  noted  by Charlson  et  al.  (1978), for the observed  range of atmospheric  particle
distributions,  the calculated  scattering  coefficient per  unit mass is  relatively  uniform.
Latimer  et  al.  (1978)  have  determined the scattering  per  unit volume for  several aerosol
distributions.  The calculated  coefficient  changes by  no  more  than 40 percent in  the size
range of 0.2 to 1.0 urn (Figure 9-10; Latimer  et al., 1978).
     The  relative  consistency  of  calculated  light  scattering per unit mass  over  a  range of
particle  distributions  and the  dominant influence of fine particles suggest that reasonably
good  approximations  of light-scattering  coefficients   can  be  obtained  by  measurements of
fine-particle mass.  Indeed, agreement from  simultaneous  monitoring of the two parameters at a
number of sites has been found  by  several  investigators.  Measurements by Waggoner  and Weiss
(1980), Weiss  (1978), Patterson and Wagman  (1977),  Macias et al.  (1975),  and Samuels et  al.
(1973) at various sites showed scattering per unit fine mass ratios of 3-5 m2/g.  Correlations
between  the  fine-particle  mass  and bscat  are consistently  high  (Table  9-1).  Figure 9-11
(Macias and Husar,  1976) shows the relationship for St. Louis.  The high correlations indicate
that at the sites studied, fine-particle mass dominates particle  scattering.


SOX9A/A                                       9-19                                      12-23-80

-------
      6
  I.
      2 -
                                         SILICA 1
 CARBON
                                            1.0
                                                              10.0
       0.01
0.1                 1.0
  DIAMETER, fim
                                                              10.0
     14
  I"
 "I
  =  6
  "    «
  jf
                                             X-0.55 Jjm
       0.01
                        0.1                1.0

                           DIAMETER, fim
                                                            10.0
Figure 9-8. (A) Calculated scattering cross-section per unit mass at a
wavelength of 55 pm for absorbing and nonabsorbing materials is
shown as a function of diameter for single-sized particles. The follow-
ing refractive indices and densities (g/cm') were used: carbon (m =
1.96-0.66J. d = 2.0). iron (m = 3.5V3.95J, d = 7.86). silica (m =  1.55,
d = 2.66), and water (m = 1.33, d = 1.0). (B) Calculated absorption
cross-section per unit mass at 0.55 pm for single-sized particles of
carbon and iron. (C) Calculated extinction cross-section per unit mass
at 0.55 pm for single-sized particles of carbon, iron, silica, and water.

(Source: (a) Faxvog (1975); (b and c) Faxvog and  Roessler (1978).
                        9-20

-------
tu  a.

5   .
D  J
-J  <

9  >
>  cc
uj  UJ
   o
   <
   E
VOLUME
                   Ig-^.,^,^,,.,  _  j
          0.01
     0.1           1.0



    PARTICLE DIAMETER,
                                                   10
0  >
C  i-
  en
5  °

SH

S  S



f|


 -O
   o
   Figure 9-9. For a typical aerosol volume (mass) distribution, the cal-

   culated light-scattering coefficient is contributed almost entirely by

   the size range 0.1-1.0 nm. The total bscat and total aerosol volume

   are proportional to the area under the respective curves.



   Source: Charlson et al. (1978).
                                   9-21

-------
     10
 D)
    i.o
"a
jf
     0.1
       0.1
                            .1.1
        i    i  i  . I i nl
1.0
                       MASS MEDIAN DIAMETER
                       10.0
 Figure 9-10. Scattering-to-volume ratios are given for various size
  distributiohs.
                                   9-22

-------
     60
     40
<  =»•
2    20
UI
     0.2
 "E
 J0.1
           4/17
4/18
   1
4/19      4/20

  TIME, days
4/21
4/22
Figure 9-11. Simultaneous monitoring of bsca* and fine-particle mass
in St. Louis in April 1973 showed a high correlation coefficient of
0.96, indicating that bscat depends primarily on the fine-particle
concentration.

Source: Macias and Husar (1976).
                           9-23

-------
                     TABLE 9-1.  LIGHT SCATTERING'PER UNIT MASS OF FINE AEROSOL

Location b ./mass,

Mesa Verde, CO
Seattle, WA (residential)
Seattle, WA (industrial)
Puget Island, WA
Portland, OR
New York, NY
St. Louis, MO
Los Angeles, CA
Oakland, CA
Sacramento, CA
(mVg)
2.9
3.1
3.2
3.0
3.2
5.0
5.0
3.7
3.2
4.4
r

--
0.95
0.97
0.97
0.95
—
0.96
0.83
0.79
0.98
N Reference9

5
58
64
26
108
—
72
39
20
6

1
1
1
1
1
2
3
4
4
4
         al, Waggoner and Weiss (1980);  2,  derived by Charlson et al.  (1978)  from
          Patterson and Wagman (1977); 3,  Macias et al.  (1975); 4, Samuels et al.  (1973).

     This was documented  in  an experiment conducted by Patterson and  Wagman  (1977),  who moni-
tored the ambient aerosol size distribution by a set of four cascade impactors in the New York
metropolitan area.   The first impactor was  operated only when the light-scattering coefficient
was between 0 and  0.15 km  ; impactor A was operated at 0.15-0.3; impactor B at 0.3-0.45; and
impactor C when  values exceeded 0.45.   The measured mass distributions (Figure 9-12;  Patterson
and Wagman, 1977)  show that  at good background visibility  levels,  the mass  concentration was
largely (70 percent) contributed by coarse particles.   At the low visibility  level C, however,
over 60 percent  of the total mass was contributed by fine particles.   Thus,  visibility in the
New York  metropolitan area  was  found to  be  lowest when the concentration  of  fine  particles
reached a maximum.
     It is conceivable that in the arid West the aerosol refractive index and relative amounts
of coarse  and fine  particles  are so different that the scattering mass  ratios  quoted above
would  not  be applicable.  However,  preliminary results  from project VISTTA  (Macias  et al.,
                                                                    O
1978)  suggest that bscat/fine  mass ratios in the  Southwest are 3 m/g as measured elsewhere.
     In  areas  where  fine-particle concentrations  are  low,  coarse particles  may  contribute
significantly to light extinction.   Coarse dust particles  are  much less efficient scatterers
per unit mass (Figure 9-7; Charlson et al.,  1978),  however, so that much higher mass concen-
trations are  required to  produce a  given  optical effect.   In windblown dust,  for example,
Patterson  and Gillette (1977) reported values  of  the ratio of light  scattering  to  mass that
were more than an order of magnitude lower than those noted above for fine particles.!   /
     With  regard to how much of extinction is due to scattering, the measurements of Waggoner
et al. (1980) made with an integrating nephelometer and the integrating plate method show that

SOX9A/A                                      9-24                                      12-23-80

-------

2.0

1.5


1.0


0.5


I I i 1 1 1 1 1 1 I i i i i 1 i i i lii
BACKGROUND VISIBILITY A
"" , ri —
"tot"44"3 M«/n»
_


—


"^ «^M f
^r\r>^^_ ^-'
1 .
T"
f
f
1
f
1


^•"i i i 1 1 1 n i i i i 1 1 1 1

_

t
i
i
\
\

\ ~
I 1 1 1
0.1   0.2
               0.5
10   20
                                                    50   100


2.0

1.5

1.0
0.5


1 1 1 II 1 11 1 | 1 1 I 1 I 1 1 III
VISIBILITY LEVEL A
' 1 —
1 M. -785 iia/nT


~
' I
t
/
	 nt-1
A
1 1


1OI "^

•"•
k xf"*^«»^ —

1 M 1 1 1 1 1 1 M 1 II 1 lTl--._
a
1
< 0
*<
^ 2.0
1.5
1.0
0.5
1
2.0

1.5
1.0
0.5
0
/
	 nt-

MINI 1 1 1 M 1 II 1 lTl--._
1 0.2 0.5 1 2 5 10 20 50 10(
1 I 1 I M III 1 1 1 1 M 1 1 1 III
VISIBILITY LEVEL B
A "~
t
1
t
/
/ _,
.x'' ,
U, :
^3— p— * — 1 '"*•»
1 1 1 1 1 1 1 i i i 1 1 1 i l i "••n— 1-»— —
.0 0.2 0.5 1 2 5 10 20 50 10
1 1 1 1 1 1 III 1 I 1 1 1 1 1 I 1 III
VISIBILITY LEVEL C _
M -212 jts/m3
tot ~

i
1
t
Si •
^ i
1
1 —
\
— ^^r~^ttl,llli,i. 	 tf^^^^^^^\ *~~^^
iiniii i iiiiiii i i ~I"l>*
.1 0.2 0.5 1 2 5 10 20 50 10
DIAMETER, pen
Figure 9-12. Aerosol mass distributions, normalized by the total
mass, for New York aerosol at different levels of light-scattering
coefficient show that at high background visibility, the fine-particle
mass mode is small compared with the coarse-particle mode. At the
low visibility level, 6,60 percent of the mass is due to fine particles.

Source: Patterson and Wagman (1977).

                          9-25

-------
in urban-industrial  areas  particle  scattering accounts for 50 to 65 percent of extinction,  in
urban residential  areas  70 to 85 percent, and in remote areas 90 to 95 percent.   A comparison
reported by Weiss (1978) which employed a nephelometer and a long path extinction device found
scatter/extinction percentages of 55 to 65 percent in urban Phoenix, Arizona and about 95 per-
cent  on  a plateau  outside Flagstaff, Arizona.   Recent  studies  by Wolff et al.  (1980)  which
used a nephelometer and absorption inferred from elemental carbon loadings produced data which
show scattering/extinction percentages of 60 to 85 percent at a variety of sites.   In general,
scattering by  particles accounts for 50 to  90 percent of extinction depending  on  location,
with urban sites in the 50 to 75 percent range and remote sites in the 80 to 95 percent range.
     Particle scattering also  plays  a role in the perception  of plumes.   Suspended particles
generally scatter  much more  in  the  forward  direction than  in  other directions.   This fact
means a plume  or haze layer can appear  bright  in forward scatter  (sun  in  front of observer)
and dark  in  back scatter (sun in back of observer)  because of the angular variation in scat-
tered air light.   This effect can vary with  background  sky and objects.   The  added air light
is both  angle  and  wavelength  (color) dependent  and the wavelength dependence  can  vary with
illumination angle.   A visible aerosol  layer will be  brighter  than an adjacent particle-free
layer for  sun  angles  (in  front of observer)  less than  30°.   At larger  angles,  the aerosols
will  usually be darker.  Aerosol  optical  effects alone are theortically capable of imparting a
reddish-brown color  to a haze layer when viewed  in  backward  scatter.   NO- would increase the
degree of  coloration  in such  a situation.   Specific circumstances of  brown  layers  must  be
examined on a case-by-case basis.
     In many pristine areas,  where  viewing distances are  50  to  100 km,  a reduction in calcu-
lated visual  range (for example,  from 350 km to 250 km) will not be the most noticeable impact
of incremental  pollution.    The  reduction  in apparent contrast  and discoloration  of nearby
objects and sky are the main effects perceived in such areas.
     Calculation  of  contrast  changes (for large  targets) accompanying  incremental  particle
levels indicate  that the maximum decrease in contrast will  occur for objects  located at dis-
tances of about  one-fourth of the visual range  from the observer  (Malm, 1979).   Thus,  in  an
initially clean  atmosphere,  a  fine  particle increment produces maximum contrast reduction for
large objects  50 to 100 km away.  A  reduction  in visual range of 5 percent would result in a
reduction in contrast of .02 for those  objects.   Such a change may be just perceptible.  The
contrast  detail  (texture,  small  objects) and  coloration  of closer objects in  contrast may,
however, be affected to a greater degree (Henry, 1979, Malm, 1979).
     The  perceived  color  of objects and sky is  also changed  by  the addition  of  aerosols.
Because of  the  difficulties  and  uncertainties  in  specifying perceived  color,  only a quali-
tative description is possible.   In general,  the  apparent color  of  any target fades toward
that  of the  horizon  sky  with  increasing distances  from the observer.   Without particles,
scattered air  light  is  blue,  and  dark  objects appear increasingly blue with distance.  The

SOX9A/A                                      9-26                                      12-23-80

-------
                                         2
addition of  small  amounts  (1  to 5  ug/m )  of fine particles throughout the  viewing  distance
tends to whiten the  horizon  sky making distant dark  objects  and the  intervening air  lig^t
(haze) appear  more  grey.   According to  Charlson  et  al. (1978), even though  the  visual  range
may be decreased only slightly from thp limit  imposed  by Rayleigh scattering the change fror
blue to  grey is an  easily perceived  discoloration.   The apparent color of  white  objects  is
less  sensitive  to  incremental  aerosol  loadings.    As  for   contrast,  incremental   aerosol
additions produce a much greater color shift in cleaner atmospheres (Malm,  1979).
     Aerosol  haze can also degrade the  view  of the  night sky.   Star brightness is diminished
by light scattering and absorption.  Perception of stars is also reduced by an increase in the
brightness of  the  night sky caused by scattering of available  light.   In or near urban areas,
night sky  brightness  is  significantly increased by particle scattering of artificial  light.
The combination of extinction of starlight and  increased sky brightness markedly decreases the
number of  stars visible  in the night sky  at fine particle concentrations of  10  to  30 ug/m
(Leonard et al., 1977).
     Thus,  the  overall impact  of aerosol   haze is to  reduce   visual  range and contrast, and
change color.   Visually the objects are "washed  out"  and the  aesthetic value of the vista is
degraded even though the distances are small relative to the visual range.   Much of the scenic
value of a vista  can be lost  when  the visual  range is reduced to a distance that is several
times greatest line-of-sight range in the scene.
9.4.2.4   Particle  Absorption—Particle  absorption  (b . )  appears  to be  on  the order  of 10
percent of particle  scattering (bscat)  in low-background areas such as Bryce Canyon (Weiss et
al., 1978).  Its contribution may rise  to  50 percent of b  .  in urban  areas, although values
of 10 to 25 percent would be more typical (Weiss et al., 1978;  Waggoner et  al., 1980; Wolff et
al., 1980).
     The apparently  large  role of  particulate absorption on  visibility  in urban  areas has
prompted much  research on  absorption  in recent years.  Weiss  (1978) fractionated aerosol at
many  sites  into coarse and fine portions  (cut at 2.5 urn) and found  that the fine particles
accounted  for   80  to  98  percent of absorption.   He  also determined the  wavelength (color)
dependence of  b .   for the samples  and  concluded that at all  U.S.  sites  bgbs is only weakly
dependent on wavelength.
     The .extinction  per unit mass raio  for  absorbing  aerosols  is a figure of  some importance
because  current estimates  show it to be significantly higher  than the extinction/mass  ratio
of scattering-only  aerosols.   Figure 9-8 shows the  theoretical value for  carbon particles to
be  higher  than  for  any  other  considered species  (Foxvog and  Roessler,  1978).   Laboratory
studies by Roessler  and Foxvog (1980) showed values of 9.8 m  /g for acetylene smoke and 10.8
n»2/g  for diesel exhaust;  they also summarized results  from  other  investigators  on various
aerosols which showed  values  of 6.1  to 9.5 m2/g.   In developing  a spectrophone, Truex and
Anderson (1979) measured  a value of 17  m2/g (at 0.417 urn) for aerosol  from a propane-oxygen

SOX9A/A                                      9-27                                      12-23-80

-------
flame.  More  recently,  as  methods for measuring elemental  carbon  have improved,  Wolff et al.
(1980) have performed  atmospheric measurements of absorption/elemental carbo,i mass  in Denver
                          2
and found values of 12.7 m /g.  While the amount of absorption per unit mass depends on chemi-
cal composition  and particle  size  distribution (Waggoner et al., 1973;  Bergstrom  1973),  the
pattern  emerging from  these empirical  and theoretical  studies  is  that absorbing  particles
probably have a more significant visibility impact than their mass would indicate.
9.4.2.5   Summary—The  extinction of  light in  rural  air is  generally dominated by  particle
scattering, while  absorption can rise to  levels comparable to scattering  in  urban  settings.
On a  regional  scale,  almost all of the  particle scattering is contributed by fine particles.
Extinction due to scattering is closely proportional  to the fine-particle mass concentration,
                                                     -6  —1      3
with extinction/ mass ratios in the range of 3-5 x 10   m  /(ug/m ).
9.4.3  Chemical Composition of Atmospheric Particles
     Given the dominance of particles in degrading visibility, it is  natural that their compo-
sition should be studied.   Such knowledge permits estimation of the roles various sources play
in visibility impairment,  so that any anthropogenic sources of impairment can be controlled in
a cost-effective manner to  a level  the public deems desirable.   Before discussing the aerosol
components currently believed  to  be of significance, it  is important to consider a number of
uncertainties in our measurements and deductions.
     Visibility is not  the  simple topic it first  seems  to  be.   Rather, it is one of the most
difficult, being determined  by the  sum of all  atmospheric  constituents,  lighting conditions,
and the  observer.   Needless  to  say, these  factors  are  extremely  variable.   Though  it is
generally accepted that fine particles  cause most visibility problems,  the concentration and
composition of these particles can  vary considerably at  different times and sites around the
country.   Furthermore,  field studies which  characterize the aerosol  need  to  include diverse
and sometimes  elaborate instrumentation  and techniques in  order to  measure all  relevant para
meters simultaneously.   There  is  evidence  of volatile aerosol  (e.g.,  ammonium nitrate) which
exists in the  atmosphere  and  thus  degrades  visibility,  but cannot always  be  retained by
conventional  filtration for  subsequent  analysis.  Some fine particles (e.g.,  diesel  exhaust)
are distinctly non-spherical  and  defy theoretical attempts to  model  their optical influence,
or even report a size distribution.   Several fine mass componenets (e.g., ammonium nitrate and
elemental carbon)  have been  either measured  inaccurately  or not at all  until  recently.  A
particularly difficult-to-hahdle fine mass component is water.  Water degrades visibility only
when  in the  liquid  phase,  but it is ubiquitous and causes particles of several common species
to grow  or shrink,  which  can  affect visibility drastically.   Thick  hazes  or  fogs  are often
dismissed as  caused  solely  by high humidity,  whereas  in some cases they may  not  have formed
without the presence of anthropogenic nuclei.
     A common  mistake  in  assessing the compounds  responsible  for  visibility impairment is to
assume that  a  compound's   contribution  to mass  is equal  to  its contribution  to visibility

SOX9A/A                                      9-28                                      12-23-80

-------
impairment.   Its  visibility contribution can  be  greater or less than its mass  contribute-,
depending primarily on  its  size distribution and refractive index.   While refractive  index  is
known (at least approximately  ) for most fine mass constituents, it is difficult to determine
the size distribution  of some  components (because of detection limits, artifact formation,  o^
volatilization) with the requisite accuracy.
     Fortunately,  there  are  reasons to believe we can  circumvent  these  uncertainties.   Suit-
ably accurate  analytical methods  are  evolving for  nitrate measurement.  Carbon  measurement
techniques  have  been   intercompared  and  applied  to  the atmosphere.   Relatively  constant
scatter/fine mass ratios are being reported for a variety of sites (Waggoner,  1980).   Extinc-
tion calculated from gross size distribution measurements is usually within a factor of two  of
measured  values   (Ensor,  1972; Patterson  and Wagman,  1977).   Regression analyses show  con-
sistent correlations  between  scattering  or visibility  and  sulfate concentration (White and
Roberts, 1977;  Cass, 1979; Trijonis, 1978a,b).
     Current knowledge  indicates  that  fine aerosol is composed of varying amounts of  sulfate,
ammonium, and  nitrate  ions;  elemental  carbon  and  organic  carbon compounds;  water and smaller
amounts of  soil dust,  lead compounds, and  trace  species.   The following discussion separates
the components, whereas  in reality they may  exist as internal mixtures  (i.e., coexist within
the same particle).
     Sulfate occurs predominately in the fine mass (Stevens et al., 1978; Tanner et al., 1979;
Lewis and Macias, 1979;  Ellestad, 1980).  Sulfate ion generally comprises 30 to 50 percent  of
the fine mass  at  a  wide variety of sites (Stevens et al., 1978; Pierson et al., 1980; Stevens
et al.,  1980;  Lewis  and Macias,  1979;  Ellestad, 1980).  Sulfate  occurs  with  cations of H2,
(NH4)H, and  (NH4)£  (Stevens, 1978; Pierson et al., 1980; Stevens et al., 1980; Tanner et al.,
1979).    Indirect  measurements of  sulfate  by  examining  the  scattering  response of ambient
aerosol to changes of relative  humidity confirm the prevalence of H^SO^ and its ammonium salts
(Weiss  et al., 1977;  Waggoner  et  al.,  1980).   Regression analyses by Cass  (1979), White and
Roberts (1977), Trijonis (1978a,b), Grojean et al. (1976), Leaderer et al. (1978), and Heisler
et al.  (1980)  show  significant correlations between  sulfate  concentrations  and visibility  or
extinction.
                                         i
     Ammonium  ion is  typically found to account  for 10 to 15 percent of the fine mass (Lewis
and Macias,  1979; Patterson and Wagman, 1977).   It  often correlates well with sulfate levels
(Tanner et  al.,  1979).   Due  to  the  possible reaction  of ammonia with previously collected
sulfuric  acid  particles,  ammonium  ion  values may  be  higher  than exist in  the atmosphere.
     It is  currently  impossible  to make  valid  statements about  ambient particulate nitrate
concentrations due to recognized sampling problems (Appel et al., 1979; Spicer and Schumacher,
1977 and 1979).  Simple  filtration (even with non-alkaline, non-reactive, high-purity filters)
may  not give  true  values,  due to  the  tendency of ammonium nitrate to  seek equilibrium with
ammonia and gaseous nitric acid during sampling and storage.

SOX9A/A                                      9-29                                      12-23-80

-------
     Concentrations of elemental  and  organip carbon have been reported by Wolff et al.  (1980)
for a  variety  of U.S.  sites.   Elemental carbon was  found to range from 1.1 ug/m  at a  remote
site to  13.3 jjg/m  in New York City, with  about 80  percent in the fine  fraction.   Based  on
concurrent Denver measurements  of b .   by a modified integrating plate method,  Wolff calcula-
ted an absorption/mass  ratio of  12.7 m /g,  indicating that elemental  carbon  may  have  four
times  as  much  visibility  impact  as  its  mass would  indicate.   Wolff further  concludes  that
elemental carbon  is the  only  significant light-absorbing specie.  This  conclusion is consis-
tent with those  of  Pierson and Russell (1969), Rosen  et al. (1980),  and Weiss  et al, (1978).
Weiss  aspirated  various  solvents  and acids through collected aerosol.   After  finishing  with
aqua regia,  he  found  that 80  percent of the mass was gone, but 90 percent of  the absorption
remained.  He reasoned that elemental  carbon was the likely specie.  Determinations of organic
carbon concentrations  will not be further discussed  because of uncertainties in  their measure-
ment from adsorption or  volatilization.   Improved techniques for organic particulate measure-
ment are being developed.
     As mentioned earlier, water  affects  visibility only when  in  the liquid phase.   Unfortu-
nately, no direct measure  of  liquid water's  contribution  is possible, due to its rapid phase
change and the  fact that typically only 1 to 100 ppm of all water in  a given volume exists  in
the liquid phase.  However, Waggoner et al.  (1980) have developed a nephelometer with humidity
control and have shown a  substantial dependence of scattering on relative humidity at a  number
of sites  (see  Figure  9-13).   Waggoner's method cycles  the  aerosol's  humidity in a very short
period, so the input aerosol  can be assumed to have  not changed in composition or concentration.
     Inasmuch as  sulfate  is a prominent,  ubiquitous,  and hygroscopic  specie of  fine particles
(Garland, 1969;  Charlson  et al.,  1978; Waggoner et  al., 1980),  water  undoubtedly has a  signi-
ficant role  in  visibility  impairment  at higher humidities.   Tang et al.  (1978)  have performed
theoretical and experimental  studies of the growth of sulfate aerosols with humidity and found
the molecular composition to be of considerable importance.   For example, for ammonium sulfate/
sulfuric  acid  ratios  above 0.95,  deliquescence occurs  at  80 percent relative humidity, for
ratios 0.75 to 0.95 at 69 percent, for ratios 0.5 to 0.75 at 39 percent, and for a ratio below
0.5 growth is  not abrupt (deliquescent) but smooth  (hygroscopic).   Thus, to predict the light
scattering behavior of a sulfate aerosol, one needs  detailed knowledge of its molecular compo-
sition  in-situ.    Compounding  the problem  of predicting  scattering-humidity   behavior,  the
electromagnetic scattering efficiency function varies greatly in the size region where sulfates
and  other secondary  particles exist.   Thus  a small  error in pre-growth  size distribution
measurement may produce a significant error  in the predicted scattering.  A further compounding
problem  is  that  as humidity  decreases salt aerosols exhibit  hysteresis (Orr  et al1., 1958;
Charlson, et al., 1978) and do not lose their water at the same humidity at which they gained it
It must be concluded that the role of water  in visibility impairment can be significant and is
difficult to predict.

SOX9A/A                                       9-30                                      12-23-80

-------
        (a)
                            TYSON, MO,
              .1 t  i j i  > «  i  * i  • j j i  i
 Q.
 V)
    o
        (b)
                   TYSON, MO.
1  . i  ,  !  . >  , I  «  !  i I  , I  .  I  i
 a
 (A
CO
1  I '  I  ' I  '  I  ' I'M!
(C)
                    TYSON, MO.
i  1 i i  j 1 i  ] i l_i l_i_J >  I -il..-i
        (d)
                            TYSON, MO.
                            t  i i  I . I  .
                     50%            100%
            RELATIVE HUMIDITY
                                                                 PT. REYES, CA.
                                                           . !  , i  i .1 i I  ,  ! ,  i ,

                                         I
                                         DC
                                          a
                                          CO
                                         .0
                                          II
                                         03
                                                         SANTA ANA CONDITIONS
                                                         PASADENA. CA.
                                                 ,  ) ,  1 .  I  i I  , I  .  I i  I .  I ,  I  .
                                               (g)
                                                                      MONTANA DEORO
                                                                      STATE BEACH, CA.
                                                        . i  »
                                                                    COMPOSITE
                                                                    •  I .  I '  I  -
                                                              50%          100%
                                                     RELATIVE HUMIDITY
    Figure 9-13. Humidograms for a number of sites show the increase in bsp which can be expected at
    elevated humidities for specific sites or aerosol types (marine, Point Reyes, CA; sulfate, Tyson, WO)
    and the range observed for a variety of urban and rural sites (composite).
    Source: Covert et al. (1980).
                                             9-31

-------
     Minor contributions to fine mass are made by soil-related elements,  lead compounds (espe-
cially in urban areas), and trace species (Stevens et al., 1978).
9.4.3.1  Summary—Current knowledge indicates that fine aerosol is composed of varying amounts
of sulfate, ammonium, and nitrate ions, elemental carbon, organic carbon  compounds,  water,  and
smaller amounts of  soil  dust,  lead compounds, and trace species.   Sulfate often predominates
the fine mass  and  visibility impairment, while elemental carbon is often the primary visibil-
ity impairing  specie  in  urban  areas.   Significant variations can occur at different times  and
sites.  Our  knowledge of the roles of  several  possibly important species is  hindered by  the
lack of sufficient good data.
9.5  HISTORICAL PATTERNS OF VISIBILITY
     Records of visual range  (prevailing  visibility) can  be used to gain insight  into  the
effects of changing  emission  patterns on visibility.   As one  example  Marians  and  Trijonis
(1979) have derived statistical relationships between light extinction (computed from visibil-
ity data)  and  historical  emission  trends.   Yearly  values of  extinction from  four Arizona
airports were  regressed  against  statewide  emissions  of  smelter  SO.  nonsmelter SO. NO ,  and
                                                                   2\                n    ^
RHC (reactive  hydrocarbons).   Smelter SO   was  found  to be  the most  significant  variable.
Particularly close  relationships  between  Arizona smelter  SO  and visibility at Tucson  and
                                                              ^\
Phoenix are shown  in Figure 9-14.
     Table 9-2 summarizes  the results  of the correlation/regression analysis between yearly
airport extinction  (visibility)  data and  Arizona  smelter  SO   emissions.    The  correlation
coefficients and Student's  t-statistics indicate significant statistical  relationships at high
                                                                                             A
confidence levels.   The  regression (extinction/emission)  coefficients  of  0.04 + 0.005 (10
m)  /(1000 tons per  day  of SO ) are remarkably consistent from site to site and represent the
                              X
change in yearly median  extinction associated with a given  change in SO  emissions; that is,
adding 1000 tons per day of SO  tended  to  increase  yearly median extinction by approximately
0.04 (104 m)"1.
     Perhaps the  best example  of  changed emission  patterns  is  a strike that  shut down  the
copper  industry over  a  9-month period  in 1967-68.    In the  Southwest  at this  time, copper
production accounted  for over  90 percent of the SOV  emissions, less than 1 percent of the NO
                                                   f\                                          r^
emissions, and less  than  10 percent  of  the conventional particulate emissions (Marians  and
Trijonis, 1979), and should therefore have affected visibility primarily through its contribu-
tion  to  sulfate  loadings    Substantial  decreases   in  sulfate  occurred  at  five  locations
(Tucson, Phoenix, Maricopa  County, White Pine,  and  Salt Lake City) within 12  to 70 miles of
                                                                                    t   i
copper smelters as  shown in Figure 9-15.  More  notably, sulfates dropped by about 60 percent
at' Grand  Canyon and!  Mesa  Verde;  these remote sites are located 325  to  500 km  from the main
smelter area in southeast  Arizona.  Comparing measurement during the strike with those during
the surrounding 4  or 6 years, Trijonis  et  al.  (1978a)  found  a large decrease in Phoenix sul-
fate loadings, accompanied by a substantial improvement  in visibility (Figure 9-16).

SOX9A/A                                      9-32                                      12-23-80

-------
  Ul
  U
  cc
  UJ
  a.
  M
  UJ on
  j 80
  I
  g 60
  VI
  > 40


  £ 20
  55
  >  n
                                                        6000
                                                              „ 2
                                 -  o
                            5000 |  §

                            4000 g  3
                                 v  3»
                                 a  m
                                 o  r-
                            3000 3  -I
                                 B:  m
                                 *  a
                            2000 ^  M
                                    O
                                   x
              50
                    55
60
65
70
75
                              YEAR
X    80
t- M
» UJ
5-j 70

-------
                                 TABLE 9-2.   CORRELATION/REGRESSION ANALYSIS BETWEEN AIRPORT

                                         EXTINCTION AND COPPER SMELTER SOV EMISSIONS
                                                                         y\
OJ
fa

Data set
Tucson (1950-75)
Tucson (1959-75)
Phoenix (1959-75)
Wins low (1948-73)
Prescott (1948-75)
Prescott (1948-69)
Correlation
coefficient
0.91
0.88
0.81
0.68
0.70
0.70
Regression coefficient
extinction/emissions,
(10* mfVdOOO TPD)
0.035
0.038
0.041
0.047
0.031
0.039
t- statistic
(t * 1.7 for 95% confidence)
(t s 2.5 for 99% confidence)
11.1
7.2
5.4
4.5
5.0
4.4

-------
?
u>
o
       ' 250 TONS/DAY SO
                    URBAN SITE
      NONURBAN SITE
        100     200

         I	I
    SCALE, milM
                                                                                 MESA VERDE

                                                                              7%l NATIONAL PARK
                                                 PHOENIX 1-62%

                                                               X
                                                            67%)TUCSON
   Figure 9-15. Seasonally adjusted changes in sulfate during the copper strike are compared with the

   geographical distribution of smelter SOX emissions.


   Source: Trijonisand Yuan (1978a).

-------
(£>
CO
cn
  •    - 250 TONS/DAY SO2





(J  URBAN AIRPORTS





       NONUR^AN AIRPORTS
                    0

                    L
        100
         I
                        SCALE. milM
 200

_J
                      Figure 9-16. Seasonally adjusted percent changes in visibility during the copper strike are compared

                      with the geographical distribution of smelter SOX emissions.
                      Source: Trijonisand Yuan (1978a).

-------
     Visibility improved at  almost all  locations during the strike,  with the  largest  improve-
ments occurring near and downwind (north) of the copper smelters in southeast  Arizona  and  near
the copper smelters  in  Nevada and Utah.  The nine locations showing  statistically  significant
improvements are all within 150 miles of a copper smelter   Attributing the improvement  in visi-
bility entirely to the drop in sulfate  levels yields an estimated extinction efficiency  of 3 9
    ~6  ~1      3
x 10   m  /(ug/m ), in agreement with data in Table 9-1.
     Altshuller (1973)  has  noted an increase over the past decade in sulfate concentration at
nonurban sites  in the  Eastern United  States,  which is not inconsistent  with  the decreasing
trend in nonurban  site  median visibilities noted by  Trijonis  et al. (1978b).   Unfortunately,
the historical  record of sulfate concentrations extends back  only to the mid-I9601s.   Within
the Eastern United  States,  over 90 percent of  the  S0x emissions are associated with  the  com-
bustion of  coal and oil.  One apparent conclusion is that visibility  reduction is currently
due  in  largp  part to  increases  in sulfate  aerosols,  which  are  formed primarily from  coal
combustion-related SO* emissions; examination of the trends and changing spatial distributions
of coal  use should be comparable with the change in the light extinction coefficient.
     Air pollutants emitted  over the Eastern United  States  result mainly from the combustion
of fossil fuels—coal,  oil  products, and gas.   The  great  spatial  and seasonal  variability of
haziness (inverse  of  visibility) prompted Husar et al.  (1979) to examine the patterns of  coal
consumption in  the  Eastern  United States over the past few decades.   For comparison with  coal
consumption estimates,  visibility data are  expressed in terms of a light extinction coeffi-
cient, b  ., via the Koschmieder formula:  b  . = 3.92/V.
     Figure 9-17  illustrates  the striking similarity between  summertime average haziness and
coal use within the Eastern United States over the past three decades.
     Further  support  for the  relationship  between  coal consumption and haziness  is  given in
Figure 9-18.
     As shown in Figure 9-18,  in 1951 the haziness was most pronounced in the wintertime,  when
the coal consumption  was the  highest.   By 1974, there was a shift toward a summer  peak, coin-
cident with the increasing  summer use  of coal.   Such coincident behavior alone cannot estab-
lish cause-effect  relationships.   It is nevertheless instructive to examine the more  detailed
spatial  and temporal patterns of coal use and haziness (specifically, extinction).
     Since 1940, the  trend  in coal consumption has been more pronounced in the summer than in
the winter (Figure  9-18;  U.S.  Bureau of Mines, 1933-74); since 1960 summer coal use has grown
by about 5.8  percent  per year compared with 2.8 percent per year for winter coal demand.   The
monthly coal combustion peaked in the winter in the early 1950's, but the seasonal  pattern had
shifted to a summer peak by 1974 (Figure 9-19).                     f
     The corresponding  regional  trends  of haziness in the Eastern United States (Figure 9-20;
Husar et al.,  1979) exhibit changes similar to those of coal combustion.
     In the Ohio  River  Valley region,   the winter (quarter 1)  average extinction (bext, km  )
decreased slightly, whereas  the spring (quarter 2) average increased.  The summer  (quarter 3)
SOX9A/A                                      9-37                                      12-23-80

-------
         176
         150
         125
      110°
      I
      1  75
      i
          50
          25
                I   I    I    I    I   I   I   I    I
                I   I    I    I    I    I   I   I    I    I    I   Ins
                                                           4.0
                                                           3.5
                                                           3.0
2.5   E
     o
2.0    i
                                                           1.5
                                                           1.0
           1940    1950    1960    1970    1980    1990    2000

                                 YEAR


         Figure 9-17. Compared here are summer trends of U.S. coal
         consumption and Eastern United States extinction coefficient.

         Source: Adapted from Husar and Patterson (1980).
 I   I   I  I  I  I   I   I   I   I   I
                 rrT"!
              ELECTRIC UTILITIES
              iimiiiiiiikiii
JFMAMJJASOND   JFMAMJ   JASOND
                                                                              g
                                                                              i
      Figure 9-18. In the 1950's the seasonal coal consumption peaked in
      the winter primarily because of increased residential and railroad use.
      By 1974, the seasonal pattern of coal usage was determined by the
      winter and summer peak of utility coal usage. The shift away from a
      winter peak toward a summer peak in coal consumption is consistent
      with a shift in extinction coefficient from a winter peak to a summer
      peak in Dayton, OH, for 1948-52.

      Source: U.S. Bureau of Mines Yearbooks 1933-74.
                            9-38

-------
2
O
t-
cc
O
I
c/3
H
O
   175
   150 -
   125 -
   100 -
                                                   175
     1940
                                          1980
                                                            " ' ' i '  ' ' ] '   • i

                                                                      WINTER
1940
                                                                                         1980
 Figure 9-19. In 1974, the U.S. winter coal consumption was well below, while the summer consumption
 was above, the 1943 peak. Since 1960 the average growth rate of summer consumption was 5.8 percent

 per year, while the winter consumption increased at only 2.8 percent per year.


 Source: U.S. Bureau of Mines, Minerals Yearbooks 1933-1974.
                                        9-39

-------
                    OHIO RIVER
                                 OUPBTE* ?
0.1
 o L.
 194050  60 70  80 90    194050  60  70 80  90

                NE. MEGALOPOLIS
    1940 50 60  70 80  90    1940 50  60 70  80  90
                 EASTERN SUNBELT
    1940 50 60  70 80  90    1940 50  60  70  80  90
                       YEAR
                                                          1940 50  60 70  80 90    1940 50 60  70 80  90


                                                                       SMOKY MOUNTAINS
                                                          1940 50  60 70  80 90   1940 50 60  70 80 90
                                                          0 L_ ..._.....,.........,		J      '		.,_		'
                                                          1940 50  60 70  80 90   1940 50 60  70  80  90
                                                                           YEAR
Figure 9-20.  Trends in the light extinction coefficient (bgxt) in the Eastern United States are shown by
region and by quarters; 1 (winter), 2 (spring), 3 (summer), 4 (fall).

Source: Husar et al. (1979).
                                                     9-40

-------
extinction increased  from  roughly 0.25 in the 1950's  (a  visibility  of 10 miles)  to abo'.t  0.-
in  the  1970's  (a  visibility of  less than  6  miles).   Fall  (quarter 4)  extinction  remain
essentially unchanged.   The summer  average  increased from  about 2  to 3, corresponding to  *
reduction in visibility  from 12 to 8  miles  (20  to 13 km).  The  Northeast megalopolis  regie-
shows a general  decline  in haziness during quarters 1 and 4, whereas quarters 2 and 3 display
a slight  increase  from b = 2.7 (9  miles,  15 km) to 3 (8 miles,  13  km) x lo'^if1.   The Smo*y
Mountain  region  displays  a strong increase in the average summer quarter extinction coeffici-
ent from about 1.6 to 4, corresponding to visibility deterioration from 15 to 6 miles.  Small-
er but still pronounced  increases are noted for quarters 2 and 4.  Evidently the Smoky Moun-
tains have become  appreciably  "smokier"  over the past 20 summers.  The Eastern Sunbelt region
has an increased haziness  for all quarters, most pronounced being the summer quarter, with an
increased from  2  (12  miles) to  3.5  (7  miles).   In the  Midwest,  extinction  during the first
quarter fluctuated slightly with no  discernible  trend.   The  spring  and fall quarters  have
increased appreciably, but summer values have nearly  doutfled,  from  1.5 to 3 (16 to 8 miles).
     The  spatial shifts  of Eastern United States  haziness  are displayed in greater detail in
Figure 9-21 (Husar et al., 1979).
9-5.1  Natural Versus Manmade Causes
     Vision in  the natural,  unpolluted atmosphere  is  restricted  by  blue sky scattering, (air
molecule  light  scattering is  often termed  Rayleigh  scattering)  by  curvature  of the earth's
surface,   and  by suspended liquid or  solid natural aerosols.   Important sources  of natural
aerosols  include water (fog, rain, snow), wind-blown dust, forest fires, volcanos, sea spray,
vegetative emissions, and decomposition processes.   The  particle  free  atmosphere scatters
light and limits visual range to about 200 miles at sea level.
     Dark objects,  such  as distant mountains, when viewed  in daytime through a particle-free
atmosphere, appear bluish  because  blue light  is scattered  preferentially  into the  line of
sight.  Bright  snow-covered  mountain tops or clouds on  the horizon  can appear yellow to pink
because the atmosphere scatters more of the blue light from bright "targets" out of the line
of sight  leaving the longer wavelength colors.   The  actual  visual range in the particle-free
atmosphere is also limited by the earth's curvature.   Thus, Rayleigh scattering is seldom the
limiting  factor  in  the   detection  of  the most distant  objects,   i.e.,  the visual  range.
Rayleigh  scattering  is,  however,  important in  reduction of  visual  texture  and  in bluish
coloration of distant dark visual targets.  Moreover, air scattering is solely responsible for
the blue color of the non-horizon sky.
     Fog is a naturally occurring phenomenon which can reduce the visual range to nearly zero.
It  is  characterized  by  high liquid water  content, typically  over  1000  ug/m  ,  dispersed in
droplets  with a mean diameter of  several micrometers  or more.   In  "natural"  fogs  all colors
are scattered  and absorbed  about equally,   so  the atmosphere  appears white (Husar,  et al.,
1979).

SOX9A/A                                      9-41                                      12-23-80

-------
              194852
                                             196064
                                                                           1970-74
                        0.277 A
                                                                                            \
                                                                   cn    t
      EXTINCTION
      COEFFICIENT, km"1

      VISIBILITY, miles
>0.36

< 6.6
0.3-0.36

 6.6-8
0.24-0.30

  810
0.18-0.24

 10-13.3
< 0.18

 >13.3
Figure 9-21. The spatial distribution of 5-year average extinction coefficients shows the substantial
increases of third-quarter extinction coefficients in the Carolines, Ohio River Valley, and Tennessee-
Kentucky area. In the summers of 1948-52, a 1000-km size multistate region centered around Atlanta,
GA, had visibility greater than 15 miles; visibility has declined to less than 8 miles by the 1970's. The
spatial trend of winter  (first quarter) visibility shows  improvements in the Northeast megalopolis
region and some worsening in the Sunbelt region. Both spring and fall quarters exhibit moderate but
detectable increases over the entire eastern United States.

Source:  Husar et al. (1979).
                                            9-42

-------
     The historical frequency  of  occurrence of fogs  in  the  continental  United States  reveals
considerable geographic variability  (Figure 9-22).   Coastal  areas experience the highest fre-
quency.   Most  inland  portions  of the United  States  west of the Applachians  can  expect  fe«er
than 20  days  of  fog  per year,  with less  than five days  of fog annually  in  the  arid  West.
     With the  exception of coastal  and  mountainous  regions, fogs are rare  during  the surfer
months.   Fogs tend  to be  localized  events  of,  at  most,  a few hours duration, commonly dj^i-c
the early morning hours.    On  an hourly basis, fogs  exist  less than one percent of the time
(Conway, 1963).   Thus,  the overall  contribution of fog  to  the  degradation of  visual  air
quality is small,  and it  is an insignificant cause  of reduced visibility during the daylight
hours.
     Thunderstorms, other rainfall, and snow can also reduce visibility.   East of Nevada, most
of the U.S.  experiences from 30 to 50 days each year with thunderstorm activity.  Such storms
are most common on summer afternoons.  Since thunderstorms are usually intense but brief, they
also contribute to  visibility  reduction less than one percent of the time on an annual basis.
     Snow is  a major  natural  impediment  to  visibility.  It  is an  important  factor  in man,
regions of  the North  and in some mountainous  areas,  where  blowing snow occurs  from  1  to 12
percent of winter  hours (Conway,  1963).   During the  winter  months snowstorms may account for
most of the hours  of  reduced visibility, and certainly may dominate the epidoses of extremely
low visibility in winter months.
     The natural  contribution  of  fog,  thunderstorms, snow, and  other  forms of precipitation
can thus cause  severe degradation of visual air quality.  With few exceptions, however,  these
intense but infrequent  events  do not dominate  the average visual range within the continental
U.S.;  typically only a small percentage of the  hours involve storms or fog.
     In the  arid West the contribution of wind-blown dust to degradation of visual air quality
is an  important problem.    Because human  activities  which  disturb  natural  soil  surfaces add
significantly to wind-blown dust, dust storms are only partially natural.
     The  suspension  of  particles   from  the  surface is  determined by  cohesiveness  of  the
particles to the underlying material, the force of the surface wind, and the topography of the
surface  layer.   The  ideal   situation  leading  to  suspension  of  surface  material  is a dry,
crumpling or disturbed  crust in flat terrain without vegetation.   Agitation of such surfaces
by strong winds and turbulence can transform a pristine arid atmosphere into a dust storm with
severely reduced visibility.   Suspended crustal material in  a  dust storm usually consists of
coarse  solid  particles  with  volume mean  diameters of tens  of micrometers  (urn) or  more.
Patterson et  al.  (1976)  found that the  optically important fugitive dust particles  include
those  up  to  40 urn  in diameter.   Orgill   and  Sehmel  (1976) have  analyzed the  frequency of
occurrence of dust  storms in the continental United States in great detail, based on National
Weather Service observations of wind-blown dust and  sand  associated with visibility of seven
miles or less.  The peak  hours for  dust are noon  to eight p.m., during the period of maximum

SOX9A/A                                      9-43                                      12-23-SC

-------
Figure 9-22. Average annual number of days with occurrence of dense fog. Coastal
and mountainous regions are most susceptible to fog.

Source:  Conway (1963).
  Figure 9-23. Annual percent frequency of occurrence of wind-blown dust when
  prevailing visibility was 7 miles or less,  1940-1970. Dust is a visibility problem
  in the Southern Great Plains and Western desert regions.

  Source: Adapted from Orgill and Sehmel (1976).

                                    9-44

-------
thermal  turbulence.   Forested, coastal  and mountainous  regions  have few,  if  any,  episodes.
The Pacific coast has high (>0.1 percent) incidence of dust only in the San Joaquin Valley *'~
the Los  Angeles  Basin.   Western desert areas in Eastern Washington, Western Nevada,  Utah, Ne«
Mexico and  Arizona are  also prone  to  dust.  The  highest dust frequency  is  in  the Southern
Great Plains, where' wind-blown dust is a  serious  problem up to 3 percent of the time (Figure
9-23).
9.5.2  Summary—The impact of sulfates in reducing visual air quality in individual plumes and
on the regional scale of 1000 km is no longer a matter of dispute.   Considerable evidence from
chemical  mass balance  methods indicates that sulfates, which make up approximately 50 percent
of the fine-aerosol mass, contribute considerably more to visibility degradation than do other
chemical  species.   Finally,  the 30-year record of the  spatial and temporal  trends  of coal
combustion and visibility  suggests that the increasing emissions of SO  since the 1950's have
been  associated  with  similar increases in  haziness.   This association is not as well defined
                                                          <
as that for sulfate concentrations and visibly decreases.
9.6 SOLAR RADIATION
      Incoming solar radiation is composed  of the direct beam and the diffuse skylight arising
from  the light-scattering atmosphere (Figure 9-24; Gates, 1966).  The relative contribution of
the skylight  is  least  at  noon and greatest at  sunrise  and sunset.  At sea  level,  and for a
clean atmosphere, skylight contributes at least 10 percent of the total radiation.
     Aerosol layers in the atmosphere scatter and absorb solar radiation (Figure 9-25).  Some
of the scattered radiation is directed upwards  and lost to space; some is directed downwards
to the earth's surface.   Most of  the  solar radiation eventually reaches the surface, but its
spectral  and  directional composition, that is,  the "quality" of the  solar radiation,  may be
changed by atmospheric haze (Weiss et al.,  1978).  A small fraction of the scattered radiation
may also be lost  back to  space,  in which  case  the amount of energy  reaching  the surface is
reduced,  contributing  to the cooling at the earth's surface.  A fraction of the radiation may
also  be  absorbed by aerosols,  further  reducing the amount of  radiation  reaching the surface
but at the same time heating the aerosol layer itself from the earth's surface.
      It should be noted that because aerosols are not uniformly distributed in the atmosphere,
their effects are  spatially nonhomogeneous.  First, the horizontal spatial scale encompassing
aerosol  source,  transport, and removal  in the lower troposphere is  variable  but often about
1000  km.   The vertical  spatial  scale of  noninfluenced aerosols is  also  quite variable, but
often the particles are concentrated in a  layer from 500 to 2000  m  deep at the earth's sur-
face.   Hence,  the  aerosol  effects should  be concentrated in the lowest  layers of the atmos-
phere, especially in industrial regions.  %
     Global-scale effects  might  also occur.  If the  effects  in industrial regions are strong
enough,  then  the heat  balance of the entire  earth could be influenced.   On  the other hand.
effects from long-lived aerosol, for example, in the stratosphere, might lead to direct physi-
cal effects on a global scale.
SOX9A/A                                      9-45                                      12-23-80

-------
         I
                         GLOBAL RADIATION

                           DIRECT SOLAR
     0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1  1.2 1.3 1.4 1.5 1.6  1.7 1.8

                          WAVELENGTH, jmm

Figure 9-24. Solar  radiation intensity spectrum at sea level in
cloudless sky peaks in the visible window, 0.4-0.7 jum wavelength
range, shows that in clean remote locations, direct solar radiation
contributes 90 percent and the skylight 10 percent of the incident
radiation on a horizontal surface. The airmass,  m, is a measure of the
amount of air the direct solar beam has to pass through.

Source: Gates (1966).
                           9-46

-------
Figure 9-25. Extinction of direct solar  radiation by aerosols is
depicted.
                           9-47

-------
9,6,1  Spectral and Directional Quality of Solar Radiation
     The spectral quality of solar radiation on a clear day and on a hazy day in Texas is shown
in  Figure  9-26 (McCree  and Keener, 1974).   On the  hazy  day, the direct  solar  radiation  is
reduced  to  about one-half  of  that  on  a clear  day,  but  most of the energy  has  reappeared  as
diffuse  skylight.   The net effect  is  that  there is an overall  loss  of up to about  10  to  20
percent of the radiation reaching the surface.
     If we take the typical backscattered fraction for haze aerosols to be 10 percent of scat-
tering,  and  the absorption to be  also  about 10 percent,  as suggested by the data  of  Weiss
(1978), then we  can estimate  the amount of  energy  lost  from the surface,  the  amount lost  to
space, and  the amount  absorbed  by the  atmosphere.   On a  day with  half of the  direct beam
transmitted, we conclude  that  10 percent of  the other half, or 5 percent,  is  lost to space,
and the  other  5  percent results in atmospheric  heating.   Together,  these phenomena lead to a
loss of  10  percent  of the radiation.   Although it is not possible to calculate  accurately the
influence this loss  might have  on surface  temperature,  rate of  thawing of frozen ground,
growing season, or  other  climatological  measures,  it is highly  probable that this loss cools
the ground  and heats  the hazy lower layers  of  the  atmosphere.   In turn,  if this  occurs,  it
must  increase  atmospheric stability,  decrease  convective mixing, and  therefore  increase the
rate at which pollutants accumulate.
     No detailed and  routine measurements  of the quality of solar radiation are available for
the United States.   However, the total  solar energy reaching the surface is monitored routine-
ly at many  meteorological  observation  sites in the United States and worldwide.   Unfortunate-
ly, the  large  variability of  such data does  not allow manmade aerosol  effects to  be distin-
guished from other natural causes.
     A data  base that  gives  some information on the  quality of solar  radiation  is  the U.S.
turbidity network  operated at about  40 stations in the country since  1961  (Flowers et a!.,
1969).  If there are no clouds  between the observer and the sun, the intensity of direct solar
radiation for  a  given solar elevation depends on the variable amount of dust, haze, and water
vapor in the atmosphere.  The extinction produced by these constituents is called "atmospheric
turbidity."   The  seasonal pattern of atmospheric turbidity in the United  States  at  29 sites
for 1961-66  is  shown  in Figure 9-27 (Flowers et al., 1969).  At all sites, the  highest turbi-
dity occurs  in the  summertime  and the lowest occurs in  the winter, which  is consistent with
the  haziness pattern  obtained  from  visibility observations  (Section 9.4.2).    However,  the
turbidity of the atmosphere in  the United  States   has  a  strong spatial  dependence. '  In the
Southwestern States  with an annual  turbidity coefficient of about 0.06,  the incoming direct
solar radiation  is  attenuated  by only 13 percent (mostly scattered) compared with Midwestern
State values of about  20 percent.   The  highest turbidity  coefficients were observed  in the
Eastern  United States,  where  summer  values  of 0.2 and winter  values of 0.1 were  typical.
This  means  that in  the summertime about 35  percent of  the direct solar  beam  is diverted  to

SOX9A/A                                      9-48                                      12-23-80

-------
   1.0 -
     0.3
0.4          0.5

       WAVELENGTH,
                                         0.6
Figure 9-26. On a cloudless but hazy day in Texas, the direct solar
radiation intensity was measured to be half that on a clean day, but
most of the lost direct radiation has reappeared as skylight. However,
there is about 20 percent of the solar radiation missing on the hazy
day, some absorbed, and some backscattered to space.

Source: McCree and Keener (1974).
                               9-49

-------
Figure 9-27. To interpret these 1961-1966 monthly average turbidity data in terms of aersol effects on
transmission of direct sunlight use the expression I/I0 = 10'B, where B is turbidity and l/lo is the fraction
transmitted.

Source: Flowers etal. (1969).
                                                9-50

-------
skylight, backscattered to space, or absorbed.  This means that about 3.5 to 7 percent is back-
scattered to space and another 3.5 to 7 percent is absorbed into the atmosphere.
     Since  the  first report of  Flowers et  al.  (1969), the turbidity data  have  been  reported
yearly  by  the  World  Meteorological  Organization  (WMO,  1977).   Comparison of  the  seasonal
turbidity pattern for 1961-66 and 1972-75 is shown in Figure 9-28.  Since the mid-I9601s there
has been  a further  increase in  Eastern  U.S. turbidity,  particularly  in  the  summer season.
Currently  the. summer  average  turbidity in  the  region including Memphis,  TN, Oak  Ridge,  TN,
Greensboro, NC, and Baltimore, MD, is about 0.3.   This corresponds to a 50 percent attenuation
of the  direct solar beam on  an  average summer day.   During  hazy episodes, turbidity coeffi-
cients of 0.6 to 1.0 are often reported, resulting in a condition in which 75 to 90 percent of
the solar  radiation  is removed  from the direct  beam,  7.5 to 18 percent is lost to space, and
7.5 to 18 percent is lost as atmospheric heating.  One of the consequences of such hazy atmos-
phere is the disappearance of shadow contrasts.  It is stropgly suspected but has not yet been
proved that there are effects on agricultural productivity.
     The  spatial  distribution and  trends  of  regional-scale  turbidity  in  the Eastern United
States are consistent with the observed pattern of haziness obtained through visibility obser-
vations.  Both the turbidity and visibility reduction by haze in  the Eastern United States can
be attributed primarily to manmade fine particles (Section 9.4.3).  Bolin  and Charlson (1976)
suggest that  many  of these radiative effects are due to sulfates and conclude that the magni-
tude of effects is comparable to that summarized here.
9.6.2  Total Solar Radiation:  Local to Regional Scale
     Changes in the total radiant energy have been observed within urban areas.  Early measure-
ments in central city  locations, primarily in Europe,  showed  levels typically 10 to 20 percent
below surrounding rural areas.   Robinson (1962) discussed some observations made in London and
in Vienna.  In London the deficit was considerably reduced after  the implementation of a clean
air act.   Measurements on  47 days in  autumn 1973  in the Los Angeles  area are summarized in
Table 9-3  (Peterson  and Flowers, 1977).  In  the St.  Louis area, however,  smaller urban-rural
differences were  observed.   On  12 cloudless  days  in summer 1972,  the  average solar and UV
fluxes at  an  urban site were only  3  and 8'percent, respectively, below those at a rural site
about 50  km from  the city.  The difference between the St. Louis and Los Angeles and European
urban areas appears to involve  both  decreased urban and  increased  rural  attenuation,  and it
may be  that neither the city of St.  Louis  nor its surroundings  over a wide area modify solar
radiation in a manner typical of other  locations.
     Recently, Angel!  and  Korshover (1975) analyzed the  solar radiation duration data (hours
of sunshine) for the eastern half of the United States.  Data for the 1950-70 time period were
obtained with on/off detectors,  and these data are believed to be more reliable for long-tern
trend analysis  than data  from  recorders  of solar  radiation  intensity.   Angel 1  and Korshover
noted some  marked  trends:   in the Southeast and South Central United States, the solar radia-

SOX9A/A                                      9-51                                      12-23-80

-------
   IU
   5
   E
            I" I'l "I'l' I'l
           MEMPHIS, TN
                                        TURBIDITY TREND
                          1961-66  --»-•
                          1972-75  ——
OAK RIDGE, TN
                             I.I.I. I.I.I .1.1,1,1.1.1
GREENSBORO, NC
      JFMAMJJASOND
             MONTH
                                                                             BALTIMORE, MD
                                                                          i.l.I.I.I.1. 1.1.1 1.1.
Figure 9-28. Seasonal turbidity patterns for 1961-66 and 1972-75 are shown for selected regions in the
Eastern United States.

Source: Flowers et al. (1969).
                                                 9-52

-------
                         TABLE 9-3.  SOME SOLAR RADIATION MEASUREMENTS IN
                                       THE LOS ANGELES AREA3
Measurement
Minimum
Average
Maximum
Total
4
11
20
UV
15
29
44
                     Values for the daily average percentage decrease of total
                     and UV solar radiation between El Monte (urban) and
                     Mt. Disappointment (rural).
                     Source:  Peterson and Flowers (1977)..

tion duration  has  decreased by about 4 to 6 percent; however, the North Central is increasing
(Figure 9-29; Angel 1 and Korshover, 1975).  Although the authors do not attribute these trends
to any  specific cause, it should be  noted that there has been an increase in haziness within
that time  period in areas with decreased solar radiation.  It is conceivable, therefore, that
increased haziness causes sunshine-duration detectors to delay the turn-on time in the morning
and advance the turn-off time.  It should also be stressed, however, that changes in the solar
radiation duration may be caused by other natural or manmade phenomena.
9.6.3  Radiative Climate:  Global Scale
     The attenuation of solar  radiation from  scattering and absorption by  particles  in the
atmosphere  is  probably an  important  factor in climatic change.   The  effect  could arise from
redirecting the photon energy, from effects on cloud structure (perturbations of the colloidal
system), or from  effects  on optical  properties,  such  as  increased photon  retention from
absorption on  soot particles.  A report by  Hobbs  et  al.  (1974) argues that aerosols are most
likely the principal agents by which pollutants may affect weather and climate, by influencing
the structure  and  distribution  of clouds.  On  a  global  scale, they noted that the effects of
manmade pollutants on climate are still a matter of debate.  Others, e.g., Twomey (1974), have
pointed up a direct connection between pollution and the number of drops in a cloud and, hence,
an influence on optical thickness and  reflectance  of the clouds (cloud albedo)  and in doing
so, affect climate.
     The importance  of particulate  matter on  climate  may be overshadowed by  that of carbon
dioxide.  A doubling of the concentration of carbon dioxide could result in a 1.5 to 3 degree
centigrade of  warming the  lower atmosphere,  according  to various predictions  viewed  by the
U.S.  Department of  Energy  (U.S.  Department of Energy,  1978).  One set  of calculations fre-
quently cited  in  the  scientific literature predicts a 2 to 3 degree centigrade  rise  in the

SOX9A/A                                      9-53                                      12-Z3-80

-------




i
s
c
u.
Z
O
<
UJ
Q
as







20
18
16
14
12
10
8

6
4
2
0

-2
-4

1<


-
—
_ —
—
—
•» ^

— —
SOUTH CENTRAL
NE\\
^L..^..---""
*** \S 1 1 1 1 1 1 1 1 1 1 1 1 ii i 1 r 1 1 1 1 1 1 1 1 1 1 1 1
>40 50 60 70 80 90 20
YEAR
















00

Figure 9-29. Analysis of the hours of solar radiation since the 1950's
shows a decrease of summer solar radiation over the Eastern United
States. There may be several causes for this trend, including an
increase of cloudiness; some of the change may also be due to haze.

Source: Angell and Korshover (1975).
                              9-54

-------
average temperature of  the lower atmosphere in the  middle  latitudes  with the doubling  of  the
C02 content  of the air and a 3 to 4  fold greater temperature increase in the polar regions
(Manabe and  Whetherald, 1975).   Current model estimates  suggest that the earth  should ha;e
experienced a  few  tenths of a degree  of warming  since the late  1880s  due  to the increase of
C02 concentrations  from about 290 to  the current 335 ppm.   However, it  appears  that natural
variations currently  are  large  enough to  mask  the expected Co2 effect  on temperature (U.S.
Department of Energy, 1980 and Stuivert, 1978).
     Effects  of volcanic  emissions  (see Chapter  4) on  weather have also  been suggested.
Although "spectacular"  sunsets have been associated with major eruptions,  for example Krakatoa
and Katmai,  the effects of  scattering  of solar  radiation from volcanic dust are  not clearly
understood.   Whether  a universal  greenhouse  effect  is created which can  result  in  cooler
weather is still  at issue.  To date, surface temperature and rainfall changes are not conclu-
sively  related to  volcanic  events.   The  large   variability  of weather  and self-preserving
aspects of the  atmospheric  system tend  to  obscure  all  but  the most  dramatic  changes in
climate.
     On local  scales  associated with urban  and industrial  areas, any  significant attenuation
of radiation  by air  pollution can,  in addition to  other  we 11-recognized  factors,  result in
changes in local  weather  (e.g.,  Landsberg,  1970).  It  is  possible that  local-  and regional-
scale  changes  in solar  radiation  caused by human activity may  ultimately  influence the heat
and water vapor  contents  of  the atmosphere  on  very large scales,  but  solar  radiation  and
aerosol levels  measured at  stations  remote from  pollutant sources have  not  as  yet displayed
any trend  that  can be  related to human causes (Fischer, 1967; Ellis and Pueschel, 1971; Hodge
et al.f 1972).
     Unfortunately, there  is little agreement about  whether the  net  effect of increased air-
borne  particulate  concentrations  is  the warming  or cooling of  the  earth as a  whole.  Most
models-can predict either an increase or a decrease in the effective albedo of the earth under
cloudless,  skies,  depending  on which  combination  of  surface albedo,  sun  angle,  particle size
distribution, and particle refractive index  is assumed.  The effects of clouds are very  impor-
tant,   and  the contributions from  infrared  radiation must be considered in order  to obtain a
              t                                 *
complete energy budget  (Wesely and Lipschutz, 1976).
9.7  CLOUDINESS AND PRECIPITATION
     The global cloud cover plays a vital  role in the  earth's radiative budget in reflecting
energy  back,to space,  in  absorbing  both solar and  longwave  (terrestrial) radiation,  and in
emitting its energy downward and outward into space.  Changes in cloud  cover, therefore, alter
the global heat balance.  Cloud- and precipitation-forming^processes may be divided into two
broad classes:  (1) macrophysical processes, which affect the rise and descent of air currents
and the amount of  water vapor  available for condensation; and  (2)  microphysical processes,
which  affect  the  inature of cloud  particles formed during condensation.    The role of  atmos-

SOX9A/A                                      9-55                                      12-23-80

-------
pheric  aerosols,  primarily those  that are  strongly  hygroscopic,  is to  influence  the  micro-
physics of cloud formation.
     On a  global  or even regional scale, the  very small  amounts of moisture that man adds by
land practices or combustion of fossil fuels are negligible in comparison with global evapora-
tion.   On  a  regional  scale,  only one  form of increasing  cloudiness  suggests  itself:   the
formation  of  aircraft contrails  (Machta and  Telegadas,  1974).   Aircraft  contrail  formation
results mainly from the injection of water vapor rather than of aerosols.
     In urban areas,  inadvertent changes of cloudiness as well  as  the quantity of precipita-
tion have  been  well  established.  Such urban impacts also include the frequencies of thunder-
storms and hail  as  well as total  amounts of rain.  In a classical  study, Changnon (1968) has
reported a notable  increase  in days of precipitation, thunderstorms, and hail occurring since
1925 at La Porte,  IN.  Since La Porte  is  30 miles east of  the  Chicago urban-industrial com-
plex, he  proposed that  the  increased precipitation results from inadvertent manmade modifi-
cations.  Figure 9-30 (Changnon,  1968) shows the 5-year running totals of days with smoke and
haze restricted visibility in  Chicago.   This measure of atmospheric pollution has a temporal
distribution  after  1930 rather  similar to  the  La Porte  precipitation curve.   A  noticeable
increase in smoke-haze days began in 1935 and became more marked after 1940, when the La Porte
precipitation curve began its sharp increase.
     A recent report by Clark (1979) critically reexamined this anomoly historically and tech-
nically through the  use of a budgetary hydrologic model.   Clark proposed several explanations
of the anomoly.   If the anomoly was  real,  it  was, at most, a microscale phenomenon), and its
disappearance by  the  mid-I9601s  arose  from  the  dissipation of the precipitation-producing
mechanism.   He  noted  that  although increases  in local  precipitation have been reported down-
wind of urban areas, no records exist of any alteration equalling the magnitude of the La Port
precipitation variation  relative to  surrounding sites.   Additionally, the 1929 to 1963 period
coincides roughly with the tenure of the sole observer.   The accuracy of the guage was also at
issue.   Clark concluded  that errors  by the  observer  and/or guage most  likely  explain the
anomaly.
     As part of project Metromex, studies by the Illinois State Water Survey suggest  increases
of about 30 percent in rain and 200  percent in thunderstorms and hail at single gauging sta-
tions downwind  of St. Louis,  with increases of about 10 percent over a two-county area.  Here
again,   the  physical  causes of the maxima are  not well  understood, but they  do appear to be
associated with perturbations  of the planetary boundary layer and enhanced cloudiness, possi-
bly  resulting from  the addition of aerosols.  It  is regrettable that the complex interactions
of cloud-  and precipitation-forming processes obscure the specific role  of manmade  aerosols.
     The incorporation of particles  into  rain  and fog droplets can change the "quality" of
precipitation by  changing  its  chemical composition.  The  most important impact on precipita-
tion water quality  is  probably that  of "acid rain," discussed in  more  detail  in chapter 8.

SOX9A/A                                      9-56                                      12-23-80

-------
   300
                                         CHICAGO!
                                         SMOKE-I  -
                                         HAZE DAYS
           OBSERVER   j
           CHANGES AT /
           LA PORTE
         1910    1920   1930    1940    1950    1960
                       I
                ENDING YEAR OF 5-YEAR PERIOD


Figure 9-30. Numbers of smoke/haze days are plotted per 5 years at
Chicago, with values plotted at end of 5-year period.

Source: Changnon (1968).
                         9-57

-------
9.S  SUW1ARY
     Traditionally, visibility has  been  defined in terms of  the  distance from an object that
is necessary to  produce  a minimum detectable contrast between that object and its background.
Although visibility is often defined in terms of this "visual  range," with its reasonably pre-
cise definition,  visibility includes more than being able to see or not see a target.   Visibility
eludes seeing targets  at shorter distances than the visual  range and being able to appreciate
the details of the target, including colors.   Visibility impairment can manifest itself in two
ways:  (1) a  layer of haze (or  a plume),  which is visible  because it has a visual discontin-
uity between itself and its background or (2) a uniform haze which reduces atmospheric clarity.
The  type  and degree  of  impairment  are  determined by  the  distribution,  concentrations,  and
characteristics of atmospheric particles  and gases, which scatter  and  absorb light traveling
through the atmosphere.  Scattering and absorption seem to determine extinction.
     The consequences  of visibility degradation  range  from the deterioration or  loss  of the
aesthetic  perception  of  scenic  vistas to annoyances  or outright  hazards to automobile  and
aircraft traffic.
     On  a  regional   scale,  the extinction  of  light  is  generally dominated  by  particle
scattering.   In   urban  areas,  absorption  by  particles becomes   important.   Extinction  by
particles  is  usually  dominated  by  particles  of diameter 0.1 to  2 urn  ("fine"  particles).
Extinction due to  scattering  is  closely proportional  to the fine-particle mass concentration,
                                                    _r _1      O
with extinction/mass ratios in the range of 3-5 x 10  m  /(pg/m ).
     Current knowledge indicates that  fine aerosol is composed of varying amounts of sulfate,
ammonium,  and  nitrate ions, elemental  carbon, organic  carbon compounds, water,  and smaller
amounts of soil  dust,  lead compounds,  and trace species. Sulfate often predominates the fine
mass  and  visibility   impairment,  while  elemental  carbon  is  often  the  primary  visibility-
impairing specie in urban areas.   Significant variations can occur at different times and sites.
Our  knowledge of the  roles of several possibly  important species  is hindered by  the lack of
sufficient good  data.  The 30-year record of  the  spatial and temporal trends of coal combus-
tion  and  visibility suggest that  the increasing  emissions of SO  since  the  1950s  have been
                                                                  J\
associated with  similar  increases  in  haziness.   Nevertheless,  the  relationship  between  SOg
emissions  and  resulting  sulfate concentrations is not  as  well defined  as  the relationships
between sulfate concentrations and visibility reduction.
     The currently available visibility  monitoring methods  measure different aspects of visi-
bility  impairment.   Generally,  contrast type  measurements  (such as  photography, telephotq-
metry, and human eye  observations) relate well to the perception of visual air quality, while
extinction or scattering measurements (such as transmissometry and nephelometry) relate to the
cause  of  visibility  degradation.    Each  of  the  above  measurement  methods can  be used  to
approximate visual  range.   No single  method  has been  proven  totally  effective in measuring
light  absorption.   The  methods  used thus far include determining  the  difference between ex-
tinction and scattering,  several  filtering methods, and  a refractive  index method.
SOX9A/A                                      9-58                                      12-23-80

-------
     Pollutants released to the atmosphere alter the environment in ways other than  visibility
reduction.  They  may lead  to slow and  subtle changes  in  the nature of the  atmosphere  ana,
possibly,  in  climate.   For example,  a  fraction of  the solar radiation may be absorbed  by
aerosols,  further  reducing the amount of  radiation reaching the earth's surface and, at the
same time,  heating the  aerosol  layer itself.  On  a  hazy day, the direct  solar  radiation  is
reduced to  about  one-half of that on a clear day, but most of the energy reappears  as diffuse
skylight.   However, there is an overall  loss of up to about 10 to 20 percent of the radiation
reaching the surface.
     If there are  no  clouds between the observer  and the sun, the intensity  of  direct  solar
radiation  for a  given solar elevation depends on the variable amount of dust, haze, and  water
vapor in the atmosphere.  The extinction produced by these constituents is called "atmospheric
turbidity."  during hazy episodes, turbidity coefficients  of 0.6 to  1.0 are often reported,
resulting  in a  condition in which 75 to 90 percent of the* solar radiation is removed from the
direct beam,  7.5  to  18 percent  is  lost  to  space 7.5  to  18 percent  is lost as atmospheric
heating.    One  of  the  consequences of such  a hazy atmosphere is  the  disappearance of shadow
contrast.
     The  attenuation  of  solar  radiation  from  scattering and absorption by  particles in the
atmosphere  is probably  an important factor in climatic change on all scales.  On local scales
associated  with  urban and  industrial  areas,  any significant  attenuation of  radiation by air
pollution can, in  addition to other well-recognized factors, result in changes in local weather.
It  is  possible  that  local- and  regional-scale  changes in  solar radiation  caused  by  human
activity  may ultimately  influence the heat and water vapor contents of the atmosphere on very
large  scales,  but  solar  radiation and  aerosol  levels measured at  stations  remote from pol-
lutant sources have not as yet displayed any  trend that can be related to human causes.
     Cloud- and precipitation-forming processes may be divided into two broad classes:  macro-
physical  and raicrophysical processes.   Macrophysical processes involve the rise and descent of
air masses and the amount  of water vapor available  for condensation.   Atmospheric aerosols,
primarily  those  that  are strongly hygroscopic, influence the microphysics of cloud formation.
The incorporation  of particles  into  rain and fog droplets can change  the "quality" of pre-
cipitation  by changing  its chemical composition.  However, the complex interactions of cloud-
and precipitation-forming processes obscure the specific role of manmade aerosols.
SOX9A/A                                      9-59                                      12-23-80

-------
9.9 REFERENCES

Allen, J.,  R.  B.  Husar, and E.  S.  Macias.   Aerosol size  and  the shape determination  Using a
     laser light scattering spectrometer.  In:  Aerosol Measurement, a Workshop,  University of
     Florida,  Gainesville,  Florida, 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.312-320.

Altshuller,  A.  P.   Atmospheric  sulfur dioxide and  sulfate,  distribution of concentration  at
     urban and  nonurban sites  in the United  States.   Environ.  Sci. Technol. 7:709-713, 1973.

Angell,  J.  K., and J.  Korshover.   Variation in  sunshine  duration over the contiguous United
     States between 1950 and 1972.  J. Appl. Meteorol.  14: 1174-1181, 1975.

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

Barnes,  R.  A.,  and  D.  0.  Lee.   Visibility  in  London   and  the long  distance  transport  of
     atmospheric sulfur.  Atmos.  Environ. 12:791-794, 1978.

Bergstrom,  R.  W.   Extinction and absorption  coefficients of  the atmospheric  aerosol  as  a
     function of particle size.  Beitr. Phys. Atmos. 4§:223-224, 1973.

Bohn, P.   An approach to the problem of estimating demand  for public good.  Swed. J.  Econ.  73:
     1971.

Bolin, B., and  R.  J.  Charlson.  On the role of the tropospheric sulfur  cycle in  the  shortwave
     radiative climate of the earth.  Ambio Ei:47-54, 1976.

Bradford,  D.   Benefit-cost analysis  and demand  curves for  public goods.  Kyklos  23: 1970.

Brookshire, D.  S.   Issues  in valuing visibility:  an overview.  In:  Proceedings of  the Work-
     shop in Visibility Values.  D. Fox, R.  J. Loomis and  T. C. Green, technical  coordinators,
     U.S. Department of Agriculture, Fort Collins, CO,  1979.  pp.  130-138.

Brookshire,  D.,  B.  C. Ives, and  W. D.  Schulze.   The  valuation  of aesthetic preferences.   J.
     Environ. Manag. 3:325-346, 1976.

Cass, G.   R.  On the relationship  between sulfate air  quality and visibility with examples  in
     Los  Angeles.  Atmos. Environ. 13:1069-1084,  1979.

Chandrasekhar, S.  Radiative Transfer.  Dover Publishers,  New York, NY,  1960.

Changnon, S.  A., Jr.  The La Porte weather anomaly—fact or fiction?  Bull. Am. Meteorol.  Soc.
     49:4-11, 1968.

Charlson,  R.  J.,  .A.   H.  Vanderpol,  D. S.  Covert,  A.  P.  Waggoner,  and  N.  C.   Ahlquist.
     H9SOA/(NH..)9SO,  Background  aerosol:    optical  detection  in  St.   Louis  region.  Atmos.
     Environ. 8:1257-1268, 1974.

Charlson, R.  J., A. P. Waggoner, J. F. Thielke.    Visibility Protection for Class  I Areas:   the
     Technical Basis.   Council on Environmental Quality, Washington, DC, August 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 Symposium, United  Nations, Dubrovnik, Yugoslavia, September 7-14,  1977.  Atmos.
     Environ. 12:39-53, 1978.
SOX9/E                               9-60                                            12-20-80

-------
Char!son, 1978.

Charlson, R.  J.,  D.  S.  Covert,  Y.  Tokiwa,  and  P.  K.  Mueller.  Multiwavelength  nephelometer
     measurements in  Los  Angeles, Smog Aerosol,  III.   Comparison to  light  extinction  by NCL
     J. Coll. Int. Sci. 39:260-265, 1972.                                                    2

Clark, R. R.   A  hydrologic reanalysis  of the La Port  anomoly.   Bull. Am. Meteorol. Soc.  60:
     1979.                                                                                  —

Conway, H. M., ed.  The Weather Handbook.   Conway Research,  Inc., Atlanta, GA,  1963.

Cornsweet, T. N.   Visual Perception.  Academic Press,  Inc.,  New York,  NY,  1970.

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.
     Adv. Environ. Sci. Technol.  9:574, 1980.

Dave, J.  V.   Scattering of electromagnetic  radiation by a large, absorbing sphere.  IBM  J.  Res.
     Develop. 13:302-313, 1969.

Davis, R. K.  Recreational planning as  an economic  problem.   Nat. Resour.  J.  3:  1963.

Duntley,  S.  Q., A. R.  Boileau, and R. W.  Preisendorfer.   Image transmission  by the troposphere
     I.  J.  Opt.  Soc.  Am. 47:499, 1957.

Eggleton, A.  E. J.,  and D. H. F. Atkins.   The chemical  composition  of atmospheric aerosols on
     Teesside and its relation to visibility.  Atmos.  Environ.  3:355-372,  1969.

Eiden, R.   Determination of  the  complex index of  refraction of spherical  aerosol  particles.
     Appl. Opt. 10:749-754, 1971.

Ellestad, T.  G.   Aerosol  composition  of urban  plumes  passing over a rural monitoring  site.
     Ann. New York Acad. Sci. 338:202-218,  1980.

Ellestad, T.  G.,  and  R. E. Speer.  Application of  a telephotometer  to visibility measurements
     in  the eastern  United  States.   Presented  at  Symposium  on Plumes and Visibility,  Grand
     Canyon, AZ,  November 1980.

Ellis, A.  T., and  R. F. Pueschel.   Absence of  air pollution trends  at  Mouana  Loa.   Science
     (Washington, D.C.) 172:845-846, 1971.

Ensor, 1972.                                '

Ensqr, D. S.,  R.  J.  Charlson,  N. C.  Ahlquist, K.  T.  Whitby, R. B. Husar,  and B. Y.  H.  Liu.
     Milltiwavelength Nephelometer Measurements in Los Angeles  Smog Aerosol.   I.   Comparison of
     calculated and measured  light scattering.  J.  Coll.  Int.  Sci. 39:242-251,  1972.

Ensor, D.,  and A.  P. Waggoner.   Angular  truncation error in  the  integrating  nephelometer.
 1    Atmos.  Environ. 4:481-487, 1970.

Faxvog, F.  R.  Optical  scattering per  unit mass  of  single  particles.  Appl. Opt.  14:269-270,
     1975.

Faxvog, F.  R., and  D. M. Roessler.'  Carbon aerosol  visibility vs. particle  size  distribution.
  .   Appl. Opt. 17:2612-2616, 1978.

Fischer,   K.   Mass absorption coefficients of  natural  aerosols  in  the 0.4-2.4 mm wavelength
     interval.  Beitr. Physik de  Atmosphiire  46,  1973.

 i
S6X9/E                               9-61                                            12-20-80

-------
Fischer,  W.  H.  Some  atmospheric turbidity  measurements  in Antarctica.   J.  Appl. Meteorol.
     6:958-959, 1967.

Fischer, 1975.

Flowers,  E.  C.,  R.  A.  McCormick,  and  K. R.  Kurfis.   Atmospheric  turbidity  over the  United
     States, 1961-1966.  J. Appl. Meteorol. 8:955-962, 1969.

Fowler, B. W., and C.  C.  Sung.   Scattering  of an electromagnetic wave from dielectric  bodies
     of irregular shape.  J. Opt. Soc. Am. 69:756-761, 1979.

Fox, D., R. J. Loomis, and T. C. Green.   Proceedings of the workshop in Visibility  Values,  Ft.
     Collins, CO, January 18-February 1, 1979.  USDA Forest Service, Washington, DC.

Garland, J. A.  Condensation on ammonium sulfate particles and its effect on visibility.  Atm.
     Environ. 3:347-354, 1969.

Gates,  D.  M.  Spectral distribution of  solar radiation  at the  Earth's  surface.   Science
     (Washington, D.C.) 151:523-529, 1966.

Green, D.  M., and F. W. Campbell.   Optical  and retinal factors  affecting visual  resolution.
     J. Physio!.  181:576-593, 1965.

Groblicki, P.  J.,  G.  T. Wolff, and R. J. Countess.  Visibility Reducing Species in the  Denver
     Brown  Cloud.    Part  One:   Relationships  Between Extinction  and  Chemical  Composition.
     Presented at  the  Symposium  on Plumes and Visibility  Measurements  and Model  Components,
     Grand Canyon, AZ, November 10-14, 1980.

Grosjean, D.  et al.  Concentration, size, distribution and modes of formulation of  particulate
     nitrate, sulfate and ammonium compounds in the eastern part of the Los Angeles air  basin.
     Presented at  the  Annual Meeting of the Air  Pollution  Control  Association,  1976.   Paper
     No. 76-20.3.

Hall, J. S., and L. A. Riley.  Basic spectrophotometric measures of air quality over long paths.
     In:   Radiative  Transfer  and  Thermal  Control,  vol.  49:   Progress  in  Astronautics  and
     Aeronautics.  A. M. Smith, ed., American  Institute of Aeronautics and Astronautics,  New
     York, NY, 1976.   pp. 205-212.

Hansen, A. D.  A.,  H. Rosen, R.  L.  Dod,  and T. Novakov.   Optical  attenuation as a tracer  for
     the  primary  component  of the  carbonaceous aerosol.   In:    Proceedings:   Carbonaceous
     Particles in the Atmosphere, Lawrence Berkeley Laboratory and National Science Foundation,
     Berkeley,  California,  March  20-22,  1978.   T.  Novakov,  ed.,  LBL-9037,  University  of
     California,  Berkeley, CA, June 1979.

Heisler, S.  L., R. C. Henry, and J. G. Watson.  The sources of the Denver Haze in November  and
     December of 1978.  Paper No. 80-58.6, Annual  Meeting of the Air Pollution Control Associa-
     tion, 1980.

Henry, R.  C.   Psychophysics and Visibility Values.  Proceedings of the Workshop on Visibility
     Values,  D.  Fox, R. J. Loomis,  and  T.  C. Greene  (technical coordinators).  Fort  Collins,
     Colorado.  U.S. Department of Agriculture, 1979.                                      jj

Henry,  R. C.  The  application  of  the  linear system theory  of visual  acuity  to visibility
     reduction by aerosols.  Atraos.  Environ.  11:697-701, 1977.

Higgins,  I.  T. T., et  al.  Airborne  Particles, Washington, DC:  National Academy of Sciences,
     1977.
SOX9/E                               9-62                                            12-20-80

-------
Hobbs,  P.  V.,  H.  Harrison,  and  E.  Robinson.   Atmospheric effects  of pollutants.   Science
     (Washington, DC) 183:909, 1974.

Hodge,  P.  W.,  N.  Laulainen, and  R.  J.  Charlson.   Astronomy  and air  pollution.   Science
     (Washington, DC) 178:1123-1124,  1972.

Hodkinson,  J.   R.   Calculations  of  colour  and  visibility  in urban  atmospheres polluted  by
     gaseous NOg.  Air and Wat. Pollut.  Int. J.  10:137-144,  1966.

Horvath,  H.,  and K.  E.  Noll.   The  relationship between  atmospheric  light  scattering  co-
     efficient and visibility.  Atmos.  Environ.  3:543-550, 1969.

Husar,  R.  B.,  and  D.  E.  Patterson.   Regional  scale  air pollution:   sources  and  effects.
     Annals of the NY Academy of Sciences  338:399-417,  1980.

Husar,  R.  B.,  D.  E. Patterson, J.  M. Holloway,  W. E.  Wilson,  and T. G.  Ellestad.   Trends of
     eastern U.S. haziness since 1948.   J_n:  Proceedings  of  the Fourth  Symposium  on  Atmospheric
     Turbulence,  Diffusion, and Air Pollution,  American Meteorological  Society, Reno, NE,  1979.
     pp. 249-256.

Koschmieder,  H.  Theoric der  horizontalen Sichtweite.   Beitr.  Phys.  Freien Atm.  12:33-53,
     171-181, 1924.                                                                   ~~

Landsberg,  H.  E.   Man-made  climatic changes.   Science (Washington, DC)  170:1265-1274,  1970.

Latimer, D. A., R.  W. Bergstrom, S.  R.  Hayes,  M. K.  Liu,  J.  H. Seinfeld,  G.  Z. Whitten,  M.  A.
     Wojcik,  and  M.  J. Hi 1Iyer.  The Development of  Mathematical  Models for the  Prediction of
     Anthropogenic  Visibility  Impairment.  EPA-450/3/78-110a, U.S. Environmental  Protection
     Agency, Research Triangle Park,  NC, September 1978.

Leaderer, B. P.,  D. M. Bernstein, J.  M.  Daisey, M.  T.  Kleinman, T.  J.  Kneip,  E. 0.  Knutson, M.
     Lippmann,  P. J.  Lioy,  K. A.  Rahney, D.  Sinclair,  R.  L. Tanner, G. T. Wolff.   Summary
     R.    J.     Charlson.     Absorption   coefficient    of    atmospheric    aerosol:      a
     method for measurement.  App.  Opt.  12:1356-1363,  1973.

Lindberg,  J.  D., and  L.  S.  Lande.    Measurement of the absorption coefficient of  atmospheric
     dust.  Appl. Opt. 13:1923-1927,  1974.

Machta, L., and K. Telegadas.  Inadvertent Large-scale Weather Modification.   In:  Weather and
     Climate  Modification.    W.  N.  Hess,  ed.,  John  Wiley &, Sons, New York,  NY,   1974.   pp.
     687-725.

Macias et al., 1975.                       !

Macias, E.  S.,  and R. B. Husar.  A review of  atmospheric particulate mass measurement via the
     beta  attenuation  technique.    In:   Fine Particles, B. Y.  H.  Liu, ed., Academic  Press, New
     York, NY,  1976.

Macias,  E.  S.,  D.  L.  Blumenthal,  J. A.  Anderson, and  B.  K. Cantrell.   Characterization of
     visibility.  Reducing aerosols in  the southwestern United States;  interim report on  Project
     VISTTA.   MRI  Report 78  IR 15-85,  Midwest Research Institute,  Kansas  City,   MO,  1978.

Malm, W.  Visibility:  A  Physical Perspective.   In:   Proceedings of the Workshop  in  Visibility
     Values,  D.  Fox,  R.  J. Loomis, and T. C.   Greene  (technical  coordinators).   Fort Collins,
     Colorado,  U.S. Department of  Agriculture, 1979.   pp. 56-68.

Malm, W.  C.   Considerations in the measurement  of visibility.  J.  Air Pollut. Control Assoc.
     29:1042-1052,  1979.


SOX9/E                              ,9-63                                           12-20-80

-------
Malm,  W.  C., E.  G.  Walther,  K.  E. O'Dell,  and M.  Kleine.   Visibility  in the  southwestern
     United  States  from  Summer  1978  through Spring  1979.    In:   Proceedings  Symposium  on
     Visibility, Grand Canyon, No. 3, 1980c.

Malm, W. C., K.  K. Kelly, and J. V. Molenar.   Human perception of visual air quality  (regional
     haze).  In:  Proceedings Symposium on Visibility, Grand Canyon, N. 33,  1980b.

Malm, W. C.,  K.  K.  Leiker, and J. V. Molenar.  Human perception of visual air  quality.  J.  Air
     Pollut. Control Assoc. 30:122-131, 1980a.

Manabe, F.,  and  R.  Whetherald.   The effects  of doubling the C02 concentration on the  climate
     of a general circulation model.  J. Atraos. Sci. 32:315, 1975.

Marians, M.,  and J.  Trijonis.   Empirical  Studies  of the  Relationship Between Emissions  and
     Visibility  in  the Southwest.   Prepared  under grant 802015 by  Technology Service  Corp.
     U.S.  Environmental Protection Agency, Cincinnati, OH, 1979.

McCree, K.  J.,  and  M.  E. Keener.   Effect of atmospheric turbidity on the photosynthetic rates
     of leaves.   Agric. Meteorol.  13:349-357,  1974.

Middleton,  W.  E.  K.   Vision Through the Atmosphere.   University  of Toronto Press, Toronto,
     Canada, 1952.

Mie, G.  Contributions of optics to cloudy media, especially colloidal metal  works.   Am.  Phys.
     25:377-445, 1908.

Mulu, et al., 1980a.

Mungnai, A., and W.  J.  Wiscombe.  Scattering of radiation by moderately nonspherical  particles.
     J. Atm. Sci. 37:1291-1307, 1980.

National Park Service.   My Eyes Need a Good Stretching, Seven Authorities  Speak  Out on Visi-
     bility — Clean Air and Unique Natural Areas.   U.S. Department of the  Interior, Washing-
     ton,  DC, 1978.

Orgill, M.  M.,  and  G.  A. Sehmel.   Frequency and diurnal variation of dust  storms in the con-
     tiguous United States.  Atmos. Environ. 10:813-825, 1976.

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

Patel,  A.  S.   Spatial  resolution  by  the human  visual  system.  The  effect of  mean  retinal
     illuminance. J.  Opt. Soc. Am. 56:689-694, 1966.

Patterson, E. M.,  and  D.  A. Gillette.  Measurements  of  visibility vs. mass-concentration for
     airborne soil particles.  Atmos. Environ. 11:193-196, 1977.

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.

Penndorf,  R.  Tables  of the refractive index  for standard air and the RayTeigh scattering co-
     efficient for the  spectral  region between 0.2 and  20.0 m and their application  to atmos-
     pheric optics.  J. Opt. Soc.  Am. 47:176-182, 1957. •

Peterson,  J.  T.,  and  E. C.  Flowers.   Interactions  between air pollution and solar radiation.
     Sol.  Energy 19:23-32, 1977.
SOX9/E                               9-64                                            12-20-80

-------
Pierson, W.  R.,  and P. B. Russell.  Aerosol  carbon  in  the  Denver  area  in  November  1973    ,>-
     Environ. 3:347-354, 1969.

Pierson, W.  R.,  W.  W. Brachaczek,  T.  J.  Truex,  J. W.  Butler,  and T. J.  Korniski.  Ambient
     sulfate  measurements  on  Alleghany Mountains and  the  question of atmospheric  sulfate  in
     the northeastern  United States.  Ann.  New York  Acad. Sci.  338:145-173,  1980.

Pinnick, R. G., D. E.  Carroll, and D. J.  Hofmann.  Polarized  light scattered from monodisperse
     randomly oriented nonspherical  aerosol  particles:   measurements.   Appl.  Opt.  15-384-393
     1976.                                                                           —

Pitchford,  M.  L.,  R.  E. Flocchini,  R.  G. Draftz, T. A.  Cahill,  L.  L.  Ashbaugh, R.  A. Eldred.
     Silicon in submicron particles  in  Southwest.  Atmos. Environ.,  in  press,  1980.

Pueschel,  R. , and  D.  L. Wollman.  On the nature  of  atmospheric background aerosol.   Presented
     at  the  14th  Conference  on Agricultural and Forest Meteorology,  Minneapolis,  MN,  April
     1978.

Rabinoff,  R., and  B.  Herman.   Effect of aerosol  size  distribution  on  the accuracy of the in-
     tegrating nephelometer.   J. Appl.  Meteorol.  12:184-186',  1973.

Randall, A.,  B.  Ives, C.  Eastman.   Bidding  games  for valuation of aesthethic environmental
     improvements.   J.  Environ.  Econ. Manag.  1:132-149, 1974.

Robinson, G. D.  Absorption of solar radiation by atmospheric aerosol,  as  revealed  by measure-
     ments at the  ground.  Arch. Meteorol.  Geophys.  Bioclimatol.  Ser. B 12:19, 1962.

Robinson, G. D.  Inadvertent Weather Modification  Workshop.   Final Report  to the MAS; The Center
     for the Environment and Man (CEM)  Report 4215-604, 1977.

Rodhe, H.,  C.  Persson, and 0. Akesson.  An  investigation  into regional transport  of soot and
     sulfate aerosols.  Atmos. Environ.  6:675-693, 1972.

Roessler,  D.  M.,  and  F. R. Faxvog.  Optical  properties of  agglomerated acetylene smoke  parti-
     cles at 0.5145  mm and 10.6  mm wavelengths. J. Opt.  Soc.  Am.  70:230-235, 1980.

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, DC)  208:741-744,  1980.

Rowe, R.  D., R. C. d'Arge, and D. S. Brookshire.   An experiment on the  economic value of visi-
     bility.  J.  Environ. Econ.  Manag.  7:1-19, 1980.

Samuels, H. J., S. A.  Twiss, E.  W. Wong,  H.1 Wong-Woo, P.  K. Mueller, J.  A. Maga, and 6.  S.  Taylor.
     Visibility, Light Scattering and Mass  Concentration of Particulate Matter.  Report  of the
     California Tri-City Aerosol Sampling  Project.   State  of California Air,  Resources  Board,
     Sacramento, CA, July 1973.

Schade, 0.  H.  Optical  and photoelectric  analog of the  eye.   J.  Opt. Soc.  Am.  46:721-739,  1956.

Schober, H. A. W..'and R. Hilz.  Contrast sensitivity of the  human eye  for square-wave grating.
     J. Opt. Soc.  Am.  55:1086-1091,  1965.

Spicer, C.  W., and P.  M. Schumacher.  Interferences  in  sampling atmospheric  particulate  nitrate.
'  «   Aim. Environ. 11:873-876, 1977.

Spicer,  C.  W.,  and  P.  M.  Schumacher.   Particulate  nitrate:   laboratory and  field studies  of
     major sampling  interferences.    Atm.  Environ.  13:543-552,  1979.



&X9/E '-SJ                          9-65                                           12-20-80

-------
Stevens,  R.  K., T.  G.  Dzubay, G.  Russworm,  and D. Rickel.   Sampling  and analysis of atmos-
     pheric sulfates and related species.  Atm. Environ. 12:55-68, 1978.

Stevens,  R.  K., T.  G.  Dzubay, R.  W.  Shaw,  W.  A.  McClenny,  C. W.  Lewis,  and W. E. Wilson.
     Characterization of  aerosol  in  the Great Smoky Mountains.   Accepted for publication  in
     Environ. Sci. Tech., 1980.

Stuiverts, M.   Atmospheric  carbon dioxide and carbon reservoir  changes.   Science  (Washington,
     DC) 199:253-258, 1980.

Tang,  I.  N.,  H.  R.  Munkelwitz, and J.  G.  Davis.   Aerosol growth  studies,  IV.   Phase trans-
     formation of mixed salt aerosols in a moist atmosphere.  J. Aerosol Sci. 9:505-511,  1978.

Tanner, R. L.,  R.  Garber,  W. Marlow, B.  P.  Leaderer.and M. A. Leyko.  Chemical composition  of
     sulfate as a  function  of particle  size  in  New York summer aerosol.  Ann. New York  Acad.
     Sci.  322:99-113, 1979.

Tombach,  I.  H., and M.  W.   Chan.   Physical,  chemical,  and radiological  characterization  of
     background particulate  matter  in  northeastern Utah.   Presented at the Annual Meeting  of
     the Air Pollution Control Association,  1977.  Paper No. 77-48.6.

Tombach, I.,  and D. Allard.   Intercomparison of visibility measurement methods.  J. Air Pollut.
     Control  Assoc. 30:134-142, 1980.

Trijonis,  J., and R.  Shapland.  Existing Visibility Levels  in the U.S.:  Isopleth  Maps of Visi-
     bility in  Suburban/Nonurban  Areas  During 1974-1976.   EPA450/5-79-010, U.S. Environmental
     Protection Agency,  Research Triangle Park, NC, 1979.

Trijonis,  J.,  K.  Yuan,  and  R. B.  Husar.  Visibility in the Southwest:  An Exploration of the
     Historical Data Base.   EPA-600/3/78/039, U.S.  Environmental  Protection Agency, Research
     Triangle Park, NC,  April 1978a.

Trijonis,  J.,  K.  Yuan,  and  R. B. Husar.  Visibility in the Northeast:   Long Term Visibility
     Trends  and  Visibility/Pollutant Relationships.    EPA-600/3-78-075,  U.S.  Environmental
     Protection Agency,  Research Triangle Park, NC, August  1978b.

Truex, T.  J., and J.  E.  Anderson.   Mass monitoring of carbonaceous aerosols with a spectrophone.
     Atm.  Environ. 13:507-509, 1979.

Twomey, S.  Pollution and the planetary albedo.  Atmos. Environ. 8:1251, 1974.

U.S. Bureau of Mines.  Minerals Yearbook, Annual Publication, 1933-74.   U.S. Department of the
     Interior, Washington, D.C.

U.S. Department of Energy.   Carbon Dioxide Effects, Research, and Assessment Program.  A com-
     prehensive plan for carbon dioxide effects, research,  and assessment, Part I:  The Global
     Carbon Cycle and Climatic Effects of Increasing Carbon Dioxide.  DOE/EV-0094, U.S. Depart-
     ment of Energy, Washington, DC, August 1980.

U.S. Department of Energy.  Office of Carbon Dioxide Effects,  Research,  and Assessment.!  A(Com-
     prehensive Plan for Carbon Dioxide  Effects, Research,  and Assessment.  U.S. Department  of
     Energy, Washington, DC, 1978.

U.S. Environmental Protection Agency.  EPA Report to Congress:   Protecting Visibility.  EPA-450/
     5-79-008, U. S. Environmental Protection Agency, Research Triangle  Park, NC,  1979.

Van Nes, F. L., and  M. A. Bouman.   Spatial modulation transfer in the human eye.   J.  Opt. Soc.
     Am. 57:401-406, 1967.


SOX9/E                               9-66                                           12-20-80

-------
Waggoner,  A.  P., A.  J.  Vanderpol,  R.  J.  Charlson,  S.  Larsen,  L.  Granat,  and  C.  Tragardh.
     Sulphate-light  scattering  ratio  as  an  index  of  the  role  of sulphur  in troposphen>
     optics.  Nature (London) 261:120-122,  1976.

Waggoner,  A.  P.,  and R.  E. Weiss.   Comparisons of fine  particle  mass concentration and light
     scattering extinction in ambient aerosol.  Atmos.  Environ. 14:623-626,  1980.

Waggoner,  A.  P., and  R.  J.  Charlson.   Measurement  of aerosol  optical  parameters.  _In:  Fine
     Particles, B. Y. H. Liu, ed., Academic Press, New  York, 1976.

Waggoner,  A.  P., et al.  Optical  absorption by atmospheric aerosols.  Appl.  Opt.  12:896, 1973.

Waggoner,  1980.

Waggoner,  A.  P., R. E. Weiss, N.  C.  Ahlquist,  D.  S. Covert,  and R. J. Charlson.   Optical Charac-
     teristics  of Atmospheric  Aerosols.    Presented  at:   Symposium on  Plumes and Visibility
     Measurements and Model Components, Grand  Canyon, Arizona, November  10-14,  1980.

Weiss, R.  E.  Ph. D. Dissertation, 1978.
                                                          t

Weiss, R.  E.  et al.  Studies  of the  optical, physical,  and chemical  properties of light-absorb-
     ing   aerosols.   In:    Proceedings  of  the Conference  on  Carbonaceous  Particles  in  the
     Atmosphere, 1978.  p. 257.

Weiss, R.  E., A.  P. Waggoner,  R.  J. Charlson, and N. C.  Ahlquist.   Sulfate aerosol:   its geo-
     graphical extent in the  midwestern and southern  United  States.   Science 195:979-980, 1977.

Wesely, M.  L., and R. C. Lipschutz.  An experimental  study of  the  effects  of aerosols  on diffuse
     and  direct  solar  radiation  received  during the  summer near  Chicago.  Atmos.  Environ.
     10:981-987, 1976.

White,  W.   H.,  and  D.  E.  Patterson.    Nitrogen  dioxide, particles,  and  the color of smoke
     plumes.  Presented at Symposium on Plumes and Visibility, Grand Canyon, AZ,  November 1980.

White, W.  H., and P. T. Roberts.   On the  nature and origins  of visibility  reducing aerosols in
     the Los  Angeles Air Basin.   Atmos. Environ.  11:803-812, 1977.

Wiscombe,  W.  J.   Improved Mie scattering  algorithms.  App. Opt. 19:1505-1509,  1980.

WMO/EPA/NOAA/UNEP.  Global Monitoring of  the Environment for Selected Atmospheric Constituents
     1975.   Environmental Data  Service National Climatic Center, Asheville, NC,  1977.

Wolff, G.  T., P. J. Groblicki,  S.  H. Cadle, and R. J. Countess.  Particulate carbon at various
     locations in the United  States.  Presented at the  General Motors Symposium on Particulate
     Carbon,  Warren, MI, October 1980.
SOX9/E                                9-67                                           12-20-80

-------
                           10.  EFFECTS ON MATERIALS

10.1  INTRODUCTION
     Beginning with associative observations in the 17th century and continuing through modern
analytical investigation, scientists have gathered evidence that air pollutants damage man-made
and natural  materials.   Pollutant-related  damage may  lead  to the need  for  measures  such as
increased maintenance, use of protective coatings, replacement of materials, or the control of
emissions.   The  decision to  pursue  one  course rather than another is  guided,  ultimately, by
weighing  the damage against  the  costs  of prevention.  Neither damage  nor  prevention  cost is
necessarily  measured  in monetary  units, especially  with  regard to works  of art  or items of
historical significance.  However, for better or worse, society tends to assign monetary units
to value gained or lost.  Thus, the objective of this report is to present and evaluate studies
useful for estimating the dollar cost of material damage caused by sulfur oxides and particulate
matter.
     The  relation  between  pollutant emissions and economic damage is depicted in Figure 10-1.
As shown  in  the  figure, one  may  (1) proceed from ambient pollutant levels to economic damage
estimates directly or (2) follow a route involving estimate of damage based on physical damage
functions.   The  latter  route,  which we  all call the damage  function  approach,  has been the
method of choice  in  the past.   More  recent studies have employed the first route.   The two
alternative  routes have an element  in  common.   This element  is  estimation of willingness to
pay-
     Economic damage  (benefit)  as a result  of  increased (decreased)  pollutant concentrations
can be estimated by willingness to pay  approaches.  All willingness to pay approaches attempt
to estimate  the  aggregate monetary  values that all affected individuals assign to the effects
of  a  change in  pollutant concentration.   Willingness  to pay approaches can be divided  into
three  classes:   damage  function   approaches,  non-market  approaches,  and  indirect  market
approaches.   Damage  function approaches  utilize the  relationship  of  pollutant  exposure to
physical   damage  as  a  first  step.  The second  step  links the  physical  damage to  a dollar
estimate  of  willingness to pay.  With proper consideration of substitution possibilities  this
damage function  approach can  be a good  estimate of willingness to pay.  Non-market approaches
generally  utilize  surveys  which  attempt to  ascertain the monetary  values assigned  to the
effects in question by all individuals.  Indirect market approaches use information about the
demand for marketed  goods  to estimate the  willingness  to  pay for non-marketed environmental
attributes which  are  closely  related to the marketed good (e.g., property value studies which
estimate  the willingness to pay for a  change  in the level of pollutant concentration through
the analysis of  the price of residential property) (Freeman, 1979).  Each of the three groups
of approaches  have different  data requirements.  The physical damage function approaches  have
been the  most  widely used in the past and will therefore receive the most extensive treatment
in  this  chapter.  Non-market and  indirect market studies will be discussed  later in Section
10.5.

SOX10A/A                                   10-1                                        12-15-80

-------
                                                  LEGISLATION
                                                  EXECUTIVE
                                                  ACTS
     CONTROL
     MEASURES
                                POLICY
                                DECISIONS
POLLUTANT
EMISSIONS
       PHYSICAL
       AND
       CHEMICAL
       INTER-
       ACTIONS
                            COSTS OF
                            CONTROL
                                                         ECONOMIC
                                                         DAMAGE
                                                         FUNCTIONS
PHYSICAL
DAMAGE
FUNCTIONS
PHYSICAL
DAMAGE
 AMBIENT
 LEVELS
                                                                       PROCESSES AND ACTIONS
                                                                       RESULTANT PRODUCTS
Figure 10-1. Relationship among emissions, air quality, damages and benefits, and policy decisions.
Source: Hershaft (1976)
                                                                       •

                                               10-2

-------
     In  the  damage function  approach,  before economic  damage is  estimated,  physical  damage
must  be  determined.    This  determination  begins  by  calculating  the  exposure  of  certain
materials in  specific  locations.   Exposure is estimated from ambient air levels over specific
time intervals.   Depending  on the pollutants and the  kinds  of material  studied, both average
short-term  and  long-term  exposure  data  may  be needed.  These  data are  used to  develop  a
physical damage  function, which is a mathematical expression linking exposure to damage.   The
damage  function  is expressed in terms appropriate  to the  interaction  of the  pollutant and
material.  For example,  the corrosion of metal might be expressed in units of thickness lost,
while  the  deterioration of paint could be  stated in units of  reflectance  or thickness lost.
The pollutant  level may be  expressed in any concentration unit.  Since material damage usually
develops over  time,  average long-term pollutant  levels  are  often used in calculating damage.
Damage studies often encompass  a yearly time scale, which facilitates economic comparisons and
allows averaging of conditions  in changeable climates.
     A  major  problem  in   establishing  reliable dama'ge  function  has been  to  separate the
influence  of  the  target   pollutant from  that  of  meteorological  parameters  and  other air
pollutants.   For the  corrosion of metals,  relative humidity  is  the most important variable.
Rainfall,  time  of  wetness,  sunlight,  and wind speed and  direction are available for  many
locations  and,  if  relevant,  can be  factored  into  damage  calculations.   The  influence of
atmospheric components should be considered in the context of the material being studied.   For
example, in  studies of SO  -related  steel corrosion,  even low atmospheric chloride  levels may
have a significant impact on  the rate of damage.
     Before  an  economic  value  is placed  on  the incremental  damage caused by  pollution,  a
critical  damage  level  must  be determined.   This   level  represents  the point  at  which the
service  life or functional  utility  of  the  material  has ended  or  is severely impaired.  When
this point is  reached, replacement or repair is  necessary, and cost  is incurred.  For example,
if a  typical  coat of  paint is  60 urn thick, the  critical  damage  level at which  repainting is
necessary may  be  around 50  urn.  By comparing the amount of surface erosion in a clean environ-
ment with  that in an  area  with a specific pollutant, the apparent damage  from the pollutant
can be calculated and used in  assessing physical damage.  The valuation  is determined through
economic damage  functions, (1) which may  be developed  from physical damage  functions; (2)
which  may  relate economic  damage directly to pollutant level by comparing the life-time costs
of a  given material in different ambient environments; (3) which may combine both approaches.
In the  first case, physical damage functions are coupled to the use  and value of the material.
Included in the  estimate are  exposure, replacement, protection, and  other data.  None of these
approaches  account for irreplaceable items,  such as works of  art,  where the only  measurable
cost   is that  of  preservation.   The  social  losses  of  artifacts  cannot  be   quantified
economically.
     As  discussed above in the context of  Figure 10-1, most  estimates  of  economic daaage to
materials  have  followed  the  damage function  approach.   This  approach  is discussed  here to
illustrate  how damage functions may be used as an integral component of cost estimation.   The

SOX10A/A                                  10-3                                        12-15-80

-------
chapter  is  organized in several parts.   Presented first are laboratory studies.   These  con-
trolled  experiments  investigate  variables  that  influence damage  to  a  material  at  various
pollutant concentrations.  The  results  of these studies are used to develop damage functions,
which include only those variables that affect material  damage.
     Field experiments are performed  to test the validity of extrapolating laboratory results
to ambient conditions.   Parameters  measured in field experiments are those found or suspected
to be significant  factors  in laboratory experiments.    For example, if relative humidity were
part of the damage function derived from laboratory results, it would be carefully measured in
the field experiment.
     Results  from   field  experiments   are  then  compared  with  laboratory-derived  damage
functions.  The comparisons may show similarities or differences.  A field  study may measure a
parameter not measured  in  a  controlled experiment but which  correlates well with damage.  In
such a  case, a  laboratory  experiment may  be designed  to  determine whether or  not  a causal
relationship between parameter  and  damage exists.   When a  damage function has  been validated
by field  and laboratory  experimental  results, it may be used as a basis for determining total
damage to materials exposed to an air pollutant.
     In the  damage function  approach  to cost estimation described  above,  the physical damage
function  is  one of several components  of an  economic damage function.  Other  necessary  com-
ponents  to   this  approach  include  a mathematical  expression  of  surface  area  of materials
exposed and cost factors associated with units of physical  damage.  Problems arising from this
approach have been (1) valid physical  damage functions do not exist for all pollutants and all
materials; (2) estimates of the amount and type of material exposed have usually been based on
some surrogate  such  as  production  figures  modified  by service  life  data  (field  surveys of
exposed material in place have not been performed); (3)  cost factors may or may  not accurately
reflect cost associated  with  pollution  as opposed to other causes.   This latter consideration
is  especially a  problem in  estimates  of  soiling  damage, where  socioeconomic  factors are
heavily involved.
     Past  estimates  of  costs  associated with  particulate matter,  sulfur oxides,  or other
pollutants must  be considered  in  light  of  the above  discussion.   The last section of this
chapter discusses cost estimates and their limitations for decision-making purposes.
10.2  SULFUR OXIDES
10.2.1  Corrosion of Exposed Metals
     Sulfur  oxides  in  the environment  accelerate the  corrosion of metals. Several  factors
other than concentration of S02 are important.  These are discussed first.                   ;
10.2.1.1  Physical and Chemical Considerations—The atmospheric  corrosion  of most metals is a
diffusion-controlled  electrochemical  process.   For electrochemical action  to take place, the
following are  necessary:   (1) an electromotive force between points on the metal surface; (2>
a mechanism  for  charge transfer between  the  electronic  conductors; and (3) a conduction path
between  the  cathode  and anode  reaction  centers.   Measurements  of the rate of SOg-accelerated
rusting of iron vary greatly  from site to site, despite careful monitoring of pollutant

SOX10A/A                                   10-4                                        12-15-80

-------
concentrations,  a  fact  that  has  often  puzzled  researchers.    Several   factors  might  oe
responsible  for  inconsistent results.   These include:   (1)  the  deposition  rate of gaseous or
dissolved  S02 and  particles;  (2) the  variability in the  electrochemical  actions  that cause
corrosion;  (3)  the  influence  of  rust on  the  rate of  subsequent corrosion;  and  (4;  t-.e
interaction  between  the  pollutant  effects  and  "wetness  time"  often  indicated by relati/e
humidity (RH) on surface electrolyte concentrations.
     According  to  Nriagu (1978),  once corrosion  has begun, the progress  of  the reaction is
largely  controlled by the sulfate  ions  formed  by oxidation of the  adsorbed SO-.   The actual
mechanism  for the  oxidation  of  S02 (and its hydrated products)  at  metal -water interfaces is
little understood.  Barton (1973) proposed the following reaction:
or
                             S02 + 02 + 2e
                                 302 + 4e
The electron is provided by the oxidation of the metal (M):
                                 M -» Mn+ + ne
Duncan and  Spedding (1974),  using an  electrophoretic  method,  found that the rates of sulfate
formation  on  iron  and zinc  surfaces  were  similar;  the  pseudo-first-order  half-life  was
determined  to be  about 24  hours.   Other workers  (Karraker  1963;  Yoshihara  et  al.,  1964)
reported  higher  oxidation rates (half-life,  10 to 100  minutes)  in  bulk solutions  using Fe3+
catalysts as  summarized in Nriagu (1978).  Nriagu also  noted that rust on  iron and steel  is
first  restricted  to  localized  sites  or  "nests"  and  then  spreads across the  entire exposed
surface.   At  an  SO, concentration of  about 260 ug/m , corrosion products were obvious on iron
                                                                       53
surfaces  after 6 to 8 weeks,  whereas at SOp  concentrations  of 4 x 10  ug/m  ,  they could  be
seen after only a few  hours.
                                                                                    2
     Barton (1976)  showed  that the critical SO, flux  for corrosion was 6 to 10 ug/m  per year
                            2
for steel and  18  to 20 pg/m  per year for zinc and copper.  The formation of rust drastically
increases the adsorption rate for S02.  For example, at  an S02 concentration of >0.001 percent
and relative  humidity (RH) of >96 percent, virtually  all  of the  S09  that comes into contact
                                       i                             f-
with the rusting surface is adsorbed or consumed in reaction.
10.2.1.1.1  Relative Humidity  and  Corrosion Rate.   According to Schwarz (1972), the corrosion
rate of  a metal  should increase by  20 percent for each increase of 1 percent in the relative
humidity (RH) above  the critical RH value.  It is evident that RH has a considerable influence
on the corrosion rate.  This influence has been established in laboratory trials by Barton and
Bartonova (1969) and Sydberger and Ericsson (1976).   It is  apparent from Figure 10-2 (Haynie
and Upham, 1974) that the corrosion rate of steel  increases with increasing RH as well as with
increasing S02 concentration.
     The  climate  of  an  area  is  usually  characterized  by  average  RH  rather  than  RH
distribution.    Since  average  RH  is  calculated from  the  distribution, there  should  be  an
empirical relationship between average RH  and the  fraction of  time  some  "critical huaidity

SOX10A/A                                   10-5                                        1^15-80

-------
I
100


 90



 80



 70
 E

 uT  60



 2  50
 g
 CO
 S  40
 ec


 6  30
 o

 t/i  20
10



 0
            I
                          SOj CONCENTRATION. jUfl/m'
              _L
J_
I
I
       0   10    20   30    40    50    60   70    80   90   100


                  AVERAGE RELATIVE HUMIDITY, %


  Figure 10-2. Steel corrosion behavior is shown as a function of aver-
  age relative  humidity at three average concentration levels of sulfur
  dioxide.


  Source: Haynie and Upham (1974).
                              10-6

-------
value"  (minimum  concentration of water  vapor required for corrosion to  proceed)  is  exceeded
(Mansfeld and  Kenkel,  1976;  Sereda, 1974).  The fraction of time that the surface is  wet must
be  zero when  the  average  RH  is  zero  and  unity when  the average  RH  is 100  percent.   The
simplest single-constant  first-order  curve that can be fitted to observed data and that meets
these conditions  is described  by  the equation  f =  [(1  -  k)RH]/(100 -  kRH), where  f  is the
fraction  of time  the  RH exceeds  the  critical  value,  RH is  the average  RH,  and  k  is  an
empirical constant  less than unity (Haynie, 1980).
     Ten quarter-year  periods of RH  data from St. Louis  International  Airport  were  analyzed
and  fitted  by the  least-squares method  to  the  above equation.   The fraction of  time  the  RH
exceeded 90 percent gave  a value of 0.86  for  k.  This fraction and the data points are plotted
in Figure 10-3 (Haynie, 1980).
     When the  temperature of a metal  is  below the ambient dewpoint, condensation of water on
the  metal  surface will take  place.   The metal temperature at which condensation  occurs will
vary with heat transfer between ground and metal and between air and metal.  Condensation will
occur when  the  RH  adjacent to the  surface  exceeds  a value in  equilibrium with the vapor
pressure of a saturated solution on the  surface.  The solution may contain corrosion products,
other hygroscopic  contaminants,  or  both.  Temperature, wind, and sunshine then become factors
in  establishing  corrosion rates,  since   they  determine whether there will  be sufficient dew
condensation.
     Haynie (1980)  reported  on  the relationship  between  diffusion theory and thermodynamics
for  the  observed  effects  of five variables:   pollution level, RH, temperature, wind velocity,
and  surface geometry.   He  observed  that  metals must  be  wet  to  corrode electrochemically.
Surfaces are wet from condensation much more  often than from precipitation.
10.2.1.1.2   Influence of Rainfall  on Corrosion.   Steel  surfaces shielded  from the  leaching
effect of rain may corrode at a higher  rate than those exposed to  rain.  The sulfate content
of  rust  has been identified as a dominant factor in corrosion and  is found at higher concen-
trations on surfaces sheltered from rain  than on exposed surfaces  because soluble sulfate is
leached  from  the  rust.   However,  sulfur deposition during rainfall  must also be considered.
Haagenrud and Ottar (1975)  noted  that  the  rate  of corrosion of unalloyed steel  and zinc
correlated with the amount of sulfur (S02 and  sulfate) in air and in precipitation.
     As  Kucera's  (1976) review of this  problem  indicates,  the mode of deposition complicates
the  analysis of acidic precipitation's effects.  For example, in an  area where dry deposits of
hydrogen and sulfate ions exceed deposits in wet precipitation, flat steel plates corrode more
rapidly on their undersides than on their upper surfaces, suggesting that rainfall has more of
a  washing  effect  than  a  corrosive  action.   However,   in  other  areas where  wet and dry
deposition were  about  equal, the upper  sides  of  the  plates corroded more quickly, suggesting
that the corrosive effect of the rainfall predominated.  Other variables, including amount and
frequency of precipitation, its pH level, humidity, and temperature, also determine the iapact
of  acidic  precipitation  (Kucera, 1976).   For a  more thorough  discussion of the  effects  of
acidic precipitation on corrosion,  see Chapter 7.

SOX10A/A                                   10-7                                        12-15-80

-------
£
3.

Ill"
C

o
55
O
DC
C


1
Q
D
100



 90



 80



 70



 60



 50



 40



 30



 20



 10
                    I
                             I
I
       0           100           200           300          400


           AVERAGE SULFUR DIOXIDE CONCENTRATION,/u 9/m3



 Figure 10-3. Steel corrosion behavior is shown as a function of aver-
 age sulfur dioxide concentration and average relative humidity (RH).


 Source: Haynie and Upham (1974).
                            10-8

-------
10.2.1.1.3  Influence of Temperature on Corrosion.   From chemical  reaction kinetics,  one might
expect corrosion rates to follow the Arrhenius law according to which the logarithm of a reac-
tion  rate is  inversely proportional  to  the absolute temperature.   However,  this  does  not
appear to  be  a major factor controlling  corrosion  rates.   In most cases, as discussed later,
the reaction  rate  is controlled by diffusion, either  in  the corrosion product film or in the
environment.    In  either case,  the  rate  of diffusion is relatively  insensitive  to changes in
ambient  temperature.   A decrease  in  temperature raises the  RH while decreasing diffusivity;
thus,  the  normal   temperature  range  will  most  likely   not observably  affect  the overall
corrosion  rate.  Even at -20°C  metal corrosion is not halted  (Sereda, 1974).
     Guttman  (1968) and  Haynie and  Upham (1974),  using  statistical techniques  of multiple
linear  regression  and nonlinear curve  fitting,  did not obtain from  their data  a significant
correlation  between corrosion  and temperature.  However,  increased  temperature  decreases the
solubility of  oxygen   in  the  electrolyte.   Moreover,  temperature  decreases  can  cause
condensation,  the  increased quantity of  condensed droplets can affect the corrosion  rate, and
the rate at which a  surface dries depends on  both temperature and air movement.
     Guttman and Sereda (1968)  made continuous measurements of sulfur oxides, time of wetness,
and temperature in their outdoor exposure tests.   The corrosion rate increased markedly with
temperature,  which may be  attributed to a  speeding  up of the  electrochemical  process.  The
effects  are  not  simply  thermodynamic  since other factors  involved include  evaporation of
surface  electrolyte and decrease in the  solubility of oxygen and gaseous pollutants.  Barton
(1976)  did find that  the  effect  of increased temperature was more pronounced  when the  rust
contained  little   water  and   sulfate.   Haynie  et al.   (1976)  found  temperature  to  be   a
significant variable in chamber studies  of weathering  steel.
10.2.1.1.4   Hygroscopicity of Metal  Sulfates.  The   sulfate  in   rust  stimulates   further
corrosion  by  a mechanism  that is  related  to  the  critical relative  humidity  at which an
electrolyte  film  is formed.   The hygroscopicity  of  iron  sulfates  in the  rust lowers the
critical   RH  for  corrosion;   however,  sulfates  are  not  the most  deliquescent  salts.   For
example,   chloride  and  nitrate salts  which  have  higher  hygroscopicity than  sulfates  make
corrosion  possible  at lower humidities.
      Surfaces  contaminated  by  sea salt  (mostly sodium  chloride) can  be expected  to be wet  when
the  RH exceeds 75  percent.   In contrast, calcium chloride keeps surfaces wet at  an  RH as low
as  30 percent.  A  saturated  solution  of zinc sulfate at  20°C is in equilibrium at  90  percent
RH.   Thus,  zinc corroded  by  sulfur dioxide is expected  to be wet when the  RH exceeds 90
percent (Haynie, 1980).
      Some  pollutants that react chemically with various materials and with other materials are
not  consumed  and  produce  catalytic effects.   The following four  mechanisms act to increase
corrosion rates without  consuming  the  pollutants:   (1)   hygroscopic materials increase the
amount of time that a corroding surface  is wet;  (2) conductivities  of solutions and corrosion
product  films  increase;   (3)   pollutants  form  soluble  intermediate  reaction  products or
activated complexes  that  destroy  the protective  nature  of  films;  and  (4)  lowering  the pH
usually increases  the solubility of corrosion products  (Haynie, 1980).
SOX10A/A                                   1°-9

-------
100


 90


 80


 70


 60


 50
u.
O   40
u.
O
ui
5
            I     I
                            I     1     I      I     I     I
t-   30


    .20


    10


     0
      0    10   20    30   40   50    60    70   80    90  100

                  AVERAGE RELATIVE HUMIDITY, %


 Figure 10-4. Empirical relationship between average relative humid-
 ity and fraction of time relative humidity exceeded 90 percent is
 shown for data from St. Louis International Airport

 Source: Haynie (1980).
                         10-10

-------
10.2.1.1.5   Electronic Conductivity of Rust.   Barton  (1976)  postulated  that  sulfate  ions
influence  the  anodic dissolution  of  iron  as  a  function  of  their  concentration  at  the
steel-rust  interface.   The corrosion  rate of  the  rust layer  is  based in  part on the  high
electronic conductivity of rust, which allows the reduction of oxygen to occur within the  rust
layer.  The  rate  is also influenced by the porosity of rust, which permits rapid diffusion of
oxygen to the cathode.
     In the presence of SO-,  ferrous  sulfate is formed before  insoluble  rust develops.   The
amount of  S02 required is small; each  SO- molecule can generate 20  to  30 molecules of rust.
Once  ferrous  sulfate is  formed, rusting can  continue  even though SO- is no longer present in
gaseous form.
10.2.1.1.6   Cathodic Reduction of Rust.  Evans  (1972)  suggests that oxidative  hydrolysis of
ferrous sulfate occurs slowly, and would  be  important only in the initial stage of corrosion.
He proposes that there is a  rate-controlling cathodic process.  Thus, the corrosion products
in the ferric state  would be  converted  to  magnetite (Fe^) by  a reaction  involving the reduc-
tion of ferric  oxyhydroxide (FeOOH):
                    Fe2"1" + SFeOOH + 2e~  -» 3Fe304 + 4H20
10.2.1.1.7   Corrosion-Protective Properties of Sulfate in Rust.  The  rust layer on  steel is
somewhat  protective against further corrosion,  though  far less  so than the corrosion layer on
zinc  and  copper.   A  limiting  factor  in rust's protection of steel  is  the content of soluble
compounds  in  rust.
      Rust   samples   investigated  by  Chandler  and  Kilcullen  (1968)  and  by  Stanners  (1970)
contained  2 to 2.5  percent soluble  sulfate and 3 to  6 percent total sulfate.   The outer  rust
layer  contained  a  small  amount  (0.04 to 0.2 percent) of  soluble  sulfate,  compared  with 2
percent  in the inner rust layer.   The concentration  of  insoluble  sulfate was  fairly uniform
throughout the  rust layers.
      The   emphasis  on  the   composition   of   the  rust  layer   has   led   to   studies  of  the
corrosion-protective properties  of rust  as  a function  of exposure  history  (Nriagu, 1978;
Sydberger,  1976).   Steel samples initially exposed to low concentrations  of  sulfur oxides and
then  moved to  sites  of  higher sulfur oxide concentrations corroded  at a  slower rate than did
samples  continuously exposed to the  higher concentrations.   Exposure tests started in summer
showed  slower  corrosion  rates during  the first years of exposure  than did those started in
wi nter.
  •    The- long-term corrosion  rate of  steel appears  to  depend on changes  in the composition and
structure of the  rust layer.   During the  initiation  period, which  varies in length with the
S02  concentration  and other accelerating  factors, the rate of corrosion increases with  time
(Barton,  1976).   Because  it  is  porous and  nonadherent,  the rust initially formed offers no
protection; in fact,  it may  accelerate corrosion by retaining  hygroscopic  sulfates and chlo-
rides,  thus producing a microenvironment  with a high  moisture  content (most often  reported in
terms of percent RH). After  the  initiation stage,  the corrosion rate decreases as  the protec-
tive  properties  of the rust  layer improve.    Satake  and Moroishi  (1974) relate this  slowing

SOX10A/A                                   10-11

-------
down  to  a decrease  in the  porosity  of the  rust layer.  During  a third and  final  stage,  a
constant  rate  of  corrosion is attained and the  amount of sulfate in rust is  proportional  to
atmospheric sulfur oxide concentrations.
     Where there  is  no appreciable deposition of sulfur or  chloride compounds, the corrosion
of steel is low, even with high atmospheric humidity and temperature (Haynie  and Upham, 1974).
The  quantitative  determination  and  subsequent  interpretation  of corrosion  rates  becomes
difficult  if   it  is  not  known how long  the  metal  has  had a  surface  layer of  electrolyte.
Variations in  the  "wet states" will occur with  RH,  temperature,  rain,  dew,  fog,  evaporation,
and wind.   The surface electrolyte layer may form on a metal  surface as  a result of rain, dew,
or adsorption  of  water from the atmosphere.   Capillary condensation in  rust  can be related to
the minimum atmospheric moisture  content that allows  corrosion to  occur (i.e., critical RH).
Centers of capillary condensation of moisture  on metals can occur in cracks,  on dust particles
on the metal surface, and in the pores of the  rust (Tomashov 1966).
10.2.1.2   Effects of Sulfur Oxide Concentrations on the Corrosion of Exposed Metals--Most  of
the  laboratory studies  reviewed   in  this section  have measured  corrosion  rates  related  to
exposure  to  sulfur dioxide alone  or  in combination with other compounds.   In  field exposure
studies,  where  sulfur oxides  almost  invariably  occur  in  combination  with   other  airborne
pollutants, an attempt is  made  to assign  separate values  to  sulfur oxides and  to describe
pollutant interactive effects on corrosion.  The discussion here unavoidably  overlaps somewhat
with a later  section on  particles, since sulfur dioxide contributes to  formation of secondary
sulfates.   Here, the emphasis  is  on the direct role of sulfur oxides in the  corrosion process
(e.g., the oxidation of SOp with moisture on a metal surface).   In the later  section, sulfates
are discussed  mainly in  terms  of their indirect role (e.g.,  their ability to increase wetness
time of a metal surface).
10.2.1.2.1  Ferrous  Metals.   Ferrous  metal  products  and structures  are exposed widely  to
ambient pollutant  levels.   Rusting of  these  metals  is  the best documented form  of metallic
corrosion  affected by sulfur  oxides.   This subsection  reviews studies  of  rusting  rates  of
ferrous metals, such as iron, steel, and steel  alloys.
     A number of investigators reported data during the period 1959-1968 which showed that the
addition  of  0.05  to  0.5  percent copper to  steel  results  in  improved  corrosion resistance
(Larrabee, 1959;  Larrabee and Coburn,  1961; Brouns  and  Kail a,  1965; and Schwenk  and Ternes,
1968).  In  an extensive  series  of exposure  tests  (Guttman and  Sereda, 1968), the American
Society for  Testing  and  Materials (ASTM) corroborated  previous  investigations, showing that
adding copper to steel  improves its resistance to corrosion.
     Stainless steels  contain more than  12 percent  chromium and are widely-used in outdoor
exposures; they are specified for use in many  industrial  processes involving  corrosive liquids
that rapidly attack  ordinary steels.   The high corrosion  resistance of stainless  steels that
incorporate chromium, molybdenum,  and nickel  is attributed to the protective  properties of the
oxide film formed on these alloys.   In heavily polluted atmospheres, however, this  film is not
completely  protective.   Particles  in  settled  dust,  including sulfates  and   chlorides,  can
                                                                                               \
SOX10A/A                                   10-12                                       12-15-80\

-------
promote rupture  of  the oxide film and cause pitting corrosion, which may be influenced by the
surface finish  (see section 10.3.1).   Smoothly polished or  electropolished surfaces are less
likely to retain solid deposits originating from airborne particles (Larrabee, 1959).
     The lowest alloyed stainless steels have little corrosion resistance.   In particular, #13
Cr  steel  suffers pitting  attack  in industrial  atmospheres.    Evgang and Rockel  (1975) report
that the austenitic steels of 18 percent Cr and 8 percent Ni  are reasonably resistant in urban
atmospheres but  have  shown slight rusting in industrial areas.  The rusting rate is decreased
when the steel surface is cleaned of atmospheric deposits.
10.2.1.2.2  Laboratory and Field Studies Emphasizing Ferrous Metals.  It is useful to consider
laboratory and  field  studies of corrosion effects separately because cause and effect is much
clearer  in  laboratory experiments;  field studies are  often  beset  by confounding variables.
Laboratory studies  develop  from  controlled experiments,  including as many  variables as are
thought  likely  to  influence damage to the material  being studied.  Data  on  materials damage
result from exposure  to various concentrations of the  air pollutant being studied.  Analysis
of  the  data  resulting from  such  laboratory  studies is  used  to  develop a  mathematical
expression  of  the  relationship  between  the  concentration  of  a pollutant  and  damage  to
materials.   Such   a  mathematical  expression  is  called  a  damage  function,  that  is,  the
quantitative expression of a relationship between exposure to  specific pollutants and the type
and  extent of  damage  to  a  target population.   Factors  that are  shown to be significant
influences in producing damage are included in the damage function.
     Information  derived  from laboratory  studies  is used to  design field experiments, which
are  performed  to  test  the  possibility  of  extrapolating  laboratory  results to  ambient
conditions.  The parameters that are measured in field studies are those found or suspected to
be  important  factors  in the laboratory experiments.  The results of laboratory experiments do
not  easily translate  to  field situations,  however, since  ambient air pollution  levels and
other  influencing  environmental   factors  vary  widely  both  in  time  and  space.   Temporal
variables  include fluctuations in  temperature,  wind moisture content,  insulation, rainfall,
and   its   chemical   characteristics   (e.g.,   acidic   rain),   and  atmospheric   pollutant
concentrations.   Spatial factors include differences such as aspect, altitude, electromagnetic
fields, and indigenous microorganisms.  Initial  conditions of the material being studied must
also be considered.
     The  results of  field  studies  are compared with the  laboratory-derived  damage function.
In some cases the results are comparable, and the laboratory damage function is validated.  In
other cases the data analysis may result in a markedly different damage function, with more or
fewer variables.
     10.2.1.2.2.1   Laboratory Studies.  Spence  and  Haynie (1974)  described  the design  of a
laboratory experiment to identify the  effects of environmental pollutants on various naterials
including  ferrous metals.   The environmental  system consisted of  five exposure chambers to
control  temperature and humidity and  chill  racks  to simulate  the  formation  of  dew.  Gaseous
pollutants  included  those  usually monitored  in field  exposures:  sulfur dioxide, nitrogen
SOX10A/A
                                           10-13                                       12-15-80

-------
dioxide, and  ozone.   Experiments  were statistically designed for analysis  of variance, and a
system was  selected  to  study the interactive effects of  pollutants  and other variables.   The
effect of particles  was  not included in the  design.  The chambers  were equipped with a xenon
arc  light  to simulate  sunlight.   The system was designed to  maintain air  contact  with the
various materials at preselected  temperatures,  relative humidities,  flow rates,  and pollutant
concentrations.   A dew-light cycle  was  used;  it produced faster deterioration than did condi-
tions of constant humidity and temperature.
     Haynie et  al.  (1976) exposed weathering steel  in  the chamber study described  above and
measured concentrations  of sulfur dioxide,  nitrogen dioxide,  and ozone in various combinations
and  at  two  levels  of pollutant concentration.  Ozone was of interest since  an  earlier field
experiment  (see following  Field-Studies  section)  had indicated that  the  presence of oxidants
inhibited metallic  corrosion.    The corrosion  rate was  measured by loss  in weight  of the
weathering steel.   Six panels each were  exposed  to 16 polluted-air and 4 clean-air conditions,
and measurements were taken  at 250, 500, and 1000  hours  of  exposure.  The  weight losses were
converted to equivalent  thickness loss values.   As expected,  corrosion was most severe at high
SOy concentrations  and high humidity.  Ozone neither inhibited nor accelerated corrosion.   The
authors concluded that  some  other oxidant  or unmeasured  factor  that  was covariant with ozone
caused the  inhibition effect.   However,  if the  data  from the  sites with high  oxidant con-
centrations  in  the  field  exposure  experiments  were  excluded,  the damage  function  from the
laboratory  study  was  an  excellent  predictor  of  the   field  results.   The coefficient  of
determination for  the  field data  using the  laboratory function was 0.986.  The  following
physical damage functions  were  developed by Haynie et al. (1976) to  relate  S00 concentrations
                                                     2
to weathering steel  and  galvanized steel  corrosion (R = 0.91).
              corrosion  = [5.64 ^ + e <55'44  '  31,150/RT)^  ^
              corrosion  = (0.0187 SO, *  e 41'85  '  23,240/RT}  t
                                    €m                         W
         where:
              corrosion  is expressed as  \im
                                      3
              SQy is expressed at ug/m
              t  is time of wetness in years
              R = gas constant (1.98 cal/gm mol/°K)
              T = temperature (°K)
     Sydberger and Ericsson  (1976)  studied the  corrosion of  mild  steel  at  1, 10, and 100 ppm
(2620,  26,200,  and  262,000 ug/m3)  S02 across  the range  of critical  humidities (80  to  96
percent RH).  The flow  rate  of the $Q2  atmosphere'was  varied,  and some  samples  were sprayed
with  water  to  simulate rain  or condensation.   The chemical  composition  of the  corrosion
products was  studied by  X-ray diffraction, infrared spectrometry, and  electron spectroscopy
for  chemical  analyses  (ESCA)  techniques.   The  flow  rates   of  the   S02  atmospheres  markedly

SOX10A/A                                   10-14                                       -12-15-80

-------
influenced the corrosion  rates.   It appears that corrosion  rates  are related not only to the
S02 concentration  in  the  atmosphere,  but also to the  supply of S02 per unit surface area and
time.    Spraying  the  samples  with  distilled  water   at  intervals  substantially  increased
corrosion.
     Sydberger and  Ericsson's  (1976)  analysis of the corrosion product (rust layer) was based
on the  concepts of  Schwarz (1972) and  Barton  (1976)  that  sulfate  is  the  primary corrosion
stimulant  in  rust  formation.   Anodic  activity is maintained by the  concentration of ferrous
sulfate  in the electrolyte.   An S02  supply  of  4 ug/cm2/hr at the  lowest humidity initiated
corrosion  at  a  low  rate.   A  rise above  50 percent RH  increased  corrosion  markedly.   Of
particular interest was  the finding that a variation  of  the flow rate  at 1 ppm (2620 ug/ra3)
S02 with 96 percent  RH  gave  significant  differences  in corrosion rates.   This  study of the
effect of  rust  on corrosion showed that even at high humidity and high sulfate content, the
corrosion  rate decreased to a low level when the S02 concentration was low.
     10.2.1.2.2.2   Field Studies.   For outdoor exposures?  the primary rate-controlling factor
in the delivery  of pollutants to a surface is eddy diffusion.  This flux is not constant and
is a  function  of the horizontal wind  velocity gradient away from the surface.  The transport
of a  pollutant  to a  surface  is  usually  expressed as  a "deposition velocity"  (u), defined as
the flux to the  surface divided  by  the ambient  pollution level  at some specific measuring
height.   Reported deposition  velocities  for   gaseous  pollutants  have usually  been within an
order of magnitude of 1 cm/sec.  These  values are consistent with calculated estimates based
on an analogy with momentum flux and measured  wind velocity profiles  (Sydberger, 1976).
     The  amount  of S02 reaching a  steel surface depends on wind direction, wind velocity, and
the orientation of the surface to the  emission source.  The concept that S02 deposition varies
with  flow direction and velocity suggests  that  data on concentration alone cannot be used to
determine  the  supply  of S02 to  metal  surfaces;  therefore  surface adsorption methods like the
lead  candle method provide valuable information  in  relating supply of  sulfur oxides to metal
surfaces  (Sydberger,  1976).   Upham's  (1967)  work indicated,  however,  that corrosion of mild
steel  at  seven Chicago  sites  increased  with   time and with  increasing mean S02 concentration
(Figure  10-5).
     Haynie and  Upham (1971) continuously monitored urban  pollutants including S02, nitrogen
dioxide,  and  ozone (oxidants)  to  determine  whether previously unconsidered  variables might
affect  steel  corrosion.   Other factors  considered were temperature,  RH,  and total suspended
particulate matter (TSP).  Their  5-year program,  begun  in 1963,  involved sites in Chicago,
Cincinnati, New Orleans, Philadelphia,  San Francisco, Washington, Detroit, and Los Angeles.
They  studied  three  types  of  steel  expected  to  show  different   levels  of  resistance to
atmospheric corrosion:   (1) a plain carbon steel containing some copper (0.1 percent copper);
(2) a copper-bearing steel (0.22 percent  copper);  and (3) a  low-alloy weathering steel (0.4
percent  copper with  0.058  percent phosphorus).  The  exposure periods  were 4,  8, 16, and 32
months.   The  same steels were exposed at rural sites  as a control.  The rural sites proved to
have  higher than  expected  corrosion  rates; however,  S02 concentrations were  not measured at
SOX10A/A
                                           10-15                                       12-15-80

-------
    18.0
 o>
2  16.0
ui
<
 9"
§
X
u
5
oc
o
u
14.0

12.0

10.0

 8.0

 6.0

 4.0

 2.0

  0
                                                   3 (months)
              I
I
I
I
I
_L
I
I
             0.02   0.04   0.06  0.08   0.10
             (50)   (100)  (ISO)  (210)   (260)
                       0.12  0.14   0.16  0.18
                       (310)  (360)  (415)  (470)
                MEAN SO2 CONCENTRATION, ppm (
 Figure 10-5. Relationship between corrosion of mild steel and cor-
 responding mean SC>2 concentration is shown for seven Chicago sites.
 (Corrosion is expressed as weight loss of panel.)
 Source: Upham (1967).
                             10-16

-------
these  sites.   Multiple  regression  analysis  established  significant  correlations  between
average S02 concentrations and corrosion of all three steels at the urban sites.
     A  first  consideration was  climatic conditions.   For  metallic  corrosion  to occur,  a
certain level  of  wetness on the metal  surface  is  required  to provide an  electrolyte  film in
which  electrochemical  action can  take  place.   The average humidities recorded by  Haynie and
Upham  were  high  enough  to  produce  this  critical  wetness  threshold  much  of  the  time.
Temperature  was  a statistically  insignificant variable, although a later  chamber  study (see
previous  section)  did  find  temperature to  be  a  significant  variable.   Average RH was also
insignificant  because  the  range was only between  62  and 77 percent (Haynie and Upham, 1971).
     Inspection of the monitored SO^ and oxidant concentrations revealed wide variations from
site  to  site.   Multiple-regression analysis  showed  that  high  concentrations  of  oxidants
correlated with lowered metallic corrosion rates.
     The  resulting  physical damage  functions for  corrosion  of carbon  steel,  copper-bearing
steel, and weathering  steel developed by Haynie and Upham'(1971) follow:
                                          0.00161 SO
              Carbon steel:  Y = 9.013   e
              Copper-bearing              0.00171 SO
              steel:
              Weathering
              steel:
Y = 8.341   e
                              (4.768t)
                                      0.7512 -  0.00582 OX
(4.351t)
                                      0.8151   0.00642 OX
             0.0045 SO,
Y = 8.876   e
(3.389t)
        0.6695 - 0.00544 OX
         where:
              Y = corrosion in urn
              SOp and OX (ozone) are expressed as ug/m
              t = time of exposure, years
The  correlation  coefficient for each  of these equations was 0.91.   A  more recent laboratory
investigation by  Haynie  et al. (1976),  however,  (see  preceding section) has shown that ozone
is not a significant corrosion controlling variable.
     Mansfeld  (1980) made  observations at  nine  test sites  in and  around St.  Louis  for 30
months beginning in October 1974 as part of the Environmental Protection Agency's Regional Air
Pollution Study to determine the effect of airborne pollutants on galvanized steel, weathering
steel, stressed  aluminum,  marble,  and house paint—essentially the same materials examined in
the  chamber  study  reported  by Haynie  et  al.  (1976).    During 1975  and 1976,  atmospheric
corrosion monitors (ACM) of the type described by Mansfeld and Kenkel (1976) were installed at
four  sites  to measure  time of wetness.   Each ACM consists of  a  copper-zinc  or copper-steel
couple that registers current flow when an electrolytic path forms between the two plates as a
result of  deposition of  water from the air, dew,  or rain on  corrosion products.   The ACM
measures the  time that the panel is wet enough for the electrochemical mechanism of corrosion
to  occur.  Parameters measured  in  Mansfeld1s study included the following:  wind  speed, wind
SOX10A/A
                                           10-17
                                                          1^15-80

-------
direction, temperature, ozone concentration, total hydrocarbon concentration, total sulfur and
NO   concentration,  hydrogen sulfide concentration, sulfur dioxide  concentration,  RH,  sulfate
concentration,  nitrate concentration,  total  suspended  particle concentration,  and  time  of
wetness.
     Mansfeld  determined  weight losses  for galvanized steel, weathering  steel,  house paint,
and  marble;  he  removed  aluminum tension  samples after failure.   Mansfeld1s data  show that
damage to a  particular material does not necessarily occur at the same corrosion rate at each
site.  Preliminary  statistical  analysis  of the results failed to show significant correlation
between  corrosivity  and  pollutant  concentrations.    There  was  substantial  error  in  the
measurement  of  RH,  an  extremely  important  corrosion  variable.   Relative  humidity  was,
therefore,  not  included  in  the  regression  analyses.   Concentrations  of  SO, measured  by
                                                                                            3
Mansfeld  (1980) were  generally an  order  of magnitude lower than the  130 ppb  (340 ug/m )
concentrations reported by Upham (1967) at urban sites in St. Louis.
     In  another  study (1974),  Haynie  and  Upham exposed  enameling  steel   containing  0.019
percent carbon  and  0.028  percent copper at 57  sites  in  the National Air Surveillance Network
(NASN).   They   measured  corrosion   by  weight  loss  and  quantitatively -determined  other
pollutants,   including  gaseous sulfur  dioxide,  total  suspended  particles,  and  the  amount of
sulfate and  nitrate in the particles.  For  a  discussion  of the measurement methodology for
NASN,  see Chapter  3.   Sulfur  dioxide  levels  ranged from  9 to 374 ug/m ,  total  suspended
particles from  11  to  182  ug/m ,  and RH  from  29  to 76  percent.   The  average  temperature
remained  within a  fairly  narrow  range and  was considered  constant.   The  temperature,  the
quantity  of  total   particles,  and  the  presence  of  nitrate  in  the  particles  did  not
significantly  affect the  corrosion  rate of  steel.   The  concentration of sulfur  dioxide was
significant  only when  sulfate  was  not  included  in  the regression analysis.  At  each site,
sulfate content of  the particles and  sulfur  dioxide concentration were closely related.   On
the basis of this study,  Haynie and Upham derived the following empirical  expression to obtain
the  best relationship between  corrosion  of  enameling steel  and  atmospheric SOp  content:

          Corrosion = 325tV0-00275 S02 '  163'2/RH>
where:     corrosion is  expressed as urn
                                  o
          S0~ is expressed as ug/m
          RH = relative humidity (percent)
          t = time
Average RH is a substitute variable for the fraction  of time the steel is wet.
     Considerable  effort  has  gone  into  isolating  environmental  variables  that  predict
long-term corrosion rates.   Empirical  expressions for corrosion  of various  steels exposed to
the  atmosphere  (see Table  10-1)  have  been  developed  by  Oma  et  al.  (1965), Chandler and
Kilcullen  (1968),  and Haynie  and  Upham  (1974).   These  equations may  be  used to relate
reduction in S02 and  sulfates to reduction in  corrosion of metals, serving  as  a  basis for a
benefit appraisal.

SOX10A/A                                   10-18                                       12-15-80

-------
                                                   TABLE 10-1.   SOME EMPIRICAL EXPRESSIONS FOR CORROSION
                                                                  OF EXPOSED FERROALLOYS
Exposure Study
Material Site Duration Empirical Equation
Steel A Sheffield 1 year y = 0.51 + O.Olx
England
Parameter Units
y = corrosion rate
in mils/year 3
x = S02 in ug/nr
Source
Chandler and
Kilcullen
(1968)
Note/Comments
Authors stated that S02 and smoke
has a major influence on the corrosion
rate of steel and accounted for about
50% of the variations found at the
different sites. Other factors, such
as time of wetness, were found equally
important in determining the corrosion
rate of steel.
95% confidence limit + 0.75 roils for
any point on regression.
Steel B      Sheffield  1 year
              England
                                     y = 0.82 + O.OOSx
                                                                  y = corrosion rate
                                                                   in mils/year  ,
                                                                  x = S02 in ug/m
                                                                               Chandler and
                                                                                Kilcullen
                                                                                (1968)
95% confidence limit + 0.5 mils for
any point in regression.
Enameling
 steel
NASN
sites
1-2
years
                                     cor =183.5 /I [e0-
Enameling
 steel
NASN
sites
1-2
years
              faO.00275 SO,-
cor = 325 J~t ie          '
                                          cor = depth of          Haynie and
                                          corrosion, um            Upman
                                          sul = average level      (1974)
                                          of sulfate in suspended
                                          particulate, pg/m
                                          RH = average relative
                                          humidity, percent
                                          t = time in years
                                                                              parameter units         Haynie and
                                                                              same as for previous     Upham
                                                                              equation except          (1974)
std. dev. 85.0; 95% conf. limits 13.4
lower, 353 upper
std. dev. 0.00451; 95% conf. limits
0.0552 lower, 0.0732 upper
std. dev. 32.8; 95% conf. limits 97.6
Sul and SO, relationship defined as
sul = 8.9 * 0.0429 SO,
± 2S = 8.2           *

-------
     Matsushima  et al.  (1974),  in  studies  of low-alloy  weathering steels,  considered  the
impact of  the  washing action of rain, the  ease with which water would drain off the surface,
and the  drying effect of sunlight to determine the effect of these variables on the retention
of particles that influence the electrolytic corrosion mechanism and the time of wetness.   The
authors  hypothesized  that  the geometry  of  unpainted  weathering  steels  may  not  favor  the
development  of a protective  oxide  film of  rust.   The model  structure used in  the exposure
trials contained  horizontal  and inclined  roofs, vertical  wall  panels, and window frames.   Two
sites were chosen:  an  industrial  location and a residential site in the Kawasaki area, which
has a cold, dry winter and a hot,  humid summer.
     The results showed that the successful  use of weathering steel is related not only to the
severity of  pollution  but  also  to  the specific  interplay between  shelter and  the  uniform
washing action  of rain.   Thus,  for areas  in which the structural factors are unfavorable,  the
optimal  rust film  is slow  to form  and  may deteriorate.   Rust  films  develop and are  then
destroyed,   and  the  surface  never  develops  a  protective  film.   Generally,   boldly  exposed
surfaces such as horizontal or inclined roofs show the least corrosion.
10.2.1.2.3   Comparison of Ferrous and Nonferrous Metals.  Sydberger  and   Vannenberg   (1972)
examined the influence of RH and rust on the adsorption  of sulfur dioxide on metal surfaces,
                                                                             o
using  radioactive sulfur.  The concentration of S02 was  0.1  ppm (262 ug/m ), and RH varied
between  50  and 98 percent.   Polished and  preexposed  samples  of iron,  zinc,   copper,  and
aluminum were  compared for  their adsorption properties.   Iron,  zinc,  and copper  were  pre-
exposed to S02 concentrations (100 ppm or  2.6 x 105 ug/m3) at 98 percent RH and 22°C for 3 hr.
The aluminum samples were preexposed for 30 hr.   The principal  corrosion product identified by
X-ray  diffractometry  was hydrated  metal  sulfate.   Adsorbed  S02  was  measured at  30-min
intervals with  a  Geiger  counter.   The corrosion rate at  90  percent RH was initially high for
zinc and copper but quite low for aluminum.   Adsorption of S02 on preexposed iron samples was
high.   At  80 percent  RH, almost all  of  the S02  was adsorbed.  The high  adsorption  rate is
perhaps explained by  the rapid oxidation of  adsorbed sulfur dioxide caused by the catalytic
effect of the  rust.   The initial  rate of adsorption  on polished  iron below 80 percent RH is
related to the absence of corrosion;  however,  at increased humidity corrosion is initiated and
the adsorption rate increases.
     Of  particular significance  is  the  observation that  S02  adsorption  can  take place at
humidities  below  the critical  humidity  (Sydberger  and Vannenberg,  1972).   This  finding
suggests that  SO- will  be adsorbed on a rusty iron surface during periods of low humidity and
will affect  the  corrosion rate when  humidity rises.   The critical  humidities  for nonferrous
metallic surfaces,  as summarized from Nriagu  (1978)  and National Academy  of  Sciences (HAS)
(1977), appear  in Table 10-2.  The corrosion products of copper and aluminum have an extremely
low adsorption  capacity  below 90 percent  RH, confirming the lower sensitivity of these metals
to corrosion by S02 (see  Figure 10-6, from Sydberger and Vannenberg, 1972).
     Aluminum  is  generally  considered  to  be corrosion  resistant.   It is  quite  resistant to
sulfur oxides  when RH  is less than  50 percent.   At higher  humidities  aluminum can  corrode

SOX10A/A                                   10-20                                       12-15-80

-------
  E
  o  2
  V)
  CD
  E
  O

                 I     I
I     I
                      34567


                         EXPOSURE TIME, hr
                                                    COPPER
                                                ALUMINUM
                               10
Figure 10-6. Adsorption of sulfur dioxide on polished metal surfaces
is shown at 90 percent relative humidity.

Source:  Adapted from Sydberger and Vannenberg (1972).
                         10-21

-------
                        TABLE 10-2.   CRITICAL HUMIDITIES FOR VARIOUS METALS
                       Metal                        Critical  humidity,  percent RH

                     Alumi num                                  75-80
                     Brass                                     60-65
                     Copper                                    65-70
                     Nickel                                    65-70
                     Zinc                                      70-75
rapidly in  the presence of very high  SO-  concentrations  forming a hydrated aluminum  sulfate
[A1_(SO«)3 • ISHpO] surface deposit.   At low concentrations of acid sulfate particles it forms
a protective film  of  aluminum oxide.   However, when the  film becomes contaminated with dirt
and soot  particles,  there  is  a change  in surface appearance  characterized  by  mottling  and
pitting.   Simpson  and  Horrobin (1970)  reported that  aluminum  undergoing long  exposure  in
industrial areas displayed  white areas of  crystalline corrosion products.  Aluminum surfaces
exposed for periods of  more than 5 years  to  a sulfate  concentration of 0.14 ppm (550 ug/m )
had pits  as deep  as  14 mils (0.36 mm).  This is,  of  course,  much higher than typical  ambient
concentrations  (see Chapter 5).
     Fink et al.  (1971)  summarized  measured corrosion rates  and  depth  of pitting of aluminum
                                           3                                              3
surfaces in  rural, mild industrial  (30 ug/m  or 0.01 ppm SOp), normal  industrial  (370 ug/m  or
0.14 ppm  SO-  and  80% RH), and severe  industrial  areas.   Their overall conclusion  was that,
although some  loss of thickness occurred in the first  2-years, structures composed of aluminum
and its alloys  are resistant to air pollutants.
     In another study,  Haynie  (1976)  found that stressed  aluminum specimens exposed to SO, at
                                                3
concentrations   of  79  (0.03 ppm)  and  1310 ug/m  (0.5  ppm)  lost  approximately  8.6 and  27.6
percent of  their  bending  strength, respectively.   He also noted  that  7005-T53  high-strength
aluminum alloy  tubing,  which  contains  very little copper,  is susceptible  to  stress-corrosion
cracking in  industrial environments.
     Abe et al. (1971)  exposed copper and copper alloys  for  2 years in marine,  rural, highly
industrial,  and urban  areas in which  there was great  variation in pollutant and  salt content.
Analyses of the surface deposits  showed the presence of  basic sulfate, sulfide, and chloride
of copper.  The green patina on copper was analyzed and'determined to be basic copper sulfate
in urban  areas and basic copper chloride  in  seacoast areas.   These  surfaces were protective
against further corrosion.
     The  formation of  these basic  copper salts, according to  Simpson and Horrobin (1970),  can
take as long as 5 or more  years and  will  vary with the  concentration  of sulfate or chloride
particles, the  humidity, and  the temperature.  They reported  the  corrosion rate of copper to

SOX10A/A                                   10-22                                       12-15-80

-------
be 0.9  to 2.2  pro/year in industrial atmospheres,  compared  with  0.1 to 0.6 urn/year  in  rural
areas.
     The  high corrosion  resistance of  nickel  and copper  compared  with  unalloyed  steel  is
attributed by Sydberger (1976)  to the ability  of  these  metals to form a  layer  of insoluble
basic sulfate  that  protects  the metal  surface.  Such layer formation does  not occur on steel.
10.2.2  Protective Coatings
     Susceptible  materials  are  generally  coated  for  protection  against  the  effects  of
exposure.  The  coatings provide  either  sacrificial  protection  or  barrier  protection.   In
galvanization,  zinc  is applied to  ferrous metal  for sacrificial  protection.   Thus, while the
galvanized surface  may suffer corrosion damage, it  helps to prevent rusting of steel products
such  as  gutters,  cables,  wire  fencing,  and  building  accessories.   Barrier protection  is
provided by varnishes,  lacquers, and paints by sealing the underlying surface material against
intrusion by moisture.
10.2.2.1  Zinc-Coated Materials—Zinc is  generally exposed-as  a  protective coating for steel
products since zinc  coating is fairly resistant  to  atmospheric corrosion.  Zinc is anodic with
respect  to  steel;  when zinc and  steel  are in  contact with an electrolyte,  the electrolytic
cell  provides current  to  protect  the  steel  from corrosion with  some  oxidation  of the zinc.
     Guttman  (1968) carried out  a long-term  exposure of zinc  panels with  measurement of the
atmospheric  factors.   He  found  that zinc  is  corroded by  SO^ and  that time of  wetness and
concentration  of  SOp are the major factors that  determine the rate of corrosion.
      Fleetwood  (1975)  conducted  5-year  exposure  studies  of  zinc  and iron  in  a number of
locations  ranging   from  dry  tropical   to industrial.   He estimated  the  service  life  of
galvanized  steel  to be 15 to  20  years  in  an  industrial area containing  pollutants and 300
years  in a  dry tropical unpolluted area.  Kucera  (1976)  noted  strong correlations  between the
corrosion rate and  (1) the adsorption  of S02  on zinc  surfaces and  (2) the concentrations of
so2.
      Haynie  and Upham  (1970) exposed zinc panels  in eight cities, continuously monitoring S02
concentration and  collecting meteorological  data, including  temperature and  RH,  from the
nearest  weather stations.   They developed the  following  empirical equation,  which correlates
corrosion  rate with average  sulfur dioxide  concentration  (for the  study  range  of 10 to 479
ug/m^) and RH:
                    Y   =     0.001028 (RH - 48.8)  S02,
         where
                    Y   =     zinc  corrosion rate (um/yr),
                    RH   =     average annual RH  in percent, and
                    S0? =     average S02 concentration  (ug/m )-
The  regression intercept  indicated that no  corrosion would  occur below  an average  relative
humidity of  48.8  percent.    This expression   gave  a reasonably good linear  fit  with the
experimental  corrosion results  obtained by Haynie and  Upham  for S02  concentration and RH.

JSGX10A/A                                  10-23                                      12-15-80

-------
     Based on  the  St.  Louis  study results (Haynie, 1980), the corrosion of small specimens of
galvanized steel follows the relationship:
     Cz   =    2.32 tw + 0.0134 v °'781 - S02 • tw>
     C    =    corrosion in urn
     t    =    time-of-wetness in years,
     v    =    wind velocity in m/s, and
     S02  =    ug/ra
     A theoretical  damage coefficient  (for  purposes  of the chapter, damage =  any measurable
adverse effect) for a pollutant can be calculated from the stoichiometry of a reaction and the
deposition  velocity.    For the  reaction  between  SOp and  zinc to  form  zinc sulfate,  the
coefficient is 0.045,  when the zinc corrosion rate is expressed fn micrometers per year, SO^
in micrograms per cubic meter, and the deposition velocity (u) in centimeters per second.   For
a  small  zinc  or galvanized steel sheet  specimen,  the material  damage coefficient  for  S07 is
                             0 78
calculated  to  be  0.0123  v "    which  agrees  well  with  the  above  determined  empirical
                                                                                 o
coefficient.   At a  wind velocity of 4 m/sec, the  value is 0.0363 (um/yr)/  (pg/m  ).   For the
same  conditions,   a  similar   calculation   for  marble   yields  a  coefficient   of   0.136
             3
(um/yr)/(ug/m ).
     Haynie  (1980)  restudied the  results  of  six  exposure  investigations   to  relate  the,
corrosion  of  zinc  and  galvanized  steel  to  the concentration of  sulfur  dioxide.    Each
investigation  was  different  and  the  data  were evaluated  differently;  thus,  no  direct
comparison of the  results  as  they were published  was  possible.   However,  in reevaluating the
data from each study, Haynie used techniques  that permitted the  comparison  of the various data
sets.
     The experimental  regression  coefficients  obtained from all of  these studies are compared
in  Table 10-3.   The  SO^  coefficient  for  the  chamber study  is low,  whereas  the  analogous
coefficients  for  the  Community  Air Monitoring  Program (CAMP)  (Haynie  and Upham,  1970)  and
Interstate Surveillance Program  (ISP,  Cavender et al., 1971) studies are  high and agree with
each  other.   The  remaining  three  S02  coefficients  are  generally  in  good agreement.   The
average of the time-of-wetness coefficients  are within a  range of  + 0.75  from a mean of 1.73
um/yr.
     The specified thickness of galvanized coating varies with intended  use.  Furthermore, the
thickness  of  a  particular coating  varies  considerably  from  one  point to another.   Bird's
measurements revealed that 5 percent of coating thickness measurements varied from the mean by
more  than 46  percent  (Bird,  1977).    Haynie  confirmed this variability  with  475  thickness
measurements  on  a  single  galvanized  steel  sheet.    The life  of  a   coating is  generally
proportional to  its thickness; thus,  rusting of  the  substrate  steel will   occur first  at the
thinnest spots and last at the thickest spots.   The American Society for Testing and Materials
(ASTM) as  reported in  Haynie (1980)  has  observed rusting at thin spots  on galvanized steel
wire, fencing,  and sheet  exposed to various  types of atmospheres  over many years.   Some of
their exposures were started in 1916 and were continued until  the test could reveal no

SOX10A/A                                   10-24                                       12-15-80

-------
                      TABLE 10-3.   EXPERIMENTAL REGRESSION COEFFICIENTS WITH
                         ESTIMATED STANDARD DEVIATIONS FOR SMALL ZINC AND
                           GALVANIZED STEEL SPECIMENS OBTAINED FROM SIX
                                          EXPOSURE SITES
         Study
Time-of-wetness
  coefficient,
     Mm/yr
S0« coefficient,
(ufn/yr)/(Mg/m )
Number of
data sets
         CAMP (Haynie and
          Upham, 1970)
         ISP (Cavender et al.,
          1971)
         Guttman, 1968
         Guttman and Sereda,
 1.15 + 0.60

 1.05 + 0.96
 1.79
 0.081 + 0.005

 0.073 + 0.007
 0.024
   37

  173
 >400
1968
Chamber study (Haynie
et al., 1976)
St. Louis (Mansfeld,
1980)
2.47 + 0.86
1.53 + 0.39
2.36 + 0.13
0.027 + 0.008
0.018 + 0.002
0.022 + 0.004
136
96
153
additional  information.   In the case  of  sheet,  the product was completely  rusted  and showed
perforations.   In  general, the  amount of  corrosion  at each site varied  linearly  with time.
Corrosion rates  at  each site were calculated  on  the  basis of time to  first rust and time to
complete rust for various zinc thicknesses, assuming +40 percent thickness variability.
     Originally, State College, Pa.,  was selected by  ASTM as a control  site  representing a
"clean"  rural  environment; however, the  corrosion  was  higher there than  at five other rural
locations including the rural-marine  environment of  Santa Cruz, Calif.,  where high relative
humidities  are  expected  to accelerate corrosion.   ASTM  made  no  pollution  measurements  but
recognized the effects of "industrial" and "severe industrial" environments.
     Zinc corroded nearly twice as fast on wire and fencing as it did on sheet,  a finding that
is  consistent  with  the  theoretically  predicted  effects  of  surface  configuration  on  S02
deposition  velocity.   One would expect a greater deposition velocity  onto  fencing than onto
sheet material.  ASTM noted that fencing corrodes  less  near the ground than it does near the
top  because wind  velocity  increases  with  height,  with  a  resultant  increase  in  deposition
velocity  and similar variation of S02 concentration.   Another factor may be the decrease in
concentration  near  the ground,  since both soil and vegetation are sinks for S02 (see Chapters
6 and 8).
 OX10A/A
                                           10-25
                                                   12-15-80

-------
     The average corrosion  rates  in the Haynie (1980) study shown in Table 10-4 correspond to
actual  corrosion  rates  that  are  two  to  four  times  greater  when  the  substrate  is  wet.
Theoretical calculations indicate  that  the average SO, levels at the Pittsburgh site  over the
                                                     3
long period of exposure were between 350 and 700 ug/m .   The average at the Altoona site could
                              3
have been as high as 1000 pg/m .
     From  the  relationships  between  theoretical  and  experimental   studies,  Haynie  (1980)
concluded the following:
     1.    Both  short-term  laboratory evidence  and  long-term exposure results  for galvanized
          steel are consistent with theoretical considerations.
     2.    Damage functions for some materials can be calculated from theoretical relationships
          that consider factors controlling time of wetness and pollutant fluxes.

           TABLE 10-4.   CORROSION  RATES OF ZINC ON GALVANIZED STEEL PRODUCTS EXPOSED TO
                                VARIOUS ENVIRONMENTS PRIOR TO 1954
                                                Mean corrosion rate and estimated standard
                                                           deviation,  urn/year
         Site
         Sheet
Wire and fencing
         Altoona,  PA
         Pittsburgh, PA
         Sandy Hook, NJ
         Bridgeport, CT
         Lafayette,  IN
         Ithaca, NY
         State College, PA
         Ames, IA
         College Station, TN
         Santa Cruz, CA
         Manhattan,  KS
         Davis, CA
     7.57 + 0.54
     5.63 + 0.34
     2.74 + 0.30
     1.27 + 0.29
 10.86 + 1.02
  4.37 + 0.45
  4.25 + 0.44
  2.94 + 0.34
  2.68 + 0.42
  2.48 + 0.24
  1.68 + 0.19
  1.22 + 0.43
  0.83 + 0.26
  0.79 + 0.27
  0.76 + 0.42
         Source:  Haynie (1980)
SOX10A/A
10-26
                12-15-80

-------
     3.    Wind speed and material geometry should be considered in evaluating
          atmospheric corrosion effects.
     Marker et al.  (1980)  examined the variables controlling the corrosion of zinc by SO,, and
sulfuric acid.  They  used  an aerosol flow reactor.   Under  steady-state conditions, they made
the following measurements:
              Environmental Measurements:
              o    Percent RH and temperature (at two points).
              o    Average flow velocity (Pitot tube).
              o    Flow velocity profile (recorded when a steady state had been estab-
                   lished).
     Aerosol Measurements:

              o    Aerosol size distribution and number concentration determined at
                   intervals during test by TSI 3050 analyzer.
              o    Two total-mass filter samples collected.
              o    Total-deposition sample collected on aluminum foil throughout each
                   experiment.
              o    TEM deposition grid samples collected continuously.
              o    X-ray photoelectron spectroscopy samples (both zinc plate and
                   aluminum foil) collected continuously during experiment.
Corrosion  rate  measurements were  recorded continuously  by an atmospheric  corrosion monitor
(ACM), which had been pretreated with either 0.1 N sulfuric acid or ammonium sulfate.  The ACM
consists  of galvanic cells  of alternating  plates  of copper and zinc  separated  by sheets of
plastic insulators.  Experimental conditions were selected from the following ranges:
              Temperature, °C                                           12-20
              Relative humidity, percent                                65-100
              Mean flow velocity, m/sec                                0.5-8
              Sulfur dioxide concentration, ppb (volume)                46-216
              Sulfate aerosol mass concentration, mg/m                 1.2
              Aerosol size distribution, urn diameter                   0.1-1.0
     The  factors  controlling  the rate of corrosion in the Marker et al. study (1980) were RH,
pollutant flux, and chemical form of the P9llutant.   Corrosion occurred only at RH high enough
(more  than 60 percent)  to wet  the  surface; temperature  did not appear  to be a controlling
factor  within  the range  12-20°C.   The results indicate that on  initial  exposure S02-induced
corrosion  of zinc proceeds at a rate  approximately  a factor of two greater than that for the
equivalent amount of deposited sulfuric acid aerosol.
     The  investigators  noted  deposition  velocities  of  0.07 cm/sec  for 0.1-1.0  pm sulfate
aerosols  and 0.93 cm/sec  for S02 at a friction velocity of 35 cm/sec.  These factors indicate
that the effects of S02 will dominate the effects of H2S04 in most urban areas.
10.2.2.2  Paint Technology and Mechanisms of Damage—Cotnpared with other environmental impacts
such  as  sun and precipitation,  paint  damage  due to  air  pollutants  is  considered  less
important.  There are at present no standard ASTM procedures for evaluating the effect of S02>


30X10A/A                                   10-27                                       12-15*80

-------
nitrogen  dioxide,  and/or ozone on paints.  Degradation  by  ultraviolet light has received the
major emphasis; outdoor test stations have been located where SCL levels are low.
     The  erosion  of  paint can be measured by  loss of thickness of the paint layer, which can
result  from  the  chemical  action of SOp and the action of light and ozone.   Film erosion rates
are used by paint manufacturers to determine the fail point for their formulations.
     In the  formulation  of paints,  the ratio  of  pigments  to film formers is of importance in
the overall  properties  of gloss,  hardness, and permeability  to water.   If the amount of film
former  is too low, soiling is increased and the paint may lose the film flexibility needed for
durability and become  brittle.   The permeability of paints  to  water has been reported by Hay
and Schurr (1971).   High-permeability  films  are desirable for  surfaces  that must allow water
to  pass  through,   such   as  wooden  exterior  walls   behind   poorly  ventilated  kitchens.
Low-permeability  coatings  are  needed  to  protect  surfaces  that  corrode  when  repeatedly
moistened.  The  low  permeability  of chlorinated rubber is advantageous for use on concrete, a
use this coating shares with styrene-aerylie.
     Paint films  permeable to water are also  susceptible  to penetration by  S02  and sulfate
aerosols.   The  absorption  of S02  was observed  by  Holbrow (1962),  who  found  sulfites  and
sulfates in paint, and by Walsh et al.  (1977),  who used radioactive S02 to determine rates and
saturation values for S02 absorption.
     Concentrations  of  SOp  encountered   in  fogs  or near industrial  sites  can  increase  the
drying and hardening times  of certain kinds of paints.   Holbrow (1962)  found that the drying
time of linseed,  tung,  and certain  castor oil paint films  increased by 50  to  100  percent on
                              3
exposure to 2620  to  5240 ug/m  (1 to  2 ppm)  SOp.   The touch-dry  and  hard-dry times of alkyl
and oleoresinous paints with titanium dioxide pigments were  also reported to increase substan-
tially;  however,  the exposure  time  of the wet films  was not reported.   Analysis of the dried
films  indicated  that  S02  had  chemically  reacted  with   the  drying  oils,  altering  the
oxidation-polymerization process.   No  studies  have been reported on the  effects of SOp on the
drying of latex paints.
     Holbrow (1962) also  studied  the effects of sulfur dioxide on dried  paint film.  In these
experiments,  paint films-were  allowed  to dry,  were refrigerated, and then were exposed for 15
rain to an atmosphere containing 1.2  percent S02-   The paint films with condensed moisture were
finally placed in  an  accelerated-weathering  chamber.  For all paints except a pentaerythritol
alkyd paint,  the  gloss  decreased  significantly  after  1  day in the  accelerated-weathering
chamber.  Without  the  accelerated weathering, the actions  of sulfur dioxide and moisture on
the paint films  produced  only a slight reduction in gloss..  Holbrow concluded that the sulfur
dioxide had sensitized the  film,  permitting water to be absorbed during the weathering cycle.
                                                                                         T
     Bluing of lead chromate pigment (green) paints has been observed during the early life of
the film.   Holbrow (1962)  reproduced  this effecttin  the laboratory by exposing the film to
sulfur dioxide and moisture  and then to  warmth and moisture.   The bluing was probably caused
by conversion of the lead chromate pigment to lead sulfate.   Holbrow/did not attempt to
SX10AA/A                                   10-28     -                                 12-15-80

-------
correlate moisture, temperature, and pollutant concentration,  or to obtain dose-response aata.
Although very high levels of sulfur dioxide were used, this experimentation does indicate trat
condensation  and moisture  evaporation are  critical   in  concentrating  the pollutant  on  the
surface of  exterior paint  films;  under these  conditions, deterioration of the  film occurs.
     Svoboda et  al.  (1973)  compared pigmented and unpigmented paint film for S02 permeability
and found  that  the  rate  of penetration of SCL  into  a paint  film was related  to  the pigment
content.  Zinc oxide and titanium dioxide pigments caused  a  50 to 70 percent decrease in the
rate of penetration of SOp into the paint film.
     Spence et al.  (1975) carried out a chamber study of the effects of gaseous pollutants on
four classes  of  paints formulated for exterior exposure:   oil-base house paint, vinyl-acrylic
latex  house  paint,  and  vinyl  and  acrylic  coil coatings  for metals.  The  house  paints were
applied to  aluminum panels by  spraying.   The coil coating panels were  cut from commercially
painted stock.   The  oil-base paint film was  58 urn thick; the acrylic latex, 45 urn; the vinyl
coil coating,  27 urn; and the acrylic  coil  coating, 20 urn'.   The  exposure chambers controlled
temperature,  humidity,  SCL, nitrogen  dioxide,  and ozone.  Each exposure chamber  had a xenon
arc lamp  to provide ultraviolet radiation.  A dew/light cycle was  included;  light exposure
time was  followed  by  a dark  period wherein  coolant circulated through racks  holding  the
specimens,  thereby  forming dew on  the  panels.   Each  dew/light  cycle  lasted  40 min  and
consisted of 20 min  of darkness with formation of dew, followed by 20 min under the xenon arc.
The total  exposure time was 1000  hr.   Damage was measured after  200-hr,  500-hr,  and 1000-hr
intervals by  loss  of weight and by  loss  of film thickness.  In evaluating the data, loss of
weight was converted to equivalent loss of film thickness.
     Visual  examination  of  the panels  coated  with  oil-base  house  paint revealed  that  all
exposure conditions  caused considerable  damage.   The  erosion  rate  varied  from  28.3 to 79.1
urn/year, with  an average of 60 urn/year.  The investigators concluded that S02 and RH markedly
affected  the  rate  of  erosion  of  oil-base  house  paint.  The presence of  nitrogen dioxide
increased the weight of the paint film.  A multiple linear regression on S02 concentration and
RH yielded the following  relation:
          E = 14.3 + 0.0151 S02 + 0.388 RH,
where
          E = erosion rate in urn/year,
                                            o
          S02 = concentration of S02 in ug/m  , and
          RH = RH in percent.
The authors  reported the 95 percent tolerance limits on 99 percent of the calculated rates to
be ± 44 urn/year.
     The formation  of  blisters  on acrylic latex house paint was noted at the high S02 levels.
The blisters  resulted from severe  pitting  and  buildup of aluminum corrosion products  on the
substrate.  The  paint  acted as  a membrane  retaining  moisture under the surface and excluding
SX10AA/A
                                           10-29                                       12-15-80

-------
oxygen which  would passivate  the  aluminum.   The  vinyl  coating and  the acrylic  coating  are
resistant  to  SO^.   The  visual  appearance of  the  vinyl coil  coating showed  no  damage.   The
average erosion rate was low, 3.29 urn/year.   The average erosion rate for a clean air exposure
was  1.29  urn/year.   The acrylic coil coating showed  an  average erosion rate of  0.57 urn/year.
     A  study  of  the  effects  of air  pollutants on paint  was conducted  by  Campbell et  al.
(1974).   The  paints  studied  included  oil  and  acrylic  latex house  paints,  a coil  coating,
automotive refinish, and an alkyd industrial  maintenance coating.   These coatings were exposed
                                       3                                3
to clean  air,  S02  at 262 and 2620 ug/m ,  and ozone at  196  and 1960 ug/m  (i.e.,  equivalent to
0.1  and  1.0  ppm  of  each  pollutant).   Other  controlled  study  variables   included  light,
temperature, and RH.   In addition,  one-half of the coatings were shaded during the laboratory
exposures.  Similar  panels (half facing  north) were exposed at field sites  in  Leeds,  North
Dakota; Valparaiso, Indiana; Research  Center,  Chicago,  Illinois;  and Los Angeles, California.
     The  laboratory  exposure chamber  operated  on  a 2-hour light-dew cycle (i.e.,  1 hour of
xenon  light at  70  percent RH and a temperature  of 66°C followed by 1 hour of darkness at 100
percent RH and  a  temperature of 49°C).  Coating erosion rates were calculated after exposure
periods of 400, 700,  and 1000  hours.   Estimated  erosion  rates  and  statistical  characteri-
zations of the results are summarized in Table  10-5.    Erosion rates  at  03 or S0£ concentra-
tions of  0.1  ppm were not significantly different from  values for clean air exposures due to
high variability  of the data.   The erosion  rates  on the shaded specimens  were  significantly
less than  the  unshaded panel  results shown in Table 10-5;  panels  facing north were also less
eroded.   At  1 ppm  pollutant  concentrations  erosion rates  were  significantly   greater  than
controls with oil-base house paint  experiencing the largest erosion rate increases, latex and
coil  coatings  moderate  increases,  and  the  industrial  maintenance  coating  and  automotive
refinish the  smallest  increases  (Yocom and Grappone 1976;  Yocom and Upham 1977; and Campbell
et al.,  1974).   Coatings  that contained extender  pigments,  particularly  calcium carbonate,
showed the greatest erosion rates  from the SO^ exposures.  Results  of  field exposures also
support these conclusions (Campbell  et al.,  1974).
10.2.3  Fabrics
     Fibers that suffer destructive action upon exposure to acids derived from S02 include (1)
cellulesic  fibers  such  as cotton  and  its  close relative  viscose rayon,  a  regenerated
cellulose, and cellulose acetate; and (2) polyamide fibers  such as  nylon 6 and 66.  Polyester,
acrylic,  and  polypropylene fibers are  not  damaged directly by SO^-   However, sulfur dioxide
concentrations can  be  a source of absorbed acid which  can accelerate the  fading of dyes and
                                                                                     i
result  in fabric  deterioration  through chemical  reactions.   The  possibility of higher acid
content due  to oxidation  of sulfur dioxide  to sulfur  trioxide  must be  considered (Salvin,
1963).
     Brysson et al.  (1967) exposed cotton fabrics at 12 different environmental sites in St.,
Louis., MO.  and  Chicago, IL, for a period of up to one year (1963-64).  The seven sites in the
St. Louis metropolitan area represented industrial, urban,  suburban, and rural-suburban
SX10AA/A                                   10-30                                       12-15-1

-------
               JABLE_ig-5._ PAINT. EROSION RAJES AND MEST.PROBABILIIY.DATA

                TABLE 10-5a.  PAINT EROSION RATES AND PROBABILITY DATA (T-TEST)
                       FOR CONTROLLED ENVIRONMENTAL LABORATORY EXPOSURES


                                 Mean erosion rate (mil  loss X 10  /hour with 95 percent
                                      confidence limits) for unshaded panels and
                                 percent probability that differences exit in parentheses
Type of paint
House paint
oil

latex

Coil coating

Automotive refinish

Industrial maintenance

Clean air
control

20.1 + 7.2

3.5 + 1.5

11.9 + 2.3

1.8 + 0.8

18^6 + 5.1

SO,
(1.0 p£m)

141.0 + 19.0
(99)
11.1 + 1.0
(99)
34.1 +4.7
(99)
3.1 + 2.6
(75)
22.4 + 7.0
(66)
°3
(1.0 ppm)

44.7 + 10.5
(99)
- 8.5 + 5.9
(93)
14.9 +2.5
(94)
5.1 + 1.3
(99)
28.1 + 14.0
(85)
                  TABLE 10-5b.  PAINT EROSION RATES AND PROBABILITY DATA (T-TEST)
                                     FOR FIELD EXPOSURES


                            Mean erosion rate (mil loss X 10  /month with 95 percent
                                 confidence limits) for panels facing south and
                            percent probability that differences exist in parentheses
Type of paint
House paint
oil

latex

Coil coating
» *»
Automotive refinish

Industrial maintenance

Rural
(clean air)

4.3 + 7.5

1.8 + 0.5

2.1 + 0.8

0.9 + 1.1

3.6 + 1.6

Suburban

14.8 + 4.9
(9973)
3.0 + 0.7
, (9972)
10.0 +1.9
(9979)
2.3 + 0.7
(9776)
8.2 + 4.2
(9773)
Urban
(SQ7 dominant,
-60 pgAr)

14.2 + 4.9
(9871)
3.8 + 0.3
(9778)
9.5 + 0.8
(9979)
1.6 + 0.4
(8672)
6.6 + 3.9
(9172)
Urban
(oxidant dominant
~40 ug/m )

21.0 + 6.2
(99.2)
6.5 + 5.6
(9473)
8.8 + 1.7
(9979)
1.7 + 0.4
(9176)
7.8 + 2.4
(99.7)
   Source:  Adapted from Yocom and Upham, 1977.
SX10AA/A
10-31
                                                                                       12-15-80

-------
environments.   Four  of the  Chicago sites  represented  downtown,  industrial,  commercial,  and
rural environments.  An  additional  site was 20 miles southwest of Chicago at Argonne National
Laboratory.
     Suspended  particulate  matter  in the  St.  Louis area was measured  using periodic 24-hour
high  volume air  samplers.   Sulfation  values were determined by  the  lead  peroxide  candle
method.   Monthly  dustfall  measurements were also  used.   At the Chicago sites,  a  high  volume
air sampler measured 24-hour total particulate matter 3 days per week and SOp was monitored by
bubbler type absorbers for 24 hours twice a week.
     Two fabric types were exposed in this study,  a desized and scoured cotton print cloth and
scoured cotton army duck.   Two samples were untreated,  two dyed, and one resin treated.   Study
results  indicate  that  there is  a  significant  relationship between  air  pollution  and  both
strength degradation  and  degree of  fabric  soiling.  As shown in  Figure 10-7,  high pollutant
                                      2                                                     3
levels (mean sulfation 5 mg SO-/100 cm /day and/or SOp  concentrations of 0.2 ppm or 520 ug/m )
reduced effective fabric strength  by one-sixth when compared with low pollution sites (0.5 mg
           2                                  3
SO-/100 cm /day and/or 0.02 ppm  or 60  ug/m  SOp  concentrations).   The  relationship between
suspended particulate matter and  fabric strength  degradation was  not as good as that for SOp.
No correlation  between dustfall and strength degradation/effective life was demonstrated and
biological  deterioration did not appear  to be a  major factor in this study  (Brysson  et al.
1967).
     In a  review  of the  Brysson  et  al.  (1967)   study,  Upham  and  Salvin  (1975)  report  a
correlation coefficient of 0.95 was obtained for breaking strength versus sulfation for cotton
duck  cloth.   The  correlation  coefficient  for the thinner  cotton print  was 0.96.   Of the
pollutants  measured,  SOp  was most  responsible for causing fabric damage  (Upham  and Salvin,
1975).
     Zeronian (1970) carried out  laboratory exposures  in which cotton  and  rayon fabrics were
exposed for seven days to clean air with and without 250 mg/m  (0.1  ppm) sulfur dioxide.  Both
controlled  environments  included  continuous exposure  to artificial  light   (xenon  arc) and a
water spray turned  on for  18 minutes  every  two  hours.   Loss  in  strength for  all  fabrics
exposed to  clean air averaged 13 percent, while the fabrics exposed  to sulfur dioxide averaged
21  percent.   Zeronian et  al.  (1971) also  exposed  fabrics  made  from  manmade fibers—nylon,
polyester,  and  modacrylic~to  controlled  conditions similar to the cotton exposures,  except
                                          o
that the sulfur dioxide level was  486 mg/m  (0.2 ppm).   They found that only the nylon fabrics
were affected,  losing  80 percent  of their strength when exposed to  sulfur dioxide and only 40
                                                                                     i
percent when exposed in clean air.                                                       i
     In a  study designed to determine effects of  air  pollution on  dye fading  on  fabrics, 67
dye-fabric  combinations  were tested.   The  test  samples  were exposed  in  the  dark  at  eleven
sites  representing  climatic  regimes  and  urban/rural  conditions.    High  temperatures  and
humidities  appeared to increase fading in the presence  of air pollutants.   Urban sites
                                                                                             \
SX10AA/A                                   10-32                                       12-15-

-------
      5.0
      4.0
E
o
§
o"
in
r  3.0
      2.0
      1.0
                              DUCK CLOTH: 12 months EXPOSURE
                              CORRELATION COEFFICIENT 0.95
                              SIGNIFICANT AT 0.5% LEVEL
                                            O -DUCK CLOTH

                                            A-PRINT CLOTH
           PRINT CLOTH: 5 monlta EXPOSURE '
           CORRELATION COEFFICIENT 0.96
           SIGNIFICANT AT 0.5% LEVEL
                  10
                             20
                                       30
                                                 40
                                                            50
                 BREAKING STRENGTH RETAINED, percent
 Figure 10-7a.  Relationship between retained breaking strength of
 cotton fabrics and corresponding mean su If at ion rate measured at
 selected sites in St. Louis area.
 Source: Bryssonet al. (1967).
     0.25
     0.20
 <  0.15
 £C


 I
 o  0.10
     0.05
                 I          T        I          I
                              • PRINT CLOTH, UNTREATED
                              Q PRINT CLOTH. DYED BLUE
                              O PRINT CLOTH. DYED ORANGE
                              A PRINT CLOTH. RESIN TESTED -
                  20        40        60        80

                BREAKING STRENGTH RETAINED, percent
                                                          100
Figure 10-7b.  Relationship between retained breaking strength of
cotton print cloth samples and mean SO2 concentration for 5-
month exposure at three Chicago sites.
Source: Brysson et al. (1967).
                       10-33

-------
produced significantly  higher  fading  than rural sites.  Fading was  highest at the site which
had the highest S02 and TSP concentrations (Beloin, 1972).
10.2.4  Building Materials
     The  deterioration  of  inorganic  building  materials   occurs  initially  through  surface
weathering.  Moisture  and salts are considered  to be the  most important  factors  in building
material damage.   Many researchers believe  that  the  mechanism  of damage  from  air pollution
involves the formation  of salts from reactions  in  the material.   Subsequently,  these surface
salts dissolve  in moist  air  and are  washed away  by  rainfall.   The  components  of inorganic
building materials can react with S02 and sulfates (Luckat,  1972;  Winkler,  1975;  Arnold et al.
1976).  Other researchers believe that the role of air pollutants  in  stone  and concrete damage
has frequently been  overestimated  (Riederer, 1974; Niesel,  1979).  These  authors  report that
inorganic  building  material   damage  is  dominantly  associated   with  RH  >65   percent  and
freeze/thaw weathering.   Some researchers indicate that microorganisms must also  be considered
in  order to quantify  damage  to building  materials due to ambient  pollutant  concentrations
(Winkler,  1966;  Riederer,  1974;  Krumbein  and  Lange,  1978;  Eckhardt,  1978; Hansen,  1980).
Sulfur  chemoautotrophs  are  well known for the  damage they  can cause  to inorganic materials.
These microorganisms  (e.g.,  Thiobacillus) convert reduced  forms  of  sulfur  to  sulfuric acid
(Anderson,   1978).   However,  the  relative  importance of biological,  chemical,  and physical
mechanisms  have   not been  systematically investigated.   Thus,   damage  functions  definitely
quantifying the relationship of pollutant concentrations to stone and concrete  deterioration
are not available in the literature.   Air  pollution damage  to  glass is  also  not presently
quantifiable (Newton, 1974).
10.2.4.1 ,  Stone—Niesel   (1979)  has  recently  completed a  literature review concerning  the
weathering  of  building  stone  in  atmospheres  containing  sulfur   oxides,   which  includes
references from 1700 to the present.   In brief, he reports  that weathering of porous building
stone containing  lime is  generally  characterized by accumulation  of  calcium sulfate dihydrate
in the near surface region.   The effect of atmospheric pollutants  on  the  rate of weathering is
believed  dominantly  controlled by  the  permeability  and   moisture   content of  the  stone.
Migrating  moisture  serves primarily as  a transport  medium.   SOg is  sorbed and  thus  can be
translocated internally while being oxidized to sulfates.  Reacting components of the building
stone are  thus  leached,  the  more soluble  compounds inward and  the  less  soluble toward the
surface, often forming a surface crust.
     Sengupta and DeGast  (1972) also report that S02 sorption causes  physical  changes in stone
involving  changes  in   porosity and  water  retention  characteristics.    Removal  of( calcium
carbonate  changes  the physical nature of  the  stone surface.  The hard, nonporous layer that
forms as a result of alternate freezing and thawing may blister,  exfoliate, and separate from
the surface.  If  the stone contains some  substances that  are unaffected  by  S02,  the surface
can deteriorate unevenly.  The conversion of calcium carbonate into calcium sulfate results in
a type of efflorescence termed crystallization spall ing.
SX10AA/A                                   10-34

-------
     Acidic  precipitation  also  contributes  to  the weathering  process of  building  ston<=r,
according  to Gauri  (1979).  He  reported that marble  that is directly exposed to  rainfall
undergoes  nearly continuous  erosion as  the  acid dissolves the  calcium carbonate,  allowing
calcite granules to break away and wash off.
10.2.4.2   Cement and Concrete—Portland  cement,  the  major active constituent of  concrete,  is
manufactured by  the  high-temperature reaction of a  mixture  of  limestone,  alumina,  silicates,
and iron salts  found in clay.  Cement, the binding agent in concrete, is an alkaline material
that reacts  with SOp and thus also  suffers erosion  and spalling effects;  it can be protected
by paint.  Concrete is also subject to damage by sodium sulfate.  Sulfate-resistant cement can
be prepared  by  reducing  the calcium  aluminate  content.   Concrete  can  also be  protected  by
nonporous paint.
     The chemical  action  of SO- or  sulfates  on  cement or concrete can be described as a dual
nature  mechanism.   Calcium hydroxide  in cement  and  concrete  can  be  converted  to calcium
sulfate,  which  reacts   to   form  calcium  sulfate  aluminum  hydrate   (ettringite),  with  a
substantial  increase in volume.  Cement for dams and culverts requires special formulation for
sulfate  resistance when exposed to  sulfate concentrations  > 200  ppm in water (Nriagu, 1978).
     Litvin  (1968) examined concrete samples containing  Portland  cement and marble aggregate
with sand at an  industrial  site in Buffington, Indiana.   Some changes were noted in the marble
aggregate,  but  a  more  observable  change  was found   in  the  cement portion.   Sealants were
evaluated  as  protective coatings;  their 'use  was accompanied  in  some   cases   by  surface
efflorescence.
10.2.5   Electrical Equipment  and Components
     Bobbins  (1970)  and  ITT Electro-Physics  Labs (1971) studied  the damaging  effects of S02
and particles  on electronic components and estimated  the cost of this  damage.   The report by
ITT Electro-Physics  Labs  considered  damage to 11 categories  of electronic components for which
a  literature  survey indicated  that  sulfur  dioxide  pollution  would be  mainly responsible.
However,  information gained directly  from manufacturers  indicated that particles  were the
major  factor in degradation and failure  of electronic  components and equipment.  (See Section
10.3.1)  Reduction of S02 and particulate concentrations  would have  little effect on costs for
the prevention of corrosion; low  concentrations of  pollutants would  still require  essentially
the same protective  measures.  Corrosion-resistant  metals are used even in environments where
air .pollution  is  minimal,   since  their  cost  is  far outweighed  by the expense  of equipment
failure.
10.2.6   Paper
     Modern  papers are  manufactured  from  cellulose.  On exposure to acids, paper is hydrolyzed
and  loses strength.   Spedding  et al.  (1971),  in work with radioactive labeling  techniques,
determined that  S02  is  readily absorbed by paper and oxidized to sulfuric acid by the metallic
impurities in  the  paper.  The reaction may also involve the  lignins in the paper, resulting in
the formation  of lignosulfonic acids.  Walsh et al.  (1977) showed that S02 is rapidly absorbed


SX10AA/A                                  10-35

-------
b\ uncoated wallpaper and less rapidly absorbed by vinyl-coated paper.   Although most paper is
used  in  objects  with a short service  life,  the preservation of documents has been of concern
in museums  and archives.   Coating paper  with polymers impervious to gases  is  an established
process.   The preservation  of  books  in  libraries is  of  major concern.   The  extent  of  the
problem  is  emphasized  by  the estimate that  50  percent of  the books printed  between  1900  and
1940  are  in  need  of  conservation.   The  New  York  Public  Library  conserves  books   by
microfilming,  lamination,  and electrostatic reproduction (Kingery, 1960).   The  library spent
$900,000  between 1952  and  1967  to  microfilm  books  that  had deteriorated  (Waddell,  1974).
10.2.7  Leather
     Leather  has a  high capacity for absorbing SO,,.   Spedding et al.  (1971) reported that  the
rate  of  SO-  diffusion  to  the leather surface  is  the controlling factor in  S00  uptake.   The
formation, in the presence of water,  of sulfuric acid is followed by hydrolysis  of the protein
(collagen)  of which  leather is  principally  composed.    This  weakening  of leather  causes
cracking and ultimately results in reduction of the leather to a red-brown powder (Spedding et
al.,  1971; Yocom and Grappone, 1976).
     The destruction of leather  by absorption of S02 has long been known and was described in
detail by Prenderleith  (1946).   The  buildup of sulfuric acid in aged leathers correlates with
deterioration, which can  be reduced  by  inactivating  the  sulfate ion  and by pH  buffering.
Deterioration  of  leather  is  important in bookbinding  and  in leather upholstery;  the  use of
artificial leathers has reduced damage costs.
10.2.8  Elastomers and Plastics
     The deterioration  of  natural  rubber  and synthetic elastomers under weathering conditions
has been  studied extensively.   Heat,  light, oxygen,  certain metallic ions,  and particularly
ozone cause deterioration; but  there is  no mention in the  literature of S02 damage to rubber.
In fact,  rubber  is  used as an acid-resistant coating.  The problem of determining ambient  air
pollution effects on  rubber  is  complicated by the presence of ozone,  which attacks the double
bonds in both natural rubber and the  butadiene-styrene and  butadiene-aery!onitrile synthetics.
     Haynie et al.  (1976) conducted a chamber  study on rubber  to determine the  effects of
ozone, S00,  and  nitrogen  dioxide  under   controlled  conditions of  temperature,  humidity,  and
                                                                                            3
light.  Exposures were made at concentrations of 0.1 and 1.0 ppm for each pollutant (in ug/m ;
262 and  2620  for S02,  196 and 1960  for Og and 188 and 1880 for N02).   As expected, ozone  was
responsible for accelerated  cracking of  the rubber.   Sulfur  dioxide  did not have any effect.
     Verdu  (1974)   presented  a  theoretical  study  of  the  effect  of  air  pollutants  on  the
weathering of  plastics.  He  attributed a  direct deteriorating effect on plastics to ozone  and
suggested  that  air pollutants  such  as   sulfur  dioxide may form  active  compounds  through
photochemical  reactions leading  to oxidation chain reactions.  In  light-exposure trials,  S02
increased the rate of degradation of polystyrene.
SX10AA/A                                   10-36                                       12-15-8J

-------
10.2.9  Works of Art
     Although works  of  art are composed of materials  already  discussed  in earlier sectior  -,*
this  chapter,  they  are  briefly treated here as  a  separate category because the  cost  of tnz
materials involved does not represent the cost of the item.
     The  deteriorating  effects  of  S02 and  particles  are well  known  to  museum  conservators
whose  function  is   to  preserve and  restore works  of  art.   The  rate  of  pollutant-related
deterioration has  increased markedly in the last 50 years.   The damage is striking in Europe,
where  ancient  buildings,  paintings,  frescoes,  stained glass windows, bronze  sculptures, and
marble statuary have suffered deterioration.
     Newton  (1974)   has  investigated  the  cause  of deterioration  of  medieval stained glass
windows.  He  found  that  the main cause of  decay is the leaching of potassium ions  from the
silicate glass  by  condensed water.   Another cause  is  S02,  which produces opaque  white  crusts
containing CaS04'H20 and syngenate (Ca-CaS04-K2S04-H20). The poor durability of medieval glass
is due to its high content  of alkaline earths such as lime and magnesia.
     Riederer (1974) conducted a study of the corrosion of bronze sculpture by air pollutants.
Sulfates were found  in the  corroded surfaces.
     The  dome   of  the  cathedral  in  Cologne,   located  in a highly  polluted urban area, has
suffered  serious  erosion  of  its sandstone  due to  the reaction of  sulfur acids  with calcium
carbonate to  form calcium sulfate,  which  is  leached  out by rain (Luckat,  1976).   Decay and
deterioration of the Taj Mahal  in India  has  also been attributed to sulfur dioxide and other
pollutants emitted from a  nearly petrochemical complex (Gajendragakar,  1977).
     Damage  occurring  in  Venice,   Florence,   Rome,  Athens,  London,  and  Cologne  has  been
attributed to the effect of S02 from industrial areas in these cities (Yocum and Upham,  1977).
The United States is also  concerned about the deterioration of public buildings and monuments.
The National Bureau  of Standards was asked by the National Park Service to investigate methods
for preservation  of stone after erosion was noted in the facade of the  Lincoln Memorial  in
Washington, DC  (SIeater, 1977).
     SI eater (1977)  investigated damage to stone  from the action of S02,  salt, sodium sulfate,
and  light.    Conservation  materials  including  epoxy  resins,  fluorosilicates,   and  silicone
resins  were evaluated.   The conservation  methods  recommended to  the National  Park Service
varied with the exposure conditions.
     The  damage  to  the   Acropolis  caused  by  S02  and  S03  has  resulted  in  a  massive
interdisciplinary effort by the Greek government to protect the ancient buildings  from further
deterioration (Yocom, 1979).
10.2.10  Review of Damage  Functions Relating SOo to Material Damage
     Even the most reliable damage functions must be used with caution.   Those listed in Table
10-6  were  selected on the  basis  of their  treatment  of independent  variables and their
inclusion  in  major  literature reviews.   Time-of-wetness  (often  expressed  as  RH  above  a
critical value) is the most important variable in these damage functions.
SX10AA/A
                                           10-37                                       12-15-80

-------
                           TABLE 10-6.  SELECTED PHYSICAL DAMAGE FUNCTIONS RELATED TO S02 EXPOSURE
           Material
                                 Reference
Dose-Response relationships
o

CO
00
       Zinc
                        Haynie and Upham, 1970   Y = 0.001028 (RH - 48.8) S0£
       Galvanized steel        Haynie, 1980
                                                 corr =  2.32 tw + 0.0134v0'781S00t
                                                               W                 c W
Oil-base house paint    Spence et al.,  1975      Y = 14.3 + 0.0151 SOy + 0.388 RH



Enameling steel         Haynie and Upham, 1974   corr = 325 t^ e(0<00275 S02 " 163-2/RH>



Weathering steel        Haynie et al.,  1976      corr = [5.64 VSOj + e(55'44 " 31>150/RT)-j ^~
       corr = corrosion, urn
       Y = corrosion/erosion rate, um/yr
       SOp ug/m                          oK\
       R - gas constant (1.98 cal/gm mol/  '
       RH = percent average annual relative humidity
                                                          t  = time-of-wetness in years
                                                          v  = wind velocity in m/s
                                                          T = °K
                                                          t = time of exposure, years
                                          0.92
       Galvanized steel        Haynie et al., 1976      corr =  (0.0187 S09 + e 41'85 " 23»240/RT)t           0.91
                                                                          £»                        W
                                                                                                              0.61
                                                                                                              0.91

-------
     Functions  for zinc  or galvanized steel  appear to  show  the best  fit,  followed by  t-.e
functions for oil-based  house  paint.   The field studies by Haynie and Upham (1970)  and Haynie
(1980) and chamber study by Haynie et al. (1976) incorporated critical  variables and pro/ided
relatively  reliable  damage functions  for  galvanized  steel.    The  functions  selected  for
weathering  and  enameling  steel  and  for  oil-based  paint  also  utilized  these  critical
environmental variables.
     When these  functions are  used to estimate damage, other factors must be considered, such
as the  amount of  exposed  (uncoated) metal,  the percentage of buildings with  oil-based (not
Latex) paint,  and  temperature  and humidity  variables  (sites in  the arid  southwest compared
with sites in the  relatively humid northeast).
10.3  PARTICULATE  MATTER
     A report  by  the  National  Academy of  Sciences (1977)  on airborne particles  notes that
deposition of dust and soot on building materials not only significantly reduces the esthetic
appeal of structures,  but also, either alone  or in conce'rt with other environmental factors,
results  in direct  chemical attack.  Because  of  the  paucity of data (see Chapter 5) regarding
total suspended particulate matter (TSP)  size distribution and composition, it is difficult to
determine the specific  types  of  particles  and chemical  constituents  that have  damaged  or
soiled a particular  structure.   As  will  be discussed  in  the  following  section,  chemical
composition  of  particulate matter is highly  important to its corrosiveness; particle size may
be important.   Since  information on both composition and size is spotty and incomplete, it is
difficult to  develop  physical  damage functions  associating  either TSP  or any particular size
function  with materials  damage or  soiling.   Research  is  under  way to address the  role  of
particle size in soiling  of paint.
10.3.1  Corrosion  and Erosion
     Early studies indicated that suspended particulate matter  played  a  significant role  in
metal  corrosion.   Sanyal  and  Singhania,  writing  in 1956,  termed  the  influence of suspended
particulate  matter  "profound."   They  ascribed  the corrosive effects  of  particles to  (1)
electrolytic, hygroscopic, and/or acidic properties and (2)  their  ability to  sorb corrosive
gases, e.g.,  sulfur dioxide.   It  has been  pointed out that it is  quite difficult  to predict
corrosion rates  separately for  sulfur dioxide  and particulate  matter  since  they  frequently
coexist  at high  levels (Chandler  and Kilcullen,  1968).   Other field studies have established
no conclusive correlation between total  suspended particulate matter and corrosion (Mansfeld,
1980;  Haynie and  Upham,  1974;  and  Upham,  1967), though  further analysis of  recent data is
still under way.
     At  an RH exceeding  60 percent, air polluted with S02 and PM results in a much more rapid
corrosion rate  than  air  polluted with S02 alone (Yocom and Grappone,  1976; Johnson et al.,
1977).   Kottori  (1980)  observed that zinc and galvanized steel corrosion rates appeared to be
related  to   the  sulfate  content  of  TSP.  Chloride  content of  dust  also may  contribute  to
accelerated corrosion of  steels (Gibbons, 1970; Bresle, 1976).


SX10AA/A                                   10-39                                       12-15-80

-------
     Barton  (1958)  reported  that dustfalls contribute to the  initial  stages of corrosion but
that  their  influence  becomes  less  important as  a  layer  of  rust  forms.   Two classes  of
particulate  matter appear  to be  definitely  associated  with  corrosion:    hygroscopic  salts
(including those of natural origin such as sodium chloride) and acid smut.
     A  review  of  atmospheric   factors  affecting  metal   corrosion provides  evidence  of  a
relationship between salinity and corrosion (Guttman and Sereda, 1968).
Corrosion  of metals can  be  accelerated by  deposition of particles due to their hygroscopic
nature.  The  influence  of hygroscopic substances on metal corrosion rates  has been previously
discussed in section 10.2.1.1.4.   As also discussed in section 10.2, particles can disrupt the
protective oxide films formed on metal surfaces such as nickel, copper,  aluminum and stainless
steel  resulting  in pitting  (Russell,  1976;  NAS,  1977).   Russell  (1976)  noted  that airborne
particles  often  play  an important role on the attack by sulfur oxides on electrical  contact
surfaces by acting as points for the concentration of active ionic species.
     Acid smut is highly corrosive, sticky material formed in and emitted mainly from furnaces
burning  liquid  fuels   containing  sulfur,  notably  in power  plants  (Ireland,   1968).   This
material  would  not usually  be  considered  suspended  particulate matter,  as  it  occurs  as
agglomerates  of  carbon, ash, and  sulfuric acid up to 0.5  cm  or more in  diameter which fall
close to the  source (Rowden,  1968; Potter, 1971).  According  to one report, acid smut mainly
settles out  within 400 meters  of the  source  under  conditions  of light  wind (Hoshizawa and
Koyata, 1970).
     Japanese investigators analyzed  a  large  (>10 drums [sic])  sample  of  acid smut and found
the sulfuric  acid  content to be 30 percent (Oyama et al., 1974).  Damage to painted surfaces,
automotive finishes, and  even agricultural  crops can be substantial.   As noted in a review of
residual oil  firing problems, "public reaction can be quite  severe"  (Exley, 1970).   A report
on the  status  of "public  nuisances" in the electric power industry of Japan reported progress
in determining the cause of acid smut and in  developing preventive techniques (Overseas Public
Nuisance Study Mission,  1965).
     Finishes on  automobiles parked  near  industrial  sites have often been severely damaged.
Staining and  even pitting of  auto finishes have been traced to  iron particles from nearby
industrial operations.   Cars parked near brick  buildings being demolished have been damaged by
alkali  mortar dust  during humid  weather.   Repainting  of  damaged auto  finishes  was  often
required  because color changes  were  not reversible by washing  or  polishing  (Fochtman and
Langer, 1957).
     Parker  (1955)  reported  that  large  numbers  of black specks collected on  freshly painted
buildings  in  industrial areas.   The  exterior  surfaces  of these  buildings became distinctly
soiled  and required  cleaning  or  repainting in 2  or 3 years,  depending  on  the particulate
concentration  in  the air.   When particulate  matter  became embedded in  the  paint  film, the
coating  was  both  esthetically  and  physically  damaged.    Embedding  of  particles  provides
nucleation sites at which other pollutants can  concentrate.   Cowling and Roberts (1954)


SX10AA/A                                   10-40                                       12-15-8d|

-------
suggest  that particles  promote the  chemical  deterioration of  paint by  acting  as wicks  to
transfer the S02 corrosive solution to the underlying surface.
10.3.2  Soiling and Discoloration
     Soiling  is the  accumulation  of  particulate matter on  the surface  of  a material.   As
defined by  Pfafflin  and Ziegler (1976), soiling  arises  from the deposition of particles less
than 10 urn-by impingement on surfaces  to  thus  mingle with settled dust.   Soiling  produces  a
change in reflection  from opaque materials and reduces light transmission through transparent
materials  (Beloin  and  Haynie,   1975;  MAS,  1977).   Soiling due to  airborne particles  from
manmade  sources results  in  increased  cleaning  costs  for building  and other materials  and
reduction in the useful life of fabrics.
10.3.2.1  Building Materials—Under  high  wind  conditions,  large  particles entrained  in  the
windstream  actually  result  in a slow erosion of surfaces similiar to sandblasting.   Particles
also  fill   surface pores of  many  sandstones   causing'them  to  become  uniformly  darkened.
Particles can contribute  to chemical  decay of  marble, limestone and dolomite stone work,  and
concrete  structures  if  they  carry  acids  and  soluble salts  (NAS,  1977).   Dose-response
relationships were developed  for suspended particulate matter and various building materials
by Beloin and Haynie (1975).
10.3.2.2  Fabrics—Although particulate matter obviously soils  fabrics, researchers have noted
that it is only damaging when the particles are highly abrasive and the fabrics are frequently
flexed.  Curtains  hanging in open windows, serving as  filters in polluted  areas,  provide a
good example.   Weakened as  a result of  such exposure,  curtains often split in parallel lines
along the folds.   The more tightly woven the cloth, the more resistant it is to soiling (NAS,
1977).
     Because  of soiling, fabrics  must be  washed more  often.   Excessive washing  may  reduce
fabric  strength,  leading  to  a  poorer  appearance  and  concomitantly  to  shortened  life
expectancy.   However, sunlight,  water vapor, SO , NO  and ozone concentrations are believed to
                                               /\    y\
affect the  service life of fabrics more significantly.   Insolation decoloration is considered
to  be  the  most important  service life  reduction  factor  (NAS,  1977).    Effects of elevated
sulfur oxide concentrations have been previously discussed.
10.3.2.3   Household  and Industrial Paints—Exterior paints can  be soiled by liquids  and by
solid  particles composed of  soot,  tarry  acids,  and various other  constituents.   Beloin  and
Haynie (1975) have compared the rates of  soiling by different levels of TSP on six different
building materials over a 2-year period.  The  mean annual PM concentrations at the five study
sites  ranged from 60  ug/m3 for a rural  residential  location  to 250 ug/m3 for  an  industrial
residential environment.  The exposed materials  included painted cedar siding, concrete block,
brick,  limestone,   asphalt   shingles,   and window  glass.   As  determined  by  reflectance
measurements, the  degree  of soiling  of  painted surfaces  was  directly proportional  to  the
square root of  the PM dose, accounting for 74  to 90 percent of the measured variability.  The
results of  regression  for soiling  of building materials  as  a function of TSP dose is included
in Table 10-7.   As one example, the linear regression results of soiling of acrylic emulsion

SX10AA/A                                   1Q-41

-------
               TABLE 10-7.   RESULTS OF REGRESSION FOR SOILING OF BUILDING MATERIALS AS A FUNCTION OF TSP DOSE
o
ro
            Material
Oil base paint
Tint base paint
Sheltered acrylic
emulsion paint
Acrylic emulsion
paint
Shingles
Concrete
Coated limestone
Uncoated limestone
Coated red brick
Uncoated red brick
Coated yellow brick
Uncoated yellow brick
Glass
400
400
400

720

48
160
80
80
80
80
80
80
45
89.43
86.13
91.54

90.79

43.50
41.75
44.57
46.99
12.95
14.88
45.05
43.21
0.2806
-0.2768
-0.2618
-0.593

-0.4131

-0.199
-0.0458
+0.0779
-0.0503
-0.0296
-0.0374
-0.1133
-0.1133
+0.0314
0.0641
0.0571
0.1156

0.0497

0.5771
0.1338
0.2464
0.1500
0.0223
0.0331
0.5337
0. 2740
0.008077
0.000069
0.000061
0.000123

0.000026

0.000258
0.000080
0.000164
0.000089
0.000013
0.000020
0.000317
0.000168
0.000007
7.6510
6.8265
13.8143

8.3791

7.6992
7.5011
6.9046
4.2035
0.6255
0.9274
14. 9533
7.6773
0.6851
0.745
0.738
0.880

0.902

0.769
0.143
0.347
0.266
0.459
0.477
0.342
0.503
0.340

         A,
         B,
Note:   Equation used in this regression analysis was reflectance = B(TSP x months of exposure/5 + A.
N,   Number of data sets (dependent upon the number of controlled variables in the factorial experiment).
     Intercept of linear regression.
     Slope of linear regression.
     Estimated variance of intercept.
     Estimated variance of slope.
     Residual variance (error).
R*",  Correlation index (fraction of variability accounted for by regression).
Source:    Abstracted from Beloin and Haynie, 1975.

-------
house paint  by  exposure to suspended particulate matter  is  shown  below:   (640 samples,  R2  =
0.896)
     ln(92.5   Y) = -0.311 + 0.345 ln(SP) + 0.612 ln(t)
     Where:  92.5 = Initial reflectance value (80 samples)
                Y = Measured percent reflectance (Photvolt Model  625)
               SP = Average TSP concentration (ug/m3)
                t = Exposure time (months)
Based on  this equation,  Figure 10-8 was  prepared  to summarize  soiling  of  acrylic  emulsion
house paint  as  a  function of  exposure  time and particulate concentrations.  Although  it  is
recognized  that  socio-economic  factors  control  ability and  motivation  to maintain  clean
surfaces  (Beloin  and Haynie  1975),  estimations of repainting frequencies  can be  formulated.
Assuming  an  individual  responds  to  a  defined change   in  reflectance  of  house  paint  by
repainting,  that  person will  repaint  a  house  twice as5  often  in an  environment  with  a TSP
concentration of  260 pg/m  in  comparison to  75 ug/m .   Specifically,  at a  35 percent change
this houseowner would  repaint every 4 years when  TSP  is 75 ug/m  and every  2 years when 260
ug/m .
10.4  SUMMARY, PHYSICAL EFFECTS OF SULFUR OXIDES AND PARTICULATE MATTER ON MATERIALS
     Reviewing the facts presented in Sections 10.2 and 10.3, it should be clear that the best
data base for association of sulfur oxides to materials effects is corrosion of metal and that
the parallel  case for particulate matter  is  soiling.   Of the damage  functions  developed for
corrosion of metals by sulfur oxides, the one for zinc appears to show the best fit.  Slightly
less well  established  are  relationships  for  sulfur  oxide exposure  and corrosion  of  other
metals.  There is  evidence  that  particulate  matter  aggravates  corrosion,  especially  when
coexistent with sulfur oxides.  This is most likely due to the hygroscopic salt content of the
particulate  matter.   However,  no mathematical  expression  of the  relative  contribution  of
particulate matter to corrosion of metal has been established.
     Both  particulate  matter  and sulfur  oxides  have  been  shown to  damage paint.   Damage
functions  for  erosion of  various  paint  by  sulfur  oxides  and  for  soiling of paint  by
parti cul ate matter  have  been developed.   The varying properties of the several types of paint
employed for  exposed materials  make it difficult to  construct  mathematical  expressions for
effect of particulate matter and/or S02 on all paints.
     Building  materials  also are  eroded by sulfur  oxides and soiled by particulate matter.
That  damage   is   attributable   to  both  pollutants  is  established;  quantitative  general
relationships  between  ambient concentration  and effect are, however,  lacking.   For erosion of
building  materials, particularly  stone  and  concrete,   the  contribution  of  sulfur  oxides
relative to other agents is not clear.   As for soiling,  though some damage functions have been
developed, the lack of  understanding  of the  role of  particle size and  composition makes it
difficult  to  generalize  about  soiling  effects of  all  suspended particulate matter on all
building materials.

SX10AA/A

-------
    45
    40
Ul

-------
     Concerning  effects  of particulate  matter  and/or sulfur oxides  on other materials,  the
data base  is even  less  well established.   There is evidence  to show that damage  occurs  to
fabrics, leather,  paper, glass,  and plastic, and  to works of  art composed of  one or  many
materials.    This evidence is mainly qualitative  and spotty.   For these materials,  there are
insufficient  data  to develop  reliable  estimates  of effects attributable to specific  ambient
pollutant concentrations.
10.5   ECONOMIC   DAMAGE  OF AIR  POLLUTION TO  MATERIALS—SULFUR  OXIDES  AND PARTICULATE  MATTER
10.5.1  Introduction
     The task of estimating  costs of air pollution effects on  materials involves  treatment of
many variables,   some of which are difficult to quantify and relate to other factors bearing on
costs.   Several  approaches to the problem  have been reported.   These are:  (1)  the physical
damage  function  approach,  (2)  the non-market approach, and (3)  the indirect market approach.
     The goal of all of the approaches is to estimate the total monetary damage  caused by air
pollution,  in this  instance, by particulate matter and/or sulfur oxides.   Monetary damage, as
defined by Maler and Wyzga (1976),  is  "the  total  compensating  variation of a change  that is
detrimental  to  the  environment."  Monetary damage comprises  two subsets of loss:   financial
loss is "that part of the monetary  damage  which  is realized  by changes in expenditure, while
amenity loss  is  the difference between monetary damage and financial loss."
     There are   several  types of financial  losses resulting  from damage  to  materials  and
soiling:   (1) reduced service  life  of  a material,  (2)  decreased utility of a material, (3)
necessity for use  of suitable substitute materials, (4)  losses due to an inferior substitute,
(5) protection  of  suspectible  materials,  and (6)  additional required  maintenance,  including
cleaning.   The major losses  of amenity, as defined by Maler and Wyzga,  are "those of enduring
and suffering soiled, damaged, or  inferior products and materials because of pollution,"  in
this case  particulate matter and/or sulfur oxides.   In addition,  amenity losses  are suffered
when pollution   damage  repair or maintenance procedures  results  in  inconvenience or other
delays  in  normal operations.   Some of these losses, such as effects on monuments  and works of
art, are especially difficult to quantify (Maler and Wyzga, 1976).
     In calculating monetary damage,  the  approach selected will vary  depending  upon  whether
financial   losses  or  losses of  amenity  are to  be  emphasized,  the  type of  damage  being
considered,  and  the  availability  of information.  Regarding pollutant effects on materials,
the literature  has  been  dominated  by  calculations  of financial  loss  based on  the physical
damage  function  approach.    Sources  of  error in  financial  loss functions using this approach
include the   following:   (1) present  techniques   do not reflect  the  recent introduction  of
resistant materials  which  last longer and require less maintenance; (2) estimates assume that
galvanized  steels  are  left uncoated;  (3)  many  materials  may  wear  out before they  are
significantly damaged by pollutants (Glass,  1978).
     One  area not  dominated  by  the  physical  damage  approach  has  been the  estimation  of
monetary  damage  associated  with soiling.   In  these  studies,   loss  of  amenity  has  been
considered,  since  socioeconomic  variables  are   heavily  involved.  The  approaches  eaployed

SX10AA/A                                   10-45                                        12-15-80

-------
reflect  the  shift in emphasis.   Non-market  and  indirect market approaches have been  used to
relate changes  in  air quality to changes in the amount of money people are willing to pay for
improvement in  air quality (reduction in air pollution).   Documentation  of this relationship
may be sought  through survey of affected individuals  or through development of relationships
between  environmental  quality and available data  on price differentials.  A major  source of
error  in these approaches  is that they demand  that all factors other than air quality that
affect cost must be accounted for.  It is also  very difficult to distinguish among  the effects
of many different air pollutants.
     Recently, the nonmarket and indirect market approaches have been increasingly  employed to
estimate both  financial  loss  and loss of amenity.   To  date,  such studies have not been found
more suitable  for regulatory  decision-making  guidance than the body  of  literature  resulting
from employment of the  more traditional  (physical damage  function)  approach.   All approaches
are limited by the difficulty in quantification of the human response to damage based upon the
ability and the incentive to pay additional  costs.   (Yocom and Grappone, 1976).
10.5.2  Economic Damage to Materials
10.5.2.1  Metals—Realistic estimation of the economic damage to metals attributable to sulfur
oxides and particles  must take into  account several  factors,  including avoidance  costs, such
as the  costs  of  specific protective treatment.   For metals,  these costs  include the  use of
anticorrosive  primers,  the practice  of  sandblasting before  painting, and the  use  of paints
that are resistant to acid.
     A recent  report  published by  the  U.S.  Department of  Commerce (Bennett  et al.,  1978)
examined the  cost of corrosion  in the  United States in 1975.   Unfortunately,  the  developed
damage costs were  not pollutant-specific and were not  associated  with ambient concentrations
of pollutants.  A total  annual  metallic corrosion  cost of $82 billion was estimated  with a
model  which  incorporated a broad  range  of  cost items  (e.g.,  materials,  labor,  energy, and
technical  capabilities).   About  40  percent of this cost,  or $33  billion,  was considered
avoidable.   Within this  avoidable  cost  is,  of  course, the cost of air pollution, a portion of
which  is  in  turn the  cost of  metallic corrosion  resulting  from particulate  matter  and/or
sulfur oxides.  The  figure of $33 billion is  thus  only useful as an  upper limit  for present
purposes.
     Fink  et  al.  (1971) estimated that  corrosion  of external metal  structures  caused  by air
pollution costs $1.45 billion annually  in the United  States,  as shown in Table 10-8.    As is
the case for the  extensive Bennett et al.  (1978) Department of Commerce report, these studies
also were not specific to single pollutants nor were the damage costs associated directlyjwith
ambient pollutant  concentrations.  Furthermore,  in some cases  material damage  resulting from
causes other than  air pollution was  included  in cost estimates (e.g., the Fink et  al.  study
included corrosion inside  pipes  of  industrial  systems).   On  the  other  hand,  Haynie  (1974)
noted  that within  Fink's  estimate  of  metal  corrosion  costs,  damage to  structural  systems
primarily  constructed  of  galvanized  steel  accounted  for more than  90 percent of  the cost.
Haynie  reasoned that,  based  on  Fink's  data,  the  accelerated  corrosion  of zinc  by sulfur
dioxide accounts for more than 90 percent of corrosion caused by air pollutants.
SX10AA/A                                   10-46                                       12-15-80

-------
         TABLE 10-8.  SUMMATION OF ANNUAL EXTRA LOSSES DUE TO CORROSION DAMAGE BY AIR POLLUTION
                                     TO EXTERNAL METAL STRUCTURES FOR 1970
Steel system or structure
Basis for calculation
    Annual
loss in $1000
Steel storage tanks
Highway and rail bridges
Power transformers
Street lighting fixtures
Outdoor metal work
Pole-line hardware
Chain-link fencing
Galvanized wire and rope
Transmission towers
     Maintenance
     Maintenance
     Maintenance
     Maintenance
     Maintenance
     Replacement
Maintenance and replacement
     Replacement
     Maintenance
  $   46,310
      30,400
       7,450
      11,910
     914,015
     161,000
     165,800
     111,800
       1,480
  $1,450,165
 Source:   Fink et al. (1971).

-------
     The economic  damage  from corrosion can also be  estimated  by determining the annual  cost
of industrial paints used for corrosion control.   This cost is estimated by totaling the costs
of  surface preparation,  paint,  and  labor.   Coatings  applied  at this  combined  cost would
protect metal structures  against corrosion for about 10 years,  whereas "ordinary" coatings of
paint would provide  protection  for only 2 years.   Painting of structural  steel in bridges was
investigated  by Moore and  O'Leary;  (1975) the  practice  involves  sandblasting  the  steel  to
produce a rust-free surface and to remove mill  scale.   Without such surface preparation, water
is immediately absorbed and sets up a corrosion system,  rusting  occurs, and the  paint surface
deteriorates  in  2 to  3  years.   The metal  surface  is protected by primer  that  inhibits  rust
formation, and the primer coat is covered with  two coats of SQp-resistant paint,  such as vinyl
resin,  which is  substantially  more  expensive  than household  paint.   Estimates  have  been
developed  by  Banov  (1973),  Michelson  and   Tourin  (1967),  and others  to  reflect  these
preparation and painting  costs  for protection  of metal  structures.  However,  sound fractional
allocation of these  costs  to sulfur oxide material  damage is not available in the literature.
     Damage  from  pollutants  during  the  manufacture  of  electrical  components   must  be
controlled  even in clean  environments.   To prevent  such problems,  parts  are  fabricated  in
"clean  rooms"  with  filtered air.   If  equipment  malfunctions   in the plant or  in  service,
additional maintenance costs are'  incurred for cleaning,  repair,  or  replacement of defective
equipment.   ITT  (1971) estimated  $15.5  million per  year  in added costs for  clean rooms and
maintenance.   Robbins  (1970)  conducted  a survey  of the  effects of S0? and  particles  on
electrical   contacts  such  as  in  switches, relays,  connectors,  and  computers.   To  reduce
corrosion, contacts are electroplated  with corrosion-resistant  metals such as gold, platinum,
palladium, and  silver.   Less expensive  metals are  susceptible  to corrosion  failure,  mostly
from the action of SOp and HpS.    Robbins  estimated  that  15 percent of the gold and platinum
used in  the United  States  for electrical  contacts  in 1970 was for  the specific  purpose of
combating SO^ corrosion, with the  remainder going for protection  against other  environmental
pollutants.   However, protection against all deleterious environmental conditions is routinely
provided in  clean environments, due to  high costs  of replacement  associated  with  electrical
system  downtime.   Thus,  costs  for  protection  of  electrical  hardware  cannot  be  allocated
directly to ambient sulfur oxide and/or particulate  matter concentrations.
10.5.2.2  Paints—Spence and  Haynie  (1972) presented a survey and economic assessment of the
deterioration of exterior  paints  ("trade paints") caused by  air  pollution.  Included in this
category were both oil-base  paints and latex  paints containing polyvinyl  acetate-acrylic as
the binder.   The  total annual economic  damage to exterior household  paints was estimated at
$540 million (1972 prices), including paint loss and a labor factor of three times the cost of
                                                                                      /
the paint.   Unfortunately,  damage  functions  were  not  available  for  use  in  developing  this
estimate.   Furthermore, the  estimated  damage cost reflects  the  impacts  of all pollutants; no
specific attribution of costs to particulate matter  or sulfur oxides is possible.
SX10AA/A                                   10-48                                       12-15-80

-------
     Salmon (1970)  estimated  that the annual cost of  soiling  of  household paint would be S35
billion  if surfaces  were  maintained as clean  as  they are  in a  clean  environment  in  this
qualitative study.  The annual cost of deterioration damage to paints was  estimated to be SI.2
billion.
     Michelson and  Tourin (1967) investigated the frequency of house repainting as a function
of  suspended  particulate  concentration.   Questionnaires  were  sent  to  residents of  three
suburbs of Washington,  DC (Suitland, Rockville, and Fairfax) and two cities in the upper Ohio
Valley  (Steubenville and Uniontown).   Data  were  compiled  from  the  questionnaires  to  show
maintenance intervals for exterior repainting in each of the five communities,  but paint types
were not  reported.   In  Steubenville, where the  mean  annual  particulate concentration was 235
ug/m ,  repainting occurred about every  year.   In Fairfax, where the mean annual  particulate
concentration  was 60 ug/m  ,  repainting  occurred every 4 years.  Thus, maintenance frequency
increased as  particulate  concentration increased.   The results of  this  investigation suggest
that  a  significant  relationship exists  between  frequency  of  repainting  and  particulate
concentration.  However,  to establish a more definite correlation, additional  maintenance data
are needed, particularly  for cities with mean  annual  particulate concentrations greater than
         2
150 M9/ra •   Any  correlation  of frequency  of   repainting  with concentration  of  particulate
matter  must  take into  account  the  fact that  other pollutants  are  usually  present  in  high
concentration  where  particle  counts  are  high.  Socioeconomic factors also must be considered.
     Booz, Allen and  Hamilton  (1970),  in a study conducted  for  EPA,  reported  on painting
maintenance frequencies  in  several  zones of the Philadelphia metropolitan area with different
population  characteristics,  climates,  and  types  of industry.   Socioeconomic factors  were
delineated by pollution  zone;  however,  paint  types were not reported.   The  percentage  of
households with  incomes  of less  than $6000 increased with pollution level, a finding that may
partially  explain why  there  was no  statistically  significant correlation between painting
frequency and  particulate level.
10.5.2.3  Economic Cost of Soiling—Studies on soiling have in some approaches  been limited to
household cleaning  costs.   Other studies employ differences in property values or willingness
to pay  as  the basis for  cost  estimates.   In the study by Booz, Allen and Hamilton (1970),  $5
billion  for  annual   household  cleaning costs was  attributed to  particulate pollution.   This
amount  did not   include  laundering,  dry cleaning,  and  personal  care  items  (i.e.,  face and
hair).  The report  has  been criticized because  it  omits  some sources of damage that might be
responsible  for  extra  maintenance.    There  is   doubt as  to  which  cleaning   and  maintenance
operations are really  sensitive to air particulate  levels.   Also,  the  labor of the homemaker
was not  added  to the cost of  the cleaning  materials, and the  number of  urban homes surveyed
was small.
     Watson and  Jaksch  (1978) estimated  physical soiling  and frequency-of-cleaning functions
using the  study  of  Beloin  and  Haynie (1975) to establish the relationship between level  of
particulate matter and soiling rate and the work of Esmen (1973) to estimate various levels of
SX10AA/A
                                           10-49                                       12-15-80

-------
particulate  soiling  for rural,  suburban, and industrial locations.  Watson  and Jaksch (1978)
reported  that  the  frequency of  the low-cost  cleaning operations depended  on the  level  of
airborne  particles  but that  the  frequency of  professional  house painting was  unaffected  by
variation in airborne particle levels.
     Watson and Jaksch  (1978) concluded that 537 to 3816 million dollars (1971 dollars) can  be
saved per year in the United States  in  the  form of net welfare benefits if the secondary TSP
standard of  60 mg/m   can be met.   Watson and Jaksch (1978) utilized data from a review report
by  Waddell  (1974).    Waddell  utilized studies  that were not  specific to particulate matter
pollution and  do  not contain dependable material exposure  allocations  (e.g.,  Gillette, 1975).
Further, Watson and  Jaksch  (1978)  determined frequency-of-cleaning functions  from  the Booz,
Allen and Hamilton study  (1970),  which surveyed 1090 households  in  Philadelphia during 1969.
     The  Beaver Report  (1954)  suggested  an annual  total  for damage  by all  forms  of air
pollution in Great Britain of 152 million pounds sterling in direct costs, of which 25 million
was  for laundry, 30 million for  painting and  decorating,  and 20 million  for  cleaning and
depreciation of buildings other  than houses;  thus, about half the total cost of pollution was
attributed to soiling.
     Michelson  and Tourin (1967)  compared the  costs  of air pollution in  the  highly polluted
area of Steubenville, OH, with  those in  the relatively clean Uniontown, PA,  area.   The per
capita  costs  for inside and  outside maintenance of houses (painting  and cleaning), laundry,
dry  cleaning,  and personal   care  (hair and facial) were $84  higher   in  Steubenville than  in
Uniontown.   Cost figures were based on data obtained from questionnaires, but the results have
been questioned on the  grounds  that socioeconomic factors  influenced  responses and that there
were insufficient statistically reliable data.
     Narayan  and  Lancaster  (1973)   conducted a questionnaire  survey  in a  rural area  and  a
polluted area  in  New South  Wales,  Australia,  to determine  the difference in cost of household
upkeep.   The cost of maintaining  a house in the polluted Mayfield area was about $90 per year
higher than in the relatively unpolluted Rotar area.  This  cost differential  was attributed  to
the  higher levels of air pollution and airborne particulate  matter  in Mayfield; however, the
accuracy of the cost data was considered questionable,  since  the attitude of the respondents
could have introduced a biased point of view.
     Waddell   (1974),  in his  review of the economic  damage  of air pollution,  noted that the
Michelson and Tourin  (1967)  and Booz, Allen and  Hamilton (1970) studies considered principally
the  costs of  household  cleaning and maintenance.  He questioned the validity of extrapolating
the  Michelson  and Tourin value  of $84 per  capita because  of insufficient  information and
socioeconomic problems  in obtaining defensible  data.   He further  asserted that "intuitively,
other than what is  implicitly  measured in property value differentials, it  is difficult  to
conclude  that there  are not significant  soiling-related costs."   In particular,  he stated
that soiling costs should include  laundering, dry  cleaning,  hair and  facial  care, washing  of
automobiles,  and costs  of cleaning commercial establishments  and  public structures.  Waddell
(1974) concluded that there  are insufficient data to assess these soiling costs.
                                                                                             I
SX10AA/A                                   10-50                                       12-15-^0

-------
     Waddell  (1974)  also reviewed the concept of  property  value  as an estimator, noting  f,e
assumption that the inconvenience of living in a  polluted area  with  soiling  and  odors  leads to
lower property  values.   He  cited the Jacksch and Stoevener (1970) study  in Toledo,  Oregon,
using dustfall measurements  as  the variable.   Their hypothesis was that  air pollution costs,
though not  quantitatively known,  were  reflected  in the value of property.  They found  that
reduction in  property  values from increasing air  pollution was greater  in  the  higher priced,
newer sections of town than in low-cost housing.
     Most property  value approaches  assume  a direct and equal relationship  between  property
values and  air  quality.   In other words, "if two sites are  similar  in all respects except air
quality,   the  difference  in their  values represents  the  market's  willingness to  pay  for
reduction in  air  pollution dosages.   That is, all  air  pollution damages  will be registered on
differential  site values  (Crocker,  1970).   However, using  areas with similar  socioeconomlc
characteristics  Gillies  and  Youston (1976)  found  no  -statistical  relationship between  air
pollution and property values.
     In  a  mere recent  study of  paired  neighborhoods  in the Los Angeles area  employing  both
differences in property value and a willingness to pay  survey,  Brookshire et al.  (1979a,b) found
that the marginal willingness to pay for an approximate 30  percent  improvement in air quality
calculated  from  property  values and  the  survey  results  ($950  million  and $650  million
respectively) were  within a  factor  of  1.5.   The  study  was, however,  not pollutant-specific,
and the  authors  noted that this  approach  to  benefit analysis  needs further refinement before
the  results "can  even  cautiously be  applied to  environmental  decision  making."  They  also
cautioned that the study may well not be generalizeable to other situations.
     The economic effects  of soiling were examined by  Liu and Yu (1976)  in a Midwest Research
Institute project undertaken for the EPA's Corvallis Environmental Research Laboratory.   The
objectives  were  to  generate physical and  economic  damage functions,  by  receptor, for SOg and
suspended  particles and  to  establish  cost/benefit relationships.   The  study  included  the
effects  of air  pollution  damage on  health, household  soiling, materials, and vegetation.
     Liu  and Yu  (1976)  used the  data  of  Booz,   Allen and  Hamilton  (1970) on soiling and
maintenance operations to  identify and quantify soiling damage.  They  surveyed costs of nine
professional  cleaning companies  in  Kansas  City,   and  developed  a methodology  for  use  on a
nationwide  basis  for  estimating  damage attributable   to  S02  and  particulate  matter.   They
calculated  the  gross cost of cleaning  and  the  net (extra) cost from increases in  suspended
particles  using  a  formula  they  derived.    For  these  calculations,  they  chose a  suspended
particulate level of 45 ug/m3 as the level  necessitating extra cleaning.
     Liu and Yu (1976) derived net and gross household  cleaning costs  for 65 large cities with
populations  greater than  500,000  and for 83 medium-size  cities of  200,000 to  500,000.   The
extra costs for Chicago, New York, and  Los  Angeles in 1970 were $516 million,  $418 million,
and $388 million, respectively.   Per capita costs for  household maintenance activities ranged
from $5  per person in San Antonio, Texas, to $104 per person  in Cleveland, Ohio.   The total
net soilage cost  for the urban areas was estimated as $5.033 billion.
SX10AA/A
                                           10-51                                       12-15-80

-------
10.5.2.4   Combined Studies—Salmon (1970)  was one  of  the  first to  provide an  estimate  of
material  damage  associated with  ambient SOp concentrations.   He calculated  economic  loss  by
determining the values of materials exposed to pollution and then multiplying by the estimated
difference  in useful  lifetime between  clean-rural  and polluted-urban  areas.   The value  of
exposed materials  was  derived by multiplying annual  production by a product lifetime estimate
and  then  applying  a  labor-factor estimate  and  an  exposure-factor estimate.   Salmon  cited
economic  damage from  sulfur oxides in the United States to the following materials, listed in
decreasing  order  of  the  extent  of  damage:   metals,  cotton, finishes  and other coatings,
building  stone, paints, paper, and leather.  Paint,  zinc, and cement/concrete accounted for 70
percent ($2.647  billion) of the  estimated  annual economic loss of all  major materials  ($3.8
billion).   Deterioration of  paints alone was estimated at $1.2 billion.   Salmon noted that it
would cost $35 billion a year to keep household painted surfaces as clean as they would  be in
an  unpolluted environment.   Carbon  steel,  which  has  virtually no  resistance  to SQy  and
sulfate, accounted for $54 million of the total loss.
     Gillette  (1975)   reported  significant reductions  in economic  damage to materials  from
sulfur  oxides attributable  to improvements  in existing  air  quality levels throughout  the
United States.  Comparing  annual  S02  concentrations  from more than  200  monitoring sites with
the  estimated inventory of  materials exposed in the proximity of these  sites,  he estimated
that nationwide material  damage  decreased  from more  than $900  million  in 1968 to less than
$100  million  in  1972.   These estimates  were derived  by carefully  distinguishing  between
physical  and  economic losses  and by attributing  current  estimates  of  losses   to  current
exposure levels.
     The  distinction  between physical  and economic  damage  to materials  was based upon  the
concept of  normal  or  economic  useful life  of materials.  Whereas physical  deterioration  to
materials  may occur at relatively low exposure levels, economic losses will occur only if the
material requires  early replacement or increased maintenance  before its normal or economically
useful life is spent.   Given the prevailing ambient  concentrations observed, Gillette reported
that most materials were not adversely affected economically except for metallic products that
were subjected to  corrosion  or paint damage.  While  material  losses were much  greater during
the  early 1960s,  the  losses in  more recent  years  are substantially lower and  reflect  the
considerable improvement in air quality (Gillette, 1975).
     Waddell  assumed that  the Gillette study is more  defensible  than  the Fink study,  that it
does  not   significantly overlap  the  Spence-Haynie   study,  and  that  it  includes  damage  to
electrical  contacts  and  components.    If  the  materials  evaluated  in  these studies  (zinc,
paints, rubber,%carbon and alloy steel, fibers, cement and concrete, plastics, building brick,
paper,  leather,  wood,  and  building  stone)  plus  those believed by  Fink  to  be  relatively
unaffected by air  pollution  (aluminum, copper, stainless steel, and lead) are subtracted from
the  Salmon study,  a total  remainder from the Salmon study of $0.4 billion is obtained.   Using
these data and the assumptions outlined above, Waddell derived an estimate of 1970 air
SX10AA/A                                   10-52                                       12-15-

-------
                    TABLE 10-9.  ESTIMATES OF TOTAL COSTS FROM AIR POLLUTION
                                   DAMAGE TO MATERIALS IN 1970
Source of
information
Mueller-Stickney
Gillette
Material
category
Elastomers
General materials
Principal
pollutants
Ozone
S00
Estimated cost,
$ billion
0.5
0.4
         Salvin
         Spence-Haynie
         Salmon


         Total
but principally
metal corrosion
Textiles and dyes
Paints
Remainder of
important materials
not accounted for
above
Particulates, S0?
Various, but
probably
mostly S02
0.2
0.7
0.4



2.2
         Source:  Waddell (1974).•

pollution  damage costs  of  $2.2 billion.   As  shown in Table 10-9, $1.5  billion  was directly
attributed to PM and SO  .
     Yocum and  Grappone  (1976) also estimated that the economic costs of air pollution damage
to  materials for  the United States  was  about  $2.2 billion  in  1970.   They estimated  the
relative  contribution  to  this  total  cost  from  SOX  at  41  percent  (about $0.9  billion);
particulate  matter,  27  percent  (about  $0.6 billion);  ozone,  26 percent; and  NOX>  6 percent.
10.5.3  Summary of Economic Damage of Air Pollution to Materials
     The damaging and soiling of materials by airborne pollutants have an economic impact,  but
this impact is difficult to quantify.  The accuracy of economic damage functions is limited by
several factors.  One  of the problems has been to separate costs related to sulfur oxides and
particles  from  those  related  to  other  pollutants,  as  well  as  those  related  to  normal
maintenance.   Cost  studies typically  involve  broad assumptions about the  kinds  of materials
that are exposed in a given area and then require complex statistical analysis to account for
a selected number  of variables.   Attitudes regarding maintenance may vary culturally, further
confounding the problem  of quantifying economic impact.
     Studies  have  used  various approaches  to determine  pollutant-related  costs  for  extra
cleaning,  early replacement,  more  frequent painting, and  protective coating  of susceptible
materials, as  well  as other  indicators "of the  adverse economic  effects of  pollutants.   Ho
SX10AA/A
                                           10-53
                                                          12-15-80

-------
study  has  produced completely  satisfactory  results, and  estimates  of cost vary  widely (see
selected examples  in Figure  10-9).   Studies of  household maintenance costs  in polluted and
unpolluted areas  have shown  annual  cost differentials  of $100 or  more,  but the results  of
these  various  studies   are   difficult  to  compare  because  of   differences  in  approach,
methodology,   and   selection   of  variables.    Damage  functions  indicate  that  reductions  in
pollutants will  decrease physical  and  therefore  economic  damage,  but  the data base  and
methodology for attribution of costs to sulfur oxides and particulate matter are  incomplete at
this time.
10.6  SUMMARY AND  CONCLUSIONS, EFFECTS ON MATERIALS
     The nature and extent of damage to materials  by sulfur oxides  and particulate  matter have
been  investigated  by field  and  laboratory  studies.    Both physical  and  economic  damage
functions have been  developed for  specific  damage/effect parameters  associated with exposure
to  these pollutants.  To date,  only  a  few of  these functions  are relatively  reliable  in
determining damage, while none have been generally accepted for  estimating  costs.
     The best documented  and most significant damage from sulfur oxides  and particulate matter
is  the acceleration  of   metal  corrosion, the  erosion and  soiling   of paint  and  soiling  of
buildings and other  structures.  Erosion  of building materials  and stone  due to  sulfur oxides
is  also  established, but the importance of sulfur  oxides  relative to other agents  is not
clear.    Although   evidence of   damage  to  fibers  (cotton  and  nylon),  paper,  leather  and
electrical  components has been reported, reliable  damage estimates  have  not.
     Relatively accurate  physical  damage functions  have been calculated  for the effects  of
sulfur dioxide on  the corrosion  of galvanized steel.   Determination  of  variables  such as time
of wetness and surface configuration affect the applicability of the functions.   Similar, but
less accurate, functions have also  been  developed  for estimating  erosion  rates of oil-based
paints from  exposure to sulfur  dioxide.   The  least  reliable  of  the "significant"  damage
functions are  those for  soiling from particulate  matter.   The poorly understood deposition
rates  and  poorly  characterized  chemical  and physical properties make general application of
the functions difficult,  if not impossible.
     The limitations of  these and other physical damage functions hinder accurate estimates of
total  material damage and  soiling.   Coupled with  these limitations is the lack  of material
exposure estimates.  These  problems  presently preclude complete and accurate estimates of the
costs  of damage  based   on  a physical  damage  function approach.    Other  approaches  to cost
estimation with  lesser   data  requirements have been  attempted,  but  the  studies  using these
approaches to date have  been inadequate for decision-making guidance.
SX10AA/A                                   10-54                                       12-15-^0

-------
POLLUTANTS
MOST
POLLUTANTS
SOX AND TSP
TSP
SOX

1.45
FINK
(1971)



METALS
0.54
SPENCE AND HAYNIE
(1972)

1.2
SALMON
H970)

PAINT
2.2
WADDELL (1974) AND
YOCOM AND GRAPPONE
(1976)
1J5
WADDELL (1974) AND
YOCOM AND GRAPPONE
(1976)
5.0
LIU AND YU (1976) AND
BOOZ-ALLEN (1970)
0.54 - 3.8
WATSON AND JACKSCH
(1978)
0.1-0.9 3.8
GILLETTE SALMON
(1975) (1970)
MOST MATERIALS
                              MATERIAL TYPE
Figure 10-9. Summary of economic damage estimates of air pollution to materials.
(in billions of dollars per year, early 1970's)
                                       10-55

-------
10.6  REFERENCES

Abe,  H.,  Y.  Ishii, and H.  Kato.   Evaluation of atmospheric  factors  by analyses of corrosion
     products and surface deposits on copper plates.  Rail. Tech. Res.  Inst. 12:170-174, 1971.

Anderson,  J.  W.   Sulphur  in  Biology.   Studies  in Biology  No.  101,  University  Park Press,
     Baltimore, MD, 1978.   pp.  17-19.

Arnold,  L.,  D.  B.  Honeyborne,  and  C.  A.  Price.   Conservation of  natural  stone.   Chem.   Ind.
     (London) 8:345-347, 1976.

Banov, A.  Paintings and Coatings.  Structures Publishing Co., Farmington, MI, 1973.

Barton, J. R.   Protection Against Atmospheric Corrosion,  Theories  and Methods.   John Wiley &
     Sons, New York, NY, 1976.

Barton,  K.    The   influence  of  dust on  atmospheric  corrosion of  metals.   Werkst.  Korros.
     8/9:547-549,  1958.

Beaver, H., Chairman.   Committee  on Air Pollution:  Report.  Her Majesty's Stationery Office,
     London,  England, 1954.

Beloin, N. J.   Fading  of dyed fabrics .by  air pollution.  Text. Chem.  Color. 4:77-82, 1972.
  -.•
Beloin,  N.  J., and F. H.  Haynie.   Soiling  of  building  materials.   J. Air  Pollut.  Control
     Assoc.  25:399-403, 1975.

Bennett,  L.  H.,  J. Kruger,  R.   L.  Parker,  E.  Passaglia,  C.  Reimann, A. W.  Ruff,  and H.
     Yakowitz.   Economic  Effects  of Metallic  Corrosion  in  the United States.   Part  I:  A
     Report to  the Congress by  the  National  Bureau of  Standards.    NBS  Special   Publication
     511-1, U.  S.  Department  of  Commerce,  National Bureau of  Standards,  Washington,  DC, May
     1978.

Bird, C. E.   Materials Protection and Performance,  16, April 1977.

Booz, Allen  and Hamilton, Inc.   Study  to  Determine Residential Soiling Costs of  Particulate
     Air Pollution.  APTD-0715, U.S. Department of Health, Education and Welfare, National Air
     Pollution Control  Administration, Raleigh, NC, October 1970.

Brauns, E., and U. Kalla.   Korrosion 17:81,  1965.

Bresle, A.   The corrosion of steel  and the dangerous chlorides.  Met.  Finish.  74:23-25, 30,
     1976.                                                                       ~

Brookshire, D.  S., T.  D.  Crocker,  and  R.  C.  D'Arge.  Methods  Development for Assessing Air
     Pollution  Control  Benefits.   Vol.  V.   EPA-600/5-79-001e,  U.S.  Environmental Protection
     Agency,  Washington, DC, February, 1979a.

Brookshire,  D.  S.,  R. C.  D'Arge,  W.   D.  Schulze.   Methods  Development  for  Assessing Air
     Pollution  Control  Benefits.   Vol.  II:   Experiments  on Valuing  Non-Market Goods:  fy  case
     Study of  Alternative Benefit  Measures of Air Pollution Control  in  the  South Coast Air
     Basin of  Southern California.  EPA-600/5-79-001b, U.S.  Environmental  Protection Agency,
     Washington, DC, 1979b.

Brysson,  R. S., B.  J.  Trask,  J.  B.  Upham,  and S.  G. Booras.  The effects of  air pollution on
     exposed cotton fabrics.  J.  Air Pollut. Control Assoc. 17:294-298,  1967.

Campbell, G.  G.,  G.  G. Shurr,  D.  E.  Slawikowski,  and J.   W.  Spence.   Assessing air pollution
     damage to coatings.  J. Paint Techno!.  46:59-71, 1974.
                                                                                              i
SOX10A/D                                  10-56                                        12-15-86

-------
Campbell, G.  C. , G.  G.  Schurr,  and D.  E.  Slawlkowski .   A  Study  to Evaluate  Techniques -..*
     Assessing Air  Pollution Damage to  Paints.   EPA-R3-73-040,  U.S.  Environmental Protect! c-
     Agency, Research Triangle Park, NC, May 1972.

Cavender, J.  H. ,  W.  M.  Cox,  M. Georgevitch,  N.  A. Huey,  G.  A.  Jutze,  and  C.  E.  Zimmer.
     Interstate  Surveillance  Project:   Measurement of Air  Pollution Using  Static Monitors.
     APTD-0666,  U.S.  Environmental  Protection  Agency,  Research Triangle  Park,  NC, May  1971.

Chandler, K.  A., and M. B.  Kilcullen.   Survey of corrosion  and atmospheric  pollution in and
     around Sheffield.  Br.  Corros.  J.  3:80-84, 1968.

Cowling,  J.   E. ,   and  M.  E.   Roberts.   Paints,  varnishes,  enamels,  and   lacquers.  In:
     Deterioration  of Materials-Causes  and Preventive Techniques.   G. A. Greathouse and  C. J.
     Wessel, eds. ,  Reinhold  Publishing  Corp., New York, NY, 1954.  pp. 596-645.

Crocker,  T.  D.   Urban  Air  Pollution Damage  Functions:   Theory and  Measurement.  California
     University, Riverside,  CA, 1970.

Duncan, J.  R. ,  and D. J.  Spedding.   The mode of initial reaction  of SO, at a  metal  surface.
     Corros. Sci. 14:241-249,  1974.                                      *

Eckhardt, F. G. W.  Microorganisms and  weathering of a sandstone monument.  In:   Environmental
     Biogeochemistry  and Geochemistry Volume 2:  The Terrestrial  Environment,  Proceedings of
     the  Third  International   Symposium, Ministry  of  Sciences and Arts  of  Niedersachsen,
     Deutsche Forschungsgemeinschaft, Wolfenburetell , Germany, 1977.  W. E. Krumbein,  ed. , Ann
     Arbor Science  Publishers, Inc., Ann Arbor, MI, 1978.  pp. 675-686.

Esmen,  N. A.   A  direct measurement  method for  dustfall.    J.  Air  Pollut.  Control. Assoc.
     23:34-36, 1973.

Evans,  U.R.   Mechanism of rusting under  different  conditions.   Br.  Corros. J.  7:10-14,  1972.

Ergang,   R. ,   and   M.   B.  Rockel.    Corrosion  resistance  of   stainless   steels  in  the
     atmosphere —  evaluation  of  the results of weathering tests of  up to 10 years duration.
     Werkst. Korros.  26:36,  1975.

Exley,  L.  M.   A practical  review of residual oil  firing problems  and solutions.  Combustion
     41:16-23, 1970.

Fink,  F.  W. ,  F. H.  Buttner, and W.  K.   Boyd.   Technical-Economic  Evaluation of  Air-Pollution
     Corrosion Costs  on Metals in the  U.S.  APTD-0654.  U.S. Environmental Protection Agency,
     Research Triangle Park, NC,  February 1971.

Fleetwood,  M.  J.   Corrosion— resisting metals  reviez-zinc  coatings.   Corros.  Prev.  Control
     22:16-19, 1975.

Fochtman, E.  G. ,  and G. Langer.  Automobile paint damaged by airborne iron particles.  J. Air
     Pollut. Control  Assoc.  6:243-247,  1957.

Freeman,  A. M. ,  III.   The Benefits  of  Environmental  Improvement,  Theory and Practice.   Johns
     Hopkins University Press, Baltimore, MD, 1979.

Gaj'endr'agadkar,  S.  K.  Effects of sulphur dioxide  on materials.   Chem.  Age India  28:673-677,
     1977.

Gauri  'K. L.   Effect of acid  rain on structures.  In:  Acid Rain,  Proceedings of the  National
     Convention Workshop Session, American Society of Civtl Engineers, Boston,  Massachusetts,
     April 2-6,  1979-   Preprint 3598, American Society of  Civil  Engineers, Boston, HA,  1979.
     pp. 55-75.
SOX10A/D
                                          10-57 ^                                    12-15-80

-------
Gibbons,  E.  V.   The Corrosion  Behaviour  of the Major Architectural  and  Structural  Metals In
     Canadian Atmospheres.   Summary  of Ten-Year  Results  of Group  I.   NRCC  11630,  National
     Research Council of Canada, Ottawa, Canada, October 1970.

Gillette, D. G.   S02 and material damage.  J.  Air Pollut.  Control Assoc. 25:1238-1243, 1975.

Gillies, K.  A.,  and D.  J. Youston.  Social effects of atmospheric emissions.  Presented at the
     69th Annual Meeting, Air Pollution Control Association, Portland, Oregon, June 27-July 1,
     1976.  Paper No. 76-19.4.

Glass, N.  Environmental Effects of Increased Coal Utilization:   Ecological Effects of Gaseous
     Emissions from  Coal  Combustion.   EPA 600/7-78-108, U.S. Environmental Protection Agency,
     Corvallis,  OR, June 1978.

Guttman,  H.   Effects  of atmospheric  factors  on  the corrosion of  rolled zinc.   In:   Metal
     Corrosion  in the  Atmosphere,  A  symposium Presented  at  the Seventieth  Annual  Meeting,
     American Society  for  Testing  and  Materials, Boston,  Massachusetts, June  25-39,  1967.
     ASTM Special Technical  Publication  No.  435, American  Society  for  Testing and Materials,
     Philadelphia, PA,  1968.  pp.  223-239.

Guttman, H., and  P.  J.  Sereda.   Measurement of atmospheric factors affecting the corrosion of
     metals.   In:  Metal  Corrosion  in the Atmosphere, A Symposium Presented at the Seventieth
     Annual  Meeting, American Society for Testing and Materials,  Boston,  Massachusetts, June
     25-30,  1967.   ASTM Special  Technical  Publication No.  435, American  Society for Testing
     and Materials, Philadelphia,  PA,  1968.   pp. 326-359.

Haagenrud, S.,  and B.   Ottar.   Long  range  transport  of air pollution  and corrosion effects.
     In:  Proceedings of the Seventh Scandinavian Corrosion Congress, Trondheim, Norway, 1975.
     pp. 102-115.

Hansen, J.  Ailing Treasures.  Sci.  80 (Eighty) 1:58-61, 1980.

Harker, A. B.,  F.  B.  Mansfeld,  D. R.  Strauss,  and D. D. Landis.   Mechanism of S02 and H^SO.
     Aerosol  Zinc Corrosion.  EPA-600/3-80-018, U.S.  Environmental Protection Agency, Research
     Triangle Park, NC, January 1980.

Hay, T.  K.,  and G. G.   Schurr.  Moisture  diffusion phenomena in practical  paint systems.  J.
     Paint Technol. 43:63-72, 1971.

Haynie, F. H.   The economics of clean air in perspective.   Mater. Prot. Perf. 13:33-38, 1974.

Haynie,  F.  H.   Air  pollution  effects on  stress  induced intergranular corrosion of 7005-T53
     aluminum  alloy.   In:    Stress  Corrosion  -  New  Approaches.   ASTM Special  Technical
     Publication  610,  American  Society for  Testing  and  Materials, Philadelphia,  PA,  1976.
     pp. 32-43.

Haynie,  F.  H.   Theoretical   air pollution  and  climate effects on  materials confirmed by zinc
     corrosion data.   In:   Durability of Building Materials and Components.  P. J. Sereda and
     G. G. Levitan, eds., ASTM Special Technical Publication 691, American Society for; Testing
     and Materials, Philadelphia, PA, 1980.  pp. 157-175.                                ,

Haynie, F. H.,  J.  W. Spence, and J.  B.  Up ham.  Effects of Gaseous Pollutants on Materials—A
     Chamber Study.  EPA-600/3-76-015, U.S. Environmental Protection Agency, Research TJriangle
     Park, NC, February 1976.                                                         ••;'-.

Haynie, F. H.,  J.  W. Spence, and J.  B.  Upham.  Technical  note:  effects of air pollutants on
     weathering  steel  and  galvanized steel:   a chamber study.   Mater.  Perform.  15:48, 1976.
SOX10A/D                                  10-58                                        12-15-*

-------
Haynie,  F.   H.,   and  0.  B.  Upham.    Correlation  between  corrosion  behavior  of  steel  and
     atmospheric  pollution  data.   ^n:   Corrosion  in  Natural   Environments.   ASTM  Special
     Technical Publication  558,  American Society for Testing and Materials, Philadelphia, PA,
     1974.  pp. 33-51.

Haynie, F.  H. ,  and J. B. Upham.   Effects  of atmospheric pollutants on the corrosion behavior
     of steels.  Mater. Prot. Perform. 10:18-21, 1971.

Haynie, F.  H., and  J.  B.  Upham.  Effects  of atmospheric sulfur dioxide  on  the corrosion of
     zinc.  Mater. Prot. Perform. 9:35-40, 1970.

Hershaft, A.  Air pollution damage functions.  Environ. Sci. Technol. 10:992-995, 1976.

Hershaft, A.,  A.  M.  Freeman, T. D. Crocker, and S. B.  Stevens.  Critical review of estimating
     benefits  of  air and  water pollution  control.   EPA-600/5-78-014, Washington,  DC,  June,
     1978.

Holbrow,  G.  L.   Atmospheric  pollution:   its  measurement and some effects on  paint.   J. Oil
     Color Chem. Assoc. 45:701-718, 1962.

Hoshizawa,  K., and  K.   Koyata.   Formation  and  their  prevention of  acid  smuts  at oil-fired
     burner.  Nenyro  Kyokaishi 49:656-665, 1970.

ITT Electro-Physics  Laboratories,  Inc.  A Survey and Economic Assessment of the Effect of Air
     Pollutants  on Electrical  Components.   Volume 1,  Sections 1 through  9.   APTD-0797, U.S.
     Environmental Protection Agency,  Research Triangle Park, NC, August 1971.

Ireland,  F.  E.  Pollution by Oxides of Sulfur.  Chem. Eng. (London) (221): CE 261-CE262, 1968.

Jacksch,  J.  A.,  and H. H.  Stoevener.  Effects of Air Pollution on Residential Property Values
     in Toledo,  Oregon.   Special Report 304, Oregon State University, Agricultural Experiment
     Station, Corvallis, OR, September 1970.

Johnson,  J.   B.,  P.  Elliott,  M.  A.   Winterbottom,  and  G.  C. Wood.   Short-term atmospheric
     corrosion  of mild  steel  at two  weather and pollution  monitored sites.   Corrosion Sci.
     17:691-700,  1977.

Karraker,  D. G.   The kinetics  of the reaction between  sulfurous  acid  and  ferric iron.  J.
     Phys. Chem.  67:871-874, 1963.

Kottori,  M.  S.   The  corrosion  rates  of metals exposed  to pulp-mill  and smelter emissions.
     Presented  -at  the  73rd Annual   Meeting,  Air  Pollution  Control  Association,  Montreal,
     Quebec,  June  22-27, 1980.   Paper  No 80-67.4.

Krumbein, W.  E.,  and C. Lange.  Decay of plaster, paintings and wall material of the  interior
     of  buildings via microbial activity.   In:  Environmental  Biochemistry and Geochemistry,
     Volume  2:  The Terrestrial  Environment, Proceedings of the Third International Symposium,
     Minister of  Sciences  and Arts  of  Niedersachsen  and  Deutsche Forschungsemeinschaft,
     Wolfbuettell,  Germany, 1977.  W.  E.  Krumbein,  ed., Ann  Arbor Science Publishers,  Inc.,
     Ann  Arbor, MI, 1978.   pp. 687-697.

Kingery,  R.  E.    The.extent  of the  paper  problem in  large research  collections  and the
     comparative   costs    of    available    solutions.     In:     Summary   of   a  Conference,
     Permanent/Durable  American Library Association and Virginia  State Library, Washington,
     DC   September 16,  I960.   Virginia State  Library  Publications  No.  16,  Virginia State
     Library,  Richmond, VA, 1960.  pp. 36-41.
SOX10A/D                                  1Q-59                                        12-15-80

-------
Kucera,  V.   Effects  of sulfur dioxide and  acid  precipitation on metals and anti-rust painted
     steel.  Ambio 5:243-248, 1976.

Larrabee,  C.   P.   Mechanisms by  which  ferrous  metals  corrode  in the  atmosphere.  Corrosion
     (Houston) 15:526, 1959.

Larrabee,  C.   P.,  and S.  K.  Coburn.   The  atmospheric corrosion  of  steels as  influenced by
     changes   in   chemical   composition.    In:    First  International  Congress  on  Metallic
     Corrosion,  London,  April 10-15,  1961.   Butterworth  and Co.   (Publishers)  Ltd.,  London,
     England, 1962.  pp. 276-285.

Liu, B., and E. S. Yu.  Physical and Economic Damage Functions for Air Pollutants by Receptor.
     EPA 600/5-76-011,  U.  S.  Environmental  Protection Agency, Corvallis,  OR,  September 1976.

Litvin,  A.   Clear  coatings  for  exposed  architectural  concrete.    J.   PCA  Res.   Dev.   Lab.
     10:49-57, 1968.

Luckat, S.   Air pollutant effects  on dome at Cologne.  Koeln Domblatt 38:39, 1976.

Luckat,  S.   Investigations  concerning the  protection against  air  pollutants of  objects of
     natural stone.  Staub Reinhalt Luft 32: May 1972.

Maler, K.  G.,  and R.  E. Wyzga.   Economic  Measurement of  Environmental  Damage:   A Technical
     Handbook.   Organization for Economic Co-operation, Paris, France, 1976.

Mansfeld,  F.  B.   Regional  Air  Pollution  Study.   Effects  of  Airborne Sulfur  Pollutants on
     Materials.   EPA-600/4-80-007,  U.S.  Environmental  Protection  Agency,  Research Triangle
     Park,  NC, January 1980.

Mansfeld,  F.,  and  J.  V.   Kenkel.   Electrochemical   monitoring  of  atmospheric  corrosion
     phenomena.  Corros. Sci.  16:111-122, 1976.

Matsushima, I., Y. Ishizu,  T.  Ueno, M. Kanazashi,  and K.  Horikawa.  Effect of structural and
     environmental factors on  the  practical  use  of low-alloy  weathering  steel.   Boshoku
     Gijutsu 23:177-182, 1974.

Michelson,  I., and B. Tourin.   Report on Study of Validity of Extension of Economic Effects of
     Air   Pollution   Damage  from  Upper  Ohio   River  Valley  to   Washington,   D.C.  Area.
     Environmental Health and Safety Research Association, August 1967.

Moore, J.  C.,  and J.  R.  O'Leary.   Evaluation  of  Structural Steel  Coatings in  Relation to
     Industrial  Atmospheric  Conditions.    West  Virginia  Department  of  Highways,  Research
     Project 23, West Virginia Department of Highways,  Charleston, WV, January 1975.

Narayanan,   R.  L., and  B.  W.  Lancaster.   Household  maintenance  costs  and  particulate air
     pollution.  Clean Air (Melbourne) 7:10-13,  1973.

National Academy of  Sciences.   Airborne  Particles.  University  Park Press,  Baltimore, MD,
     1979.

Newton,  R.  G.   Cathedral  chemistry—conserving  the  stained glass.   Chem.  Br.  10:89-91, 1974.

Niesel,  K.   The weathering of  building  materials  in  atmospheres  containing sulfur oxides-a
     literature discussion.  Fortsch.  Mineral. 57:68-124, 1979.

Nriagu,  J.  0.   Deteriorative effects  of sulfur pollution  on materials.   In:   Sulfur, in the
     Environment.  Part II:  Ecological Impacts.   John Wiley & Sons,  Inc., New York, NY, 1978.
     pp.  1-59.


SOX10A/D                                  10-60                                        12-15-8C

-------
Oma,  K.,  T.  Sugano,  and  Y.  Hirai.   Studies  on atmospheric corrosion  of steels  related tc
     meteorological  factors  in  Japan.   II.   Multiple  correlation  of  meteorological  *rA
     atmospheric pollution  substances  on corrosion of steel.  Boshoku Gijutsu 14:16-19, 1965.

Overseas Public  Nuisance  Study Mission.  Present Status of  Public  Nuisance in Electric Power
     Industry of Japan.  Central Power Council, Tokyo, Japan, September,  1965.

Oyama, M.,  A. Moriyama,  H. Miyazato,  and  H. Sakugawa.   Examples  on the acid-smut fallout.
     Okinawa-ken Kogai Eisei Kenkysho-ho 8:32-33, 1974.

Parker A.   The  destructive effects  of  air  pollution on materials.   In:   Proceedings  of the
     22nd Annual Conference, National Smoke Abatement Society, Bournemouth, England, September
     28, 1955.  National Smoke Abatement Society, Brighton, England, 1955.  pp. 1-15.

Potter, P.   Chimney corrosion and acid smut emission.  N.Z. J.  Dairy Sci.  Technol. 6:11-14,
     1971.

Phafflin and Ziegler, 1976.

Prenderleith,  H.  J.  The  Preservation of Leather Bookbindings.  The  British Museum,  London,
     England, 1957.

Riederer, J.   Pollution  damage to works of art.   In:  New Concepts in Air Pollution Research.
     Experientia Suppl. (20):73-85. 1974.

Robbins,  R.  C.  Inquiry  into the Economic  Effects of Air  Pollution  on Electrical Contacts.
     Stanford Research Institute, Menlo Park, CA, April 1970.

Rowden E.   Acid corrosion and smut  emission by kiln exhaust gases.  Trans.  Br.  Ceram.  Soc.
     68:227-229, 1968.

Russell,  C.  A.  How  environmental  pollutants diminish  contact  reliability.   Insul. Circuits
     22:43-44, 1976.

Salmon, R. L.  Systems Analysis of the  Effects of Air Pollution on Materials.  APTD-0943,  U.S.
     Department   of  Health,   Education  and  Welfare,   National    Air   Pollution   Control
     Administration, Raleigh, NC, January 1970.

Salvin,  V.   S.   Effect  of air  pollutants  on dyed fabrics.  J.  Air Pollut.  Control  Assoc.
     13:416-422, 1963.

Sanyal, B., and G.  K. Singhania.  Atmospheric corrosion of metals:  Part I.  J. Sci. Ind.  Res.
     Sect. B. 15:248-255,  1956.

Satake, -J.,  and T. Moroishi.   Iji:   Proceedings  of the  Fifth International  Congress on Metal
     Corrosion, 1974.  National Association of Corrosion Engineers, Houston, TX.  pp. 744-749.

Schwarz, H.  On the effect  on magnetite on atmospheric rust and on rust under a coat of paint.
     Werkst. Korros. 23:648-663, 1972.

Schwenk   W    and   H.   Ternes.   Rate  of corrosion  of plain carbon and  low-alloy  structural
     steels.' Stahl Eisen 88:318-321, 1968.

Senqupta  M   and  A   A  de  Gast.   Environmental  deterioration  and  evaluation for dimension
     stone. "Can. Min. Metall. Bull.  65:54-58, 1972.

Sereda   P   L    Weather  factors  affecting  corrosion  pf metals.   In:  Corrosion  in  Natural
     Environments.  ASTM  Special  Technical  Publication 558, American  Society for Testing and
     Materials, Pittsburg,  PA, 1974.   pp. 7-22.
SOX10A/D
                                          10-61                                         12-15-80

-------
Simpson,  J.  W.,   and  P.  J.  Horrobin,  eds.    The  Weathering  and  Performance  of Building
     Materials.  Wiley-Interscience, New York, NY, 1970. p. 286.

Sleater,  G.   A.    Stone  Preservatives:    Methods  of  Laboratory  Testing  and  Preliminary
     Performance  Criteria.   NBS  Technical  Note 941,  U.S.  Department of  Commerce, National
     Bureau of Standards, Washington, DC, 1977.

Spedding, D.  J.,   P.  P. Rowlands,  and  J.  E.  Taylor.   Sorption of  sulphur dioxide by  indoor
     surfaces:  leather.  J. Appl. Chem. Biotechnol.  21:68-70, 1971.

Spence,  J.  W., F.  H. Haynie,  and J. B.  Upham.   Effects of gaseous  pollutants on paints:  A
     chamber study.  J. Paint Technol. 47:57-63, 1975.

Spence, J. W., and F. H. Haynie.  Design of a laboratory experiment to identify the effects of
     environmental pollutants on materials.  In:  Corrosion in Natural Environments.  STP  558.
     American Society for Testing and Materials, Philadelphia, PA, 1974.  pp. 279-291.

Spence, J.  W., and F. H. Haynie.   Paint Technology  and Air Pollution:  A  Survey and Economic
     Assessment.    AP-103,  U.S.  Environmental  Protection Agency,  Research  Triangle Park, NC,
     February 1972.

Stanners, J.  F.   Use of environmental data  in  atmospheric corrosion studies.  Br. Corros. J.
     5:117-121, 1970.

Svoboda, M.,  B. Knapek, and H.  Kilcoua.   The  effect of pigments on the penetration of  sulfur
     dioxide into coatings.   J.  Oil Colour Chem. Assoc. 5_6:172174, 1973.

Sydberger,  T.   Influence  of  Sulphur  Pollution  on  the  Atmospheric  Corrosion  of   Steel.
     University of Gothenburg,  Department of  Inorganic Chemistry,  Gothenburg, Sweden, 1976.
     pp. 3-8.

Sydberger,  T., and  R.  Ericsson.   Laboratory testing  of the  atmospheric corrosion of  steel.
     Werkst. Korros. 28:154-158, ,1976.

Sydberger,  T., and  N.  G.  Vannerberg.  The  influence of the  relative humidity and corrosion
     products on  the adsorption of sulfur dioxide on metal surfaces.  Corros. Sci. 12:775-784,
     1972.

Tomashov, N. D.  The Science of Corrosion.   The MacMillan Co., New York, NY, 1966.

Upham, J. B.  Atmospheric corrosion studies in two metropolitan areas.  J.  Air  Pollut. Control
     Assoc.  17:398-402, 1967.

Upham,  J.   B., and  V.  S.  Salvin.   Effects  of Air Pollutants  on  Textile  Fibers  and Dyes.
     EPA-650/3-74-008,  U.S.  Environmental  Protection Agency,  Research  Triangle  Park,  NC,
     February 1975.

Vannerberg,   N. G., and T.  Sydberger.   Reaction between SO, and  wet metal  surfaces.  Corros.
     Sci. 10:43-49, 1970.                                  *                            ,

Verdu.   Effect of air  pollutants on aging of plastic  materials.   Trib.  CEBEDEAU 27:360-370,
     1974.                                                                         ~~

Waddell, T.   E.  The Economic Damages of Air Pollution.  EPA-600/5-74-012, U.S. Environmental
     Protection Agency, Washington, DC, May 1974.

Walsh, M.,  A.  Black, A. Morgan, and G. Crashaw.  Sorption of  SO, by  indoor surfaces including
     carpets, wallpaper, and paint.  Atmos. Environ.  11:1107-1111, 1977.


SOX10A/D                                   10-62                                        12-15-80

-------
Watson, W.  D. ,  Jr., and J. A. Jaksch.  Household cleaning costs and air pollution.  Presented
     at  the  71st  Annual  Meeting,  Air  Pollution Control  Association, Houston,  Texas, June
     25-30, 1978.   Paper No. 78-52.3.

Winkler,  E.  M.   Stone:   Properties, Durability  in Man's  Environment.  Second revised edition,
     Springer-Verlag, New  York, NY, 1975.

Winkler,  E.  M.   Important agents of weathering  for building and monumental stone.  Eng. Geol.
     (Amsterdam) 1:381-400, 1966.

Yocom,  J.  E.  and  Grappone.   Effects of  Power  Plant  Emissions  on Materials.   EPRI  EC-139,
     Electric Power Research  Institute, Palo Alto, CA, July 1976.

Yocom,  J. E.   Air  pollution  damage  to  buildings on  the Acropolis.   J.  Air Pollut.  Control
     Assoc. 29:333-338,  1979.

Yocom,  J.  E.,  and  J.   B.  Upham.   Effects  on  economic  materials and  structures.   In:  Air
     Pollution.  Third  Edition.   Volume II:   The Effects of Air Pollution.  A. C. Stern, ed.,
     Academic Press,  Inc.,  New York, NY, 1977.   pp. 65-116.

Yoshihara,  K.,  T. C.  Huang, H. Ebihora, and H. Shibata.  Chemical  effects  of the  35Cl(n,p)35  S
     reaction.   Radiochim.  Acta 3:185-191, 1964.

Zeronian,  S. H.   Reactions of  cellulosic  fabrics to  air contaminated with  sulfur dioxide.
     Text.  Res.  J.  40:695-698, 1970.

Zeronian,  S.  H.,  K.  W. Alger,  and S.  T.  Omaye.  Reactions  of  fabrics  made  from  synthetic
     fibers   to air  contaminated  with  nitrogen dioxide,  ozone,  or  sulfur dioxide.  In:
     Proceedings of  the Second  International Clean Air  Congress, International  Union  of Air
     Pollution,  Washington, DC,  December  6-11,  1970.   H.  M. Englund  and W.  T.  Beery, eds.,
     Academic Press,  New York, NY,  1971.  pp. 468-476.
 SOX10A/D                                 10~63

-------
                                 GLOSSARY FOR VOLUME IV
Adsorption:  Adhesion of a thin layer of molecules to a liquid or solid surface.
Aerosol:  Solid particles or liquid droplets which are dispersed or suspended in air.
Albedo:   That fraction  of  the  total light  incident on a  reflecting surface,  especially  a
     celestial body, which is reflected back in all directions.
Anode:  Positive electrode.
Austenite:  Gamma iron with carbon in solution.
Austenitic steel:   An   alloy   whose  structure  is  typically  that  of  austenite  at  room
     temperature.
Colloidal  system:   An  intimate mixture of two  substances, one  of which, called the dispersed
     phase (or colloid), is uniformly distributed in a finely divided state through the second
     substance, called  the dispersion medium (or dispersing medium); the dispersion medium or
     dispersed phase  may  be  a gas,  liquid,  or solid.   Also known as  colloidal dispersion;
     colloidal suspension.
Convective mixing:  Atmospheric motions that are predominantly vertical, resulting in vertical
     transport and mixing of atmospheric properties.
Crystalization spall ing:   The  reduction  of stone surfaces worn  away by a crystal line-forming
     process.
Ecosystem:  The interacting system of a biological community and its environment.
Efflorescence  (mineral):   A  whitish powder,  consisting of one or several minerals produced as
     an encrustation on the surface of a rock in an arid region.   Also known as bloom.
Electrophoretic  coating:   A surface  coating on  a  metal deposited  by electric  discharge of
     particles from a colloidal solution.
Exfoliate:  Flake away or peel off in scales.
Ferrous (metals):  Relating to or containing iron.
Hygroscopic:   Pertaining  to a marked ability  to accelerate the condensation  of water vapor.
Insolation decoloration:  Removal or fading of color due to exposure to the sun.
Isopleth:  A line on a map or chart connecting points of equal value.
Iterative bidding:  A research technique for subjectively establishing value.
Mesoscale:  Of or relating to meteorological phenomena from 1 to 100 kilometers in horizontal
     extent.
Nephelometer-  A type  of instrument  that measures  at  more than  one angle, the  scattering
     function  of  particles  suspended  in  a  medium; information   obtained  may  be   used to
     determine the size of suspended particles and the visual range through the medium.
Nuclei (plural  of  nucleus):   A  central   point,  group,  or  mass  about  which  gathering,
     concentration, or accretion takes place.
SOXFM/G

-------
Participate matter  (participates,  particles,  etc.):   Fine liquid or solid particles within an
     aerosol such as dust, smoke, mist, fumes or smog.
Passivate  (metal):   To  render passive; to reduce the reactivity of a chemically active metal
     surface by electrochemical polarization or by immersion in a passivating solution.
Rayleigh scattering:  Coherent scattering in which the intensity of the light of wavelength g,
     scattered  in  any  direction making  an  angle  with  the  incident direction,  is directly
     proportional to 1 + cos r and inversely proportional to g .
Sorb:  To take up and hold by absorption or adsorption.
Sorbed:  Past tense of "sorb."
Sorption:  The process of being sorbed.
Spatial:   Relating to, occupying, or of the nature of space.
Spectral:  Relating to the electromagnetic spectrum.
Spectroscopy:    The  branch  of  physics  concerned  with  the  production,  measurement,  and
     interpretation of electromagnetic  spectra  arising from either emission  or absorption of
     radiant energy by various substances.
Stoichiometry:   The numerical relationship of elements and compounds as reactants and products
     in chemical reactions.
Sulfur dioxide  (^O^):   Colorless  gas with  pungent  odor released  primarily from  burning of
     fossil fuels, 'such as coal, containing sulfur.
Transmissometer:  An instrument for measuring the extinction coefficient of the atmosphere and
     for the determination  of  visual  range.   Also known  as  hazemeter or transmittance meter.
Troposphere:   That portion  of the  atmosphere  in  which  temperature decreases  rapidly with
     altitude,  clouds  form, and mixing  of  air  masses by convection  takes  place.   Generally
     extends to about 7 to 10 miles above the earth's surface.
SOXFM/G                                     G-2                                       12-17-80

-------