vvEPA
          United States
          Environmental Protection
          Agency
                 Office of Air and Radiation
                 Washington D.C. 20460
Assessing the Risks of
Trace Gases That Can
Modify the Stratosphere
EPA 400/1-87/001C
December 1987
          Volume III:
          Chapters 6-18

-------
Assessing The Risks of Trace Gases
 That Can Modify The Stratosphere


      Volume III: Chapters 6-18


   Senior Editor and Author: John S. Hoffman
         Office of Air and Radiation
     U.S. Environmental Protection Agency
          Washington, D.C. 20460
            December 1987

-------
                                ACKNOWLEDGEMENTS
     Many people made  this  document possible.

     A Science Advisory Board  (SAB) panel chaired by Dr. Margaret Kripke and co-
 chaired by Dr. Warner North conducted an extensive and constructive review of
 this document.  Members of the panel provided  important insights and assistance
 in the assessments development.  Members of the panel are:

     Dr.  Martyn Caldwell (Utah State University)
     Dr.  Leo T. Chylack, Jr. (Center for Clinical Cataract Research)
     Dr.  Nien Dak Sze  (A.E.R., Inc.)
     Dr.  Robert Dean
     Dr.  Thomas Fitzpatrick (Massachusetts General Hospital)
     Dr.  James Friend  (Drexel  University)
     Dr.  Donald Hunten (University of Arizona)
     Dr.  Warren Johnson (National Center for Atmospheric Research)
     Dr.  Margaret Kripke (Anderson Hospital and Tumor Institute)
     Dr.  Lester Lave (Carnegie-Melon University)
     Dr.  Irving Mintzer (World Resources Institute)
     Dr.  Warner North  (Decision Focus, Inc.)
     Dr.  Robert'Watson (National Aeronautics and Space Administration)
     Dr.  Charles Yentsch (Bigelow Laboratory)
     Dr.  Terry F. Yosie (U.S.  Environmental Protection Agency)

 The  panel's contribution to the process of protecting stratospheric ozone has
 been critical.  We also want  to thank Terry Yosie, Director of the Science
 Advisory Board, for setting up and helping to  run the panels, and Joanna
 Foellmer for helping  to organize meetings.

     Other scientists  and analysts, too numerous to name, provided reviews of
 early drafts of the chapters.

     Production assistance, including editing,  typing, and graphics, was
 provided by the staff of ICF  Incorporated, including:

     Bonita Bailey
     Susan MacMillan
    Mary O'Connor

Maria Tikoff of the U.S.  Environmental Protection Agency,  coordinated
 logistical parts of this  document.

    The  cover photograph was  supplied by the National Aeronautics and Space
Administration.

    Technical support documents (Volumes VI,  VII,  and VIII) have been published
along with the first five  volumes of the Risk Assessment.   These documents are
not part of the official  Risk Assessment,  and have not been reviewed by the
SAB.  Their publication is simply to assist readers who wish more background
than available in the Risk Assessment.

-------
                       List of Contributors
Craig Ebert
ICF Incorporated,
9300 Lee Highway,
Fairfax, VA  22031

Sarah Foster
ICF-Clement,
9300 Lee Highway,
Fairfax, VA  22031

Michael J. Gibbs
ICF Incorporated,
9300 Lee Highway,
Fairfax, VA  22031

Kevin Hearle
ICF Incorporated,
9300 Lee Highway,
Fairfax, VA  22031

Brian Hicks
ICF Incorporated,
9300 Lee Highway,
Fairfax, VA  22031

Patsy H. Lill
Department of Pathology,
University of South Carolina School of Medicine,
VA Bldg No. 1, Garnet Ferry Rd,
Columbia, SC  29208

Janice Longstreth
ICF-Clement,
9300 Lee Highway,
Fairfax, VA  22031

Neil Patel
U.S. Environmental Protection Agency,
401 M Street, S.W.,
Washington, DC  20460

Hugh M. Pitcher
U.S. Environmental Protection Agency,
401 M Street, S.W.,
Washington, DC  20460

-------
Alan F. Teramura
Department of Botany,
University of Maryland,
College Park, MD  20742

Dennis Tirpak
U.S. Environmental Protection Agency,
401 M Street, S.W.,
Washington, DC  20460

Jim Titus
U.S. Environmental Protection Agency,
401 M Street, S.W.,
Washington, DC  20460

John B. Wells
The Bruce Company,
Suite 410, 3701 Massachusetts Ave., N.W.,
Washington, DC  20016

G. Z. Whitten
Systems Applications, Inc.,
101 Lucus Valley Road,
San Rafael, CA 94903

Robert Worrest
Corvallis Environmental Research Laboratory,
200 Southwest 35th Street,
Corvallis, OR  97333

-------
                               TABLE OF CONTENTS

                                                                        PAGE
VOLUME I

ACKNOWLEDGMENTS 	   i

ORGANIZATION 	  ES-1

INTRODUCTION 	  ES-2

SUMMARY FINDINGS 	  ES-5

CHANGES IN ATMOSPHERIC COMPOSITION 	  ES-15

POTENTIAL CHANGES IN OZONE AND CLIMATE 	  ES-23

HUMAN HEALTH, WELFARE, AND ENVIRONMENTAL EFFECTS 	  ES-32

QUANTITATIVE ASSESSMENT OF RISKS WITH INTEGRATED MODEL  	  ES-54

VOLUME II

ACKNOWLEDGEMENTS 	  i

INTRODUCTION 	  1

    The Rise of Concern About Stratospheric Change 	  1
    Concern About Public Health and Welfare Effects of Global
        Atmospheric Change 	  1
    Need for Assessments 	  2

1.  GOALS AND APPROACH OF THIS RISK ASSESSMENT 	  1-1

    Analytic Framework 	  1-1
    Supporting Documents and Analysis for this Review 	  1-2
    Chapter Outlines 	  1_2

2.  STRATOSPHERIC PERTURBANTS:   PAST CHANGES IN CONCENTRATIONS
        AND FACTORS THAT DETERMINE CONCENTRATIONS 	  2-1

    Summary 	  2-1
    Findings 	  2-3
    Measured Increases in Tropospheric Concentrations of
        Potential Ozone Depleters	  2-4
    Measured Increases in Tropospheric Concentrations of
        Potential Ozone Increasers 	 2-13
    Factors that Influence Trace Gas Lifetimes 	  '2-21
    Long-Lived Trace Gases 	  2-22
    Trace Gases with Shorter Lifetimes 	  2-26
    Carbon Dioxide  and the Carbon Cycle 	.'	  2-26
    Source Gases for Stratospheric Sulfate Aerosol (OCS, CS2) 	  2-26
    Appendix A:  CFC Emissions-Concentrations Model 	  2-28
    References  	  2-30

-------
                               TABLE OF CONTENTS
                                  (Continued)

                                                                        PAGE
3.   EMISSIONS OF INDUSTRIALLY PRODUCED POTENTIAL OZONE MODIFIERS 	   3-1

    Summary 	   3-1
    Findings 	   3-3
    Introduction 	   3-6
    Chlorofluorocarbons 	   3-6
    Chlorocarbons 	   3-58
    Halons 	   3-59
    References 	   3-66

    Appendix A:   Chemical Use Estimate Made Available
            Since Publication of the Risk Assessment 	   A-l

    Appendix A:   References 	   A-10

4.   FUTURE EMISSIONS AND CONCENTRATIONS OF TRACE GASES WITH
        PARTLY BIOGENIC SOURCES 	   4-1

    Summary 	   4-1
    Findings 	   4-2
    The Influence of Trace Gases on the Stratosphere and
        Troposphere 	   4-4
    Trace Gas Scenarios 	   4-4
    Effects of Possible Future Limits on Global Warming 	   4-23
    Conclusion 	   4-23
    References 	   4-25

5.   ASSESSMENT OF THE RISK OF OZONE MODIFICATION 	   5-1

    Summary  	   5-1
    Findings 	   5-3
    Introduction 	   5-6
    Equilibrium Predictions for Two-Dimensional Models 	   5-18
    Time Dependent Predictions for One-Dimensional
        Models for Different Scenarios of Trace Gases  	   5-32
    Time Dependent Predictions for Two-Dimensional Models
        with Different Scenarios of Trace Gases  	   5-40
    Models Fail to Represent All Processes That Govern
        Stratospheric Change in a Complete and Accurate Manner  	   5-61
    The Implications of Ozone Monitoring for Assessing Risks
        of Ozone Modification	   5-80
    References  	   5-104

VOLUME III

6.  CLIMATE  	   6-1

    Summary  	   6-1
    Findings  	   6-2
    The Greenhouse Theory  	   6-7
    Radiative Forcing by Increases in Greenhouse Gases 	   6-8
    Ultimate Temperature Sensitivity  	   6-15
    The Timing of Global Warming  	   6-18
    Regional Changes in Climate Due to Global Warming  	   6-22

-------
                               TABLE OF CONTENTS
                                  (Continued)
                                                                        PAGE
    Effects  on the  Stratosphere of Possible Control of Greenhouse
        Gases  	  6-26
    Attachment A:   Description of Model to be Used in Integrating
                   Chapter 	  6-28
    Attachment B:   Trace Gas Scenarios 	  6-32
    References 	  6-33

7.   NONMELANOMA SKIN TUMORS 	  7-1

    Summary  	  7-1
    Findings 	  7-2
    Background on Solar Radiation and the Concept of Dose 	  7-5
    Introduction 	  7-5
    Biology  of Nonmelanoma Skin Tumors: Links to UV-B 	  7-11
    Epidemiological Evidence 	  7-27
    Dose-Response Relationships 	  7-40
    Attachment A	  7-49
    References 	  7-58

8.   CUTANEOUS  MALIGNANT MELANOMA  	  8-1

    Summary  	•	  8-1
    Findings 	  8-3
    Introduction 	  8-7
    Epidemiologic Evidence  	  8-11
    Experimental Evidence  	  8-28
    Dose-Response Relationships 	  8-29
    References 	  8-41

9.   UVR-INDUCED IMMUNOSUPPRESSION: CHARACTERISTICS AND POTENTIAL
        IMPACTS 	  9-1

    Summary 	  9-1
    Findings  	  9-3
    Introduction 	  9-5
    Basic Concepts in Immunology  	  9-5
    Salt: Skin-Associated Lymphoid Tissues  	  9-7
    Effects  of Ultraviolet Radiation on Immunological Reactions  	  9-8
    Human Studies  	  9-14
    Effects  of Ultraviolet Radiation on Infectious Diseases  	  9-15
    References 	  9-18

10. CATARACTS AND OTHER EYE DISORDERS  	  10-1

    Summary 	  10-1
    Findings  	  10-2
    Cataracts  	  10-3
    Potential Changes in Senile Cataract  Prevalence  for
        Changes in UV-B  	  10-29
    Other Eye Disorders  	  10-33
    References 	  10-37

-------
                               TABLE OF CONTENTS
                                  (Continued)
                                                                        PAGE
11.  RISKS TO CROPS  AND TERRESTRIAL ECOSYSTEMS FROM ENHANCED
        UV-B RADIATION 	   11-1

    Summary 	   11-1
    Findings 	   11-2
    Introduction 	   11-5
    Issues and Uncertainties in Assessing the Effects
        of UV-B Radiation on Plants 	   11-5
    Issues Concerning UV Dose and Current Action Spectra
        for UV-B Impact Assessment 	   11-5
    Issues Concerning Natural Plant Adaptations to UV Radiation 	   11-7
    Issues Associated with the Extrapolation of Data from
        Controlled Environments to the Field 	   11-10
    Uncertainties in Our Current Knowledge of UV-B Effects on
        Terrestrial Ecosystems and Plant Growth Forms 	   11-11
    Uncertainties with the Ability to Extrapolate Knowledge to Higher
        Ambient C02 Environment and Other Atmospheric Pollutants 	   11-13
    Risks to Crop Yield Resulting from an Increase in
        Solar UV-B Radiation 	   11-15
    Risks to Yield Due to a Decrease in Quality 	   11-20
    Risks to Yield Due to Possible Increases in
        Disease or Pest Attack 	   11-20
    Risks to -Yield Due to Competition with Other Plants 	   11-22
    Risks to Yield Due to Changes in Pollination and Flowering  	   11-23
    References  	   11-25

12. AN ASSESSMENT OF THE EFFECTS OF ULTRAVIOLET-B
        RADIATION ON AQUATIC ORGANISMS  	   12-1

    Summary  	   12-1
    Findings  	   12-2
    Introduction 	   12-4
    Background on Marine Organisms and Solar Ultraviolet
        Radiation  	   12-4
    Effects of UV-B Radiation in Phytoplankton  	   12-9
    Effects on Invertebrate Zooplankton  	   12-11
    Effects on Ichthyoplankton (Fisheries)  	   12-23
    Conclusions  	   12-28
    References  	   12-29

-------
                               TABLE OF CONTENTS
                                  (Continued)

                                                                        PAGE

13.  EFFECTS OF UV-B ON POLYMERS 	  13-1

    Summary 	  13-1
    Findings 	  13-2
    Photodegradation of Polymers 	  13-4
    Polymers in Outdoor Uses and the Potential for Degradation 	  13-7
    Damage Functions and Response to Damage 	  13-16
    Effect of Temperature and Humidity on Photodegradation 	  13-29
    Future Research 	  13-31
    References 	  13-32

14.  POTENTIAL EFFECTS OF STRATOSPHERIC OZONE DEPLETION ON
        TROPOSPHERIC OZONE 	  14-1

    Summary 	  14-1
    Findings 	  14-2
    Introduction 	  14-3
    Potential Effects of Ultraviolet Radiation and Increased
        Temperatures on Urban Smog  	  14-5
    Conclusions and Future Research Directions 	  14-9
    References 	  14-14

15.  CAUSES AND EFFECTS OF SEA LEVEL RISE  	  15-1

    Summary 	  15-1
    Findings 	  15-2
    Causes of Sea Level Rise 	  15-5
    Effects of Sea Level Rise 	  15-15
    Conclusion 	  15-32
    Notes  	  15-33
    References 	  15-34

16.  POTENTIAL EFFECTS OF FUTURE CLIMATE CHANGES ON FORESTS
        AND VEGETATION, AGRICULTURE, WATER RESOURCES
        AND HUMAN HEALTH 	  16-1

    Summary 	  16-1
    Findings 	  16-5
    References 	  16-10

-------
                               TABLE OF CONTENTS
                                  (Continued)
                                                                        PAGE
17.  MODELS FOR INTEGRATING THE ANALYSES  OF HEALTH AND
        ENVIRONMENTAL RISKS ASSOCIATED WITH OZONE MODIFICATION 	  17-1

    Summary 	  17-1
    Introduction 	  17-2
    The Model as a Framework 	  17-2
    Analysis Procedure 	  17-4
    Model Limitations 	  17-9
    References 	  17-11
    Appendix A:   Model Design and Model  Flow 	  A-l
    Appendix B:   Scenarios of Chemical Production, Population,
        and GNP 	  B-l
    Appendix C:   Evaluation of Policy Alternatives 	  C-l
    Appendix D:   Emissions of Potential  Ozone-Depleting Compounds 	  D-l
    Appendix E:   Atmospheric Science Module 	  E-l
    Appendix F:   Health and Environmental Impacts of Ozone
        Depletion 	  F-l

18.  HUMAN HELATH AND ENVIRONMENTAL EFFECTS 	  18-1

    Summary 	  18-1
    Findings 	  18-2
    Introduction 	  18-6
    Methods for Estimating Health and Environmental Risks 	  18-11
    Description of Range of Production,  Emissions, and Concentrations
        Scenarios for Evaluating Risks 	  18-12
    Sensitivity of Health and Environmental Effects to Differences
        in Emissions of Ozone Depleters 	  18-18
    Sensitivity of Results to Alternative Atmospheric Assumptions 	  18-23
    Sensitivity of Effects to Uncertainty in Dose Response 	  18-54
    Relative Importance of Key Uncertainties 	  18-61
    Summary 	  18-62
    References  	  18-65

VOLUME IV

    Appendix A

    Ultraviolet Radiation and Melanoma

VOLUME V

    Appendix B

    Potential Effects of Future Climate Changes on Forests and
        Vegetation, Agriculture, Water Resources, and Human Health

VOLUME VI

    Technical Support Documents

    Appendix C

    Projecting  Production of Ozone Depleting Substances

-------
                               TABLE OF CONTENTS
                                  (Continued)
VOLUME VII

    Technical Support Document

    Appendix D

    Scientific Papers

VOLUME VIII

    Technical Support Document

    Appendix E

    Current Risks and Uncertainties of Stratospheric Ozone Depletion
        Upon Plants

-------
                               LIST  OF EXHIBITS

                                                                    Page

1-1    Relationships Among the Chapters  	  1-3

2-1    Measured Increases in Tropospheric Concentrations
       of CFC-11 (C7C-13) 	  2-5

2-2    Measured Increases in Tropospheric Concentrations
       of CFC-12 (CF2CL2) 	  2-6

2-3    Measured Increases in Tropospheric Concentrations
       of HCFC-22  (CHC1F2)	  2-7

2-4    Measured Increases in Tropospheric Concentrations
       of CFC-113  (C2CL3F3)  	  2-7

2-5    Measured Increased in Tropospheric Concentrations
       of Carbon Tetrachloride  (CC14)  	  2-9

2-6    Measured Increases in Tropospheric Concentrations
       of Methyl Chloroform  (CH3CC13)  	  2-10

2-7    Measured Increases in Tropospheric Concentrations
       of Halon-1211  (CF2ClBr)  	  2-11

2-8    Measured Increases in Tropospheric Concentrations
       of Nitrous  Oxide  (N20)  	  2-12

2-9    Measured Increases in Tropospheric Concentrations
       of Nitrous  Oxide  (N20)  	  2-14

2-10  Ice  Core Measurements of Historical  Nitrous  Oxide
        (N20)  Concentrations  	  2-15

2-11  Measured Increases in Tropospheric Concentrations
       of Carbon Dioxide (C02)  	  2-16

2-12  Ice  Core Measurements of Historical  Carbon Dioxide
        (C02)  Concentrations  	  2-17

2-13  Measured Increases in Tropospheric Concentrations
       of Methane  (CH4)  	  2-19

 2-14   Ice  Core Measurements of Historical  Methane  (CH4)
       Concentrations 	  2-20

2-15a  CFC-12:  Constant Emissions  	  2-23

2-15b   CFC-12:  Atmospheric  Concentrations  	  2-23

2-16a   CFC-12:  Emissions  	  2-24

-------
                               LIST OF  EXHIBITS
                                  (Continued)

                                                                     Page

2-16b   CFC-12:  Atmospheric Concentrations 	  2-24

2-17    CFC-12:  Atmospheric Concentrations from Different
        Emission Trajectories 	  2-25

A-l     Concentrations of Fluorocarbons 	  2-29

A-2     Locations of Stations 	  2-29

3-1     Selected Properties of CFCS 	  3-8

3-2     CFC Characteristics and Substitutes 	  3-10

3-3     Companies Reporting Data to CMA 	  3-11

3-4     Production of CFC-11 and CFC-12 Reported to CMA 	  3-13

3-5     Historical Production of CFC-11 and CFC-12 	  3-15

3-6     CFC-11 and CFC-12 Used in Aerosol and Nonaerosol
        Applications in the EEC 	  3-16

3-7     Comparison of Estimated CFC-11 Use: 1985 	  3-18

3-8     Comparison of Estimated CFC-12 Use: 1985 	  3-19

3-9     Estimates of Production and Emissions of CFC-11
        and CFC-12 	  3-21

3-10    Published Estimates of U.S.S.R. Production of CFC-11
        and CFC-12 	  3-22

3-11    Historical Production of CFC-11 and CFC-12 in the U.S	  3-24

3-12    EEC Production and Sales Data 	  3-25

3-13    The Bottom Up Approach 	  3-27

3-14    Range  of Population and GNP Per Capita Projections  	  3-30

3-15    Summary of Demand Projection Estimates 	  3-32

3-16    Summary of Demand Projection Estimates
        (Average annual rate of growth  in percent) 	  3-34

  3-17   Long Term Projections CFC-11 and CFC-12
        World  Production (2000-2050) 	  3-35

-------
                              LIST OF EXHIBITS
                                 (Continued)

                                                                    Page

3-18   Bevington's Projections for Use  of CFCs in the EEC 	   3-38

3-19   Camm Projections of World Use  	   3-40

3-20   Summary of EFCTC Projections 	   3-42

3-21   Global Population and GNP Scenarios Used in
       Gibbs' Analysis 	   3-43

3-22   Gibbs Scenarios of World CFG Use 	   3-45

3-23   Hammitt Projections of World CFC Use 	   3-47

3-24   Summary of Hedenstrom Projections for  Sweden 	   3-49

3-25   Knollys Projections 	   3-51

3-26   Summary of Kurosawa Projections  for Japan 	   3-53

3-27   Nordhaus Scenarios of World Nonaerosol CFC Consumption ....   3-55

3-28   Summary of Sheffield Projections for Canada 	   3-57
       (1984-2005)

3-29   Global Halon Projections for Quinn 	   3-62

3-30   Hammitt and Camm Global Halon  Projections 	   3-64

A-la   Global Annual Production in Millions of Kilograms 	   A-4

A-lb   U.S. Annual Production in Millions of Kilograms 	   A-4

A-2    Assumptions for lEc Halon Projections   	   A-5

A-3    lEc Projections of Sales and Emissions for Halon-1301
       and Halon-1211 	   A-6

A-4    Global Halon-1301 and Halon-1211 Growth Rates 	   A-7

A-5    Dupont Estimates of CFC per Capita Use	   A-8

4-1    Effects of Changes in Composition of Atmosphere 	   4-5

4-2    Historical Carbon Dioxide Emissions from Fossil Fuels
       and Cement 	   4-7

4-3    A Schematic of the Carbon Cycle  	   4-8

-------
                              LIST OF EXHIBITS
                                 (Continued)
                                                                    Page
4-4    Projected Carbon Dioxide Emissions and Doubling
       Time of Concentrations 	  4-10

                                         12
4-5    Estimated CH4 Emission Sources (10   grams per year) 	  4-12

4-6    Two Ways That CH4 Concentrations Could Have Changed 	  4-14

4-7    Possible Changes in CH4 Sources and in
       Emission Factors 	  4-15

4-8    Current Sources and Sinks of Carbon Monoxide
       (1984 concentration of CO:   30-200 ppb) 	  4-17

4-9    Scenarios of Carbon Monoxide (CO) Emissions from
       Combustion 	  4-18

4-10   Preliminary Scenario of Future Growth in N20
       Emissions by Source 	  4-21

4-11   Projected Nitrous Oxide (N20) Concentrations  	  4-22

4-12   Summary of Standard Scenarios Proposed for Assessment  	  4-24

5-1    Temperature Profile and'Ozone Distribution in the
       Atmosphere 	  5-7

5-2    Steady-State Scenarios Used in International Assessment  ...  5-10

5-3    Change in Total Ozone from Representative One-Dimensional
       Models for Steady-State Scenarios Containing  Clx
       Perturbations  	  5-12

5-4    Change in Total Ozone at 40 Kilometers for Steady-State
       Scenarios Containing  Clx Perturbations 	  5-13

5-5    Change in Total Ozone for Steady-State Scenarios  	  5-14

5-6    Effect of Stratospheric Nitrogen (NOy) on Chlorine-Induced
       Ozone Depletion  	  5-15

5-7    Effect of Doubled C02 Concentrations  on Ozone
       Temperature  	  5-16

5-8    Calculated Changes  in Ozone Versus  Altitude  	  5-17

5-9    Two-Dimensional Model Scenarios  Used  in International
       Assessment  	  5-19

-------
                              LIST OF EXHIBITS
                                 (Continued)

                                                                    Page
5-10   2-Dimensional Model Results:  Globally and Seasonally-
       Averaged Ozone Depletion 	  5-20

5-11   Ozone Depletion by Latitude,  Altitude, and Season for Clx
       Increase of 6.8 ppbv (MPIC 2-D Model) 	  5-21

5-12   Ozone Depletion by Latitude,  Altitude, and Month for Clx
       Increase of 6.8 ppbv (AER 2-D Model)  	  5-22

5-13   Ozone Depletion by Latitude,  Altitude, and Month for Clx
       Increase of 14.2 ppbv (AER 2-D Model) 	  5-23

5-14   Ozone Depletion of Latitude,  Altitude, and Month for Clx
       Increase of 6.8 ppbv (GS 2-D Model)  	  5-24

5-15   Ozone Depletion by Latitude and Season for Clx Increase
       of 6.0 ppbv (IS 2-D Model) (AER 2-D Model) 	  5-25

5-16   Change in Ozone by Latitude and Season for Clx
       Perturbations  (MPIC 2-D Model) 	  5-26

5-17   Change in Ozone by Latitude and Season for Clx
       Perturbations  (AER 2-D Model) 	  5-27

5-18   Latitudinal Dependence of AER and MPIC 2-D Models  	 5-28

5-19   Change in Ozone by Latitude, Altitude, and Month for
       Coupled Perturbations (GS 2-D Model)  	  5-29

5-20   Changes in Ozone by Latitude, Altitude, and Season for
       Coupled Perturbations (MPIC 2-D Model) 	  5-30

5-21   Changes in Ozone by Latitude and Altitude in Winter  for
       Coupled Perturbations (MPIC 2-D Model) 	  5-31

5-22   Models With Reported Time Dependent  Runs  	  5-33

5-23   LLNL 1-D Model Versus Parameterization Fit 	  5-34

5-24   Trace Gas Assumptions for Results  in Exhibit 5-25
        (Brasseur and  DeRudder 1-D Model,  1986)	  5-35

5-25   Time-Dependent Change in Ozone for CFC Growth and Coupled
       Perturbations  (Brasseur and DeRudder 1-D Model)  	  5-36

5-25   Time-Dependent Change in Ozone for Constant CFC  Emissions
       and Growth in  Other Trace Gases (Brasseur and DeRudder
       1-D Model) 	  5-37

-------
                              LIST OF EXHIBITS
                                 (Continued)
                                                                    Page
5-27   Sensitivity of 1-D Models to Representation of Radiative
       Processes (Brasseur and DeRudder 1-D Model) 	  5-38

5-28   Model Comparison:  Time-Dependent Change in Ozone for CFC
       Growth and Coupled Perturbations 	  5-39

5-29   Trace Gas Assumptions for Results in Exhibit 5-30
       (AER 1-D Model, 1986) 	  5-41

5-30   Time-Dependent Change in Ozone for Various Scenarios of
       Coupled Perturbations (AER 1-D Model) 	  5-42

5-31   Trace Gas Scenarios Tested in LLNL 1-D Model 	  5-43

5-32   Time-Dependent, Globally Averaged Change in Ozone for
       Coupled Perturbations (LLNL 1-D Model) "Reference Case"  ...  5-45

5-33   Time Dependent, Globally Averaged Change in Ozone for
       Coupled Perturbations (LLNL 1-D Model) 	  5-46

5-34   Effect of Potential Greenhouse Gas Controls on Ozone
       Depletion (Results from 1-D Parameterization) 	  5-47

5-35   Calculated Ozone Depletion for 1970 to 1980 Versus
       Umkehr Measurements	  5-49

5-36   Time-Dependent Globally and Seasonally Averaged Changes
       in Ozone for Coupled Perturbations (IS 2-D Model)  	  5-50

5-37   Time-Dependent Globally and Seasonally Averaged Changes
       in Ozone for Coupled Perturbations (IS 2-D Model)  	  5-51

5-38   Time-Dependent Seasonally Averaged Change  in Ozone  for
       1980 CFC Emissions and Coupled Perturbations  (IS 2-D
       Model)  	  5-52

5-39   Time-Dependent Seasonally Averaged Change  in Ozone  for
       1.2% Growth in CFC Emissions  and Coupled Perturbations
       (IS 2-D Model)  	  5-53

5-40   Time-Dependent Seasonally Averaged Change  in Ozone  for
       3% Growth  in CFC  Emissions  and Coupled Perturbations
       (IS 2-D Model)  	  5-54

5-41   Time-Dependent Seasonally Averaged Change  in Ozone  for
       3.8% Growth in CFC Emissions  and Coupled Perturbations
       (IS 2-D Model)  	  5-55

-------
                              LIST OF EXHIBITS
                                 (Continued)
5-42   Temperature Feedback Experiment:  Time-Dependent, Globally
       and Seasonally Averaged Change in Ozone for 3% Growth in
       CFC Emissions and Coupled Perturbations (IS 2-D Model) ....  5-57

5-43a  Two-Dimensional,  Time-Dependent Simulation for Constant
       CFC Emissions (AER 2-D Model)  	  5-58

5-43b  Two-Dimensional,  Time-Dependent Simulation for CFC Growth
       of 62 Percent Per Year (AER 2-D Model) 	  5-59

5-44   Model Comparison for Coupled Perturbations Scenario 	  5-60

5-45   Calcualted Ozone -- Column Change to Steady-State for Two
       Standard Assumed Perturbations 	  5-63

5-46   Latitudinal Gradients in Odd Nitrogen: Models vs
       Measurements 	  5-65

5-47   Logical Flow Diagram for Monte Carlo Calculations 	  5-67

5-48   Histogram of Measurements for a Rate Constant  	  5-68

5-49   Recommended Rate Constants and Uncertainties Used in
       Monte Carlo Analyses 	  5-70

5-50   Monte Carlo Results: Change in Ozone Versus Fluorocarbon
       Flux 	  5-71

5-51   Monte Carlo Results: Change in Ozone Versus Fluorocarbon
       Flux 	  5-72

5-52   Monte Carlo Results: Ozone Depletion for Coupled
       Pertubations  	  5-73

5-53   A Monte Carlo Distribution of Column Ozone Changes
       for Changes in CFC Production  	  5-74

5-54   Monte Carlo Results: Changes in Ozone by Altitude  	  5-76

5-55   Monte Carlo Results: Changes in Ozone by. Column and
       Altitude, Unscreen Data  	  5-77

5-56   Monte Carlo Analysis With the LLNL  1-D Model  	  5-78

5-57   Monte Carlo Results:  Changes in Ozone by Altitude  	  5-79

5-58   Ozone Trend Estimates by Latitude  	  5-83

-------
                              LIST OF EXHIBITS
                                 (Continued)
                                                                    Page
5-59   Changes in Ozone from 1970 to 1980:   Umkehr Measurements
       and Model Calculations 	   5-85

5-60   SBUV Zonal Trends Estimates Versus "Umkehr Station
       Blocks" 	   5-86

5-61   Ozone Trend Estimates and 95% Confidence Intervals 	   5-88

5-62   Ozone Balloonsonde Stations 	   5-90

5-63   Correction Factors for Balloonsonde Measurements 	   5-91

5-64   Ozone Trend Emissions (% per year) As Determined from
       Balloon Ozonesondes Versus Those Determined from Dodson
       Measurements (Tiao, et al., personal communication) 	   5-93

5-65   Monthly Means of Total Ozone at Halley Bay 	   5-96

5-66   Nimbus 7 Antarctic Ozone Measurements:  12-Day Sequence ...   5-97

5-67   Nimbus 7 Antarctic Ozone Measurements:  Mean Total
       Ozone in October  	   5-98

5-68   Global (60°N-60°S) Monthly Ozone Determined from
       NOAA TOVS System  	   5-100

5-69   Preliminary Ozone Trend Data (Health versus 2-D Model
       Results)  (Isaksen)  	    5-102

6-1    Stratospheric Perturbants and Their Effects 	   6-9

6-2    Absorption Characteristics of Trace Gases  	   6-10

6-3    Radiative Forcing for a Uniform Increase in Trace Gases ...   6-11

6-4    Effects of Vertical Ozone Distribution on
       Surface Temperature  	   6-13

6-5    Water Vapor, Altitude, and Radiative  Forcing  	   6-14

6-6    Temperature Sensitivity to Climatic Feedback Mechanisms  ...   6-16

6-7    Empirical Estimates of Climate Sensitivity are Sensitive
       to  Estimates of Historical Temperature Increases and
       Trace Gas Concentrations  	  6-17

6-8    Relationship of Radiative Forcing, Ocean Heat Uptake,
       and Realized and  Unrealized Warming  	'.	  6-20

-------
                              LIST OF EXHIBITS
                                 (Continued)

                                                                    Page

6-9    Transient Estimates of Global  Warming  	   6-21

6-10   Expected Temperature Increases 	   6-23

6-11   Results of Transient Analysis  Using a  General
       Circulation Model 	   6-24

6-12   Regions of U.S. :   Change in Runoff 	   6-25

7-1    Variation in UV  Radiation by Latitude  as Percent of Levels
       at the Equator on March 21 at Noon 	   7-6

7-2    UV Radiation by  Month in Washington,  D.C	   7-8

7-3    Ratio of Instantaneous Flux Throughout the Day to
       Flux at 5:15 a.m. in Washington on June 21 (Assumes
       a Clear Day) 	   7-9

7-4    Average DNA-Damage Action Spectrum 	   7-10

7-5    Organization of  the Adult Skin 	   7-12

7-6    Ultraviolet Absorption Spectra of Major Epidermal
       Chromophores 	   7-14

7-7    Skin Types and Skin Tanning Responses  	   7-16

7-8    Ultraviolet Action Spectra for DNA Dimer Induction,
       Lethality and Mutagenicity 	   7-19

7-9    Effectiveness of UVR at Inducing Pyrimidine Dimers
       and Transformation  	   7-21

7-10   Action Spectrum  for the Induction of Single-Strand Breaks
       in DNA 	   7-24

7-11   Action Spectrum  of Mouse Edema (MEE48) as Compared
       to that of DNA Damage and the Robertson-Berger
       Meter 	   7-26

7-12   Comparison of Age-Adjusted Incidence Rates Per 100,000
       Persons for Squamous Cell Carcinoma (SCC) and Basal Cell
       Carcinoma (BCC)  Among White Males and Females in the
       United States 	   7-29

7-13   Percentage of Tumors by Anatomic Site for Nonmelanoma
       Skin Cancer Among White Males and Females in the United
       States (1977-1978 NCI Survey Data)	   7-31

-------
                              LIST OF EXHIBITS
                                 (Continued)
                                                                    Page
7-14   Distribution by Sex and Anatomic Site of Nonmelanoma
       Skin Tumors: Canton of Vaud,  Switzerland (1974-1978) 	   7-32

7-15   Annual Age Adjusted Incidence Rates for Basal and Squamous
       Cell Carcinoma (1977-1978 NCI Survey Data)  and Melanoma
       (1973-1976 SEER Data)  Among White Males 	   7-35

7-16   Annual Age Adjusted Incidence Rates for Basal and
       Squamous Cell Carcinomas (1977-1978 NCI Survey Data) and
       Melanoma (1973-1976 SEER Data) Among White  Females 	   7-36

7-17   Estimated of  Relative Risks of Basal and Squamous Cell
       Carcinomas for 32 Combinations of Risk Factors 	   7-38

7-18   Relative Mutagenicity of UV-B as a Function of Wavelength .   7-46

 A-l   Correlation of Alternative Measurements of UV-B Radiation
       for Ten Locations in the United States 	   7-51

 A-2   Population Weights for Ten Locations in the United
       States 	   7-52

 A-3   Estimated Dose-Response Coefficeints (and t-Statistics) for
       Basal and Squamous Cell Skin Cancers (UV-B Dose-Skin
       Cancer Incidence) 	   7-53

 A-4   Estimated Percentage Changes in UV-B Radiation in San
       Francisco for a Two and Ten Percent Depletion in Ozone ....   7-55

 A-5   Percentage Change in Incidence of Basal and Squamous Cell
       Skin Cancers for a Two Percent Depletion in Ozone for
       San Francisco 	   7-56

 A-6   Percentage Change in Incidence of Basal and Squamous Cell
       Skin Cancers for a Ten Percent Depletion in Ozone for
       San Francisco 	   7-57

8-1    Location of Melanocyte in the Epidermis 	   8-8

8-2    Comparative Transmittance of UV Radiation 	   8-9

8-3    Increases in Incidence and Mortality Rates from
       Malignant Melanoma in Different Countries 	   8-17

8-4    Anatomic Site Distribution of Cutaneous Malignant
       Melanoma  	   8-19

-------
                               LIST OF EXHIBITS
                                 (Continued)
8-5    Anatomic Site Distribution of Cutaneous Malignant
       Melanoma by Gender  	  8-20

8-6    Malignant Melanoma Risk Factors by Measures of Skin
       Pigmentation Within the Caucasian Population  	  8-26

8-7    Multiple Ways in which UVR can Play a Role in Melanoma
       Development  	  8-30

8-8    Summary Statistics  for Regressions of Skin Cancer
       Incidence and Mortality on Latitude  	  8-32

8-9    Estimated Relative  Increases  in Melanoma  Skin Cancer
       Incidence and Mortality Associated with Changes  in
       Erythema Dose  	  8-33

8-10   Summary of Fears, Scotto, and Schneiderman  (1977)
       Regression Analyses of Melanoma Incidence Dose-Response  ...  8-35

8-11   Biological Amplification  Factors  for Skin Melanoma by
       Sex and Anatomical  Site Groups, Adjusting for Age and
       Selected Constitutional and  Exposure Variables  	  8-35

8-12   Biological Amplification  Factors  for Melanoma Incidence
       by Sex and Anatomical Site Groups, Adjusting  for Age
       and Combinations  of Selected Constitutional and
       Exposure Variables  	  8-36

8-13   Percentage Increase in Melanoma Death Rates for  a One
       Percent Decline in  Ozone  	  8-39

 9-1    Action Spectra for  Local  Suppression of  Contact
       Hypersensitivity Assuming Either  One-Hit  or Multi-Hit
       Mechanisms  	  9-11

 9-2    Action Spectrum For Systemic Suppression of Contact
       Hypersensivity 	  9-12

10-1     Cataract  Prevalence by UV Zone  	  10-7

10-2     Comparison of Cataract  Prevalence for Aborigines and
       Non-Aborigines 	  10-7

10-3     Composite  Transmittance  Curves  for the  Rabbit 	  10-9

10-4     Calculated Total Transmittance  of the Human Eye  	  10-10

10-5     Percent  Transmissivity  Through the Entire Rhesus 'Eye 	  10-11

-------
                               LIST OF EXHIBITS
                                  (Continued)
                                                                     Page
10-6    Transmittance of the Total Rabbit Cornea,  the Total
        Human Cornea, and the Rabbit Corneal Epithalium 	 10-12

10-7    Transmittance of the Anterior Ocular Structures of
        the Human and Rabbit Eyes 	 10-13

10-8    UV Radiant Exposure Threshold Data for the Cornea
        H£, Lens HL Cataracts, and Retina 1L.
        for the Rabbit and Primate 	 10-14

10-9    The Action Spectra for Photokeratitis and Cataracts
        for the Primate and Rabbit 	 10-15

10-10   Free Radicals and Oxidation: Reduction Systems 	 10-17

10-11   Enzyme Systems Involved in Oxidation: Reduction 	 10-18

10-12   Standardized Regression Coefficients for Cataract  	 10-31

10-13   Estimated Relationship Between Risk of Cataract
        and UV-B Flux  	 10-32

11-1    A Summary of Studies Examining the Sensitivity of  Cultivars
        to UV-B Radiation  	 11-8

11-2    Survey of UV Studies by Major Terrestrial Plant
        Ecosystems  (after Whittaker 1975)  	 11-12

11-3    Summary of UV-B and C02 Effects on Plants  	 11-14

11-4    Summary of  Field Studies Examining the Effects of
        UV-B Radiation on Crop Yields  	 11-16

11-5    Details of  Field Study by Teramura  (1981-1985)  	 11-17

11-6    Summary of  Changes  in Yield Quality  in Soybean
        Between the  1982 and  1985 Growing  Seasons
        (Teramura 1982-1985)  	 11-21

12-1    Solar  Irradiance Outside  the  Earth's Atmosphere and at
        the  Surface of the  Earth  for  a Solar Zenith  Angle  of 60"... 12-5

12-2    Relationship Between  Ozone  Depletion and  Biological
        Effectiveness  of Increased  UV-B  Radiation 	 12-6

12-3    Solar  Spectral Irradiance at  the  Surface  of  the Ocean
        and  at Four Depths  	 12-7

-------
                               LIST OF EXHIBITS
                                  (Continued)
                                                                     Page
12-4    Lethal Effects on Shrimp Larvae for Various Combinations
        of UV-R,  Dose-Rate,  and Total Dose 	  12-13

12-5    Estimated Effective UV-B Solar Daily Dose at Various
        Atmospheric Ozone Concentrations Based on a 4-Year
        Mean of Medians, Manchester, Washington,  1977-1980 	 12-14

12-6    Estimated Biologically Effective UV-B Doses Leading to
        Significant Effects in Major Marine Zooplankton Groups .... 12-17

12-7    Percentage of Total Dose Limit to be Reached on Any
        Particular Day:  Lethal Doses Accumulated Only After
        Dose-Rate Threshold is Exceeded 	 12-20

12-8    Effect of Increased Levels of Solar UV-B Radiation on
        the Predicted Loss of Larval Northern Anchovy from Annual
        Populations, Considering the Dose/Dose-Rate Threshold
        and Three Vertical Mixing Models 	 12-27

13-1    Wavelengths of UV Radiation and Polymers with Maximum
        Sensitivity and Corresponding Photon Energies 	 13-4

13-2    Plastics Used in Applications Where Exposure of the
        Material to Sunlight Might Be Expected 	 13-8

13-3    Modes of Damage Experienced by Polymers Used in
        Outdoor Application 	 13-11

13-4    PVC Siding Compound Composition 	 13-13

13-5    UV Screening Effectiveness of Selected Pigments 	 13-14

13-6    Domestic Consumption of Light Stabilizers, 1984-85 	 13-15

13-7    Increased Stabilization Market (1970-2020) 	 13-18

13-8    Ozone Depletion Estimates 	 13-19

13-9    Cumulative Added Cost  	 13-20

13-10   Diagrammatic Representation of the Effect of
        Pigment/Fillers as Light Shielders
        (Monodisperse Spherical Filler) 	 13-22

13-11   Relative Damage Indices for Yellowing of PVC Under Miami
        (March 22nd) Conditions, at Different Extents of
        Ozone Layer Deterioration 	 13-25

13-12   Estimated Ranges of Factor Increase in Damage and the
        Factor Increase in Stabilizer Needed to Counter the
        Change of Yellowing of Rigid PVC Compositions 	 13-28

-------
                               LIST OF EXHIBITS
                                  (Continued)

                                                                     Page

13-13   PVC Damage with Ozone Depletion 	 13-29

13-14   Projections of Future Demand for Selected Years
        (Thousands of Metric Tons)  	 13-30

14-1    Ozone Concentrations for Short-Term Exposure That
        Produce 5% to 20% Injury to Vegetation Growth Under
        Sensitive Conditions 	 14-6

14-2    Ozone Concentrations at Which Significant Yield
        Losses Have Been Noted for  a Variety of Plant
        Species Exposed Under Various Experimental Conditions 	 14-7

14-3    Ozone Concentrations Predicted for Changes in
        Dobson Number and Temperature for Three Cities (ppm) 	 14-10

14-4    Global Warming Would Exacerbate Effects of Depletion
        on Ground-Based Ozone in Nashville 	 14-11

15-1    Snow and Ice Components 	 15-6

15-2    Worldwide Sea Level in the  Last Century 	 15-8

15-3    Temperature Increase At Various Depths and Latitudes 	 15-10

15-4    Estimates of Future Sea Level Rise 	 15-13

15-5    Local Sea Level Rise 	 15-14

15-6    Evolution of Marsh as Sea Level Rises 	 15-17

15-7    Composite Transect -- Charleston, S.C	 15-18

15-8    Louisiana Shoreline in the  Year 2030 	 15-20

15-9    Distribution of Population in Bangladesh  	 15-21

15-10   The Bruun Rule 	 15-23

15-11   Percent of Tidal Cycles in Which Specified
        Concentration is Exceeded at Torresdale .During a
        Recurrence of the 1960's Drought for Three Sea Level
        Scenarios  	 15-28

15-12   Estimates of Flood Damages  for Charleston and Galveston
        Resulting From Sea Level Rise 	 15-30

16-1    Summary of Findings from the WMO/UNEP/ICSU Conference
        on Global Climate Held in Villach, Austria, October 1985  .. 16-1

17-1    Modular Structure  	 17-3

-------
                              LIST OF EXHIBITS
                                 (Continued)

                                                                    Page

17-2   Major Model Input Choices 	 17-5

17-3   Effects Not Quantified 	 17-10

 A-l   Flow of Analysis Program 	  A-4

 A-2   Files Required to Specify a Run 	  A-6

 B-l   Future Global Production Scenarios: Middle Scenario 	  B-3

 B-2   Regional Use Shares for CFC-11 	  B-4

 B-3   U.S. End Use Shares for CFC-11 	  B-6

 B-4   Middle U.S. Population Scenario 	  B-7

 B-5   Middle U.S. GNP Scenario 	  B-7

 B-6   User-Modified Scenario Specifying a Growth Rate of 2
       Percent Annually for CFC-11 in the U.S	  B-9

 B-7   A User-Modified Scenario: Production as a Function of
       Population 	  B-9

 D-l   Release Tables for CFC-11 	  D-2

 D-2   Emissions from a Hypothetical 100 Million Kilograms of
       Production in 1985 	  D-4

 D-3   Emissions from Production Over a Series of Years
       (Millions of Kilograms)  	  D-5

 D-4   Sample Table of Exogenously Specified Emissions 	  D-6

 E-l   Trace Gas Assumptions Used to Develop the Ozone Depletion
       Relationship 	  E-2

 E-2   Comparison of Total Column Ozone Depletion Results from
       the 1-D Model and the Parameterized Numerical Fit 	  E-3

 E-3   Hypothetical Table of User-Specified Ozone Depletion 	  E-6

 E-4   Example Ozone Depletion Scaling Factors 	  E-7

 F-l   Cities Used to Evaluate Changes in UV Flux for the Three
       Regions of the U. S	  F- 3

 F-2   States Included in the Three Regions of the U.S. .'	  F-5

 F-3   Percent Change in UV as a Function of Change in Ozone
       Abundance for Three U.S. Regions 	  F-6

-------
                              LIST OF EXHIBITS
                                 (Continued)
                                                                    Page
 F-4   Age Distribution  of  the U.S.  Population Over Time  in the
      North Region  	   F-8

 F-5   Baseline  Incidence for Nonmelanoma  Skin Cancers  	    F-10

 F-6   Basal Incidence for  Melanoma  Skin Cancers  	   F-12

 F-7   Mortality Rates for  Melanoma  Skin Cancers  	   F-14

 F-8   Standardized  Regression Coefficients  for Cataract  	   F-16

 F-9   Estimated Relationship Between Risk of Cataract  and UV-B
      Flux  	   F-17

 F-10  Sample  Table  for  Specifying Relative  Weights for Exposure
      During  a  Person's Lifetime  	   F-19

 F-ll  Coefficients  Relating Percent Change  in UV to Percent
      Change  in Skin Cancer Incidence 	   F-22

 F-12  Coefficients  Relating Percent Change  in UV to Percent
      Change  in Melanoma Mortality  	   F-23

 F-13  Dose-Response Coefficients  	   F-25

 F-14  Damage  Index  and  Increase in  Stabilizer for Ranges of
      Ozone Depletion 	   F-29

18-1   Types of  Human Health and Environmental Effects
      Estimated 	  18-8

18-2A  Real  World Equivalent to  the  No-Growth Scenario  in Risk
      Assessment 	  18-15

18-2B  Real  World Equivalent to  the  1.2% Growth Scenario  in Risk
      Risk  Assessment 	  18-16

18-3   Global  Average Ozone Depletion:  Central Case 	  18-20

18-4   Additional Cases  of  Nonmelanoma Skin Cancer by Type of
      Nonmelanoma 	  18-21

18-5   Additional Mortality From Nonmelanoma Skin Cancer  by Type
      of Nonmelanoma 	  18-22

18-6   Additional Cases  of  Melanoma  Skin Cancer by Cohort  	  18-23

18-7   Additional Mortality From Melanoma Skin Cancer by Cohort ..  18-24

18-8   Additional Senile Cataract Cases by Cohort  	  18-25

18-9    Environmental Effects Estimated Quantitatively for  the U.S.  18-26

-------
                               LIST OF EXHIBITS
                                  (Continued)
                                                                     Page
18-10   Environmental Effects Based on Case Studies and Research
        in Early Stages 	 18-27

18-11   Equilibrium Temperature Change for the Emissions Scenarios
        Assuming 3°C Warming for Doubled C02 	 18-28

18-12   Global Average Ozone Depletion:   Comparison to Results
        with a 2-D Dimensional Atmospheric Model 	 18-30

18-13   Climate, and Other Effects:  Sensitivity to Relationship
        Between Climate Change and C02 Emissions 	 18-31

18-14   Summary of Effects of Greenhouse Gases on Ozone Depletion
        and Global Equilibrium Temperature 	 18-33

18-15   Global Average Ozone Depletion:   Scenario of Limits
        to Future Global Warming 	•	 18-34

18-16   Human Health Effects:  Scenarios of Limits to Future
        Global Warming.  Additional Cumulative Cases and Deaths
        Over Lifetimes of People Alive Today 	 18-36

18-17   Human Health Effects:  Scenarios of Limits to Future
        Global Warming.  Additional Cumulative Cases and Deaths
        Over Lifetimes of People Born 1986-2029 	 18-37

18-18   Human Health Effects:  Scenarios of Limits to Future
        Global Warming.  Additional Cumulative Cases and Deaths
        Over Lifetimes of People Born 2030-2074 	 18-38

18-19   Materials, Climate,  and Other Effects:   Scenarios
        of Limits to Future  Global Warming 	 18-39

18-20   Global Average Ozone Depletion:   Methane Scenarios  	 18-41

18-21   Human Health Effects:  Methane Scenarios Additional
        Cumulative Cases and Deaths Over Lifetimes
        of People Alive Today 	 18-42

18-22   Human Health Effects:  Methane Scenarios
        Additional Cumulative Cases and Deaths Over Lifetimes
        of People Born 1986-2029 	 18-43

18-23   Human Health Effects:  Methane Emissions Cases
        Additional Cumulative Cases and Deaths Over Lifetimes
        of People Born 2030-2074 	 18-44

18-24   Materials, Climate and Other Effects:   Methane Scenarios .. 18-45

18-25   Global Average Ozone Depletion:   Sensitivity to
        Relationship Between Ozone Depletion and Emissions  	 18-46

-------
                               LIST OF EXHIBITS
                                  (Continued)
                                                                     Page
18-26   Human Health Effects:   Sensitivity to Relationship
        Between Ozone Depletion and Emissions.  Additional
        Cumulative Cases and Deaths Over Lifetimes of People
        Alive Today 	  18-48

18-27   Human Health Effects:   Sensitivity to Relationship
        Between Ozone Depletion and Emissions.  Additional
        Cumulative Cases and Deaths Over Lifetimes of People
        Born 1985-2029 	 18-49

18-28   Human Health Effects:   Sensitivity to Relationship
        Between Ozone Depletion and Emissions.  Additional
        Cumulative Cases and Deaths Over Lifetimes of People
        Born 2030-2074 	 18-50

18-29   Human Health Effects:   Maximum Depletion of 95 Percent.
        Additional Cumulative Cases and Deaths Over Lifetimes
        of People Alive Today 	 18-51

18-30   Human Health Effects:   Maximum Depletion of 95 Percent.
        Additional Cumulative Cases and Deaths Over Lifetimes of
        People Born 1986-2029 	 18-52

18-31   Human Health Effects:   Maximum Depletion of 95 Percent.
        Additional Cumulative Cases and Deaths Over Lifetimes of
        People Born 2030-2074 	 18-53

18-32   Human Health Effects:   Sensitivity to Dose-Response
        Relationship.  Additional Cumulative Cases and
        Deaths Over Lifetimes of People Alive Today  	 18-55

18-33   Human Health Effects:  Sensitivity to Dose-Response
        Relationship Additional Cumulative Cases and
        Deaths Over Lifetimes of People Born 1986-2029  	 18-56

18-34   Human Health Effects:  Sensitivity to Dose-Response
        Relationship Additional Cumulative Cases and
        Deaths Over Lifetimes of People Born 2030-2074  	 18-57

18-35   Human Health Effects:  Sensitivity to Relationship Between
        Ozone Depletion and Action  Spectrum.  Additional  Cumulative
        Cases and Deaths Over Lifetimes of People Alive Today  	 18-58

18-36   Human Health Effects:  Sensitivity to Relationship Between
        Ozone Depletion and Action  Spectrum.  Additional  Cumulative
        Cases and Deaths Over Lifetimes of People Born  1986-2029  	 18-59

18-37   Human Health Effects:  Sensitivity to Relationship Between
        Ozone Depletion and Action  Spectrum.  Additional  Cases  and
        Deaths Over Liftimes of People  Born  2030-2074  	 18-60

-------
                               LIST OF  EXHIBITS
                                  (Continued)
                                                                     Page
18-38   Comparative Sensitivity of Mortality From Skin Cancer to
        Various Factors 	 18-63

-------
                                   CHAPTER 6

                                    CLIMATE
SUMMARY

    Stratospheric perturbants are trace gases that are emitted by industrial or
biogenic processes that can change the amount and location of ozone or water
vapor in the stratosphere.  Increases in these trace gases can be expected to
warm the earth directly by reducing the escape of infrared radiation to space.
As stratospheric perturbants, they will indirectly warm the earth by altering
the vertical distribution of ozone-'- and the abundance of water vapor in the
stratosphere.

    The magnitude of warming anticipated from the expected increases in
greenhouse gases is quite large compared to past natural variations in global
temperature.  The anticipated speed of the warming is more rapid than has
occurred before.  The temperature increase from the Wisconsin Ice Age, 18,000
years before present (B.P.), is believed to be 4°C.  Global average temperature
increase in the next 100 years, without curtailment of the growth in the
concentration of greenhouse gases, is likely to be somewhat comparable.

    The major uncertainty in predicting the magnitude of global warming is the
extent to which clouds will amplify or possibly reduce the warming.  The major
uncertainty in predicting the timing is the rate at which the oceans will absorb
heat.  Despite the numerous uncertainties, the realized global temperature rise
would be expected to exceed 1.6°C by 2075 and be as high as 4.5°C if no steps
are taken to limit the rise in the concentrations of greenhouse gases.

    Major weather and climatic changes would be expected to accompany global
warming, although it is impossible to specify details of how the changes will
occur in different regions of the world.  In addition to temperature increases,
changes can be expected in precipitation, storms, and weather patterns.  Current
evidence suggests that significant changes will occur in the dryness or wetness
of many regions and in the frequency and location of extreme weather.

    Policies aimed at reducing chlorofluorocarbons (CFCs) in order to limit
ozone depletion would also reduce global warming.  Increases in other
stratospheric perturbants (e.g., methane and carbon dioxide), while also
.contributing to the greenhouse effect, are expected to buffer the earth from
ozone depletion.  Reducing emissions of these gases to limit global warming
would make the stratosphere more vulnerable to depletion from CFCs.
       If ozone depletion grows to a sufficient magnitude that there is a
decrease in ozone at altitudes below 28 kilometers, something that occurs in
models only in cases where depletion is large, the net effect of the ozone
change would, from that point on, produce a cooling effect.

-------
                                      6-2
FINDINGS

1.  INCREASES IN THE ABUNDANCE OF CFCs AND OTHER TRACE GASES CAN INCREASE GLOBAL
    TROPOSPHERIC SURFACE TEMPERATURES.  THESE GASES CAN ALTER THE VERTICAL
    DISTRIBUTION OF OZONE AND INCREASE STRATOSPHERIC WATER VAPOR. THEREBY
    INFLUENCING GLOBAL WARMING.

    la.  Trace gases that act as stratospheric perturbants also are greenhouse
         gases--as their concentrations increase in the troposphere they will
         retard the escape of infrared radiation from earth, causing global
         warming.

    Ib.  Increases in methane (CH4) will also add water vapor to the
         stratosphere, thereby enhancing global warming.  Methane increases will
         also add ozone to the troposphere, where it acts as a strong greenhouse
         gas that will further increase global warming.

    Ic.  In all model-generated scenarios of ozone depletion, ozone decreases in
         the stratosphere above 28 km.  This allows more ultraviolet radiation
         to penetrate to lower altitudes where the "self-healing effect"
         increases ozone to partially compensate for the ozone loss above.  In
         some scenarios sufficient depletion occurs so that ozone eventually
         decreases at all altitudes.

    Id.  Decreases in ozone at approximately 28 km or above will have a warming
         effect on the Earth.  There  is a small net gain in energy because the
         increase in ultraviolet radiation (UV-B) allowed to reach the earth's
         surface more than compensates for the infrared radiation that is
         allowed to escape due to depletion of ozone above that altitude.

    le.  Below approximately 28 km, increases in ozone are more effective as
         absorbers of infrared radiation.  Consequently, increases in ozone
         below 28 km also will produce a net warming.  In this case, the
         additional UV blocked by more ozone is less than the additional
         infrared that is blocked from escaping the earth.  Conversely, a
         decrease in ozone below 28 km will tend to cool the Earth's surface.

    If.  The direct effect of column  depletion of ozone on global temperatures
         will depend on the magnitude of the depletion.  Until the depletion  is
         of sufficient magnitude that it occurs at the lower part of the column,
         ozone depletion will be a net contributor to global warming.  If the
         stratosphere continues to deplete so that ozone is depleted below 28
         km, this depletion will cause a cooling.  One-dimensional models differ
         from two-dimensional models  in the vertical distribution of ozone
         change, with depletion occurring  at all altitudes  in the higher
         latitudes  in two-dimensional models, rather than just at high
         altitudes.  Thus, according  to 2-D models,  the changes  in radiative
         balance will be  latitude dependent.  At the current time, no studies
         have been  undertaken to determine the net radiative forcing of changes
         projected  by 2-D models.

    Ig.  Radiative  forcing may vary strongly with  changes  in ozone at different
         altitudes  and  latitudes.  Consequently, until  comparisons are made
         between  the models  in  terms  of their global  impact, estimates of  the
         effects  of changes  in  the vertical  column of  ozone  on global warming
         made with  1-D models must be viewed cautiously.   In addition, changed

-------
                                      6-3
         vertical distribution of ozone could influence stratospheric dynamics.

2.   INCREASES IN TRACE GAS CONCENTRATIONS ASSOCIATED WITH STRATOSPHERIC
    MODIFICATION ARE LIKELY TO WARM THE EARTH SIGNIFICANTLY.

    2a.   Two National Academy of Sciences panels have concluded that the
         equilibrium warming for doubling atmospheric concentrations of C02,  or
         for an equivalent increase in the radiative forcing of other trace
         gases, will most likely be between 1.5° and 4.5°C.

    2b.   The magnitude of warming that would be directly associated with
         radiative forcing from increases in trace gases without feedback
         enhancement would increase temperature by approximately 1.2°C for a
         doubling of C02, and approximately an additional 0.45°C for a
         simultaneous doubling of N20 and CH4.  Direct radiative forcing from a
         uniform 1 ppb increase in both CFC-11 and CFC-12 would increase
         temperature by about 0.15°C.

    2c.   The initial warming from direct radiative forcing would change some of
         the geophysical factors that determine the earth's radiative balance
         (i.e., feedbacks will occur)  and these changes would amplify the
         initial warming.  Increased water vapor and altered albedo effects
         (snow and ice melting, reducing the reflection of radiation back to
         space) have been projected by several modeling groups to increase the
         warming by as much as 2.5°C for doubled C02 or its radiative
         equivalent.  Large uncertainties exist about the feedbacks between
         global warming and clouds, which could further amplify, or possibly
         reduce, the magnitude of warming.

    2d.   The three major general circulation modeling groups in the U.S.
         estimate an average global warming of around 4°C for doubled C02 or its
         radiative equivalent.  However, because of uncertainties in the
         representation of the cloud contributions, greater or lesser
         amplifications, including a negative feedback that would reduce the
         warming to 2°C or an even lower value, cannot be ruled out.

    2e.   Global average temperature has been estimated as having risen about
         0.6°C over the last century.   This increase is consistent with general
         predictions of climate models.  Attempts to use these data to derive
         empirically the temperature sensitivity of the earth to a greenhouse
         forcing are not likely to succeed.  Uncertainty about the past
         concentrations of trace gases in the atmosphere, other exogenous
         factors that affect the climate (such as aerosols or solar input), and
         oscillation and instabilities in the internal dynamics of the climate
         system (such as ocean circulation), currently prevent the derivation of
         the earth's temperature sensitivity from examination of the historic
         rise of temperature.  This limitation is likely to remain for more than
         another decade.

    2f.   The global warming associated with increases in ozone-modifying gases
         varies with scenarios of future growth in these -gases.  If the use of
         CFCs grows at 2.5 percent per year through 2050, C02 concentrations
         grow at the 50th percentile rate defined by the NAS (approximately 0.6
         percent per year from 1985 to 2050), N20 concentrations grow at 0.20
         percent per year, and CH4 concentrations grow at 0.017 ppm per year
         (approximately 1.0 percent of current concentrations), then equilibrium

-------
                                  6-4
     temperatures would rise by about 5.6°C by 2075 (relative to observed
     temperature in 1985),  based on a temperature sensitivity of 3°C for
     doubled C02.  Values would be about 50 percent higher for a 4.5°C-based
     temperature sensitivity and about 50 percent lower for 1.5°C.   If CFC
     use remains constant through 2050,  the projected warming would be about
     4.3°C by 2075 (+ 50%), and if use were phased out by 2010,  projected
     warming would be about 4.0°C (+ 50%).

2g.  Efforts to gather worldwide time series data for clouds have begun.  If
     adequate, these data may narrow estimates of the cloud contribution to
     temperature sensitivity within the next decade.  However, because of
     the complexity of this issue, this effort may fail to resolve the large
     uncertainties affecting this aspect of climate.

THE TIMING OF GLOBAL WARMING DEPENDS ON THE RATES AT WHICH GREENHOUSE GASES
INCREASE. THE RATES AT WHICH OTHER FORCINGS SUCH AS VOLCANOES AND SOLAR
RADIATION CHANGE. AND THE RATE AT WHICH OCEANS TAKE UP HEAT AND PARTIALLY
DELAY TEMPERATURE EFFECTS.   A GLOBAL WARMING GREATER THAN VARIATIONS THAT
OCCURRED THIS PAST CENTURY IS EXPECTED IN THE NEXT TEN YEARS IF VOLCANIC AND
SOLAR FACTORS DO NOT SUBSTANTIALLY CHANGE.

3a.  The delay in temperature rise introduced by absorption of heat by the
     oceans can only be roughly estimated.  The simple one-dimensional
     models of oceans that have been used for this purpose do not
     realistically portray the mechanisms for heat transport into the
     oceans.  Instead, these models use eddy diffusion to treat heat in a
     parameterized manner so that heat absorption is consistent with data
     from the paths of transient tracers.  These models indicate that the
     earth will experience substantial delays (on the order of several
     decades) in experiencing the full warming from greenhouse gases.

3b.  The earth's current average temperature is not in equilibrium with the
     radiative forcing from current concentrations of greenhouse gases.
     Consequently, global average temperature would increase in the future
     even if concentrations of gases did not rise any further.  For example,
     if 2°C is the actual sensitivity of the earth's climate system to a C02
     doubling, simple models estimate the current "unrealized warming" to be
     approximately 0.34°C;  for a. 4°C  temperature sensitivity, the current
     unrealized warming would be approximately 1.0°C.

3c.  Only one three-dimensional general circulation model has been used to
     simulate changes in temperature as concentrations of greenhouse gases
     increase over time.  This simulation shows a faster warming than
     predicted by simpler one-dimensional models that use ocean box models
     to simulate time-dependent warming.

3d.  Future uptake of heat by the oceans may change as global warming alters
     ocean circulation, possibly altering the delaying effect of the oceans
     as well as reducing their uptake of C02.

3e.  Inadequate information exists to predict how volcanic or solar forcings
     may change over time.   Analyses done of transient warming assume that
     past levels of volcanic aerosols will continue into the future and that
     solar forcing changes will average out over relatively short periods of
     time.

-------
                                      6-5
4.  WITH A FEW GENERALIZED EXCEPTIONS.  THE CLIMATIC CHANGE ASSOCIATED WITH
    GLOBAL WARMING CANNOT BE RELIABLY PREDICTED ON A REGIONAL BASIS.

    4a.   In general,  as the earth warms,  temperature increases will be greater
         with increasing distance from the equator.

    4b.   Global warming also can be expected to increase precipitation and
         evaporation,  intensifying the hydrological cycle.  While models lack
         sufficient reliability to make projections for any single region, all
         perturbation studies with three-dimensional models (general circulation
         models) show significant regional shifts in dryness and wetness, which
         suggests that shifts in hydrologic conditions will occur throughout the
         world.

    4c.   Current general circulation models represent oceanic, biospheric, and
         cloud processes with insufficient realism to determine how extreme
         weather events and climatic norms are likely to change on a regional
         basis.  For example, one analysis of general circulation model outputs
         suggests that the frequency of extreme climatic conditions will change
         in many regions of the world.   Another model projects increased summer
         drying in mid-latitudes for perturbation studies, utilizing either of
         two different representations of clouds.  Still another analysis
         suggests changes in latitudinal gradients of sea surface temperature
         will play a critical role in determining regional climatic effects.

5.  LIMITING GLOBAL WARMING BY REDUCING EMISSIONS OF STRATOSPHERIC PERTURBANTS
    THAT TEND TO INCREASE OZONE WOULD INCREASE THE STRATOSPHERE'S VULNERABILITY
    TO OZONE DEPLETION.  UNDER SCENARIOS IN WHICH CONTINUED BUFFERING OF OZONE
    DEPLETION BY OTHER TRACE GASES IS ASSUMED. SUBSTANTIAL GLOBAL WARMING
    RESULTS.

    5a.   Decreases in substances with the potential to deplete stratospheric
         ozone--that is, chlorofluorocarbons and nitrous oxides--would decrease
         the rate and magnitude of global warming.

    5b.   Decreases in methane emissions, which have the potential to increase
         stratospheric and tropospheric ozone and thereby buffer ozone
         depletion, would decrease warming in three ways:  by reducing direct
         radiative effects from its presence in the troposphere; by lowering
         water vapor in the stratosphere; and by reducing ozone build-up below
         28 km.

    5c.   Decreases in C02 emissions would decrease global warming, but would
         also have the effect of increasing the "stratosphere's vulnerability to
         ozone depletion.

    5d.   Decreases in carbon monoxide concentrations, which may occur as energy
         production practices change, could result in decreases in methane
         concentrations by increasing OH-radical abundance which, in turn, would
         shorten the lifetime of methane and could shorten the lifetime of
         methyl chloroform and CFC-22.

-------
                                      6-6
6.   ADDITIONAL RESEARCH IS NEEDED ON CLIMATE TO REDUCE UNCERTAINTIES ABOUT
    GLOBAL WARMING ASSOCIATED WITH TRACE GAS GROWTH.

    6a.  The key to improving the accuracy of estimates of global temperature
         sensitivity is to acquire a better understanding of the effect of
         clouds.  This recommendation has been made by numerous groups over the
         last decade;  yet research devoted to this issue remains relatively
         small.

    6b.  An increased understanding of ocean circulation is critical to
         improving estimates of timing and regional projections,

    6c.  The effect of climate on biological systems and soils and their impact
         on climate must be modeled if regional estimates of climate change are
         to be developed.

    6d.  A better understanding of the radiative properties of CFC-113 and other
         compounds is needed for estimating the effects of this compound on
         climate.

    6e.  Experiments with three-dimensional models that have altered scenarios
         of vertical ozone need to be undertaken to assess the possible impacts
         on the magnitude of global warming and on general circulation.

-------
                                      6-7
THE GREENHOUSE THEORY

    The earth is heated by solar radiation that penetrates the atmosphere in a
variety of wavelengths; the greatest quantity of energy that reaches the earth
is in the visible part of the spectrum.  In turn, the earth is cooled by the
infrared radiation given off to space.  The average surface temperature that
results is a balance of this heating and cooling.

    Since Tyndall discovered in 1861 that water vapcr can absorb infrared
radiation (Tyndall 1861),  it has been understood that increases in the abundance
of infrared absorbing gases (the "greenhouse" gases) in the atmosphere can raise
the earth's temperature by reducing the escape of this energy to space.  It has
also long been known that other aspects of atmospheric composition, including
ozone and aerosol abundance, can influence the quantity of radiation
(ultraviolet and visible)  that reaches the earth's surface.

    A variety of studies and international assessments (NAS 1979; NAS 1982; WHO
1986a) have been conducted over the last two decades concerning the potential
for increasing concentrations of trace gases to raise global temperature and
alter weather and climate.  All of the reviews conducted have come to general
agreement that increases in greenhouse gases will lead to a global warming.

    This chapter reviews studies on how the gases that can alter the
stratosphere create a greenhouse effect and how the alterations in the
stratosphere itself can add to or reduce the magnitude of the greenhouse
warming.  The chapter is divided into five sections:

    The section on radiative forcing examines evidence that gases that could
perturb the stratosphere are also gases that could retard the escape of infrared
radiation, thereby forcing an increase in the earth's surface temperature.  In
particular, this section analyzes potential indirect greenhouse effects of these
gases by reviewing how changes they cause in the stratosphere can add to or
subtract from the direct radiative forcing of any of the gases.  This section
also reviews studies that estimate the magnitude of warming that would directly
result from radiative forcing in the absence of feedbacks (that is, changes in
the earth's surface or atmosphere that could alter the warming).

    A section on the earth's ultimate temperature sensitivity to radiative
forcing reviews evidence that bears on the magnitude of feedbacks that would
amplify or dampen the initial radiative warming.  Uncertainties in current
knowledge about the magnitude of the feedbacks is highlighted.

    A section on timing reviews those factors in the earth's geophysical system
that will delay the full experience of the temperature equilibrium associated
with specific increases in greenhouse gases.  In particular, this section
reviews evidence on the uncertainties associated with oceanic heat absorbtion.

    A section on regional  climate changes associated with global warming focuses
on changes in regional weather and climatic patterns which could alter the basic
environmental conditions of the planet.

    The final section on the effects of possible control of greenhouse gases,
analyzes the possible impact that curtailing emissions of various greenhouse
gases (C02 and CH4) to limit global warming would have on stratospheric ozone.

    Together the five sections cover the major issues that will determine how

-------
                                      6-8
the greenhouse effect will unfold with time and how that evolution is inter-
related to risks to the stratosphere.

RADIATIVE FORCING BY INCREASES IN GREENHOUSE GASES

    By their influence on chemical or physical processes, a large number of
gases can influence the structure and composition of the stratosphere.  During
their residence in the troposphere,  these gases, alternatively called
stratospheric perturbants, trace species, radiatively important trace species,
or greenhouse gases, retard the flux of infrared radiation from the earth's
surface to space.  Exhibit 6-1 shows the multiple effects that these gases have
on atmospheric processes.  Carbon dioxide (C02), the most important greenhouse
gas, cools the stratosphere and, in most models, increases ozone (NAS 1984 and
WHO 1986b).   Methane (CH4) has three major effects on warming:  its direct
radiative effect in the troposphere; its effect in adding ozone to the
atmosphere below 28 kilometers; and its effect  in adding water vapor to the
stratosphere (NAS 1984).   Nitrous oxide (N02) warms the surface by its presence
in the stratosphere and depletes ozone in the stratosphere unless interfered
with by high chlorine levels (Ramanathan et al. 1985 and Stolarski, personal
communication).

    While small uncertainties still exist in our understanding of the radiation
that is absorbed by each gas, with a few exceptions (CFC-113, for instance), the
absorption feature of these gases are well characterized, as shown in Exhibit
6-2 (WHO 1986b; Luther and Ellingson 1985; and  Ramanathan et al. 1985 for
discussion).  Consequently, increases in the concentrations of these gases will
reduce the escape of infrared radiation from earth, providing a radiative
forcing of the earth's climate system.  When computing the magnitude of that
forcing for multiple perturbants, care must be  taken in considering possible
overlaps that would reduce net forcing (Wang, Wuebbles, and Washington 1985;
Luther and Ellingson 1985).

    Studies of the direct radiative effects of  increases in infrared-absorbing
gases all produce approximately the same result -- that a doubling of C02  (or
equivalent) eventually will raise global temperatures  1.26°C if no other factors
in the geosystem are changed (Ramanathan et al. 1985;  Schlesinger and Mitchell
1985; Hansen et al. 1984).

    Exhibit 6-3 shows the relative effectiveness of a  1 ppm increase  in such
greenhouse gases.  Notice that chlorofluorocarbons and methane have much greater
impact per molecule than C02.  The reason is that they 'cover' a part of the
infrared spectrum that is  'open' or 'transparent' to much of the infrared
radiation given off by earth, whereas the region of C02  absorption is already
somewhat  'opaque' or  'closed'  to part of the infrared  spectrum.

Changes In Stratospheric  Structure and Composition Can Change Radiative Forcing

    In photochemical models, the structure and  composition of the  stratosphere
changes as gases such as  CFC and methane increase in the atmosphere.  These
changes can influence the radiative forcing of  the earth by altering  fluxes of
infrared or ultraviolet  radiation from or to its surface.-  The critical factors
of  the stratosphere's organization  that  influence radiative balance are the

-------
                                      6-9
                                  EXHIBIT 6-1

                  Stratospheric Perturbants and Their Effects
                         Direct Effect
                           on Global
                        Temperature from
                          Tropospheric
                            Presence
  Physical
 Effect on
Stratosphere
  Effect on
Column Ozone
Carbon Dioxide Increases-'-
Me thane Increases *-
Nitrous oxide Increases-*-
Chlorofluorocarbons Increases
Other Trace Gases Increases^
(methyl chloroform,
carbon tetrachloride ,
Halons)
Cools2
Adds water
vapor ,
hydrogen
Adds nitrogen1'
Adds chlorine4
Adds catalytic
species to
! 4
stratosphere^
Increases differently
at different
latitudes-^
Increases at some
latitudes^
Decreases"
Decreases4
Decreases^
1 Ramanathan et al.  1985.

2 Connell and Wuebbles 1986.

o
J Isaksen, personal communication.

4 NAS 1984.

^ Isaksen and Stordal 1986.

  NAS (1984) notes the direct effect of N20 on column ozone.  In the presence of
high levels of chlorine, N20 may interfere with the catalytic cycle of chlorine,
reducing net depletion (Stolarski, personal communication).

-------
                                       6-10


                                   EXHIBIT 6-2

                   Absorption Characteristics  of Trace Gases
                                       A (jam)
        20  18  16   14
                         12
                                 10
            I   I
                      FI3
                   FII6
              FI4
                     CHC!3

                   CH2CI2  CH2CI2
                        =>   <— >
                        F22
     CHCI5
                                      F22
                                                 CHF3
                                                C2H2
            F22
                 CH,CCI3CCU
         SO,

         I*-"
                         fll
                              ,2
            NjO
                                                           CH4
                                        03
              t
                        l
                             l
                                  I
-H20-
 J	1	I	L
        500   600   700   800   900  1000  1100  1200  1300  1400  1500  1600  1700
Atmospheric  trace gases absorb outgoing infrared radiation  of various
wavelengths.   Carbon dioxide absorbs  much of the-radiation  in the spectral
region between 13/u m and 20^ m.   Several trace gases,  including
chlorofluorocarbons 11 and 12  (Fll  and F12)  have strong  absorption features in
the spectral region of 7/i m to 13^  m.   This  spectral region is called the
atmospheric  "window," since in this region the atmosphere  is relatively
transparent.
Source:  WHO  1986.

-------
                          6-11


                      EXHIBIT 6-3

Radiative Forcing for a Uniform Increase in Trace Gases
Temperature Sensitivity
Compound (°C/ppb)
C02
CH4
N20
CFC-11
CFC-12
CFC-13
Halon 1301
F-116
CC1
4
CHC1
3
F-14
HCFC-22
CH Cl
2 2
CH CC1
3 3
C H
2 2
SO
2
.000004
.0001
.001
.07
.08
.10
.10
.08
.05

.04

.04
.03
.02

.01

.01

.01
          Source:  Adapted from Ramanathan
                   et al. 1985.

-------
                                      6-12
amount and vertical distribution of water vapor located in the stratosphere and
the vertical distribution of ozone.  One-dimensional and two-dimensional models
of the stratosphere project changes in the vertical organization of ozone in
response to increasing concentrations of CFCs and other gases (Connell and
Wuebbles 1986; Stordal and Isaksen 1986).  These changes are significant.  Two-
dimensional models also show that the changes are latitude dependent (Stordal
and Isaksen 1986).   Increases in water vapor (due to increasing methane) are
also projected for most scenarios of trace gas growth (Connell and Wuebbles
1986).

    Changes in ozone abundance can have two effects on radiation.  They can
alter the penetration of UV-B and visible radiation from space and the escape of
infrared from the surface.  Ozone decreases anywhere in the column will always
allow more UV-B and visible radiation (in the Hartley-Muggins and Chanpuis
bands) to reach the earth's surface.  Ozone increases will decrease UV-B
penetration.  Ozone decreases will allow more infrared to escape, but the effect
of ozone changes on infrared escape varies substantially with altitude.  Ozone
increments added near the troposphere produce the largest increases in surface
temperature.  This is because greenhouse efficiency is proportional to the
difference in the temperature of the radiation that is absorbed by the ozone
(essentially the ground temperature) and the radiation that is emitted by the
ozone (the local atmospheric temperature) (Lacis, Wuebbles, and Logan 1986).  As
a result, a net column depletion that consists of depletion of ozone in the
upper atmosphere and of increases of ozone in the lower atmosphere, will
increase the infrared radiation that is blocked from escaping from earth.

    Exhibit 6-4 shows the relationship of ozone and altitude to increases in
temperature (Lacis, Wuebbles and Logan 1986).  At 12 km, the net sensitivity of
the surface warming to a given increase in ozone is approximately 15 times
greater than at 20 km.  Above 28 km an ozone decrease actually would produce a
small net temperature decrease.  Exhibit 6-5 shows the relationship between
water vapor, altitude, and radiative forcing.  However, increases in water vapor
always produce a warming, with maximum effect for water vapor increases at 14 km
(Lacis and Wuebbles 1986).

    Scenarios in which total column ozone depletion is small combine two
effects:  a large decrease above 28 km and a smaller increase in the lower
stratosphere  (self-healing effect and methane) and troposphere (from methane
increases).  In these cases, the total net radiative forcing of the planet will
be increased  (Lacis and Wuebbles 1986).

    In scenarios in which total depletion becomes large enough that ozone
decreases at altitudes below 28 kilometers, the total radiative forcing from
changes in vertical ozone will begin to favor cooling.

    Care must be taken in estimating vertical changes in ozone from
one-dimensional models; the forcing may be different than predicted.  The reason
is that, as two-dimensional chemical models show, depletion levels will vary
with  latitude.  At higher latitudes, there may be less ozone increase at low
altitudes.  At mid-latitudes, where industrial pollution is important, ozone
increases are already evident.  Nearer the equator, tropospheric increases in
ozone from CH4 will be more important and help balance depletion more.
Consequently, the radiative forcing is not likely to be equal at all latitudes.

-------
                                      6-13


                                  EXHIBIT 6-4

         Effects of Vertical Ozone Distribution on Surface Temperature
             60.DL i  i  .  i
             50.0
   Height    30. D
(Kilometers)
             20.0
             10.0 -
                                                  REGION II
                                               (ozone increases
                                                cause cooling)
                       REGION I
                    (ozone increases
                     cause warming)
               -0.01
o.oo
0.01
                Relative Temperature  Change (°C per Dobson Unit)
Sensitivity of  global  surface temperature to changes  in vertical ozone
distribution.   The  heavy solid line is a least  square fit to model results for
an increase in  ozone of 10 Dobson units (approximately 3%)  added to each
vertical layer.   Above 28 kilometers, increases  in ozone cause a net cooling of
the surface.  Below 28 kilometers, increases in ozone cause a net warming of the
surface.
Source:  Lacis,  Wuebbles,  and Logan 1986.

-------
                                     6-14
                                  EXHIBIT 6-5
                 Water Vapor, Altitude,  and Radiative  Forcing
              60.0
              50.0 -
              40.0 -
   Height
(Kilometers)
             10.0
              n.
             30.0-
             20. Or~
                -o.i    -a.a    o.i     0.2    0.3     0.4    0.5

                     Temperature  Sensitivity  (°G/CM(STP))
Increases in water vapor at all altitudes will warm the surface.
warming occurs for increases at 14 kilometers.
   0-6     0-7
The maximum
Source:   Lacis,  personal communication.

-------
                                      6-15
As a consequence, the positive feedbacks from albedo changes (snow and ice
melting) may be different (Lacis and Wuebbles 1986).

ULTIMATE TEMPERATURE SENSITIVITY

    To understand the possible role of feedbacks on the warming caused solely by
the initial radiative forcing, one must model the effects of that warming on the
earth.  A number of modeling analyses have been used to examine the earth's
responses to radiative forcing and how those responses can increase or decrease
the global temperature rise (Hansen et al. 1984; Washington and Meehl 1983;
Manabe and Wetherald 1975).   Exhibit 6-6 shows an analysis by Hansen et al.
(1984) of the changes that would occur with the initial warming.  While the
exact responses are peculiar to their model, they are representative of the
feedbacks found by other modeling groups.  Thus, as Exhibit 6-6 shows, increases
in water vapor due to an initial warming and snow and ice melting will enhance
the warming significantly.  Water vapor has a warming effect because it is a
greenhouse gas.  The melting of snow and ice enhances warming because snow and
ice reflect more visible radiation back to space than do land and water --so
snow and ice melting will reduce the amount of visible radiation reflected back
to space (Hansen et al. 1984).

    Consistency between models is quite strong.  In Hansen's model (Hansen et
al. 1984),  the magnitude of the global warming is 2.0°C without cloud or snow
and ice feedback.  Manabe and Wetherald (1975) obtain a 2°C warming for
perturbation experiments in which their model assumes fixed clouds.  When Manabe
allows clouds to vary, his model shows that the warming is amplified to 4°C
(Manabe and Wetherald 1986) .   Washington and Meehl (1984), who also allow clouds
to vary, obtain a global temperature sensitivity of 3.5°C.

    Despite this apparent consensus, however, care should be taken in
interpreting this agreement.   The exact nature of future cloud changes remain
uncertain.   Somerville and Remer (1984) propose a mechanism by which clouds
might dampen the warming.  Manabe, Washington, and Hansen all counsel caution
(personal communications), and each is uncertain of his model's portrayal of
cloud feedback processes.  The uncertainty associated with clouds is, of course,
completely consistent with the National Academy of Sciences (NAS) estimate of
temperature sensitivity.  Instead of using the range of temperature responses
produced by models, 2° to 4°C, the NAS has recommended a range of 1.5°C to
4.5°C, which characterizes their judgment about the possible range of
temperature sensitivity.  Unfortunately, the possibility that the earth's
temperature sensitivity lies outside this range cannot be excluded (NAS 1982).

Fast: and Near-Term Rises in Temperature Cannot be Expected
to Resolve Uncertainties about Actual Climate Sensitivity

    A different approach other than modeling has been suggested as a means for
estimating the temperature sensitivity of the earth to a given radiative forcing
--to compare historical temperature rises to past increases in greenhouse gases
(Webb and Wigley 1985; Broecker 1986).  Unfortunately, this approach cannot
provide precise estimates of temperature sensitivity because many of the factors
that may have contributed to past temperature changes are themselves uncertain.
Exhibit 6-7 from Hansen et al. (1984) shows the extent to which the derived
temperature sensitivity would depend on knowing initial trace gas concentrations
and the rise in temperature.   Uncertainty exists about both past temperature

-------
                                       6-16


                                   EXHIBIT 6-6

            Temperature Sensitivity  to Climatic Feedback Mechanisms
                                   Lapse
                                   Rate
                                 (-0.2K/km)
             C02
             (X2)
  H20     H?0
(xl.32) Vertical
      Distribution
Ground  Cloud   Cloud
Albedo  Height   Cover
J-0.01)  ?>10mb) (-1.7%)
                                                      Fc~1.3
The earth's responses to an  initial  radiative forcing will determine the final
temperature change.  The direct  radiative forcing of doubled C02 would increase
the global average temperature by 1.26°C.   Subsequent melting of snow and ice
would affect water vapor (H20) concentrations and distributions, amplifying the
initial warming by approximately 1.6°C.   The effect of melting on the earth's
reflectivity, or ground albedo,  would amplify the warming by 0.4°C.  Changes in
cloud cover would add 0.8°C.  Total  warming would thus equal approximately 4°C.
Source:  Hansen et al. 1984.

-------
                                     6-17
                                 EXHIBIT 6-7

         Empirical Estimates of Climate Sensitivity are Sensitive  to
                Estimates of Historical Temperature Increases
                         and Trace Gas Concentrations
              12
              •c
              10
            go*
            2
            2
             04
             02
              00
                ovc
                                                    *C 0
                                                  to)
              OO
The vertical axis measures the increase in global average temperature that
.occurred from 1850 to 1980.   The horizontal axis indicates the earth's
temperature sensitivity to doubled C02 concentrations.  The curves represent
temperature changes consistent with different historical changes in trace gas
concentrations.   Panel "a" shows the uncertainty for C02 alone.  Panel "b"
incorporates other trace gases.  If one assumes a historical temperature rise of
0.5°C, Panel "b" shows that temperature sensitivity could vary from
approximately 2.3°C to 4.5°C, within a reasonably accepted range of 260 ppm to
280 ppm for pre-industrial C02 concentrations.
Source:  Hansen et al.  1984.

-------
                                     6-18
change and past concentrations of C02 and trace gases.   Estimates of
pre-Industrial values of C02 based on ice core and other data range from 260 to
290 ppm (Oeschger et al. 1982; Neftel et al.  1982; Pearman et al. 1986; and Peng
et al. 1983).   To complicate matters further,  uncertainty also exists about
changes in other kinds of forcing that can influence radiative balance, such as
changes in solar radiation and aerosols from volcanoes  (Bradley and Jones 1985).
Failure to include these possible changes in forcing could bias estimates of
temperature sensitivity.  In addition, uncertainties about instabilities and
oscillations in the internal dynamics of the climate/ocean system are also
unknown, and could produce variation in observed temperatures (Hoffert and
Flannery 1985).  For example, in a recent run of the Goddard Institute for Space
Studies (GISS) general circulation model, without any change for radiative
forcing, global average temperature varied 0.2°C from the 100-year mean in a
period of 100 years (Hansen et al. 1986).  Based on this estimate of "natural
variation which is entirely expected according to model runs (Hansen et al.
1986b)," an accurate assessment of temperature sensitivity will not be possible
even after the temperature has risen another 0.5°C.

Clouds are the Major Uncertainty

    The International Satellite Cloud Climatology Project is seeking to collect
a global database of radiation information that can be used to improve our
understanding of clouds and their response to global warming (WHO 1984).
Several research obstacles including insufficient satellite coverage,
difficulties in converting radiation data to accurate representation of clouds,
and the complexity of modeling cloud systems and their effects, must be overcome
for this endeavor to be successful.  The goal of this effort is to provide data
that allows us to narrow considerably our uncertainty about clouds.  However,
the possibility exists  that the data from this international project will
provide a useful beginning for resolving our questions about this important
climate feedback (Luther and MacCracken 1985) .

THE TIMING OF GLOBAL WARMING

    The timing of global warming  is of particular importance to assessing  the
associated risks.  Knowing the radiative forcing and the temperature sensitivity
of earth to that forcing are  insufficient for assessing the time-dependent
evolution of the climate system.  To determine timing, oceanic heat absorption
must be considered.

    In  this section, we review the results of two research teams that  have
considered all three components of climate change: changes in greenhouse gases,
temperature sensitivity, and  oceanic heat absorption.

    The timing estimates represented here can be broken down into efforts  that
utilized one-dimensional and  three-dimensional models.  Hansen et al.  (1981)
used  a  one-dimensional  model  to estimate the  rise of temperatures from historic
levels.  Details of the model, a  parameterized version of which was also used by
Hoffman, Wells, and Titus  (1985),  are presented  in Attachment A.  Before
proceeding, however, a  brief  discussion  of the radiative forcings and  how  the
model deals with oceanic heat uptake  is warranted.

    The rate at which infrared-absorbing gases are expected to increase has been
presented  elsewhere in  this  document  (Chapters 2,  3, and 4).  While changes  in
volcanic aerosols and solar  radiation could accelerate or retard the increase  in
net radiative  forcing,  changes in these  factors  will not be examined here,

-------
                                      6-19


except to say that the radiation output of the sun has been measured since early
1980 by the NASA Solar Maximum Mission Satellite, and unpublished data show a
decrease in the last four years.  Previous studies have speculated on various
cycles for changes in solar radiation (Newkirk Jr. 1983; Hoyt 1979; and
Gilliland and Schneider 1984).  At this time, there is no basis for predicting
future trends (Hoffert and Flannery 1985).

    Aerosols from volcanoes depend on the rate and nature of volcanic eruptions.
In general, the eruptions must propel aerosols into the stratosphere in order
to produce a climatic effect.  For example, despite its magnitude the Mt. St.
Helens eruption had negligible effects on climate because of the lack of sulfur
in the explosion and because of its sideways ballistics.  The 1982 El Chichon
eruption had a very large climatic effect, possibly placing more aerosols into
the stratosphere than any eruption since Krakatoa (MacCracken and Luther 1984;
NASA 1982; and Pollack 1983).  Future eruptions cannot yet be predicted, but
volcanic aerosols from single eruptions are unlikely to have more than a two- to
three-year influence on radiative forcing (Hoffert and Flannery 1985).

    Oceanic heat absorption is the major factor that could influence the timing
of the warming (NAS 1979).  At this time, it is impossible to describe
accurately all the processes that will influence the ocean system's ability to
take up or give off heat.  Data and knowledge are both lacking.  The ocean is a
complex three-dimensional system in which heat-exchange processes operate by
many means, including both  convection and diffusion.  The specific topography
of ocean floors is important, as is the salt and water density.  Initial
attempts are now being made to model the ocean in more of its complexity
(Hoffert and Flannery 1985; Sarmiento and Bryan 1982; and Woods 1985).  However,
an adequate ocean general circulation model does not yet exist.

    In the absence of an adequate three-dimensional representation of oceanic
processes and heat uptake, representation of the oceans has been relegated to
simple parametric models of oceanic heat uptake that use diffusion as a
surrogate for other mechanisms to absorb heat.  Such models, while clearly not
portraying mechanisms, have the advantage of allowing different assumptions to
be used about the quantity of heat uptake in perturbation studies, thereby
providing some understanding of the variation in estimates of warming that is
attributable to uncertainties in oceanic heat absorption.  One-dimensional (1-D)
ocean models that use eddy diffusion to transport heat have been developed for
this purpose.  The models have been calibrated to reproduce the behavior of
transient tracers (Broecker and Peng 1982).  Such models have allowed
experiments to be conducted that allow for changes in the rate of heat
.absorption.  Exhibit 6-8 shows the logical flow of how these models can be used
for estimating transient atmospheric temperatures.  Note the parameter "k,"
which can be exogenously manipulated to test differing rates of heat absorption.

    Hansen et al. (1984) used a one-dimensional radiative convective model
coupled to an ocean box model to estimate future realized and unrealized warming
through time (Exhibit 6-9).  The scenarios used for trace gas growth are
approximately the same as those discussed in Chapters 2 and 4.  (see Attachment
B for details).  The distance between the dotted and solid lines represents the

-------
                                                  6-20
                                              EXHIBIT 6-8


                     Relationship  of Radiative  Forcing,  Ocean Keat Uptake,
                                and  Realized and  Unrealized Wanning
                                            Input:

                                                Set
                                             Temperature
                                              Sensitivity
Input:
  Scenarios of
  Greenhouse
   Gas Rise
 Atmospheric
Concentrations
    Rise
               2°C to 4° C
 Equilibrium
Temperature
 Estimated
                                    Heat Flux
                                  Air
                                Warms
                                Slightly
                             Heat
                                               Heat
                                 Air
                             Temperature
                             Slowly Rises
                                                                               Difference at Any Time
                                                                                 Unrealized Warming
           Input:
           Set Diffusivity Surrogate for Oceanic Heat Absorption
           (1.7 cm2/sec used in this analysis)
Top Layers of
Ocean Slowly
Absorb Heat and
Warm
1 ,
Heat
Passed to
Deeper Layers
of Ocean
       The  model  allows  input parameters to  be set  and  then consistently  simulates
       time-dependent evolution of the system.

-------
                                       6-21


                                   EXHIBIT 6-9

                      Transient Estimates of Global Warming
                    30
                   25
                   20
                 u
                    10
                   0.5
—— No Ocean Heal Coooc''»
— — — Mned Lsye' (nQml "eat Capacity
	Mned Layer » Thermocline

      (k  m cm* i"')
                    1850  '830
        1920
        Date
i960
                                              2000
Change in global  average surface temperature due  to  increases in C02 and other
trace gases from  1850  to 2007.   The solid line  shows  the warming that would
occur if oceanic  heat  absorption did not delay  global warming.   The dashed line
represents oceanic heat  absorption by only the  trop layers of the ocean.  The
dotted lines represent heat absorption by the lower  layers of the ocean, and
test two assumptions about the  rate of heat uptake,  "k".   The distance between
the solid line, which  shows "equilibrium" temperature,  and the dotted lines,
which show "transient" temperature, equals the  "unrealized" warming.  Note that
changes in this exhibit  are relative to 1850.
Source:  Hansen  et  al.  1984.

-------
                                      6-22
unrealized warming for two different assumptions about heat uptake.  Note that
in the time period examined, that is, until the year 2007, the projected
temperatures are quite similar for different heat absorption.  Hoffman,  Wells,
and Titus (1985) also use a modified version of this model to make estimates of
temperature change farther into the future (see Exhibit 6-10).  They simulated a
similar scenario, using 2°C and 4°C as the earth's temperature sensitivities to
doubled C02.

    Only two time-dependent simulations have been conducted using a general
circulation model (Hansen et al.  1986).  The representation of the oceans in
that model, while still unrealistic, does use three-dimensional heat absorption
based on transient tracers.  Exhibit 6-11 shows their results, which predict a
faster warming than do their 1-D models.  Since the results are from the same
modeling group, the question arises of whether 1-D time-dependent
representations underestimate the actual rate of global warming.  Another
question that deserves close attention is whether other parameterizations,  such
as upwelling diffusion, should be used; these would give smaller results
(Hoffert and Flannery 1985).

REGIONAL CHANGES IN CLIMATE DUE TO GLOBAL WARMING

    Only one change in regional climate can be projected with a high degree of
confidence.  Temperatures will rise more for latitude bands as distance from the
equator increases.  For model runs in which the atmospheric C02 concentration is
doubled, Hansen et al. (1986) and Washington and Meehl (1983) find a ratio of
approximately two to one between warming near the poles and near the equator.
Manabe and Wetherald (1975) show a larger gradient of up to seven to one.
Almost as certain as the latitudinal gradient is that global warming can be
expected to intensify the hydrological cycle, increasing global mean
precipitation and evaporation (Schlesinger and Mitchell 1985).  In response to a
quadrupling of C02 (with a model with a 2°C sensitivity; no cloud feedback
allowed), Manabe and Wetherald (1985) find a. 7% increase in both precipitation
and evaporation.  In a similar experiment Washington and Meehl  (1984) showed an
increase of 6% in precipitation.  For a doubled C02 experiment, which yields a
temperature increase of approximately 4°C, the Goddard Institute for Space
Studies model predicts an 11% increase in precipitation and evaporation (Rind
and Lebedeff 1984).

    All three-dimensional models predict significant shifts  in dryness and
wetness throughout the world.  For the U.S., the Goddard Institute for Space
Studies model  (Rind and Lebedeff 1984) predicts an average increase in runoff of
18%.  Exhibit  6-12 shows the distribution of changes predicted by that model.
Estimates of the distribution of runoff change from the general circulation
model of Princeton's Geophysical Fluid Dynamics Laboratory and National Center
for Atmospheric Research are similar to the Goddard Institute for Space Studies
run, even though they vary  greatly in regional detail  (Manabe and Wetherald
1985; Washington and Meehl  1984).

    The reliability of regional projections from general circulation models  is
weak, however.  Current general circulation models lack realism in representing
important ocean, biospheric, and cloud processes that  are- critical to
establishing precise estimates of hydrological change  for particular regions.
At this time,  no  interactive climate/ocean model has been successfully developed
and published.  Yet we know from events such as the 1983  El  Nino that changes in

-------
                                         6-23



                                     EXHIBIT 6-10


                            Expected Temperature Increases
              5 i-
            O
            TJ
            «
            U
            Ol
            c
            O
            U
            O
            L
            (T
Unrealized
V7arraing
(in pipeline)
                                                                       Equilibrium
                                                                       Terrperature
                                                                       Rise
                                     Realized
                                     Tertperature
                                     Rise
I960
                        199O
20OO
2010
2O20
203O
   Change in global average  surface temperature due ~to increases  in C02 and trace

   gases from 1980 to 2030.  A modified version of the Goddard Institute for Space

   Studies model (Hansen et  al.  1981)  was used.  The model's  thermal sensitivity is
   4°C for doubled C02 concentrations.   Note that changes  in  this figure are

   relative to 1980.
   Source:   Hoffman, Wells,  and Titus 1985.

-------
                                      6-24


                                  EXHIBIT 6-11

        Results of Transient Analysis Using a General Circulation Model
 _
 i
 o
 B
 Q
 cr
 UJ
 LJ
           OBSERVATIONS

           SCENARIO A

           SCENARIO B
                        ILL
_LL
                                      DATE
Only two time-dependent simulations have been  conducted using a general
circulation model.  The results, shown above,  indicate  an increase in global
average temperature of approximately 0.9°C by  the  year  2000 for Scenario A
(which is a continuation of current rates of growth  in  trace gases).   Scenario B
(which reflects reduced rates of trace gas growth) indicates a warming of about
0.5°C by 2000.  Scenario A achieves a radiative  forcing equivalent to that of
doubled C02 about 40 years from now; Scenario  B  requires 75 years.
Source:  Hansen et al.  1986.

-------
                                     6-25


                                 EXHIBIT 6-12

                      Regions of U.S.:  Change in Runoff
      60

      50

      40

      30
  ui
  a
  x
X
      |0


       0


     -10

     -20

     -90
The general circulation model of the Goddard Institute for Space Studies was
used to simulate changes in hydrology due to increases in C02 and other trace
gases.  Estimates were made for 23 gridded regions of the United States.  The
graph above shows the distribution of changes in precipitation runoff:
increases in 14 regions and decreases in 9 regions.
Source:  Rind and Lebedeff 1984.

-------
                                     6-26
oceanic circulation can influence regional climates (Wigley, Angell,  and Jones
1985).   None of the models adequately represents the hydrological process,
especially the role of the biosystem.  In fact,  in all models the biosystem
remains constant as climate changes and C02 increases.  In reality we can expect
large alterations in ecosystems in response to the effects of climate change and
the response to the direct physiological effects of C02 on photosynthesis
(Strain 1985).   These changes would alter hydrological processes in important
ways.

    The possibility still exists, however, that for large enough areas, some
robust projections can be made.  The general circulation model at Princeton's
Geophysical Fluid Dynamics Laboratory has been run using two very different
cloud regimes - - one in which clouds are prescribed and one in which they
change.  Manabe and Wetherald (1986) find that in both simulations summer drying
occurs at mid-latitudes:  "... it seems significant that all the experiments
discussed in this report indicate C02-induced summer reduction and winter
enhancement of soil wetness over extensive, mid-continental regions in middle
and high latitudes ... it is likely that the basic conclusion is valid despite
the imperfections of the model."  If Manabe is correct, it means that, while
regional changes are difficult to state with confidence, some can be projected.

EFFECTS ON THE STRATOSPHERE OF POSSIBLE CONTROL OF GREENHOUSE GASES

    Stratospheric perturbants that are greenhouse gases may also alter
stratospheric structure and composition.  Limiting chlorofluorocarbon emissions
in the future would decrease both stratospheric ozone depletion and global
warming.  This relationship is straightforward.  For other perturbants, however,
the situation is more ambiguous.

    Miller and Mintzer  (1986) point out that emissions of some stratospheric
perturbants may eventually be reduced by governments in an effort to limit
global warming.  Targets could include nitrous oxide  (N20), carbon dioxide
(C02), or methane  (CH4).  Decreases in nitrous oxide would lower global
temperatures (Ramanathan et al. 1985).  The net effect on ozone depends on
chlorine concentrations (Stolarski, personal communication).  In many scenarios
considered plausible, lowering N20 to reduce global warming might actually
exacerbate ozone depletion.  Decreases in methane would reduce global warming  in
three ways:  by reducing its direct radiative effects while it's in the
troposphere  (Ramanathan et al. 1985); by altering its indirect effects on
increasing tropospheric and lower stratospheric ozone  (NAS  1984); and by
reducing the amount of water vapor added  to the stratosphere  (NAS 1984).
Methane emissions  might be limited by altering rice cropping practices, by
altering livestock rearing practices  (and  through biotechnology to reduce
livestock methaneogenis), by reducing forest burning, or by controlling pipeline
leaks.  Methane emissions might also be limited through indirect means, that  is
by controlling  carbon monoxide emissions  that would reduce  the lifetime of
methane molecules  (Miller and Mintzer 1986).  As a consequence, however, a
decrease in  methane would allow  increasing CFC concentrations to deplete column
ozone more effectively.

    Carbon dioxide cools  the  stratosphere, thus slowing the process of ozone
loss.  Consequently, model scenarios  that  include C02  growth  show reductions  in
ozone  depletion (Connell  and Wuebbles 1986).  Efforts  to  limit the growth of  C02
could  include  altering  energy mixes  (e.g., using more  nuclear or solar energy),
conservation (less energy to  do  the  same  work), and altering  land clearing
practices  (Seidel  and Keyes  1983).   Such  actions to limit C02 could

-------
                                      6-27
inadvertently increase the vulnerability of the stratospheric ozone layer.

    Decreases in carbon monoxide (CO) emissions from combustion are projected to
occur (Kavanaugh 1986).  Decreases in CO emissions may also be possible if
forest burning decreases (Hoffman and Wells 1986).  Since carbon monoxide plays
a large role in OH concentrations, which, in turn, influence the lifetime of
methane, changes in carbon monoxide emissions could alter methane concentrations
(Khalil and Rasmussen 1985).  If CO concentrations decrease, the lifetime of
methane (CH4) would decrease, and there would be a tendency to lower
concentrations (Thompson and Cicerone 1986).   Decreasing CH4 concentrations
would not only decrease global warming but would also make the ozone layer more
vulnerable to depletion.

-------
                                      6-28


                                  ATTACHMENT A

             DESCRIPTION OF MODEL TO BE USED IN INTEGRATING CHAPTER


     The model used in the  integrated assessment chapter  at  the  end of this
 document was adapted from  a one-dimensional  radiative-convective  model developed
 by  the Goddard Institute for Space  Studies for estimating temperature increases
 associated with atmospheric C02 rises  (Hansen et  al.  1981).

     The model computes vertical temperature  profiles  over time  from net
 radiative and convective energy fluxes.   Radiative  fluxes,  in turn,  depend  on
 changes in atmospheric gases, and on the  associated feedback  effects.   The
 parameterized equation used here is based on an empirical fit to  the
 radiative-convective model.

     The fitted equation, described  in Hansen et al. (1981), was developed for
 C02  only.  It was later modified by Lacis (personal communication)  to
 incorporate CFC-11, CFC-12, CH4, and N20.  Other  minor trace  gases  were
 incorporated based on the  work of Ramanthan  et al.  (1985).

 Radiative Forcing Equations

     The equations in the model are:

 C02:  FC02 - In [1 + .942*C02/(1+.00062*002) + .0088*C022 + 3.26*10"6*C023  +
      .156AC021-3 exp (-C02/760)]

 CH4:  FCH4 = [.394*CH4-66  + .16*CH4 exp (-1.6*CH4]  /  (1  + .169*CH4'62)

 N20:  FN20 = 1.556 In [1 + 1.098 (1+.032*N20)*N20'77 / (1+.0014*N202)]

 Overlap:  FOVL = .14 In [1 + .636 (CH4*N20)-75 +  (.007*CH4  (CH4*N20)1-52]

 Fll:  FF11 = .066*F11

 F12:  FF12 = .084*F12

 ATrad = FC02(now)-FC02(ref) = FCH4(now)-FCH4(ref) + FN20(now)-FN20(ref)  -
 FOVL(now)+FOVL(ref)  = FFll(now)-FFll(ref) + FF12(now)-FF12(ref)

 For  002 from 300 to 600:   ATrad = FC02(600)   - FC02(300)  = 1.26°

Where:   ATrad is the radiative forcing in year t, expressed in  degrees
 centigrade (C).   The model is initialized to the  year 1880  ("ref" =  1880).

 002  is the 002 concentration in year t, in parts  per million  (ppm)
C02(ref) - 270.

CH4  is the CH4 concentration in year t, in ppm.    CH4(ref) = 1.02.

N20  is the N20 concentration in year t, in ppm.    N20(ref) = 0.2853.

Fll  is the CFC-11  concentration in year t, in parts per billion (ppb)
Fll(ref) = 0.0

-------
                                      6-29
F12(t) is the CFG-12 concentration in year t, in ppb. F12(ref) =0.0

  For the other trace gases incorporated in the model, the radiative forcing in
any year is computed in the same fashion as CFG-11 and CFG-12:
                             A Trad...  = a.  * X.
                                   (J)     J     Jt
where: ATrad.   = radiative forcing in degrees C in year t for compound "j".

        a.      = sensitivity to a uniform 1 ppb increase in the trace gas
                  concentration (0 to 1 ppb) for compound "j".

        x.t     = the concentration in ppb in year t for compound "j".


The sensitivities were extracted from the work of Ramanathan et al. (1985)  and
are:

                                               "a"
                                     Direct radiative forcing
                                     for a 0 to 1 ppb increase
                 Trace Gas                 (degrees C)

                 S02                           0.01
                 F14                           0.06
                 F116                          0.13
                 HCFC-22                       0.05
                 CFC-13                        0.22
                 CH2C12                        0.03
                 CHC13                         0.06
                 CC14                          0.08
                 CH3CC13                       0.02
                 Halon 1301                    0.17
                 C2H2                          0.02


    The total radiative forcing (the sum of the radiative forcing from each
trace gas) is the basis for estimating heat flux into the earth's surface.  The
following equations, from Hansen et al. (1984), are used:

ATeq = f.ATrad

F (W/m2) -  Fo(2*C02)  (ATeq - AT) = global flux into the ocean
           ATeq(2*C02)

AT = is the current global temperature change from initial reference  state

We use the flux conversion efficiency of Model II for global flux into the  ocean
as determined by the 2*C02 experiment.

-------
                                      6-30


F (W/m2) = 3.58.(ATrad - AT/f)*(0.7)

F (W/m2)/697.4 = F (calkin'1*™"2)

    The heat flux is estimated for time periods ranging from each month to each
year (a semimonthly time step was used in this study).  The appropriate AT value
for calculating F(t) in each time period (t=n) is the value estimated for the
previous period (t=n-l).  For a simple one-layer ocean model, AT is obtained by
solving the following differential equation:

                                d A T - Fft)
                                 dt      C
                                          o

Where:   C  is the heat capacity of the mixed layer of the ocean per unit area
(cal cm  ;.

Diffusion of Heat in the Ocean

    The ocean model consists of a mixed layer of depth Hm = 100m and a
thermocline with 63 layers and depth H = 900m.  The mixed layer temperature is
assumed to be independent of depth, while the thermocline temperature is defined
by a diffusion equation with constant thermal diffusivity.

    The temperature change in the mixed layer (ATm) is a solution of the
equation:

                 CHm  d A Tm  = F(t) + FD(t)
                        dt
where C is the heat capacity of water, Hm is the depth of the mixed layer, F(t)
is the heat flux from the atmosphere into the ocean, and

                 FD(t) - - A3A   T |
                           	 I
is the heat flux from the thermocline into the mixed layer.  Note that our
z-axis is directed toward the bottom of the ocean.  Also, we use g, cm, sec and
cal,  so the heat conductivity lamda is numerically equal to the heat diffusivity
K.  Different values for diffusivity may be chosen.

    The temperature change in the thermocline (A T) is determined by the
diffusion equation:


             c   dA Tfz.t)  =   A d2AT(z.t)

                                    fiz2

-------
                                     6-31
    The boundary conditions for   AT are:
                              AT =  AT   at z = H
                                      m          m
and zero heat flux at the bottom of the thermocline;
            A3A  T

               az
0 at z = H + Hm.
Thus it is assumed that no energy escapes through the lower boundary of the
thermocline.   Note that  AT  and  AT are temperature changes of the mixed layer
and the thermocline between the initial time (1880) and time t.  It is assumed
that in the year 1880  AT  =  AT = 0 and thus that the ocean temperature was in
a state of equilibrium with the atmosphere at that time.

    The use of diffusivity coefficients as a surrogate for all circulation
processes that transport heat may fail to describe the time paths well, with the
downward heat transport in the latter period probably being overestimated
because of increasing oceanic stability.  Nevertheless this method simplified
the problem substantially, and a range of different possible coefficients can be
used to investigate the sensitivity of the overall estimate to the rate of
downward heat transport.

    Using transient tracers, Broecker and Peng (1982) developed an average eddy
diffusion coefficient for the ocean of 1.7 cm^/sec.  This value, accepted by
many as a reasonable surrogate measure for estimating heat absorption, is the
reference value in our model.  To test the sensitivity of the results to this
choice, other values may be chosen.
                                                         f\
    The range of coefficients tested, from 0.85 to 3.4 cm^/sec, covers the range
of mean ocean-wide mixing rates as determined by the National Science
Foundation-sponsored transient tracer experiment and others.  Variation in
estimates of the exact value of the data depend on the tracer used and the
statistical method of computing the global mean.  The values of 0.2 and 4.0
cm^/sec were tested in earlier work to account for the possibility of dramatic
changes in ocean transports due to deglaciation and climate change (Hoffman et
al. 1983).

-------
                                  6-32







                              ATTACHMENT B




      Trace Gas Scenarios Used in Hansen et al. 1984; Exhibit 6-9
YEAR
1850
1900
1950
1960
1970
1980
1990
2000
2010
C02
(ppm)
270
291
312
317
326
338
353
372
396
CFC-11
(ppt)
0
0
7
33
126
308
479
638
787
CFC-12
(ppt)
0
0
1
11
62
178
280
369
447
CH4
(ppb)
1400
1400
1400
1416
1500
1650
1815
1996
2196
N20
(ppb)
295
295
295
295
295
301
307
313
320
Trace Gas Scenarios Used in Hoffman, Wells,  and Titus 1985;  Exhibit 6-10
YEAR
1880
1900
1950
1960
1970
1980
1990
2000
2010
2020
2030
C02
(ppm)
280
288
311
316
326
339
355
373
398
431
470
CFC-11
(ppt)
0
0
1
11
64
185
341
599
992
1500
2116
CFC-12
(ppt)
0
0
8
36
137
337
577
903
1371
2023
2863
CH4
(ppb)
1402
1448
1568
1593
1619
1645
1822
2093
2406
2764
3177
N20
(ppb)
295
296
298
299
299
300
307
313
320
328
338

-------
                                     6-33
REFERENCES
Bradley, R.S., and P.O.  Jones (1985).   "Data Bases for Isolating the Effects of
the Increasing Carbon Dioxide Concentration."  In MacCracken, M.C., and P.M.
Luther (eds.), Protecting the Climatic Effects of Increasing Carbon Dioxide.
DOE/ER-0235, U.S. Department of Energy, Washington, D.C.

Broecker, W.S., (1986).  "Statement for Ozone Depletion, the Greenhouse Effect,
and Climate Change," Hearings before the Subcommittee on Environmental Pollution
of the Committee on Environment and Public Works, United States Senate,
Ninety-Ninth Congress, Second Session, June 10 and 11, 1986.

Broecker, W.S., and T. Peng (1982).  Tracers in the Sea.  Palisades, NY:
Lament-Doherty Geophysical Observatory.  Columbia University.

Connell, P.S., and O.J.  Wuebbles (1986).  Ozone Perturbations in the LLNL
One-Dimensional Model -- Calculated Effects of Projected Trends in CFC's. CH4.
C02.  N20. and Halons Over 90 Years. Lawrence Livermore National Laboratory,
Livermore, CA.

Dickinson, R.E., and R.J. Cicerone (1986).  "Future Global Warming from
Atmospheric Trace Gases," Nature 3.19:109-115.

Gilliland, R.L., and S.H. Schneider (1984).  "Volcanic, C02 and Solar Forcing of
Recent Climatic Changes," Nature 310:38-41.

Hansen, J.E.  (1986). "Statement for Ozone Depletion, the Greenhouse Effect, and
Climate Change," Hearings before the Subcommittee on Environmental Pollution of
the Committee on Environment and Public Works, United States Senate,
Ninety-Ninth Congress, Second Session, June 10 and 11, 1986.

Hansen, J., D. Johnson,  A. Lacis, S. Lebedeff, P. Lee, D. Rind, and G. Russell
(1981).  "Climate Impacts of Increasing Atmospheric C02, Science 213:957-966.

Hansen, J., A. Lacis, D. Rind, G. Russell, I. Fung, and S. Lebedeff (1986).
"Evidence for Future Warming:  How Large and When," presented at the C02
Conference of the National Forest Products Association, June 25-27, 1984,
Boulder, Colorado, in press.

Hansen, J., A. Lacis, D. Rind, G. Russell, P. Stone, I. Fung, R. Ruedy, and J.
Lerner  (1984).  "Climate S'ensitivity:   Analysis of Feedback Mechanisms."  In
Hansen, J.E., and T. Takahashi (eds.), Climate Processes and Climate
Sensitivity. Geophysical Monograph 29, Maurice Ewing Volume 5, American
Geophysical Union, Washington, D.C.

Hoffert, M.I., and B.P.  Flannery (1985).  "Model Projections of the Equilibrium
Climatic Response to Increased Carbon Dioxide."  In MacCracken, M.C., and F.M.
Luther  (eds.), Projecting the Climatic Effects of Increasing Carbon Dioxide.
DOE/ER-0237, U.S. Department of Energy, Washington, D.C.

-------
                                      6-34
Hoffman, J.S., and J.B. Wells (1986).  "Trace Gas Scenarios," presented at the
UNEP Economic Workshop on Protection of the Ozone Layer, May 26-30,  1986, Rome,
Italy.

Hoffman, J.S., J.B. Wells, and J.G.  Titus (1985).  "Future Global Warming and
Sea Level Rise."  In G. Sigbjarnarson (ed.), Iceland Coastal and River
Symposium. Proceedings. Library of the National Energy Authority, Reykjavik,
Iceland.

Hoyt, D.V. (1979).  "Variations in Sunspot Structure and Climate," Climatic
Change 21:79-92.

Isaksen, I.S.A., and F. Stordal (1986).  "Ozone Perturbations by Enhanced Levels
of CFCs, N20, and CH4:  A Two-Dimensional Diabatic Circulation Study Including
Uncertainty Estimates," Journal of Geophysical Research. 91(D4): 5249-5263.

Kavanaugh, M. (1986).  "Estimates of Future CO, N20, and NOx Emissions from
Energy Combustion," to appear in Atmospheric Environment (in press).

Khalil, M.A.K.,  and R.A. Rasmussen (1985).  "Causes of Increasing Atmospheric
Methane:  Depletion of Hydroxyl Radicals and the Rise of Emissions," Atmospheric
Environment 19(3): 397-407.

Lacis, A.A. and D.J. Wuebbles (1986).   "Radiative Forcing of Global Climate
Changes in Stratospheric Water Vapor," to be submitted to Geophysical Research
Letters.

Lacis, A.A.,  D.J. Wuebbles, and J.A. Logan  (1986).  "Radiative Forcing of Global
Climate by Changes in the Vertical Distribution of Atmospheric Ozone," to be
submitted to  Geophysical Research Letters.

Luther, P.M., and R.G. Ellingson (1985).  "Carbon Dioxide and the Radiation
Budget."  In  MacCracken, M.C., and P.M. Luther  (eds.), Projecting the Climatic
Effects of Increasing Carbon Dioxide.  DOE/ER-0237, U.S. Department of Energy,
Washington, D.C.

Luther, F.M., and M.C. MacCracken (1985).   "Recommendations for Research and
Modeling Activities for Projecting the Climatic Effects of Increasing Carbon
Dioxide."  In MacCracken, M.C., and F.M. Luther  (eds.),  Projecting the Climatic
Effects of Increasing Carbon Dioxide.  DOE/ER-0237, U.S. Department of Energy,
Washington, D.C.

MacCracken, M.C. and F.M. Luther (1984).  "Radiative and Climatic Effects of the
El Chichon Eruption," Geofisica International  23{3): 385-401.

Manabe, S., and R.T. Wetherald (1975).  "The Effect of Doubling  the C02
Concentration in  the Climate of a General Circulation Model," Journal of
Atmospheric Sciences 32:3-15.

Manabe, S., and R.T. Wetherald (1986).  "Reduction  in Summer Soil Wetness
Induced by an Increase  in Atmospheric Carbon Dioxide,"  Science  232:626-628.

-------
                                      6-35
Miller, A.S.,  and I.M. Mintzer (1986).  "'Draining the Sink' Policy Implications
of CFC Growth Trajectories," discussion draft prepared for the U.S.
Environmental Protection Agency Workshop,  "'Protecting the Ozone Layer' Workshop
of Demand and Control Technologies", March 6-7, 1986, Washington, D.C.

National Academy of Sciences (NAS)  (1979).  Carbon Dioxide and Climate: A
Scientific Assessment. National Academy Press, Washington, D.C.

National Academy of Sciences (NAS)  (1982).  Carbon Dioxide and Climate: A Second
Assessment. National Academy Press, Washington, D.C.

National Academy of Sciences (NAS)  (1984).  Causes and Effects of Changes in
Stratospheric Ozone and Update 1983. National Academy Press, Washington, D.C.

National Aeronautics and Space Administration  (NASA) (1982).  "Radiative Effect
of the El Chichon Volcanic Eruption:  Preliminary Results Concerning Remote
Sensing" (NASA TM 84959), NASA, Washington, D.C.

Neftel, A., H. Oeschger, J. Schwander, B.  Stauffer, and R. Zumbrunn (1982).
"Ice Core Sample Measurements Give Atmospheric C02 Content During the Past
40,000 Yr.,11 Nature 295:220-223.

Newkirk, G., Jr. (1983).  "Variations in  Solar Luminosity," Annual Reviews of
Astronomy and Astrophysics  21:429-467.

Oeschger, H., B. Stauffer, A. Neftel, J.  Schwander, and R. Zumbrunn (1982).
"Atmospheric C02 Content in the Past Deduced from Ice-Core Analyses,"  Annals of
Glaciologv 3:227-232.

Pearman, G.I., D. Etheridge, F. deSilva,  and P.J. Fraser  (1986).  "Evidence of
Changing Concentrations of Atmospheric C02, N20, and CH4  from Air Bubbles in
Antarctic Ice," Nature. 320:248-250.

Peng, T.H., W.S. Broecker, H.D. Freyer, and S. Trumbore (1983).  "A
Deconvolution of the Tree Ring Based d!3C Record," Journal of Geophysical
Research 88:3609-3620.

Pollack, J.B., O.B. Toon, E.F. Hoffman, O.J. Rosen, andJ.M. Rosen (1983).  "The
El Chichon Volcanic Cloud:  An Introduction,"  Geophysical Research Letters
10:989-992.

£amanathan, V., R.J. Cicerone, H.B. Singh, and J.T. Kiehl (1985).  "Trace Gas
Trends and their Potential Role in  Climate Change," Journal of Geophysical
Research. 90(D3): 5557-5566.

Rind, D., and S. Lebedeff (1984).   Potential Climatic Impacts of Increasing
Atmospheric C02 with Emphasis on Water Availability and Hydrology in the United
States. U.S. Environmental Protection Agency,  Washington, D.C.

Sarmiento, J., and K. Bryan (1982).  "An  Ocean Transport Model for the North
Atlantic," Journal of Geophysical Research 87:9694-9698.,

-------
                                     6-36
Schlesinger,  M.E.,  and J.F.B.  Mitchell (1985).   "Model Projections of the
Equilibrium Climatic Response to Increased Carbon Dioxide."  In MacCracken,
M.C., and F.M.  Luther (eds.),  Projecting the Climatic Effects of Increasing
Carbon Dioxide. DOE/ER-0237,  U.S.  Department of Energy,  Washington, D.C.

Seidel, S. and D.  Keyes (1983).   Can We Delay a Greenhouse Warming?.  U.S.
Environmental Protection Agency, Washington, D.C.

Somerville, R.C.J.,  and L.A.  Remer (1984).   "Cloud Optical Thickness Feedbacks
in the C02 Climate Problem,"  Journal of Geophysical Research 89(d6):  9668-9672.

Stordal,  F.,  and I.S.A.  Isaksen (1986). "Ozone Perturbations Due to Increases in
N20, CH4, and Chlorocarbons:   Two-Dimensional Time Dependent Calculations,"
prepared for UNEP Economic Workshop on the Ozone Layer,  Rome, Italy,  May 26-30,
1986.

Strain, B.R.  (1985).  "Background on the Response of Vegetation to Atmospheric
Carbon Dioxide Enrichment."  In Strain, B.R.,  and J.D. Cure (eds.), Direct
Effects of Increasing Carbon Dioxide on Vegetation. DOE/ER-0238, U.S. Department
of Energy, Washington, D.C.

Thompson, A.M., and R.J. Cicerone (1986).  "Atmospheric CH4, CO and OH from 1860
to 1985," Nature.  in press.

Tyndall,  J. (1861).   "On the Absorption and Radiation of Heat by Gases and
Vapours,  and on the Physical Coxection of Radiation, Absorption, and
Conduction," Philosophical Magazine and Journal of Science 22(146): 169-194.

Wang, W.C., D.J. Wuebbles, and W.M. Washington (1985).  "Potential Effects of
Perturbations Other than Carbon Dioxide."  In MacCracken,  M.C., and F.M.  Luther
(eds.), Projecting the Climatic Effects of Increasing Carbon Dioxide.
DOE/ER-0237,  U.S.  Department of Energy, Washington, D.C.

Washington, W.M.,  and G.A. Meehl (1984).  "Seasonal Cycle Experiment on the
Climate Sensitivity Due to a Doubling of C02 With an Atmospheric General
Circulation Model Coupled to a Simple Mixed-Layer Ocean Model," Journal of
Geophysical Research 89(D6):  9475-9503.

Webb, T., Ill and T.M.L. Wigley (1985).  "What Past Climates Can Indicate about
a Warmer World."  In MacCracken, M.C., and F.M. Luther (eds.), Projecting
the Climatic Effects of Increasing Carbon Dioxide. DOE/ER-0237, U.S. Department
of Energy, Washington, D.C.

Wigley, T.M.L., J. Angell, and P.O. Jones (1985)'.  "Analysis of the Temperature
Record."  In MacCracken, M.C., and F.M. Luther (eds.), Projecting the Climatic
Effects of Increasing Carbon Dioxide. DOE/ER-0237, U.S.  Department of Energy,
Washington, D.C.

Woods, J.D. (1985).   "The World Ocean Circulation Experiment," Nature
314:501-511.

-------
                                      6-37
World Meteorological Organization (WHO) (1984).   The International Satellite
Cloud Climatology Project (ISCCP)." report of the Third Session of the
International Working Group on Data Management,  March 6-8,  1984,  Tokyo.   World
Climate Programme Report WCP-82, WMO, Geneva, Switzerland.

World Meteorological Organization (WMO) (1986a).  Report of the International
Conference on the Assessment of the Role of Carbon Dioxide and of other
Greenhouse Gases in Climate Variations and Associated Impacts. Villach.  Austria.
9-15 October 1984. WMO-No. 661, International Council of Scientific Unions,
United Nations Environment Programme, and World Meteorological Organization.

World Meteorological Organization (WMO) (1986b).  Atmospheric Ozone 1985.
Assessment of Our Understanding of the Processes Controlling its Present
Distribution and Change. Global Ozone Research and Monitoring Project -- Report
No. 16, WMO, Geneva, Switzerland.

-------
                                   CHAPTER 7

                            NONMELANOMA SKIN TUMORS
SUMMARY

    Nonmelanoma skin tumors are the most common cancers occurring in white
populations.  The two major forms of nonmelanoma skin tumors are basal cell
carcinoma (BCC) and squamous cell carcinoma (SCC).   Although the incidence of
BCC is generally several times greater than the incidence of SCC, SCCs account
for as much as four-fifths of all nonmelanoma skin cancer deaths.  Prolonged
sunlight exposure is considered to be the dominant (but not only) risk factor
for nonmelanoma skin tumors.  As a result, it is believed that increased
exposure to solar ultra-violet (UV) radiation (due to the depletion of the ozone
layer) will increase the incidence of nonmelanoma skin tumors among susceptible
populations.  SCC would have a much larger percentage increase than BCC.

    Individuals with predisposing genetic traits such as albinism and xeroderma
pigmentosum (a genetically inherited inability to repair UV-induced DNA damage)
are at the highest risk of developing nonmelanoma skin tumors.   The much larger
population consisting of light-skinned individuals (skin phenotypes I and II) is
considered to be at high risk, especially those with a susceptibility to
sunburn, poor tanning ability, red or blond hair, blue or green eyes, and a
Celtic heritage.  Populations with pigmented skin,  such as blacks,  are at
significantly lower (but not zero) risk.

-------
                                      7-2
FINDINGS
    1.   BASED ON SURVEYS (PARTICULARLY IN THE UNITED STATES AND IN AUSTRALIA).
        PROLONGED SUN EXPOSURE IS CONSIDERED TO BE THE DOMINANT RISK FACTOR FOR
        NONMELANOMA SKIN TUMORS.

        la.   Nonmelanoma skin tumors tend to develop in sun-exposed sites (e.g.,
             the head,  face,  and neck).

        Ib.   Higher incidence rates occur among groups subject to greater
             exposure to the  sun's rays because of occupations that necessitate
             their working outdoors.

        Ic.   A latitudinal gradient exists for UV-B radiation, and higher
             incidence rates  of nonmelanoma skin tumors generally occur in
             geographic areas of relatively high UV radiation exposure.

        Id.   Skin pigmentation provides a protective barrier that reduces the
             risk of developing nonmelanoma skin tumors.

        le.   The risk of nonmelanoma skin tumors is highest among genetically
             predisposed individuals (e.g., those with xeroderma pigmentosum).

        If.   A predisposition to develop nonmelanoma skin tumors exists among
             light-skinned individuals (skin phenotypes I and II) who are
             susceptible to sunburn and who have red/blond hair, blue/green
             eyes, and a Celtic heritage.

    2.   AVAILABLE EPIDEMIOLOGICAL EVIDENCE INDICATES THAT THE TWO MAJOR TYPES  OF
        NONMELANOMA SKIN TUMORS.  SQUAMOUS CELL CARCINOMA (SCO AND BASAL CELL
        CARCINOMA (BCC). RESPOND DIFFERENTLY TO SOLAR EXPOSURE.  IT HAS BEEN
        SUGGESTED THAT CUMULATIVE UV RADIATION HAS A GREATER EFFECT ON THE
        DEVELOPMENT OF SCC THAN ON BCC.

        2a.   The BCC/SCC incidence ratio decreases with decreasing latitude and
             therefore, increasing UV levels.

        2b.   BCC is more likely to develop on normally unexposed sites (e.g.,
             the trunk) compared to SCC.

        2c.   SCC is more likely than BCC to develop on sites receiving the
             highest cumulative UV radiation doses (e.g., the nose).

        2d.   For a given cumulative level of sunlight exposure, the risk of
             developing SCC may be greater than the risk of developing BCC.

-------
                                  7-3
3.  THE RESULTS FROM SEVERAL EXPERIMENTAL STUDIES SUGGEST THAT UV-B MAY BE
    THE MOST IMPORTANT COMPONENT OF SOLAR RADIATION THAT CAUSES VARIATIONS
    IN THE INCIDENCE OF NONMELANOMA SKIN TUMORS.

    3a.  UV radiation produces nonmelanoma skin tumors in animals.   UV-B
         wavelengths have been shown to be most effective in producing these
         tumors.

    3b.  UV-B has been shown to cause a variety of DNA lesions, to induce
         neoplastic transformation in cells, and to be a mutagen in both
         animal and bacterial cells.

4.  SEVERAL RESEARCHERS HAVE INVESTIGATED THE CHANGES IN THE INCIDENCE OF
    NONMELANOMA SKIN TUMORS THAT MAY RESULT FROM INCREASES IN EXPOSURE TO
    SOLAR UV RADIATION.  GIVEN UNCERTAINTIES. RANGES OF ESTIMATES OF
    INCREASED INCIDENCE THAT COULD OCCUR WITH DEPLETION ARE ESTIMATED.

    4a.  The action spectra for initiation and promotion of basal cell and
         squamous cell skin cancer have not been precisely determined.
         Photocarcinogenic studies indicate that the erythema and DNA action
         spectra span a range likely to encompass that of squamous cell and
         basal cell skin cancer.  The Robertson-Berger (R-B) meter, while
         providing useful data for describing ambient UV radiation, does not
         relate as closely to those wavelengths thought to promote sunburn
         and skin cancer..

    4b.  Several studies have provided estimates of a biological
         amplification factor (BAF), which is defined as the percent change
         in tumor incidence that results from a 1 percent change in UV-B
         radiation.  The results from six studies produced an overall BAF
         range that is 1.8-2.85 for all nonmelanoma skin tumors.

    4c.  BAF estimates are generally higher for males than for females and
         generally increase with decreasing latitude.  In addition, the BAF
         estimates for SCC are higher than the BAF estimates for BCC.  This
         finding is consistent with observations that the BCC/SCC ratio
         decreases with decreasing latitude and that BCC is more likely to
         develop on unexposed sites.

    4d.  Optical amplification (the change in UV-B radiation related to
         ozone depletion) increases the response of these cancers to ozone
         depletion, because the relevant action spectra increase more than 1
         percent for a 1 percent depletion.  For example, a 1 percent
         depletion has an optical amplification of over 2 for the DNA action
         spectrum.

    4e.  Uncertainty exists in the actual doses of solar UV radiation
         received by populations and in the statistical estimates of the
         dose-response coefficients.  Therefore,  a range of estimates must
         be developed for changes in incidence associated with changes in
         dose .

-------
                              7-4
4f.   Currently available nonmelanoma mortality data are of uncertain
     accuracy because of the discrepancy of reporting between death
     certificates and hospital diagnoses and the low proportion of
     deaths reported on both hospital diagnoses and death certificates.
     Based on published studies,  the rates of metastasis among SCCs and
     BCCs have been estimated to  be 2-20% and 0.0028-0.55%,
     respectively.  The overall case fatality rate for nonmelanoma skin
     tumors is approximately 1-2% with three-fourths to four-fifths of
     the deaths attributable to SCC.

4g.   Changes in behavior have tended to increase skin cancer incidence
     and mortality.  While some evidence exists that this is reaching a
     limit, skin cancer rates, even in the absence of ozone depletion,
     would be likely to rise.  Future rates of skin cancer could be
     reduced if people changed their behavior.  Care should be taken,
     however, in interpreting such a change as a cost-free response.

-------
                                      7-5
INTRODUCTION

    There are two major types of nonmelanoma skin tumors:  basal cell carcinoma
(BCC) and squamous cell carcinoma (SCC).  These tumors are among the most common
cancers occurring in white populations.  In the United States alone, nonmelanoma
skin cancers develop in more than 400,000 white Americans each year and
represent over half of all cancers occurring in U.S. whites.

    Prolonged sunlight exposure is considered to be the dominant, but not the
only, risk factor for these tumors based on several observations.  Nonmelanoma
skin tumors tend to develop in sun-exposed sites (e.g., head and face) and occur
at a higher rate among outdoor workers compared to indoor workers.   Nonmelanoma
incidence rates are higher in areas with high ultraviolet radiation
(UVR) exposure (e.g., closer to the equator) than in areas with low UVR
exposure.  Light-skinned individuals who are susceptible to sunburn and who have
red or blond hair, blue or green eyes,  and a Celtic heritage are at high risk of
developing nonmelanoma skin tumors.

    To evaluate the role of solar radiation, and UV-B in particular, in the
development of nonmelanoma skin tumors, this chapter presents background
information on solar radiation as it relates to the concept of dose.  Then
reviews the biology of the skin and nonmelanoma skin cancers (NMSC), focusing on
how skin responses to solar radiation can result in NMSC.  Evidence from
molecular and animal studies relating nonmelanoma skin tumors to UVR is
discussed.  Epidemiologic evidence linking exposure to solar radiation and
development of nonmelanoma skin tumors among human populations is also reviewed.
Finally, the available dose-response relationships that estimate potential
changes in the incidence of nonmelanoma skin tumors due to changes in exposures
to UVR are presented.

BACKGROUND ON SOLAR RADIATION AND THE CONCEPT OF DOSE

    The total amount of energy from UVR that any target receives in a given
amount of time will depend, in part, on variations in ambient radiation that
occur in the natural environment.  One effect of the ozone layer and the earth's
rotation and revolution is the modification of the quality and quantity of solar
energy delivered from place to place over time.  Ambient solar radiation
incident on various sites on the earth's surface varies significantly with
latitude, altitude, season (day of the year), time of day, cloudiness,
reflectiveness of surfaces (albedo), and atmospheric aerosols.   More
importantly, these variations differ quantitatively for different wavelengths of
UVR.

    The latitude of a particular location on the Earth's surface determines the
average angle between the location and the sun.  This angle, in turn, determines
the thickness of the atmosphere that photons must pass through before reaching
the Earth's surface.  The lower the angle, the longer the path through the
ozone-containing stratosphere that photons must travel and the greater the
amount of UVR absorption that can occur.  A longer path also allows longer
exposure to aerosols and, thus, more scatter -- thereby additionally reducing
the amount of energy reaching a location.   Because ozone absorbs different
wavebands differently, however, all photons are not retarded equally; instead
differential gradients for various wavelengths occur.  Exhibit 7-1 shows

-------
                                             7-6
                                         EXHIBIT 7-1

                       Variation in UV Radiation by Latitude as Percent
                         of Levels  at the Equator on March 21 at Noon
  100.0
                                                   r   r   i  ~ r
Cfl
hH
H
>
W
CM
O
EH
3
O
«
w
                                        r
FT   T
                                                                            375-379nm

                                                                            335-339nm
                                                                            K
                                                                            315-319n*i
                                                                            305-309nll
     1.0 i-
         0
                                                 295-299n>
                                                           i
-I	I ._J ._!.._ J..__l	I _. J	L....J_  J_  .!._ _!.._. J	I	I
20                 40                  60                 80
            LATITUDE (°N)
        This  graph  shows the variation estimated by  the NASA model for five wavebands  of
        ultraviolet radiation for a single day,  high noon, in March.  The variation
        would be  different for other days.

-------
                                      7-7
estimates from the NASA UV Model, developed by Serafino and Frederick  (1986),
which predict that at 12 noon on March 21, the flux of UV-B at 295-299 nm can be
expected to vary from the equator to 70°N by a factor of over 100, while UV-A at
335-339 nm can be expected to vary by a factor of 5.

    Season also has a tremendous impact on the flux of solar radiation at a
given location.  Again, this occurs because the relationship of a location to
the sun changes with season.  Exhibit 7-2 shows model estimates of how the
various wavelengths vary by month for Washington, D.C.  Energy at 295 nm
increases by about a factor of 10 from winter (December) to Spring (March) and
by about another factor of 10 by midsummer (July).  UV-B at 305 nm shows much
less variation and UV-A very little, if any, variation.

    Exhibit 7-3 shows the impact that varying time of day has on estimates of
UV-B.  Clearly, the time of day of exposure is a very important factor in
determining the amount of UV energy received at the various wavelengths.
Variations in cloud cover and surface albedo also cause variations in ambient
solar energy levels, however, there is not much differential variation by
wavelength (for additional details, see Chapter 2 of Appendix A of the Risk
Assessment).

    In addition to factors that modify the ambient levels of UVR in the
environment, there are a number of factors that can influence the biologically
effective dose (i.e., the amount of energy received by a target in the skin that
results in the development of skin cancer).  These include an individual's
clothing and sun-exposure behavioral habits, the degree of skin pigmentation and
thickening, and the action spectrum for the biologically important effect.

    The amount of energy that is actually effective is determined by the action
spectrum of the effect of concern.  If, for instance, damage to DNA is the
underlying effect linking solar radiation to skin cancer, then not all of the
photons delivered to the target molecules will be effective.  Instead, the
absolute amounts of energy delivered in the various wavelengths need to be
weighted according to the action spectrum presented in Exhibit 7-4.  There are a
number of different effects that might underlie the relationship between solar
radiation and skin cancer (details of which are presented in Chapters 2 and 3 of
Appendix A); a key characteristic of the effects is that energy in the UV-B
wavelengths (295-320 nm) tends to be much more biologically effective.

    How much energy target cells can potentially receive depends on the time a
person spends outside during certain periods of the day and certain days of the
year.  The total amount of energy delivered at a given location provides an
upper bound of exposure, not the actual exposure.  Few individuals are out in
the sun during all daylight hours, therefore actual exposure is correspondingly
reduced.   As indicated above,  the seasonal and hourly variations in incident
energy from solar radiation can be considerable,  particularly in those
wavelengths that are the most biologically effective (295-299 nm),  so all hours
of the day cannot be considered equally.   In addition, people living in areas
having the same number of daylight hours may have additional behavioral
differences that modify the amount of radiation reaching the skin (the actual
exposure).   For example, some people wear lots of clothing,  others do not.  Some
people wear sunscreens,  while others use sun reflectors to gather more solar
radiation.

-------
                                      7-8


                                 EXHIBIT  7-2

                  UV Radiation by Month in Washington, D.C.
                                                             r T~T  r  r  i
                          MONTH (1 = JANUARY)
This graph shows that the NASA model estimates  that radiation at 295 nm has  a
much larger proportional gain than at 305 nm or higher wavelengths.   Note that
the 335 nm line is almost coincident with the x axis,  indicating low monthly
variation in Washington, D.C.

-------
                                      7-9
                                  EXHIBIT 7-3

                           Ratio of Instantaneous Flux
                          Throughout the Day to Flux at
                        5:15 am in Washington on June 21
                              (Assumes a Clear Day)
25^-r
20-
15
10-
 J !
i—i—-r-r
                  8
            11
1-4
17
                                                                             1
                                                                             H
                                                                             J
20
                               TIME  OF DAY (HOURS')

-------
                                    7-10





                                EXHIBIT 7-4



                     Average DNA-Damage Action Spectrum
       SH
       EH
          to
S  10'
H
W
z

w
       u
       I—I

       I   J

       o
       H
       ffl

       M
       > 10"
                    \
                     \
  /


/ — SUN LIGHT
                                  • — ERYTHEMA

                                  \t'.
                       ONA —
               260    28O   300    320   34O

                         WAVELENGTH (nm)
                                         360
                                                       10
                                                         )2
                                                      -10
                                               -10'
                                                        10
«
H
0.
                                                           0)
                                                         fN
                                                   EH
                                                   Z

                                                   g
                                                      !0
                                                      10'
Source: Adapted from Setlow  (1974).

-------
                                     7-11
    Finally, the transition from the amount of energy to which an individual is
exposed to the amount of energy that is delivered to a target molecule (the
biologically effective dose),  involves the interplay of factors such as skin
thickness and pigmentation.  The greater the skin thickness and degree of
pigmentation, the smaller the amount of energy penetrating to the potential
targets.

BIOLOGY OF NONMELANOMA SKIN TUMORS: LINKS TO UV-B

    A general understanding of the biology and photobiology of the skin provides
a useful background against which to examine the association between solar
radiation and nonmelanoma skin tumors.

Biology of the Skin

    There are three principal layers of skin: the epidermis, the dermis,  and the
panniculus adiposus (Exhibit 7-5).  These layers vary in thickness depending on
their location.  The epidermis normally consists of about 6-10 cell layers and
is typically 60 -100 mm thick.  The thinner epidermis is found on the head,
trunk, and upper limbs and the thicker epidermis is found on the lower limbs.
The epidermis on the palms and soles is as much as ten times thicker than that
on the head and trunk.  The stratum corneum is the outermost, more dense layer
of the epidermis and generally comprises between 8 :m and 15 :m of the
epidermis' thickness (Pearl 1984; Fitzpatrick and Soter 1985).   The dermis is
typically between 1700 :m and 2000 :m thick and the subcutaneous layer, the
panniculus adiposus, is typically 4000-9000 :m thick (Fitzpatrick and Soter
1985).

    The epidermis is principally a mixture of three cell types of different
embryonic origin and function; these are (in order of percent composition) the
keratinocyte (80 percent), the melanocyte (5-10 percent), and the Langerhans
cell  (5-10 percent).  The dermis is composed mainly of connective tissue fibers.
These fibers, which are secreted by cells called fibroblasts, are responsible
for the skin's resilience and elasticity.  Many skin changes associated with
aging are caused by the impact of solar radiation on these dermal fibers.   The
subcutaneous layer is a specialized layer of connective tissue that functions as
a cushion between the bone and the epidermis and dermis.  It consists primarily
of fat cells.

    The major cell type of the epidermis is the keratinocyte.  In the epidermis,
keratinocytes are organized vertically by various stages of differentiation.  In
its first differentiation state, the keratinocyte is termed a basal cell.   Basal
cells are organized in layers along a basement membrane that marks the interface
of the dermis and the epidermis.  There are structurally and functionally
distinct populations of basal cells; some may represent the epidermal stem cell
population, whereas others may serve as an anchoring function (Fitzpatrick and
Soter 1985).

    Both basal cell and squamous cell carcinomas are of keratinocyte origin.  It
is currently believed that BCC is derived from undifferentiated pluripotent stem
cells; ultrastructural studies indicate that BCC cells are very similar to
primitive ectodermal cells and can simulate both epidermal and adnexal
development.  In contrast, SCC consists of cells that are differentiated to the
point of producing keratin (Laerum and Iversen 1981) .

-------
                                        7-12

                                    EXHIBIT 7-5

                          Organization of  the Adult Skin
                                                    Slrofum oorrwum
                                                    Oronulor c*ll
                                                            cwl lay*
                                                    6fl»ol
                                                                      ,PAf>ILLA*Y
                                                                      SUBCUTIS
Source:   Fitzpatrick  and Soter  (1985).

-------
                                      7-13
    Keratinocytes produce at least three biologically important proteins:
keratin, histidine-rich proteins, and interleukin 1 (IL-1, also known as
epidermal-derived thymocyte activation factor [ETAF]).  Keratins, the primary
structural proteins of the epidermis, consist of a family of polypeptides that
have strong absorption properties in the UV-C and UV-B range (Harber and Bickers
1981).   Histidine-rich proteins (HRPs) are produced by keratinocytes and
sequestered with keratin into keratohyalin granules.  As keratinocytes
terminally differentiate into corneocytes (enucleated keratinocytes that
comprise the major component of the stratum corneum),  the HRPs break down and
release free histidine, which is converted into urocanic acid (Scott et al.
1982).   IL-1/ETAF functions as a major mediator in both immune and inflammatory
responses and may play a role in the immunosuppression induced by sunlight.
Once thought to be solely a macrophage-produced lymphokine, it is now known to
be produced by other cells (e.g., keratinocytes).  Investigation into its
biologic properties indicates that it is a potent inducer of lymphocyte
activation and chemotaxis,  that it enhances the production of acute phase
proteins by the liver, increases the number of circulating macrophages, and can
result in elevated temperatures (fevers) (Daynes et al. 1986).

    Photobiology of the Skin

    The interaction of sunlight with the skin is a complex process.  It involves
the transfer of energy from sunlight to various molecules in the skin layers and
the subsequent cellular responses to this energy transfer.  About 95 percent of
incident radiation penetrates the skin, while the remaining 5 percent is
reflected by the stratum corneum.  Two processes, scatter and absorption,
determine the penetration of radiation into the skin.   Measurements
of the transmission of UVR through isolated epidermis tissue from medium-
complexion Caucasian skin (Bruls et al. 1984; Kaidbey et al. 1979) indicate that
between 1 percent and 20 percent of 295 run UV-B radiation would reach the basal
layer.   For 340 nm UV-A radiation, as much as 20 to 60 percent may penetrate to
the basal layer (Kubitschek et al. n.d.).  Above 400 nm (in the visible range),
transmission approaches 90 percent (Anderson 1983).   The amount
of radiation reaching the basal layer is a function of the thickness of the
epidermis and stratum corneum as well as their content of radiation-absorbing
molecules called chromophores.   There are a large number of chromophore types in
the skin, one of which is melanin.  Most of the optical absorbance within the
skin, however, is attributable to melanin,  as graphically shown in Exhibit 7-6.

    The skin has three response mechanisms for dealing with exposure to solar
radiation:  (1) dose-reduction through the production of increased melanin
(tanning) or through keratinocyte hyperplasia; (2) damage repair in which the
cell's DNA repair mechanisms remove photoinduced damage; and (3) cell removal.

    Solar Radiation Dose Reduction

    Dose-reduction may occur via one of two mechanisms:  tanning or skin
thickening.   Tanning involves the de novo synthesis of new melanosomes that are
subsequently transported to keratinocytes where they produce a tan within about
10 hours after exposure.  This response may continue for several days, with the
maximal tan being achieved in about one week.  With time the tan wears off as
the keratinocytes containing the extra pigment are sloughed off.  In addition to
the increased production of melanosomes, the number of active melanocytes in the

-------
                                      7-14


                                  EXHIBIT 7-6

        Ultraviolet Absorption Spectra of Major Epidermal Chromophores
         r
         3
              2 00
                     220
                            240
                                  2*0      280

                                 WAVELENGTH, NM
Experimental Conditions:

        DOPA-melanin, 1.5 mg% in HO;

        urocanic acid, 10"  M in HO;

        calf thymus DNA, 10 mg% in HO  (pH  4.5)

        tryptophan (TRP), 2 x 10"  M  (pH  7);

        tyrosine (TYR),  2 x 10    M  (pH 7).
Source:  Anderson (1983).

-------
                                      7-15
skin also increases, probably as a result of increased mitosis and recruitment
of dormant melanocytes.  The delayed tanning response is induced predominantly
by wavelengths in the UV-B range, although UV-C, UV-A, and visible light can
also induce the response.   There is a wide variation in the actual tanning
response of individuals to sunlight exposure.  Skin tanning responses have been
generally classified into six skin type categories as outlined in Exhibit 7-7
(Fitzpatrick 1986).

    The other dose-reduction response involves the increased production of
keratinocytes and a subsequent thickening of the epidermis (keratinocyte
hyperplasia).   Studies in humans have shown that after a single exposure to
UV-B, a sustained increase in epidermal mitoses occurs, which leads to a 1.5- to
3-fold increase in the thickness of the epidermis and stratum corneum over the
course of one to three weeks.  The UV dose first induces a transient depression
in macromolecular synthesis in which DNA, RNA, and protein synthesis are
markedly reduced and then elevated.  The response is maximal within 24 to 48
hours after irradiation but may continue for as long as one week.  UV-B and UV-C
are the most effective wavelengths for inducing this response.  Keratinocyte
hyperplasia protects individuals with a poor tanning ability because the
disulfide-rich keratin synthesized by the keratinocytes absorbs photons in the
UV-C and UV-B ranges.  Among lightly pigmented individuals, this skin thickening
is probably the most important dose-reducing response, whereas among
dark-skinned individuals,  tanning is the more important response (Gange and
Parrish 1983).

    DNA Repair Mechanisms

    A different response mechanism for dealing with exposure to solar radiation
is DNA repair.  UVR can either directly or indirectly damage DNA.  Direct damage
results from the absorption of UVR by DNA and the subsequent formation of DNA
lesions.  The most studied form of direct damage is the development of
cyclobutyl pyrimidine dimers (Spikes 1983) .   These dimers are formed between
adjacent pyrimidines on the same DNA strand, and their presence renders the
phosphodiester bond joining the deoxyribose moieties resistant to nuclease
digestion.  Indirect DNA damage is mediated by a reactive oxygen species, such
as a superoxide radical, which absorbs radiative energy and transfers it to the
DNA.  UV-B principally inflicts direct damage on DNA but can also cause indirect
damage at some wavelengths.  UV-A and UV-C can inflict direct or indirect DNA
damage on cells.  DNA has  an absorption maximum at 265 nm (UV-C range).

    When DNA is damaged, the lesions may cause cell death or may merely disturb
DNA transcription and replication.   Based on the somatic mutation theory of
carcinogenesis,  these lesions may also result in mutations that subsequently
result in neoplastic transformation.  Alternatively, damage to the DNA may
result in oncogene activation.   DNA damage that adversely affects cell function
and survival needs to be repaired in order to assure continuation of a species.

    The cell possesses three DNA repair mechanisms to respond to UV-induced
damage:  photoreactivation, excision repair, and postreplication gap repair.
Repair mechanisms that result in unchanged DNA are called "error-free," whereas
mechanisms that generate altered segments of DNA are generally called
"error-prone."  The inaccuracy of a repair mechanism is likely to result in
secondary structural changes in DNA, some of which may lead to mutations.

-------
                             7-16



                          EXHIBIT 7-7


            Skin Types and Skin Tanning Responses
Skin Color
(unexposed skin)

White


Brown
Black
Skin
Types
I
II
III
IV
V
VI
' Sunburn'
yes
yes
yes
no
no
no
'Tan'
no
minimal
yes
yes
yes
yes
a /
   Sun-reactive skin types:   based on verbal response regarding
first, moderate 45- to 60-min.  (3 MED)1 unprotected sun exposure.
MED = Minimal erythema dose.
Source: Fitzpatrick (1986).

-------
                                      7-17
    Photoreactivation repair of pyrimidine dimers occurs when an enzyme binds  to
the dimer, forming an enzyme-substrate complex.  The enzyme-substrate complex
absorbs photons of UV-A light and the dimer  is then monomerized (Spikes 1983).
Photoreactivation is an error-free, nonmutagenic repair pathway, which offers  a
number of advantages to the cell:  It uses UV-A photons as an energy source,
there is no incision into the DNA phosphodiester backbone, and therefore no risk
of DNA degradation; and there is no polymerization, and thus, no chance for the
introduction of coding errors (Spikes 1983) .

    Another efficient DNA repair mechanism is called excision repair.  Excision
repair works more slowly than photoreactivation but may repair not only dimers
but also other kinds of damage, such as, methylations.  In the case of dimers,
excision repair involves the activity of three enzymes, an endonuclease specific
for dimers, a polymerase, and a ligase.  The endonuclease attacks the DNA at or
adjacent to the dimer and introduces a nick  in the DNA.  The DNA polymerase
removes the damaged DNA segments while utilizing the opposite strand as a
template for new synthesis, and finally, a ligase joins the newly synthesized
DNA to the preexisting strand.  Other kinds  of DNA damage, such as methylations,
require different first enzymes, but the activities of the polymerase and the
ligase remain the same (Spikes 1983).

    The third cell repair mechanism is termed postreplication gap repair.  This
mechanism is invoked when, following DNA replication, there are gaps left in the
DNA opposite the dimer.  The cell invokes a DNA recombinational mechanism, which
provides a good copy of the DNA needed to repair the damage.   Using this
information, the cell can repair the damage with an excision repair mechanism
(Spikes 1983).  This mechanism is, however, fairly error prone.

Biology of Nonmelanoma Skin Tumors

    Although most skin cancer statistics combine both BCC and SCC cell types,
the limited data available from the U.S. and several other countries indicate
differences in a number of characteristics.  Basal cell carcinomas are neoplasms
of the germinal layers of the epidermis and the appendages that differentiate
toward glandular structures (Scotto and Fraumeni 1982).  As a rule,  these tumors
are slow growing and follow a relatively benign course, although on rare
occasions they may result in extreme morbidity, mutilation, or, if they
metastasize, death (Pollack et al. 1982).

    Basal cell carcinomas are believed to arise from a pluripotent epithelial
cell present in the epidermis.   It has been hypothesized that basal cell
carcinomas arise because an abnormal interaction between these pluripotent stem
cells and the surrounding connective tissue induces the cells to differentiate
neoplastically (Pollack et al.  1982;  Kent 1976).   These tumors appear to be very
stromal dependent,  however,  and it has been suggested that they will rarely
metastasize to a foreign tissue bed unless they take along a portion of their
stroma (Pinkus 1953).   This hypothesis has been confirmed in part by studies
that have shown that  basal cell tumor cells cultured in vitro in the absence of
accompanying connective tissue convert to a keratinizing epithelium (Flaxman
1972;  Kubilus et al.  1980).   It has also been shown that autotransplants fail if
the cells are transferred without stroma (Epstein et al.  1984).

    Squamous cell carcinomas are neoplasms of the epidermis that differentiate
towards keratin formation (Scotto and Fraumeni 1982).   SCCs are far  less

-------
                                     7-18
fastidious in their growth requirements in vitro and will grow under a variety
of conditions.   They are generally more aggressive than BCCs and account for
about three-fourths to four-fifths of the deaths attributable to nonmelanoma
skin cancer (Dunn et al.  1965).

    The most prevalent precursor skin lesions for solar-radiation-induced BCC or
SCC are actinic hypertrophy and keratoses (Urbach 1981, Schwartz and Stoll
1987).   Actinic keratosis affects nearly 100 percent of the elderly white
population and may be found among younger light-skinned individuals living in
sunny areas (Schwartz and Stoll 1987).   Both epidemiologic and experimental
evidence indicate that BCC and SCC may develop at sites of actinic keratosis
(Carter 1987; Schwartz and Stoll 1987).  Stoll and Schwartz (1987), in a review
of SCC, reported that approximately 12 percent of untreated actinic keratoses
may develop into SCC, but they noted that this figure may be as high as 25
percent.  The likelihood of metastasis from SCC arising in actinic keratoses
(excluding SCCs of the lip) appears to be low, with 0.1 percent to 2.6 percent
of the SCCs potentially becoming metastatic (Stoll and Schwartz 1987).

EXPERIMENTAL EVIDENCE

Cellular Studies

    Solar radiation and its component UV and visible wavelengths have been
observed to cause cellular changes, such as the induction of cell division
and/or differentiation, the loss of specialized functions (e.g., antigen
presentation),  mutation,  transformation, and death.  Evidence from cellular and
molecular studies clearly indicate that UVR can cause transformation in
mammalian cells in the absence of any confounding immunological, hormonal, or
physiological effects encountered in vivo.  Shorter UV-B wavelengths are
generally more effective in transforming mammalian cells.

    As already described, a single in vivo exposure of human skin to UV-B
radiation can cause the epidermis to thicken, thereby increasing the tolerance
of skin to subsequent radiation.  UV-B exposure also induces melanocyte division
(Rosdahl and Szabo 1978).  Some investigators studying the response of
melanocytes in trunk epidermis of black mice to UVR have concluded, however,
that the increased number of active melanocytes results from both the
proliferation and recruitment of amelanogenic melanocytes (Miyazaki et al.
1974).

    Ultraviolet light is mutagenic in both bacterial and mammalian cells.  In
UVR-treated bacteria, researchers have concluded that pyrimidine dimers are
responsible for much of the observed mutagenesis (Hall and Mount 1981).  UVR-
treated cells,  if subsequently treated with photoreactivating light, show a
substantially decreased frequency of mutants per survivor compared to
UVR-treated cells without subsequent photoreactivation treatment.  Peak et al.
(1984) derived action spectra for DNA dimer induction, lethality, and
mutagenicity in E. coli over wavelengths from 254 nm to 405 nm  (Exhibit 7-8) and
found  that all three endpoints decreased in efficiency in a similar fashion as
the wavelengths of radiation increased.  From 300 nm to 320 nm,
all characteristics showed differences of about 2.5 orders of magnitude.
Furthermore between about 250 nm and 320 nm, the values for the three  endpoints
either coincided with or closely paralleled Setlow's  (1974) proposed average DNA
action spectrum.

-------
                                          7-19
                                      EXHIBIT  7-8

                Ultraviolet Action Spectra for DNA Dimer Induction:
                              Lethality and Mutagenicity
                 ID'1  r
 o, •  Relative  dimer yield per quantum
A—A Relative  lethality  per quantum
•—• Relative  mutagenicity per quantum
 	Average  ONA spectrum
      (Setlow, 1974)
 A,«  Xenon lamp
 A, o  Hg lines
                                              P'\     A
                                                (Tyrrell, 1973)
                     250       300       350        400
                                      WAVELENGTH (ml
                          450
Source:  Peak  et al.  (1984).

-------
                                      7-20
    Experiments have also been performed using fibroblasts from xeroderma
pigmentOGum (XP) patients.  These patients have a genetic inability to repair
UV-induced pyrimidine dimers.   Setlow et al. (1969) compared the responses of XP
fibroblasts to normal human fibroblasts upon irradiation to UVR.  They observed
that normal human fibroblasts that were UVR irradiated, then stimulated to
undergo cell division by immediate replating, showed much higher numbers of
mutations than similarly irradiated normal cells that were incubated for seven
days as nondividing monolayer cultures.  These results indicated that normal
human cells can repair the damage induced by UVR (presumably by excision
repair).  XP cells, by comparison, produced a comparable number of mutants
whether replated immediately or given seven days of incubation.  Since XP cells
are unable to excise pyrimidine dimers, these experiments suggest that unexcised
dimers present at the time of cell division may have been responsible for the
production of mutations.

    In vitro transformation is thought to be correlated, albeit imperfectly, to
in vivo tumorigenesis (Heidelberg 1977).  As such, it has been used extensively
to characterize the potential carcinogenicity of a wide variety of chemicals and
physical agents.  Cells that have been induced in vitro to lose certain normal
growth controls are frequently, although not always, tumorigenic
in mice.  A hierarchy of transformational changes is recognized, and the ability
of cells to grow without attachment to a solid substrate (loss of anchorage
dependence) is generally accepted as.the best correlate to tumorigenicity
(Freedman and Shin 1974).

    Doniger et al. (1981) developed action spectra for transformation and
thymine dimer formation using a monochromatic light source (Exhibit 7-9).  In
dose-response studies comparing pyrimidine dimer formation and transformation of
Syrian hamster embryo cells, the slopes of the dose-response curves were not
always parallel.  The discordance was greatest at 290 run.  The lowest exposure
required for equivalent cell transformation, lethality, and pyrimidine dimer
formation was at 270 nm.  Comparing results at 290 nm, 297 nm, and 302 nm, the
respective doses required were 2.3 J/m  , 8.7 J/m  , and 25 J/m  for dimer
formation and 7.4 J/m^, 47 J/m^, and 97 J/Tcr for transformation.  A similar
spectrum was found for induction of anchorage independent growth in human
fibroblasts (Sutherland et al. 1981).  These authors also found a maximum
effectiveness at 265 nm and that transformation at 290 nm was  six times more
effective per photon than that at 297 nm.

Molecular Studies:  DNA and UV Radiation

    Chemical changes and biological damage  induced by ultraviolet light require
the absorption  of  light  energy (photons) by molecules within the target.  Each
type of molecule is capable of absorbing radiation only in specific wavelength
ranges.  Examination of  the absorption  spectra of molecules in biological
systems indicates  that a number of biomolecules can absorb radiation  in the  220-
to 400-nm region and thus may be critical targets  for  detrimental UV  effects
(Spikes 1983).  This section will review information only on the most important
molecule, DNA,  because of its putative  role  in mutagenesis, transformation,  and
carcinogenesis.

    As  already  mentioned, two possible  types of mechanisms can induce DNA damage
--a direct mechanism  resulting from absorption of  energy by DNA and  an  indirect
mechanism involving reactive  oxygen species.  A number  of different lesions  may

-------
                                      7-21
                                   EXHIBIT 7-9


                    Effectiveness of 297, 302 and 313*  NM  UVR

           at  Inducing Pyrimidine Dimers (D) and  Transformation  (•)
                                                                                 z
                                                                                 O
                               60    120    180


                                 Exposure, J/m2
4000   8000  12,000
                                                                             9  «
                                                                           - 6
                                                                                 O
                                                                                 z
                                                                                 K
                                                                                 a
                                                                                 3>^
                                                                                 Q
a
z
M
Q
M
s
M
K

a.
Source:   Adapted from Doniger et al.  (1981).

-------
                                      7-22
be induced in DNA by UV irradiation.  These include (1) pyrimidine diraers,  (2)
pyrimidine adducts, (3) single-strand breaks,  (4) double-strand breaks, and (5)
protein-DNA crosslinks.  Different wavelengths have different efficiencies  for
the production of these lesions.

    Pyrimidine Dimers and Adducts

    As described in an earlier section, pyrimidine dimers may be formed by
direct absorption of photons in the UV-B wavelength range.  Adjacent pyrimidine
molecules on the same strand of DNA become linked together by a cyclobutane ring
between the 5 and 6 carbon atoms  of each residue.  Thymine dimers are the most
likely pyrimidine dimers because  the excited state of thymine is a much lower
energy state than that of cytosine.  Cytosine co-dimers are also possible,  as
are heterodimers between thymine  and cytosine or uracil (Spikes 1983).

    Action spectra for four-membered ring pyrimidine adduct formation between
two consecutive bases on the same strand of DNA resemble the action spectra for
pyrimidine dimer formation (Patrick and Rahn 1976).   The efficiency of formation
of these photoproducts is about two to ten times lower than for cytosine-
cytosine and thymine-thymine dimers.  The proportion of these adducts varies
according to the base content of  DNA and becomes much higher when the ratio of
guanine-cytosine/adenine-thymine  is greater than or equal to one.

    Thymine glycols, another form of pyrimidine adduct, are defined as a group
of ring-saturated lesions of the  5,6-dihydroxydihydrothymine type and have been
detected in the DNA of human cells after irradiation at 254 run and 313 nm
(Hariharan and Cerruti 1977).   The formation of these lesions is probably caused
by the'action of hydroxyl radicals on the 5-6 double bond of thymine.  In both
the UV-B and UV-A range, thymine  glycols may result from the action of reactive
oxygen species produced by endogenous sensitizers.

    Thymine glycol lesions occur  with almost the same frequency as thymine
dimers at 313 nm, indicating their possible significance in the UV-B range
(Cerutti and Netrawali 1979).   Glycol lesions may undergo spontaneous decay to
form apyrimidinic sites in a fashion similar to that described for gamma-
radiation produced saturated thymine glycols (Dunlap and Cerutti 1975).  Only
about one-third of the thymine glycols are released from the DNA backbone,  but
there is little evidence to suggest that either the ring-saturated thymine or
the apyrimidinic decay products are lethal lesions in UV-irradiated DNA.

    Brash and Haseltine (1982) have shown that there is a linear relationship
between base damage incidence and mutation incidence.  For shorter wavelengths
(UV-B) it seems clear that pyrimidine dimers and 6-4 photoadducts are involved
in mutagenesis, but at 365 nm the correlation between dimers or adducts and
mutagenesis is not known (Peak et al. 1984).

    DNA Single-Strand Breaks

    DNA single-strand breaks can be induced directly by UV-B radiation or
indirectly by UV-A and visible light.  Analysis of the relative efficiencies for
the induction of single-strand breaks reveals an action spectrum that
corresponds with nucleic acid absorption at or below 313 nm (Peak and Peak
1986).  Above 313 nm to wavelengths as long as 546 nm, single-strand breaks have
been detected in human fibroblasts with an action spectrum in the visible range

-------
                                      7-23
resembling that of riboflavin  (Rosenstein and Ducore 1983)  (Exhibit  7-10).
There is, however, some suggestion that this latter observation may  be an
artifact of bilirubin photosensitization (Peak and Peak  1986).  Relative to
thymine dimers, single-strand breaks are induced only  to a  small extent by 254-
nm radiation, but as the wavelength is increased, the  proportion of  single-
strand breaks to thymine dimers increases.  At 313 nm  in E. coli. one single-
strand break is induced for every nine thymine dimers  (Cerutti and Netrawalli
1979).  Some single-strand breaks induced at 313 nm may  be  due to indirect
effects from photosensitizers and oxygen-dependent mechanisms  (Miguel and Tyrell
1983), but the spectral analysis indicates that the majority of single- strand
breaks induced at this wavelength are due to direct effects (Peak and Peak
1986).

    DNA Double-Strand Breaks

    DNA double-strand breaks occur about 80 times less frequently than
single-strand breaks at 313 nm in bacteria (Tyrrell 1984).  Both single- and
double-strand breaks are resealed within one hour in both prokaryotes and
eukaryotes.  Two repair processes, one slow and the other fast, repair damage
induced in E. coli by ionizing radiation.  There is some evidence that repair of
breaks induced at 313 nm may proceed in a similar fashion (Tyrell 1984).

    DNA-Protein Crosslinks

    DNA-protein crosslinks can be induced in human cellular DNA following
borderline UV-B radiation (290 nm) via a direct photon-absorbing mechanism.
Below 320 nm, there are approximately 40 DNA-protein crosslinks per  lethal
event.  As cells can survive 2 x 10^ DNA-protein crosslinks induced  at longer
wavelengths (405 nm),  it appears that such a small number of DNA-protein
crosslinks is not important in UV-induced cell lethality.  This assumes,
however,  that there are no interactions between dimers and DNA-protein
crosslinks (Peak et al. 1985).

    DNA-protein crosslinking is demonstrable in normal human fibroblasts
immediately after UV irradiation,  but this crosslinking  is partially reversed
after about 12 hours.   In fibroblasts from XP patients, crosslinking after
UV-exposure was not reversed and actually progressed with time.  This is
possibly a secondary change due to severe cell damage  (Fornace and Kohn 1976).
The abnormal sensitivity of XP cells to UVR has generally been attributed to be
a defective capacity to repair cyclobutylpyrimidine dimers in cellular DNA
(Smith and Paterson 1981).   The inability of XP cells to repair DNA crosslinks
suggests,  however,  that the lethality of UVR to XP cells may not be fully
explained by their inability to excise dimers but may also relate to the
persistence of DNA-protein crosslinks.

    Although the biological significance of DNA-protein crosslinks is not clear,
it would seem that these lesions are not lethal to the cell.  In normal cells,
the number of DNA-protein crosslinks per genome per lethal hit is greater than
900 (Peak and Peak 1986).   It can reasonably be concluded,  however,  that normal
cells  possess an ability to repair these lesions.

-------
       10
        ,-1
o
l_
Q.
c 10"

"o
    <0
       10
        -5
                                  7-24
                              EXHIBIT 7-10

                    Action Spectrum for the Induction
                      of Single-Strand Breaks in DNA
         250
                           »,° Alkaline Sucrose
                            o Alkaline Elution
                            ° D20 Experiments
                      \  Rosenstein and Ducore, 1983
                Peak and Peak,
                1982
                     350
                                         450
                                                    550
                          Wavelength (nm)
Source:  Peak and Peak (1986).

-------
                                      7-25
Animal Studies

    Observations that outdoor workers had a greater tendency to develop skin
cancer promoted Findlay (1928) to examine UV-induced careinogenesis in mice.
Findlay reported that not only did UVR alone induce skin tumors in mice,  but
tumors induced by topically applied tar appeared more rapidly if those mice were
subsequently exposed to UVR.  Since Findlay's early works, researchers have
found that tumors induced in mice by UVR are primarily squamous cell
carcinomas and fibrosarcomas (Epstein and Epstein 1962; Hsu et al. 1975;  Kligman
and Kligman 1981; Kripke 1977; Spikes et al.  1977; Stenback 1975; Strickland et
al. 1979; Winkelman et al. 1963) and are mostly monoclonal in origin (Burnham et
al. 1986).  There is no animal model for basal cell carcinomas.

    A number of studies have attempted to identify specifically those UV
wavelengths responsible for the observed carcinogenicity with varying results.
In early studies filters were used to remove shorter wavelengths from broadband
UVR so that the effect of different wavelengths could be studied.  In some of
the original studies on the carcinogenic action spectrum of UVR, Rusch et al.
(1941) reported that the carcinogenic wavelengths were between 290 nm and 334
nm; wavelengths greater than 334 nm and those at 254 nm were not found to be
carcinogenic.  Blum (1943), in contrast, concluded from his experiments that the
most effective wavelengths for producing tumors in mice were between 260 nm and
300 nm.

    Freeman (1975),  using monochromatic UVR,  exposed the ears of albino mice to
a weekly dose of 420 J/m" at 290 nm, 600 J/m2 at 300 nm, 7500 J/m2 at 310 nm,
and 49,500 J/m  at 320 nm.  Mice exposed to 290 nm developed no tumors, those
exposed to 300 nm and 310 nm developed tumors with the same median latent
period, and those exposed to 320 nm developed fewer tumors.  When the mice were
exposed to the same incident radiation at 300 nm and 310 nm, only the mice given
radiation at 300 nm developed tumors.  This result indicated that UVR at 300 nm
is a more potent carcinogen than UVR at 310 nm.

    In a more recent study, Cole et al.  (1986) investigated the ability of
wavelength weighting schemes based on three different action spectra to predict
the tumorigenicity of UVR (Exhibit 7-11).   Use of a weighting scheme for an
averaged spectrum based on DNA damage overestimated the importance of the
shorter wavelengths in inducing tumors.   Use of an action spectrum based on the
sensitivity of the Robertson-Berger (R-B)  meter underestimated the contribution
of the shorter wavelengths.  These authors found that the best predictor of UVR
effectiveness at inducing tumorigenesis in hairless mice was an action spectrum
based on the induction of "cutaneous edema" 48 hours after a single acute dose
(MEE 48).

    There has been a great deal of experimentation to determine if there is
dose-rate reciprocity in UV carcinogenesis.   Blum et al. (1942) studied the
effect of the intensity of the dose on the tumor latency period and reported
that for a 10-fold dose range of between 4.3  J/m /sec and 0.42 J/m /sec,  no
significant differences were found.  In a second study in which a wider range of
intensities and greater numbers of animals were used, he reported that above
0.4 J/m^/sec the effectiveness of UVR in inducing tumors was almost independent
of dose-rate,  but below 0.4 J/m2/sec the effectiveness fell off rapidly with
intensity.

-------
                                  7-26
                              EXHIBIT 7-11



                  Action Spectrum of Mouse Edema (MEE48)

       as Compared to that of DNA Damage and the Robertson-Berger Meter
            100.0
          w
          w
          2
          H
          EH
          U
          w
          H

          EH
          W
          OS
             10.0
               1.0
0.1
                    DNA
                                         \  R-B
                    260   280   300    320   340

                            WAVELENGTH (nm)
Source:  Cole et al.  (1986).

-------
                                      7-27
    To determine if the dose-rate affects tumor development, the amount of
energy necessary to produce tumors in 50 percent of the tested animals can also
be calculated.  De Gruij1 et al. (1983) stated that the total dose delivered to
hairless mice to induce tumors must be greater if a high daily dose is given
than if a low daily dose is given.  The greater dose given to the mice was
9.4 x 103 J/m2/wk.  This same trend was reported by Spikes et al. (1977) in C3H
mice.

    Experiments on photocarcinogenesis in laboratory animals may shed light on
the role of UVR in the development of human nonmelanoma skin tumors.  Results
from animal studies clearly show that UVR is carcinogenic and that UV-B
wavelengths (290-320 nm) are most effective in inducing a response.

Immunosuppression

    The risk of developing nonmelanoma skin tumors may be influenced by the
immune status of a UVR-exposed animal or human.  For example, immunosuppressed
mice were observed to be less able to reject skin treated with the carcinogen
3-methylcholanthrene than normal mice.  In addition, mice exposed to UV light
and treated with immunosuppressive drugs developed more tumors than animals
exposed to UV light alone or treated only with immunosuppressants (Smith and
Brysk 1981).  UVR itself is selectively immunosuppressive (Kripke et al. 1977;
Spellman and Daynes 1977), inducing the production of a class of lymphocytes,
T-suppressor cells, which depress the host's ability to respond to UVR-induced
tumor antigens (Fisher and Kripke 1977, 1978; Spellman and Daynes 1977, 1978;
Daynes et al. 1979).  Exposure to UVR has also been found to result in the
impairment of antigen presenting cell function in the mouse (Greene et al.
1979).  In animals with primary UVR induced tumors, this impairment is
associated with the appearance of UV-tumor antigen-specific suppressor
lymphocytes (Kripke et al. 1977; Daynes et al. 1977).

    The T-suppressor cell is thought to prevent the generation of the cytotoxic
("T-killer") lymphocytes (Romerdahl and Kripke 1986), which would normally kill
UVR-induced tumor cells, having recognized them as foreign by the new antigens
present on their surfaces (Kripke 1977).  Other forms of immunosuppression
(e.g., via administration of an antilymphocyte serum or 6-mercaptopurine) give
conflicting results, sometimes enhancing and sometimes depressing
photocarcinogenesis (Nathanson et al. 1976).  (A more detailed discussion of
this subject is presented in Chapter 9.)

EPIDEMIOLOGIC EVIDENCE

Occurrence and Trends

    Nonmelanoma skin tumors are among the most common neoplasms occurring in
white populations.  Based on a one-year survey (1977-1978) conducted by the
National Cancer Institute (NCI), it was determined that nonmelanoma skin cancers
developed in approximately 400,000 white Americans each year (Scotto and
Fraumeni 1982).  The annual age-adjusted incidence rate for this survey period
was estimated to be 232.6/10^ among whites.  For comparison, the estimated
incidence rate for all other cancers among whites in the United States based on
1973-1976 data from the Surveillance, Epidemiology and End Results (SEER)
Program was 318.9/10^ (NCI 1985).  Among blacks, the annual age-adjusted

-------
                                      7-28
incidence rates for nonmelanoma skin tumors and all other cancers (NCI 1985)
were 3.4/1CH and 347.3/10 ,  respectively.

    The incidence of BCC is generally several times greater than the incidence
of SCC.  For example, based on data collected for 62 skin cancer cases
registered from 1956-1960 in three public hospitals in New Zealand,  Eastcott
(1963) observed that 73 percent of the cases had BCC,  15 percent had SCC, and 7
percent had cutaneous malignant melanoma (CMM).   Lee (1982) presented data on
2,019 skin cancer cases in Switzerland for 1974-1978 and showed that
approximately 69 percent of the cases had BCC, 20 percent had SCC, and 11
percent had CMM.  Among 2,000 individuals whose head,  neck, forearms, and dorsa
of hands were examined in a survey in Alfred Hospital in Victoria from
1982-1983, 71 percent of the histologically confirmed skin tumors were BCC and
29 percent were SCC (Goodman et al. 1984).  A similar distribution of SCC and
BCC was observed in a one-week study of 2,113 adults in 1982 in Maryborough,
Australia, whose heads, necks, forearms, and dorsa of hands were examined (Marks
et al. 1983).

    These relative differences in BCC and SCC persist when males and females are
examined separately even though the incidence in males generally exceeds that in
females.  Exhibit 7-12 presents age-adjusted BCC and SCC incidence data for
white American males and females (Scotto and Fraumeni 1982).  The incidence of
BCC was approximately four to six times greater than the incidence of SCC.

    The overall case fatality rate for nonmelanoma skin tumors is approximately
1-2 percent  (Epstein et al.  1984).  SCCs are, however, generally more aggressive
than BCCs, accounting for about-three-fourths to four-fifths of the deaths
attributable to nonmelanoma skin cancer (Dunn et al. 1965; NAS 1982).  SCCs may
metastasize  shortly after their appearance or even at the time of diagnosis
(Epstein 1983).  The rate of metastasis among SCC patients has been estimated by
several researchers to range between 2 percent and 20 percent.  Epstein  (1983)
reported that 25 percent of SCC patients with metastasis were alive 5 years
after diagnosis, 13 percent were alive 10 years after diagnosis, and only 8
percent were alive after 15 years.  Approximately two-thirds of metastasizing
SCCs were observed to occur on the face and dorsa of the hands.   BCCs, in
contrast, rarely metastasize even when present for many years (Pollack et al.
1982).  Their aggressiveness may be correlated with the site of tumor origin.
The rate of  metastasis among BCC cases in the United States has been reported to
be 0.1%, 9 of 9,050 cases (Cotran 1961); 0.4%, 2 of 499 cases (Hughes 1973); and
0.55%, 1 of  183 cases (Cade 1940).  In an Australian study (Paver et al. 1973),
the rate of  metastasis was estimated at 0.0028% (total number of cases not
given).  The time between primary tumor onset and metastasis has been estimated
to range from 0 to 43 years with a mean time of approximately 10-11 years
(Blewitt 1980; Farmer and Helwig 1980; Cotran 1961).  Once metastasis occurs,
the patient's prognosis is poor, with the time to death ranging from 0.8 to 1.6
years  (Farmer and Helwig 1980; Coker et al. 1983).  As of 1977,  only about 109
cases of metastatic BCC had been reported in the United States  (Safai and Good
1977).

    Two sources of mortality data are available for nonmelanoma skin cancer,
hospital records and death certificates.  Hospital record data have been
obtained, for example, as part of the Third National Cancer Survey  (TNCS) which
was conducted in two states and seven metropolitan areas from 1969-1971.  Death
certificates may be obtained from the National Center for Health Statistics

-------
                              7-29
                          EXHIBIT 7-12

 Comparison of Age-Adjusted Incidence Rates Per 100,000 Persons
for Squamous Cell Carcinoma (SCC) and Basal Cell Carcinoma (BCC)
       Among White Hales and Females in the United States
                                    Age-Adjusted
                                  Incidence Rate/10^
         Skin Tumor Type          Males      Females
SCC
BCC
65.4
246.6
23.6
150.1
         Source:  Scotto and Fraumeni  (1982).

-------
                                      7-30
(NCHS) or from state health departments.  The validity of these mortality data,
however, is uncertain (Dunn et al. 1965).  The reporting of deaths due to
nonmelanoma skin cancer may be biased due to the common perception that
nonmelanoma skin cancer is not a. lethal disease.  Also, since nonmelanoma tends
to be a more prevalent disease among older persons, the mortality due to
nonmelanoma skin cancer may be underreported due to competing risks of death
(Scott and Straf 1977).  The reporting of nonmelanoma deaths in hospital records
may also be biased downwards because affected patients are often not treated in
hospitals.

    Time Trends

    Several researchers have observed that the incidence of BCC and SCC have
been increasing over the past several decades (Lee 1982; NAS 1982).  Epstein et
al. (1984)  have pointed out that the rate of increase of SCC is greater than
that of BCC.  Scotto and Fraumeni (1982), however, noted when comparing the NCI
1971-1972 and 1977-1978 survey data that the observed incidence increases
applied mainly to BCC.  The authors noted that the incidence of BCC among United
States whites increased by approximately 15-20 percent over the six-year period
between surveys.

    In general, older individuals (e.g., over 60 years of age) are at higher
risk of developing nonmelanoma skin tumors than are younger individuals.
Vitaliano and Urbach (1980) observed that only three percent of a total 424 BCC
and SCC cases from the Tumor Clinic of the Skin and Cancer Hospital in
Philadelphia were under 40 years of age.  However, Emmett (1982) and Harris
(1982) have observed that SCC and BCC are no longer only diseases of old age
since an increasing number of younger individuals have been presenting with
nonmelanoma skin tumors.  Harris (1982) suggested that these occurrences were
consistent with increased sunlight exposure among these age groups.

    The age-specific incidence patterns for SCC and BCC are not identical,
suggesting that different etiological mechanisms may exist.  Although incidence
rates for SCC and BCC have been reported to rise with age and level off at the
oldest age groups (Scotto and Fraumeni 1982), the increase with age has been
observed to be sharper for SCC than BCC.  Laerum and Iversen (1981) summarized
study results indicating that among a group of BCC cases, 15 percent were less
than 50 years of age and 65 percent were less than 70 years of age.  Among a
group of SCC cases,  in contrast, Laerum and Iversen (1981) observed that 70
percent were over 70 years of age.

    Anatomical Distribution

    The predominant anatomical sites for both SCC and BCC and for both males and
females are generally the face, head, and neck.  Exhibit 7-13 presents the
distribution of BCC and SCC by sex for tumors occurring in whites in the United
States (Scotto and Fraumeni 1982).  The data were collected as part of the
1977-1978 NCI survey.  The face, head, and neck accounted for 60 percent or more
of the total nonmelanoma skin tumors.  Most of the remaining BCCs occurred on
the trunk,  whereas most of the remaining SCCs occurred on the upper extremities.
Skin tumor data from 1974-1978 from Switzerland (Exhibit 7-14) shows similar
tumor site distributions (Lee 1982).  Scotto and Fraumeni (1982) noted that the
tendency for nonmelanoma skin tumors to develop in exposed areas was consistent
with the belief that solar radiation is a dominant risk factor for nonmelanoma.

-------
                         7-31
                     EXHIBIT  7-13

       Percentage of Tumors by Anatomic Site for
     Nonmelanoma Skin Cancer Among White Males and
             Females in the United States
              (1977-1978 NCI Survey Data)
BCC
Anatomic Site
Face, Head and Neck
Trunk
Upper Extremities
Lower Extremities
Other Sites
Male
81
12
4
1
0
.2
.0
.9
.3
.5
Female
84
8
3
2
0
.1
.9
.4
.9
.7
sec
Male
74
4
18
1
1
.8
.5
.1
.3
.4
Female
60.
5.
25.
5.
3.
1
3
8
7
2
Source:   Scotto and Fraumeni (1982).

-------
                          7-32
                      EXHIBIT 7-14

  Distribution by Sex and Anatomic Site of Nonmelanoma
        Skin Tumors:   Canton of Vaud,  Switzerland
                       (1974-1978)
BCC
Anatomic Site
Head and Neck
Trunk
Upper Limbs
Lower Limbs
Other
Total
Male

540
(71.2)
130
(17.1)
14
(1.8)
15
(2.1)
58
(7.6)
758
(100.0)
Female
505
(78.7)
76
(11.9)
'17
(2.6)
18
(2.8)
26
(4.0)
642
(100.0)
sec
Male

202
(80.5)
8
(3.2)
26
(10.3)
5
(2.0)
10
(4.0)
251
(100.0)
Female
103
(72.5)
11
(7.7)
19
(13.4)
7
(4.9)
2
(1.4)
142
(100.0)
Source:   Levi and Chapallaz (1981)  as cited in Lee (1982)

-------
                                      7-33
    There are, again, some general differences in the distributions of SCO and
BCC, which suggests that these nonmelanoma skin tumors may respond differently
to different dosages of sun exposure.  The available epidemiologic data
generally show that BCC is more likely than SCC to develop on regularly
unexposed sites.  Pollack et al.  (1982) noted that although it is well
established that sun-exposed areas are more prone to BCC, in at least one study
(Urbach et al. 1972), approximately one-third of the BCCs occurred in
light-protected regions.  Laerum and Iversen (1981) observed that among a group
of SCC and BCC cases, 90 percent of the BCCs and approximately 50 percent of the
SCCs occurred on the head and neck.  They estimated that of these tumors,
two-thirds of the BCCs occurred on sites of the head and neck receiving the
highest UV radiation doses (e.g., the nose) whereas all the SCCs occurred at
these sites.  Hillstrom and Swanbeck (1970) examined only SCC cases and observed
that about 80 percent of the SCCs occurred on the head.  Facial SCCs occurred
relatively equally between males and females, but of the SCCs on the external
ear, 90 percent occurred among males and only 10 percent among females.   Emmett
(1982) examined only BCC cases and observed that 75.5 percent occurred on the
head and neck, 16 percent on the limbs, and 8.4 percent on the trunk.  Lee
(1982) noted that although BCC and SCC tend to be concentrated on exposed sites,
the distribution of BCC did not precisely correspond with sun-exposed areas.

Exposure Factors

    Several exposure factors considered to be associated with nonmelanoma skin
tumors have been identified in epidemiologic studies, including latitudinal
gradient, prolonged exposure to sunlight, and treatment with UVR.

    Prolonged sun exposure is considered to be the dominant risk factor for
nonmelanoma skin tumors among light-skinned populations (NAS 1982;  Scotto and
Fraumeni 1982; Greene and O'Rourke 1985; Lee 1982; Beral and Robinson 1981).  In
its 1982 report, the National Academy of Sciences (NAS) wrote that experts agree
that exposure to sunlight causes  90 percent or more of the BCCs and SCCs in the
United States.  Several observations supporting this hypothesis were cited in
the NAS report (NAS 1982) and by other researchers (Scotto and Fraumeni 1982;
Urbach 1981; Beral and Robinson 1981; Scotto et al. 1981; Laerum and Iversen
1982; Emmett 1982).  These observations include:

        o   the tendency for nonmelanoma skin tumors to develop in
            sun-exposed sites (e.g., head, face,  and neck);

        o   the higher incidence  rates among occupational groups with
            outdoor exposures compared to those with indoor exposures;

        o   the latitudinal and UVR gradient showing the highest
            incidence rates in geographic areas of relatively high UVR
            exposure;

        o   the increase in incidence rates with increasing age;

        o   the inverse correlation between nonmelanoma skin tumor
            incidence and degree  of skin pigmentation;

        o   the high risk among genetically predisposed individuals
            (e.g.,  those with xeroderma pigmentosum);

-------
                                     7-34
        o   the predisposition for nonmelanoma skin tumors to develop
            among light-skinned individuals who are susceptible to
            sunburn and who have red or blond hair, blue or light
            eyes, and a Celtic heritage;

        o   the increased incidence of SCC in individuals treated with
            high intensity UV-B sources and oral methoxsalen or
            topical coal tar;  and

        o   the capacity of UVR to induce nonmelanoma skin tumors in
            experimental animals.

    The observed variations in nonmelanoma incidence by latitude in particular
support an association with sun exposure.  For example, Scotto and Fraumeni
(1982) observed, based on the  1977-1978 NCI survey data, that the incidence of
SCC and BCC in the United States showed a latitudinal gradient with higher rates
in the south.   These results are displayed in Exhibits 7-15 and 7-16 for white
males and white females,  respectively.  (Also shown in these figures are
1973-1976 CMM data from the SEER program.  The latitudinal gradients for CMM
were least pronounced.)  It should be noted, however, that only a small number
of data points (as shown in Exhibits 7-15 and 7-16) were used to examine the
latitudinal gradients of BCC and SCC.

    Variations in the incidence ratio of BCC to SCC by latitude further suggest
that the two forms of nonmelanoma skin tumors respond differently to solar
exposure.  The results of MacDonald and Bubendorf (1964, as cited in Vitaliano
and Urbach 1980) showed that the BCC:SCC ratio decreased from approximately 10:1
in northern United States cities to approximately 2:1 or 3:1 in southern rural
areas.  Similarly, Scotto and Fraumeni  (1982) noted that the ratio of SCC
incidence to BCC incidence increased with decreasing latitude and increasing
sunlight exposure (see Exhibits 7-15 and 7-16).  Pollack et al. (1982) commented
that dosimetry studies revealed a poor correlation between BCC density in a site
and UVR dose.   They suggested that etiological factors for BCC, other than UVR,
such as the presence of areas of scarring or epidermal nevi, may exist.  Among
blacks, in whom nonmelanoma skin tumors occur rarely compared to whites, areas
of trauma or scarring may be important sites for the development of SCC (Scotto
and Fraumeni 1982).

    Vitaliano and Urbach (1980) examined several risk factors for SCC and BCC in
a case-control study.  The study included 366 BCC and 58 SCC cases seen at the
Tumor Clinic of the Skin and Cancer Hospital in Philadelphia (dates were not
specified).  A group of 294 white controls without carcinoma were selected from
the skin and cancer outpatient department.  Information compiled on each case
and control included cumulative solar exposure based on vocational and military
history, time spent sunbathing, and participation in outdoor sports (as a
spectator or participant).  Exposure was divided into four categories:  El =
0-9,999 hours, E2 = 10,000-19,999 hours, E3 = 20,000-29,999 hours, and
E4 = 30,000 or more hours.  The host factors that were examined included
complexion  (pale or mild-dark), age  (0-59 years or 60 and over), and ability to
tan (tans or burns-sensitive).

-------
                                         7-35
                                    EXHIBIT 7-15


              Annual Age-Adjusted Incidence Rates for  Basal Cell and

               Squamous Cell Carcinomas  (1977-1978 NCI Survey Data)

              and Melanoma  (1973-1976 SEER Data) Among White Males"'
      600

      500


      400


      300




      200
      100  -
   DC
   <
   cc
BASAL CELL
CARCINOMA
SQUAMOUS CELL
CARCINOMA
      10  -
                                                                   MELANOMA
-

1
,1
91

•1
I?
|


|
I
6
B
£
la
S3
L
1

ul
c
9
CO
, |


a
"c
i
(A
O

O
1
I ,
j_
1

CT
T ,
                     120
                                140          ISO          180



                             SOLAR ULTRAVIOLET (UV-B) RADIATION INDEX
   According to annual  UV-B measurements at selected areas  of the United  States,

with regression lines based on an  exponential model.
Source:  Scotto and Fraumeni (1982).

-------
                                       7-36
                                  EXHIBIT  7-16
             Annual Age-Adjusted Incidence  Rates for Basal Cell and
              Squamous Cell Carcinomas  (1977-1978 NCI Survey Data)
            and Melanoma (1973-1976 SEER Data)  Among White Females"/
  a.
  <
  
-------
                                      7-37
    Based on a statistical logistic regression of the case-control data,
Vitaliano and Urbach (1980) identified the most important risk factors for BCC
and SCC as follows:

        BCC:    cumulative solar exposure (p < 0.001) » ability to
                tan (p < 0.001) » age (p < 0.005) » complexion
                (p < 0.025)

        SCC:    cumulative solar exposure (p < 0.001) » age
                (p < 0.001) » ability to tan (p < 0.005).

Cumulative solar exposure was the most important risk factor for both SCC and
BCC.  Ability to tan was also important even at low levels of exposure.
Complexion was a less important risk factor for SCC than for BCC.

    Exhibit 7-17 presents the estimated relative risks (RRs) of BCC and SCC for
the combinations of risk factors considered in the Vitaliano and Urbach (1980)
study.  Vitaliano and Urbach (1980) concluded that the most important difference
between SCC and BCC was their relationship with cumulative exposure.  As shown
in Exhibit 7-17, a higher exposure level was required for BCC than for SCC to
reach similar RRs.  The authors noted that the maximum response of BCC to solar
exposure occurred in exposure category E4 (30,000 or more hours), whereas for
SCC it occurred in exposure category E3 (20,000-29,999 hours).  They observed
that the results were consistent with the belief that exposure to UVR has a
greater effect on the development of SCC than on BCC, although an association
between BCC and sunlight does exist.

    Additional evidence of an association between UVR and nonmelanoma skin
tumors is provided by cohort studies of psoriasis patients treated with high-
intensity UV-B radiation.  High-intensity UV-B sources would be expected to
damage DNA in the same way as 290-300 nm solar wavelengths.  The NAS (1982)
report noted that in one cohort study, elevated risks were associated with UV-B
exposure and topical crude coal tar.  Although treatment with photochemotherapy
(oral methoxsalen and UV-A) damages DNA in a different way than solar radiation,
BCC and SCC rates in 1,373 patients were three times higher than expected (NAS
1982).  In a more recent study by Stern et al. (1984) of 1,286 psoriasis
patients followed-up for an average of 5.7 years, the risk of SCC 22 months
after the first exposure to psoralen and UV-A (PUV-A) among those exposed to a
high dose was 12.8 times that for those exposed to a low dose, even after
adjustment for exposure to ionizing radiation and topical tar preparations.  No
substantial dose-related increase was observed for BCC.  Stern et al.   (1984)
concluded that the results showed a substantial dose-dependent increase in the
risk of SCC, even in those with neither a prior history of skin cancer nor
substantial exposure to cutaneous carcinogens.  The authors suggested that PUV-A
may act as both a co-carcinogen and an independent carcinogen in the development
of SCC.  In two European trials (Finland and Vienna-Innsbruck) no increased risk
of skin cancer was observed among patients treated with PUV-A (Stern 1984).  The
absence of a relationship may, however, have been due to different study
methods, shorter follow-up periods, less sun or occupational exposure, and lower
PUV-A doses in the European patients.

-------
                                      7-38
                                  EXHIBIT 7-17

              Estimated Relative Risks of Basal Cell and Squamous
              Cell Carcinomas for 32 Combinations of Risk Factors
   Exposure Grade
(Total Exposure, hr)
                                     Tans
                                   Burns -S ens11 ive
  0-59 yrs    	
Dark   Pale   Dark
         60+ yrs
                0-59 vrs
                             60+ vrs
               Pale   Dark
                      Pale   Dark    Pale
                                        Basal  Cell  Carcinoma
E4 (30,000 or more)
E3 (20,000-29,999)
E2 (10,000-19,999)
El (0-9,999)
   19
   86
   77
 1.00
4.94
4.43
2.75
1.55
4.
4.
2.
99
49
79
1.57
7.76
6.95
4.32
2.43
6.10
5.47
3.39
1.91
9.43
8.47
5.26
2.96
9.57
8.58
5.32
3.00
14.80
13.29
 8.25
 4.65
                                       Squamous  Cell  Carcinoma
E4
E3
E2
El
7
7
4
1
.09 22
.09 22
.42 5
.92 1
.79
.79
.72
.00
28
28
17
7
.61
.61
.94
.76
90.12
90.12
23.08
4.03
26.61
26.61
16.60
7.19
84.66
84.66
21.41
3.74
107.
107.
66.
29.
70
17
99
19
347.08
347.08
86.52
15.06
Source:  Vitaliano and Urbach (1980).

-------
                                      7-39
Host Factors

    As already mentioned, in addition to exposure factors, host factors play an
important role in the risk of nonmelanoma skin tumors.  Characteristics
associated with pigmentation, including eye, hair, and skin color, ability to
tan, tendency to sunburn, and ethnic origin are particularly important.
Evidence presented in the 1982 NAS report and by numerous other researchers has
identified those with fair skin, blue or green eyes, red or blond hair, a
tendency to burn and to rarely tan, and a Celtic heritage as being at high risk
of developing nonmelanoma skin tumors (NAS 1982).

ImmunosuppressIon

    One of the first reports of cancer occurring de novo in immunosuppressed
transplant patients was published in 1968 (Doak et al. 1968).  Since that time,
a substantial body of evidence has accumulated showing that impairment of an
individual's immune defense system may permit or facilitate the development of
skin cancers (Smith and Brysk 1981; Penn 1981; Blohme and Larko 1984).  In fact,
the most common malignancies developing in immunosuppressed individuals are
cancers of the skin, with the greatest frequency occurring, as in the general
population, in sun-exposed areas (Shell 1986).  Those at risk include organ
transplant patients and others receiving immunosuppressive drugs, patients with
lymphoma, leukemia, and other diseases who have defective immune responses, and
even those with age- or dietary-related decreases in immune responsiveness (Penn
1981; McMurray 1984; Smith and Brysk 1981).   Gupta et al.  (1986) estimate that
over 7,000 individuals will receive renal transplants each year in North America
alone.

    The epidemiologic studies in this area have predominantly focused on
individuals receiving immunosuppressive treatment, particularly organ transplant
patients.  The risk of developing skin cancer has been observed to be greater
among transplant recipients than in the general population, and the risks appear
to also be greater in geographical areas receiving higher amounts of UVR (Smith
and Brysk 1981;  Blohme and Larko 1984; Gupta et al.  1986;  Sheil 1986; Penn 1981;
Kinlen 1985) .   Skin cancer incidence rates among organ transplant recipients
have been observed to be approximately 3-8 times greater than in the general
population in low sun exposure areas (e.g.,  the United States,  United Kingdom,
Canada, and Sweden) and approximately 16-21  times greater in high sun exposure
areas (e.g., Australia and South Africa) (Hoxtell et al.  1977;  Hardie et al.
1980; Gupta et al.  1986; Penn 1981; Blohme and Larko 1984; Smith and Brysk
1981).

    Most of the skin cancers that develop in immunosuppressed individuals are
SCCs, in contrast to the observed predominance of BCC over SCC (approximately
4-6 BCC:1 SCC)  in the general population.   The ratio of SCC:BCC among renal
transplant patients has been observed to range from 1.33  (Minnesota) and 2.3
(Toronto) to 11 (Sydney, Australia, and Denver,  Colorado)  and 16 (New South
Wales)  (Gupta et al.  1986).   Only one Swedish study has reported an SCC:BCC
ratio of less  than one (Blohme and Larko 1984).   In a study of 523 white renal
transplant patients in Toronto,  Gupta et al.  (1986)  estimated that the risks of
developing all  skin cancers,  SCC,  and BCC were 3.2,  18.4,  and 1.4 times higher,
respectively,  than in the general population.   The elevated risks for all skin
cancers and SCC were statistically significant.   Sheil (1986) observed that SCC
occurs  about 40 times more frequently among  transplant patients than among the

-------
                                      7-40
general population, whereas CMM occurs about five times more frequently.   Sheil
(1986) noted that it was not clear whether BCC occurs more frequently among
transplant patients.  Among 1,349 British transplant recipients given
immunosuppressive therapy, the incidence of SCC was observed to be five times
that in the general population (Penn 1981).

    The SCCs that develop in immunosuppressed individuals tend to metastasize
more frequently and be more aggressive than the SCCs observed in the general
population (Sheil 1986; Penn 1981).  Furthermore, the average age at which
immunosuppressed patients develop skin cancer is about 30 years less than the
average age among skin cancer cases in the general population (Penn 1981; Gupta
et al. 1986).  Sheil (1986) has reported that malignancies occur within a few
months following transplant, with the average presentation occurring in two to
three years post-transplant and the increased risk persisting indefinitely.
Gupta et al.  (1986) observed a mean time to development of skin cancer among 523
renal transplant patients of 86.6 months (24 to 181-month range).

    The available epidemiologic evidence on the incidence of skin cancer among
immunosuppressed individuals clearly indicates that immune mechanisms play an
important role in limiting the development of malignant skin tumors.  However,
the different patterns of disease among immune depressed populations compared
with the general population, e.g. (the disproportionate increase in SCC over
BCC, the abbreviated time to tumor development, the increased tumor
aggressiveness, and the increased tendency for tumor metastasis) indicate that
other more complex factors may be involved.   One potential factor may involve
viruses of human papilloma and herpes types.  It has been observed that at least
one-third of SCC patients among .immunosuppressed transplant recipients carry the
genome of the papilloma virus (Sheil 1986).   Gupta et al. (1986) suggest that
immunosuppression, UV light, and the papilloma virus may act synergistically in
the production of skin cancer.

DOSE-RESPONSE RELATIONSHIPS

    This section reviews  the results of several important studies that have
examined the quantitative relationship between exposure to UVR and the incidence
of nonmelanoma skin cancer.  The methodological issues that add uncertainty to
these estimates are also  discussed.  Much of the discussion in this section is
based on review papers by the National Research Council (NAS 1982), van der Leun
(1984), and Zeger et al.  (1986).

    The available evidence from experimental and epidemiologic studies clearly
indicates that UVR exposure is associated with the development of nonmelanoma
skin tumors.  In an effort to evaluate the potential impacts on nonmelanoma of
increased UVR exposures due to depletion of the ozone layer, several researchers
have attempted to quantify the relationship between UVR dose and the incidence
of nonmelanoma skin tumors.  Dose-response relationships between UVR and the
incidence of nonmelanoma  skin tumors can then be combined with estimates of the
effects of changing ozone levels on UVR to predict the potential effects of
ozone changes on the incidence of nonmelanoma skin tumors.

Previous Analyses

    Dose-response relationships  for exposure to UVR and nonmelanoma skin tumors
have been developed in both epidemiologic and experimental studies.  Six major

-------
                                      7-41
dose-response analyses are reviewed below; three are based on epidemiologic
(ecological) studies, two are based on animal experiments, and the last is based
on a molecular experiment.  In each of these studies, UV-B exposure is estimated
by weighting individual wavelengths in the UV-B spectrum according to biological
effectiveness.  A single weighting approach is referred to as an action
spectrum.  Several action spectra based on the UV-B measurement approach are
available including Robinson-Berger (R-B) meter, human erythema, and Setlow DNA.
The erythema and DNA action spectra weight those wavelengths thought to promote
sunburn and damage to DNA more heavily than other wavelengths.

    Early epidemiologic research in dose-response of UVR exposure and
nonmelanoma skin cancer was conducted by the National Research Council (NRC
1975), Scott and Straf (1977) and Scotto et al.  (1974).  These studies all used
nonmelanoma skin cancer incidence data from the 1971-1972 National Cancer
Institute (NCI) survey and UV doses measured by R-B meters at four locations
in the United States.  The NRC considered these early studies to provide only
crude estimates of the dose-response relationship (NAS 1982).

    Since these early works, additional epidemiologic data were made available
by NCI.  A new survey of skin cancer incidence was conducted by NCI in 1977-1978
for eight locations across the U.S.  These data consist of newly diagnosed cases
of nonmelanoma skin cancer, and are age-, sex-,  and cell type-specific for
whites only.  In addition, new readings from R-B meters at five locations in the
U.S. were provided.  These readings combined with the earlier R-B meter readings
provided UV-B data for all eight locations included in the 1977-1978 NCI survey.

    All three epidemiologic studies reviewed below use the 1977-1978 NCI survey
data for nonmelanoma cancer incidence.  The studies differ, however, in their
choice of an action spectrum and in their specification of the model used to
estimate the dose-response relationship.

    The relationship between UV-B and nonmelanoma skin tumors in the
epidemiologic studies is expressed as a biological amplification factor (BAF),
which is defined as the percent change in tumor incidence that results from a
one percent change in exposure to UV-B radiation.  Estimated BAF values can be
combined with predictions of the changes in UV-B radiation due to changes in
ozone to obtain estimates of the effects of ozone changes on nonmelanoma skin
tumor incidence rates.

    Scotto et al.  (1981) and Fears and Scotto (1983)

    Scotto et al.  (1981) estimated values for the BAF by correlating the
incidence of nonmelanoma skin cancer with UV-B dosage for eight locations in the
U.S.  Separate estimates were presented by location, sex, and cell type.  The
functional form of the model used was:

                             ln(R)  =  ln(a)  +  bU  + e

where R is nonmelanoma skin cancer incidence rate, U is UV-B radiation dosage,  a
and b are constants, and e is an error term.   The constant b is the BAF that
describes the UV-B dose-skin cancer incidence relationship.  Incidence rates
were obtained from the 1977-1978 NCI survey.   Age-adjusted incidence rates were
used for all estimates.  UV-B dosages were obtained from R-B meters located at
major metropolitan airports at each location.  The coefficients were estimated

-------
                                      7-42
using ordinary least squares regression.

    Scotto et al. (1981) found that although incidence rates varied by age, BAFs
did not vary significantly by age.   Their estimates of the BAFs for BGC ranged
from 1.32 to 2.59 for males and from 1.06 to 2.07 for females.   The estimates of
the BAFs for SCC were slightly greater,  from 2.08 to 4.09 for males and from
2.18 to 4.30 for females.   In both cases, higher estimates were obtained for
locations at lower latitudes.

    In a later study, Fears and Scotto (1983) again estimated BAFs by
correlating nonmelanoma skin cancer incidence with UV-B radiation dosage.  The
data used were the same as those used in the earlier study.  In this study Fears
and Scotto (1983) used two functional forms of the model.  The first is the
log-linear form used by Scotto et al. (1981) as described above.  The second is
the log-log form:

                           ln(R)  =  ln(a)  + bln(U)  +  e

Separate estimates were presented by sex, location,  and percent increase in UV-B
dosage but not by cell type.  The BAF estimates presented in this study ranged
from 1.4 to 3.2 for males and from 1.1 to 2.3 for females.  Higher estimates
were obtained for locations at lower latitudes and for higher percent changes in
UV-B dosage.

    Fears and Scotto (1983) concluded that the relationship between UV-B
radiation and nonmelannoma skin cancer incidence could be modeled using either
the log-linear or the log-log form.  They noted that there was no strong reason
to prefer one form of the model over the other.

    Scott (1981)

    Scott used the log-linear equation as specified in the Scotto et al.  (1981)
study discussed above.  Incidence data were also obtained from the 1977-1978 NCI
survey.  However, rather than using the R-B meter data for UV-B dosage, Scott
(1981) used the Setlow (1974) action spectrum for DNA damage.

    Scott (1981) concluded that the values of UV-B dose obtained from the DNA
damage weighting method were approximately linear with those obtained from the
R-B meter.  Therefore, the estimates for the BAFs were approximately the same as
the estimates made by Scotto et al.  (1981) using the R-B meter data.  Estimates
of the change in UV-B dosage associated with changes in ozone concentration,
however, were found to be sensitive to the method used.  Thus,  even though the
estimate of the BAFs were insensitive to the method used to measure UV-B dosage,
the estimated effect of changes in ozone concentration on nonmelanoma skin
cancer is sensitive to the choice of action spectrum.

    For example, using the DNA action spectrum, Scott (1981) estimated that a 5
percent ozone reduction would lead to an increase in BCC in Minneapolis-St. Paul
of 7.7 percent for males and 5.8 percent for females.  Using the R-B meter, with
the same functional model and cancer incidence data, Scotto et al. (1981)
estimated corresponding increases to be 5.6 percent for males and 4.4 percent
for females.  The reason for this difference lies in the differential
sensitivity of different weighting functions for ozone depletion  -- the optical
amplification factor.  The energy delivered in wavelengths from 295-300 nm

-------
                                      7-43
increases much more with ozone depletion than energy in wavelengths from 315-320
nm.  Thus the shape of the action spectrum significantly influences the effects
of a given ozone depletion.

    In an earlier paper, Scott and Straf (1977), using pooled data, pointed out
that estimates of the BAF will be biased because actual doses vary by
individual.  Using survey data from one location only, they estimated that to
correct this bias, the BAF's should be increased by 30 percent.  More recent
estimates of the bias indicate that BAF estimates should be increased by 70
percent (NRC 1982).

    Rundel (1983)

    The Rundel study is the third epidemiologic study to use the 1977-1978 NCI
survey data.  Measurements from the R-B meter were used to estimate UV-B
exposure.  However, Rundel used a different dose-response model than that used
by Fears and Scotto (1983) or Scott (1981) to estimate the BAFs.   The Rundel
model used age as a proxy for cumulative sun exposure and estimated skin cancer
incidence at a single location as a function of this proxy variable.  Separate
BAF estimates were made by sex, cell type, and location.

    The Rundel model was based on the assumption of reciprocity;  that is, the
probability of developing nonmelanoma skin cancer depends only on the cumulative
dose received and not on the time course of that dose.  The assumption of
reciprocity tends to bias the estimates of the BAFs upward.  It should be
recognized, however, that there is evidence that reciprocity may not be a valid
assumption (Fears et al. 1977; Green 1978; de Gruijl 1983).

    Rundel (1983) concluded that the probability of a nonmelanoma skin tumor
becoming observable in an individual was well described by a log-normal
distribution.  Estimates of the BAF for males ranged from 1.8 to 2.2 for BCC and
from 2.4 to 2.8 for SCC.  The corresponding ranges for females were 1.1 to 1.5
for BCC and 1.6 to 2.1 for SCC.  The higher values were again obtained for lower
latitudes.

    Rundel (1983) estimated that a 1 percent depletion in ozone would lead to a
1.78 percent increase in BCC and a 2.3 percent increase in SCC.  He concluded
that these estimates were insensitive to the choice of action spectrum.
Comparing dosage weighted by the R-B meter with dosage weighted by the DNA
damage spectrum, Rundel (1983) showed that with the R-B meter the effect of
ozone changes on UV-B radiation would be small and the effect of UV-B dosage on
skin cancer incidence would be larger.  Thus, the total effect of ozone changes
on skin cancer incidence would be approximately the same.

    Forbes et al. (19821

    In this study, hairless (Skh:HR) mice were exposed 5 days/week to UVR from a
xenon arc lamp attenuated by five different filter thicknesses (0.64, 1.0,  1.3,
2.0,  and 3.0 mm).  The thinner filters transmit increasing amounts of UVR at
wavelengths less than 295 nm.   For each filter thickness, four different
exposure periods were tested (43.6,  61.2, 85.5,  and 120 minutes/day).  Forbes et
al.  observed that the proportion of affected animals,  the mean number of tumors
per affected animal, and the size of individual tumors all tended to increase
with time exposed.

-------
                                     7-44
    The effects of varying filter thicknesses and action spectra were determined
by examining the median time to the appearance of the first tumor on each animal
(T50 or latent period).   When the reciprocal of T50 was compared to the daily
UVR dose expressed as R-B meter counts for all filter groups combined, a high
correlation was observed (r = 0.957).   However, when the reciprocal of T50 was
compared to dose (R-B meter counts) for each filter thickness separately,
different slopes were observed.  The latent period was observed to decrease at a
faster rate with increasing exposure for thinner filters than for thicker
filters.  These results indicated that incremental additions of shorter
wavelength UVR had a relatively greater carcinogenic effect than the R-B meter
predicted.  The authors concluded that, although the R-B meter is a reasonably
good model for the UV wavelength dependence of UV photocarcinogenesis, it may
underestimate the consequences of some environmental changes.

    A similar evaluation of the T50 data was carried out using the MEE48 action
spectrum (based on the acute "delayed" edema response in hairless mice).   For
all filter thicknesses combined, a better correlation between latent period and
daily effective dose (r = 0.995) was observed for the MEE48 action spectrum than
for the R-B meter spectrum.  When the filters were analyzed separately,
significant differences between separate filters and the combined filter data
were not observed, in contrast to the R-B meter analysis.  The authors noted
that these results suggested that the MEE48 action spectrum is a better model
for mouse photocarcinogenesis for simulated solar spectra at wavelengths greater
than 295 nm.

    de Gruiil et al. (1983)

    The dose-response analysis presented by de Gruij1 et al. (1983) was also
based on data from an animal experiment.  Mice were exposed to radiation from a
sunlamp for 3.25 hours per day at seven different levels ranging from 0 mJ/cm^
to 190 mJ/cm  .  The response was measured by observing the  initiation and growth
of tumors on the mice.  The results of this study showed that the initiation of
tumors was dependent on the dose, but that the growth and yield of tumors were
independent of dose.  These authors also concluded that reciprocity was not
supported, except at high dosage levels.

    Kubitschek et al. (1986)

    This  study presented a method for calculating the effects of a reduction in
atmospheric ozone upon the incidence of nonmelanoma skin tumors by calculating
the incidence of mutations in  E. coli.  The model that was  applied had the
following form:

            Sm =   E(6,A,W) T(A) e(A) dA

        where

            Sm       = Rate of induction of mutation
            0        = Solar zenith angle
            A        = Wavelength
            W        = Amount  of stratospheric ozone
            E(6,A,W) = irradiation at  earth's  surface
            T(A)     = rate of transmission  through  skin

-------
                                      7-45
            e(A)     = action spectrum from E. coli

    The Kubitschek et al. (1986) study predicted that a one percent reduction in
atmospheric ozone would lead to a 1.7 percent increase in mutation rate.  This
estimate was constant for solar zenith angles up to 70 degrees, a range that
includes most heavily populated areas of the globe.  Their conclusions were
based on the action spectrum for mutagenesis in E.  coli (Exhibit 7-18),
unpublished information in similar studies in mammalian cells, information on
the epidermal transmission rate of the various UV wavelengths, and consideration
of a BAF taken from van der Leun (1984).  They estimated that a 3-5 percent
depletion of stratospheric ozone would lead to an increase in nonmelanoma skin
cancer rates of about 10-20 percent.

Uncertainties in Estimates of Dose-Response

    There are a number of issues related to the studies described above that add
uncertainty to dose-response estimates.  The degree of uncertainty that is added
has not been fully quantified.

    A major factor contributing uncertainty to dose-response estimates is the
limited availability of epidemiologic data.  All of the epidemiologic studies
described above use the 1977-1978 NCI survey for nonmelanoma-tumor incidence
data.  These researchers, thus, used a single measure of UV-B dosage for all
individuals in a given location.  Actual dosages received by individuals will,
however, vary because of behavioral factors such as time spent outdoors.
Unfortunately, data on individual UV-B exposures are not currently available.
The incidence of nonmelanoma skin cancer will also vary by individual because of
genetic characteristics such as skin color.  If these factors are correlated
with location, the estimates could be biased.

    The issue of peak versus cumulative dose has not been addressed in these
studies.  The failure to consider this factor may be important since BCC and SCC
are believed to respond differently to solar exposure.  As described earlier in
this chapter, exposure to UVR is believed to have a greater effect on the
development of SCC than on BCC.

    Furthermore, the three epidemiologic studies discussed above cannot be
considered independent because they are all based on the same skin cancer
incidence database.  As more comprehensive data are collected for more
locations, across longer time periods, and for individuals, estimates of
dose-response will hopefully become more precise.

    Another important issue is the choice of action spectrum for measuring UV-B
dosage.  The estimate of the effect of ozone depletion on skin cancer incidence
is sensitive to the choice of action spectrum.  There is,  however, no current
consensus on what is the best action spectrum to use to estimate the
dose-response relationship.

-------
                                     7-46


                                 EXHIBIT 7-18

           Relative Mutagenicity of UV-B  as a Function of Wavelength
                         300                350
                           WAVELENGTH (X),nm
400
    The smooth line shows  the  fit  to  these  data by  the sum of two exponentials
at wavelengths above 293 nm, as  described in  the  text.  The broken line is the
fit after correction for spectrophotometer  slit width.

Source:  Kubitschek et al.  (1986).

-------
                                      7-47
    There  is also statistical uncertainty  in  the measurement  of UV-B  flux.   This
uncertainty is not captured  in regression  analysis, which  considers the UV-B
values as  certain.  As a result, dose-response coefficients are subject to  a
greater degree of statistical uncertainty  than is  indicated in the regressions.
Incorporating statistical uncertainty in the  estimates of  UV-B would  probably
result in  different dose-response coefficients, but the  direction and magnitude
of  this difference is not known  (Zeger 1986).

    Another issue is whether the models used  are valid for all scenarios  of
ozone change.  For large levels of ozone depletion, for  example, the  amount of
UV-B radiation reaching the  Earth's surface may be outside the range  of UV-B
data used  to estimate the dose-response relationship in  the models, or may  occur
in  a different pattern.

    Because it is impossible to experiment on humans, experimental (e.g.,
animal) studies may be used  to examine the association between UV-B and
nonmelanoma skin cancer.  The extent to which the animal response to  UV-B
resembles  the human response is not well known.  Furthermore, the intensity,
duration,  schedule, and range of the doses applied in an animal experiment  is
likely to  differ from the exposures received  by humans in  the ambient
environment.

    In conclusion, while quantified dose-response estimates provide critical
information relating skin cancer incidence rates to changes in the ozone  level,
they should be regarded with caution.  As briefly described above, there  are
numerous uncertainties associated with such dose-response  estimates.

Sensitivity of Dose-Response Estimates to Choice of Action Spectrum

    As noted above, one important methodological issue that adds uncertainty to
the estimates of dose-response is the choice  of action spectrum.  In  comparing
the DNA damage weighting method with the R-B meter method  used by Scotto  et  al.
(1981), Scott (1981) found that estimates of  the BAFs were approximately  the
same but that estimates of the effect of ozone depletion on UV-B dosages  were
different.  On the other hand,  Rundel (1983)  concluded that the effect of ozone
depletion  on skin cancer incidence was insensitive to the  choice of action
spectrum.

    An independent analysis of the sensitivity of dose-response estimates to the
choice of  action spectrum was conducted and is summarized  in Attachment A.   The
methodology used in this analysis follows that used by Fears and Scotto (1983).
The dose-response is estimated using three alternative action spectra --  R-B
meter,  human erythema,  and Setlow DNA.

    The major result of this analysis was that the estimated BAF is not
sensitive  to the choice of action spectrum.  The estimated coefficient using the
human erythema and Setlow DNA action spectra all fell well within one standard
error of the coefficients estimated for the R-B meter action spectrum (see
Exhibit A-3 in Attachment A).

    However,  because the optical amplification factor is sensitive to ozone
depletion,  (that is,  because UV-B at 295-300 ran increases more than at 315-320
nm)  estimates  of the effect of  ozone depletion on UV-B dosage were found  to
differ  significantly by choice  of action spectrum.   Use of an action spectrum

-------
                                      7-48
based on the effects to a biological target, of course, makes sense.
Nevertheless, an issue still exists about whether DNA, erythema,  cutaneous
edema, or some other biological action spectrum should be used.   The Setlow DNA
action spectrum was found to give the highest estimates of the effect of ozone
depletion on UV-B dosage of the three examined (see Exhibit A-4 in Attachment
A).  Thus, the choice of an action spectrum will influence estimates of the
effect of ozone depletion on skin cancer incidence.

-------
                                      7-49
                                  ATTACHMENT A

                Sensitivity Analysis of the Selection of Action
                       Spectra on Dose-Response Estimates
               Relating UVR Exposures to Nonmelanoma Skin Tumors


    This attachment presents the results of an analysis of the relationship
between changes in ozone levels and incidence rates of nonmelanoma skin tumors.
Estimating the impact of ozone change on tumor incidence requires two pieces of
information: (1) the response of UV-B radiation to changes in ozone, and (2) the
dose-response relationship between UV-B exposure and the incidence of
nonmelanoma skin tumors.  This analysis focuses on the sensitivity of the
selection of an action spectrum used to describe the relationship between UV
exposure and the incidence of nonmelanoma skin cancer.

    UV-B Exposure and Nonmelanoma Skin Cancer

    The relationship between UV-B exposure and the incidence of nonmelanoma skin
tumors was estimated from the power model used by Fears and Scotto (1983):

                ln(R) = a + bln(UV) + e

where R is age-adjusted incidence, e is an error term, and a and b are
constants.  The constant b is the BAF that describes the UV-B dose-skin cancer
incidence relationship.

    The coefficients a and b were estimated using weighted ordinary least
squares regression.  Two sets of data were required to develop these estimates:
(1) incidence rates for nonmelanoma skin tumors, and (2) measurements of UV-B
e>posure for individuals developing these tumors.  The sources of each data set
are discussed below.

    Incidence of Nonmelanoma Skin Cancer

    Age-specific incidence rates (new cases per 100,000 population) for
nonmelanoma skin cancer among whites were obtained from Scotto et al. (1981) for
eight locations in the United States (1977-1978 NCI survey data).

    UV-B Radiation

    Although there is substantial agreement that wavelengths within the UV-B
band are primarily involved in the development of nonmelanoma skin cancers,
there is uncertainty about which individual wavelengths are most important.  In
this analysis,  the relative effect of using three alternative action spectra in
regressions to quantify the coefficients relating UV-B dose to skin cancer
incidence relationship were:   Robertson-Berger (R-B) meter, human erythema, and
Setlow DNA.  The erythema and DNA action spectra weight those wavelengths
thought to promote sunburn and damage to DNA more heavily than other
wavelengths.

    Data on UV-B radiation were obtained from the NASA UV-B model (Serafino and
Frederick 1986).   This model uses mathematical relationships to estimate UV
solar flux along different wavelengths (UV-A and UV-B) reaching the Earth's

-------
                                      7-50
surface at any time of year and location on the globe.  To estimate UV flux, the
model uses information about global ozone abundance, terrain height, cloud
cover, cloud transmission, and surface characteristics.

    The NASA model provided estimates of UV-B radiation for each of the eight
U.S. locations.  The data were provided for all three action spectra, and for
each action spectrum, data were presented for three different time periods --
annual, a day in the month of June, and a clear day.  In addition to UV-B data
at current ozone levels, the NASA model provided estimates of radiation reaching
the Earth's surface at different ozone levels.

    To analyze the significance of measuring UV-B radiation using different
action spectra and time periods, Exhibit A-l shows correlations among the NASA
UV-B data for the three action spectra and for annual, month of June, and clear
day time periods.  The exhibit also shows correlations among actual UV-B
measurements (Scotto et al.  1981) and the modeled data.  The strong correlation
among these UV-B data suggests that dose-response coefficients are not highly
sensitive to the different measures.  Nonetheless, a regression analysis was
performed using dose-response coefficients for the annual time period for all
three action spectra.  Before performing the regression analysis, two weights
were applied to the data.  First, the data were weighted by the population of
each respective location.  Exhibit A-2 shows the populations of these locations.
Second, as a correction for heteroskedasticity, the data were weighted by the
inverse of the estimated variance of the logarithm of the age-adjusted incidence
rates.

    Exhibit A-3 presents regression estimates of dose-response coefficients for
basal cell and squamous cell skin cancers by sex.  Separate estimates are given
for all three action spectra (annual only).  The exhibit shows a low, middle,
and high estimate.  The low value represents the estimated coefficient minus one
standard error.  The high coefficient adds one standard error.  The
dose-response coefficients for SCC are higher than those for BCC for each of the
action spectra.  In addition, the coefficients for SCC were found to be of
greater statistical significance.  These findings are consistent with the belief
that cumulative UV-B radiation has a greater effect on the development of SCC
than on the development of BCC.  As already mentioned, epidemiologic
observations indicate that the ratio of basal cell to squamous cell skin cancer
decreases with decreasing latitude and that BCC is more likely to develop on
regularly unexposed sites.  In addition, the sensitivity of basal cell cancer
incidence rates to UV-B was higher for males than for females.  The relative
sensitivity by sex is reversed for SCC.

    It is important to recognize that the NASA model's UV-B flux values are
subject to statistical uncertainty.  This uncertainty is not captured in the
regression analysis, which treats the UV-B values as certain.  As a result, the
dose-response coefficients are subject to a greater degree of statistical
uncertainty than is indicated in the regressions.  Incorporating statistical
uncertainty into the estimates of UV-B would probably result in different
dose-response coefficients,  but the direction and magnitude of this difference
is not known.

-------
                                                  7-51
                                              EXHIBIT A-l
                      Correlation of Alternative Measurements of DV-B Radiation
                                                                     a/
                                for Ten Locations in the United States
b/ b/ b/
Annual Clear Day Month of June

Annual c/
R-B Actual
R-B Meter
Erythema
DNA
Clear Day
R-B Meter
Erythema
DNA
June
R-B Meter
Erythema
DNA
R-B
Actual R-B Erythema DNA R-B Erythema DNA R-B Erythema

1.00 0.979 0.988 0.987 0.940 0.977 0.977 0.882 0.955
1.00 0.990 0.984 0.964 0.975 0.970 0.935 0.975
1.00 0.999 0.941 0.985 0.986 0.883 0.959
1.00 0.932 0.984 0.987 0.866 0.949

1.00 0.986 0.955 0.938 0.967
1.00 0.999 0.869 0.947
1.00 0.850 0.936

1.00 0.978
1.00

DNA

0.967
0.989
0.971
0.963

0.968
0.960
0.951

0.967
0.999
1.00
  The ten locations include:   Seattle,  Washington;  Minneapolis,  Minnesota;  Detroit,  Michigan;  Des
  Moines, Iowa; Salt Lake City, Utah;  San Francisco-Oakland,  California;  Atlanta,  Georgia;  Dallas-Ft.
  Worth, Texas; New Orleans,  Louisiana; Albuquerque,  New Mexico.   Data for  Des  Moines  and Dallas-Ft.
  Worth were included.in the correlations; however, because nonmelanoma skin cancer  rates were
  available only for 1971-1972, these locations were excluded from the dose-response regression
  analysis.
b/
  Model data from the NASA UV model.
  Annual UV-B data from Scotto,  Fears and Fraumeni  (1981).   These data are actual UV-B measurements
  taken at National Weather Service stations.

-------
                          7-52


                       EXHIBIT A-?

Population Weights for Ten Locations in the United States
                  Location             Weight


           Seattle (King County)       1,000,000

           Minneapolis-St.  Paul       2,113,533

           Detroit                    4,353,413
                         a/
           Salt Lake City             1,461,000
                                b/
           San Francisco-Oakland      3,250,630

           Atlanta                    2,029,710

           New Orleans                1,076,204
                      c/
           Albuquerque                1,303,000
           a/
             State of Utah.

           b/
             San Francisco.
             State of New Mexico.
           Source:  State and Metropolitan Area Data
                    Book. 1982. U.S. Department of
                    Commerce,  Bureau of the Census,
                    U.S. Government Printing Office,
                    August 1982.

-------
                                      7-53
                                  EXHIBIT A-3

            Estimated Dose-Response Coefficients (and t-Statistics)
                 for Basal Cell and Squamous Cell Skin Cancers
                       (DV-B Dose-Skin Cancer Incidence)
           	R-B Meter	      Human Erythema       	Setlow DNA	
              a/               b/     a/               b/     a/               b/
           Low      Mid    High    Low      Mid    High    Low      Mid    High


Basal Cell

  Male     1.300   1.810   2.330   1.020   1.410   1.800   0.932   1.290   1.650
                   (3.53)                  (3.60)                  (3.60)

  Female   0.443   1.050   1.670   0.346   0.809   1.270   0.316   0.739   1.160
                   (1.73)                  (1.75)                  (1.75)

Squamous Cell

  Male     1.940   2.840   3.730   1.540   2.210   2.880   1.420   2.030   2.640
                   (3.20)                  (3.31)                  (3.33)

  Female   1.880   3.060   4.230   1.570   2.420   3.260   1.470   2.220   2.980
                   (2.60)                  (2.87)                  (2.94)


a/
  Estimated coefficient minus one standard error.

b/
  Estimated coefficient plus one standard error.

-------
                                     7-54
    Ozone Levels and Changes in UV-B Exposure

    The NASA UV-B model was used to estimate changes in UV-B radiation
associated with changes in ozone levels.   Exhibit A-4 shows representative
estimates for San Francisco.  In the exhibit, a 2 percent reduction in ozone
resulted in an estimated 1.6 percent increase in UV-B radiation as measured on
an R-B meter.  The estimated percentage increase is greater using the human
erythema or Setlow DNA action spectra.

   -Relationship Between Ozone Changes  and the Incidence of Nonmelanoma
    Skin Tumors

    The NASA estimates of the sensitivity of UV-B radiation to changes in ozone
were combined with the UV-B dose-skin cancer incidence coefficients.  Exhibits
A-5 and A-6 present the results of this analysis for an estimated 2 percent and
10 percent depletion in the ozone layer for San Francisco.  For example,  Exhibit
A-5 indicates that the steady-state incidence rate of basal cell skin cancer
could increase by 2.98 percent for males  and 1.72 percent for females in
response to a 2 percent depletion in ozone.

-------
                             7-55
                          EXHIBIT A-4

Estimated Percentage Changes in UV-B Radiation in San Francisco
      For  a Two Percent and Ten Percent Depletion in Ozone
Weighting
Function
R-B Meter
Human Erythema
Setlow DNA
Ozone
2 percent
1.6
3.5
4.3
Depletion
10 percent
8.6
19.0
23.0

-------
                                    7-56
                                 EXHIBIT A-5

       Percentage Change in Incidence of Basal Cell and Squamous Cell
     Skin Cancers for a Two Percent Depletion in Ozone for San Francisco
               R-B Meter	       Human Erythema            Setlow DNA
           Low    Mid    High     Low    Mid    High      Low    Mid    High
Basal

  Male     2.13   2.98   3.86     3.55   4.95    6.36     3.99   5.56    7.17
  Female   0.72   1.72   2.75     1.19   2.81    4.44     1.33   3.15    4.99

Squamous

  Male     3.20   4.72   6.24     5.41   7.86   10.36     6.14   8.89   11.72
  Female   3.10   5.09   7.11     5.52   8.64   11.81     6.36   9.76   13.32

-------
                                      7-57
                                  EXHIBIT A-6

        Percentage Change  in Incidence  of Basal  Cell  and Squamous  Cell
      Skin Cancers for a Ten Percent Depletion in Ozone for San Francisco
           	R-B Meter	      Human Erythema       	Setlow DNA	
            Low     Mid    High     Low     Mid    High     Low     Mid    High


Basal

  Male     11.34   16.14   21.24   19.10   27.34   36.14   21.15   30.41   40.45
  Female    3.73    9.07   14.80    6.11   14.87   24.32    6.72   16.43   26.97

Squamous

  Male     17.39   26.46   36.12   30.20   46.05   63.82   33.95   51.87   72.19
  Female   16.81   28.78   41.86   30.88   51.40   74.84   35.34   57.93   84.68

-------
                                      7-58
                                   REFERENCES
Anderson, R.R.  Tissue optics and photoimmunology.  In:  Photoimmunology. J.A.
    Parrish, M.L. Kripke, and W.L. Morison, (eds).  New York: Plenum Medical
    Book Co. pp 61-76 (1983).

Beral, V.,  and N. Robinson.  The relationship of malignant melanoma, basal and
    squamous skin cancers to indoor and outdoor work.  Br J Cancer 44:886-891
    (1981).

Blewitt, R.W.  Why does basal cell carcinoma metastasize so rarely?  Int J
    Dermatol 19:144-146  (1980).

Blohme, I.  and 0. Larko.  Premalignant and malignant skin lesions in renal
    transplant patients.  Transpl 37:165-167 (1984).

Blum, H.F.   Wavelength dependence of tumor induction by ultraviolet radiation.
    JNCI 3:533-537 (1943).

Blum, H.F., H.G. Grady, and J.S. Kirby-Smith.   Relationships between dosage
    and rate of tumor induction by ultraviolet radiation.  JNCI 3:91-97 (1942).

Brash, D.E., and W.A. Haseltine.  UV induced mutation hotspots occur at DNA
    damage hotspots.   Nature 298:189-192 (1982).

Briele, H.A. and T.K. das Gupta.  Natural history of cutaneous malignant
    melanoma.  World J Surg 3:255-270 (1979).

Bruls, W.A.G., H. Slaper, J.C.  van der Leun, and L.  Berrens.   Transmission of
    human epidermis and stratum corneum as a function of thickness in the
    ultraviolet and visible wavelengths.  Photochem Photobiol 40(4):485-494
    (1984).

Burnham, D.K., L.C. Gahring, and R.A. Daynes.   Clonal origin of tumors
    induced by ultraviolet radiation.  JNCI 76:151-158 (1986).

Cade, S.  Malignant Disease and Its Treatment By Radium.   John Wright & Sons,
    Ltd., Bristol, England, (1940).  As cited in Safai and Good (1977).

Carter, D.M. Basal cell carcinoma.  In:   Dermatology in General Medicine. 3rd
    ed., Fitzpatrick, T.B., A.Z. Eisen,  K.  Wolf, I.M. Freedberg, andK.F.
    Austen, (eds.), New York:  McGraw-Hill Book Co.  pp. 759-772 (1987).

Cerutti, P.A. and M.  Netrawali.  Formation and repair of DNA damage induced by
    indirect action of ultraviolet light in normal and xeroderma pigmentosum
    skin fibroblasts.  Proc 6th Int'l Long Rad Res:   424-432 (1979).

Coker, D.D., E.G. Elias, T. Viravathana, E. McCrea,  and M.  Hafiz.   Chemotherapy
    for metastatic basal cell carcinoma.  Arch Dermatol 119:44-50 (1983).

Cole,  C.A., D.  Forbes, and R.E. Davies.   An action spectrum for UV
    photocarcinogenesis.  Photochem Photobiol  43:275-284 (1986).

-------
                                     7-59
Cotran,  R.S.  Metastasizing basal cell carcinoma.   Cancer 14:1036-1040 (1961).

Daynes,  R.A.,  and C.W Spellman.   Evidence for the generation of suppressor
    cells by ultraviolet light.   Cell Immunol 31:182-	 (1977).

Daynes,  R.A.,  M.K. Schmitt, L.K.  Roberts, and C.W.  Spellman.  Phenotypic and
    physical characteristics of the lymphoid cells  involved in the immunity to
    syngeneic UV-induced tumors.   J Immunol 122:2458-2464 (1979).

Daynes,  R.A.,  W.E. Samlowski,  O.K. Burnham, L.C.  Gahring, and L.K. Roberts.
    Immunobiological consequences of acute and chronic UV exposure.  Curr Probl
    Derm 15:176-194 (1986).

Doak, P.B., J.Z. Montgomerie,  J.D.K. North, and F.  Smith.  Reticulum cell
    sarcoma after renal homotransplantation and azothioprine and prednisone
    therapy.  Brit Med J 4:746 (1968).  As cited in Sheil (1986).

Doniger, J., E.D. Jacobson, K. Krell, and J.A. DiPaolo.  Ultraviolet light
    action spectrum for neoplastic transformation and lethality of Syrian
    hamster embryo cells correlate with spectrum for pyrimidine dimer formation
    in cellular DNA.  Proc Natl Acad Sci USA 78:2378-2382 (1981).

Dunlap,  B., and P. Cerutti.  Apyrimidinic sites in gamma-irradiated DNA.  FEES
    Letters 51:188-190 (1975).

Dunn, J.E., Jr., E.A. Levin, G.  Linden, and L. Harzfeld.  Skin cancer as a
    cause of death.  Calif Med 102:361-363 (1965).

Eastcott, D.F.  Epidemiology of skin cancer in New Zealand.  NCI Monograph No.
    10:141-151  (1963).

Emmett,  A.J.  Basal cell carcinoma.  In:  Malignant Skin Tumors.   A.J. Emmett,
    and M.G.E. O'Rourke (eds).  Churchill Livingston, New York (1982).

Epstein, J.H., and W.L. Epstein.   Cocarcinogenic effects of ultraviolet light
    on DMBA tumor initiation in albino mice.  J Invest Dermatol 39:455-460
     (1962) .

Epstein, W.L.  Malignant sun induced squamous cell carcinoma of the skin.
    J Dermatol Surg Oncl 9:505-506  (1983).

Epstein, W.L., J.C. Bystryn, R. Edelson, P.M. Elias, D.R. Lowy, and S. Yuspa.
    Nonmelanoma skin cancer melanomas, warts and viral oncogenesis.  J  Amer
    Acad Dermatol 5(2):960-970 (1984).

Farmer,  E.R., and E.B. Helwig.  Metastatic basal cell carcinoma:  A
    clinicopathologic study of seventeen cases.  Cancer 46:748-757 (1980).

Fears, T.R., and J. Scotto.  Estimating  increases in skin cancer  morbidity due
    to increased  in ultraviolet radiation  exposure.  Cancer Investig l(2):119-26
     (1983).

-------
                                      7-60
Fears, T.R., J. Scotto,  and M.A. Schneiderman.  Mathematical models of age
    and ultraviolet effects on the incidence of skin cancer among whites in the
    United  States.  Am J Epidemiol 105(5):420-27 (1977).

Findlay, G.M.  Ultraviolet light and skin cancer.  Lancet (ii):1070-1073
    (1928) .

Fisher, M.S., and M.L. Kripke.  Systemic alteration induced in mice by
    ultraviolet radiation and its relationship to ultraviolet carcinogenesis.
    Proc Natl Acad Sci 74:1688-1692 (1977).

Fisher, M.S., and M.L. Kripke.  Further studies on the tumor-specific
    suppressor cells induced by ultraviolet radiation.  J Immunol 121:1139-1144
    (1978).

Fitzpatrick, T.B., and N.A. Soter.  Pathophysiology of the skin.  In:
    Pathophysiology:  The Biological Principles of Disease.   L.H. Smith, Jr.,
    and S.O. Thier (eds),  Philadelphia:  W.B. Saunders Co. pp. 1297-1330 (1985).

Fitzpatrick, T.B.  Ultraviolet-induced pigmentary changes:  Benefits and hazards
    In:  Therapeutic Photomedicine.  Honigsmann H. and Stingl G.  (eds.) Vienna,
    KARGER, Basel (1986).

Flaxman, B.A.  Growth in vitro and induction of differentiation in cells of
    basal cell cancer.  Cancer Res 32:462-469 (1972).

Forbes, P.D., R.E. Davies, F. Urbach,  D. Berger, and C. Cole.  Simulated
    stratospheric ozone depletion and increased ultraviolet radiation:  Effects
    on photocarcinogenesis in hairless mice.  Cancer Res 42:2796-2803 (1982).

Fornace, A.J. and K.W. Kohn.   DNA-protein cross-linking by ultraviolet
    radiation in normal human and xeroderma pigmentosum fibroblasts.  Biochim
    Biophys Acta 435:95-103 (1976).

Freedman, V.H., and S. Shin.   Cellular tumorigenicity in nude mice:
    correlation with cell growth in a sensi-solid medium.  Cell 3:355-359
    (1974).

Freeman, R.G.  Data on the action spectrum for ultraviolet carcinogenesis
    JNCI 55:1119-1122 (1975).

Gange, R.W., and J.A. Parrish.  Acute effects of ultraviolet radiation upon
    the skin.  In: Photoimmunology.   J.A.  Parrish,  M.L. Kripke,  and W.L. Morison
    (eds). New York:   Plenum Medical Book Co. pp. 77-94 (1983).

Goodman, G.J., R. Marks, T.S. Selwood,  M.W. Ponsford, and W.  Pakes.
    Nonmelanotic skin cancer and solar keratoses in Victoria --  clinical studies
    II.  Aust J Dermatol 25:103-106  (1984).

Green, A.C. and M.G.E.  O'Rourke.   Cutaneous malignant melanoma in
    association with other skin cancers.  JNCI 74(5):977-980 (1985).

-------
                                      7-61
Green, A.C..  Ultraviolet exposure and skin cancer: Response.  Am J Epidemiol
    107:277-80 (1978).

Greene, M.I., M.S. Sy, M. Kripke, and B. Benacerraf.  Impairment of antigen-
    presenting cell function by ultraviolet radiation.  Proc Natl Acad Sci USA
    76(12)-.6591-6595 (1979).

Gupta, A.K., C.J. Cardella, and H.F. Haberman.  Cutaneous malignant neoplasms
    in patients with renal transplants.  Arch Dermatol 122:1288-1293 (1986).

Hall, J.D,.  and D.W. Mount.  Mechanisms of DNA replication and mutagenesis in
    ultraviolet irradiated bacterial and mammalian cells.  Prog Nucleic Acid Res
    Mol Biol 25:53-126 (1981).

Harber, L.C., and D.R. Bickers.  Photoprotection.  In:  Pho to s ens i t ivi ty
    Diseases:  Principles of Diagnosis and Treatment. Harber, L.C., and D.R.
    Bickers, (eds).   Philadelphia:  W.B. Saunders Co. (1981).

Hardie, I.R., R.W. Strong, L.C.J. Hartley, P.W.H. Woodruff, andG.J.A. Clurie.
    Skin cancer in Caucasian renal allograft recipients living in a subtropical
    climate.  Surg 87:177-183 (1980).

Hariharan, P.V.,  and P.A. Cerutti. Excision of ultraviolet and gamma ray
    products of the 5,6-dihydrothymine type by nuclear preparations of xeroderma
    pigmentosum cells.  Biochim Biophys Acta 447:375-378 (1976).

Hariharan, P.C.,  and Cerutti, P.A.  Formation of products of the
    5,6-dihydroxydihydrothymine type by ultraviolet light in HeLa cells.
    Biochemistry 16(12):2791-2795 (1977).

Harris, T.J.  Squamous cell carcinoma.  In:  Malignant Skin Tumors.  A.J.
    Emmett andM.G.E. O'Rourke  (eds.).  New York:  Churchill Livingston (1982).

Heidelberg,  C.   Oncogene transformation of rodent cell lines by chemical
    carcinogens.   In:  Origins of Human Cancer.  H.H. Hiatt, T.D. Watson,  and
    J.A. Winster. (eds).   New York:   Cold Spring Harbor Laboratory (1977).

Hillstrom, L.,  and G. Swanbeck.  Analysis of etiological factors of squamous
    cell skin cancer of different locations.  Acta Derm Venereol 50:129-133
    (1970).

Hoxtell, E.O.,  J.S.  Mandel, S.S. Murray, L.M. Schuman, and R.W. Goltz.
    Incidence of skin carcinoma after renal transplantation.  Arch Dermatol
    113:436-438 (1977).

Hughes, J.M.  Metastatic basal cell carcinoma:  A report of two cases and a
    review of literature.  Clin Radiol 34:392-393 (1973).

Hsu, J., P.D. Forbes, L.C., Harber,  and E. Lakow.  Induction of skin tumors
    in hairless mice by a single exposure to UVR.  Photochem Photobiol
    21:185-288 (1975).

-------
                                      7-62
Hu, F.  Melanocyte cytology in normal skin, melanocytic nevi and malignant
    melanomas:  A review.  In:  The Pathology of Malignant Melanoma. A.B.
    Ackerman, (ed.), New York:  Masson Publishing, pp. 1-22  (1981).

Kaidbey, K.H., P.P. Agin, M.M. Sayre, and A.H. Kligman.  Photoprotection
    by melanin -- A comparison of black and Caucasian skin.  J Am Acad Dermatol
    1:249-260 (1979).

Kent, A.' Pathology of basal cell carcinoma.  In:  Cancer of the Skin.
    R. Andrade,  S.L. Gumport, and G.L. Popkin (eds.).  Philadelphia:  W.B.
    Sanders Co.,  pp. 845-882 (1976).

Kinlen, L.J.  Incidence of cancer in rheumatoid arthritis and other disorders
    after immunosuppressive treatment.  Am J Med 78 (Suppl lA):44-49 (1985).

Kligman, L.H., and A.M. Kligman.  Histogenesis and progression of ultraviolet
    light-induced tumors in hairless mice.  JNCI 67:1289-1297 (1981).

Kripke, M.L.  Latency, histology, and antigenicity of tumors induced by
    ultraviolet light in three inbred mouse strains.  Cancer Res 37:1395-1400
    (1977).

Kripke, M.L., J.S. Lofgreen, J. Beard, J.M. Jessup, M.S. Fisher.  In vivo
    immune responses of mice during carcinogenesis by ultraviolet irradiation.
    J Natl Cancer Inst, 1977, Oct:59(4):1227-30.

Kubilus, J., H.P. Baden, and N. McGilvray.  Filamentous protein of basal cell
    epithelioma:   Characteristics in vivo and in vitro. JNCI, 65:869-875 (1980).

Kubitschek, H.E., K.S. Baker, and M.J. Peak.  Enhancement of mutagenesis and
    human skin cancer rates resulting from increased fluences of solar
    ultraviolet radiation.  Photochem Photobiol (in press).

Laerum, O.D., and O.H. Iversen (eds.).  Biology of Skin Cancer (Excluding
    Melanomas') :   A series of workshops on the biology of human cancer.  Report
    No. 15, International Union Against Cancer, UICC Technical Report Series,
    Volume 63, Geneva (1981).

Lee, J.H.  Melanoma and Exposure to Sunlight.  Epidemiol Rev 4:110-136 (1982).

Marks, R.,  M.W.  Ponsford, T.S. Selwood, G. Goodman, and G., Mason
    Non-melanotic skin cancer and solar keratoses in Victoria.   Med J Aust
    2:619-622 (1983).

McMurray, D.N.  Cell-mediated immunity in nutritional deficiency.  Prog Food
    Nut Sci 8:193-228 (1984).

Miguel, A.G., and R.M. Tyrell.  Induction of oxygen dependent lethal damage by
    monochromatic UV-B (313 run) radiation:  Strand breakage,  repair, and cell
    death.   Carcinogenesis 4:375-380 (1983).

-------
                                     7-63
Miyazaki H.,  A. Kawada, Y. Tahaki,  T Sato,  and M.  Matsutani.   Effects of
    ultraviolet light on epidermal  dendritic cells of hairless mice.   In:
    Sunlight and Man:  M. Pathak, L Harber,  M. Seiji, and A.  Kukita,  (eds.),
    Tokyo:  University of Tokyo Press,  pp.  217-229 (1974).

Nathanson, R.B., P.D. Forbes,  and F. Urbach.  Modification of photocarcino-
    genesis by two immunosuppressive agents.  Cancer Lett 1:243-247 (1976).

National Academy of Sciences (NAS).  Causes and Effects of Stratospheric Ozone
    Reduction:  An Update.  Washington, D.C.:  National Academy Press (1982).

National Cancer Institute (NCI).   1985 Annual Cancer Statistics Review.
    Presented to the National Cancer Advisory Board, National Institutes of
    Health, Bethesda, Maryland.  (1985).

National Research Council (NRC).   Environmental Impact of Stratospheric Flight:
    Biological and Climatic Effects of Aircraft Emissions in the Stratosphere.
    Washington, B.C.:  National Academy of Sciences (1975).

Nordlund, J.J.  Genetic basis of pigmentation and the disorders of
    pigmentation.  In:  Pathology of Malignant Melanoma. A.B. Ackerman (ed.),
    New York:  Masson Publishing, pp. 23-45 (1981).

Pathak, M.A.   Activation of the melanocyte system by ultraviolet radiation
    and cell transformation.  Ann NY Acad Sci 453:328-339 (1985).

Patrick, M.H., and R.O. Rahn.   In:   Photochemistry and Photobiology of Nucleic
    Acids. Vol II, Chap 2.  S.Y.  Wang (ed.).  New York:  Academic Press (1976).

Paver, K., K. Poyzer, N. Burry, and M.  Deakin.  The incidence of basal cell
    carcinoma and their metastases in Australia and New Zealand.  Aust J
    Dermatol 14:53 (1973).

Peak, M.J., J.G. Peak, M.P. Moehring, andR.B. Webb.  Ultraviolet action
    spectra for DNA dimer induction, lethality, and mutagenesis in Escherichia
    coli with emphasis on the UV-B region.  Photochem Photobiol 40(5):613-620
    (1984) .

Peak, M.J., and J.G. Peak.  Induction of single-strand breaks in human
    cell DNA by far-ultraviolet,  near-ultraviolet, and visible radiations:
    Action spectrum and mechanisms.  (unpublished manuscript).

Peak, M.J., J.G. Peak, and C.A. Jones.  Different  (direct and indirect)
    mechanisms for the induction of DNA-protein crosslinks in human cells by
    far- and near-ultraviolet radiation (290  and 405 nm).  Photochem Photobiol
    42:141-146  (1985).

Pearl, D.H.  A stochastic cancer model involving the environment of
    transforming cells with application to  skin cancer.   Ph.D. Thesis,
    University of California, Berkeley  (1984).

Penn,  I.  Depressed  immunity and the development of cancer.  Clin Exp Immunol
    46:459-474  (1981).

-------
                                      7-64
Percy, C.,  E. Stanek, and L. Gloeckler.  Accuracy of cancer death certificates
    and its effect on cancer mortality statistics.  Am J Publ Health 71:242-250
    (1981).

Pinkus, H.   Premalignant fibroepithelial tumors of skin.  Arch Dermatol
    Syphilol 67:98-615 (1953).

Pollack, S.V., J.B. Gaslen, E.F. Sherertz,  andB.V. Jegasothy.  The biology
    of basal cell carcinoma:  A review.  J Amer Acad Dermatol 7:569-577 (1982).

Romerdahl,  C.A.,  and M.L. Kripke.  Regulation of the immune response against
    UV-induced skin cancers:  Specificity of helper cells and their
    susceptibility to UV-induced suppressor cells.  J Immunol 137:3031-3035
    (1986).

Romsdahl, M.M.,  and I.S.  Cox.  Biological aspects of pigment cells and
    malignant melanoma.  In:  Neoplasms of the Skin and Malignant Melanoma, from
    the 20th Annual Clinical Conference on Cancer, New York:  Year Book Medical
    Publishers,  Inc., pp. 251-277 (1976).

Rosdahl, I.K., and G. Szabo.  Mitotic activity of epidermal melanocytes in UV
    irradiated mouse skin.   J Invest Dermatol 70: 143-148 (1978).

Rosenstein, B.S., and J.M.  Ducore.   Induction of DNA strand breaks in normal
    human fibroblasts exposed to monochromatic ultraviolet and visible
    wavelengths in the 260-546 run range.  Photochem Photobiol 38:51-56 (1983).

Rundel, R.D.  Action spectra and estimation of biologically effective UV
    radiation.  Physiol Plant 58: 360-366 (1983).

Rusch, H.P., B.E. Kline,  and C.A. Baumann.   Carcinogenesis by ultraviolet rays
    with reference to wavelength and energy.   Arch Pathol 31:135-146 (1941).

Safai, B.,  and R.A. Good.  Basal cell carcinoma with metastasis.   Arch Pathol
    Lab Med 101:327-331 (1977).

Schwartz, R.A.,  and H.L.  Stoll,  Jr.   Epithelial precancerous lesions.  In:
    Dermatology in General Medicine. 3rd ed.   T.B. Fitzpatrick,  A.Z. Eisen, K.
    Wolff,  I.M.  Freedberg,  K.F.  Austen (eds.).  New York:  McGraw-Hill Book Co.
    pp. 733-796 (1987).

Scott, E.L.  Summary of studies  made since December 1979:  Estimating the
    increase in skin cancer that will result from the depletion of stratospheric
    ozone.   University of California, Berkeley (1981) (transmitted to J.R.
    Hibbs,  U.S.  Environmental Protection Agency,  Washington, D.C.)
    (unpublished).

Scott, E.L., and M.L. Straf.  Ultraviolet radiation as a cause of
    cancer.  In:   Origins of Human Cancer.  Book A. Incidence of Cancer in
    Humans. Vol.  4.  H.H. Hiatt, J.D. Watson, and J.A. Winsten,  (eds.).  New
    York:  The Cold Spring Harbor Laboratory, pp. 529-549 (1977).

-------
                                     7-65
Scott,  I.R.,  Harding,  C.R.,  and Barrett,  J.G.   Histidine-rich protein of the
    keratohyalin granules source of the free amino acids,  urocanic acid and
    pyrrolidone carboxylic acid in the stratum corneum.   Biochem Biophys Acta
    719:110-117 (1982).

Scotto,  J.,  and J.F.  Fraumeni,  Jr.   Skin (other than melanoma).   In:   Cancer
    Epidemiology and Prevention.  D.  Schottenfeld and J.F.  Fraumeni,  Jr.,
    (eds.),  Philadelphia:  W.B.  Saunders Co.,  pp.  996-1011 (1982).

Scotto,  J.,  T.R. Fears,  and J.F. Fraumeni,  Jr.   Incidence of Nonmelanoma Skin
    Cancer in the United States.  National  Cancer Institute,  U.S.  Department of
    Health and Human Services,  (NIH)  82-2433 (1981).

Scotto,  J.,  A.W. Kopf,  and F.  Urbach.   Nonmelanoma skin cancer among
    Caucasians in four areas of the United  States.  Cancer 34:1333-1338 (1974).

Setlow,  R.B., J.D. Regan, J. German,  and W.L.  Carrier.   Evidence that
    xeroderma pigmentosum cells do not perform the first step in the repair of
    ultraviolet damage to DNA.   Proc Natl Acad Sci USA 64:  1035-1041 (1969).

Setlow,  R.B.   The wavelengths in sunlight effective in producing skin cancer:
    A theoretical analysis.   Proc Natl Acad Sci USA 69:  3363-3366 (1974).

Sheil,  A.G.R.  Cancer after transplantation.  World J Surg 10:389-396 (1986).

Smith,  E.B.,  and M.M.  Brysk.  Immunity and  skin cancer.   South Med J 74:44-46
    (1981).

Smith,  P.J.,  and M.C.  Patterson.  Abnormal  responses to mid-ultraviolet light
    of cultured fibroblasts from patients with disorders featuring sunlight
    sensitivity.  Cancer Res 41: 511-518 (1981).

Spellman, C.W., and R.A. Daynes.  Modification of immunologic potential by
    ultraviolet radiation.  II.  Generation of suppressor cells in short term
    UV-irradiated mice.   Transplantation 24:120-126 (1977).

Spellman, C.W., and R.A. Daynes.  Properties of ultraviolet light-induced
    suppressor lymphocytes within a syngeneic tumor system.  Cell Immunol
    36:383-387  (1978).

Spikes, J.D.   UV-damage in biological systems.  In:  Experimental and
    Clinical Photoimmunology. Daynes,  R.A., and Spikes,  J.D. (eds.).  Boca
    Raton:  CRC Press Inc. pp. 28-41 (1983).

Spikes, J.D., M.L. Kripke, R.J. Conner, and E.J. Eichwald.  Time of
    appearance and histology of tumors induced in the dorsal skin of C3Hf mice
    by ultraviolet radiation from a mercury arc lamp.  JNCI 59:1637-1643 (1977)

Stenback, F.   Species-specific neoplastic progression by ultraviolet light
    on the skin of rats, guinea pigs,  hamsters and mice.  Oncology 31:309-225
    (1975).

-------
                                      7-66
Stern, R.S., N. Laird, J. Melski, J.A. Parrish, T.B. Fitzpatrick, and H.L.
    Bleich.  Cutaneous squamous-cell carcinoma in patients treated with PUVA.
    N Engl J Med 310(18):1156-1161 (1984).

Stoll, H.L., R.A. Schwartz.  Squamous cell carcinoma.  In:  Dermatology in
    General Medicine. 3rd, T.B. Fitzpatrick, A.Z. Eisen, K. Wolf, I.M.
    Freedberg, and K.F. Austen.  New York:  McGraw-Hill Book Co., pp. 746-758
    (1987).

Strickland, P.T., F.J. Burns, and R.E. Albert.  Induction of skin tumors in
    the rat by single exposure to ultraviolet radiation.  Photochem Photobiol
    30:683-688 (1979).

Sutherland, B.M., N.C. Delihas, R.P. Oliver, andJ.C. Sutherland.  Action
    spectra for ultraviolet light-induced transformation of human cells to
    anchorage independent growth.  Cancer Res 41:2211-2214 (1981).

Tyrrell.  Damage and repair from non-ionizing radiation, In:  Repairable
    Lesions in Microorganisms. A. Hurst, and A. Nasim (eds).  London:  Academic
    Press (1984).

Urbach, F., D.B. Rose, and M. Bonnem.  Genetic and environmental interactions
    in skin carcinogenesis.  In:  Environment and Cancer, (eds).  Baltimore:
    Williams and Wilkins Co., pp. 355-371 (1972).

Urbach, F., C.J. Witcop, and O.D. Laerum.  Skin cancer in man.   In:   Biology
    of Skin Cancer (Excluding Melanomas). O.D. Laerum and O.H.  Iverson (eds.),
    Geneva:  International Union Against Cancer,  pp. 58-86 (1981).

van der Leun, J.C.  Yearly Review:  UV-Carcinogensis.  Photochem Photobiol
    39(6):861-868 (1984).

Vitaliano, P.P., and F. Urbach.  The relative importance of risk factors in
    nonmelanoma carcinoma.  Arch Dermatol 116:454-456 (1980).

Winkelman, R.K., P.E. Zollman, and E.J.  Baldes.  Squamous cell carcinoma
    produced by ultraviolet light in hairless mice.   J Invest Derm 40:217-224
    (1963).

Zeger, S.L., L. Cameron, E.  Gold, and E. Emmett.   Nonmelanoma skin cancer and
    ultraviolet radiation.  Johns Hopkins School  of Hygiene and Public Health
    (1986),  (unpublished).

-------
                                   CHAPTER 8

                         CUTANEOUS MALIGNANT MELANOMA
SUMMARY

    In 1987, it is estimated that there will be 25,800 cases and 5,800 deaths
from cutaneous malignant melanoma (CMM) in the United States.  For more than a
decade, there has been serious concern that CMM is at least partially caused by
ultraviolet-B (UV-B) radiation.  However,  several aspects of the scientific
information about CMM have puzzled researchers and have contributed to
uncertainty about the relationship of CMM to solar radiation, in particular
UV-B.  In the past several years, some progress has been made in understanding
CMM and its possible relationship to solar radiation, and there now exists an
array of evidence that indicates that exposure to solar radiation, and to UV-B
in particular, is a likely cause of CMM.

    Supporting evidence includes (1) for this relationship the fact that people
who lack the protection of pigmentation which blocks penetration of ultraviolet
radiation (UVR) in the skin have higher CMM incidence rates; (2) a correlation,
in well-designed ecological studies, of higher CMM rates with decreasing
latitude and increasing UV-B levels; (3) the discovery of an association between
freckling and nevus formation  (risk factors for CMM) and solar exposure; (4)
differences in CMM rates found between natives and immigrants to sunny climates;
(5) high rates of CMM in patients who are genetically deficient at repairing DNA
damage induced by UV-B; and (6) the indication, in controlled studies, that sun
exposure at early ages and of an intermittent nature results in higher CMM
risks.

    The results of some studies have created uncertainty about the relationship
between solar radiation and CMM:  several econological studies failed to find
latitude gradients for CMM; outdoor workers have been found to have lower CMM
rates than  indoor workers (but higher non-melanoma skin cancer rates); and
anatomic sites with lower sun exposure have high CMM rates.

    The latter evidence has made it clear that the relationship between solar
radiation and CMM is not a simple one.  Indeed other information indicates that
the relationship is a complex one.  Part of the complexity lies in the nature of
CMM.  It is a multifactorial disease which includes several histologically
different types of tumors with different biologic behavior and clearly different
relationships to solar exposure.  Additional complexity is contributed by the
nature of the relationship of UV exposure to CMM; not only is there likely to be
a direct relationship via an initiating (DNA damaging) step, but an indirect
relationship via promotional events such as immunosuppression, is also likely.
Indeed the  information contributing to the uncertainty about the relationship
indicates that cumulative dose based on total number of hours in the sun may not
be the appropriate dose parameter to correlate with CMM development.  This has
led to the  emergence of hypotheses that postulate the importance of intermittent
exposure to high fluxes of UV-B (resulting in large UV-B doses) in the causation
and/or promotion of CMM.

-------
                                      8-2
    Unfortunately,  at this time,  there have been no studies which have examined
the correlation between CMM and actual measured UV-B doses received by
individuals, thus making it impossible to prove various hypotheses concerning
CMM.  The balance of available evidence appears to support the conclusion that
solar radiation is at least a major cause of CMM.   Other evidence appears to
indicate that cumulative sun-exposed hours is not an appropriate measure of dose
for CMM, and that CMM has a complicated etiology.

    A variety of different kinds  of evidence supports the hypothesis that UV-B
is a component of solar radiation that causes CMM.  Most important is the fact
that xeroderma pigmentosum patients, who are genetically deficient in repairing
UV-B-induced DNA damage, have significantly higher rates of CMM than the general
population.  Another fact that supports this hypothesis is that, in experimental
studies, UV-B is the most mutagenic and carcinogenic waveband as well as being
the most active at causing immunosuppression.  All of this information suggests
that UV-B is the primary component of solar radiation that causes CMM.
Together, this evidence does not provide proof that UV-B is the waveband of
solar radiation that causes CMM;  however, the preponderance of evidence makes
UV-B a likely cause of CMM.

    Finally, although much remains uncertain about the relationships among solar
radiation, UV-B, and CMM, enough information exists to make reasonable estimates
of the dose-response relationships that exist between UV-B and CMM.  These can
then be used to estimate future incidence and mortality of CMM if ozone
depletion occurs.  The possible relationships based on an improving but still
deficient database in the United States indicate that each 1 percent ozone
depletion will produce around a 1 to 2 percent increase in incidence and 0.8 to
1.5 percent increase in mortality.  At this time,  it is impossible to estimate
how incidence and mortality of CMM will change in other countries whose
populations have unknown skin pigmentation and sun exposure behavior.

-------
                                      1-3
FINDINGS

1.  CUTANEOUS MALIGNANT MELANOMA (CMM)  IS A SERIOUS LIFE-THREATENING DISEASE
    THAT AFFECTS A LARGE NUMBER OF PEOPLE IN THE UNITED STATES.   THERE ARE
    SEVERAL HISTOLOGICAL FORMS OF MELANOMA THAT ARE LIKELY TO HAVE SOMEWHAT
    DIFFERENT ETIOLOGIES AND RELATIONSHIPS TO SOLAR AND UV-B RADIATION.

    la.  CMM incidence and mortality is increasing among fair-skinned
         populations.  These increases  appear not to be merely the result of
         improved diagnosis and reporting.

    Ib.  In 1987, it is estimated that  there will be an estimated 25,800 cases
         of CMM and 5,800 fatalities related to melanoma in the United States.
         In the absence of ozone depletion, the lifetime risk of CMM in the
         United States is expected to be about 1 in 150.

2.  LIMITATIONS IN THE DATABASE PREVENT ABSOLUTE CERTAINTY ABOUT THE
    RELATIONSHIP OF SOLAR RADIATION. UV-B. AND CUTANEOUS MALIGNANT MELANOMA
    (CMM).

    2a.  There currently is no animal model in which exposure to UV-B radiation
         experimentally induces melanomas.

    2b.  There is also no experimental  in vitro model for malignant
         transformation of melanocytes.

    2c.  No epidemiologic studies of CMM have been conducted in which individual
         human UV-B exposures (and biologically effective doses of solar
         radiation) have been adequately assessed.

3.  EVALUATION OF THE EPIDEMIOLOGICAL AND EXPERIMENTAL DATABASES FOR MELANOMA
    REQUIRES CLOSE ATTENTION TO THE RELATIONSHIP OF WAVELENGTH AND DOSE AND TO
    THE VARIATIONS OF SOLAR RADIATION IN THE AMBIENT ENVIRONMENT.

    3a.  Ozone differentially removes wavelengths of UV-B between 295 and 320
         nm; UV-A (320-400 nm) in wavelengths above 350 nm is not removed, nor
         is visible light (400-900 nm).  Ozone removes all UV-C (i.e.,
         wavelengths less than 295 nm).

    3b.  Wavelengths between 295 nm and 300 nm are generally more biologically
         effective (i.e., damage target molecules in the skin, including DNA)
         than other wavelengths in UV-B and even more so than UV-A radiation.

    3c.  Latitudinal variations exist in solar radiation; model predictions
         indicate that the greatest variability is seen in cumulative UV-B
         (e.g., monthly doses) followed by peak UV-B (highest one-day doses) and
         then cumulative UV-A.  Peak UV-A does not vary significantly across
         latitudes up to 60°N.  Greater ambient variation also exists in UV-B
         than in UV-A by time of day.

    3d.  The biologically effective dose of radiation that actually reaches
         target molecules depends on the duration of exposure at particular
         locations, time of day, time of year, behavior (i.e., in terms of
         clothes and sunscreens), pigmentation, and other characteristics of the

-------
                                      8-4


         skin including temporal variations (e.g.,  changes  in pigmentation due
         to tanning).

    3e.   Cloudiness and albedo,  although causing large variations  in the  amount
         of exposure  to UV-B and UV-A,  do not greatly change the ratio of UV-B
         to UV-A.

    3f.   Ozone depletion is predicted to cause the  largest  increases in
         .radiation in the 295-299 nm UV-B range, less in the 300-320 nm UV-B
         range;  UV-A is virtually unaffected by ozone depletion.

    3g.   Cutaneous malignant melanoma has a number  of different histologic types
         that vary in their relationship to sunlight, site,  racial preference,
         and possibly in their precursor lesions.   Assessment of incidence by
         types is not consistent among registries,  thus complicating attempts to
         evaluate the relationship between CMM and  solar radiation.

    3h.   Melanin is the principal pigment in skin that gives it color; melanin
         effectively absorbs UV radiation; the darker the skin, the more  the
         basal layer is protected from UV radiation.

4.   A LARGE ARRAY OF EVIDENCE SUPPORTS THE CONCLUSION THAT  SOLAR RADIATION IS
    ONE OF THE CAUSES OF CUTANEOUS MALIGNANT MELANOMA.

    4a.   Whites, whose skin contains less protective melanin, have higher
         incidence and mortality rates from CMM than do blacks.

    4b.   Light-skinned whites, including those who  are unable to tan or who tan
         poorly, have a higher incidence of CMM than do darker-skinned whites.

    4c.   Sun exposure leading to sunburn apparently induces melanocytic nevi.

    4d.   Individuals who have more melanocytic nevi have a higher incidence of
         CMM; the greatest risk is associated with a particular type of nevus --
         the dysplastic nevus.

    4e.   Sunlight induces freckling, and freckling is an important risk factor
         for CMM.

    4f.   Incidence has been increasing in cohorts in a manner  consistent with
         changes  in patterns of sun exposure, particularly with respect to
         increasing intermittent exposure of certain anatomical sites.

    4g.   Immigrants who move to sunnier climates have higher rates of CMM than
         populations who remain in their country of  origin.  Immigrants develop
         rates  approaching  those of prior  (but  native born)  immigrants to the
         adopted  country; this  is particularly  accentuated  in  individuals
         arriving before the age of puberty  (10-14 years).

    4h.  It has been suggested  that CMM risk may be  associated with childhood
         sunburn; other  evidence suggests  that  childhood sunburn may  reflect an
         individual's pigmentary characteristics or  may be  related to nevus
         development, rather  than being a  separate risk factor.

-------
    41.   Most studies that have used latitude as a surrogate for sunlight or
         UV-B exposure have found an increase in the incidence or mortality of
         CMM correlated to proximity to the equator.   A recent study of
         incidence using measured UV-B and CMM survey data found a strong
         relationship between UV-B and incidence of CMM.   Another study that
         used modeled UV-B data and an expanded database  on mortality found a
         strong UV-B/mortality relationship.

    4j.   One form of CMM, Hutchinson's melanotic freckle,  appears almost
         invariably on the chronically sun-damaged skin of older people.

5.   SOME EVIDENCE CREATES UNCERTAINTY ABOUT THE RELATIONSHIP BETWEEN SOLAR
    RADIATION AND CUTANEOUS MALIGNANT MELANOMA.

    5a.   Some ecologic epidemiology studies,  primarily in Europe or close to the
         equator, have failed to find a latitudinal gradient for CMM.

    5b.   Outdoor workers generally have lower incidence and mortality rates for
         CMM than indoor workers, which appears incompatible with a hypothesis
         that cumulative dose from solar exposure causes  CMM.

    5c.   Unlike basal cell and squamous cell carcinomas,  most CMM occurs on
         sites that are not habitually exposed to sunlight; this contrast
         suggests that cumulative exposure to solar radiation or UV-B is not
         solely responsible for variations in CMM.

6.   UV-B RADIATION IS A LIKELY COMPONENT OF SOLAR RADIATION THAT CAUSES
    CUTANEOUS MALIGNANT MELANOMA (CMM).  EITHER THROUGH INITIATION OF TUMORS OR
    THROUGH SUPPRESSION OF THE IMMUNE SYSTEM.

    6a.   Xeroderma pigmentosum patients who fail to repair UV-B-induced
         pyrimidine dimers in their DNA have a 2,000-fold excess rate of CMM by
         the time they are 20.

    6b.   UV-B is the most active part of the solar spectrum in the induction of
         mutagenesis and transformation in vitro.

    6c.   UV-B is the most active part of the solar spectrum in the i.nduction of
         carcinogenesis in experimental animals and is considered by most to be
         a causative agent of nonmelanoma skin cancer in humans.

    6d.   UV-B is the most active portion of the solar spectrum in inducing
         immunosuppression, which may have a role in melanoma development.

    6e.   The limitations in the epidemologic and experimental database leave
         some doubt as to the effectiveness of UV-B wavelengths in causing CMM.
    WHILE UNCERTAINTY EXISTS.  INCREASES IN THE INCIDENCE AND MORTALITY OF
    CUTANEOUS MALIGNANT MELANOMA ARE LIKELY AS A RESULT OF OZONE DEPLETION.
    WHILE MANY UNCERTAINTIES EXIST (E.G..  REGARDING ACTION SPECTRA.  PEAK VERSUS
    CUMULATIVE DOSE.  ETC.) ABOUT THE NATURE OF THE RELATIONSHIP BETWEEN UV-B AND
    MELANOMA. THE FACT THAT UV-B RADIATION VARIES ACROSS THE ENVIRONMENT IN THE
    RANGE OF VARIATION EXPECTED FROM DEPLETION PROVIDES INFORMATION USUALLY

-------
                                  8-6
UNAVAILABLE TO RESEARCHERS MAKING QUANTITATIVE RISK ESTIMATES.   THUS
ALTHOUGH IMPERFECT. EPIDEMIOLOGIC INFORMATION EXISTS TO ESTIMATE A RANGE OF
CHANGES IN INCIDENCE AND MORTALITY IF THE OZONE LAYER IS DEPLETED.

7a.  Uncertainty exists about the appropriate action spectrum to be used in
     estimating dose, the best functional form for dose- response,  and the
     best way to characterize dose (peak value, cumulative summer exposure,
     etc.)-  Histologically different CMMs (or possibly CMM located at
     different anatomical sites) are likely to have different dose-response
     relationships.  Most estimates of CMM dose-response relationships fail
     to consider these histological or site differences.  Nonetheless, by
     encompassing a range of possibilities, it is possible to estimate
     dose-response because of the systematic variations in UV-B.

7b.  A recent study by the NIH presents a well-designed ecological study of
     melanoma and UV-B using survey data and measured UV-B at ground level.
     While uncertainties exist, this dose-response relationship, when used
     with different action spectra and assumptions about the importance of
     peak versus cumulative exposure, can be utilized to estimate a range of
     values for cases.  The relationship estimates that a 1 percent change
     in ozone is likely to increase incidence by between slightly less than
     1 to 2 percent, depending on the choice of action spectrum.  The
     appropriate action spectrum is likely to be encompassed in the range of
     erythema and DNA.

7c.  Melanoma mortality is estimated at about 25'percent of all cases.  This
     result is consistent with the projections of a dose-response model of
     mortality developed by EPA/NCI.  It is estimated that a 1 percent
     change in ozone would result in between a 0.3 and a 2.0 percent change
     in CMM mortality depending on the assumptions about the appropriate
     dose and UV weighting functions used in the model.

7d.  Additional uncertainties for projecting future incidence and mortality
     of CMM in the U.S. include the lack of an adequate database describing
     variations in skin pigmentation and human sun-exposure behavior among
     different populations and estimates of how these relationships may
     change in the future.

-------
                                      8-7
INTRODUCTION

    This chapter is designed to examine the current state of knowledge with
regard to the relationship between human cutaneous malignant melanoma and solar
radiation, in particular UV-B.  Two previous reviews examining this question
came to the following somewhat contradictory conclusions:

        "The NCI [National Cancer Institute] is collecting skin
        melanoma incidence from ten locations in the United States.
        Earlier data analyses indicate that skin melanomas,  which
        predominate in Caucasians, are related to UV-B exposure.  A
        preliminary finding is that the incidence is increasing and
        that the rate of increase may be greater than expected from
        earlier surveys.  The dose-response relationship between skin
        melanoma and UV-B appears to be more complicated than that
        observed for nonmelanomas of the skin." (NAS 1979)

        "...since 1976, the case for an association between UV-B and
        melanoma has been weakened rather than strengthened by the
        results of additional clinical, pathological and
        epidemiological studies.  Furthermore (with the exception of a
        single animal), it has not been possible to use UV-B alone to
        induce melanomas in experimental animals." (NAS 1982)

These contrasting conclusions require that a closer examination be made of the
relationship between melanoma and UV-B.

    Since publication of the second NAS document in 1982, more information has
been gathered; this chapter reviews briefly both the old and the new information
in order to ascertain what can be said at the current time about the
relationship of UV-B to melanoma.  A much more detailed review of the relevant
information is presented in Appendix A.

Background on Melanoma and the Concept of Response

    Malignant melanoma is a rare tumor that arises as the result of the
neoplastic transformation of a melanocyte.   Melanocytes are pigment-producing
cells.  Exhibit 8-1 shows the relationship of the melanocyte to other cells in
the epidermis and other skin layers.

    The major function of the melanocyte is the production and distribution of
melanin to keratinocytes -- the major cell population of the skin.  Melanin is a
pigment that absorbs UV light over a broad range of wavelengths (25 nm to 1200
nm); it is the principal chromophore responsible for the difference between
black and white skin in the transmittance of UV light (shown in Exhibit 8-2).

    Melanin is a pigment that absorbs light in the broad range of 250-1200 nm;
however, its absorption increases steadily towards the shorter wavelengths
(Anderson 1983) .   It is synthesized in melanocytes from tyrosine and deposited
in a protein matrix in organelles termed melanosomes (Romsdahl and Cox 1976).
Melanosomes,  in turn, are eventually destributed to keratinocytes which are
vertically arranged in the skin.  In humans and other mammals there are two
predominant forms of melanin:  eumelanin, which is brown or black, and
phaeomelanin, which is yellow or auburn (the pigment responsible for red hair).

-------
                                          8-8


                                      EXHIBIT 8-1

                      Location of Melanocyte in  the Epidermis
                              Corniflcd calls;

                            Melanin duat

                         Malanoaomaa.

                        Karatlnocyta
     Stratum corn*um
     Qranular lay

  Karatlnocyt* layer
     Mcianoeyt**

Baa«m«nt m«mbra
                                i can layer
                          -aaaal evil layer  ™
                                           -EPIDERMIS
                                          ^PAPILLARY OERMIS
                                            RETICULAR DERMIS
                                          . SUBCUTIS

-------
                                         8-9




                                     EXHIBIT 8-2



                      Comparative Transmittance of DV Radiation
    100
u
c

     50
c
o
         (IlilJIIilllilllllillfllllllllllllllllilllllliillll (I  I
      250
                            I  1 1 I I  I I t t I  I I I  » I I II I I 1 I  I 1 I  L LI 1 1_ 1.- I t I LL t !__!_
4OO
600
80O
                            Wavelength  ( nm)
  A:  Caucasians; B:   dark blacks.


  Source:  Wan et al.  1981.

-------
                                      8-10
The melanosomes containing these two types of melanin differ structurally in
that those with phaeomelanin are round and have a protein matrix with a
"tangled" appearance, whereas eumelanin-containing melanosomes are round and
have a characteristic lamellar structure (Norlund 1981) .

    It is the degree of melanosome aggregation and their number which determine
skin color.  Melanosomes in keratinocytes from light-skinned individuals are
characteristically found as aggregates, whereas melanosomes in keratinocytes
from dark-skinned individuals, because they are larger, generally occur singly.
Extensive studies have indicated nonsignificant differences in the number of
melanocytes in various racial groups, although there are variations by anatomic
region.  The highest concentration of melanocytes is in the cheek (2,310 + 150
melanocytes/mm^) and the lowest is in the back and thigh (890 + 70/mm  and 1,000
+ 70/mm , respectively) (Briele and das Gupta 1979).

    The biology of cutaneous malignant melanoma is complex; there are several
morphologic types of tumor that may have different pathways of histogenesis;
they tend to behave differently in terms of age at appearance and characteristic
location, and yet have common elements in their progression.  The majority >85%
of melanomas are superficial spreading and nodular; these are located on the
habitually or intermittantly exposed (trunk) areas, while the melanomas on the
feet and hands, etc. are <10% in whites (Fitzpatrick 1986).  Brief information
is presented below; more detailed information is found in Appendix A.

    Terminology for the four principal classes of melanomas was proposed by
McGovern et al. (1973) and has been subsequently modified (Smith 1976; Elder et
al. 1980).  The four principal types of melanoma are given below:

        (1)  Melanoma arising in Hutchinson's melanotic freckle (HMF);
             it occurs predominantly on sites receiving the greatest
             cumulative exposure and is thought to have the strongest
             relationship to cumulative solar exposure.

        (2)  Superficial spreading melanoma (SSM); it shows the
             strongest preference for sites that are intermittently
             exposed (trunks in males, lower extremities in females).
             It also shows a stronger association with nevi-possible
             precursor lesion.

        (3)  Nodular melanoma (NM);  it has growth characteristics that
             differ from those of SSM and HMFM, although these types
             may progress to NM.

        (4)  Unclassifiable melanoma (UCM).

A fifth type of melanoma - - acral lentiginous melanoma (ALM) - - has been
distinguished with a site preference for soles, palms, and  subungual  surfaces.
It is virtually the only form observed in blacks and in people who have never
lived south of the Artie Circle (Clarke et al. 1986).  Thus in characterizing
response to dose, aggregation of all types of melanoma will tend to obfuscate
real relationships.

-------
                                      8-11
EPIDEMIOLOGIC EVIDENCE

    With the above background, the remainder of this chapter will review the
epidemiologic and experimental evidence relevant to assessing the role of solar
radiation and UV-B in the development of CMM.  Since the National Academy of
Sciences published its latest document (NAS 1983) evaluating the impact of UV-B
on skin cancer, several major epidemiologic studies have been completed by
Green, Holman, Elwood, Dubin, and their colleagues.  These authors' results have
provided valuable insights into the potential relationships between sunlight and
melanoma.  At the same time, these results have illustrated the extremely
complex nature of CMM.  In this section,  the major findings presented in these
recent studies will be briefly reviewed.   Then, because of the extensive
quantity of epidemiologic information available, five areas of evidence--time
trends, anatomical site distribution, exposure factors, host factors, and
miscellaneous factors--will each be reviewed.  This discussion is followed by a
brief review of the experimental evidence and a discussion of possible
dose-response relationships.

Recent Epidemiologic Studies

    The epidemiologic data described below were based on information obtained
from four different parts of the world.  Green and her colleagues examined
epidemiologic data from Queensland in Northeast Australia, while Holman and his
colleagues focused on data from Western Australia.   Elwood and his colleagues
analyzed data from Western Canada.  Data obtained from three New York City
hospitals were evaluated by Dubin and his colleagues.

    Green and Colleagues - Queensland Melanoma Data

    The Queensland patient series consists of 633 cases of first primary CMM
diagnosed at the 24 Queensland pathology laboratories between 1 July 1979 and 30
June 1980, directly age-standardized to 1979 Australian Bureau of Statistics
population estimates.  Patient information included age, sex,  residential
address at time of diagnosis, and histologically classified lesion according to
the Clark method.  In a series of seven publications to date,  Green and her
colleagues have examined several subsets  of the Queensland patient series.   The
four major studies present case-control comparisons for randomly selected
patient subsets and age-,  sex-,  and residence- matched controls randomly
selected from the Queensland electoral roll.  Information obtained on each
subject included lifetime sun exposure (occupational or recreational),  acute and
chronic response to sun exposure,  episodes of severe sunburn,  complexion (skin,
eye, and hair color), number of nevi (2 mm or more  in diameter) on the left arm,
family melanoma history, social class (based on occupation),  ethnic origin,  and
nonmelanotic facial skin cancers.

    In the first study of 183 matched case-control  pairs (Green 1984),  the risk
of melanoma was observed to increase as cumulative  hours of sun exposure
throughout life increased,  even after adjusting for the effects of exact age,
presence of nevi on the arms, hair color,  and sunburn propensity in a
multivariate model.   For less than 2,000  cumulative hours of sun exposure,  the
relative risk (RR) was 1.0.   For 2,000 to 50,000 hours, the RR was 3.2 (95%  C.I.
0.9-12.4) and for more than 50,000 hours,  the RR was 5.3 (95%  C.I.  0.9-30.8).   A
subsequent analysis of the same 183 matched case-control pairs (Green et al.
1985) indicated a strong association between the presence of any pigmented nevi

-------
                                      8-12
(check) on the arms and CMM.  The RR,  after adjusting for hair color,  propensity
to sunburn, and lifetime sun exposure, was 30.1.  The authors observed that,
among all the phenotypic characteristics evaluated,  the major CMM determinants
were nevi on the arms, propensity to sunburn upon acute exposure,  and hair1
color.

    The third published examination of the 183 matched case-control pairs
(Green, et al.  1985b) revealed an association between multiple sunburns and CMM.
After controlling for age and the number of nevi on the arms, the relative risk
of developing CMM was 1.5 for those with 2-5 sunburns per lifetime and 2.4 for 6
or more sunburns.  The dose-response relationship of increasing CMM risk with
increasing number of sunburns per lifetime was statistically significant (p <
0.05).  When the presence of skin cancers, migrant status, and social class were
taken into account, the relative risks remained essentially unchanged.

    In the fourth case-control study of 232 matched case-control pairs (Green
and O'Rourke 1985), increased risks were associated with the presence of facial
actinic tumors (RR =3.6 after adjusting for age and nevi).  Risks also
increased with increasing lifetime sun exposure.  After adjusting for age, nevi,
hair color, and sunburn propensity, the relative risks were 3.2 for 2,000-50,000
hours cumulative exposure and 5.3 for 50,000 or more hours exposure.

    Based on the results of these studies, Green and her colleagues have made
several observations regarding the potential associations between CMM and
sunlight.  They have indicated that the data

         o    "strongly suggest that melanoma does have an association
              with high doses of solar UV radiation" (Green 1984), and

         o    in conjunction with "similar findings regarding the risk
              of melanoma in sun sensitive persons... provide
              excellent circumstantial evidence that sun exposure has
              a causal association with disease [CMM]" (Green
              et al. 1985).

    Holman and Colleagues - Western Australia Melanoma Data

    Holman and his colleagues have presented several important case-control
comparisons based on 511 melanoma patients presenting from 1 January  1980 to
5 November 1981 in Western Australia and  511 age-, sex-, and residence-matched
controls.  In addition to determining histogenic tumor type, level of  invasion,
and tumor thickness, information corrected on each case and control by interview
consisted of skin, hair, and eye color, number  of palpable nevi on the arms,
history of sun exposure, hormone use, diet, measurements of weight, height,
hairiness, and extent of actinic damage in skin (based on cutaneous
microtopographs), acute and chronic skin  reactions to sunlight, ethnicity of
grandparents, family history of melanoma  or xeroderma pigmentosum, history of
mole  excisions, and  treatment of nonskin  cancers.  Skin color was measured at
the dorsum of the left hand  (continuous sun exposure), tip of left  shoulder
(intermittently exposed), and left upper  inner  arm  (not usually exposed) and was
graded according to  ranges  of reflectance values  (%).

    In one analysis  of this case-control  data  (Holman and Armstrong 1984), the
strongest  risk factor identified was  the  number of palpable benign nevi  on the

-------
                                      8-13
arms, with crude RRs of 2.0 for 1-4 nevi, 4.0 for 5-9 nevi, and 11.3 for 10 or
more nevi (p < 0.0001).  The authors noted that nevi are probably important
"either as an early stage in the pathogenesis of non-HMF melanomas or as
indicators of increased risk."  Inability to tan was the most important
pigmentary trait associated with the risk of CMM, followed by susceptibility to
sunburn, and then hair color.  After taking these three factors into account,
skin and eye color had no additional effects, leading the authors to conclude
"that the ability to tan quickly in response to sunlight is of prime importance
in reducing risk of skin cancers and is more important than the base-line skin
color."

    Another evaluation of the case-control data (Holman and Armstrong 1984a)
focused on indicators of total accumulated exposure to the sun and separate
histogenic tumor types.  Two important factors were positively associated with
CMM risk:  duration of residence in Australia of migrants and mean annual hours
of bright sunlight (averaged over a lifetime) among native-born Australians.
For all melanomas combined and SSM alone among migrants, age at arrival was a
better predictor of melanoma risk than was duration of residence.  The authors
suggested that "it is possible that exposure to sunlight in childhood is a
factor in the production of benign nevi, which have their strongest relationship
with SSM, probably as precursor lesions."  Among native Australians, the risk of
developing CMM significantly increased with increasing mean annual hours of
bright sunlight for all melanomas combined (p = 0.003) and for SSM alone (p =
0.02).  The gradient of increasing risk was steepest, however, for HMFM.

    Holman and Armstrong (1984a) also observed that persons with a history of
nonmelanotic skin cancer had a threefold higher risk of developing CMM
(p < 0.00001) and SSM,  and a fivefold higher risk of developing HMF
(p = 0.001).  After controlling for the effects of chronic and acute skin
reactions to sunlight,  hair color,  and number of European, African,  and Asian
grandparents, the risks associated with nonmelanotic skin cancers dropped
slightly, suggesting that the association with nonmelanotic skin cancer was
explained only partly by constitutional factors.   Based on the results of this
study, Holman and Armstrong (1984a)  concluded that HMF "may be related to the
total accumulated dose of sunlight received on exposed body sites."   They also
suggested that their results generally supported the hypothesis that "an
individual's maximum potential to develop SSM would be fixed by the  number of
initiated nevus cells induced by UV radiation and other agents in childhood and
young adulthood."  Finally,  they concluded that "a role of sunlight  in the
causation of NM is supported by these results."

    A third study of the case-control data focused on the relationship between
CMM and individual sunlight-exposure habits (Holman et al.  1986).  The authors
identified the presence,  and absence,  of several  associations:

         o    Increased risks of SMM were associated with low total
              outdoor exposure in early adulthood and frequent
              participation in boating and fishing.   SSM of the trunk
              was related to the frequency of sunbathing at 15-24
              years of  age and to exposure of the trunk while working
              outdoors.

-------
                                     8-14
         o    The  risks  of  developing  CMM  (except for SSM) were higher
              in persons if the body site  was sometimes exposed rather
              than usually  exposed or  usually covered.

         o    In females, the  risks of developing CMM on  the  trunk
              were significantly  elevated  (odds ratio = 12.97) in
              those who  work a bikini  or bathed nude at 15-24 years  of
              age  compared  to  those who wore a more conservative
              one-piece  bathing suit.

         o    After controlling for confounding factors (e.g., hair
              color,  skin reaction to  sunlight, ethnic origin, age at
              arrival in Australia), only  HMF showed a relationship  to
              the  highest severity of  past sunburn.  No relationship
              was  observed  between childhood sunburn or sunburn in
              early adulthood  and any  histologic type of  melanoma.

         o    Little support for  the hypothesis that CMM  may  be
              related to occasional bursts of recreational sun
              exposure during  early adult  life was obtained when
              recreational  sun exposure was expressed as  a proportion
              of total outdoor exposure  (a proportion which had been
              considered, a priori, to be  an index of intermittent
              sunlight exposure).

    Holman et al.  (1986) noted that "some  but not all" of their results
supported the hypothesis that  intermittent exposure plays an  important  role in
the etiology of SSM.   They  concluded that  their results suggested that  "nevi
and occasional or  recreational sun  exposure interact to produce an effect on the
rate of SSM greater than the addition  of the two independent  effects but...  less
than expected [based on] multiplication  of effects."

    Holman and his colleagues  published  one more important study since  1982 on
the relationship between CMM and  exogenous sex hormones and  reproductive factors
(Holman et al. 1984d).  The case-control study was based  on  the 276  female CMM
patients identified in the  West Australia  Lions Melanoma  Research Project from
1980 to 1981 and 276 age- and  residence-matched female controls.  No consistent
evidence was observed of a  relationship  between incidence rates of different
histogenic tumor types and  age at menarche, duration of menstral  life,  degree of
obesity, number of pregnancies of more than 20 weeks'  duration, or use  of oral
contraceptives.  However, borderline evidence was  shown of an association
between SSM and duration of estrogen use.

    Elwood and colleagues - Western Canada Melanoma  Study

    This major case-control study was  based on  data  from  four western Canadian
provinces (British Columbia, Alberta,  Saskatchewan,  and Manitoba).   The cases
consist of 595 histologically  confirmed primary  CMM  patients registered in each
province's cancer registry from  1 April 1979 to  31 March  1981 (excluding acral
lentiginous melanoma and lentigo  maligna).  Each  patient  was matched by sex and
age with a control randomly selected  from  medical  insurance  plan  subscriber
lists.   Information obtained by  interview  from each  case  and control included
pigmentation, skin freckling in childhood, sensitivity to sunlight,  tanning and
sunbathing behavior  (both as an adult  and  as a child),  residence,  occupational

-------
                                       1-15
 history,  recreational  activities with  specific  reference  to  sunlight  exposure,
 medical history,  chronic  drug use,  family history,  diet,  smoking  and  alcohol
 consumption.   In  addition,  for women participants,  information was  gathered on
 reproductive history and  the use of oral contraceptives and  menopausal
 estrogents.  Skin and  hair  color were  determined by direct comparison with
 prosthesis and wigmaker samples made specifically  for  the project.  Eye  color
 was  recorded based on  direct observation.

     In the first  published  analysis of this  data (Elwood  et  al. 1984), CMM risks
 were observed  to  be significantly  increased  among  individuals with  increasingly
 light hair and skin color (pigmentation factors),  many freckles in  adolescence,
 and  a tendency to burn and  rarely  tan.  The  associations  with these risk factors
 did  not substantially  change after  adjustment for  ethnic  origin or  UV exposure
 during occupational or recreational activities.  Elwood et al. (1984) concluded
 that their results "are consistent  with the  hypothesis that  SSM and NM [which
 were diagnosed in the  majority of  the  cases] develop from nevi and  suggest that
 SSM  and NM have similar relations  to characteristics of pigmentation."   After
 observing that the associations between CMM  and a  history of severe or frequent
 sunburns  became weaker and  non-significant by adjusting for  pigmentation
 factors,  the authors suggested "that rather  than the occurrence of  sunburn
 itself increasing the  risk  of melanoma, the  risk is due to the characteristics
 of pigmentation associated  with poor sun tolerance."

     In a  subsequent analysis of the case-control data, which focused  on
 intermittent and  chronic  sun exposure  (Elwood et al. 1985a), factors
 significantly  associated  with CMM were  vacation and recreational  activities
 during which a bathing suit or very light clothing  is  worn and increasing number
 of sunny  vacations  per decade (after adjusting  for skin and  hair  color,
 freckling in adolescence, and ethnic origin).   The risk of developing CMM was
 also  significantly elevated among individuals with mild occupational  sun
 exposure  (roughly 8 hours/week) but not among those with  higher occupational sun
 exposures (over 8  hours/week).   The authors  concluded  that their  results
 "suggest  that  intermittent  and continuous exposures have  different  effects 0and!
 are  consistent with the hypothesis  that intermittent sun  exposure raises the
 risk  of melanoma,   while long-term constant exposure may have no effect or may
 reduce risk."

    The third published analysis to date focused on the effects of  sunburn and
 suntan to CMM  risk  (Elwood  et al.  1985b).   The  results showed that  a mild suntan
 in winter and  summer and  a history of severe vacation  sunburns are both
 independently associated with elevated CMM risks.   However,   neither factor is a
 significant predictor  of  risk when usual reaction to sun  exposure (e.g.,  tan
with no burn or burn,   rarely tan)  is concurrently evaluated.   In a multivariate
 analysis of history of vacation sunburns,  usual degree of suntan,  usual  reaction
 to sun,  and a host  factor score (representing skin and hair  color and freckling
 in childhood),  the usual  reaction to sun variable and host factor score were the
only statistically  significant  predictors of CMM risk.   Individuals with a
tendency to burn easily were at nine times greater risk of developing CMM than
those with a tendency  to  tan and not burn.   Individuals with light hair and skin
color and many freckles in adolescence were at  37  times greater risk than those
with dark hair and skin color and few freckles  in adolescence.   The results
indicate that the  tendency to burn easily and tan poorly  is  more  strongly
associated with CMM than  is  the history of sunburn or suntan.  "The
implication," state the authors,  "is the factor contributing to the risk  of

-------
                                     8-16
melanoma is the individual's tendency to burn rather than the history of having
had burns."

    Dubin Colleagues - New York University Medical Center Data

    This large case-control study was based on 1,103 primary CMM patients seen
at the New York University Medical Center from 1972 to 1982 and 585 non-matched
controls (Dubin et al. 1986).   In a multiple logistic regression of the data,
Dubin et al. observed elevated risks for those with no ability to tan (relative
to average ability), no tendency to burn (relative to tendency for painful
burn), a history of freckling, red hair color, blue eyes, and a previous medical
history of solar keratosis.

Time Trends

    As indicated in Exhibit 8-3, there have been sharp increases in incidence
and mortality due to CMM reported in Caucasian populations worldwide (Magnus
1982; Roush et al. 1985b; Osterlind and Jensen 1986).  Based on observational
and analytical evidence, most experts agree that these trends in CMM incidence
and mortality are genuine, and not due to increases in the registration and
diagnosis of the disease (Magnus 1975; Pakkenen 1977; Malec et al.  1977).  It
has been estimated that in 1987 there will be 25,800 cases and 5,800 deaths from
CMM (Kopf 1987).  At the current rate of increase, the lifetime risk of CMM for
individuals born in the United States in the year 2000 has been predicted to be
1 in 150 (Kopf et al.  1984).

    Birth cohort analyses of CMM incidence rates across six nations show
consistent increases in CMM for cohorts born between 1910 and 1930, whereas
later birth cohorts (1940-1950) that show these increases do so to a lesser
extent (Muir and Nectoux 1982).  Most authors that have conducted cohort
analyses of CMM incidence and mortality rates conclude that virtually all
secular increases in CMM are due to cohort effects (Magnus 1981a 1981b; Houghton
et al. 1980; Muir and Nectoux 1982; Holman et al. 1980a; Cooke et al. 1983).

    In most countries, the first signs of increasing rates are seen in cohorts
born around 1900, though increases in cohorts born as early as 1865 are observed
in Australia and New Zealand (Holman et al. 1980a; Cooke et al.  1983).  In
Norway and several other countries, there is a slight tendency for a slowing of
increase in incidence in cohorts born around and after 1930 (Magnus 1981a).
Stabilization of mortality rates is also occurring in cohorts born 1925-1939 and
later in countries such as Australia, New Zealand, England, Wales,  and Finland
(Lee and Carter 1970;  Teppo 1978; Holman et al. 1980a; Cooke et al. 1983).

    Age-adjusted incidence rates of CMM of the head, neck, and face for birth
cohorts born after approximately 1900 have not differed markedly; however,
age-adjusted incidence rates of CMM of the lower extremities among females and
the trunk among males have increased in successively younger birth cohorts born
during the first half of this century (Magnus 1981a; Houghton et al. 1980; Cooke
et al. 1984; Stevens and Moolgavkar 1984).

    On the basis of the Connecticut Tumor Registry data, modeling  of the CMM
incidence rates observed in subsequent birth cohorts indicated that the
incidence rate of CMM in the 1955 birth cohort will rival those for colon

-------
                                                    8-17
                                               EXHIBIT  8-3
                                Increases IB Incidence and Mortality Rates Frcai
                                   Malignant Melanoma in Different Countries
First Period
of Observation

Incidence

Incidence

Mortality

Mortality

Mortality
Mortality

Mortality

Mortality

Incidence

Incidence

Mortality

Country Sex
New York State M
F
Norway M
F
Norway M
F
Canada M
F
United Kingdom Both
Australia M
F
Denmark M
F
Sweden M
F
Connecticut M
F
U.S.A. WM
WF
U.S.A. WM
WF
Time
1941-1943
1941-1943
1955
1955
1956-1960
1956-1960
1951-1955
1951-1955
1950
1931-1940
1931-1940
1956-1960
1956-1960
1956-1960
1956-1960
1935-1939
1935-1939
1974
1974
1950
1950
Rate
per ID6
1.2
1.8
1.8
2.6
1.6
1.3
0.7
0.6
0.5
1.0
0.8
1.6
1.6
1.7
1.1
1.1
0.9
6.7
6.0
1.0
0.8
Second Period
of Observation
Time
1967
1967
1970
1970
1966-1970
1966-1970
1966-1970
1966-1970
1967
1961-1970
1961-1970
1966-1969
1966-1969
1966-1968
1966-1968
1975-1979
1975-1979
1983
1983
1977
1977
Rate
per 106
3.4
2.9
6.3
6.8
2.7
1.8
1.4
1.2
1.0
3.6
2.5
2.4
2.1
2.1
1.5
8.2
6.8
9.6
8.3
2.6
1.6
Percent
Increase
176
65
264
195
69
36
93
107
100
267
227
49
32
.0
40
645
656
43
38
160
200
Total
Number of
Years
25
25
15
15
10
10
15
15
16
30
30
10
10
9
9
40
40
10
10
27
27
Average
Annual
Percent
Increase*
7.0
2.6
17.6
13.0
6.9
3.6
6.2
7.1
6.3
8.9
7.6
4.9
3.2
3.3
4.4
16.1
16.4
4.3
3.8
5.9
7.4
*  Computed by  dividing total percent increase in incidence  or mortality rate by the number of years between
   the first and  second periods of observations.   The figures in this column do not represent compounded
   annual growth  rates.
Adapted from:   Elwood  and Lee (1975); Sondik et al.  (1985);  Heston et al. 1985.

-------
                                      8-18
cancer, currently the third most common cancer in Connecticut (Roush et al.
1985a).

    Those anatomical sites that show the greatest amount of increase are not the
sites that receive the greatest amount of cumulative solar exposure.  This
finding tends to lead to a conclusion that cumulative solar exposure is not
causally associated with melanoma; as discussed in more detail below, cutaneous
melanoma may be related to some other measure of exposure to solar radiation.

Anatomical Site Distribution

    Cutaneous malignant melanoma has a unique anatomical distribution that has
been the subject of numerous epidemiologic investigations.  Research efforts
have focused on the site-specific trends in CMM incidence related to sex, age,
race, histogenic type, birth cohort, and season.  The most pronounced
differences have been associated with gender, race, and birth cohort.

    The overall site distribution is presented in Exhibit 8-4.  There is wide
variability among the studies presented in the exhibit, making it difficult to
generalize about these data.  However, as a rule, among white populations, the
upper extremities have the lowest overall proportion of melanomas, while the
lower extremities have the highest proportion.  The variability is most probably
due to differences among the study populations,  such as their sex and racial
distribution.

    Gender is one of the factors associated with the most pronounced differences
in CMM site distribution.  Exhibit 8-5 lists data from 14 studies that reported
CMM site distribution by sex.  Data in the table indicate that most of the
studies observed higher incidences of CMM on the lower extremities among females
and on the trunk among males than on other parts of the body.  Most authors have
concluded that the observed differences in site distribution by gender are not
incompatible with the role of sunlight as a major etiologic factor  (Pathak et
al. 1982, Crombie 1981; Hinds and Kolonel 1980).

    Race and gender also are associated with pronounced effects in  site
distribution of CMM.  The predominant difference is a higher percentage of CMM
emphasize that increase occurring on the feet and, in some cases, on the hands
of darker-skinned ethnic groups.  These groups have a much lower incidence than
the light-skinned ethnic groups, so that the higher percentages on  feet may not
represent a higher incidence.  Trauma has been suggested as a possible reason
for the increased percentage of melanomas occurring on feet (Lewis  1967; Hinds
1979).  Although in a recent review of the subject, Briggs (1984) concluded that
there  is no unequivocal evidence for a role of trauma in the vast majority of
melanomas.  One group has concluded that sunlight is not an important risk
factor for melanoma (Hinds 1979; Hinds and Kolonel 1980) for any site among
non-Caucasians.

    There appear to be site-dependent differences in melanoma incidence by age,
with melanomas of the face showing one pattern and melanoma of trunks and  lower
extremities, another.  In several instances the differences have been identified
as being between continuously exposed sites  (e.g., head) and  intermittently
exposed sites  (e.g.,  trunk and lower extremities).  As a general rule the
pattern observed for  the exposed sites is characteristic of that observed  for
many other cancers  -- a slow increase with age, up to the age of 40  or 50,

-------
                                                  8-19
                                                EXHIBIT 8-4

                        Anatomic Site Distribution of Cutaneous Malignant Melanoma
                                       (Percentage of Total Tunrs)
Extremities
Location Years Sample Size Head/Neck Trunk Upper Lower Source
United States3 1978-1981 4,864
(Caucasian)
Texas 1944-1966 911
Texas 1954-1970 510
Alabama 1955-1980 537
New Mexico 1966-1977 403
New South Wales 1955-1980 1,110
Queensland 1977 690
New Zealand 1963-1981 24
(Maori/Polynesian)
Israel 1960-1972 966
Norway 1955-1970 2,541
Finland 1953-1973 2,501
Japan 1961-1982 546
Hong Kong 1964-1982 43
Uganda 1963-1966 152
20 35 23 22 Scotto and Fears 1986

22 25 19 13 MacDonald 1976
25 16 21 37 Smith 1976
27 28 19 23 Balch et al. 1982
27 29 19 25 Pathak et al. 1982
14 37 14 33 Balch et al. 1982
21 34 20 24 Little et al. 1980
13 13 0 54 Moss 1984

16 25 21 38 Anaise et al. 1978
22b 43C 8 18 Magnus 1973
19 37 10 26 Teppo et al. 1978
15 20 13 46 Takahashi 1983
7 7 21d 63e Collins 1984
8 llf 5 72s Kiryabwire et al. 1968
  Based on data from Seattle,  Detroit,  Iowa, Utah, San Francisco/Oakland, Atlanta, and New Mexico.
b Face.
c Neck/truck.
  17 percent of the total  on the hands.
e 56 percent of the total  on the feet.
  "Skin" and genitals.
8 Feet and legs.

-------
                         8-20
                    EXHIBIT  8-5
Anatomic Site Distribution of Cutaneous Malignant Melanoma by Gender
              (Percentage of total tutors by gender*)
Location
(Ethnic Group) Years Sample Size
Sex
Head/Neck
Trunk
Predominantly Caucasian Study
United States 1980 4,545

b
United States 1978-1981 4,864


North America
(Caucasian)

Europe
(Caucasian)
Texas (White) 1944-1966 1,252

Hawaii (Caucasian) 1960-1977 262

New Mexico (Anglos) 1966-1977 403

Israel 1960-1972 966

Scotland 1961-1976 477

Finland 1953-1973 2,501
Denmark 1943-1957 1,204

Queensland 1977 713

Texas 1954-1970 510

F
M

F
M

F
M

F
M
F
M
F
M
F
M
F
M
F
M
F
M
F
M
F
M
F
M
15
23

16
24

16
27

16
23
17
23
9
24
22
33
16
16
25
32
18
19
24
32
19
23
17
34
31
53

25
45

29
44

23
42
18
25
28
37
20
41
19
32
14
27
28
48
19
25
21
45
10
23
Extremities
Upper Lower
Source
Populations
19
13

26
21

18
17

13
13
25
20
28
24
21
16
15
21
14
11
11
9
14
9
22
17
24
18
35
11

33
11

36
12

47
23
28
11
35
15
37
10
50
31
47
30
36
17
33
22
35
13
50
24
Balch et al. 1984


Scotto and Fears
(in press) 1986
c
Crombie 1981

c
Crombie 1981
MacDonald 1976

Hinds and Kolonel 1980

Pathak et al. 1982

Anaise et al. 1978

Pondes et al. 1981

Teppo et al. 1978
Clemmesen 1965
(cited in Lee 1982)
Little et al. 1980

Smith 1976


-------
                                                          8-21
                                                      EXHIBIT 8-5 (Continued)
                          Anatomic Site Distribution of Cutaneous Malignant HeLancBa by Gender
                                         (Percentage of total tnors by gender*)
   Location
(Ethnic  Group)
Years
                                         Extremities
Sample Size   Sex   Head/Neck   Trunk   Upper   Lower
                                                                          Source
                                                Mixed Study Populations
Israel (Mixed Race)

Japan

Texas (Spanish
surname)
Texas (Non-White)

New Mexico
(Hispanics)
Hawaii
(Non -Caucasian)
Hawaii
(Non-Caucasian)
Uganda

1961-1967

1961-1982

1944-1966

1944-1946

1969-1977

1960-1970

1960-1977

1963-1966

368 F
M
546 F
M
206 F
M
30 F
M
35 F
M
66 F
M
64 F
M
152 F
M
15
17
20
13
22
18
6
31
23
23
26
16
26
16
6
10
21
35
19
23
16
22
6
—
36
46
17
19
17
20
llh
llh
15
5
18
11
18
11
18
8
32
23
13
19
13
20
4
5
42
25
42d
51e
30
30
53
54
9
8
44
47
44f
44s
781
74J
Movshovitz and Modan
1973
Takahashi 1983

MacDonald 1976

MacDonald 1976

Pathak et al. 1982

Hinds and Kolonel 1980

Hinds 1979

Lewis 1967

       Percentages may not total 100 percent because of rounding errors and exclusion of "other sites" of MM.
       Based on data  from Seattle, Detroit, Iowa, Utah, San Francisco/Oakland, Atlanta, and New Mexico.
     c Trunk includes scrotum and "unspecified" melanomas.
       29 percent of  the total were on the feet.
     e 39 percent of  the total were on the feet.
       22 percent of  the total were on the feet.
     8 42 percent of  the total were on the feet.
       "
        Skin"  and  genitals.
       63 percent  of the total were on the feet.
       64 percent  of the total were on the feet.

-------
                                      8-22
followed by a rapid increase with age thereafter (Magnus 1981b;  Teppo 1982;
Holman et al. 1980b).   For the intermittently exposed sites,  the pattern is
quite different a relatively steep increase that peaks between 40 and 60
(depending on the data set) and levels off or decreases thereafter (Magnus
1981a,b; Teppo 1982; Holman et al.  1980b).

    There are also differences in the anatomic distribution of melanomas
associated with different histologic (histogenetic) types of CMM.  HMFM and its
precursor lesion, HMF, are observed predominantly on the face and other exposed
sites, whereas SSM was most commonly observed on less exposed sites (trunk and
lower extremities) (Holman et al. 1980b, Smith 1976, Adler and Gaeta 1979,
Pondes et al. 1981).

    Very few studies have examined the relationship of geographic area to
differences in a anatomic site distribution.  In Norway, Magnus (1973) noted
that the only site lacking a definite north/south gradient was the foot, and a
comparison of Alabama patients to Australian patients revealed that whereas
melanomas of the trunk were the most common in both sets of patients, melanomas
of the lower extremities were more common in Australians than in the Alabama
patients, and head and neck tumors showed the reverse trend (Balch et al. 1982).

    A final important parameter that has been associated with differences in
site distribution of CMM is the temporal pattern of exposure.  This is a complex
parameter that is, as indicated earlier, difficult to separate from the observed
age effects, but it is fairly clear that there are differences in the anatomic
distribution of melanomas, which are associated with the year of diagnosis
and/or birth cohort.  It is difficult to generalize, however, the conclusion
drawn by Boyle et al.  (1983) that there were cohort effects of melanoma
incidence that differed by site and that the site effects differed by sex.  The
overall impression from the studies that looked at these cohort effects is, as
well, that birth cohort-site differences are greater for sites that are
intermittently exposed (e.g., trunk and lower extremities) than for those that
are always exposed (e.g., face).

Exposure Factors

    A number of studies have examined how various indicators or measures of
exposure are related to melanoma incidence or mortality.  This section reviews
information  from ecological studies that have looked at the correlation of
melanoma incidence or mortality on the one hand, and geographic location and
migration on the other.  Studies that have examined the relationship between
type of exposure  (e.g., early, intermittent) and melanoma incidence are also
discussed.

    Geographic Location

    Within predominantly Caucasian nations, most of the ecological studies of
melanoma and latitude show increasing melanoma  incidence and/or melanoma
mortality with decreasing latitude, leading to  the hypothesis that melanoma is
associated with ultraviolet radiation because of the strong correlation between
UV and  latitude.  As a general rule, those  studies  that failed to find  this
association  did not adequately account  for pigmentation differences  (Crombie
1979b)  or had other serious methodological  flaws (poor measurement of UV-B)
(Baker-Blocker 1980).

-------
                                      8-23
    In general, melanoma rates  (incidence and/or mortality) were found to be
higher in areas closer to the equator, in coastal rather than  inland areas, and
in urban rather than rural areas within various nations (Holman and Armstrong
1984; Green and Siskind 1983; Magnus 1981a, 1981b).

    Migrant Behavior

    In an Australian study (Holman and Armstrong 1984b), age at arrival in
Australia was more important than duration of residence with respect to the risk
of SSM, with arrival before age 10 incurring a risk approaching or exceeding the
risk of SSM for those born in Australia.  The odds ratio decreased when the age
at arrival was 10-14 years and stabilized at a significantly lower level for
those arriving at or after the age of 15.

    Immigrants moving to sunnier climates in which the native  CMM incidence
rates exceed those of the immigrant's country of origin initially tend to have
lower CMM risks than the native population.  These risks increase, however, with
increasing duration of residence or earlier age of arrival in  the adopted
homeland (Holman and Armstrong 1984b).

    Type of Sun Exposure

    Early Exposures.  CMM risk is associated with childhood sunburn (Green,
et al. 1985b, Elwood et al. 1985b) but this association appears only to reflect
an individual's pigmentary characteristics as related to poor  sun tolerance.
One study that evaluated the risk associated with outdoor work before college
found a significantly elevated odds ratio for those who had worked outdoors
compared with those who had not (Paffenbarger et al.  1978).

    Intermittent ("Summer Sun") Exposure.   One commonly examined surrogate for
intermittent exposure has been a parameter related to a history of sunny
vacations (Eklund and Malec 1978;  Klepp and Magnus 1979; Adam et al.  1981).
Different studies have used slightly different measures of this type of exposure
behavior (e.g., one study evaluated those who "tanned themselves while on
holiday abroad" (Adam et al.  1981),  another evaluated those with a history of
long (30 days or more),  sunny vacations as a child (Lew et al.  1983).   In
general,  however,  there was a positive association between CMM risk and
increasing number of "sunny vactions" taken.

    The strongest evidence for an association between a measure of sunny
vacation exposure comes from a case-control study conducted in Western Canada
(Elwood et al.  1985a,  1985b).   Analysis of data from this  study of 595 melanoma
case-control pairs indicated that there was a significant  (p < 0.001)  trend of-
increasing risk of CMM with increasing numbers of sunny vacations per decade,
with a relative risk of 1.8 associated with four or more sunny vacations per
decade.   In addition,  this relative risk remained significant  (RR =1.7,  95%
C.I.  1.2-2.3)  even after controlling for significant host  factors (hair color,
skin color,  history of freckles) and ethnic origin.

    At least one study that evaluated a different measure  of intermittent
exposure  found no  significant association between intermittent exposure and an
increased incidence of any histologic type of CMM (Holman  et al.  1986).   This
study of  507 pairs (matched according to sex,  age,  and residence)  used as its
surrogate for intermittent exposure  a variable based on the ratio of

-------
                                     8-24
recreational outdoor sun exposure in summer to total outdoor sun exposure in
summer.

    In the same study (Holman et al.  1986), an increased SSM risk for some
summer sun activities at early ages (from history or estimated)  was observed for
boating, fishing,  and female sunbathing (on trunk).   This study  also found some
increased risks for summer sun exposure by clothing habits.   The primary CMM
site was more frequently "sometimes exposed to sun"  than usually exposed or
usually covered, except for SSM, which showed increasing risk with increasing
exposure.  The risk of developing CMM of the trunk (especially SSM) was found to
be greater for women whose bathing suits covered less of their bodies between 15
and 24 years of age.

    A seasonal pattern with a summertime peak was found for CMM  incidence rates
(female in U.S. [Scotto and Nam 1980], both sexes in Hawaii [Hinds et al.
1981]).  This seasonal pattern  may be related to greater awareness of skin
changes in summer months or could indicate promotional effects of UV-B
exposures.

    Sunburn.  A number of case-control studies that have examined the
relationship of CMM to history of sunburn have found statistically significant
associations.  (Beral et al. 1983, Elwood et al.  1984) Most of these studies did
not, however, adjust for pigmentary factors.  One study that controlled for
pigmentary factors found a significant association (Lew et al. 1983); however,
in this study, biased selection of controls was likely, leaving  this association
in doubt.  The conclusion drawn from studies that appeared to be subject to the
fewest biases was that the presence or absence of factors associated with a
tendency to sunburn, rather than the sunburn itself,  determine the CMM risk.  In
a very recent report, Armstrong and his colleagues (1986) observed a significant
correlation between number of burns received up to the age of ten and the number
of nevi.  Thus, the relationship of CMM to sunburn could be via  an intermediate
risk factor - number of nevi.

    A study from Western Australia (Holman, et al. 1980b) found  significantly
increased total CMM risk, SSM risk, and HMFM risk with increased annual hours of
bright sunlight at residence, with increasing actinic skin damage, and with
previous nonmelanoma skin cancer.

    Assessment of individual exposure through questionnaires.  In Western
Canada, an analysis of total sun exposure showed some increased  risks in higher
exposure groups compared with the lowest exposure group but none were
statistically significant nor was there a significant trend of increasing risk
(Holman et al. 1986).

    In Queensland, Australia, elevated CMM risks were associated with increasing
estimated total hours of sun exposure at age 10 years and older  after adjustment
for exact age, presence of nevi on arms, hair color,  and sunburn propensity.
The confidence intervals for intermediate and higher levels of exposure include
unity.

    Indirect measures of cumulative sun exposure (i.e., assessment of numbers of
hours exposed) show correlations with melanoma risk that were significant,
borderline significant, and not significant.  Several studies have evaluated a
direct measure of received dose (i.e., actinic skin damage), which is also a

-------
                                      8-25
measure of the skin's responsiveness to  insult.  In these studies, melanoma risk
increased significantly with the degree  of actinic skin damage.

    Exposure During Sunspot Activity.  Most studies that examined the
relationship between CMM incidence rates and sunspot cycles found high
correlations (Houghton et al. 1978; Wigle 1978).  Different studies have
observed differences in the lag time between the period of sunspot activity and
changes in CMM incidence rates.

    Most of the information presented above tends to support a relationship
between melanoma and some function of increased exposure to solar radiation.
What the exact function of sunlight exposure may be is unclear.  There is some
indication that early (childhood) exposure may be important, and other
information implicates intensive exposure such as that which occurs during
summertime and times of sunspot activity; however, no single kind of exposure
clearly stands out.  None of these studies considered UV-B, which varies
significantly with time of day.  Thus, by using sun-exposed hours as a dose
variable, such studies may be inaccurately assessing dose, especially when they
compare across occupations whose sun-exposure behavior may vary significantly.

Host Factors

    There is clear evidence that certain host factors are associated with the
risk of developing CMM.  It is possible that the factors described below reflect
the capability to modify the dose of solar radiation potentially received by a
target cell or modify the ability of the cell to respond to insult.

    Epidemiologic studies have shown that Caucasian populations have much higher
rates of CMM incidence and mortality than other racial groups.  Crombie (1979a)
analyzed data from the International Agency on Cancer dataset on cancer
incidence on five continents and found a statistically significant three-fold
increase in CMM incidence in whites over that of non-whites.  In South Africa,
CMM was approximately six times more frequent in whites than in blacks.  Broken
down by sex, the ratio of white to black CMM incidence was greater for males
(13:1) than for females (4:1) (Rippey and Rippey 1984).  Black/white differences
in the United States population were even greater, although this may be due to
differences in the quality of cancer registry data between South Africa and the
United States.   Based on 1983 SEER data, white to black ratios were 19:1 and 8:1
in males and females,  respectively (Sondik et al.  1985).  In addition, whites in
New Zealand experience much higher incidence rates than New Zealand Maoris and
Polynesians.  Likewise,  American Indians experience much lower rates of CMM than
American whites.

    Within white populations, rates of CMM differ according to country of
origin.  CMM incidence rates for Hispanic whites in New Mexico, for example, are
much lower than those for non-Hispanic whites;  individuals from the
Mediterranean countries in southern Europe tend to have lower rates than
Caucasians from northern Europe;  individuals of Celtic origin in Australia tend
to have higher rates than non-Celtic individuals.   Variation in the incidence of
CMM within the Caucasian race is commonly thought to be a function of variation
in genetically determined pigmentary traits across ethnic groups.

    Numerous epidemiologic studies have focused on identification of important
pigmentary characteristics in the etiology of CMM.  Exhibit 8-6 shows a summary

-------
                                     8-26
                                  EXHIBIT 8-6

                Malignant Melanoma Risk Factors by Measures of
               Skin Pigmentation Within the Caucasian Population
     Study Reference
                                       Measures of Skin Pigmentation
                                        Within Caucasian Population
                                Reaction
                                 to Sun
Skin   Hair    Eye              (Tanning/
Color  Color  Color  Freckling  Sunburn)   Ethnicity
Lancaster and Nelson 1957     +

Gellin et al. 1969            +

Lane-Brown et al.  1971

IARC 1976

MacDonald 1976

Klepp and Magnus 1979

Mackie and Atchinson 1982

Beral et al.  1983             +

Hinds and Kolonel 1983

Lew et al. 1983

Elwood et al. 1984            +

Holman and Armstrong 1984b    +

Graham et al. 1985            +
         +

         +

         +
+

+
         +

         +
+

+

+

+
NOTE:      + = Significant risk factor.
           - = Not significant risk factor.
       Blank = Not included in study.

-------
                                      8-27
of findings from a number of these studies.  The following associations between
pigmentary traits and CMM risk have been found:

         (a)     Skin color --in all studies reviewed, fair
                complexions were associated with higher risks of
                CMM than were dark complexions.

         (b)     Hair and eye color -- in most studies, red and
                blond hair in childhood were associated with
                increased risk of CMM relative to dark hair.  Blue
                eyes were an independent risk factor in only one
                of four well-controlled epidemiologic studies;
                however, this could be due to the homogeneous
                nature of most of the study populations.

         (c)     Freckling -- those who freckled readily were at
                consistently elevated CMM risk relative to other
                individuals.

         (d)     Reaction to sun exposure -- in most studies
                reviewed, individuals who usually burned and were
                unable to tan were at significantly higher risk of
                CMM than those who tanned easily.

Exposure to sunlight appears to encourage the appearance (and possibly the
disappearance) of nevi (including dysplastic nevi) from the skin (Kopf et al.
1978, 1985; Armstrong et al. 1986).  Dysplastic nevi are clearly a risk factor
for melanoma independent of freckling or other pigmentary characteristics; it
also appears that the pressure of congenital nevi as well as acquired
melanocytic nevi may also be risk factors (Elder et al.  1981).

    Information from xeroderma pigmentosum (XP) patients indicates that
individuals who have an inability to repair solar-radiation-induced DNA damage
also have a high incidence of CMM relative to the normal population (Kraemer et
al.  1984).  The best characterized defect in XP patients is an inability to
excise pyrimidine dimers (Cleaver 1983),  which suggests that the repair of such
lesions can be important to prevention of CMM development.

Miscellaneous Factors

    Although a number of possible etiologic agents for melanoma have been
evaluated (e.g., exposure to chemicals or ionizing radiation), no strong
candidate has emerged.   It has been suggested that the risk of developing CMM
may be elevated among individuals exposed to fluorescent lighting at work (Beral
et al.  1982); however,  several recent well-controlled studies have failed to
find such an association (Rigel et al.  1983;  Durbin et al.  1986;  Elwood et al.
1986).   Indeed, one of these studies initially found a positive association
based on personal interviews,  but a self-administered postal survey of the same
study subjects did not confirm the initial findings Pasternack et al.  1983;
Durbin et al. 1986) .

    Total CMM incidence has been found to be higher among professional and
administrative indoor office workers,  but not among other indoor workers
relative to outdoor workers (Lee and Strickland 1980;  Holman et al.  1980a);

-------
                                      8-28
however, the implications of this finding may be confounded by differences in
socioeconomic status, melanoma site, and histologic type.  There is a higher
risk of indoor office workers developing CMM on an anatomical site that is
usually covered (e.g., the trunk) than on a site that is usually exposed (e.g.,
the face) (Beral and Robinson 1981; Vagero et al.  1986).  For usually covered
parts of the body, the incidence of CMM among indoor office workers is higher
than for outdoor workers.  And the reverse is also true; for usually uncovered
parts of the body, the incidence of CMM is higher among outdoor workers than
among indpor office workers (Beral and Robinson 1981; Vagero et al. 1986).

    It is not clear why certain indoor workers are at greater risk for
developing CMM than outdoor workers, nor why in these individuals the affected
sites are those not normally exposed.  One speculation is that indoor workers
have more leisure time and that they spend that time under conditions of solar
exposure that maximize intensive UV-B exposures of areas that are not protected
by a tan; however, there is no information currently available from a published
study that would allow this hypothesis to be examined.  From current data it is
clear that for certain indoor workers,  cumulative sun hours of exposure are not
associated with melanoma risk.  This information does not rule out that some
other measure of exposure (e.g., number of high UV-B exposures received at a
time of low tan) may be associated with melanoma risk.

Weighing and Balancing the Epidemiologic Evidence

    In summary, while there is no single piece of evidence that proves that
solar radiation causes CMM, the weight of the available epidemiologic evidence
clearly supports the conclusion that solar radiation plays a major role in the
development of CMM for susceptible populations.  Cumulative dose (i.e., total
sun-exposed hours) clearly does not explain important variations in incidence
rates.  Although from existing evidence it is not possible to determine the most
accurate method of calculating dose, the collective evidence does suggest that
exposure to solar radiation in some form produces cutaneous malignant melanomas
in Caucasian populations.

EXPERIMENTAL EVIDENCE

    Much of the experimental evidence relevant to the relationship of CMM and
exposure to UV-B has been reviewed on non-melanoma skin cancer.  To briefly
recapitulate, there are five important points:

        (1)  UV-B is the most active portion of solar radiation
             responsible for the induction of adverse effects on
             mammalian systems; it has been shown in vivo and in vitro
             to induce transformation of mammalian epidermal cells,
             (albeit not melanocytes), and to be mutagenic to them as
             well.  All of these effects are thought to occur by a
             mechanism that involves damage to the DNA  (Doniger et al.
             1981)

        (2)  UV-B at 295-300 nm is the most active waveband of UVR for
             these effects, as well as for the induction of pyrimidine
             dimers.  (See pg 7 of Chapter 7.)

-------
                                     8-29
        (3)   Patients with xeroderma pigmentosum,  who cannot repair
             pyrimidine dimers induced in DNA by UV-B (which are
             thought to be important to skin cancer development) have
             a 2000-fold higher rate of CMM by the time they are 20
             than normal individuals.  (Kraemer et al. 1984)

        (4)   UVR is carcinogenic in animals and it is the UV-B
             wavelengths that are the most effective at inducing
             cancer -- the shorter UV-B wavelenths being more
             carcinogenic than the longer ones (Forbes et al. 1981).
             The tumors induced are principally fibrosarcomas and
             squamous cell carcinomas; melanomas have not yet been
             experimentally induced with any consistency by UVR.
             Suppression of the immune response in humans (e.g., in
             renal transplant patients or those on immunosuppressive
             drugs) have increased incidence of malignant melanoma.

        (5)   UVR, and specifically UV-B, can induce the generation of
             T suppressor cells, which specifically suppress the
             immune response to UV-induced tumors.  The suppressor
             cells have been shown to be capable of shortening the
             latent period of tumors induced by UV-B, and UV-B-treated
             animals also appear to have increased susceptibility to
             transplanted melanomas.  (For detailed discussion see
             chapter .9.)

Weighing and Balancing the Experimental Evidence

    The experimental data indicate that UV-B irradiation damages DNA and induces
carcinogenesis and immunosuppression in animals.  CMM rates are elevated in
patients who cannot repair UVR-induced lesions in DNA.  Thus solar ultraviolet
radiation has the potential to effect CMM development in multiple ways.  A
pictorial representation of this, developed by Dr. Thomas Fitzpatrick (1987) of
Harvard University, is presented in Exhibit 8-7.

DOSE-RESPONSE RELATIONSHIPS

    Empirical relationships between exposure to UV-B and the incidence and
mortality of CMM have been estimated in several studies using population-based
mortality or incidence information.

    Animal studies are inadequate for estimating CMM dose-response because there
is no animal model for melanoma induced by UV-B.  Because the limitations in
animal studies appear insurmountable at this time, epidemiologic
studies provide the best candidate dose-response relationships for risk
estimation.   There are no epidemiologic studies in which individual human UVB
exposures have been measured.  Instead, population-based estimates of UVB
exposure have generally been used to assess dose-response relationships.  This
section reviews four such studies.  Three of the studies -- Fears et al. (1976,
1977) and Scotto and Fears (1986) -- develop dose-response relationships
associating UV-B and melanoma incidence.  Two studies -- Fears et al. (1976) and
Pitcher (1987) -- estimate dose-response relationships for melanoma mortality.

-------
                                               Exhibit 8-7
Solar UVR
        \
           Initiator
                                        Multiple Ways in Which UVR
                                   Can Play a Role  in Melanoma Development
                         Inability to tan
                         (Skin types I & II)
                  ?;T*;C Normal melanocyte
                        v
       Promoter
                                        Transformed cell
Clark's (dysplastic)
melanocytic nevus

Common acquired
melanocytic nevus
Altered
melanocyte
              Defective DNA repair
              (e.g. Xeroderma
              pigmentosum)
                                                                       Defective immunologic
                                                                       surveillance
                                                                       (e.g. following
                                                                       renal transplant)
                                                                                                          00

                                                                                                          o
                                         Melanoma cell
"in the white population
                                                                             © T. B. Fitzpatrick

-------
                                      8-31
    Fears et al. (1976) present one epidemiologic analysis that is potentially
useful for estimating dose -response.  These authors use two types of data to
represent UV-B:  (1) latitude -- latitude and UV-B radiation weighted for
erythema effectiveness correlate at 0.97, and (2) monthly totals of erythema-
producing UV radiation expressed as Biologically Effective Units (BEU) and
derived from Schultze (1974).  Both types of data were correlated with CMM
mortality and incidence data.  Incidence information for four cities was
obtained from the Third National Cancer Survey (TNCS) (1975), and mortality data
were obtained from the U.S. Cancer Mortality by County database (Mason and McKay
(1973).

    Exhibit 8-8 shows the results of a simple correlation model based on
latitude:

                                 log R =  a  + B>L

where R is the age -adjusted rate of incidence or mortality, a and & are
constants, and L is latitude.  The estimated coefficients were statistically
significant at p > 0.01.  Exhibit 8-9 shows the results estimated using BEU's.
These estimates were based on an exponential model and represent the percentage
changes in incidence and mortality estimated to occur with increases in UV
radiation dose of between 10 and 30 percent.  Note that the use of an
exponential model results in higher dose -response relationships for higher base
exposures .

    There are several important limitations in this work.  In particular, only
four cities were used for melanoma incidence.  With a sample this small,
variations in confounding factors such as behavior, skin pigmentation, and cloud
cover may bias dose-response estimates.  As a result, caution must be used in
evaluating the error of estimate.

    In another study, the same authors estimated a dose -response relationship
for melanoma incidence using a power model  (Fears et al. 1977)
                    P.. = b  (U.
where :
    P.. = probability of developing melanoma for jth age group at
          location i;

    U.  = annual UV-B count at ith location;

    A.  = midrange of jth age group; and

    b, c, and k are constants.

This equation was estimated by fitting a log form of the model (In R. . = a + c
In U  + kLnA. + E..) to age-specific melanoma incidence rates for botn males and
    i        i    ij      t>   t-

females after weighting by the observed number of cases.  Annual counts from
Robertson-Berger (R-B) meters were used for UV-B; the R-B meter attaches greater

-------
                                                  8-32
                                                EXHIBIT  8-8

                                    Sumary Statistics  for Regressions of
                               Skin Cancer Incidence and  Mortality on Latitude
            	MALES	     	FEMALES	

            Correlation     Regression   Doubling     Correlation    Regression     Doubling
            Coefficient*   Slope + S.D.   Latitude     Coefficient*   Slope + S.D.   Latitude
Melanoma       -.86        -.031 +  .007     -  9.8U
  Incidence
  Mortality
*Simple correlation coefficient between log of incidence/mortality  rate  and latitude.
Source:  Fears, Scotto,  and Schneiderman (1976).

-------
                                     8-33
                                     EXHIBIT 8-9

         Estimated Relative  Increases  in Melanoma  Skin  Cancer  Incidence  and
                Mortality Associated with Changes in Erythema Dose a/
             (figures in parentheses  are 95 percent confidence  intervals)



Melanoma
Incidence



Melanoma
Mortality




BASE
BED


650
850
1050


650
850
1050
MALES
FEMALES
Increase in Total Dose
10% 20%


15 (7-24) 32
20 44 (18-75)
25 57


8 (6-10) 16
10 22 (16-28)
13 28
30%


52
72
96 b/ (37-180)


26
35
45 (33-58)
Increase in Total Dose
10%


13
18
22


6
9
11
20%


29
39
50


13
18
22
30%


46
64
84 b/


21
28
36
a/ A sample computation is as follows:

A 10% increase in total dose at 48.25'N, where exposure is 650 biologically effective
units (BEU),  equals 65 BEU.

A change of 65 BEU is equivalent to a reduction in latitude of 1.93'.

A change of 1.93' at 48.25'N is associated with an 18% increase in nonmelanoma
incidence.

b/ These estimates require extrapolation well beyond the range of the data.
Source: Fears, Scotto, and Schneiderman (1976).

-------
                                     8-34
weight to longer wavelengths compared to the erythema action spectrum.
Incidence data also were from the TNCS (1975) .

    Exhibit 8-10 presents summary regression coefficients and statistics for
this age-specific analysis.   Applying the results in this exhibit, a one percent
increase in UVR results in an estimated 100*((1.01) -1) percentage increase in
melanoma incidence, or 2.47  for females and 2.24 for males.  The authors stress,
however, that these estimates may be biased by the omission of location-specific
demographic and environmental variables.

    Scotto and Fears (1986)  estimate a dose-response relationship for melanoma
incidence using updated information about populations in a greater number of
cities.  The authors also introduce a new term (VAR) to adjust for the presence
of some host or environmental characteristics:

            Ln R..  = a + b In (Age.) + c In UV-B + dVAR + e

Annual counts obtained from Robertson-Berger meters were used for UV-B.  Seven
areas were included:  Detroit, Seattle, Iowa, Utah, San Francisco, Atlanta, and
New Mexico.  Separate estimates were provided for:  (1) anatomical sites of
melanoma -- FHN (face, hands, neck), UE (upper extremity), LE (trunk and lower
extremity); and (2) age groups -- 20-39,  40-54, 55-64, 65-74.

    Scotto and Fears (1986)  found that after controlling for confounding
variables each 1 percent increase in UV-B causes melanoma incidence to increase
by less than 1 percent.  Exhibit 8-11 shows results for different anatomical
sites.  The results are statistically significant (p < 0.1).  Exhibit 8-12
describes the effects of adding different constitutional and environmental
variables derived using step-wise multiple regression.  If the DNA action
spectrum is used, Scotto's results indicate slightly less than a 2 percent
increase in incidence for each 1 percent ozone depletion.

    Pitcher (1987)  used data on melanoma death rates by county for a 30-year
period and exposure data from a satellite-data-based National Air and Space
Administration model (the NASA Model; Serafino and Frederick 1986) to estimate
the relationship between CMM death rates and several alternative estimates of
UV-B doses.  These were either peak (a clear day in June) or mean annual dose
estimates using weighting functions derived either from the average DNA action
spectrum proposed by Setlow  (1974) or an erythema action spectrum.

    Data on melanoma mortality by county were obtained from an EPA/NCI data base
(Riggan et al. 1983) for the period 1950 to 1979.  The NASA model estimated UV
exposures in these counties by using modeling relationships based on latitude,
longitude, altitude, surface albedo, total column ozone, and cloud
cover to estimate the amount of UV delivered.  Weighting functions derived with
various biological action spectra were then applied to these total UV energy
estimates to get estimates of "effective" UV energy termed "weighted UV doses."
An extensive effort to validate the exposure data obtained from the UV model
using ground-based measurements from Robertson-Berger meters has been made and
will be reported in a subsequent report  (Pitcher 1987b) .

    An  examination of Pitcher's  (1987a) data indicated that  the relationship
between age and melanoma death rates is approximately exponential.  Further, the
rates of increase  for males  and females were different.  These observations led

-------
                                     8-35
                                     EXHIBIT  8-10

                   Summary of Fears, Scotto, and Schneiderman (1977)
               Regression Analyses  of Melanoma  Incidence Dose-Response*
                       t-Statistic       a +  SD          K +  SD          c +  SD
SEX R- Squared
Males 0.74
Females 0.62
(UV Coefficient) (constant) (age coefficient) (UV coefficient
5.4 -35.6 + 6.5 .80 + .31 2.45 + .45
4.6 -30.5 + 6.9 .29 + .34 2.23 + .48
* In Rp = a + c In U. + K In A. +
Source:  Fears, Scotto, and Schneiderman (1977)

-------
                                      8-36
                                 EXHIBIT  8-11
              Biological Amplification Factors  for"   Skin Melanoma
               by  Sex and Anatomical  Site  Groups, Adjusting for
            Age and Selected Constitutional and Exposure Variables*
                                   	MALE	     	FEMALE	
    VARIABLE                       Trunk/LE    FHN/UE    Trunk/LE      FHN/UE


Age                                  6%          8%          5%           10%
Sunburn
Freckles
Scandinavian ancestry
Light hair color
Scot/Irish ancestry
Moles
Light eye color
Fair skin
Sunscreen use
Suntan lotion use
Radiation protection
Protective clothes
Hours outdoors during weekdays
Hours outdoors during weekends
4
5
5
5
6
6
6
6
4
5
5
8
7
5
8
7
9
8
6
7
8
8
7
8
8
10
9
7
5
5
5
5
4
5
4
4
4
4
6
5
5
8
10
10
10
10
8
11
9
9
10
10
10
10
10
13
   A biological amplification factor indicates the relative change in melanoma
incidence associated with a 10 percent relative increase in UV-B; abbreviations
used for anatomical sites are LE:  lower extremities; FHN:  face, hands, neck;
and UE:  upper extremities.

Source:  Scotto and Fears (1986).

-------
                                      8-37
                                  EXHIBIT  8-12

           Biological Amplification Factors for~  Melanoma Incidence
               by Sex and Anatomical  Site Groups, Adjusting  for
    Age and Combinations of Selected Constitutional and Exposure Variables*
ANATOMICAL
   SITE
            MALE
            FEMALE
Trunk/LE

  Variables
  included
  in model
FHN/UE

  Variables
  included
  in model
          3 Percent

Suntan lotion use
Scandinavian ancestry
Light hair color
UV-B radiation index
Hours outdoors during weekdays

          5 Percent

Scot/Irish
Suntan lotion use
Fair skin
UV-B radiation index
Hours outdoors during weekends
           4 Percent

Suntan lotion use
Scandinavian ancestry
Light hair color
UV-B radiation index
Hours outdoors during weekends

        6 Percent

Scot/Irish
suntan lotion use
fair skin
UV-B radiation index
a /
   A biological amplification factor is a number that indicates the percentage
increase in melanoma incidence for a 10 percent relative increase in UV-B;
abbreviations used for anatomical sites are LE:  lower extremities; FHN:  face,
hands, neck; and UE:  upper extremities.

Source:  Scotto and Fears (1986).

-------
                                      8-38
Pitcher to fit both a power and an exponential model to data for each sex group.
For either model DRM..,  is defined as the death rate for the ith cohort
in the jth location in the kth time period.  The exponential model then can be
expressed as

        DRM...  = exp(a. + a-AGE..  + a0WUV.) + e. „   (Model 1)
           ijk     ^  0    lik    2   j     ijk

    and the power model can be expressed as

        DRM...  = exp (b- + bn .AGE.,  + b.log(WUV.)) + e. ..  or,
           ijk     K   0    li   ik    2  6    j      ijk

        DRMi.k = (WUV.)b2 exp(bQ + b  AGEik) + e     (Model 2)


where AGE.,  is the age of ith cohort in the kth time period, WUV. is the UV dose
         llc                                                     J
in the ith location (SMA) and e...  is the error term.
       J                       ijk
    In model 1, the percentage change in melanoma mortality associated with
variations in UV flux is higher the greater the baseline exposure.  Model 2
differs from the first only in the use of the log of the exposure variable.  In
it, a 1 percent increase in UV dose generates the same percentage increase in
melanoma death rates regardless of the baseline death rate.

    Pitcher (1987a) found that many of the birth cohorts had zero deaths in any
given 5-year period, especially in smaller cities.  Further, because rates
differed significantly across cohorts, the variances of the cohort death rates
varied under the assumption that the probability of death had a binomial
distribution.  The different population sizes within cohorts also caused
variation in the variance of the cohort death rate.  These issues made the use
of normal linear regression techniques unfeasible.  Normal weighting techniques
could not be used because zero rates for some cohorts prevented computation of
weights.  Therefore, weights were computed using expected rates rather than
actual rates.

    Using both the power and the exponential models, Pitcher (1987a) estimated
the empirical relationships between melanoma mortality and a variety of
different estimates of UV dose derived through the combination of two exposure
scenarios (peak dose versus annual dose) and two different weighting functions
(based on the DNA damage or erythema action spectra).  Exhibit 8-13 presents the
percentage increase in melanoma death rates predicted for a one percent decline
in stratospheric ozone using the two models and four different dose estimates
(DNA-weighted peak, DNA weighted annual, erythema-weighted peak and
erythema-weighted annual) for males and females.  Clearly choice of the dose
estimate, particularly between annual or peak exposure durations, will make
large differences in the prediction of increases in melanoma death rate.
Unfortunately, the strength of the statistical association between any of  these
dose estimates and melanoma mortality are approximately the same so that a
choice cannot be made on that basis.  Furthermore, the mechanism of melanoma
induction, particularly the role of UV, is sufficiently unclear, that a choice
between the various options is difficult.  Further investigations are planned in
this area (Pitcher 1987b) .

-------
                                  8-39
                              EXHIBIT  8-13

              PERCENTAGE INCREASE IN MELANOMA DEATH RATES
                   FOR A ONE PERCENT DECLINE  IN OZONE
                                             MALES
El Paso
    DNA
    Erythema

San Francisco
    DNA
    Erythema

Minneapolis
    DNA
    Erythema
El Paso
    DNA
    Erythema

San Francisco
    DNA
    Erythema

Minneapolis
    DNA
    Erythema
Peak
Exponential
1.99
1.90
1.59
1.54
1.40
1.34
Peak
Exponential
1.30
1.27
1.04
1.03
.92
.90

Power
1.65
1.60
1.64
1.63
1.72
1.77
FEMALES

Power
1.11
1.10
1.10
1.09
1.20
1.12
Annual
Exponential
1.18
1.07
.82
.76
.52
.50
Annual
Exponential
.76
.70
.53
.50
.34
.33

Power
.82
.75
.82
.75
.86
.77

Power
.56
.53
.57
.53
.59
.54

-------
                                      8-40
Limitations of Epidemiologic Studies

    A variety of problems exist in applying the results of the epidemiologic
studies to an analysis of melanoma risks associated with ozone depletion; many
of these problems have been described in earlier sections.  In particular,
applying the dose-response relationships requires an assumption that certain key
factors will not change.   However,  the proportion of individuals in the
population who are most susceptible to UVR, such as certain ethnic groups, may
change over time.  Migration was not considered in the studies; if southern
states experience large increases in population, then the sensitivity of future
populations to UV-B will be underestimated.  Individuals might also change their
use of sunscreens or their propensity to work outdoors.

    The epidemiologic studies do not measure actual ambient UV-B dose.
Therefore, UV-B data used to estimate dose-response relationships will differ
from the actual potential ambient exposure of individuals at any location.
Other factors will also affect actual ambient exposures, such as skin color,
patterns of dress, work exposure, recreation, eating, and medical care and
intervention.  These factors can be expected, in varying degrees, to introduce
unexplained variation (i.e., noise) into any analysis of incidence or mortality.
Finally, few data are available to assess the significance of the genetic
factors that effect melanoma incidence but are not captured in epidemiologic
studies.

-------
                                      8-41
                                   REFERENCES
Acquavella, J.F., Wilkinson, G.S., Tietjen, G.L., Key, C.R., Stebbings, J.H.,
and Voelz, G.L.  A melanoma case-control  study at the Los Alamos National
Laboratory.  Health Phys 45:587-592  (1983).

Adam, S.A., Sheaves, J.K., Wright, N.H.,  Mosser, G., Harris, R.W., and Vessey,
M.P.  A case-control study of the possible association between oral
contraceptives and malignant melanoma.  Br J Cancer 44:45-50 (1981).

Adler, S., and Gaeta, J.F.  Malignant melanoma.  In:  Cancer Dermatology. Helm,
F. (ed). Philadelphia:  Lea & Febiger pp. 141-157 (1979).

Anaise, D., Steinitz, R., and Ben Hur, N.  Solar radiation:  A possible
etiological factor in malignant melanoma  in Israel:  A retrospective study
(1960-1972).  Cancer 42:299-304 (1978).

Anderson, R.R.  Tissue optics and photoimmunology.  In:  Photoimmunology.
Parrish, J.A., Kripke, M.L., and Morison, W.L. (eds).  New York:  Plenum Medical
Book Co. pp. 61-76 (1983).

Armstrong B.K.  Melanoma of the skin.  Br Med Bull 40:346-350 (1984).

Armstrong, B.K., de Klerk, N.H., and Holman, C.D.J.  Etiology of common acquired
melanocytic nevi:  Constitutional variables, sun exposure, and diet.  JNCI
77(2):329-335 (1986).

Armstrong, B.K., Holman, C.D.J., Ford, J.M., and Wooding, T.L.   Trends in
melanoma incidence and mortality in Australia.  In:  Trends in Cancer Incidence.
Causes and Practical Implications. Magnus, K., ed., Hemisphere Publications
(1981).

Austin, D.F., Reynolds, P.J., Snyder, M.A., Biggs, M.W., and Stubbs, H.A.
Malignant melanoma among employees of Lawrence Livermore National Laboratory.
Lancet 2:712-716 (1981).

Bahr, A.K., Rosewaike, I., Hermann, N., Grover, P., Stellman, J., and O'Leary,
K.  Letter to the editor:   Melanoma after exposures to PCBs.  N Engl J Med,
August 9, p. 450 (1976).

Baker-Blocker, A.  Ultraviolet radiation  and melanoma mortality in the United
States.  Env Res 23:24-28 (1980).

Balch, C.M., Karakousis, C., Mettlin, C., Natarajan, N., Donegan, W.L., Smart,
C.R., and Murphy, G.P.  Management of cutaneous melanoma in the United States.
Surg Gynecol Obstet 158:311-318 (1984).

Balch, C.M., Soong,  S., Milton,  G.W., Shaw,  H.M., McGovern, V.J., McCarthy,
W.H., Murad, T.M.,  and Maddox,  W.A.  A comparison of prognostic factors and
surgical results in 1,786 patients with localized (Stage I) melanoma treated in
Alabama,  USA,  and New South Wales,  Australia.   Ann Surg 6:677-684 (1982).

-------
                                     8-42
Balch, C.M., Soong,  S.,  Milton,  G.W.,  Shaw,  H.M.,  McGovern,  V.J.,  McCarthy,
W.H.,  Murad, T.M.,  and Maddox,  W.A.   Changing trends in cutaneous  melanoma over
a quarter century in Alabama,  USA,  and New South Wales, Australia.  Cancer
52(9): 1748-1753 (1983).

Beral, V., Evans,  S., Shaw,  H.,  and Milton,  G.   Cutaneous factors  related to the
risk of malignant melanoma.   Br J Dermatol 109:165-172 (1983).

Beral, V., Evans,  S., Shaw,  H.  and Milton, G.  Milignant melanoma  and exposure
to fluorescent lighting at work.  Lancet 1:290-293 (1982)

Beral, V., Ramcharan, S.,  and Faris,  R.  Malignant melanoma and oral
contraceptive use among women in California.  Br J Cancer 36:804-809 (1977) .

Beral, V., and Robinson, N.   The relationship of malignant melanoma, basal and
squamous skin cancers to indoor and outdoor work.   Br J Cancer 44:886-891
(1981).

Boyle, P., Day, N.E., and Magnus, K.   Mathematical modeling of malignant
melanoma trends in Norway, 1953-1978.   Am J Epidemiol 118:887-896  (1983).

Briele, H.A., and Das Gupta, T.K.  Natural history of cutaneous malignant
melanoma.  World J Surg 3:255-270 (1979).

Briggs, J.C.  The role of trauma in the aetiology of malignant melanoma:  A
review article.  Br J Plastic Surg 37:517-516 (1984).

Cleaver, J.E.  Xeroderma pigmentosum.   Chapter 57 in Stanbury, J.B., Wyngarten,
J.B.,  Frederickson,  D.S.,  Goldsteen,  J.L. and Brown, M.S. (eds.)  The Metabolic
Basis of Disease. McGraw-Hill Book Company, New York, pp. 1227-1248 (1983).

Clark, W.H. , Jr., Elder, D.E., and van Horn, M.  The biologic forms of malignant
melanoma. Hum Pathol 17:443-450  (1986).

Collins, R.J.  Melanoma in the Chinese of Hong Kong: Emphasis on volar and
subungual sites.  Cancer 54:1482-1488  (1984).

Cook,  P.J., Doll, R., Fellingham, S.A.  A mathematical model  for  the age
distribution of cancer  in man.   Int J Cancer 4:93-112  (1969).

Cooke, K.R., and Fraser, J.  Migration and death from malignant melanoma.  Int J
Cancer 36:175-178 (1985).

Cooke, K.R., Skegg, D.C.G., and  Fraser, J.  Trends in malignant melanoma of skin
in New Zealand.  Int J  Cancer 31:715-718  (1983).

Cooke, K.R., Skegg, D.C.G., and  Fraser, J.   Socio-economic status,  indoor and
outdoor work and malignant melanoma.   Int J  Cancer 34:57-62  (1984).

Crombie,  I.K.  Racial differences in melanoma incidence.  Br  J Cancer 40:185-193
(1979a)

Crombie,  I.K.  Variation of melanoma incidence with  latitude  in North America
and Europe.  Br J Cancer 40:774-781 (1979b).

-------
                                      8-43
Crombie, I.K.  Distribution of malignant melanoma on the body surface.  Br J
Cancer 43:842-849 (1981).

Dobson, A.J., and Leeder, S.R.  Mortality from malignant melanoma in Australia:
Effects due to country of birth.  Int J Epidemiol 11(3):207-211 (1982).

Doniger, J., Jacobson, E.D., Krell,  K.,  Di Paolo, J.A.  Ultraviolet light action
spectra for neoplastic transformation and lethality of syrian hamster embryo
cells correlate with spectrum for pyrimidine dimer formation in cellular DNA.
Proc Nat Acad Sci (USA) 78(4):2378-2382 (1981).

Durbin, N., Moseson, M.,  and Pasternack, B.S.  Epidemiology of malignant
melanoma:  Pigmentary traits, ultraviolet radiation, and the identification of
high-risk populations.  Recent Results in Cancer Research 102:56-75 (1986).

Eklund, G., and Malec, E.  Sunlight and incidence of cutaneous malignant
melanoma.  Scand J Plast Reconstr Surg 12:231-241 (1978).

Elder, D.E., Jucovy, P.M., Tuthill,  R.J.,  and Clark, W.H.  The classification of
malignant melanoma.  Am J Dermatopathol 2:315-320 (1980).

Elder, D.E., Greene, M.H., Bondi, E.E.,   and Clark,  W.H.    Acquired melanocytic
nevi and melanoma, The.dysplatic nevus syndrome, Chap. 11,
In: Pathology of Malignant Melanoma. Ackerman, A.B.  (ed)  New York: Masson
Publishing, USA, pp 185-215, (1981).

Elwood, J.M., Gallagher,  R.P., Davison,  J., and Hill,  G.B.  Sunburn, suntan and
the risk of cutaneous malignant melanoma:   The Western Canada melanoma study.
Br J Cancer 51:543-549 (1985b).

Elwood, J.M., Gallagher,  R.P., Hill, G.B., and Pearson, J.C.G.  Cutaneous
melanoma in relation to intermittent and constant sun exposure:   The Western
Canada melanoma study.  Int J Cancer 35:427-433 (1985a).

Elwood, J.M., Gallagher,  R.P., Hill, G.B., Spinelli, J.J., Pearson,  J.C.G., and
Threlfall, W.  Pigmentation and skin reaction to sun as risk factors for
cutaneous melanoma:  Western Canada melanoma study.   Br Med J 288:99-102 (1984).

Elwood, J.M., and Hislop, T.G.  Solar radiation in the etiology of cutaneous
malignant melanoma in Caucasians.  Natl Cancer Inst Monogr 62:167-171 (1982).

Elwood, J.M., and Lee, J.A.H.  Trends in mortality from primary tumors of skin
in Canada.  CMA Journal 110:913-915  (1974).

Elwood, J.M., and Lee, J.A.H.  Recent data on the epidemiology of malignant
melanoma.  Semin Oncol 2(2):149-154  (1975).

Elwcod, J.M., Williamson, C., and Stapleton,  P.J.   Maligant melanoma in relation
to moles, pigmentation and exposure  to fluorescent and other lighting sources.
Br J Cancer 53:65-74 (1986).

-------
                                     8-44
Fears, T.R.,  Scotto,  J.,  and Schneiderman,  M.A.   Skin cancer,  melanoma and
sunlight.  Am J Public Health 66:461-464 (1976).

Fears, T.R.,  Scotto,  J.,  and Schneiderman,  M.H.   Mathematical models of age and
ultraviolet effects on the incidence of skin cancer among whites in the United
States.  Am J of Epidemiol 105:420-427 (1977).

Fitzpatrick,  T.B.,  personal communication,  1986.

Fitzpatrick,  T.B.,  personal communication,  1987.

Fitzpatrick,  T.B.,  and Soter, N.A.   Pathophysiology of the skin.  In:
Pathophysiology:  The Biological Principles of Disease.  Smith, L.H., Jr.,  and
Thier, S.O. (eds.), pp. 1297-1330,  1985.

Forbes, P.D., Blum, H.F.,  Davies,  R.E.  Photocarcinogenesis in hairless mice:
dose-response and the influence of dose-delivery.  Photochem Photobiol
34:361-365 (1981).

Gellin, G.A., Kopf, A.W.,  and Garfinkel, L.  Malignant melanoma:  a controlled
study of possibly associated factors.  Arch Derm 99:43-48 (1969).

Gordon, T., Crittenden, M., and Haenszel, W.  End results and mortality trends
in cancer: II.  Cancer mortality trends in the United States.   Natl Cancer Inst
Monogr 6:133-350 (1961).

Graham, S., Marshall, J.,  Haughey,  B., Stoll, H., Zielezny, M., Erasure, J., and
West, D.  An inquiry into the epidemiology of melanoma.  Am J Epidemiol
122(4):606-619  (1985).

Green, A.  Sun exposure and the risk of melanoma.  Br J Derm 25:99-102  (1984).

Green, A., MacLennan, R.,  and Siskind, V.  Common acquired naevi and the risk of
malignant melanoma.  Int J Cancer 35:297-300 (1985a).

Green, A., and O'Rourke,  M.G.E.  Cutaneous malignant melanoma  in association
with other skin cancers.   JNCI 74(5):977-980 (1985).

Green, A., and  Siskind, V.  Geographical distribution of cutaneous melanoma in
Queensland.  Med J Aust 1:407-410 (1983).

Green, A., Siskind, V., Bain, C., and Alexander,  J.  Sunburn and malignant
melanoma.  Br J Cancer 51:393-397 (1985b).

Haenszel, W.  Variations in skin cancer merdena within the United States.  Natl
Cancer Inst Monogr 10:225-243  (1963).

Herron, J.  The geographical distribution of malignant melanoma  in Queensland.
Med J Aust 1:892-894  (1969).

Hinds, M.W.  Anatomic distribution of malignant melanoma of the  skin among
non-Caucasians  in Hawaii.  Br J Cancer 40:497-499  (1979).

-------
                                      8-45
Hinds, M.W., and Kolonel, L.B.  Malignant melanoma of the skin in Hawaii,
1960-1977.  Cancer 45:811-817  (1980).

Heston, J.F., Kelly J.B., Meigs, J.W., and Flannery J.T.  Forty-five years of
cancer incidence in Connecticut:  1935-1979.  NCI Monograph No. 70  (1985).

Hinds, M.W., Lee, J., and Kolonel, L.N.  Seasonal patterns of skin melanoma
incidence in Hawaii.  Am J Public Health 71:496-499 (1981).

Hinds, M.W., and Kolonel, L.B.  Cutaneous Malignant Melanoma in Hawaii --An
update.  West J Med 138(1):50-54 (1983).

Hoar, S.K., and Pell, S.  A retrospective cohort study of mortality and cancer
incidence among chemists.  J Occup Med 23:483-494 (1981).

Holman, C.D.J., and Armstrong, B.K.  Skin melanoma and seasonal patterns.  Am J
Epidemiol 113:202 (1981).

Holman, C.D.J., and Armstrong, B.K.  Cutaneous malignant melanoma and indicators
of total accumulated exposure  to the sun:  An analysis separating histogenetic
types.  JNCI 73:75-82 (1984a).

Holman, C.D.J., and Armstrong, B.K.  Pigmentary traits, ethnic origin, benign
nevi, and family history as risk factors for cutaneous malignant melanoma.  JNCI
72:257-266  (1984b).

Holman, C.D.J., Armstrong, B.K., Evans, P.R., Lumsden, G.J., Dallimore, K.J.,
Meehan, C.J., Beagley, J., and Gibson, I.M.  Relationship of solar keratosis and
history of  skin cancer to objective measures of actinic skin damage.  Br J
Dermatol 110(2):129-38 (1984c).

Holman, C.D.K., Armstrong, B.K., and Heenan, P.J.  Guest editorial:  A theory of
the etiology and pathogenesis of human cutaneous malignant melanoma.  JNCI
71:651-656  (1983).

Holman, C.D.J., Armstrong, B.K., and Heenan, P.J.  Cutaneous malignant melanoma
in women:   Exogenous sex hormones and reproductive factors.  Br J Cancer
50:673-680  (1984d).

Holman, C.D.J., Armstrong, B.K., and Heenan, P.J.  Relationship of cutaneous
malignant melanoma to individual sunlight-exposure habits.  JNCI 76:403-414
(1986).

Holman, C.D.J., Heenan,  P.J.,  Caruso, V., Clancy, R.J., and Armstrong, B.K.
Seasonal variation in the junctional component of pigmented nevi.  Int J Cancer
31:213-215  (1983).

Holman, C.D.J., James, I.R.,  Gattey,  P.H.,  and Armstrong, B.K.   An analysis of
trends in mortality from malignant melanoma of the skin in Australia.   Int J
Cancer 26:703-709 (1980a).

Holman, C.D.J., Mulroney, C.D., and Armstrong,  B.K.   Epidemiology of
pre-invasive and invasive malignant melanoma in Western Australia.   Br J Cancer
25:317-323  (1980b).

-------
                                     8-46
Holmbert,  B.,  Westerholm,  P.,  Massing,  R.,  Kestrup,  L.,  Gumaelius, H.,  Holmlund,
L.,  and Englund, A.  Retrospective cohort study of two plants in the Swedish
rubber industry.  Scand J  Work Environ Health 9(2):59-68 (1983).

Houghton,  A.,  Flannery, J., and Viola,  M.V.  Malignant melanoma in Connecticut
and Denmark.  Am J Cancer  25:95-104 (1980).

Houghton,  A.,  Munster, E.W.,  and Viola, M.V.  Increased incidence of malignant
melanoma after peaks of sunspot activity.   Lancet (April 8):759-760 (1978).

Houghton,  A.N.,  and Viola, M.V.  Solar radiation and malignant melanoma of the
skin.  J Am Acad Dermatol  5:477-483 (1981).

IARC monographs.  Biological data relevant to the evaluation of carcinogenic
risk to humans.   Appendix to Vol 40, in press, (1986).

Katz, L.,  Ben-Turia, S., and Steinitz,  R.   Malignant melanoma of the skin in
Israel:  Effect of migration.   In: Trends in Cancer Incidence:  Causes and
Practical Implications. Magnus, K. (ed), Washington:  Hemisphere Pub.  Corp.
(1982).

Kiryabwire, J.W., Lewis, M.G., Ziegler, J.L., and Loefler, I.  Malignant
melanoma in Uganda.  East Afr Med J 45:498-507 (1968).

Kleinbaum, D.G., Kupper, L.L., and Morgenstern, H.  Epidemiologic Research:
Principles and Quantitative Methods.  New York:  Van Nostrand Reinhold (1982).

Klepp, 0., and Magnus, K.   Some environmental and bodily characteristics of
melanoma patients:  A case-control study.  Int J Cancer 23:482-486  (1979).

Kopf, A.W.  Personal communication (1987)

Kopf, A.W., Lazar, M., Bart, R.S., Dubin, N., and Bromberg, J.  Prevalence of
nevocytic nevi on  lateral and medial aspects of arms.  J Dermatol Surg Oncol
4124  (1978).

Kopf, A.W., Kripke, M.L.,  and  Stern, R.S. Sun and malignant melanoma .  Am Acad
Dermatol 11:674-684  (1984).

Kopf, A.W., Lindsay, A.C., Rogers, G.S., Friedman, R.J., Rigel, D.S., and
Levenstein, M.  Relationship of nevocytic nevi to sun exposure  in dyplastic
nevus syndrome.  J Am Acad Dermatol 12:656-62  (1985).

Kraemer, K.H.,  Lee, M.M.,  and  Scotto, J.  DNA repair protects against cutaneous
and internal  neoplasia:   Evidence from xeroderma pigmentosum.   Carcinogenesis
5:511-514  (1984).

Lancaster,  H.O., and Nelson, J.   Sunlight  as a cause of melanoma:   A clinical
survey.  Med  J  Aust  6:452-456  (1957).

Lane-Brown, M.M.,  Sharpe,  C.A.B., MacMillen, D.S.,  and McGovern,  V.J.  Genetic
predisposition  to  melanoma and other skin  cancers in Australians.   Med J Aust
1:852-853  (1971).

-------
                                      8-47
Larsen, I.E., Little, J.H., Orell, S.R., and Prade, M.  International
pathologists' congruence survey on quantitation of malignant melanoma.
Pathology 12:245-253 (1980).

Lee J.A.H.  Melanoma.  In:  Cancer Epidemiology and Prevention. Shottenfeld, D.,
and Fraumeni, J.F., Jr., (Eds.).  Philadelphia:  W.B. Saunders Company (1982a).

Lee, J.A.H.  Melanoma and exposure to sunlight.  Epidemiol Rev 4:110-136
(1982b).

Lee, J.A.H.  Epidemiology of malignant melanoma:  10 Years' Progress.  Pigment
Cell 6:1-21 (1983).

Lee, J.A.H., and Carter A.P.  Secular trends in mortality from malignant
melanoma.  J Natl Cancer Inst 45:19-97  (1970).

Lee, J.A.H., and Issenberg, H.J.  A comparison between England and Wales and
Sweden in the incidence and mortality of malignant skin tumors.  Br J Cancer
26:59-66 (1972).  .

Lee, J.A.H., Petersen,  G.R., Stevens, R.G.,  and Vesanen, K.  The influence of
age, year of birth, and date on mortality from malignant melanoma in the
populations of England and Wales, Canada, and the white population of the United
States.  Am J Epidemiol 110:734-739 (1979).

Lee, J.A.H., and Storer, B.E.  Further studies on skin melanomas apparently
dependent on female sex hormones.  Int J Epidemiol 11:127-131 (1982).

Lee, J.A.H., and Strickland, D.  Malignant melanoma:  Social status and outdoor
work.  Br J Cancer 41:757-763 (1980).

Lew, R.A., Sober, A.J., Cook, N. , Marvell, R., and Fitzpatrick, T.B.  Sun
exposure habits in patients with cutaneous melanoma:  A case control study.  J
Dermatol Surg Oncol 9:981-986 (1983).

Lewis, M.G.  Malignant melanoma in Uganda:  The relationship between
pigmentation and malignant melanoma on the soles of the feet.  Br J Cancer
21:483-495 (1967).

Lilienfeld, A.M., and Lilienfeld, D.E.  Foundations of Epidemiology.  New York:
Oxford University Press (1980).

Little, J.H., Holt, J., and Davis, N.  Changing epidemiology of malignant
melanoma in Queensland.  Med J Aust 1:66-69  (1980)

MacDonald, E.J.   Incidence and epidemiology of melanoma in Texas.  In:
Neoplasms of the Skin and Malignant Melanoma. 20th Annual Clinical Conference on
Cancer, 1975.  Chicago: Year Book Medical Publishers, Inc. (1976).

Mackie, R.M., and Aitchinson, T.  Severe sunburn and subsequent risk of primary
cutaneous malignant melanoma in Scotland.  Br J Cancer 46:955-950 (1982).

-------
                                      8-48
Magnus, K.  Incidence of malignant melanoma of the skin in Norway,  1955-1970:
Variation in time -,nd space and solar radiat/on.   Cancer 32(5) : 1275-1286
(1973).

Magnus, K.  Epidemiology of malignant melanoma of the skin in Norway with
special reference to the effect of solar radiation.  In:  Biological
Characterization of Human Tumors, pp. 249-259, Exerpta Medica International
Congress Series No. 375, Amsterdam (1975).
         ^
Magnus, K.  Incidence of malignant melanoma of the skin in the five Nordic
countries:  Significance of solar radiation.  Int J Cancer 20:477-485 (1977).

Magnus, K.  Incidence trends in the Nordic countries:  Effects of sun exposure.
In:   Trends in Cancer Incidence. Causes and Practical Implications (K.  Magnus,
ed.), Washington:  Hemisphere Publication, (1981a).

Magnus, K.  Habits of sun exposure and risk of malignant melanoma:   An analysis
of incidence rates in Norway 1955-1977 by cohort,  sex, age, and primary tumor
site.  Cancer 48:2329-2335 (1981b).

Magnus, K.  Incidence Trends in the Nordic countries:  effects of sun exposure.
In:   Trends in cancer incidence causes and practical implicauions.   K.  Mangus
(ed).  Hemisphere Publishing Corp. pp 387-392 (1982).

Malec, E., and Eklund, G.  The changing incidence of malignant melanoma of the
skin in Sweden, 1959-1968.  Scand J Plast Reconstr Surg 12:19-27 (1978).

Malec, E., Eklund, G., and Langerlof, B.  Reappraisal of malignant melanoma
diagnosis in the Swedish Cancer Registry.  Acta Path Microbiol Immunol Scand 85:
707-712 (1977).

Mason, T., and McKay, F.  U.S. Cancer Mortality by County, DHEW Publ. No.  (NIH)
74-615, Govt. Printing Office, Washington, D.C. (1973).

McCarthy, W.H., Blach, A.L., and Milton, G.W.  Melanoma in New South Wales:  An
epidemiologic survey 1970-76.  Cancer 46:427-432  (1980).

McGovern, V.J., Mihm, M.C., Bailly, C., Booth, J.C., Clark, W.H., Jr. Cochran,
A.J., Hardy, E.G., Hicks, J.D., Levene, A., Lewis, M.G., Little, J.H.,  and
Milton, G.W.  The classification of malignant melanoma and its histologic
reporting.  Cancer 32:1446-1457  (1973).

McKay, F.W., Hanson, M.R., and Miller, R.W.  Cancer mortality  in the United
States:  1950-1977.  Natl Cancer Inst Monog 59 (1982).

Moss, A.L.H.  Malignant melanoma in Maoris and Polynesians in  New Zealand.  Br J
Plastic Surg 37:73-75 (1984).

Movshovitz, M., and Modan, B.  Role of sun exposure  in  the etiology of malignant
melanoma:  Epidemiologic  inference.  J Natl Cancer Inst 51(3):777-779 (1973).

-------
                                      8-49
Muir, C.S., and Nectoux, J.  Time trends:  Malignant melanoma of the skin.  In:
Trends in Cancer Incidence. Magnus, K. (ed).   Washington:  Hemisphere Publishing
Corp. pp. 365-385 (1982).

NAS.   Protection Against Depletion of Stratospheric Ozone by
Chlorofluorocarbons. Washington:  National Academy of Sciences (1979).

NAS.   Causes and Effects of Stratospheric Ozone Reduction:  An Update.  National
Research Council, Washington:  National Academy of Sciences (1982).

National Institute  for Occupational Safety and Health (NIOSH).  Unpublished
data:  melanoma after exposure to PCBS (1976),

Nordlund, J.J.  Genetic basis of pigmentation and the disorders of pigmentation.
In:  Pathology of Malignant Melanoma. Ackerman, A.B. (ed).  New York:  Masson
Publishing pp. 23-45 (1981).

Osterlind, A., and  Jensen O.M.  Trends in incidence of malignant melanoma of the
skin in Denmark 1943-1982.  Recent Results Cancer Res 102:8-17 (1986).

Paffenbarger, R.S., Wing, A.L., and Hyde, R.T.  Characteristics in youth
predictive of adult-onset malignant lymphomas, melanomas, and leukemias:   Brief
communication.  J Natl Cancer Inst 60:89-92 (1978).

Pakkanen, M.  Clinical appearance and treatment of malignant melanoma of the
skin.  Ann Chir Gynaecol 66:21-30 (1977).

Pasternack, B.S., Durbin, N., and Moseson, M.   Malignant melanoma and exposure
to fluorescent lighting at work.  Lancet 1:704 (1983).

Pathak, D.R., Samet, J.M., Howard, C.A.,  and Key, C.R.  Malignant melanoma of
the skin in New Mexico 1969-1977.  Cancer 50:1440-46 (1982).

Pitcher, H.M.  Examination of the empirical relationship between melanoma death
rates in the United States 1950-1979 and satellite-based estimates of exposure
to ultraviolet radiation.  In press (1987a).

Pitcher, H.M.  Personal communication (1987b).

Pitcher H.M. and Longstreth, J.D.  Ultraviolet Radiation.

Pondes, S., Hunter, J.A.A., White, H., Mclntyre,  M.A., and Prescott, R.J.
Cutaneous malignant melanoma in south-east Scotland.  Q J Med, New Series L.
197:103-121 (1981).

Riggan, W.B., Briggan,  J.B., Acquavella,  J.F., Beaubier, J., and Mason, T.J.
U.S.  Cancer Mortality Rates and Trends,  1950-1979.  USEPA (1983).

Rigel, D.S. Fredman, R.J., Bernstein, M.  and Greenwald,  D.J.  Letter to the
editor:  Malignant melanoma and exposure to fluorescent lighting at work.
Lancet 2:704 (1983).

-------
                                      8-50
Rippey, J.J., and Rippey, E.  Epidemiology of malignant melanoma of the skin in
South Africa.  SA Med J 65:595-598 (1984)

Romsdahl, M.M.,  and Cox, I.S.  Biological aspects of pigment cells and malignant
melanoma.  In:  Neoplasms of the Skin and Malignant Melanoma. 20th Annual
Clinical Conference on Cancer, 1975.   Chicago:  Year Book Medical Publishers
(1976).

Roush, G.C., Schymura, M.J., and Holford, T.R.  Risk for cutaneous melanoma in
recent Connecticut birth cohorts.  Am J Public Health 75:679-682 (1985a).

Roush, G.C., Schymura, M.J., Holford, T.R.,  White, C.,  and Flannery, J.T.   Time
period compared to birth cohort in Connecticut incidence rates for twenty-five
malignant neoplasms.  JNCI 74:779-788 (1985b).

Rushton, L., Alderson, M.R.  An epidemiological survey of eight oil refineries
in Britain.  Br J Ind Med 38:225-234 (1981).

Scotto, J., and Fears, T.R.  The association of solar ultraviolet radiation and
skin melanoma among Caucasians in the United States.  Cancer Invest (in press
1986).

Scotto, J., and Nam, J.  Skin melanoma and seasonal patterns.  Am J Epidemiol
111(3):309-314 (1980).

Schultze, R.  Increase of carcinogenic ultraviolet radiation due to reduction in
ozone concentration in the atmosphere.  Proceedings of the International
Conference on Structure.  Compos'ition and General Circulation of the Upper and
Lower Atmospheres and Possible Anthropogenic Perturbations.  Atmospheric
Environment Service, Ontario, Canda,  Volume 1:479-493 (1974).

Serafino, G., and J. Frederick.  Global modeling of the ultraviolet solar flux
incident on the biosphere, prepared for the U.S. Environmental Protection
Agency, Washington, D.C., (1986).

Setlow, R.B.  The wavelengths in sunlight effective in producing skin cancer:  A
theoretical analysis.  Proc Natl Acad Sci USA 71(9):  3363-3366 (1974).

Smith, J.L.  Histopathology and biologic behavior of malignant melanoma.  In:
Neoplasms of the Skin and Malignant Melanoma.  Chicago:  Year Book Medical
Publishers, Inc., pp. 293-330 ( 1976).

Sondik, E., Young, J.L., Horm, J.W.,  and Ries, L.A.G.  1985 Annual Cancer
Statistics Review.  National Cancer Institute (1985).

Stevens, R.G., and Moolgavkar, S.H.  Malignant melanoma:  Dependence of
site-specific risk on age.  Am J Epidemiol 119:890-895 (1984).

Swerdlow, A.J.  Incidence of malignant melanoma of the skin  in England and Wales
and its relationship to sunshine.  Br Med J 2:1324-1327 (1979).

-------
                                      8-51
Takahashi, M.,  and Seiji, M.  Acral melanoma in Japan.  Pigment cell, 6:150-165
(1983).

Teppo, L.  Incidence trends of melanoma of the skin in Finland by histological
type.  In:  Trends in Cancer Incidence:  Causes and Practical Implications.  K.
Magnus (ed.), Hemisphere Publishing Corp., pp. 693-697 (1982).

Teppo, L. , Pakkanen, M.,  and  Hakulinen, T.  Sunlight as a risk factor of
malignant melanoma of the skin.  Cancer 41:2018-2027 (1978).

Teppo, L. , Pukkala, E., Hakama, M.,  et al.  Way of life and cancer incidence in
Finland.   Scand J Soc Med (Suppl.) 19:50-54 (1980).  As cited in Lee (1982).

TNCS.  Third National Cancer Survey:  Incidence Data.  DREW Publication No.
(NIH) 75-787, Natl Cancer Inst Monogr 41 (1975).

Vagero, D., Ringback, G., and Kiviranta, H.  Melanoma and other tumours of the
skin among office, other indoor and outdoor workers in Sweden 1961-1979.  Br J
Cancer 53:507-512 (1986).

Venzon, D.J., and Moolgavkar, S.H.  Cohort analysis of malignant melanoma in
five countries.  Am J Epidemiol 119:62-70 (1984).

Viola, M.V., and Houghton A.N.  Solar radiation and cutaneous malignant
melanoma.  Hosp Pract,  June:97-106 (1982).

Wan, S.,  Anderson, R.R.,  and Parrish, J.A.  Analytical modeling for the optical
properties of the skin with in vitro and in vivo applications.  Photochem
Photobiol 34:493-499 (1981).

Wigle, D.T.  Malignant melanoma of skin and sunspot activity.  Lancet 2:38
(1978).

-------
                                   CHAPTER 9

                        UVR-INDUCED  IMMUNOSUPPRESSION:
                     CHARACTERISTICS AND POTENTIAL IMPACTS
SUMMARY

    Solar radiation has been found to have a detrimental effect on the immune
system of both humans and experimental animals.  In particular, UV radiation
(UVR) reduces the ability of the cell-mediated arm of the immune response to
respond adequately to antigens.  UV irradiation of the skin predisposes the
immune system to form suppressor T cells (Ts) in response to antigens
encountered via the skin or injected subcutaneously.  These Ts are specific for
antigen and prevent the development of immune responses.   In a tumor bearing
animal, this immunosuppresslon may result in the out growth of tumor cells which
in normal animals would have been destroyed by the immune system.

    One cell population in the skin known to be severely damaged are the
Langerhans cells.  Langerhans cells are skin-resident, antigen-processing cells
which are required to induce cell mediated or antibody-based immune responses to
antigens which are introduced through the skin.  Although it is not certain that
the damage to the Langerhans cells is entirely responsible for the
immunosuppressive effect of UVR, it is clear that UV irradiation of skin reduces
the immune response in that skin.  Even more important, it is also clear that it
is the UV-B portion of the UV spectrum that is responsible for the depression of
the immune response.

    One important facet of this UVR-induced immunosuppression is the role it
plays in UV-B-induced carcinogenesis;  it affects the host's ability to respond
to tumor-specific antigens present on the UVR-induced tumor cells thereby
increasing the risk for tumor development in animal models, probably by allowing
the tumor to escape the normal immune surveillance mechanisms.  Thus the tumor
cells are allowed to divide and establish a large and growing tumor in the host.

    The immunosuppressive effects of UVR also are very likely to have a
deleterious effect on the immune response to those infectious diseases that
enter through the skin, especially if the initial immune response to the agent
takes place in the skin.   As yet, little research has been done in this area.
Preliminary evidence indicates, however, that UV irradiation during a first
cutaneous infection with two very different organisms, the parasite Leishmania
sp. and Herpes simplex virus,  may result in an impairment of the immune response
of the host to subsequent infections.   In the case of leishmaniasis,  this could
lead to the development of the more lethal form of the disease, visceral
leishmaniasis.

    Although there are no experimental data which specifically address these
issues several hypotheses about UVR induced immunosuppression seem reasonable:

        1)  All populations,  black and white may be at risk,
            particularly if the mechanism of induction for UVR
            induced immunosuppression lies above the melanin as one
            theory suggests.

-------
                                      9-2
        2)  Individuals who are already inununosuppressed, e.g.,
            transplant patients or individuals with immunodeficiency
            diseases,  could via additive effects be at greater risk
            than the normal population, and

        3)  In developing countries, particularly near the Equator,
            high solar insulation could result in an
            immunosuppression which could exacerbate parasitic
            infections of the skin such as leishmaniasis.

    The effects of UV-B and solar radiation on the human immune system have not
been studied in sufficient detail to allow estimation of dose-response
relationships for these effects.  Qualitatively, it is known from animal studies
that the doses of UV-B needed to induce immunosuppression are much lower than
those required for carcinogenesis.  This may mean that exposure to low doses of
UVR, even doses that do not cause a sunburn, may decrease the ability of the
human immune system to provide an effective defense against neoplastic skin
cells or skin infections.

-------
                                      9-3
FINDINGS

1.  UV-B SUPPRESSES THE IMMUNE SYSTEM IN ANIMAL EXPERIMENTS.

    la.  UV radiation administered at relatively low doses causes a depression
         in local contact hypersensitivity (a form of cell-mediated immunity)
         resulting from an inability to respond to an antigen presented through
         UV-irradiated skin

    Ib.  High doses of UV radiation cause a depression in systemic contact and
         delayed type hypersensitivity reactions,  that result in an inability of
         the animal to respond to an antigen which is presented to the animal
         through unirradiated skin

    Ic.  Both the local and systemic effects on contact hypersensitivity are
         mediated by a T suppressor cell which prevents the development of
         active immunity to the antigen

    Id.  The immunosuppressive effects of UVR have been found to reside almost
         entirely in the UV-B portion of the ground level solar radiation

2.  SUPPRESSION OF THE IMMUNE SYSTEM MAY PLAY AN IMPORTANT ROLE IN
    CARCINOGENESIS.

    2a.  Animals which are UV-irradiated develop T suppressor cells which
         interfere with the immune response to UV-induced tumors in such a way
         that the animals are more susceptible to the growth of autochthonous
         UV-induced tumors.  The contribution of the suppression of the immune
         system to cancer incidence that would result from ozone depletion is
         reflected in the dose-response estimates of photocarcinogenesis
         assuming that the action spectra for the two phenomena are the same.
         If these two impacts have different action spectra,  the estimates could
         be either high or low.

3.  LIMITED EXPERIMENTAL DATA INDICATE UV-B SUPPRESSES THE HUMAN IMMUNE SYSTEM.

    3a.  Although there is limited information about the effects of UV radiation
         on humans, several studies indicate that the immune response of humans
         is depressed by UV radiation and is depressed in UV-irradiated skin.

4.  UV-B-INDUCED SUPPRESSION OF THE HUMAN IMMUNE SYSTEM COULD HAVE A DELETERIOUS
    EFFECT WITH REGARD TO MANY HUMAN DISEASES.

    4a.  Preliminary studies indicate that UV radiation may prevent an effective
         immune response to micro-organisms that infect via the skin, thus
         predisposing to reexpression or chronic infection.

    4b.  Two human diseases that may be influenced by UV-B-induced immune
         suppression are herpes virus infections and leishmaniasis.

    4c.  Almost no research has been conducted on the influence of UV-B on other
         infectious diseases; additional investigation is clearly warranted.

    4d.  For at least one theory of the mechanisms of UV-B-induced suppression

-------
                                  9-4
     of the immune system (that involving urocanic acid),  a possibility
     exists that non-whites,  as well as whites,  would be  vulnerable to
     increased immune suppression caused by ozone depletion.

4e.   Because UV-B can produce systemic immunologic change, the possibility
     exists that changes in UV-B could have resulted in effects on diseases
     whose control requires systemic rather than local immunity.

4f.   Immunologic studies to date have not assessed the effects of long-
     term, low-dose UV-B irradiation.  Consequently, the  magnitude of this
     risk cannot be assessed.

-------
                                      9-5
                         UVR-INDUCED IMMUNOSUPPRESSION:
                      CHARACTERISTICS AND POTENTIAL IMPACTS
INTRODUCTION

    As indicated in Chapter 7,  it has been known for years that ultraviolet
radiation (UVR) is carcinogenic.  In the past decade or so, studies designed to
explore the mechanism of UVR-induced carcinogenesis revealed that UVR is a
potent but specific immunosuppressive agent which can abrogate the ability of an
animal to reject its own or a transplanted syngeneic UV-induced tumor.
Subsequent work investigating the mechanism of UVR-induced immunosuppression has
indicated that UVR can have a number of effects on the immune system, effects
which have the potential to modify a variety of disease processes.  This chapter
has four goals: 1) to briefly review some of the basic concepts in immunology
necessary to understand the effects of UVR on the immune system, 2)to provide an
introduction to that portion of the immune system first affected by solar
radiation - the skin-associated lymphoid tissues (SALT); 3) to review in some
detail what specifically is known about the impact of UVR on immune responses
including a discussion of dosimetry and action spectra and 4) to close with a
discussion of the potential impacts UVR-induced immunosuppression could have on
infectious diseases.

BASIC CONCEPTS IN IMMUNOLOGY

    A complete history of immunology is beyond the scope of this review; what
follows is a very cursory review which focuses solely on the key concepts
required for understanding the impact of ultraviolet radiation on the immune
system as it is currently understood.  The field is constantly changing however.
For entertaining and readable reviews of the immune response in general, the
reader is referred to two recent texts by Jan Klein (1982) and Bill Paul (1986).
For the most current information on the subject of photoimmunology the reader is
referred to the most recent issues of the journals Photoimmunology, and
Photobiology and Photochemistry and to recent reviews by Margaret Kripke (1984,
1986), Raymond Daynes and colleagues (1986) and Richard Granstein (1987).

    The immune system is a complex network of specialized organs and cells that
has evolved to defend the body against "foreign invaders" e.g., infectious
agents such as bacteria and viruses.  The immune system shows several remarkable
characteristics.  It is able to distinguish between "self" and "nonself".  It is
able to remember previous experiences and react accordingly.  It displays
enormous diversity and extraordinary specificity; not only is it able to
recognize many millions of "nonself" molecules, it can produce molecules to
match up with and counteract each one of them.

    The success of this system in defending the body depends on an incredibly
elaborate and dynamic regulatory communications network.  Millions and millions
of cells organized into sets and subsets pass information back and forth.  The
result is a sensitive system of checks and balances that produces an immune
response which is prompt, appropriate,  effective and self-limiting.

    By the 1950's, it was generally accepted by immunologists that there were
two types of immune responses:  those mediated through antibodies (humoral
responses) and those mediated by cells (cell-mediated responses).  However,

-------
                                      9-6
since the antibodies are produced by cells,  there was some question as to how
the cells involved in producing antibodies differed from those involved in cell
mediated responses.  Over the decade from mid 1950's to the mid 1960's, this
question was answered as the cells important to immune response were
characterized.   There are certain types of cells and anatomic structures which
are critical to the induction and elicitation of the immune response and which
are common to immunological phenomena throughout an organism.  Given below is a
brief introduction to key concepts and terminology.

    Antigens

    An antigen is any substance that is recognized as foreign ("nonself") by the
body's immune system.  A molecule is recognized as "self" or "nonself" by virtue
of unique structurally-conferred characteristics called epitopes.  A substance
such as a protein molecule, or an entity such as a bacterium or a virus particle
can display many different epitopes all of which can induce unique and specific
immune responses.  Certain (generally very small) epitopes termed "haptens" will
not alone induce an immune response but when displayed on a larger molecules,
termed "carriers", are able to induce immune responses specific for themselves
and not for the carrier molecule.

    Cell Types

    The key cell types involved in generating immune responses affected by UVR
are lymphocytes and antigen-presenting cells (APC's).   Lymphocytes are
generally considered to be the ultimate effector cells in the immune response;
they comprise a functionally heterogeneous population of two major lineages.  B-
lymphocytes are important to humoral immune (antibody) responses; their terminal
differentiation state is the antibody-exporting plasma cell.  T-lymphocytes are
important both to cell-mediated and humoral responses.  There are several
subsets of T-lymphocytes; T-helper/inducer cells provide help to B-lymphocytes
in the development of antibody responses, T-killer cells function as effector
cells in cytotoxicity (cell killing) responses and T-suppressor cells perform
regulatory functions, suppressing both humoral and cell-mediated responses.

    Upon exposure to antigen, the first step in the immune response is uptake of
antigen by cells capable of processing and/or otherwise focusing the antigen for
effective presentation to immunocompetent lymphocytes.  Cells which perform this
function are termed antigen-presenting cells and there is reason to believe that
they comprise a heterogeneous population that at a minimum contains a certain
subset of macrophages termed dendritic cells.  There is some uncertainty as
whether or not all macrophages process antigen; evidence is accumulating to
suggest that this function is restricted to the "dendritic" subpopulation, but
the final judgement is not yet in (Klein 1982) .

Local Versus Systemic Effects

    Much of what is known about immunology comes from experiments which
characterize systemic immune responses.  These are immune response demonstrable
throughout the body which occur following antigen processing/presentation by the
macrophages, recognition of processed/presented antigen by T-lymphocytes,
interaction of the T-lymphocyte antigen complex with B-lymphocytes followed by
clonal proliferation and differentiation of T and B lymphocytes.  These events
are usually considered to take place within the cortical region of lymph nodes

-------
                                      9-7
and the equivalent region of the splenic periarteriolar sheath.  The cellular
and molecular consequences of T cell activation (lymphokines,  cytotoxic cells,
cells that mediate delayed-type hypersensitivity,  and cells that regulate the
immune response) and of B cell activation (specific antibodies, memory B cells,
regulating anti-antibodies) then disseminate systemically through the lymphatics
and the blood.

    In contrast to systemic immune responses, there exist immune responses which
are localized to a particular microenvironment such as the gut, the respiratory
tract, the breast, external eye and the skin.  The lymphoid tissues responsible
for the immune responses observed in these various tissues have been identified
respectively as the gut associated lymphoid tissue (GALT), the bronchial
associated lymphoid tissue (BALT), the mammary associated lymphoid tissue (MALT)
the conjunctival associated lymphoid tissue (CALT) and the skin associated
lymphoid tissue (SALT).  In each case, the responses observed in each of these
sites are dependent on circuits of lymphocytes specific to the particular milieu
provided at these anatomic locations.

SALT: SKIN-ASSOCIATED LYMPHOID TISSUES

    Streilein and his colleagues (1983,1984) have proposed that the skin and its
component cell types comprise a unique regional sphere of immunologic influence
which serves a surveillance function for the skin allowing it to successfully
fulfil its role as the main barrier between the organism and the environment.
As indicated in chapter 7,the skin has several cell types which function in
immune responses including a rather unique version of an antigen-presenting cell
(the Langerhans cell), a series of T and probably non-T lymphocytes which can
effect immune responses and keratinocytes which produce a substance, epidermal
cell derived thymocyte activating factor (ETAF),  that may also affect immune
responses in the skin.

    On virtually a daily basis, the skin protects the body from a plethora of
infectious agents such as bacteria, fungi and viruses, as well as to numerous
chemical arid physical carcinogens the most prevalent one being solar radiation.
The daily bombardment by the UVR present in solar radiation poses the risk of
neoplastic transformation of epidermal cells; such transformation is generally
associated with the expression of unique new antigens on the surface of the
neoplastic cells.  Streilein and his colleagues (1983, 1984) have proposed that
SALT represents the physiologic mechanism created to deal with the special
demand this process places upon the skin.  The process whereby SALT controls
neoplasia might be stated as follows:  neoplastic transformation of keratinocyte
results in the display of new surface antigens.  The ubiquitous Langerhans cells
pick up or receive this new antigen, process it and present it in immunogenic
form.  One of two subsequent events may then transpire: wandering T lymphocytes
with predetermined affinity for the epidermis and with immunologic specificity
for the new antigen migrate into the epidermis, recognize the new antigen
presented on LC, transform into effector cells and directly destroy the adjacent
malignant cells, or 2) the Langerhans cell bearing the new antigen migrates via
the  draining lymph into the regional lymph node, there to attract antigen-
specific lymphocytes that transform in response to the antigen, proliferate, and
then disseminate systemically albeit returning predominantly to the skin in
order to effect destruction of the neoplastic keratinocytes.

-------
                                      9-8
    Evidence which Streilein and his colleagues (1983) feel support this
proposition include:  1) Some,  but not all,  T lymphocytes display special
affinity for skin and draining lymph nodes  (Rose et al. 1976). 2) Skin contains
immunocompetent lymphocytes (Parker and Billingham 1972) and antigen presenting
cells (Stingl et al.  1978)  and produces inununoregulatory molecules (Sauder et
al. 1981, 1982), 3) Immune recognition of antigen takes place within skin
(Streilein 1983);  4)  Antigen introduced via the skin which escapes recognition
induces specific unresponsiveness (Toews et al. 1980).

EFFECTS OF UVR ON IMMDNOLOGICAL REACTIONS

    The effects of UVR on the immune system were first described in
investigations of mechanism of carcinogenesis induced by chronic administration
of UVR.  Subsequent work explored the specific effects of UVR on the
immunological reactions - in the absence of carcinogenesis.  This section will
briefly review the initial work done to characterize the immunologic reactions
evaluated during the carcinogenic process (for a detailed review see Chapter 	
of Appendix A) and will then focus on the studies designed to explore the exact
impact of UVR on the immune response.

    Effects of UVR on Tumor Growth

    The first indication that UVR might have an impact on the immune system came
from studies by Kripke (1974) who reported that most murine tumors induced by
UVR were rejected when transplanted into normal syngeneic recipients whereas the
same tumors grew progressively in immunosuppressed mice.  This phenomenon was
originally identified in a C3H mouse system in which tumors were induced using a
high intensity mercury arc lamp (Kripke 1974), subsequent studies indicated that
it held true for other inbred strains of mice irradiated using FS40 fluorescent
sunlamps (Kripke 1977).

    Experiments designed to determine if the mechanism of this effect was that
UV irradiation of mice led to a generalized immunosuppression, evaluated a
variety of immune parameters using a number of different exposure regimens
(albeit all involving chronic administration)  (Kripke et al. 1977; Spellman et
al. 1977; Norbury et al. 1977; Lynch and Daynes 1983).  At the doses used to
induce carcinogenicity, most of the functions examined  (e.g., response to
mitogens, allograft rejection, antibody production) remained normal, however, in
some instances  (e.g., graft-versus-UVR-treated host reactions (Kripke et al.
1977), reactivity to dinitrochlorobenzene (DNCB) (Kripke et al.  1977), and
naturally occurring cell-mediated cytotoxicity in UV-irradiated  mice (Lynch and
Daynes 1983) were found to be transiently suppressed but returned to normal
after between 2 and 4 months of UVR treatment.

    It is clear from the above information that mice which have  been chronically
irradiated with UV light are not generally immunosuppressed but  have normal
immunity when tested over a very wide number of immune  functions.  Nevertheless
they are unable to reject UV-induced tumors.   Further  experiments revealed that
the mechanism for the  lack of immune reactivity to UV-induced tumors in
UV-irradiated mice was a radiosensitive, la positive, Lyt-1+2-  (Ullrich and
Kripke 1984) T  suppressor cell (Fisher and Kripke 1977, 1978, Spellman and
Daynes 1977, 1978,  Daynes et al. 1979).  These cells were specific for an
antigen  common  to UV-induced tumors and did not affect  the ability of immune
mice to  reject  tumors  as a function of immunity to tumor specific

-------
                                      9-9
transplantation antigens (Roberts et al 1980) or to histocompatibility antigens
(Kripke and Fisher 1976).  Preliminary evidence indicates that UV-irradiated
mice inununologically recognize UV-induced tumors (Romerdahl and Kripke, 1986),
but that the suppressor cell acts to prevent the development of cytotoxic
effector cells.

    Other characteristics of this response were 1) that with the exception of a
transplantable spontaneous melanoma (B-16), the immunosuppression was specific
for the transplantation of UVR induced tumors (Kripke et al 1979); 2)that it can
be induced with natural sunlight (as opposed to artificial sources) (Morison and
Kelley 1985); 3)that the active wavelengths lie below 315 nm; 4) that there is
dose reciprocity (i.e.,  reducing the dose rate but keeping the total dose
constant results in the same carcinogenic impact (De Fabo and Kripke 1979,
1980); 5)that immunosuppression, when transferred via lymphoid cells could
result in enhanced UVR-induced tumorigenesis (Fisher and Kripke 1982); and 6)
that the effect on tumorigenesis was systemic, i.e., tumor development was
enhanced on both exposed and non-exposed skin in irradiated animals (De Gruijl
and Van Der Leun 1982, 1983)

    Characterization of the Impacts of DVR on the Immune System

    In normal mice, application of a contact sensitizing agent such as
trinitrochlorobenzene (TNCB) or dinitrofluorobenezene (DNFB) to shaved abdominal
skin induces a cell-mediated immune response.  A second exposure to the contact
sensitizing agent via painting on the ears elicits a contact hypersensitivity
(CHS) response characterized by swelling of the ears 24 hours later.  The
response is specific to the initial sensitizing agent in that TNCB will not
sensitize for DNFB and vice versa.  A similar phenomenon, delayed type
hypersensitivity, differs from CHS only in the way in which the ultimate
response is elicited.   Whereas in CHS skin painting on the ear elicits the
ultimate response, in DTK injection of the sensitizing agent in
dimethylsulfoxide into the footpad or injection of haptenated splenic
lymphocytes into the footpad or ear elicits the ultimate response.

    The specific impacts of UVR on the immune system can generally be
characterized one of three ways: 1) suppression of contact hypersensitivity
(CHS) at the site of irradiation (also known as local suppression), 2) systemic
(generalized) suppression of contact hypersensitivity ,  and 3) systemic
suppression of delayed type hypersensitivity (DHS).  These three impacts have
different action spectra and dose requirements; what follows is a discussion of
the pertinent differences as well as an attempt to relate these differences to
possible impacts of UVR in human populations.

    Local suppression of contact hypersensitivity.  The first description of the
impact that ultraviolet radiation could have on contact hypersensitivity was
that of Haniszko and Suskind (1963) who demonstrated that, in guinea pigs, prior
exposure of skin to UV-B (280-320) resulted in a reduced CHS response.  The dose
given was approximately a minimum erythemal dose (MED)-1-.  Unfortunately, the
     1 In humans an MED from exposure to UV-B is generally achieved with about
20 to 70 mjoules/cm^ (equivalent to 200 to 700 joules/nr depending on location
and time of year (Pathak 1982); this is the amount that a fair-skinned
individual would receive in about 12 to 15 minutes on a clear day in June in

-------
                                      9-10
work of Haniszko and Suskind (1963) went unremarked for nearly 20 years until
the work of Toews and his colleagues (1980) confirmed the ability of UVR to
induce immunological tolerance to skin applied antigens and demonstrated that
the defect responsible for this was a local (i.e., occurred at the site of
irradiation and antigen administration) impact on the antigen-presenting
Langerhans cells.  It was considered to be a local impact in that in irradiated
animals the effect could only be demonstrated if the sensitizing agent were
applied to irradiated skin; application to unirradiated skin in irradiated
animals failed to show suppression of the CHS response.  Subsequent work showed
that the defect in antigen-presenting cell function was accompanied by the
development of hapten-specific suppressor T cells (Elmets et al 1983).   Further
investigation of this phenomenon by Elmets and his colleagues (Elmets et al.
1985) first demonstrated that mice exposed to 700 J/m  broad-band UV-B developed
a suppressed contact hypersensitivity to dinitrofluorobenzene (DNFB) and then
characterized the wavelength dependence of this response.  Exhibit 9.1 presents
the two action spectra developed in that work; one assumes single hit kinetics
and the other assumes multi-hit kinetics.  In the single hit case radiation
between 260-300 nm was most efficient in inducing immunological
unresponsiveness.  The dose of 260 nm energy to achieve a 50% inhibition (an
ED-^) of contact hypersensitivity was about half the dose of 270, 280 and 290 nm
energy and at 310 the ED50 was more than 30 times the ED50 at 260 nm.  Work by
Noonan and her colleagues  (1984) on the impact of UVR on the number and function
of Langerhans cells appears to show a similar wavelength dependency; (and as
discussed in some detail below) these authors were able to distinguish a
different wavelength dependence for systemic suppression of contact
hypersensitivity.

    Systemic suppression of contact hypersensitivity.  The amounts of broad-band
UV-B used to induce local unresponsiveness in the experiments described above
were in the range of a human MED (400 to 700 J/nr) whereas when narrowband
energy was used an ED50 could be achieved at doses as low as 200 J/m .   When
much larger doses of broad-band UV-B are used (£050 of around 30 kJ/m  , Noonan
et al 1981a) a systemic suppression of contact hypersensitivity is induced.  The
action spectrum  (Exhibit 9-2) for systemic suppression of CHS is quite different
from that found for local  suppression (Exhibit 9-1).  In the Noonan et al.
(1984) study, these authors were able to demonstrate that 320 nm radiation was
very much more efficient at systemic suppression as compared to its impact on
the number and function of Langerhans cells whereas the converse was true for
270 nm radiation. In an earlier report (from which the action spectrum was
derived) De Fabo and Noonan (1983) point out that the action spectrum they
derived for systemic suppression of contact hypersensitivity is extremely
similar to the absorption  spectrum for urocanic acid (UCA), an amino acid found
principally in the stratum corneum above most of the skin's protective melanin
layer.  UCA has  some interesting properties in that it is a photoreceptor which
when it absorbs a photon undergoes a cis to trans isomerization and
concomitantly becomes much more soluble thus allowing its migration out of the
stratum corneum.  In its cis form it is virtually absolutely restricted (in the
skin) to the stratum corneum. The correspondence of the UCA absorption spectrum
with that of systemic suppression led De Fabo and Noonan to propose that UCA has
a role in the induction of that suppression.  At that time they proposed that
Philadelphia  (Forbes 1987). Note however,  that  the guinea pig is about 2 to 3
times less sensitive than the human  (Forbes  1987).

-------
                                       9-11
                                   EXHIBIT 9-1

      ACTION SPECTRA FOR LOCAL  SUPPRESSION OF CONTACT HYPERSENSITIVITY
                ASSUMING EITHER ONE-HIT  OR MULTI-HIT MECHANISMS
               UJ
               cc
               o
               \u
               at

               I
               S
               K
               Ul
               0>
S
111
o


                  -10
                     .-5
                      260   270  280   290   300   310   320
                                  WAVELENGTH (NK<)
                      Action spectrum for the induction of unrespon-
             siveness to DNFB in C3H mice by in vivo low dose UV
             radiation. The solid line represents the action spectrum
             curve if single hit kinetics are observed. The dashed line
             represents the action spectrum curve if multiple component
                            kinetics are observed.
Source:   Elmets et al  1985.

-------
                                            9-12
                                        EXHIBIT 9-2

                ACTION SPECTRUM FOR SYSTEMIC SUPPRESSION OF CONTACT
                                     HYPERSENSITIVITY
                10-1 :
z
JJ
cfl
            a
            8
            o
            a
            UJ  10-2;
               10-3:
               10-4
                                  WAVELENGTH (nm)
           In vi\o action spectrum for the induction of systemic suppression of CHS (•).
Points on the action spectrum represent the reciprot.il of the number of photons (mj/photon)
ir<|uired to produce 509?  suppression, normalised m 'J70 nm. Bars represent ±1 SEM  For
purposes of comparison the DN A action spectrum (—) (54) and the absorption spectrum of
urocanic jcid (LCA) (O)
 Source:   De Fabo and Noonan 1983.

-------
                                      9-13
UCA was in some way responsible for the antigen presenting cell defect.  The
fact that they later discovered that the two phenomena had different action
spectra led them to discard this notion with regard to the local effect observed
on APC; however, this hypothesis is still viable with regard to systemic
immunosuppression.

    A role for UCA in systemic immunosuppression has as a corollary the fact
that UVR-induced immunosuppression may not be restricted to the same fair-
skinned individuals who show the greatest sensitivity to UVR-associated skin
cancer.  This is because UCA's position in the stratum corneum puts it above
most of the skin's melanin so that UCA will not be protected from incoming
photons by melanin the way that keratinocytes and melanocytes are.  In addition,
if UCA is important to systemic immunosuppression, then use of sunscreens (which
tend to be absorbed into the epidermis rather than staying localized in the
stratum corneum) might not be protective against all of the immunosuppressive
effects of UV radiation.  [As discussed in detail below, there is some evidence
that sunscreens may not be protective for the adverse impacts of UVR on cells of
the human immune response.]

    Systemic suppression of delayed type hypersensitivitry.  Because of the
similarities between DTK and CHS, most authors have assumed that these two
reactions are equivalent and thus that the impact of UVR on these responses
would be the same.  Recently however, work by Kripke and Morison (1986a,b)
provides some indication that this is not the case but rather that UVR impacts
these two reactions quite differently.  Using three one-day exposures  (total
dose of 48.6 kJ/m^) these authors found that although suppression of DHS could
be overcome by immunizing with hapten-coupled antigen presenting cells from
normal donors this was not the case for suppression of CHS, treatment of UV-B
irradiated mice with an anti-inflammatory agent (methylprednisolone) before
immunization prevented the suppression of delayed hypersensitivity but had no
effect on contact hypersensitivity, and that the hyporesponsiveness to CHS in
UV-B-irradiated mice could be transferred to x-irradiated mice via the use
spleen cells from UV-B irradiated mice, although the induction of hapten
specific suppressor cells requires both UV-B irradiation and priming with
hapten.  Based on their findings, these authors postulated that UV-B irradiation
produces a population of suppressor inducer cells with specificity for a
modified skin antigen which can serve as a carrier molecule for haptens that
induce contact hypersensitivity and for tumor- specific transplantation antigens
on UV-B-induced tumors.

    Much lower doses have been used by another group to suppress the DTH
response to herpes simplex virus (HSV) (Howie et al. 1986a,b).  This group used
a single dose of 960 J/m^ to suppress the systemic DTH response to HSV.  Timing
of the administration was crucial in that maximal suppression occurred if
irradiation was given 2 to 3 days prior to sensitization with the virus, was
less if given 5 or 7 days prior to sensitization and did not occur if given 14
days before, on the day of, or 3 days after sensitization.  These authors
further characterized this immunosuppression to show that two types of
suppressor cells were involved and that it was the efferent arm (i.e.,
development of effector cells) of the immune response which was affected.
Unfortunately, neither a dose-response nor an action spectrum for this impact on
the DTH response has yet been characterized.

-------
                                      9-14
    In summary there appear to be at least two,  and possibly three,  distinct
ways in which UVR may affect the immune system.   There clearly exists a low
dose, shorter UV-B waveband-dependent suppression of local hypersensitivity
which has an action spectrum consistent with that described for local effects on
Langerhans cells and there is as well a high dose,  slightly longer wavelength,
UV-B-dependent process which induces systemic immunosuppression  of contact
hypersensitivity and has an action spectrum similar to the absorption spectrum
of UCA.  In addition, there appears to be a high dose systemic immunosuppression
of delayed type hypersensitivity which requires  further investigation in order
to characterize the dose-response and action spectrum and determine if the DTK
impact is a unique phenomenon.

HUMAN STUDIES

    There have been fewer studies of the effects of UVR on the immune response
in humans.  Morison et al (1979) irradiated human volunteers with 1.5 minimal
erythemal doses (MED) to produce what they termed asymptomatic erythema or 3 MED
to produce what they termed symptomatic erythema.  Post irradiation, all
subjects showed a significant increase in the proportion of circulating
polymorphonuclear leukocytes and a corresponding increase in the proportion of
circulating lymphocytes.   Subjects who received 3 MED also had a significant
decrease in the proportion of circulating E-rosette forming cells, i.e. T cells,
and an increase in the proportion of null cells.  The changes in the absolute
numbers of these cells were not significant.  Thirty minutes and 4 hours after
UV-irradiation, the response of lymphocytes to PHA was increased, then
decreased, reaching a minimum 12 hours after UVR, and returning to normal 72
hours post exposure.

    Hersey et al (1983a) reported on the effects of solarium exposure on normal
human subj ects.  Normal volunteers were given a standard course of treatment in
a solarium to acquire a suntan; twelve 1/2 hr exposures.  The energy delivered
was 10 J/m2 for UV-A and about 1% of total irradiation was UV-B (which was not
quantitated).  Immune functions were analyzed before, on completion, and 2 weeks
after the end of irradiation.  The test subjects had reduced skin test responses
to DNCB and slightly reduced numbers of blood lymphocytes.  There was a relative
increase in T (OKT3+) cells which was attributable to an increase of OKT8+
(suppressor) cells.  There was a significant increase in suppressor T-cell
activity against pokeweed mitogen induced IgG production in vitro and a
depression of NK activity.  The changes were still present in some subjects 2
weeks after solarium exposure.  The NK data is a little uncertain since the
control subjects had lower levels of NK activity throughout the studies than the
UV-irradiated subjects.  Some patients had increased NK activity after treatment
and some did not change.  In a second study (Hersey, 1983b), normal human
volunteers were exposed to sunlight for 12 days in a 2 week period.  Tests of
immune function were made before, on completion and 2 weeks after the end of the
UVR.  There were 15 UV-irradiated subjects and 13 age- and sex-matched controls.
There were no significant changes in hemoglobin levels, total leukocyte count,
and lymphocyte, or neutrophil counts. In test subjects there was a small but
significant drop in T cell numbers which returned to normal by 2 weeks.  There
was, however, a marked increase in the ratio of suppressor cells to cytotoxic
helper cells which had not returned to normal at 2 weeks after UV-irradiation.
After UVR there was an increase in the activity of pokeweed mitogen-induced
gamma radiation-sensitive suppressor T cell activity against IgG and IgM
production.  Although immunoglobulin production  in vitro has too wide a range

-------
                                      9-15
for statistical analysis of the small numbers of patients to be of value, gamma
irradiation of T cells prior to culture increased the amount of immunoglobulin
produced in patients exposed to sunlight but not in controls.  There were 5
persons who did not show these changes, but they were largely accounted for by
persons who did not receive the full dose of UVR.  Immunoglobulin levels in vivo
did not change nor did NK activity change significantly in this study.  In a
very recent study, Hersey et al. (1987) confirmed the suppression of NK activity
following exposure to radiation from solarium lamps (12 1/2 hr exposures; 18
J/m2 per exposure equivalent to 185 mJ/m2 of UV-B and 17.8 J/m2 of UV-A) with
the greatest depression demonstrated against a melanoma target. In the same
experiments, designed to evaluate the effectiveness of a sunscreen, these
authors were surprised to discover that use of a sunscreen did not prevent
depression of NK activity.  Although, as the authors point out, this observation
could be due to the fact that the dose of UV-A delivered was sufficient to cause
the effect, they also suggest that caution should be exercised in the use of
such agents.  Until the effectiveness of sunscreens in preventing systemic
effects of UVR on the immune system is more clearly established,  it should be
considered that the ability of such agents to prevent erythemal responses may
encourage longer exposure to solar radiation and hence lead to greater damage to
those elements of the immune system which function in immune surveillance
against melanoma.

    O'Dell et al (1980) reported that there was a diminished immune response in
sun-damaged skin.  The concentration of DNCB required to elicit a positive patch
test was greater in sun-damaged skin than in skin which was normal.  In
addition, the delayed-type hypersensitivity to intradermal injection of Candida,
mumps, and PPD antigens was decreased in sun-damaged skin so that the
differences were not due to a difference in percutaneous absorption of antigen
through sun-damaged skin.  The response to a primary irritant was the same in
both sun-damaged skin and normal skin and there was no difference in the two
tested sites (back of the neck and the back) in volunteers without sun-damaged
skin at the back of the neck.  Therefore, there is apparently a local
suppression of contact hypersensitivity in sun damaged skin.

    There is also evidence that the inability to repair UV-induced damage may
also play a role in impairment of immune function.  Patients with xeroderma
pigmentosum, who suffer from sun sensitivity and increased numbers of
sunlight-induced cancers, also have been shown to have impaired ability to
remove and repair lesions in their DNA caused by UVR.   Morison et al. (1985)
reported that xeroderma pigmentosum patients had impaired ability to develop
contact hypersensitivity to DNCB in sun-exposed skin when compared to normal
control subjects.  The numbers and morphology of Langerhans cells in the skin
were the same in both groups.  Unfortunately, the study did not include an
assessment of ability to develop contact hypersensitivity in skin not usually
exposed to sunlight.

EFFECTS OF ULTRAVIOLET RADIATION ON INFECTIOUS DISEASES

    Upon the realization that irradiation with UV-B can result in both local and
systemic suppression of contact hypersensitivity, concern developed that the
immune response to diseases which have a cutaneous phase might be adversely
affected by UV irradiation.  This was especially true since only low doses of UV
irradiation had been found to be necessary to suppress an immune response to
antigens applied to UV-irradiated skin.  As yet, not a great deal of research

-------
                                     9-16
has been published on this problem,  although there are at least two infectious
diseases,  Herpes virus infections and cutaneous leishmaniasis in which research
has begun to indicate that UVR-induced immunosuppression may play a role.

    It has been reported for some time that exacerbation or recurrences of
herpes simplex virus infections can be caused by sun exposure.  An exposure of 5
to 6 MED has been reported to cause recurrence of oral herpes (Spruance 1985)
and Morison (1984) noted that it is common to report exacerbation following sun
exposure or UV-B phototherapy although the mechanism is unknown.  In fact,
exacerbation of genital herpes has also been reported to occur following
exposure to the penis to experimental UVR (Klein and Linnemann, Jr. 1986) and
there is a. very recent report (Perna et al.  1987) that administration of UV-B at
three times an individual's previously determined MED to patients infected with
herpes simplex virus can reactivate the infection at sites occurring close to or
within the erythema induced by irradiation.

    Although, the mechanism of the activation is unknown, recent investigations
on the effects of UV radiation on primary herpes infection in mice are yielding
very interesting information.  Low doses of UVR from sunlamps given 7 days prior
to infection of mice with herpes simplex virus suppressed the delayed-type
hypersensitivity to the virus (Howie et al.  1986a).   Once it was induced, immune
suppression to the virus persisted for at least 3 months after the UVR.  The
suppression has been shown to be due to two different subsets of T lymphocytes
(Howie et al. 1986b).  Thus in this animal model, exposure to UV radiation at an
appropriate time prior to infection, suppressed the immune response made to that
organism for some time and may have profound effects on the outcome of the
infection.

    The observations of Howie et al. (1986a,b) particularly in light of the
findings of Perna et al. (1987), are of particular importance in the context of
the potential role that UV-B-induced immunosuppression might play in infectious
diseases.   The dose given by Howie and her colleagues was only slightly in
excess of a human MED yet it was sufficient to ensure that rather than
developing immunity to HSV, mice became tolerant to the virus.  In addition, the
dose necessary to induce recurrence in the study by Perna et al. was about
equivalent to a moderately severe sunburn and recurrence occurred in 3 to 6 days
- a timespan not that difference from the one required in order to ensure
immunosuppression (in the Howie study the virus had to be given 3 to 7 days
post-irradiation in order to observe any immunosuppression.)  Conceivably a
moderate sunburn could not only induce recurrences of herpes infections but at
the same time induce the development of suppressor cells thus preventing the
host from making an efficient immune response.  A similar scenario applied to
immunization programs might indicate that immunizations given 3 to 7 days post
an intense solar exposure could induce tolerance rather than immunity.

    In addition to the impact of UVR on herpes infections, there exists another
system where preliminary data suggest an impact of UVR on the progression on an
infectious disease.  Giannini  (1986) has reported that suberythematous doses of
UV-B blocked the development of the initial skin lesions in experimental
leishmaniasis in mice.  In addition, much as Howie et al. (1985) showed with
herpes, the UV-B irradiation also abrogated the development of  delayed type
hypersensitivity to the leishmania antigens, so that animals which would
normally be  immune following primary infection were made tolerant.   In a very
recent report  (Giannini and De Fabo 1987), further investigation of  the  impact

-------
                                      9-17
of UV-B irradiation on leishinaniasis in mice revealed that the waveband which
was most active in blocking the development of skin lesions (i.e.,  in modifying
the pathogenesis of a primary infections) was longer length UV-B, 320 nm as
compared to 290 nm.  What was more interesting, however, was the observation
that the waveband which was most effective at affecting immunity was 290 nm (as
contrasted to 320 nm) and the effect (albeit looked for in only a small number
of animals) was increased parasite dissemination.  Thus UV-B irradiation
concommitant with an animal's first encounter with Leishmania appears to result
in hosts that are non-immune (and possibly even tolerant), heavily parasitized
yet free of the skin ulcers which normally announce such infections.  Infection
of humans through recently UV-irradiated skin might suppress the immune response
to leishmanial antigens following a primary infection and thus predispose to a
more chronic, or more severe disease in subsequent exposures.

    The data reviewed above indicate that UV-B is immunosuppressive both locally
and systemically for antigens introduced via the skin either by the epicutaneous
or subcutaneous routes.  There are as yet no data to indicate that UV-B
treatment abrogates immunity to an antigen first introduced by a route other
than the skin.  The immunosuppression is for cell-mediated rather than humoral
responses and is associated with the appearance of suppressor T cells specific
for the skin-introduced antigens.  UV-B-induced immunosuppression has been
demonstrated for tumor, protein, parasite, viral, cellular and contact-
sensitizing antigens.  In the tumor system, the immunosuppression significantly
contributes to tumorigenesis; however, very little is known about the
contribution of UV-B-induced immunosuppression to the pathogenesis of infectious
disease,  In leishmaniasis, irradiation by UV-B results in less severe skin
lesions in primary infections; preliminary data suggest that this is at a cost
of depressed resistance to subsequent infection.

    There are a number of questions which still need to be answered.  For
instance, is UV-B induced suppression limited to antigens administered via the
skin?  If an antigen is administered via skin capillaries (e.g., as by mosquito)
is that more like being administered via cutaneous administration or intravenous
(systemic) administration?  Other questions include:  1) Does UV-B induced
immunosuppression increase the incidence or severity of infectious diseases?  2)
What are the long term consequences of low level UV exposures on the immune
system?  3) Does UV-B affect ocular immunity?  4) Are IgE responses suppressed
by UV-B irradiation?  And, lastly, 5) Can UV-B treatment modify an already
established secondary response?  If for instance, an individual is immunized as
a child, can a subsequent booster shot given after a sizable dose of UV-B
convert an immune state to tolerance?

-------
                                      9-18
                                   REFERENCES

Barker, C.F. and Billingham, R.E. Immunologically competent passenger cells in
mouse skin. Transplantation 14:525-527 (1972)

Daynes, R.A.,  Samlowski, W.E.,  Burnham, D.K., Gahring, L.C., and Roberts, L.K.
Immunobiological consequences of acute and chronic UV exposure. Curr Probl Derm
15:174-194  (1986)

Daynes, kR.A.,  Schmitt, M.K., Roberts, L.K.,and Spellman, C.W. Phenotypic and
physical characteristics of the lymphoid cells involved in the immunity to
syngeneic UV-induced tumors.  J Immunol 122:2458-2464 (1979)

Daynes, R.A.,  Spangrude, G.J.,  Roberts, L.K., and Krueger, G.G.  Regulation by
the skin of lymphoid cell recirculation and localization properties.  J Invest
Derm 85:14s-20s (1985)

De Fabo, E.G., and Kripke, M.L.  Dose-response characteristics of immunologic
unresponsiveness to UV-induced tumors produced by UV-irradiation of mice.
Photochem Photobiol 30:385-390 (1979)

De Fabo, E.G., and Kripke, M.L.  Wavelength dependence and dose-rate
independence of UV radiation-induced immunologic unresponsiveness of mice to a
UV-induced fibrosarcoma. Photochem Photobiol 32:183-188 (1980)

De Fabo, E.G., and  Noonan, F.P.  Mechanism of immune suppression by ultraviolet
irradiation in vivo. I. Evidence for the existence of a unique photoreceptor in
skin and its role in photoimmunology.  J Exp Med 158:84-98 (1983)

De Gruijl, F.R., and Van Der Leun, J.C.  Systemic influence of pre-irradiation
of a limited skin area on UV-tumorigenesis.   Photochem Photobiol 35:379-383
(1982)

De Gruijl, F.R., Van Der Leun,  J.C.  Follow up on systemic influence of partial
pre-irradiation on UV-tumorigenesis.   Photochem Photobiol  38:381-383 (1983)

Elmets, C.A.,  Bergstresser, P.R., Tigelaar,  R.E., Wood, P.J., and Streilein,
J.W.  Analysis of the mechanism of unresponsiveness produced by haptens painted
on skin exposed to low dose ultraviolet radiation. J Exp Med 158:781-794 (1983)

Fisher, M.S.,  and Kripke, M.L.   Systemic alteration induced in mice by
ultraviolet radiation and its relationship to ultraviolet carcinogenesis. Proc
Natl Acad Sci USA. 74:1688-1692  (1977)

Fisher, M.S.,  and Kripke, M.L.   Further studies on the tumor-specific suppressor
cells induced by ultraviolet radiation.  J Immunol 121:1139-1144 (1978)

Fisher, M.S.,  and Kripke, M.L.   Suppressor T lymphocytes control the development
of primary skin cancers in ultraviolet-irradiated mice. Science 216:1133-1134
(1982)

-------
                                     9-19
Giannini,  M.S.H.   Suppression of pathogenesis in cutaneous leishmaniasis by UV-
irradiation.  Infect Iimnun 51:838-43 (1986a).

Giannini,  S.H., Effects of UV-B on infectious disease in:  Effects of Changes -
Stratospheric Ozone and Global Climate.   Titus,  T.G.  UNEP/EPA Conference
Proceedings, Vol 2, pp 101-112 (1986).

Giannini,  S.H., and De Fabo,  E.G.  Abrogation of skin lesions in cutaneous
leishmaniasis by ultraviolet B irradiation. In:  Leishmaniasis: The First
Centenary (1885-1985") New Strategies for Control. Hart, D.T. (ed.)
NATO ASI Series A: Life Sciences, London: Plenum Publishing Company Limited,
(1987) in press.

Granstein, R.D.  Photoimmunology.  In:  Dermatology in General Medicine. 3rd ed.
Fitzpatrick, T.B., Eisen, A.Z., Wolff,  K.,  Freedberg, I.M., and Austen, K.F.
(eds.) New York:  McGraw-Hill Book Co.,  pp 733-796 (1987)

Haniszko,  J., and Suskind, R.R.  The effect of ultraviolet radiation on
experimental cutaneous sensitization in guinea pigs.   J Invest Derm 40:183-191
(1963)

Hersey, P., Bradley, M., Hasic, E., Haran,  G.,  Edwards, A., and McCarthy, W.H.
Immunological effects of solarium exposure.  Lancet 1:545-548 (1983a)

Hersey, P., Haran, G., Hasic, E., and Edwards,  A.  Alteration of T cell subsets
and induction of  suppressor T cell activity in normal subjects after exposure to
sunlight. J Immunol 131:171-174  (1983b)

Hersey, P., MacDonald, M, Burns, C., Schibeci,  S., Matthews, H., and Wilkinson,
F.J. Analysis of  the effect of a sunscreen agent on the suppression of natural
killer cell activity induced in human subjects by radiation from solarium lamps.
J Invest Dermatol 88:271-276 (1987)

Howie, S., Norval, M. and Maingay J. Exposure to low-dose ultraviolet radiation
suppresses delayed-type hypersensitivity to herpes simplex virus in mice. J
Invest Dermatol 86:125-8  (1986a)

Howie, S.E., Norval, M, Maingay, J., and Ross J.A. Two phenotypically distinct T
cells  (Lyl+2- and Lyl-2+) are involved in ultraviolet-B light-induced
suppression of the efferent DTH  response to HSV-1 in vivo. Immunology 58:653-8
(1986b)

Kripke, M.L.  Antigenicity of murine skin  tumors induced by ultraviolet  light.
J Natl Cancer  Inst 53:1333-1336  (1974)

Kripke, M.L.   Target organ for a systemic  effect of ultraviolet radiation.
Photochem Photobiol  24:599-600  (1976)

Kripke, M.L.   Latency, histology,  and antigenicity of  tumors  induced by
ultraviolet  light in three inbred  mouse  strains.  Cancer Res  37:1395-1400  (1977)

-------
                                      9-20
Kripke, M.L., Immunological unresponsiveness induced by ultraviolet radiation.
Immunol Rev 80:87-102 (1984)

Kripke, M.L.  Photoimmunology: The first decade. Curr Probl Derm 15:164-175
(1986)

Kripke, M.L. and Fisher, M.S.  Immunologic parameters of ultraviolet
carcinogenesis.  J Natl Cancer Inst 57:211-215  (1976)

Kripke, M.L. and Morison W.L. Studies on the mechanism of systemic suppression
of contact hypersensitivity by UVB radiation. II. Differences in the suppression
of delayed and contact hypersensitivity in mice. J Invest Dermatol 86:543-9
(1986)

Kripke, M.L. and Morison W.L. Studies on the mechanism of systemic suppression
of contact hypersensitivity by ultraviolet B radiation.  Photodermatol 3:4-14
(1985)

Kripke, M.L., Lofgreen, J.S., Beard, J., Jessup, J.M., and Fisher, M.S.  In vivo
immune responses of mice during carcinogenesis by ultraviolet irradiation.  J
Natl Cancer Inst 59:1227-1230 (1977)

Kripke, M.L., Thorn, R.M., Lill, P.H., Civin, C.I., Pazmino, N.H., and Fisher,
M.S.  Further characterization of immunological unresponsiveness induced in mice
by ultraviolet radiation.  Growth and induction of non-ultraviolet-induced
tumors in ultraviolet-irradiated mice. Transplantation 28:212-217 (1979)

Luger, T.A., Stadler, B.M., Katz,  S.I. and Oppenheim, J.J. Epidermal
(keratinocyte)-derived thymocyte activating factor (ETAF). J Immunol 127:1493-
1498 (1981)

Lynch, D.H., and Daynes, R.A.  Evaluation of naturally occurring cell-mediated
cytotoxic activity in normal and UV-irradiated mice.  Transplantation 35:216-223
(1983)

Lynch, D.H., Gurish, M.F., and Daynes, R.A.  The effects of high-dose UV
exposure on murine Langerhans cell function at exposed and unexposed sites as
assessed using in vivo and in vitro assays.  J Invest Derm 81:336-341 (1983)

Lynch, D.H., Gurish, M.F., and Daynes, R.A.  The effects of ultraviolet
irradiation on the generation of antitumor cytotoxic effector cell responses in
vitro.  J Immunol 127:1163-1168 (1981)

Morison, W.L., and Kelley, S.P.   Sunlight suppressing rejection of 280- to
320-nm UV-radiation-induced skin tumors in mice.  JNCI 74:525-527 (1985)

Noonan, F.P., Bucana, C., Sauder,  D.N.,  and De Fabo, E.G.  Mechanism of systemic
immune suppression by UV radiation in vivo. II.  The UV effects on number and
morphology of epidermal Langerhans cells and the UV-induced suppression of
contact hypersensitivity have different wavelengths.  J Immunol 132:2408-2416
(1984)

-------
                                      9-21
Noonan, P.P., De Fabo, E.G., and Kripke, M.L.  Suppression of contact
hypersensitivity by UV radiation and its relationship to UV-induced suppression
of tumor immunity.  Photochem Photobiol 34:683-689 (1981a)

Noonan, P.P., Kripke, M.L., Pedersen, G.M., and Greene, M.I. Suppression of
contact hypersensitivity in mice by ultraviolet irradiation is associated with
defective antigen presentation.  Immunol 43:527-533 (1981b)

Norbury, K.C., Kripke, M.L., and Budmen, M.B.  In vitro reactivity of
macrophages and lymphocytes from ultraviolet -irradiated mice.  J Natl Cancer
Inst 59:1231-1235 (1977)

Perna, J.J., Mannix, M.L., Rooney, J.E., Notkins, A.L. and Straus, S.E.
Reactivation of latent herpes simplex virus infection by ultraviolet light: A
human model (submitted for publication; 1987)

Roberts, L.K., and Daynes, R.A.  Modification of the immunologic properties of
chemically-induced tumors arising in hosts treated concomitantly with
ultraviolet light.  J Immunol 125:438-447  (1980)

Roberts, L.K., Spellman, C.W.,  and Daynes, R.A.  Modulation of immunoregulatory
responses directed toward various tumor antigens within hosts possessing
distinct immunologic potentials.  J Immunol 125:663-672 (1980)

Roberts, L.K., Spellman, C.W.,  and Daynes, R.A.  Establishment of a continuous T
cell line capable of suppressing anti-tumor immune responses in vivo.  J Immunol
131:514-519 (1983)

Romerdahl,  C.A., and Kripke, M.L.  Detection of UV tumor-specific T helper
activity in vitro.  J Immunol,  In press.

Rose, M.L., Parrot, D.M.V.  and Bruce, R.G. Divergent migration of mesenteric and
peripheral immunoblasts to  sites of inflammation in the mouse. Cell Invmunol
27:36-46 (1976)

Sauder, D.N., Carter, C.S., Katz, S.I. and Oppenheim, J.J. Epidermal cell
production of thymocyte activating factor  (ETAF). J Invest Dermatol 79:34-
(1982)

Sauder, D.N., Noonan, P.P., De Fabo, E.G., and Katz, S.I.    Ultraviolet
radiation inhibits alloantigen presentation by epidermal cells:  Partial
reversal by the soluble epidermal cell product, epidermal cell-derived
thymocyte-activating factor  (epidermal T-cell activating factor).    J Invest
Derm 80:485-489 (1983)

Spellman, C.W., Woodward, C.W., and Daynes, R.A.  Modification of immunological
potential by ultraviolet radiation.  I. Immune status of short-term UV-
irradiated mice.  Transplantation 24:112-119 (1977)

-------
                                      9-22
Spellman, C.W., and Daynes, R.A.  Modification of immunologic potential by
ultraviolet radiation.  II.  Generation of suppressor cells in short-term
UV-irradiated mice.  Transplantation  24:120-126 (1977)

Spellman, C.W., and Daynes, R.A.  Properties of ultraviolet light-induced
suppressor lymphocytes within a syngeneic tumor system.  Cell Immunol 36:383-387
(1978)
         i
Spellman, C.W., and Daynes, R.A.  Cross-reactive transplantation antigens
between UV-irradiated skin and UV-induced tumors.  Photoderm 1:164-169 (1984)

Spikes, J.D. Comments on light, light sources and light measurements, pp. 5-21,
In:  Experimental and Clinical Photoimmunology. Daynes, R.A., and Spikes, J.D.
Eds., CRC Press, Boca Raton, FL, 1983

Spruance, S.L.  Pathogenesis of Herpes simplex labialis:  Experimental induction
of lesions with UV light.  J Clin Microbiol 22:366-368 (1985)

Stingl, G., Katz, S.I., Clement, L.,  Green, I., and Shevach, E.M. Immunologic
functions of la-bearing epidermal Langerhans cells.  J Immunol 121:2005-2013
(1978)

Stingl, L.A., Sauder, D.N., lijima, M.,  Wolff, K.,  Pehamberger, H., and Stingl,
G.  Mechanism of UV-B-induced impairment of the antigen-presenting capacity of
murine epidermal cells.  J Immunol 130:1586-1591 (1983)

Streilein, J. W. and Tigelaar, R.E. SALT: Skin-associated lymphoid tissues,  In
Photoimmunology Kripke, M.L.,  Parrish, J.A. and Morison W. (eds)

Streilein, J.W. The skin as an immune organ,  In The Effect of Ultraviolet
Radiation on the Immune System. Parrish,  J.A (ed) New Jersey:  Johnson and
Johnson Baby Products, pp 21-34, (1983)

Streilein, J.W., Toews, G.B.,  and Bergstresser, P.R.   Langerhans cells:
Functional aspects revealed by in vivo grafting studies. J Invest Dermatol
75:17-21 (1980)

Toews, G.B., Bergstresser, P.R., Streilein, J.W., and Sullivan, S. Epidermal
Langerhans cell density determines whether contact hypersensitivity or
unresponsiveness follows skin painting with DNFB. J Immunol 124:445-453  (1980)

Ullrich,  S.E., and Kripke, M.L.  Mechanisms in the suppression of tumor
rejection produced in mice by repeated UV irradiation.  J Immunol 133:2786-2790
(1984)

-------
                                  CHAPTER 10

                       CATARACTS AND OTHER EYE DISORDERS
SUMMARY

    Cataracts are opacities that develop in the lens of the eye and impair
vision.  In the United States and other developed countries, cataract operations
prevent most cataracts from causing blindness.  However, in the U.S. cataract
remains the third leading cause of legal blindness.  In developing countries
where such operations are not always available, cataracts often result in
blindness.

    Scientific understanding of the physical mechanisms which cause cataracts is
incomplete; it is likely that more than one mechanism operates.  Epidemiological
studies, laboratory animal studies, and biochemical analysis support the belief
that some cataracts are caused by ultraviolet radiation B (UV-B).   Ultraviolet
radiation A and other causes are also likely.  A change in the amount of UV-B
radiation is reasonably likely to alter the incidence of cataracts.  UV-B may
also play a role in causing or exacerbating other eye disorders.  Ozone
modification that alters the amount of UV-B reaching the earth's surface is
likely to change the prevalence (and incidence) of cataracts.

-------
                                     10-2
FINDINGS
    1.   THERE APPEARS TO BE A REASONABLE PROBABILITY THAT CATARACT INCIDENCE
        WILL CHANGE WITH ALTERATIONS IN THE FLUX OF UV-B CAUSED BY OZONE
        MODIFICATION.

        la)  Many possible mechanisms exist for formation of cataracts.  UV-B may
             play an important role in some mechanisms.

        Ib)  Although the cornea and aqueous of the human eye screen out
             significant amounts of UV-A and UV-B radiation, nearly 50 percent
             of radiation at 320 nm is transmitted through to the lens.
             Transmittance declines substantially below 320 nm, so that less
             than one percent is transmitted below approximately 290 to 300 nm.
             However, the results of laboratory experiments on animals indicate
             that short wavelength UV-B (i.e., below 290 nm) is perhaps 250
             times more effective than long wavelength UV-B (i.e., 320 nm)  in
             inducing cataract.

        Ic)  In laboratory animal experiments, the action spectrum for cataracts
             is weighted heavily in the UV-B range.

        Id)  Human cataract prevalence appears to vary with latitude and UV
             radiation; brunescent nuclear cataracts show the strongest
             relationship.

    2.   INCREASES IN THE AMOUNT OF UV-B THAT CAN REACH THE RETINA APPEAR CAPABLE
        OF CAUSING STABLE RETINAL DISORDERS AND RETINAL DEGENERATION. TWO CAUSES
        OF BLINDNESS.

    3.   UV-B MAY PLAY A ROLE IN DISORDERS OF THE EYES AS WELL AS IN
        DEVELOPMENTAL DISORDERS.

-------
                                      10-3


 CATARACTS1

    The longest standing hypothesis which may account for the development of
 senile cataracts is that radiant energy,  particularly sunlight, is a major
 factor in the etiology of the disease.  This concept apparently originated from
 observations reported by a number of individuals indicating that cataracts
 occurred more frequently or earlier in persons whose occupations kept them
 outdoors or that populations living in areas with longer hours of sunshine have
 a higher frequency of cataract than populations from areas where there is less
 sunshine.  While the early studies were severely flawed by their failure to
 consider adequately the possible effects  of a variety of socioeconomic and other
 variables, Duke-Elder (1926, 1972) proposed that "the fundamental cause of
 cataract in all its forms may be traced to the incidence of radiant energy
 directly on the lens itself."

 Definition of Cataract

    Cataract is defined as an opacity in  the normally transparent lens of the
 eye which produces an impairment of vision.  Cataracts may occur as a result of
 a wide variety of factors including metabolic disorders, exposure to toxic
 agents, trauma, exposure to radiation, and hereditary factors.  The great
 majority of cataracts, however, are the so-called senile cataracts which occur
 in older individuals and for which no specific causative factor can be
 identified.  Cataract is a major cause of visual impairment and blindness
 particularly in developing countries where access to modern surgical facilities
 is limited.  Even in the United States, cataract is the third leading cause of
 legal blindness.  Some 60 percent of people aged 60 to 74 have at least some
 cataractous changes in their lenses.  The only efficacious treatment for
 cataract at present is the surgical removal of the opaque lens and nearly
 660,000 such operations were performed in 1982 in the U.S.

    It is clear therefore that senile cataract is a very significant health
 problem, both in terms of its impact on the affected individuals and on society
 at large.  While considerable progress has been made in elucidating the
 biochemical etiology of certain specific cataract types, such as sugar
 cataracts, there is little conclusive data on the causes of senile cataracts.
 It is likely that there are a variety of potential risk factors and that, in
 general, senile cataracts have a multifactorial etiology.  This conclusion is
 supported by the great variability observed clinically in the time of onset,  the
 rate of maturation, and the morphological appearance and location within the
 lens of these opacities.   It appears that many, if not all, of the processes
 contributing to senile cataractogenesis are normal aging processes which for
 whatever reason are accelerated in certain individuals.

 Cataract Classification

    There are now rather sophisticated systems for the classification of
 cataracts.  Chylack et al.  (1978,  1983) have devised an in vitro system which is
based on photographic documentation of opacification and nuclear color.   Almost
 2500 cataracts have been studied and classified since the original methods were
 adopted (Chylack et al.,  1983).   In addition,  Marcantonio et al.  (1980)  have
 suggested a system of classification of human senile cataracts by using
photography for in vivo and in vitro analysis and determining the sodium and
       This section in boldface type is taken from Pitts,  D.G.  et al.  (in press).

-------
                                     10-4
protein content of the extracted lens.   The photographic method gave consistent
involves 'it of th«"  IDHS nucleus but it was n- - always possible to relate sodium
changes to light scattering.  Marcantonio et al.  (1980) argue that a dual
classification system is necessary because most cataracts are mixed in nature
and the osmotic and nuclear mechanisms provide quite different changes in
protein distribution.

    The present state of knowledge concerning the human senile cataract probably
justifies a classification into only two major types in spite of the elegant
classification systems mentioned above.  The first and most common type of human
senile cataract is the cortical cataract.  Cortical cataracts are characterized
by imbalances in cation levels within the lens cells which produce osmotic
swelling and ultimate opacity in the lens cortex.  Altered cation balance may
result from damage to the Na+, K^-ATPase or from compromise of the normal
permeability characteristics of the lens membranes.  Any of a great variety of
potential insults could be the ultimate cause of such cataracts.  The second
major type of senile cataract is the nuclear cataract which is characterized
primarily by very marked modifications to the structural proteins of the lens,
the crystalline, in the central region of the lens.  These two types of
cataracts are not at all mutually exclusive; in many instances, senile cataracts
contain both cortical and nuclear opacities.  It is advisable for those who are
involved in cataract research to become knowledgeable and use a recommended
classification system.

Epidemiological Studies

    A few general statements should be made relative to the cataract
epidemiological studies.  Most epidemiological studies have reported an increase
in the prevalence of cataracts with an increase in age and at 65 years of age
and above there is an acceleration in the prevalence of cataracts.  Women have
shown a larger prevalence of cataracts than men; however, this difference has
not been shown to be statistically significant.

    In recent years there have been a number of epidemiological studies reported
which have attempted to establish that there is an association between cataract
prevalence and exposure to sunlight or the ultraviolet component of sunlight.
Killer, Giacometti and Yuen (1977) utilized data on the populations sampled in
two independent health surveys, the Model Reporting Area for Blindness
Statistics (MRA) and the National Health and Nutrition Examination Survey
(HANES) to compare the prevalence of cataract in geographical areas with varying
total annual hours of sunlight.  The total annual hours of sunlight was obtained
from U.S. Weather Bureau statistics and ranged from 1800 to 3800 hours.  For
each health survey ocular diseases other than cataract were used as control and
the data reported as age-specific ratios of cataract cases to control cases for
areas of differing total annual insolation.  The findings suggest that in the
youngest age group studies  (20-44 yrs) there is no greater incidence of cataract
in areas of high annual sunlight, but that with increasing age there appears to
develop an increasing association of cataract with sunlight exposure.  In the
65-74 yr sample population, there was at least a doubling of the ratio of
cataract to each control disease between the lowest and highest sunshine area
and in persons 75 years and older this trend was even more pronounced.  The
study did not consider the possible effects of genetic or socioeconomic
differences among the populations from different areas nor did it consider the
actual sunlight exposure of the individuals (e.g. indoor vs outdoor occupation).

    Zigman, Datiles, and Torczynski (1979) have studied populations from three

-------
                                      10-5
widely separated areas (Manila, the Philippines; Tampa, Florida; and Rochester,
New York) which differ considerably in the yearly levels of ultraviolet
radiation in sunlight.  This study considered not only the age of the
individuals studied, but also considered separately those with indoor and
outdoor occupations.  The findings which were based on study of extracted
cataracts included analysis of the cataract type for each lens.  The results
indicated that no correlation existed between geographic location and
distribution of cataracts except for the brunescent cataracts; i.e., those
nuclear cataracts with significantly increased pigmentation levels.  In Manila,
the area with greatest UV, such cataracts accounted for 43% of total cataracts
while in Tampa 20% of cataracts were brunescent and in Rochester, the area of
least UV, only 9%.  In all three populations, the percentage of brunescent
cataracts extracted in persons with outdoor occupations was markedly higher than
in those persons working indoors.  Thus, both the latitudinal variation and the
individual differences within geographic regions suggested a strong relationship
between UV exposure and brunescent cataract.  These data are consistent with
numerous reports that tropical areas have higher cataract incidence than areas
at higher latitudes and that the percentage of brunescent cataracts is higher in
tropical areas (Pirie 1972 and references in Taylor 1980).  The confounding
factors which were pointed out by Zigman, Datiles, and Torczynski (1979) as not
being controlled in their study included the economic, nutritional and genetic
backgrounds of the individuals in the respective populations.  While these
differences would probably be substantial, particularly between Manila and the
two U.S. sites, this study is significant in that it involved some biochemical
evaluation of the cataracts as well as the epidemiological data.

    Two recent epidemiological studies have concerned cataracts in Australian
aborigines, a rural population exposed to relatively intense solar radiation.
Taylor (1980) studied 350 individuals from which detailed personal histories
were obtained.  The possible role of a variety of personal and environmental
factors in cataractogenesis were investigated.  Among the 350 individuals, all
of whom were over 30 years of age, 116 had lens opacities as determined by
slit-lamp examination.  The major findings relative to the possible influence of
radiation exposure to cataractogenesis were a trend toward association of
cataracts with increased hours of sunlight and with higher annual mean UV-B
levels.  These trends were reflected by a strong association of cataract with
lower latitudes.  No other environmental factors studies appeared to be
associated with cataract nor did any personal factors other than age.  No
correlation would found between occupation and cataract; it is not clear whether
this finding is inconsistent with Zigman, Datiles, and Torczynski (1979) since
it is not reported whether any of the individuals studied had indoor
occupations.  Unfortunately, the types of cataracts present were not reported in
this study (Taylor 1980).

    In a second study involving a much greater geographical area and over 50%
(64,307) of the total Aborigine population, Hollows and Moran (1981) also found
a statistically significant correlation between environmental UV irradiation and
the prevalence of cataract (p<0.005).  Additionally cataracts occurred much more
frequently in the younger age group (40-59 yrs) in geographical zones of high UV
radiation than in zones of low UV irradiation.  The authors suggest that the
Aborigines from remote rural Australia are a suitable population for such
studies since their lifestyle tends to be highly uniform throughout the country.
In contrast to these results, a large sample of non-Aboriginal people from the
same areas showed much lower levels of cataract, particularly at younger ages,
and there was no correlation between cataract prevalence and sunlight in this
group.  This was attributed to the much higher standard of living and the

-------
                                     10-6
greater likelihood of indoor occupations in this group.

    The data of Hollows and Moran (1981) may be used to  estimate the number of
senile cataracts which are caused by sunlight if we can  accept the confounding
factors mentioned above.  Exhibit 10-1 presents the prevalence of cataracts
found in aborigines by UV zone for three different age groups.  Exhibit 10-2
compares the prevalence of cataracts found in non-aborigines and aborigines who
live in the same regions.

    Exhibit 10-1 shows that 13.6% of the aborigines in the lowest solar region
develop cataracts while about 30% in the more solar intense zones develop
cataracts.  These data appear to indicate that about 15% of the senile cataracts
are due to solar exposure.  If the aborigine and non-aborigine are compared
(Exhibit 10-2) there was about a 13% increase in cataracts which are probably
due to solar radiation.  More recently, Weale (1982) has presented a method of
estimating the risk attributed to light for the incidence of senile cataracts
and reported a risk factor of 5.  If the risk factor may be expressed in percent
form or 20%, the results are not too different from the  above.  It is
interesting that the two sets of diverse data present such interesting results.
Incidentally, the 18.3% prevalence for cataracts for the non-aborigine compared
quite favorably to the Framingham Study (Kahn et al. 1977a and b) and the
Gisborne Study (Martinez et al. 1982).

    Thus, there appears to be a consensus from these epidemiclogical studies in
support of the notion that senile cataract or at least a particular segment of
the heterogeneous mixture of opacity types which comprise senile cataracts, is
associated with higher exposure to sunlight.  Substantive questions could
certainly be raised concerning each of the studies cited above; however, taken
in aggregate they represent a variety of approaches, using different types of
populations, different criteria for cataract, different sampling and statistical
methods, and different variables to test the same hypothesis.  It is striking
that the general conclusions of each study are so similar.

Ocular Transmittance

    The effect of nonionizing radiation on a cell depends on the specific
chemical composition within the cell, that is, on the presence of absorbing
molecules or chromophores (Lerman 1980a).  This type of radiation must be
absorbed in order to cause a change in the molecule since absorbed energy  is
required to promote a chemical change.  Molecules in excited electronic states
have different chemical  and physical properties than their counterparts in the
ground  state  (prior to absorption of energy).  Thus cells that do not contain
chemical compounds absorbing at certain wavelengths will transmit these
wavelengths.  For example, the nucleic acids and most proteins in a cell are
essentially transparent  to and completely transmit visible light but absorb
certain wavelengths in  the UV region  (between 250 and 295 nm) and can be damaged
by this form  of radiation, while other macromolecules in a cell such as

-------
                10-7


            EXHIBIT 10-1

   Cataract Prevalence by UV Zone

UV Zone
1
2
3
4
5

0-39
0
0
0.1%
0.1%
0.2%
Age
40-59
1.7%
2.6%
3.7%
3.8%
5.1%

60+
13.6%
24.2%
29.5%
30.5%
29.8%
Source:  After Hollows and Moran 1981.
          EXHIBIT 10-2

 Comparison of Cataract Prevalence
 for Aborigines and Non-Aborigines
                         Age
 Ethnic Group      50-59      60+
Non-aborigine      0.8%      18.3%
Aborigine          4.4%      29.3%
Source:  After Hollows and Moran
         1981.

-------
                                      10-8
rhodopsin, which absorbs at 498  nm, and hemoglobin, which has absorption peaks
in the TJV and visible region (275, 400, and 540 to 576 nm), appear colored since
they absorb visible light.  These latter macromolecules can be damaged by high
intensities of visible radiation at their specific absorption wavelengths.

    The transmittance data for the rabbit, primate and human eye and ocular
media for the wavelength range of 200nm to 2500nm are given in Exhibits 10-3,
10-4 and 10-5.  Exhibits 10-6 and 10-7 present transmittance data for the 200 to
400 wavelength range (Kinsey 1948; Boettner and Wolter 1962; Maher 1978; Barker
1979).

The Action Spectrum

    Since the absorption of nonionizing radiation is determined by the chemical
composition of the tissue being exposed, the more radiation that the molecules
absorb the greater will be the effect of the radiation.  The term "action
spectrum" is used as a measure of the relative effect of different wavelengths
of radiation on a chemical compound, macromolecule, cell, or entire organism.
The action spectrum is a plot of the dose or radiant exposure necessary to
produce the defined effect versus the wavelength.  For example, the maximum
efficiency for experimental photokeratitis has been shown to occur at
approximately 300 nm, with a smaller peak at 295 and 320 nm (Pitts 1978 and
Pitts and Cullen 1981).  The action spectra for photokeratitis, cataracts and
retinal lesions for the rabbit, primate and human are presented in Exhibits 10-8
and 10-9.

    The eye is the only organ or tissue in the body (aside from the skin) that
is particularly sensitive to the non-ionizing wavelengths of radiation  (longer
than 280 nm) normally present in our environment.  In addition to infrared and
visible radiation, we are constantly exposed to ultraviolet radiation (solar and
man-made) throughout life.  It is estimated that approximately 8% (11 mW/cm2) of
solar radiation above the atmosphere is in the ultraviolet region (280-400 nm).
At sea level this is decreased to 2-5 mW/cm2, depending on geographic location
and season (Lerman 1980b).

    Nature has provided us with transparent ocular media which are essentially
avascular and contain very few visible wavelength absorbing chromophores  in
order to effectively transmit (as well as refract) the specific wavelengths
required to initiate the visual process by photochemical reactions.  However,
these tissues do have the ability to absorb varying amounts of ultraviolet
radiation (particularly the ocular lens). The shorter the wavelengths of
radiation absorbed the greater the potential for photic damage since there is an
inverse relationship between a wavelength and the photon energy association with
it.  Thus, UV radiation is the non-ionizing portion of the electromagnetic
spectrum which could cause the most damage, provided that  it  is absorbed.  This
axiom applies to all the  ocular tissues  including the retina  in the very  young
eye where the lens has not as yet become as effective UV filter but, in
particular, the ocular lens sustains the greatest amount of photochemical change
during a lifetime of exposure to  ambient UV radiation.

Some Biochemical Mechanisms

    There are several mechanisms which are biochemically related  to radiation
damage to the eye and the lens in particular.  These mechanisms include

-------
                                                     10-9
                                                EXHIBIT 10-3

                          Composite  Transmittance Curves for  the Rabbit With
                     Representative Data  for Boettner  and Wolter  (human,  1962),
                        Weisinger  (rabbit,  1956)  and Kinsey (rabbit,  1948)  for
                                     Comparison (after Barker,  1979)
Ul
u

-------
                                      10-10
                                   EXHIBIT 10-4

                    Calculated Total Transmittance of the Human Eye
                          (after Boettner and Wolter 1962)
100
                      TOTAL TRANSMITTANCE AT THE
                      VARIOUS ANTERIOR SURFACES
                                          VITREOUS
                                          RETINA
I    AQUEOUS
2    LENS
        300     400    500  600
                        WAVELENGTH
                 800   1000  1200
                MILLIMICRONS
1600  2000

-------
    IBD


     SB
                                             10-11
                                         EXHIBIT 10-5



                     Percent Transmissivity Through the Entire Rhesus Eye

                                       (after Maher 1978)
     TB .
in   G2
in


m   SB

oc
£   43
     33




     23 .




     10 .




      0
       .2
 i'ii>i>i'i


.0     1.0    1.2    I.H    I.E    I .B    2.0   2.2    2.M
                        WRVETL.ENEXH

-------
                                                 10-12
                                            EXHIBIT 10-6

            Transnittance of the Total Rabbit  Cornea, the Total Human Cornea,
                 and  the Rabbit Corneal Epithelium (after Kinsey [1948] and
                                    Boettner and Wolter  [1962]
i.o
0.9
0.8
0.7
O MUIT CORNEAL IPITMCLIIM. KINSEY

» MUIT TOTAL CORNEA. KINSEY

O MUIT TOTAL CORNEA. BACHEN

• HUMAN TOTAL CORNEA. METTNER AND MOLTERS
    2JO    210    250   260   270    280   290   300    J10   320   330    310   350   360   370   380    390   100

                                        WAVELENGTH IN NANOMETERS

-------
                                         10-13
                                     EXHIBIT 10-7


            Transmittance of the Anterior Ocular  Structures of the Human
              and Rabbit Eyes.  The Symbols Indicate the Percentage of
                   Radiant Energy Incident On  a Certain Structure
                         (after Pitts, Hacker,  and Parr 1977)
O
LU
O
    1.0


   0.9



   0.8



   0.7



   0.6
CO
2
<
cr
   0.3
   Q2


   01
  0.0
         OINCIOENTON AQUEOUS, HUMAN
         X INCIDENT ON LENS, HUMAN
         • INCIDENT ON VITREOUS, HUMAN
         D INCIDENT ON AQUEOUS, RABBIT
         A INCIDENT ON LENS, RABBIT
         V INCIDENTONVITREOUS.RABBIT
                              32O   330   340  350  36O  37D  38O  39O  400
      270  280  290  300  310

                   WAVELENGTH IN  NANOMETERS

-------
                                     10-14
                                 EXHIBIT 10-8
             DV Radiant Exposure Threshold Data for the Cornea Hc,
          Lens HL Cataracts,  and Retina HR for the Rabbit and Primate
WAVELENGTH
nm
210
220
230
240
250
260
270
280
290
295
300
305
310
315
320
325
330
335
340
345
350
355
360
365
370
375
380
385
390
395
405
441
CORNEAL
Radiant Exposure
Hc in J/cm2
Rabbit
0.17
0.046
0.03
0.033
0.041
0.018
0.005
0.11
0.012

0.022
0.07
0.05
7.25
7.5
18.0
30.0
30.0

• - -

50.0

65.0








Primate
0.33
0.021
0.022
0.012
0.020
0.011
0.004
0.006
0.007

0.01

0.02

9.6

41.1

58.3

61.5

88.4

130.

170.

258.



LENS RETINA
Radiant Exposure Radiant Exposure
HL in J/cm2 HR in J/cm2
Rabbit Primate +Rabbit *Primate








3.00
0.75
0.15 0.12 0.225 0.12
0.30
0.75
4.50
12.60 5.0
50.00




5.4





8.1



15.0
30.0
* for 100s exposure duration in aphakic primate producing retinal lesion.

+ for 650s exposure duration in phakic rabbit eye producing changes in the
retina.

-------
       THRESHOLD RADIANT  EXPOSURE IN   Jm~2 X I0~4
. o
o
o
 .0
 O<3
    ~
                     0
                     O
0

*>
0-
O
P
O
•J-  O
o  p
b  o
                                                    O
                                                    o
                                                    b
                                                                 o
                                                                 b
                             oo
                             o  o
>•" o1
^-. o
PI
*~ 2
m °
^X** fO
O §
-I 3
in o
9 °
ni °
	i w
m o
co S
•1 •
                                          31 H Tl I  T> J< <^
                                          p. t. 3) C  - C O
                                          01 01 ? Z  3: a> o
                                          gj u> ^ i>  JJ H >;
                                                                           f
                                                                         S'S
                                                                                   ?
                                                     n
                                                                           M
                                                                         i'«
                                                                         B?
                                                                           81
                                                                           H
                                                     O* ft

                                                     *!
                                                     ^S
                                                       H-
                                                     .- rt
                                                     Htr
                                                     "^
                                                     H-
                                                                         *—*



                                                                         I
                                                                         ?
                                                                         VO
                                                                         00
                                                                           ?
                                                                           8
                                                                           rt
                                                                           CO
                                                                              w

-------
                                     10-16
photo-oxidation of free and protein bound tryptophan, photosynthesis processes
involving the activated species of oxygen, disruption of the cation transport
system and damage to the nucleic acids (DNA) of the lens epithelium.  Some of
these mechanisms have only been studied only in cultures, some in vitro and
others in vivo.  It is important to be able to project the mechanism to a "real
live" situation in order to be able to evaluate the relative importance of the
factors involved in inducing the senile cataract.  Prior to discussing these
mechanisms it may be desirable to review some of the basic concepts in the
biochemical mechanisms of radiation induced damage; therefore, some current
interpretations of lens free radicals and oxidation reduction reactions relative
to cataract follows.  Exposure to "UV radiation initiates enzymatic activity
involved in cellular protection from oxidative processes which may be due to
both light and metabolism (Exhibit 10-10).  Catalase destroys hydrogen peroxide,
which is produced in all cells by metabolism; superoxide dismutase (SOD) has the
responsibility of destroying the superoxide radical, a very toxic radical due to
its powerful oxidative nature.  In the lens cortex, there is very little SOD or
catalase, but they are highly concentrated in the epithelial cells.  Of all the
ocular tissues, the greatest concentrations of these enzymes are present in the
retina.  In the lens cortex there are other agents to protect against oxidation
such as glutathione (GSH) and ascorbic acid.  There is only a small amount of
vitamin E in the lens, but attempting to protect animals from ocular tissue
oxidative damage by feeding them high levels of vitamin E or other anti-oxidants
has not succeeded.  The sura total of all of the anti-oxidants that are present
in the ocular tissues still does not protect against the formation of free
radical oxidative reactions to near-UV radiation.  Another toxic oxidant that
can form is singlet oxygen.

    The content of the stable free radicals previously described is highest in
the cortex and seems to decrease in the nucleus, of the lens.  The process of
aggregation of soluble proteins (TSP) occurs both in the cortex and in the
nucleus.  Because of the conservative growth process of the lens, much of the
aggregated material remains in the nucleus indefinitely and often in a form that
is associated with the fiber cell membranes.  The lens fraction called the
insoluble fraction contains both membranes and the aggregated proteins.  There
is a chemical reaction between the free radicals and proteins so that the
protein binds to free radicals chemically, which quenches the ESR (electron spin
resonance) signal.  Multiple molecular species of proteins can be associated
through new crosslinks leading to aggregation and, therefore, light scattering.
In the nucleus of the lens, the free radicals are quenched by their reactions
with proteins.  Since all enzymes are proteins, this may illustrate a universal
process whereby enzyme activity is inhibited by near-UV radiation.  These
changes would eventually cause malfunction of all cells and tissues.

    Exhibit 10-11 shows a scheme of a series of enzymatic oxidation:  reduction
reactions going on in all cells that are influenced by aging and exposure to
near-ultraviolet radiation.  These reactions are all linked together.  The
co-factor NADH/NADP system is linked to the oxidation and reduction of
glutathione (GSH).  When oxidation occurs, reduced GSH becomes oxidized GSSG,
changes which also relate to protein oxidation involving SS-crosslinks.  Many
enzymes, in order to remain active, must maintain sulfhydryl groups (SH) of
cysteine in reduced form in order for the protein portion to maintain its
activities as an enzyme.  Many enzymes require free-SH groups in their active
sites to retain their activity.  If these enzymes are oxidized so that

-------
                                     10-17


                                 EXHIBIT 10-10

                Free Radicals and Oxidation:  Reduction Systems
               SOD

        EPITHELIUM
       UV
                                                       CORTEX
                                                      NUCLEUS
                                                      LOW FREE
                                                       RADICAL
                                                      GSH

                                                      ASCCREATE

                                                      HIGH FREE
SOD represents superoxide dismutase, GSH represents glutathione.   Epithelial SOD
and catalase protect  these cells against oxidation from excessive  hydrogen
peroxide and superoxide anion.  High levels of ascorbic acid and glutathione
protect cortical  fibers against oxidation.  TSP represents soluble and INS
insoluble protein.  TSP converts to INS protein both in the cortex and in the
nucleus.   But INS accumulates more in the nucleus than in the cortex.  Free
radicals are quenched in nucleus due to binding to proteins.

-------
                                     10-18



                                 EXHIBIT  10-11


               Enzyme Systems Involved in Oxidation:  Reduction
        CoA-SH / PROVEIH-SH
           H20 / ROM
GSSG
OXIDATION on

'un ox 11) AT i ON

   CoA-SS-CoA / P
               ROOH
                      PRQT
                          pnoi,
                                           NADPH
            NADI't
6-PHOSPHATE-
GLUCOMOLACTONE
                                                                      TO
                                      PENTOSE

                                      SIIUNT
                                                            D-GLUCOSE-6-
                                                              PHQSPHATE
             SUPEHOXIDE     OLUrATHIONE   GMITATHIONE    GLUCOSE 6-PHOSPIIATG
             CiUBjAifim                  I'   ' jU
-------
                                     10-19
disulfides form, they are inactivated.  For example, if high concentrations of
hydrogen peroxide are generated, protein SH will convert to SS thereby
inhibiting the activity.  Several of the enzymes referred to in Exhibit 10-11
are known to be reduced with enhanced aging in most living cells.  In many cases
the cells turn over, so that cells' enzymatic activities are maintained, but
where there are populations of cells that do not turn over very actively, a
cumulative process can lead to a great loss of enzyme activities.  This may
occur in the lens.

    One enzyme that is diminished with aging is glutathione reductase and it
also diminishes under in vitro circumstances in living tissues when near-UV
radiation is provided in the presence of excess tryptophan.  There has not yet
been an aging or photoinactivation study of glutathione peroxidase.  The SOD is
very important as a protective agent because it destroys the very toxic
superoxide radical.  SOD activity does diminish with aging, but is not sensitive
to destruction from near-UV radiation (with tryptophan as a sensitizer).
Because the lens substance contains little oxygen to serve as a superoxide
source, the site of lens damage would be the epithelium if SOD activity were to
be diminished drastically.  Damage to the lens epithelium would certainly lead
to both physiological and developmental anomalies in the lens and cataract.  The
SOD activity must be maintained in the retina, however, due to its high oxygen
tension and the potential for oxidation to occur more readily.

    Hydrogen peroxide (H202) is another very powerful oxidant that is present in
ocular tissues and fluids by virtue of the action of SOD on superoxide and other
metabolic systems, such as the glutathione and ascorbic acid cycles.  Recently,
high levels of H202 (mM concentrations) have been observed in the human aqueous
humor by Garner and Spector (1980) .  This concentration of H202 is suspected of
being capable of causing cataracts by poisoning important enzymes and by
crosslinking proteins (Spector and Garner 1982).  In eyes without lenses, H202
from the aqueous humor could more easily diffuse even to the retina (Zigman
1981, Kramer 1980), a highly hazardous circumstance.  Ocular tissues utilize the
enzyme catalase to destroy and detoxify H202.  Since this enzyme is diminished
with aging and has been shown to be inhibited by near-UV radiation plus
tryptophan (Zigman, Yulo, and Griess 1976), its activity loss would allow
heightened levels of H202 to damage ocular tissues.  A large imbalance in favor
of H202 accumulation would lead to extensively altered ocular tissue proteins
and much enzyme inactivation.

    Glutathione reductase is another enzyme of great functional importance to
the lens and retina, and it is diminished with aging and UV exposure (Kalustian,
Sun, and Zigman 1978).  Its loss of activity would estimate oxidation of small
molecules and proteins in most ocular tissues.

    Another enzyme  (not involved in oxidation: reduction, but significant with
regard to its photoinactivation) is Na+IC* ATPase.  This enzyme is also sensitive
to near-UV radiation in the presence of tryptophan and loses its activity
accordingly.  In the lens, it maintains the osmotic balance by controlling the
sodium and potassium cation exchange and prevents water inhibition and swelling.
In some genetic cataracts in mice, it has been found that the cause of cataract
is the high concentration of an inhibitor of ATPase in the lens  (Kinoshita
1974).  It is likely then that a strong inhibition of lens epithelial cell
ATPase by light-sensitized action could be a major factor in cataract formation
osmotically.  In the retina, the ATPase activity is very important in terms of
the chemical reactions that support the visual process.  Should this enzyme be
photochemically inhibited in the retina, the functioning of the visual process

-------
                                     10-20


would be markedly reduced.  Retinal Na"*TC*" ATPase is also sensitive to inhibition
by near-UV plus tryptophan.

    Before discussing the biochemical studies which pertain to the possible
cataractogenic effects of optical radiation, it may be pertinent to review some
contrasts between human lenses and those of the most commonly used laboratory
animals.  All vertebrate lenses are composed essentially of a single cell type,
the lens fiber, which differentiates from a single layer of epithelial cells
present only at the bow or equator of the lens.  New fibers are continually laid
down at the periphery; thus, the oldest tissue is located at the center or
nucleus of the lens and the cells become progressively younger as one moves from
the center toward the lens capsule in the cortex.  Cells are never sloughed from
the lens and this makes the cells in the lens nucleus as old as the animal.
Furthermore, differentiated lens fibers gradually lose virtually all cell
organelles, including nuclei, and lose the capacity to synthesize protein as
they age and are forced toward the nucleus.  Therefore, the proteins, primarily
lens crystallins, present in the aging human lens may be the longest-lived
proteins in the organism.  This means that unlike most other tissues the central
portion of the lens does not have the ability to replace damaged proteins with
newly synthesize proteins.  This may be an important factor in the presumptive
long-term effects of chronic exposure to near UV radiation.  There are very
clear differences between the lens nucleus and cortex in terms of their
biochemistry and the loss of the capacity to repair or replace damaged proteins
probably accounts for much of the difference.

    There is at least one very significant difference between human lenses and
the common laboratory animals and that is the presence of pigmented compounds in
the human lens.  While the lenses of most non-primate mammals have no pigment,
the lenses of diurnal primates and a few other strongly diurnal species are
yellow.  Cooper and Robson (1969) demonstrated that in the human lens there are
two classes of pigmented compounds.  One group that is present even before birth
is of low molecular weight, is water soluble and absorbs maximally at about 365
nm.  A second class of colored compounds appears later, increases with age, is
bound to the lens proteins and absorbs maximally near 320 nm.  This latter class
of chromophores is localized primarily in the lens nucleus and may be
responsible for the age-related increase in pigmentation in human lenses.  Since
these chromophores absorb in the near UV, they may be major determinants of the
effects of such radiation on the human lens.

Studies at the Biochemical Level

    Studies at the biochemical level have generally been concerned with the
structural modifications which crystallins undergo during aging and
cataractogenesis and have attempted to explain these reactions in terms of
photo-oxidative mechanisms.  It is well-documented that crystallins,
particularly in the lens nucleus, accumulate a variety of modifications.  These
include the formation of disulfides and other covalent crosslinks, the
development of a novel blue fluorescence, progressive pigmentation, oxidation of
methionine, racemization of aspartate residues, polypeptide chain cleavages,
deamidations, aggregation and ultimate insolubilization.  The lack of turnover
of protein in the lens nucleus accounts for the accumulation of these
modifications and they have been the subject of several recent reviews (Zigler
and Goosey 1981; Hoenders and Bloemendal 1981; Harding 1981).  It is clear that
most of the protein changes are the result of oxidative stress, and the possible
role of radiation in that stress has been the subject of much study over the
last 15 years.

-------
                                     10-21
    Pirie (1968, 1972) studied the effects of sunlight on solutions of lens
proteins and of other proteins.  The proteins became brown following irradiation
and analysis of absorbance changes indicated similarity to those found in lens
proteins from cataracts.  Pirie also found decreased levels of the oxidation
sensitive amino acids, histidine and tryptophan in protein from cataracts
relative to normal lenses.  Although the data for tryptophan was subsequently
retracted as an artifact, Dilley and Pirie (1974) suggested that photooxidation
of tryptophan residues, perhaps with formation of N'-formyl kynurenine, was a
primary step in cataractogenesis.  These studies stimulated a number of other
investigators.

    Kurzel (1973) performed fluorescence and phosphorescence measurements and
Weiter and Finch (1975) ESR studies on human lenses and found signals which they
believed to be due to tryptophan free radical species in lens proteins.
Van Heyningen (1971) was able to identify several of the components contributing
to the color of human lenses as kynurenine derivatives, species which can be
derived from tryptophan either metabolically or photooxidatively.  These were
part of the low molecular weight colored material present in human lenses and
Van Heyningen (1973) subsequently showed that lens proteins exposed to sunlight
in the presence of these compounds were photo-oxidized more extensively than in
their absence.  The mechanisms of this accelerated photo-oxidation was not
determined.

    In addition to oxidation of protein bound tryptophan other possible
mechanisms were explored.  Zigman et al. (1973) and Zigman and Vaughan (1974)
demonstrated that photo-oxidation of free tryptophan yielded pigmented and
fluorescent species which would bind to lens crystallins in vitro.  This raised
the possibility that the target of photo-oxidation could be either free or
protein-bound tryptophan.  Numerous investigators turned to the study of
brunescent nuclear cataracts since these lenses have the greatest concentration
of the pigment and of the non-tryptophan fluorescence.  The search for clear
decreases in the levels of tryptophan in the proteins of such lenses or in their
free amino acid pool has not been successful to date (Dilley and Pirie 1974;
Pirie and Dilley 1974; Zigler et al. 1976).  It should be noted however that
tryptophan comprises less than 2% of total amino acid in crystallins and, thus,
small changes would be difficult to detect especially in view of the inherent
problems of tryptophan analysis.

    Lerman (1980b) suggests that in the normal lens less than 20% of the protein
tryptophan is susceptible to photo-oxidative damage.  Studies on the novel blue
fluorescence of aging crystallins, particularly the insoluble fraction from
brunescent lens nuclei suggest the presence of a number of related species
(Lerman 1980b).  The species with emissions in the visible are generally more
concentrated in nuclear cataracts.  There is some evidence that this
fluorescence may be concentrated in certain crystallin polypeptides (Zigman
1981).  It has also been demonstrated that the formation of non-disulf ide
covalent crosslinks between crystallin polypeptides is associated with the
heavily pigmented protein fraction and can be generated in vitro by irradiation
(Buckingham and Pirie 1972) .  Additional photoproducts of tryptophan, including
B-carbolines (Dillon, Spector and Nakanishi 1976) and anthranilic acid
(Truscott, Faull, and Augusteyn 1977) have been identified from the crystallins
of nuclear cataracts.  Dillon and Spector (1980), Dillion et al. (1982) and
Borkman, Tassin, and Lerman (1981) have studied the photolysis of free
tryptophan, or tryptophan containing peptides, and of isolated lens crystallins.
Analysis of these data suggests that a variety of products are possible and that

-------
                                     10-22
the microenvironment of individual tryptophan residues is of paramount
importance.  Additionally, photolysis is much faster in the presence of oxygen.

    Harding and Dilley (1976), however, raised two objections to the idea that
sunlight caused brown nuclear cataracts.  First, they pointed out that the lens
damage is in the nucleus whereas the shortest wavelengths reaching the lens;
i.e., those which might be absorbed by tryptophan are probably absorbed in the
anterior lens cortex.  Indeed as noted above cataracts induced in animals by
UV-B are located in the anterior cortex.  Secondly, the lack of
evidence for loss of tryptophan in brown cataracts was cited.  While these
arguments are difficult to rebut in terms of mechanisms in which UV oxidation of
tryptophan is the central event, there is another mechanism of photo-oxidation
for which these objections may be less significant.

    Recently, there has been increasing interest in the possible role of
photosensitized processes, particularly with the involvement of activated
species of oxygen, in the oxidative damage observed in the human lens.  This
work was spurred by data demonstrating light-mediated lens damage with such
photosensitizing drugs as 8-methoxypsoralen, phenothiazines, and tetracycline
and by the studies cited above by Pirie and Van Heyningen showing accelerated DV
effects on crystallins in the presence of kynurenines and related compounds.
Zigler and Goosey (1981) have demonstrated that several of these compounds
endogenous to human lenses are capable of generating singlet molecular oxygen, a
highly reactive species capable of damaging proteins as well as other biological
molecules and structures.  The ability of this oxidant to induce the oxidative
changes characteristic of human crystallins has been established in vitro.  It
has also been demonstrated that such photosensitizing activity is present in the
heavily altered insoluble lens crystallins from brunescent cataracts and to a
lesser extent in the soluble crystallins of normal human lenses as well.  Based
on experiments in vitro, such a photosensitized process could account for the
generation of each of the oxidative modifications presently known to occur in
lens crystallins.  Furthermore such a process seems consistent with nuclear
localization of damage.  In the lens nucleus there is no repair or replacement
of altered molecules, thus allowing progressive accumulation of crystallins with
oxidized residues some of which are photodynamic sensitizers.  The UV-A which
these species absorb will readily penetrate to the nucleus.  Additionally while
the lens cortex is protected by a battery of antioxidant defenses it is known
that these defenses are markedly reduced in the nucleus (Hata and Hockwin 1977;
Fecondo and Augusteyn 1983).

    One could envision a system in which there was a slow rate of
photo-oxidation ongoing in the lens controlled by antioxidant defenses and by
the greatly reduced oxygen tension in the lens nucleus.  The oxidative stress
might have several components including direct UV oxidation of tryptophan
(primarily UV-B), the high levels of H202 in the aqueous humor (Spector and
Gardner 1982) and photosensitized oxidation involving activated species of
oxygen.  It has been recently demonstrated that even fetal human lenses contain
low molecular weight chromophores which can generate singlet oxygen when
irradiated with UV-A.  The initial photochemical event could be absorption by
such chromophores or direct photo-oxidation of tryptophan to produce N-formyl
kynurenine (NFK) a known photodynamic sensitizer.  In either case, it seems
likely that the continued build-up of oxidized products in the nucleus is likely
due primarily to a sensitized process, since such a process would not require
large-scale tryptophan loss nor would it require penetration of UV-B into the
nucleus.  A relatively small number of stable sensitizing species such as NFK,
bound within the long-lived nuclear crystallins, could continue to generate

-------
                                     10-23
singlet oxygen indefinitely with the gradual accumulation of the various
oxidative protein changes outlined above culminating in aggregation and
insolubilization of much of the nuclear crystallins in advanced brunescent
cataracts.

    Several suggestions have been made that attempt to reconcile the finding
that tryptophan is not reduced in senile cataracts (as compared to normal
lenses) with the hypothesis that it acts as an important endogenous
photosensitizer.  For example, it has been suggested that tryptophan is oxidized
to a reactive molecule that simply transfers its energy to some other cell
component from the excited state.  It then returns to the ground state without
being destroyed in the process, yet photosensitizes damage to other lens
macromolecules (Welter and Finch 1975; Weiter and Subramanian 1978).  Another
explanation is that free tryptophan rather than protein bound tryptophan may be
destroyed when it acts as a photosensitizer, whereas tryptophan incorporated
within lens proteins is unaffected.  The level of free tryptophan may be reduced
in cataract (Zigler et al. 1976).  The failure of protein-incorporated
tryptophans to be photo-oxidized might be explained by the fact that they lie in
very different micro-environments that render them less susceptible to
photo-oxidation (Lerman, 1980b).  A third suggestion to explain the finding that
tryptophan is not decreased in cataractous lenses is that only a small percent
of the protein-incorporated tryptophan may be susceptible to photo-oxidation and
small losses in these susceptible tryptophan might not be detectable with
current analytical capabilities.  Tryptophan and its photoproducts have become
the most frequent culprits implicated in causing photo-oxidative changes in lens
proteins, changes that are hypothesized to be responsible for senile
cataractogenesis.

    Photo-oxidation of lens proteins has become a major topic in lens research
and has been shown to be induced by photosensitizers that are either endogenous
(Zigler and Goosey 1981) or externally applied (Goosey, Zigler, and Kinoshita
1980) to the lens.  One effect frequently found to occur as a direct effect of
UV or as an effect of UV plus a photosensitizer is crosslinking of lens proteins
(Buckingham and Pirie 1972; Goosey, Zigler, and Kinoshita 1980).

    Interest in agents that can cross-link or otherwise aggregate lens proteins
has been high since Benedek (1971) first proposed that the mechanisms underlying
senile cataract formation was aggregation and insolubilization of lens
crystallins.  They suggested that aggregated lens proteins would serve as light
scattering particles in the lens.  The search was then underway for these high
molecular weight aggregates, but unfortunately, the research produced very
inconsistent and conflicting results.  Some research groups found an increase in
the amount of high molecular weight proteins in senile cataractous lenses while
other groups found no difference in the amount of high molecular weight proteins
in cataractous lenses compared to that present in lenses in age matched controls
(Harding and Dilley 1976).  It was also suggested that aggregation was an
artifact resulting from the extraction procedure (Harding 1972).  Analogous
studies have more recently been carried out on lenses in which the cataracts
were separated according to the nature of the changes present in the lens.  The
results of those studies have suggested that nuclear brunescent, but not
cortical cataract, is associated with formation of increasing amounts of water
insoluble proteins (Truscott and Augusteyn 1977).  Some researchers have
attributed this increase to cross-linking of lens proteins induced by
photo-oxidation.  It has been further suggested that the reason these changes
occur more readily in the lens nucleus than in the cortex is that the cortex has
a higher concentration of protective anti-oxidants.

-------
                                     10-24
    If Benedek's (1971) original hypothesis that the mechanism underlying senile
cataract formation was aggregation and insolubilization of lens crystallins is
correct, then one would expect to find an increase in protein cross-linking and
aggregation in lenses that show light scattering.  To the contrary, lenses with
light scattering opacities in the cortex do not show an increase in high
molecular weight proteins or crosslinks (Truscott and Augusteyn 1977; Anderson,
Wright, and Spector 1979).  Instead, reports in the literature suggest that
there may be an increase in protein aggregates in lenses with substantial
nuclear brunescence.  This is rather a surprise because nuclear brunescence per
se is not a light scattering change in the lens; rather, the lens turns a yellow
or a brown color that absorbs rather than scatters light, reducing its intensity
on the retina (Lerman and Borkman 1976).  The effect of brunescence is,
therefore, analogous to placing a transparent filter before the eye.

    Unfortunately, only man and a very few other diurnal species have been found
to develop lens browning so that it has been difficult to find an animal model
in which to test the hypothesis that browning can be induced in vivo by exposure
to UV radiation.  Therefore, those studies that have examined the effects of
chronic UV exposure in experimental animals have not demonstrated that UV
induces lens browning.  They have, however, consistently demonstrated that UV
exposure induces light-scattering cortical opacities (Bachem 1956; Zigman and
Vaughan 1974; Pitts, Hacker, and Parr 1977).  Therefore, though we may lack a
suitable model for nuclear brunescence, there are animal models that can be used
to study the induction of cortical opacities by UV radiation.  These models
mimic the common and considerably more visually disturbing senile cortical
cataracts in man.

    In those studies on UV-induced lens changes in which histology of the lenses
were examined, epithelial cell changes were a uniform finding (Zigman and
Vaughan 1974; Pitts, Hacker, and Parr 1977).  In their study of chronically
exposed mice, Zigman and Vaughan (1974) noted the similarity of the lens changes
to changes induced by X-irradiation.  For example, lens cells appeared to have
lost their capacity to differentiate and were found to have migrated to the
posterior pole, a situation that has also been found to be associated with
senile cortical cataracts (Streeten and Eshaghian 1978).

    While biochemical approaches have generally concentrated on nuclear
cataracts with respect to UV-mediated effects, there has been recent interest in
photosensitized reactions in the aqueous humor as sources of damage to lens
membranes and, hence, as a possible initiator of cortical (osmotic) cataracts.
Varma, Kumar, and Richards (1979) have been investigating the possibility that a
photochemical conversion of molecular oxygen present in the aqueous humor and
lens into superoxide and subsequent derivatization of superoxide to other potent
oxidants such as hydrogen peroxide, hydroxyl radical and singlet oxygen may be
involved in initiating a cascade of toxic biochemical reactions leading to the
formation of cataracts.  Thus, according to this hypothesis, cataractogenic
influence of light is mediated by a photochemical generation of superoxide from
the ambient oxygen.  Spin restriction offered by the molecular oxygen makes the
formation of superoxide a necessary event in most oxidation reactions involving
oxygen.  It is commonly understood that, like many other free radicals,
superoxide and its derivatives, if allowed to remain unscavenged for any length
of time in a biological milieu, will initiate many nonspecific and deleterious
reactions such as an upsetting of the normal redox chain, oxidation of vitally
important protein and nonprotein - SH, peroxidation of membrane and cytosolar
lipids, and polymerization and depolymerization of macromolecules such as

-------
                                     10-25
proteins and hyaluronic acids.  The cataractogenic influence of these oxidants
is likely to be modulated by certain endogenous protective mechanisms.
Superoxide dismutase (SOD), catalase and perioxidase constitute the first line
of defense against the toxic effects of those oxygen species.  Jernigan et al.
(1981) showed that singlet oxygen generated in the medium had similar effects.
Varma, Beachy, and Richards (1982) have demonstrated lipid peroxidation in
cultured lenses irradiated with fluorescent light.  While such studies are of
great interest in terms of the effects of these oxidants on the lens, at the
present ^time there is no real evidence that they are produced insignificant
quantities in the ocular humors by light mediated processes.

    Considerable effort has been directed at elucidating mechanisms by which UV
radiation (alone or in combination with exogenous photosensitizers) might induce
cataract formation.  As previously mentioned, it has been shown that DV
radiation can induce cross-linking and aggregation of proteins.  As has also
been mentioned, protein aggregation may be associated with nuclear changes but
it is not a characteristic associated with cortical cataracts.  It may therefore
be more relevant to consider the possibility that UV damage to lens proteins
induces local perturbations in macromolecular structure and function.  This
might in turn result in localized changes in lens structure of function.  For
example, transport systems and ATPase (Varma, Kumar, and Richards 1979) in the
lens have been shown to be inhibited by UV radiation.  If the lens transport
systems are damaged by UV radiation, the osmotic balance will be disrupted
which, in turn, would produce major changes in lens morphology.

    The UV effects on transport systems may occur due to direct UV absorption by
and damage to the related enzymes; however, other mechanisms may also
contributed to this inhibition.  For example, the activity of transport enzymes
could be dramatically altered secondary to UV induced disruption of lipids in
lens membranes; and, lipid peroxidation has been proposed as a mechanisms of
cataract formation (Goosey, Allison, and Garcia 1983).  The resulting
alterations in lipid structure could of themselves make the membranes leaky or,
as just mentioned, such alterations could disrupt the structure and function of
membrane proteins.  If one can extrapolate from the large amount of research
that has been done on other cell systems, then it may be predicted that fairly
long wavelengths of UV will similarly be able to produce membrane damaging
effects in the lens (Moss and Smith 1981; Imbrie and Murphy 1982; Sprott,
Martin, and Schneider 1976).

    If we are to determine the role of lipid peroxidation or inactivation of
specific enzyme systems in the development of UV cataracts, it is critically
important that lenses be sampled several intervals prior to the onset and during
the development of lens opacities.  Once the lens is fully opacified,
retrospectively, it is difficult to state with any assurance that some observed
biochemical change was the specific change that initiated the cataract.

    For example, if an ATPase were found to be inactivated after a particular
UV-induced cataract was fully developed, one might conclude that this
inactivation had caused cell swelling, membrane lysis, and an "osmotic
cataract."  But it is also possible that the primary cause of the cataract was
direct membrane damage or inhibition of protein synthesis and, that, only
secondarily was ATPase activity involved.  Therefore, to show that ATPase
inactivation is of any real significance in the etiology of the cataract, it is
necessary to demonstrate that ATPase inactivation precedes development of the
lens opacities.  In the case of UV-induced cataracts, one also needs to
demonstrate that those wavelengths that inactivate the ATPase are those same

-------
                                     10-26
wavelengths that most readily induce cataract formation.

    It may be misleading to examine effects of UV (with or without added
photosensltizers) on lens macromolecules isolated in test tubes because UV
effects in vitro may be very different from the effects of UV on the intact
lens.  For example, it is possible to cause tryptophan destruction in lens
proteins irradiated at a concentration of 2.5 mg/ml in a test tube, but the same
type of irradiation was not found to produce such effects at the concentration
of protein in the lens (Dillon et al. 1982).  Also, the action spectrum may
differ considerably from the absorption spectrum (Turro and Lamola 1977).   The
action spectrum shifts markedly in situ in the case of the photosensitizer,
psoralen.  Psoralen, riboflavin, methylene blue and rose bengal have all been
found to induce photo-oxidative changes in isolated lens proteins when these
proteins are irradiated with UV.  However, each of these photosensitizers
requires the presence of oxygen, otherwise lens proteins are not damaged.   The
amount of oxygen in the lens is not high (Kwan, Ninikoski, and Hunt 1972;  Barr
and Roetman 1974).  Thus, these findings must be extrapolated to the in vivo
lens with considerable discretion and it remains to be determined just how much
photodynamic damage (that is damage where oxygen is involved as an intermediate)
can be produced in the lens in situ.  This is a very important research question
because any effect that is of significance as a mechanisms of cataract formation
must be inducible in a lens that is still in the eye.  For example, it might be
interesting to use 254 nm radiation as a toll in vitro to demonstrate that the
lens epithelium has the capacity to repair damaged DNA (Jose and Yielding 1977).
However, 254 nm radiation does not penetrate the cornea, so it will not be
hazard for the in vivo lens and it will not cause cataracts.  Therefore, it is a
far more significant finding that 300 nm radiation can induce DNA repair
synthesis in the lens epithelium when the lens is exposed in situ through the
intact cornea (Jose, Kock, and Respondek 1982; Brenner and Grabner 1982).   This
finding shows that 300nm UV radiation penetrates through the cornea and
demonstrates that epithelial cell DNA is a target of its effects.

    Therefore, if we are to establish mechanisms of UV damage in the lens, we
must be able to show that a given damaging reaction can be produced by UV in the
in situ lens.  An effect that can be produced on lens macromolecules isolated in
a test tube may be interesting and stimulate further research, but it must not
be taken as sufficient proof that such an effect can occur in vitro.  This is
especially true for those reactions that require oxygen.

    One photosensitizer that may provide important clues as to mechanisms
involved in cataract formation is the drug psoralen which in combination with UV
radiation is used in treating psoriasis.  Two different mechanisms have been
proposed to underlie the psoralen cataract.  According to one hypothesis, the
cataracts occur as a consequence of psoralen binding covalently to lens proteins
which results in formation of additional damaging photosensitizers in the lens
(Lerman, Megaw, and Willis 1980).  The second hypothesis is that the cataracts
occur as a consequence of psoralen induced damage to lens nucleic acids (Jose
and Yielding 1979).

    It has been found that intense exposure to UV radiation will cause psoralen
to bind to isolated lens proteins, specifically the tryptophan moieties (Lerman,
Megaw, and Willis 1980); however, the reaction requires oxygen  (Megaw, Lee, and
Lerman 1980).  Therefore, it is  important to determine whether such protein
binding occurs in the intact lens.  Although spectra have been presented that
demonstrate that UV radiation induces psoralen binding  in the in vivo lens, it
is not clear that the binding observed represents protein specific binding or

-------
                                     10-27
whether the binding is to some other lens macromolecule.  Lens protein fractions
have been extracted from psoralen-treated humans or laboratory animals and
fluorescence spectra taken from these.  These spectra were interpreted to
demonstrate photobinding of psoralen to lens proteins; however, no effort was
made in these studies to extract nucleic acids from these "protein"
preparations.  It is very possible that a significant amount of the binding that
was observed is accounted for by psoralen binding to UNA or DNA present in the
degenerating nuclei of the lens cortex.  The possibility that the spectra
represented binding to nucleic acids in the differentiating fibers was
discounted by the researchers because there is such an overwhelming
concentration of protein compared to DNA in lens fibers (Lerman, Megaw, and
Gardner 1982a).  This is a reasonable assumption if a significant percentage of
the psoralen were actually bound to the protein.  However, the argument would be
nullified if the majority of the psoralen were found to be bound to the small
amount of nucleic acids in the fibers.  Autoradiograms showing psoralen binding
in the lens suggest that the latter is in fact the case (Lerman et al, 1981).
Examination of those autoradiograms shows pronounced binding of psoralen in the
epithelium plus discrete binding to the degenerating fiber nuclei.  In
comparison, binding to any other portions of the fibers is not distinguishable
from background grains.

    Autoradiography has also demonstrated psoralen binding in the nucleated
layers of the retina and cornea but no binding to the non-nucleated layers of
these tissues (Lerman et al, 1981).  It is rather difficult to reconcile the
failure of psoralen to bind to the non-nucleated regions of the retina and
cornea with the suggestion that it binds significantly to cortical fiber
proteins.  A trivial explanation for this difference is that lens proteins are
somehow "different" from corneal and retinal proteins, making the former
susceptible to psoralen binding and the latter not.  This argument is rather
difficult to reconcile with the requirement of oxygen for photo-induction of
psoralen linking to proteins, a requirement that would lead one to expect
greater binding to the proteins in the highly oxygenated retina and cornea
compared to the relatively anoxic lens.

    If lens proteins are the major target of psoralen's action on the lens, then
one would expect that the cortical fibers would be disrupted as a primary event.
In fact, however, histologic observations have shown that the first target of
psoralen's effects are the epithelial cells (Jose, Kock, and Respondek 1982).
Only after very considerable damage is observable in the epithelium is any
damage detectable in the lens cortex which is consistent with DNA as a primary
target of psoralen damage.  That psoralen plus UVB can induce DNA damage in the
lens is shown by the finding that the combination induces DNA repair in lens
epithelial cells (Jose and Yielding 1979).  It will be interesting to resolve
the question as to the relative roles of specific macromolecular targets in
development of psoralen-induced cataracts.

    As with psoralen, most other studies of UV effects on the lens have directed
their major emphasis toward examinations of changes in lens proteins.  But, as
with the psoralen study, there are compelling reasons to consider other
macromolecules a targets of UV damage.  Lens lipids have been mentioned
previously and interest in lipids peroxidizing effects is of current interest in
many laboratories.   However, interest in the effect of UV and photosensitizers
on lens DNA is currently very limited.  The role of DNA damage in development of
lens cortical opacities is generally overlooked or rationalized away.  Some have
assumed that there is so little DNA in the lens that even if it were a target,
any effects on it would be overwhelmed by effects on lens proteins,  There are,

-------
                                     10-28


however, good reasons to consider the lens to he similar to other biological
systems in which DNA is a significant target of UV damage.

    If one considers the action spectrum that has been determined in those
studies in which specific wavelengths of UV were isolated to examine their
cataractogenic potential, it is noted that these lie in the range of 290 to 320
nm (Bachem 1956; Pitts, Hacker, and Parr 1977).  These are the same wavelengths
that most readily produce thymidine dimers in skin (Pathak, Kramer, and
Guengerich 1972).  Shorter wavelengths are blocked from deeper layers of the
skin by the stratum corneum, quite analogous to corneal UV absorption which
protects the lens.  These are the same wavelengths which most readily induce
malignant transformation and skin cancer (Freeman 1975).  UV radiation up to 320
nm also can induce DNA repair synthesis in cultured fibroblasts (Ichihashi and
Ramsey, 1976).  Longer wavelengths of UV do not induce skin cancer, although
they may, if applied very intensely, induce strand breaks and other damage in
DNA (Webb and Peak 1981, Harm 1978).  Endogenous photosensitizers are likely to
be involved since DNA itself does not absorb appreciably at that wavelength.
Long wave UV may potentiate the action of short wave UV in cancer induction and
it can also be noted that 364 nm has been found to inactivate DNA repair
mechanisms (Tyrell 1976).

    With all the interest in tryptophan as an endogenous photosensitizer of
protein damage in the lens, it is interesting to point out parenthetically that
photo-oxidation of tryptophan has also been found to induce DNA damage in lower
organisms as a consequence of production of H202 (Ananthaswamy and Eisenstark,
1976).  The possibility that those free radical species that have been found to
be generated from UV excitation of tryptophan in the lens may also act on lens
DNA as a target should not be overlooked.  Furthermore, tryptophan photoproducts
have also been found to bind to DNA (Glazer, Rincon, and Eisenstark 1976) and to
inhibit strand rej oining in damaged DNA (Yoakum et al. 1974) .

    Another reason to consider that DNA is a target of UV damage in the lens is
the finding that 300 nm radiation can induce DNA repair synthesis in the lens
epithelium, even when the irradiation is applied through the intact cornea
(Jose, Kock, and Respondek 1982; Brenner and Grabner 1982) and DNA repair
synthesis is an indication that DNA was damaged.  This point is often
misunderstood, repair is not a perfect process and it should not be assumed that
lens DNA will be absolutely protected.  Mutational and lethal events can occur
any time that DNA is damaged and repair is undertaken; thus, anything that can
induce DNA repair must be recognized for its primary damaging effects.  Errors
induced in the genome of lens cells may manifest themselves as mutational events
in the target epithelial cell.  Such mutational events would be cumulative over
an individual's lifetime.  Those cells that have undergone mutations might loose
their capacity to differentiate into normal lens fibers and, for example, pile
up at the posterior pole.  The later is seen in the senile cataract (Streeten
and Eshaghian 1978) as well as in animals exposed to near UV radiation (Zigman
and Vaughan 1974) .  The cells may also not carry out their normal functions such
as maintaining less osmolarity.  Then we might develop what appeared to be an
"osmotic" cataract, when in fact, the underlying mechanism was genetic.

POTENTIAL CHANGES IN SENILE CATARACT PREVALENCE FOR CHANGES IN UV-B
    Epidemiological studies have identified a correlation between the prevalence
of various types of cataracts in humans and the flux of sunlight or ultraviolet
radiation reaching the earth's surface (Hiller, Giacometti and Yuen 1977,
Zigman, Datiler, and Torczynski 1979; Taylor 1980, Hollows and Moran 1981).
Hiller, Sperduto and Ederer (1983) developed a multivariate logistic risk

-------
                                      10-29
 function  that  describes  the  correlation  found between  the prevalence  of  senile
 cataracts and  the  flux of UV-B and other risk factors.  The results of this
 study may indicate the magnitude of change  in the prevalence of  senile cataracts
 that could be  associated with changes  in UV-B flux due to ozone  depletion.

    Killer,  Sperduto and Ederer based  their analysis on the 1971-1972 National
 Health and Nutrition Examination Survey  (HANES) general medical  and
 ophthalmological examinations of over  10,000 persons,  ages 1-74  years.   Using
 the HANE.S data, persons  were assigned  to one of two mutually exclusive groups:
 (1) cataract or aphakia  (either eye),  and (2) neither  cataract nor aphakia
 (either eye).  Cataract  was  defined as "senile lens changes (cortical, nuclear,
 posterior-subcapsular, or other) consistent with best  corrected  visual acuity of
 20/30 (6/9)  or worse" (Hiller, Sperduto  and Ederer (1983) p. 240).  This
 definition differs from  the  HANES survey, which used a visual
 acuity of 20/25 (6/7.5)  or worse.  Aphakia  was diagnosed when "the lens  had been
 surgically removed and there was no history of congenital, traumatic  or
 secondary cataract" (Hiller, Sperduto  and Ederer (1983) p. 240).  The study by
 Hiller, Sperduto and Ederer  included HANES  data on a total of 2,225 persons
 between the  ages of 45 and 74 years who  had resided at least one half of their
 life in the  state  where  the  HANES examination took place.  Of these 2,225
 people, 413  (18.6  percent) were placed in the cataract or aphakia outcome
 category.

    The UV-B data  were developed by NOAA for the 35 HANES locations used in the
 study based  on a statistical analysis of UV-B data collected at  10 locations
 using Robertson-Berger meters (RB-meters).  The statistical analysis
 incorporates season, latitude, elevation, weather (clouds), and  haze.
 Subsequent validation of the estimates at six locations indicated that the
 differences  between the  estimated and observed mean daily flux average about
 seven percent.

    These  data on  UV-B and outcome (i.e., cataract) were used in conjunction
 with demographic and medical history data to estimate  the following multivariate
 logistic  risk  function:
                       1 + exp(-a-bX-. . .-
                                     i
where P is the probability (or risk) of having a cataract, and X. are risk
factors.  In addition to UV-B, the following risk factors were analyzed:  age;
race; sex; education; diabetes; systolic blood pressure ;  and residence  (urban,
rural) .

    Exhibit 10-12 displays the standardized regression coefficients estimated
for each of the risk factors.  Positive coefficients indicate factors that are
correlated with increased risk, negative coefficients indicate factors  that are
correlated with decreased risk.  The coefficients presented in the exhibit are
"standardized," meaning that they represent the expected change in the  logit of
P (equal to In (P/l-P)) for a one standard deviation change in the risk factor.
Standardization of the coefficients allows the relative importance of the risk
factors  to be identified by the relative size of the standardized coefficients.

    As shown in Exhibit 10-12, three risk functions were estimated:  (1)
univariate (outcomes as a function of the risk factor); (2) bivariate (outcome

-------
                                     10-30
as a function of the risk factor and age);  and (3) outcome as a function of all
the risk factors simultaneously.  For all three formulations, UV-B is
statistically significant,  and positively correlated with the increased risk of
being in the cataract outcome category.

    Using the multivariate risk function coefficients, and the mean values for
all the risk factors other than UV-B, the change in the prevalence of cataract
for each 1.0 percent change in UV-B is estimated to be approximately 0.5
percent.  This relationship holds for changes in UV-B as large as minus 20
percent to plus 30 percent.  Outside of this range, reductions in UV-B are
associated with less of a reduction in cataract prevalence, and increases in
UV-B are associated with larger increases.

    Of note is that this estimated relationship between UV-B and cataract
prevalence varies with age; UV-B has a larger effect on prevalence (on a
percentage basis) among younger individuals.  Exhibit 10-13 displays the percent
increase in cataract prevalence expected due to increases in UV-B, for peoples
of different ages.  As shown in the exhibit, the percentage increase in
prevalence due to changes in UV-B are estimated to be larger for 50 year olds
than for 70 year olds.

    Although the effect of UV-B on prevalence is estimated to be larger at
younger ages (on a percentage basis) using the multivariate risk function, the
prevalence of senile cataracts is known to increase substantially with age.
Leske and Sperduto (1983) report the prevalence of senile cataracts in both
sexes found in the Framingham Eye Study to be as follows:  52 to 64 years old --
4.5 percent; 65 to 74 years old -- 18.0 percent; 75 to 85 years old -- 45.9
percent.  These prevalence estimates use the same definition of cataracts as
used by Hiller, Sperduto, and Ederer.  (Larger prevalence rates are reported by
Leske and Sperduto based on HANES data.   These estimates, however, use a
definition of cataract that includes a decrease in vision to 20/25 (6/7.5),
instead of the 20/30 (6/9) used in both the Framingham study and the Hiller,
Sperduto, and Ederer risk study.)  Because cataracts are more prevalent in older
individuals, increases in the actual number of cases of cataracts would likely
be larger for older individuals, even though the percentage increase in risk has
been estimated to be larger for younger individuals.

    Using the prevalence data cited above,  the prevalence of cataracts in the
U.S. population is on the order of 9.3 million.  Using the multivariate risk
function (with all values set to their means except age and UV-B) the
hypothetical increased prevalence for 1985 that would have occurred had the
entire population experienced 1.0 percent ozone depletion can be estimated.  For
a 1.0 percent depletion, the annual UV-B flux measured on the RB-meter has been
estimated to increase by approximately 0.83 percent (see Serafino and Frederick
(in press) and Chapter 17 for a discussion of the relationship between ozone
depletion and UV flux).  A 0.83 percent increase  in UV-B is associated with
increases in cataract prevalence that varies by age.  Across all the ages,
prevalence would be expected to be about 0.26 percent higher, or about 24,000
cases, had ozone been depleted by 1.0 percent.

    Of note is that the RB-meter measure of UV radiation may not be the
appropriate action spectrum to use to evaluate the potential biological effects
of increased UV-B such as cataract.  For example,  the DNA action spectrum may be
preferred.  Even though the RB-meter and DNA action spectra are highly

-------
                                     10-31


                                EXHIBIT  10-12

               Standardized Regression Coefficients for Cataract
fi/
Risk Factor
Age
Race
Sex
Education
Diabetes
Systolic Blood Pressure
UV-B
Residence
b/
Mean
61.97
1.22
1.52
2.87
0.08
146.3
3.59
1.37
c/
Univariate
1.22^
0.18
0.07f
-0.43*
0.25*
0.33*
0.19*
0.18
d/
Bivariate
f
0.20
0.08
-0.25*
0.23
0.15s
0.20*
0.21
e/
Multivariate
1.20f
n
0.13s
0.08
-0.14s
f
0.21
0.08
0.13s
0.19
  a/ Values for categorial risk factors:   race:   1 — white,  2 = black;
sex:  1 - male, 2 - female;  education:  1  = <5 grades; 2=5-8 grades; 3 = 9-11
grades; 4 - 12 grades;  5 = college;  diabetes:  0= absent,  1 = present;
residence:  1 = urban;  2 = rural.

  b/ Mean value for the risk factor  in the 2,225 persons in the study.

  c/ Each risk factor analyzed separately.

  d/ Each risk factor analyzed with  age only.

  e/ All risk factors analyzed simultaneously.

  f/ p(two sided) <0.005.

  £/ p(two sided) <0.05.


  Source:   Hiller,  R.,  R.  Sperduto and F.  Ederer (1983), "Epidemiologic
           Associations with Cataract in  the 1971-1972 National Health and
           Nutrition Examination Survey,"  American Journal of Epidemiology.
           Vol. 118,  No. 2,  pp.  239-249.

-------
Percent
Increase
in
Cataract
Prevalence
                                        10-32



                                    EXHIBIT 10-13


            Estimated Relationship Between Risk of Cataract and UV-B Flux
20


18-


16-


14


12


10


 8


 6


 4-


 2-


 0
                                                                      AGE

                                                                      AGE
                                   50

                                   60
                             AGE = 70
                                10
15
                                     20
25
30
                              Percent Increase in UV-B Flux
   Increased UV-B flux (measured with an RB-meter)  is  associated with increased
   prevalence of cataract.   The percent change  in prevalence varies by age.


   Source:   Developed from data presented  in  Killer R.,  R.  Sperduto, and F.  Ederer
   (1983),  "Epidemiologic Associations with Cataract in the 1971-1972 National
   Health and Nutrition Examination Survey,"  American Journal of Epidemiology, Vol.
   118,  No.  2, pp. 239-249.

-------
                                     10-33
correlated in the range of current observations,  because the DNA action spectrum
is more heavily weighted toward shorter wavelengths,  it increases more rapidly
with decreases in ozone levels;  a 1.0 percent depletion would lead to
approximately a 2.0 percent increase in UV-B.  Using the 2.0 percent increase in
UV-B would yield about a 0.6 percent increase in prevalence, or about 57,000
cases.

    There are various important limitations in the use of these estimates and
data.  The correlation between UV-B and cataracts reported by Killer, Sperduto
and Ederer does not prove a causal connection - -  other (unknown) factors could
be playing a role.  These (unknown) factors would have to be correlated with
UV-B flux.  Also, the study does not have estimates of individual lifetime UV-B
exposure, thereby limiting the strength of the evidence for the association
between UV-B exposure and cataracts.   Additionally, the sample population
analyzed may not be representative of the entire U.S. population.  Finally, the
outcome category used in the study does not differentiate between different
types of cataracts, some of which may be more strongly related to UV-B exposure
than others.

    Confidence in the estimates developed here are strengthened by several
considerations.  The correlation between UV-B flux and sunlight flux is high,
and a correlation between sunlight and cataracts has also been found in
Australia (Taylor 1980) and in China (Mao and Hu 1982).  An association between
UV-B exposure and cataract has also been demonstrated in laboratory animals.
Therefore, although considerable investigation remains to be performed,
indications are that the association between UV-B and cataract is a reasonable
basis for evaluating potential impacts due to increased UV-B flux associated
with ozone depletion.

OTHER EYE DISORDERS2

Stable Retinal Disorders

    Evidence has been adduced which provides a chain of evidence on which to
postulate that optical radiation may contribute to human retinal disorders which
are not progressive, that is, stable retinal problems.  This chain starts with
the knowledge that UV-B, UV-A and light are absorbed by various retinal tissues
(Wolbarsht 1976; Warner 1982).  Moreover, the amount of radiant energy incident
on the retina is dependent on the transparency of the cornea and lens, and on
pupil diameter.  Next a large number of bioeffects of these wavelengths have
been identified which are hazardous to the retina (Marshal 1970; Williams and
Baker 1980).  Further, empirical estimates have been made of the domain of time,
intensity, and wavelength which can induce damage to the pigment epithelium, the
photoreceptors and the inner retina.   These estimates provide sufficient
quantitative evidence to establish safety standards to protect people from the
short-term retinal damage which could be induced by lasers and other intense
sources of optical radiation (Sliney and Wolbarsht 1980; Pitts 1973).  There is,
moreover, sufficient evidence of additive and persistent changes in the monkey
and human retina to hypothesize that some stable long-term visual impairments in
humans may be related to optical radiation.  There are reports that optical
radiation produces retinal damage hours, days, months and years after
irradiation.  Although sparse, this evidence is instructive.  For example,
sungazing by humans has been reported to produce persistent, evolving and
        This section in boldface type is taken from Waxier, M. (in press).

-------
                                     10-34
delayed alterations in the visual system.  Cavonius, Elgin, and Robbins (1974)
reported that aniseikonia may persist for years.  Pigmentary patches and blind
spots have been reported to evolve in a complicated fashion over many months
after solar retinopathy (Lowenstein and Steel 1941) .  In addition, late
complications resulting in reduced visual acuity have been described (McFaul
1969).  Furthermore there are several reports in humans of prolonged, and even
cumulative, impairments in dark adaptation following extended viewing by humans
of skylight (Hecht et al. 1948; Clark, Johnson, and Dreher 1946) or bright blue,
violet (Brindley 1953) and ultraviolet (Wolfe 1949) radiation.  Also there is at
least one report of persistent visual impairment resulting from the accidental
exposure to ultraviolet radiation from a welding arc (Naidoff and Sliney 1974).
Even more important is the description by Tso and Woodford (1983) of sub-RPE
neovascularization and late impairment in fluid transport after several years of
evolving changes in retinal pathology following excessive exposure to optical
radiation.

    Some of these effects are the result of injury to the retina which was
ophthalmoscopically visible, some of these effects resulted from exposure to
optical radiation which did not produce ophthalmoscopically visible damage.
Most important are two demonstrations that the addition of subthreshold
exposures separated by several days induces retinal damage (Greiss and
Blankenstein 1981; Kuwabara and Okisake 1976).  The data strongly suggest that
retinal damage induced by optical radiation can cumulate over days.  This is
especially important since the concern about long-term visual health problems
necessarily involves consideration of the intermittent nature of exposure to
emissions from multiple sources over an extended period of time.

Retinal Degeneration

    There are some reasons to suspect that retinal damage induced by ultraviolet
radiation can be unstable, that is, produce progressive damage.  The retina of
monkeys can be damaged by ultraviolet radiation (Ham 1984; Sykes et al. 1981).
This damage to the retinal pigment epithelium, to the photoreceptors and the
neurons is dose dependent, photochemically mediated and cumulative over hours
and days.  Although very few long-term observations have been made on the
primate retina following excessive exposure to optical radiation, there are
theoretical reasons to suspect that a retina compromised by disease or age would
be more vulnerable to any degenerative changes which could be induced by optical
radiation  (Young 1981).  The combination of short-wavelength optical radiation,
oxygen, chromophores, and photosensitizers in the retina is potent with
possibilities for producing retinal degeneration, although the only direct
evidence of the exacerbation of retinal degeneration by optical radiation has
been obtained in rodents (LaVail and Batelle 1975).  In addition, the higher
retinal doses of UV received by aphakes and pseudophakes make these individuals
especially vulnerable to cystoid macular edema  (Kraff et al. 1985).  Fluorescein
leakage occurs following exposure of the retina of the aphakic monkey eye to
UV-A radiation, but comparable studies using UV-B radiation have not been
conducted.  Therefore, retinal degeneration should be considered a risk of
excessive exposure to ultraviolet radiation.

    Very little research has been conducted on primates, on diurnal animal
models of human retinal degenerations, or on retinas compromised by chemicals,
age or other agents.  The striking evidence of long-delayed and persistent
changes in the monkey retina following optical radiation, coupled with some
clinical evidence in humans, suggests that research on primates needs emphasis.

-------
                                     10-35
Aging Disorders

    Deterioration of the visual system as individuals age is a fact (Owsley et
al. 1986).  The likelihood that optical radiation contributes to this
deterioration is high for several reasons.  First, the cellular pathology
observed in older individuals is similar in appearance to radiation induced
damage (Kuwabara 1978).  Second, the ophthalmoscopic appearance of the aged
retina is sometimes similar to that induced by optical radiation (Klein 1958).
Third, some of the kinds of vision loss described in the aged are consistent
with the losses which could be produced by optical radiation damage (Jaffe, de
Nonestario, and Podgor 1982; Harweth and Sperling 1975).  Fourth, the
accumulation of lipofuscin materials in the RPE and other changes in the retina
suggest that the aged retina has a lessened capability of repair (Feeney-Burns,
Herman, and Rothman 1980).

    Optical radiation may have quite different visual health consequences for
eyes with and without ocular disease.  Optical radiation may accelerate the
deterioration of central vision in older individuals with central retinal
disease (Hyman et al. 1983), whereas in those free of ocular disease optical
radiation may impair paramacular and peripheral vision more than foveal vision.
This interpretation is offered to reconcile the facts of visual loss outside the
macular in disease-free aged eyes (Owsley et al. 1986) with Young's prediction
that the central retina will exhibit the major effects of lifetime retinal
exposure to optical radiation (Young 1981).  This ad hoc explanation is not very
satisfactory but it does indicate the need to temporize theory with data.  Other
interpretations also are possible, e.g., those individuals with central retinal
disease may have had a different exposure history or some genetic vulnerability.

    Given these facts and theoretical considerations, exposure of the aged
retina to ultraviolet radiation in excess of amounts permitted by the phakic eye
probably is hazardous.  The retina of the aged individual is exposed to more
radiation than phakics receive when the crystalline lens is extracted without
placing a comparable filter in front of the retina, even when ordinary
sunglasses are used.

    Research is needed to determine the extent to which optical radiation is a
factor in the deterioration of the vision of the elderly.  In the meantime, we
should take practical steps to protect our eyes from excessive optical
radiation, and thus prolong our visual lifetime.

Development Disorders

    While there are several reasons to suspect hazardous effects of optical
radiation on visual development, almost no research has been conducted on the
developing primate retina.  In fact, there is very little experimental exposure
data on the developing retina of any species.  Glass et al. (1985) have reported
that the probability of retinopathy of prematurity (ROP) is highest in infants
exposed to more light during their hospital stay.  This evidence confirms a
prediction by Wolbarsht et al. (1983) about ROP and light and is consistent with
the evidence for the interaction of oxygen and light in retinal damage (Ruffolo
et al. 1984).  The action spectrum for this effect is unknown but UV-B and UV-A
probably play a role.  Furthermore,  there are a large variety of sources and
situations in which infants are exposed to optical radiation.  Some of the same
evidence which gave rise to suspicions about the role of optical radiation in
retinal degeneration and visual aging is relevant to the potential hazards to
visual development:  short-wavelength radiant energy incident on the retina,

-------
                                     10-36
short-wavelength chromophores in the retina,  dose-dependence of retinal damage,
cumulative effects,  long duration effects,  a^d delayed effects.  In addition,
there are special properties of the developing visual system which might
decrease injury thresholds, such as more transparent ocular media and different
densities of chromophores and screening pigments, or which might increase the
vulnerability of infants, such as the long period of post-natal retinal
development (especially foveal) and early dependence on non-foveal visual
fields.  Thus, the long periods required for many visual processes to mature
suggests a window-of-vulnerability during which time optical radiation might
alter later visual performance, especially if optical radiation alters the
spatial and temporal summation properties of the neural retina.

    Even though there is little experimental data on the effects of optical
radiation on the developing retina, the photobiological and visual science
evidence is sufficient to postulate that disorders of visual development are
risks of optical radiation.

Retinal Problems

    The only human data from which one could derive an estimate of the radiant
exposure of UV which could damage the retina is Hecht et al. (1948) and Clark,
Johnson, and Dreher (1946).  These data showed that staring at skylight for
several hours induced an abnormal retardation of dark adaptation.  Since
rhodopsin has an absorption band in the UVA (Kurzel, Wolbarsht, and Yananashi
1977), this deficit could have been due to the ambient outdoor UVA.  If the
ambient level of UVA at the cornea was about 9 x 10   Jem  , as suggested by Ham
andJUueller (1982), then the retinal radiant exposure might have been about 9 x
10   Jem   in the Hecht and Clark studies.  Experimental estimates of UV retinal
damage in monkeys have been few.  Schmidt and Zuclich (1980) found the threshold
at 325 nm to be 10 Jem  .  Recent evidence in phakic rabbits (Pitts, Bergmanson,
and Chu 1983) and rats (Rapp, Jose, and Pitts 1985) suggest that the radiant
exposure necessary to damage the retina is even lower at 300 nm.  The threshold
for 300 nm damage to the monkey retina currently is approximately 0.6 Jem „.
Ham and Mueller £1982) found the aphakic monkey has thresholds of 5.0 Jem   at
325 nm, 5.4 Jcnf  at 350 and 8.1 Jem"  at 380 nm.

-------
                                     10-37
REFERENCES
Ananthaswamy HN and A Eisenstark (1976)  "Near-UV-induced breaks in phage DNA:
sensitization by hydrogen peroxide (a tryptophan photoproduct)."  Photochem.
Photobiol. 24:439-442.

Anderson El, DD Wright and A Spector (1979)  "The state of protein sulfhydryl
groups in normal and cataractous proteins II cortical and nuclear regions." Exp.
Eye Res. 29: 233-243.

Augusteyn RC (1981)  In Mechanisms of Cataract Formation in the Human Lens.  G.
Duncan (ed).  Academic Press.

Augusteyne R (1974)  "Human lens albuminoid."  Jap J. Ophthalmol. 18: 127-134.

Augusteyne RC (1975)  "Distribution of fluorescence in the human cataractous
lens."  Ophthalmic Res. 7: 217-224.

Bachem A (1956)  "Ophthalmic ultraviolet action spectrum."  Amer. J. Ophthal.
41: 969-980.

Bando M, A Nakauima and K Satoh (1975)  "Coloration of human lens protein."
Exp. Eye Res. 20: 489-492.

Barker FM II (1979)  The transmittance of the EMS from 200nm to 2500nm through
the optical tissues of the eye of the pigmented rabbit.  MS Thesis University of
Houston, Houston, Texas.

Barr RE and EL Roetman (1974) "Oxygen gradients in the anterior chamber of
anesthetized rabbits." Invest. Ophthal. 13: 386-392.

Benedek GB (1971) "Theory of transparency of the eye." Applied Optics 10:
459-473.

Berler D and R Peyser (1982)  "Light intensity and visual acuity following
cataract surgery." Ophthalmology 89 (Suppl.) L 117.

Bhuyan KG and DK Bhuyan (1977)  "The relative functions of superoxide dismutase
and catalase in the eye."  IRCS Medical Science 5: 279.

Boettner EA and JR Wolter (1962)  "Transmission of the ocular media."  Invest.
Ophth. 1(6): 766-783.

Borkman RF (1977)  "Ultraviolet action spectrum for tryptophan destruction in
aqueous solution."  Photochem. Photobiol. 26: 163-166.

Borkman RF and S Lerman (1977)  "Evidence for a free radical mechanism in aging
and UV irradiated ocular lenses."  Ep. Eye Res. 25:  303-309.

Borkman RF, A Dalrymple and S Lerman (1977)  "Ultraviolet action spectrum for
fluorogen production in the ocular lens."  Photochem. Photobiol. 26: 129-132.

-------
                                     10-38
Borkman RF, JD Tassin and S Lerman (1981) Exp. Eye. Res. 32: 747-754.

Brenner W and G Grabner (1982)  "Unscheduled DNA repair in human lens epithelium
following in vivo and in vitro UV irradiation."  Ophthal. Res. 14: 160-166.

Brilliant LB, NC Grasset, RP Pokhrel, A Kolstad, JM Lepkowski, GE Brilliant, WN
Hawks and B Pararajasegaram (1983)  "Associations among cataract prevalence,
sunlight hours and altitude in the himalayas." Amer. J. Epidem. 118: 250-264.

Brindley GC (1953)  "The effects on colour vision of adaptation to very bright
lights."  J. Physiol. 122:  332.

Buckingham RH and A Pirie (1972).   "The effect of light on lens proteins in
vitro."  Exp. Eye Res. 14: 297.

Castineiras SG, J Dillon and A Spector (1979)  "Effects of reduction on
absorption and fluorescence of human lens proteins."  Exp. Eye Res. 29: 573-575.

Cavonius, CR, S Elgin, and DO Robbins (1974)  "Thresholds for damage to the
human retina by white light."  Exp. Eye Res. 19:  543-548.

Chand D, HK El-Aguizy, RD Richards and SD Varma (1982)  "Sugar cataracts in
vitro: implications of oxidative stress and aldose reductase I."  Exp.  Eye Res.
35: 491-497.

Chand D and SD Varma (1983)  "Human lens fatty acids composition."  Invest.
Ophthal. and Vis. Sci. (ARVO  suppl.) 24: 31.

Chand D, SD Varma said RD Richards (1982)  "Lens lipid peroxidation.  Conjugated
acylpolyene and schiff bases in human cataracts."  Invest. Ophthal. and Vis.
Sci. (ARVO suppl.) 20: 218.

Chylack LT Jr. (1978)  "Classification of human cataracts."  Arch. Ophthal. 96:
888.

Chylack LT, MR Lee, WH Tung and H-M Cheng (1983) "Classification of human senile
cataractous change by the American cooperative Cataract Research Group (CCRG)
Method." Invest Ophthalmol. and Vis Sci, 24: 424-431.

Clark, BA, ML Johnson and RE Dreher (1946)  "The effect of sunlight on dark
adaptation."  Am. J. Ophthal. 29:   823-836.

Cooper GF and JG Robson (1969) J.  Physiol. 203: 411-17.

Crylin MN, A Pedvis-Leftick and J Sugar (1980)  "Cataract formation in
association with ultraviolet photosensitivity." Annals of Ophthalmol. 12:
786-790.

Dilley KJ and A Pirie (1974)  "Changes to the proteins of the human lens nucleus
in cataract."  Exp. Eye Res. 19: 59-72.

Dillon J and A Spector (1980)  "A comparison of aerobic and anaerobic
photolysis of lens proteins."  Exp. Eye Res. 31: 591-599.

Dillon J, A Spector and K Nakanishi (1976) Nature 259: 422-433.

-------
                                     10-39
Dillon J, M Garner, D Roy and A Spector (1982)  "The photolysis of lens protein:
molecular changes."  Exp. Eye Res. 34: 651-658.

Doniger J, ED Jacobson, K Krell and JAD Paolo (1981)  "UV light action spectra
for neoplastic transformation and lethality of Syrian hamster embryo cells
correlate with spectrum for pyrimidine dimer formation in cellular DNA."  PNAS
78: 2378-2382.

Duke-Elder S and PA MacFaul (1972) System of Ophthalmology V xiv pt 929-932.

Duke-Elder WS (1926) Lancet 1: 1188-91.

Fecondo JV and RC Augusteyn (1983) Exp. Eye Res. 36: 15-23.

Feeney-Burns L, ER Herman and H Rothman (1980)  "Lipofuscin of human retinal
pigment epithelium."  Am. J. Ophth. 90: 783-791.

Fraunfelder F, J Megaw and S Lerman (1983)  "Further studies on allopurinol in
human cataracts."  Invest. Ophthalmol. Vis. Sci. 24 (Suppl.): 3.

Freeman RG (1975)  "Data on the action spectrum for UV carcinogenesis."  J. Nat.
Cancer Inst. 55: 1119-1121.

Garner MH and A Spector (1980)  "Selective oxidation of cysteine and methionine
in normal and senile cataractous lenses."  Proc. Natl. Acad. Sci. 77: 1274.

Girgus JS, S Coren and C Porac (1977)  "Independence of in vivo human lens
pigmentation from UV-light exposure."  Vision Res. 17: 749-750.

Glass PA, G Avery, KN Sivasubramanian, M Keys, AN Sostek and DS Friendly (1985)
"Effect of bright light in the hospital nursery on the incidence of retinopathy
of prematurity."  N. Engl. J. Med. 313: 401-

Glazer LD, A Rincon and A Eisenstark (1976)  "The binding of near UV
light-induced tryptophan photoproducts to DNA."  Biochem. Biophys. Acta 418:
137-145.

Goosey JD, JS Zigler and JH Kinoshita (1980)  "Cross-linking of lens crystallins
in a photodynamic system: a process mediated by a singlet oxygen." Science 208:
1278-1280.

Goosey JD, ME Allison and CA Garcia (1983)  "Lipid peroxidative mechanism for
posterior subcapsular cataract formation in the rabbit."  ARVO Abstracts,
Invest. Ophthal. Vis. Sci. 24: 75.

Greiss GA and MF Blakenstein (1981)  "Additivity and repair of actinic retinal
lesions."  Invest. Ophth. and Vis. Sci. 20: 803-807.

Ham WT Jr (1984) "Basic mechanisms underlying the production of photochemical
lesions in the mammalian retina."  Curr.  Eye Res. 3: 165-174.

Ham WT Jr and HA Mueller  "Action spectrum for retinal injury from near
ultraviolet radiation in the aphakic monkey."  Am. J.  Ophthalmol. 93:299-

Harding JJ (1972)  "The nature and origin of the urea-insoluble protein of human
lens."  Exp. Eye Res. 13: 33-40.

-------
                                     10-40
Harding JJ (1981)  In Molecular and Cellular Biology of the Eye Lens.  H.
Bloemendal (ed).  Wiley Interscience.

Harding JJ and KJ Dilley (1976) "Structural proteins of the mammalian lens: a
review with emphasis on changes in development, aging and cataract."  Exp. Eye
Res. 22: 1-73.

Harm, W (1978) "Evidence for photoenzymatically repairable, lethal 'non-dimer'
photoproducts formed in E. coli cells by 265 nm radiation or by sunlight 360
nm."  Mutat. Res. 51: 301-310.

Harweth RS and HG Sperling (1975) "Effects of intense visible radiation on the
increment-threshold spectral sensitivity of the rhesus monkey eye."  Vis. Res.
15: 1193-1204.

Hata N and Hockwin O (1977) Ophthal. Res. 9: 256-262.

Hecht S, CD Hendley, S Ross, and PV Richmond (1984)  "The effect of exposure to
sunlight on night vision."  Am. J. Ophthal. 31: 1573-1580.

Killer R, L Giacometti and K Yuen (1977)  "Sunlight and cataract:  An
epidemiological investigation."  Amer. J. Epidemiol. 105: 450-459.

Hiller R, RD Sperduto and F Ederer (1983)  "Epidemiologic associations with
cataract in the 1971-1972 national health and nutrition examination survey."
118: 239-249.

Hochheimer BF (1981)  "A possible cause of chronic cystyc maculopathy the
operating microscope."  Ann. Ophthalmol. 13: 153-155.

Hockwin O, V Dragomirescu and S Lerman (1983)  "In vivo age-related changes in
normal and cataractous human lens density." Acta XXIV International Cong.
Ophthalmol. In Print.

Hoenders HL and H Bloemendal (1981)  In Molecular and Cellular Biology of the
Eye Lens.  H. Bloemendal (ed).  Wiley Intersciences.

Hollows F and D Moran (1981)  "Cataract -- The Ultraviolet Risk Factor."  Lancet
ii: 1249-1251.

Hymen LG, AM Lilienfeld, et al. (1983)  "Senile macular degeneraton:  A
case-control study."  Am. J. Epidemiol. 118: 213-227.

Ichihashi M and CA Ramsey (1976)  "The action spectrum and dose response studies
of unscheduled DMA synthesis in normal human fibroblasts."  Photochem.
Photobiol. 23: 103-106.

Imbrie CW and TM Murphy (1982)  "UV action spectrum (254-405nm) for  inhibition
of K-stimulated and  adenosine triphophosphatase from the pasma membrane of Rosa
damascena." Photochem. Photobiol. 36: 537-599.

Jaffe MJ, FM de Monestario and MJ Podgor  (1982)  "Retinal aging and blue-
sensitivity cone functioning."  Invest. Ophth. Vis.  Sci.  22  (Suppl.): 218.

-------
                                     10-41
Jernigan HM, HN Fuki, JD Goosey and JH Kinoshita  (1981)  "Photodynamic effects
of rose bengal or riboflavin on carrier mediated  transport systems in rat lens."
Exp. Eye Res. 32: 461-466.

Jose JG (1978)  "The role of DNA damage, its repair and its misrepair in the
etiology of cataract: a review."  Ophthal. Res. 10: 52-62.

Jose JG and KL Yielding (1977)  "Unscheduled DNA  synthesis in lens epithelium
following ultraviolet irradiation."  Exp. Eye Res. 24: 113-119.

Jose JG and KL Yielding (1979)  "Photosensitive cataractogens, chlorpromazine
and methoxypsoralen cause DNA repair synthesis in lens epithelial cells."
Invest. Opthal. Vis. Sci. 17: 687-691.

Jose JG, HR Kock and A Respondek (1982)  "Histological observations of the
lenses of psoralem •*- DV-A treated albino rats and a theory as to the underlying
mechanism."  Graefe's Arch. Clin. Exp. Ophthalmol. 219-44-

Kahn HA and RC Milton, Revised Framingham eye study prevalence of glaucoma and
diabetic retinopathy.

Kahn HA, HH Leibowitz, JP Ganley, MM Kini, T Colton, RS Wickerson and TR Dawaber
(1977a)  "The Framingham eye study:  I.  Outline  and major prevalence findings."
Am. J. Epidemiol. 106: 17-32.

Kahn HA, HM Leibowitz, JP Ganley, MM Kini, T Colton, RS Wickerson and TR Dawaber
(1977b)  The Framingham eye study:  II.  Association of ophthalmic pathology
with single variables previously measured in the  Framingham Heart study."  Am.
J. Epidemiol. 106: 33-41.

Kalustian A, M Sun and S Zigman (1978)  "N-UV light sensitized inactivation of
glutathione reductase."  Am. Soc. Photobiol.  (Abstracts-June) 112-113.

Kinoshita J (1974)  "Mechanisms of cataract formation."  Invest. Ophth. Vis.
Sci. 18: 713-724.

Kinsey VE (1948)  "Spectral transmission of the eye to ultraviolet."  Arch.
Ophth. 39: 508-513.

Klien B (1958)  "Diseases of the macula."  Arch. Ophthalmol. 60: 175-186.

Kraff MC,  DR Sanders, LM Jampol and HL Lieberman  (1985)  "Effect of an
ultraviolet-filtering intraocular lens on cystoid macular edema."  Ophthalmology
92:366-

Kramer SG (1980)  "Epinephrine distribution after topical administration to
phakic and aphakic eyes."  Trans-Am Ophthalmol. Soc. 75: 947-982.
Kurtin WE and JA Zuclich (1978) "Action spectrum for oxygen-dependent
near-ultraviolet induced corneal damage."  Photochem. Photobiol. 27:  329.

Kurzel R,  ML Wolbarsht and BS Yamanashi (1977)  "UV radiation effects on the
human eye."  Photochem.  Photobiol.  Rev. 2:  133-167.

Kurzel RB (1973)  "Trypotophan excited states and cataracts in the human lens."
Nature 241:132-133.

-------
                                     10-42
Kuwabara T and S Okisake (1976)  "Effect of electronic strobe flashlight on the
monkey retina."  Jap. J. Ophthal. 20: 9-18.

Kuwabara T (1978)  "Photic and photo-thermal effects on the retinal pigment
epithelium."  In the Retinal Figment Epithelium, Harvard University Press,
Cambridge.

Kwan M, J Ninikoski and TK Hunt (1972)  "In vivo measurements of oxygen tension
in the cornea, aqueous human and anterior lens of the open eye." Invest.
Ophthal. 11: 108-111.

Lakhanpthal V et al. (1982) Arch Ophthalmol 100: 1068-1070.

LaVail MM and BA Battelle (1975)  "Influence of eye pigmentation and light
deprivation on inherited retinal dystrophy in the rat."  Exp. Eye Res. 21:167.

Lerman S (1980a)  "Human ultraviolet light cataracts."  Ophth. Res. 12: 303-314.

Lerman S (1980b)  "Radiant energy and the eye."  MacMillan Publishing Company,
New York.

Lerman S (1982)  "Ocular photoxicity and PUVA therapy:  An experimental and
clinical evaluation."  FDA Photochemical Toxicity Symp.  J. Natl. Cancer Inst.
69: 287-302.

Lerman S (1983)  "Psoralens and ocular effects in animals and man:  In vivo
monitoring of human ocular and cataneous manifestations."  Conf. on
Photobiologic, Toxicologic and Pharmacologic Aspects of Psoralens, March 1-3,
1982.  J. Natl. Cancer Inst., In Print.

Lerman S and R Borkman (1976)  "Spectroscopic evaluation and classification of
the normal, againing and cataractous lens."  Ophthal. Res. 8: 335-353.

Lerman S, J Megaw, and K Gardner (1982a)   "Psoralen-long-wave ultraviolet
therapy and human cataractogenesis." Invest. Ophthal. Vis. Sci. 23: 801-803.

Lerman S, J Megaw and K Gardner  (1982b)  "Allopurinol therapy and human
cataractogenesis." Amer. J. Ophthalmol. 94: 141-146.

Lerman S, JF Kuck, R Borkman and E Saker (1976)  "Acceleration of an aging
parameter (Fluorogen) in the ocular lens."  Ann. Ophthal. 8: 558-562.

Lerman S, JF Kuck, R Borkman and E Saker (1976)  "Induction, acceleration and
prevention  (in vitro) of an aging parameter in the ocular lens."  Ophthal. Res.
8: 213-226.
Lerman S, JM Megaw, K Gardner, Y Takei and I Willis  (1981)   "Localization of
8-methoxypsoralen in ocular tissues."  Opthal. Res.  13: 106-116.

Lerman S, K Gardner, J Megaw and R Borkman (1981)  "The prevention of direct and
photosensitized UV radiation damage to the ocular lens."  Ophthalmic Res. 13:
284-292.

Lerman S, J Megaw and I Willis  (1980)  "The photreactions of 8-methoxypsoralen
with  tryptophan and  lens proteins."  Photochm. Photobiol. 31: 235-242.

-------
                                     10-43
Leske MC and RD Sperduto (1983)  "The epidemiology of senile cataracts:  a
review," American Journal of Epidemiology.  118: 152-165.

Lowenstein A and J Steel (1941)  "Macular burning and ring scotoma."  Glasgow
Med. J. 135: 73-85.

Maher EF (1978)  "Transmission and absorption coefficients for ocular media of
the rhesus monkey."  Report SAM-TR-78-32, USAF School of Aerospace Medicine,
Brooks AFB, Tx.

Marcantonio JM, G Duncan, PD Davis, and AR Bushell (1980)  "Classification of
human senile cataracts by nuclear color and sodium content."  Exp. Eye Res.  31:
227-237.

Marshal, J (1970)  "Thermal and mechanical mechanisms in laser damage to the
retina."  Invest. Opthalmol. 9: 97-115.

Mao W and T Hu (1982), "An Epidemiologic Survey of Senile Cataract in China,"
Chinese Medical Journal, 95(11): 813-818.

Martinez GS, AJ Campbell, J Reiken and BC Allan (1982)  "Prevalence of ocular
disease in a population study of subjects 65 years old and older."  Am. J.
Ophthalmol. 94: 181.

McCanna, P. et al (1982) Arch Ophthalmol 100: 1071-1073.

McFaul PA (1969)  "Visual prognosis after solar retinopathy."  Brit. J.
Opthalmol. 53: 534-541.

Megaw J, J Lee and S Lerman (1980)  "NMR analyses of tryptophan-8-
methoxypsoralen photoreaction products formed in the presence of oxygen."
Photochem. Photobiol. 32:  265-269.

Moss SH and KC Smith (1981)  "Membrane damage can be a significant factor in the
inactivation of E. coli by near UV radiation."  Photochem. Photobiol. 32:
203-210.

Naidoff MA and DA Sliney (1974)  "Ritinal injury from a welding arc."  Am. J.
Ophthal 77: 663-668.

Owsley C, DW Kline, JS Werner, V Greenstein and J Marshall (1986)  "Optical
radiation effects on aging and visual perception."  In Light Toxicity.  M Wahler
and VM Kitchens (eds.).  CRC Press, Boca Raton, Florida.

Parrish JS, RR Anderson, R Urback and D Pitts (1978)  UV-A Biological Effects of
Radiation with Emphasis on Human Responses to Longwave Radiation.  Plenum Press,
New York.

Pathak MA, DM Kraemer and U Guengerich (1972) "Formation of thymine dimers in
mammalian skin by UV radiation in vivo" Photochem. Photobiol. 15: 177-185.

Pirie A (1968) Invest. Ophthalmol. 7: 634-649.

Pirie A (1972) "Photo-oxidation of Proteins and Comparisons of Photo-oxidized
Proteins with Those of the Cataractous Human Lens."  Israeli J. Med Sci. 8:
1567-.

-------
                                     10-44
Pirie A and KJ Dilley (1974)  "Photooxidation of N-fornylkynurenine and
tryptophan peptides by sunlight or simulated sunlight."  Fhotochem. Fhotobiol.
19: 115-118.

Pitts DG (1970)  "A comparative study of the effects of ultraviolet radiation on
the eye."  Am. J. Optorn. 47: 535-546.

Pitts DG (1973)  "The ocular ultraviolet action spectrum and protection
criteria."  Health Physics 15: 559-566.

Pitts DG (1974)  "The human ultraviolet action spectrum."  Am. J. Op torn, and
Physiol. Optics 51: 946-960.

Pitts DG (1978) "The ocular effects of ultraviolet radiation."  Amer. J. Optorn.
and Physiol. Optics 55: 19.

Pitts DG (1981)  "The threat of ultraviolet radiation to the eye - how to
protect against it."  J.Amer. Optorn. Assn., 52: 949-957.

Pitts DG, PD Hacker and BH Parr (1977)  "Ocular ultraviolet effects from 295 nm
and 400 nm in the rabbit eye."  DHEW (NIOSH) Publication no. 77-175.

Pitts DG, JPG Bergmanson and LWF Chu (1983)  "Rabbit eye exposure to
broad-spectrum fluorescent light."  Acta Ophthalmol (Copenhagen) Supp. 159.

Pitts DG and KR Kay (1969) "The photoophthalmic threshold for the rabbit."  Am.
J. Optom. 46: 561-572.

Pitts DG and AP Cullen  (1981)  "Determination of infrared radiation levels for
acute ocular cataractogenesis."  Graefes Arch. Klin. Exp. Ophthalmol. 217:285-.

Pitts DG and TJ Tredici (1971)  "The effects of ultraviolet on the eye."  Am.
Ind. Hyg. Assoc. J. 32: 235-246.

Pitts DG, et al. (in press),  "Optical Radiation and Cataracts," in Vicusal
Health and Optical Radiation. M. Waxier and V. Hitchens, eds., CRC Press, Inc.,
Boca Raton, Florida.

Rapp LM, JG Jose and DG Pitts (1985)  "DNA reapir synthesis in the rat retina
following in vivo exposure to 300 nm radiation."  Invest. Ophthalmol. Vis. Sci.
26: 384.

Ruffolo JJ Jr, WT Jr Ham, HA Mueller and JE Miller  (1984)  "Photochemical
lesions in the primate retina under conditions of elevated blood oxygen."
Invest. Ophthalmol. Vis. Sci. 25: 893-898.

Satoh K, M Bando and A Nakajima (1973)  "Fluorescence  in human lens."  Exp. Eye
Res., 16: 167-172.

Schmidt RE and JA Zuclich  (1980)  "Retinal lesion due  to ultraviolet laser
exposure."  Invest. Ophthalmol. Vis. Sci. 19: 1166-1175.

-------
                                     10-45
Serafino GM and Frederick JE (in press)  "Global modeling of the ultraviolet
solar flux incident on the biosphere," prepared for the U.S. Environmental
Protection Agency, Washington,  D.C.

Sliney, D and M Wolbarsht (1980)  Safety with Lasers and Other Optical Sources.
Plenum Press, New York.

Spector A, D Roy and J Stauffer (1975)  "Isolation and characterization of an
age-dependent polypeptide from human lens with non-tryptophan fluorescence."
Exp. Eye Res. 21: 9-24.

Spector A, O Roy and J Stauffer (1975)  "Isolation and characterization of an
age-dependent polypeptide from human lens with non-tryptophan fluorescence."
Exp. Eye Res. 21: 9-24.

Spector A and WH Gardner (1982)  "Hydrogen perioxide and human cataract."  Exp.
Eye Res. 33: 673-681.

Sprott GD, WG Martin and H Schneider (1976) "Differential effects of near DV and
visible light on active transport and other membrane process in E. coli."
Photochem. Photobiol. 24: 21-27.

Streeten BW and J Eshaghian (1978)  "Human posterior subcapsular cataract: a
gross and flat preparation study."  Arch. Ophthal. 96: 1653-1658.

Sykes SM, WG Jr Robison, M Waxier and T Kuwabara (1981)  "Damage to the monkey
retina by broad-spectrum fluorescent light.  "Invest. Ophth. Vis. Sci. 20:
425-434.

Taylor HR (1980) "The Environment and the Lens."  Brit. J. Ophthal. 64: 303-310.

Thomas DM and KL Schepler (1980)  "Raman spectra of normal and ultraviolet-
induced cataractous rabbit lens."  Invest. Ophthalmol. Vis. Sci. 19: 904-912.

Trevithick JR, MO Creighton and WM Ross (1980)  "Vitamin E prevents
cataractogenesis in diabetic rats."  Invest. Ophth. and Vis. Sci. (Suppl.) 19:
207-208.

Truscott RJW and RC Augusteyn (1977)  "Oxidative changes inhuman lens protein
during senile nuclear cataract formation."  Biochem. Biophys. Acta 492: 43-52.

Truscott RJW, K Faull and RC Augusteyn (1977) Ophthal. Res. 9: 263-8.

Turro NJ and A Lamola (1977)  Photochemistry in the Science of Photobiology.  KG
Smith (ed), Plenum, New York, pp. 63-86.

Tso MOM and BJ Woodford (1983)   "Effect of photic injury on the retinal
tissues."  Ophthalmology 90: 952.

Tyrell RM (1976) Photochem. Photobiol. 23: 13-20.

Van Heyningen R (1971) Nature 230: 393-394.

Van Heyningen R (1973) Exp. Eye Res. 17: 137-147.

-------
                                     10-46
Van Heyningen R (1976)  "What happens to the human lens in cataract."  Sci.
Amer. 233: 70-81.

Varma SD (1981) "Superoxide and lens of the eye: a new theory of
cataractogenesis."  Inter. J. Quantum Chein. XX: 479-484.

Varma SD, A Tina and RD Richards (1977)  "Protection against superoxide radicals
in rat lens.1"  Ophthal. Res. 9: 421-431.

Varma SD, D Chand, YR Sharma and JF Kuck (1983)  "Effect of vitamin E on
cataract development in emory mice."  Invest. Ophthal. Vis. Sci. (ARVO Suppl.)
24: 31.

Varma SD, D Chand, YR Sharma, JF Kuck and RD Richards (1983)  "Oxidative stress
on lens and cataract formation:  role of light and oxygen."  Presented at the
1st International Symposium on Light and Oxygen Effects on the Eye.  Current Eye
Research, In press.

Varma SD, NA Beachy and RD Richards (1982)  "Photoperoxidation of lens lipids:
prevention by vitamin E. Photochem. Photohiol. 36: 623-626.

Varma SD, S Kumar and RD Richards (1979)  "Light-induced damage to ocular len
cation pump: prevention by vitamin C."  Proc. Natl. Acad. Sci. 76: 3504-3506.

Varma SD, VK Srivastava and RD Richards (1982)  "Photoperoxidation in lens and
cataract formation: preventive role of superoxide dismutase, catalase and
vitamin C."  Ophthal. Res. 14: 167-175.

Warner, JS (1982)  "Development of scotopic sensitivity and the absorption
spectrum of the human ocular media."  J. Optical. Soc. Am. 72: 247-258.

Waxier, M. (in press), "Long-Term Visual Health Problems:  Optical Radiation
Risks," in Visual Health and Optical Radiation. M. Waxier and V. Kitchens, eds.,
CRC Press, Inc., Boca Raton, Florida.

Weale RA (1982)  "The age variation of senile cataract in various parts of the
world."  Br. J. Ophthalmol. 66: 31-

Webb RB and MJ Peak (1981)  "The effect of stray light in the 300-320 nm range
and the role of cyclobutyl pyrimidine dimers in 365 nm lethality."  Mutat. Res.
91: 177-182.

Weiter JJ and ED Finch (1975)  "Matters arising: UV light and human cataract."
Nature 254: 536-537.

Weiter JJ and S Subramanian (1978)  "Free radicals produced in human lenses by a
biophotonic process." Invest. Ophthal. Vis. Sci. 17: 869-873.

Wiesinger H et al. (1956)  "Transmission of light through the ocular media of
the rabbit eye."  Am. J. Ophthalmol. 42: 907-910.

Williams, TP and BN Baker eds. (1980)  The Effects of Constant Light on Visual
Processes.  Plenum Press, New York.

Wolbarsht, ML (1976)  "The function of intraocular color filters."  fed. Proc.
35: 44-50.

-------
                                     10-47
Wolbarsht ML, GS Goerge, WA Jr. Shearin, JA Klystra and MB III Landers (1983)
"Retinopathy of Prematurity:  A new look at an old disease."  Paper presented at
1st Int. Synp. on Light and Oxygen Toxicity, Baltimore.

Wolfe E (1949)  "Effects on visual threshold of exposure to radiation below 4000
angstroms."  Trans Am. Acad. Ophthalmol. Otolaryrgol. 53: 400-414.

Yoakum G, W Ferron, A Eisenstark and RB Webb (1974)  "Inhibition of replication
gap closure in E. coli by near TJV photoproducts of L-tryptophan." J. Bacterial
119:62-69.

Young RW (1981)  "A theory of central retinal disease."  In Future Directions of
Ophthalmic Research.  ML Sears (ed.).  Yale University Press, New Haven, Conn.

Yu N-T, JFR Kuck and CC Askren (1979) "Red fluorescence in older and brunescent
human lenses."  Invest. Ophthalmol. Vis. Sci. 18: 1278-1280.

Zigler JS and JD Goosey (1981)  "Aging of protein molecules:  lens crystallins
as a model system."  Trends in Biochem. Sci. 6: 133-136.

Zigler JS and JD Goosey (1981)  "Photosensitized oxidation in the ocular lens:
evidence for photosensitizers endogenous to the human lens."  Photochem. and
Photobiol. 33: 869-874.

Zigler JS, JB Sidbury, BS Yamanashi and M Wohlbarsht (1976) "Studies on
brunescent cataract.  I. Analysis of free protein-bound amino acids."  Ophthal.
Res. 8: 379-387.

Zigman S (1970)  "Near-UV light and cataracts."  Photochem. and Photobiol. 26:
437-441.

Zigman S (1981) "Photochemical mechanisms in cataract formation."  In Mechanisms
of Cataract Formation in the Human Lens."  G. Duncan (ed.).  Publ. Acad. Press,
Inc. London.  Pp. 117-149.

Zigman S and R Vaughan (1974)  "Effects of near-UV light on the lenses and
retinas of mice."  Invest. Ophth. Vis. Sci. 13: 462-465.

Zigman S, Griess, T Yulo and J Schultz (1973)  "Ocular protein alterations by
near UV light."  Exp. Eye Res. 15: 255-264.

Zigman S, M Datiles and E Torczynski (1979)  "Sunlight and Human Cataracts."
Invest. Ophthal. Vis. Sci. 18: 462-467.

Zigman S, T Yulo and G Griess (1976)  "Inactivation of catalase by
near-ultraviolet light and tryptophan photoproducts."  Mol. Cell. Biochem. 11:
149-154.

Zuclich JA (1977)  "Lenticular effects of near-UV laser radiation."  In Research
on the Ocular Effects of Laser Radiation.  Part II.  Technology Incorporated,
Final Report, Contrast F41609-73-C-0017.  USAF School of Aerospace Medicine.

Zuclich JA (1980)  "Cumulative effects of near-ultraviolet induced corneal
damage."  Health Phys. 38: 833.

-------
                                     10-48
Zuclich JA and JS Connolly (1976)  "Ocular damage induced by near-ultraviolet
laser radiation." 15: 760-764.

-------
                                   CHAPTER 11

                       RISKS TO CROPS AND TERRESTRIAL
                   ECOSYSTEMS FROM ENHANCED UV-B RADIATION
SUMMARY

    In making an assessment of the risk to crops and ecosystems of increased
ultraviolet-B (UV-B) radiation, it must be recognized that existing knowledge
is in many ways deficient.  The effects of enhanced levels of UV-B radiation
have been studied in only four of the ten major terrestrial plant ecosystems.
Most of our knowledge is derived from studies focused upon agricultural crops
and conducted at mid-latitudes, not in the tropics or at more poleward
latitudes.  Trees have not been subject to experimentation.  Experimental
protocols often have had flaws; too often, single-year studies have been done,
rather than long-term ones, and too much of the existing data comes from
growth chambers, in which plants grow under unrealistic conditions, rather
than from field studies.  Therefore, the full extent of the potential impacts
of enhanced levels of UV-B radiation on a global basis cannot be adequately
assessed.

    Despite these limitations, a broad range of experimental results
demonstrated that, in nearly half of the plant species examined, UV-B
radiation deleteriously affected crop yield and quality.  Data exist that
indicate that it may be reasonably anticipated that if UV-B radiation
increases, crop yield and quality will decline for at least some cultivars.
Existing data also suggest that increased UV-B radiation will alter the
distribution and abundance of plants and potentially disrupt ecosystems.
Unfortunately, a qualitative prediction of how these ecosystems would be
altered cannot be determined from the current knowledge base.

-------
                                     11-2
FINDINGS

1. LIMITED STUDIES ANALYZING THE EFFECT OF INCREASED UV-B RADIATION ON CROPS
   GENERALLY SHOW ADVERSE IMPACTS.   HOWEVER.  CONCLUSIONS ABOUT THE AMOUNT OF
   YIELD LOSSES ATTRIBUTABLE TO UV-B CANNOT BE DRAWN.

   la.  Difficulties in experimental design,  the large number of species and
        cultivars, and complex interactions between plants and their
        environment have prevented quantification of total crop loss from
        increases in UV-B.

   Ib.  Action spectra for UV damage to higher plants are limited, but
        indicate a strong weighting toward shorter UV-B wavelengths which are
        those most affected by ozone reduction.

2. OF PLANT CULTIVARS TESTED IN THE LABORATORY. APPROXIMATELY 70 PERCENT WERE
   DETERMINED TO BE SENSITIVE TO UV-B: CARE MUST BE TAKEN IN INTERPRETING THIS
   FINDING.

   2a.  Different cultivars within a species have exhibited different degrees
        of UV-B sensitivity.  While this suggests selective breeding could
        limit damage, neither the basis for selectivity nor the potential
        effect on other aspects of growth has been studied.

   2b.  Laboratory experiments have been shown to inadequately replicate
        effects in the field, thus the implications of cultivar sensitivity
        are not certain.

   2c.  In some species, mitigation responses more readily apparent in the
        field (e.g., increased production of UV absorbing flavonoids) have
        reduced adverse impacts.

3.  THE EFFECTS OF UV-B JRADIATION HAVE _BEEN EXAMINED FOR_ONLY J~OUR_OF THE TEN
    MAJOR TERRESTRIAL ECOSYSTEMS AND FOR ONLY A THIRD OF THE PLANT GROWTH
    FORMS.

    3a.   Little or no data exist on enhanced UV-B effects on trees, woody
          shrubs, vines, or lower vascular plants.

4.  LARGE UNCERTAINTIES EXIST AS A RESULT OF AN IMPERFECT EXPERIMENTAL DESIGN
    OR DOSIMETRY.  EXISTING EXPERIMENTAL FIELD DATA SUGGEST A POTENTIAL
    REDUCTION IN CROP YIELD FOR SOME CROPS DUE TO ENHANCED UV-B RADIATION.

    4a.   Field experiments in which UV-B radiation has been supplemented are
          limited.  Several of the earlier field experiments are of limited
          value since UV-B doses or other factors such as soil temperature
          were not sufficiently controlled or representative of field
          conditions.  Dose-response studies in the field are particularly
          different.

    4b.   The only long-term field studies of a crop involved soybeans.  These
          studies have found that enhanced levels of UV-B, simulating between
          16 and 25 percent ozone depletion, caused crop yield reductions of

-------
                                    11-3


          up  to  25  percent  in a  particular  cultivar.   Smaller  reductions  in
          yield  were  experienced in years where  drought  conditions  existed.

    4c.    Soybean (CV Essex)  yield could be accurately predicted when total
          UV-B dose,  daily  maximum temperature,  and number  of  days  of
          precipitation were  included in a  regression model.

    4d.    The lipid and protein  content  of  soybean was reduced up to 10
          percent;  however, higher UV-B  doses  alone did  not consistently
          result in the largest  reductions.

    4e.    While  only  several  cultivars have been tested  in  the field,  two out
          of  three  soybean  cultivars  tested under laboratory conditions were
          sensitive to UV-B.   If this relationship holds true  in the field,  it
          suggests  (when considered  in light of  yield reduction experiments)
          that UV-B increases could harm the potential of the  world
          agricultural system to produce soybeans.

5.   THE EFFECTS  OF UV-B ON  FUNGAL OR  VIRAL  PATHOGENS  VARY WITH PATHOGEN.  PLANT
    SPECIES.  AND CULTIVAR.

    5a.    Current evidence  on possible interactions with pathogens  is very
          limited.

    5b.    Reduced vigor in  UV-sensitive  plants could  render the plants more  or
          less susceptible  to pest or disease  damage  and thus  result in
          changes in  crop yield.

6.   CHANGES IN UV-B LEVELS  MAY INDUCE, SHIFTS IN  INTERSPECIFIC  COMPETITION.

    6a.    If enhanced UV-B  favors weeds  over crops, agricultural costs (e.g.,
          for increased tilling and  herbicide  application)  could increase.
          However,  insufficient evidence exists  to form  a basis for evaluating
          this effect.

    6b.    Increases in UV-B could alter  the results of the  competition in
          natural ecosystems  and thus shift community composition.   Since UV-B
          changes would be  both global and  long  term, possible UV-induced
          alterations of plant species balances  could result in large-scale
          changes in  the character  and equilibrium of vegetation in
          nonagricultural areas such as  forests  and grasslands.

7.   UV-B RADIATION INHIBITS AND STIMULATES  FLOWERING. DEPENDING ON  THE SPECIES
    AND GROWTH CONDITIONS.

    la.    The timing  of flowering may also  be  influenced by UV-B radiation,
          and there is limited evidence  that pollen may  be  susceptible to UV
          damage upon germination.

    7b.    Reproductive structures enclosed  within the ovary appear  to be
          well-protected from UV-B  radiation.

-------
                                    11-4
8.   INTERACTIONS BETWEEN UV-B RADIATION AND OTHER ENVIRONMENTAL FACTORS  ARE
    IMPORTANT IN DETERMINING POTENTIAL UV-B EFFECTS  ON PLANTS.

    8a.    UV-B effects may be worsened under low light regimes  or less
          apparent under conditions  of limited nutrients  or water.

    8b.    Interactions with other environmental effects make extrapolation of
          data from growth chambers  or greenhouses to  field conditions
          difficult and often unreliable.

    8c.    The combined effect of higher UV-B and other environmental  changes
          cannot be adequately assessed by current data.   Extensive,  long-term
          studies would be required.

-------
                                     11-5
INTRODUCTION

    This chapter examines the published and unpublished material currently
available to assess the likely impact of projected increases in UV-B radiation
upon global crop productivity and the distribution and abundance of plants in
natural ecosystems.  The limitations to this assessment are formidable because
of the paucity of experimental data and the slow development of appropriate
experimental technology.  Therefore, actual risks may be far greater or
somewhat less than current knowledge suggests.

ISSUES AND UNCERTAINTIES IN ASSESSING THE EFFECTS OF
UV-B RADIATION ON PLANTS

    Ideally, experiments should be designed to develop a data base that
perfectly simulates future conditions for all plant species.  These conditions
should include all direct effects of enhanced UV-B radiation, in addition to
all of the possible significant combinations with other effects
(interactions).  Such interactions should include, but not be limited to, the
effects of increased atmospheric levels of carbon dioxide (G02),  drought, and
mineral deficiency.  A perfect simulation of all possible future environments
would make it possible to accurately assess the potential impacts of global
UV-B radiation increases.  Unfortunately, such ideal circumstances do not
exist.  In reality, we must make assessments based upon imperfect experimental
designs, which include only very selective and sometimes unrealistic growing
conditions and the testing of only a few representative plant species.
Therefore, the existing data base for this assessment will allow us to examine
only some of the potential effects of UV-B changes and cannot be regarded as
conclusive.

ISSUES CONCERNING UV DOSE AND CURRENT ACTION SPECTRA FOR
UV-B IMPACT ASSESSMENT

    Total global UV-B irradiance is dependent on a number of factors,
including solar angle, latitude and altitude, stratospheric ozone
concentration, atmospheric turbidity, and cloud cover.  The earth-sun distance
and minor solar fluctuations also contribute to annual variations in
irradiance (Caldwell 1971).   Because of diurnal and seasonal variations in
many of these factors, the spectral composition of solar radiation also varies
substantially.  On a daily basis, solar UV-B irradiance is sinusoidal, peaking
at solar noon.  Annually, UV-B irradiance is maximum during summer and minimum
during winter.  Experiments evaluating the effectiveness of UV-B radiation on
plants typically do not account for such changes because of practical
difficulties in monitoring and supplementing UV-B radiation.  Generally,
supplemental UV-B radiation is provided using filtered sunlamps as a
squarewave function by using timers.  This system provides a proportionately
greater UV irradiance during morning and late afternoons and under cloudy
skies than would be anticipated outdoors.  Caldwell et al.  (1983) have
designed a modulated system to monitor ambient UV-B and provide the desired
supplemental UV-B dose.   This system provides a more realistic simulation of
anticipated ozone depletion because it modulates lamp output in accordance
with actual levels of incoming solar UV radiation.  Despite substantial
expenses,  it is highly recommended that such a system be utilized to improve
field simulations and sensitivity of validation studies.

-------
                                     11-6
    The source of UV-B radiation most commonly used in plant effects research
is the fluorescent sunlamp, which is a low-pressure mercury vapor lamp
containing a phosphor that fluoresces primarily in the UV-B region, and, to a
lesser extent, in the UV-C and UV-A regions.  Although the energy emitted
principally comes from the fluorescing phosphor,  some emission from mercury
vapor is superimposed upon this, producing distinct lines in the spectrum.  A
weighting function is absolutely essential to determine the biological
effectiveness of the spectral energy emitted from various types of lamps.  The
biologically effective irradiance (IRP) i-s given by the following
relationship:

                                IBE  -  * A EA  dA

where IA is the lamp spectral irradiance and EA is the relative effectiveness
of the energy to produce a response at wavelength A.  Thus, the biologically
effective irradiance is the product of the action spectrum and the spectral
irradiance at each wavelength.

    Several UV action spectra have been developed that share the common
feature of decreasing effectiveness as wavelength increases, but have
considerable variation in the rate of this decrease.  UV-B effectiveness in an
isolated organelle can differ considerably from an intact plant because of the
cellular shielding effects in the plant and inherent repair mechanisms.
Caldwell (1971) has developed a generalized plant damage spectrum based upon
the combined responses of a number of different plant species.  Although this
represents a step forward in understanding plant responses, an action spectrum
developed on intact plants under polychromatic radiation would be preferable.
Caldwell et al. (1986) have recently attempted to develop such an action
spectrum, and although there are experimental limitations, they were able to
show that it is technologically possible.

    Another essential reason for developing appropriate action spectra is to
evaluate radiation amplification factors (RAF), i.e., the relative increase in
biologically effective UV-B radiation associated with a specific ozone
reduction.  RAF is a complex function taking into consideration initial ozone
layer thickness, percent ozone layer reduction, latitude, season, and
biological weighting function (NAS 1979).   The increase in solar UV-B
radiation as a result of ozone reduction becomes appropriate only when the
biological effectiveness of this radiation is known.  Without an RAF, the
absolute increase of total solar UV-B radiation resulting from even an
appreciable ozone reduction is insignificant.

    Since solar spectral irradiance increases by orders of magnitude with
increasing wavelength in the UV-B region,  the tails of the action spectra have
a profound effect on the net RAF.  The RAF values are much higher for those
tails with steep slopes than for those with shallow slopes.  Thus, the
computed biological effectiveness of solar radiation could either be under- or
overestimated if the action spectra are not representative of true plant
responses.

-------
                                    11-7
    It is evident from the above discussion that a more realistic action
spectrum is needed for a proper assessment of the possible consequences of
ozone depletion.

ISSUES CONCERNING NATURAL PLANT ADAPTATIONS TO UV RADIATION

    Tremendous variability exists in species' sensitivity to UV-B radiation
(Krizek 1978; Van, Gerrard, and West 1976; Hashimoto and Tajima 1980;  Tevini,
Iwanzik, and Thoma 1981 and 1982; Tevini and Iwanzik 1982; Teramura 1983)
(Exhibit 11-1).   Some plants show sensitivity to ambient levels of UV-B
radiation (Teramura, Biggs, and Kossuth 1980; Bogenrieder and Klein 1978;
Sisson and Caldwell 1976), and others are apparently unaffected by rather
massive UV enhancements (Becwar, Moore, and Burke 1982; Ambler, Rowland, and
Maher 1978).  Similarly, large differences have been reported among cultivars
of a given species (Biggs, Kossuth, and Teramura 1981;  Dumpert and Boscher
1982; Murali and Teramura 1986a,b; Murali, Teramura, and Randall 1986).  The
mechanisms for these inherent differences have not been well documented.

    Three main categories of natural mechanisms that protect against the
effects of UV radiation mechanisms may be considered (Beggs,
Schneider-Ziebert, and Wellmann 1986).  The first includes repair mechanisms
such as photoreactivation, a light-activated enzyme-mediated process whereby
pyrimidine dimers produced by UV absorption are split (Rupert 1984).   Indirect
evidence (Beggs, Schneider-Ziebert, and Wellman 1986; Tanada and Hendricks
1953; Bridge and Klarman 1973) suggests that photoreactivation is a widespread
phenomenon in plants.  Excision repair, the replacement of deleterious
photoproducts by new, correct DNA sequences, has also been documented in plant
tissues (Rowland, Hart, and Yette 1975, Soyfer 1983).  Finally, quenching and
free radical scavenging of oxygen singlets produced by photo-oxidation also
alleviate some types of UV-induced damage.

    The second category of protective mechanisms includes those that tend to
minimize the damaging effects of UV-B radiation.  Probably the most important
of these mechanisms in plants is growth delay.  If cell division stops or is
slowed upon UV irradiation, other repair mechanisms could help ameliorate the
damage before it becomes lethal.  Although many plants show a growth
inhibition upon UV-B radiation exposure (Teramura 1983) ,  it is not clear
whether this inhibition is the direct effect of the damage or due to this
protective mechanism.

    The third category of protective mechanisms involves those that
effectively reduce the amount of UV radiation actually reaching sensitive
plant targets.  Structural attenuation by the cuticle and cell wall may play
only a minor role since they offer little UV absorption (Caldwell, Robberecht,
and Flint 1983;  Steinmueller and Tevini 1985).  The principal mechanism is
probably the production of UV-absorbing pigments in outer tissue layers.
Flavonoids and other related polyphenolic compounds, which occur in epidermal
cells and have high UV absorption coefficients (Caldwell, Robberecht,  and
Flint 1983), are likely candidates.  It has been shown that flavonoid
concentrations in plant leaves substantially increase upon exposure to UV
radiation (Wellman 1982; Murali and Teramura 1985a and 1986a; Robberecht and
Caldwell 1978; Tevini, Iwanzik, and Teramura 1981 and 1983; Flint, Jordan, and
Caldwell 1985),  but it has not been established whether such an increase can

-------
                       EXHIBIT 11-1
A Sumnary of Studies Examining The Sensitivity of Cultivar
                    to UV-B Radiation
Number of
Crop Cultivars
Glycine max 19


2

23


5

Phaseolus vulgaris 2

3


Brassica oleracea 2
(?)
Cucumis sativus 5
sensitive
2


Tritium aestivum 4
7
2
Zea mays 4


Oryza sativa 5

Hordeum vulsare 4

3
a/
Growth Basis of
b/ c/
Condition Conclusion Comparison"
G.C. 20% tolerant d.w.
60% intermediate
20% sensitive
G.H. Cultivar Altona more d.w.
sensitive than Bragg
G.H. 8% tolerant d.w.
33% unaffected
59% sensitive
F 20% sensitive seed d.w.
80% unaffected
G.H. & G.C. BEL 290 more leaf
sensitive than Astro resistance
G.H. Maxidor sensitive
Saxa, Favorite
tolerant
G.H. & G.C. no difference p.s. & d.w.
Garrard et al. (1976)
G.H. 20% tolerant d.w.
(1986a)
G.H. Poinsett extremely d.w.
sensitive, Ashley
slightly sensitive
G.H. no difference (?) d.w.
G.C. no difference d.w.
F no difference d.w.
G.C. 25% extremely d.w.
sensitive
75% sensitive
G.C. 60% tolerant d.w.
40% sensitive
G.H. 25% tolerant
75% sensitive d.w.
G.C. no difference d.w.
Reference
Biggs et al. 1981


Vu et al. 1978

Teramura and Murali 1986




Bennett (1981)

Dumpert and Boscher (1982)


Van et al (1976) 2G.C.

Murali and Teramura 80%

Krizek (1978)


Dumpert and Boscher (1982)
Biggs and Kossuth (1978)
Biggs et al. (1984)
Biggs and Kossuth (1978)


Biggs and Kossuth (1978)


Dumpert and Boscher (1982)
Biggs and Kossuth (1978)
                                                                                                                  CO

-------
                                                      EXHIBIT 11-1
to UV-B Radiation
(continued)
Number of Growth Basis of
Crop Cultivars Condition a/ Conclusion b/
Spinacia oleracea 2 G.H both sensitive
Gossypium hirsutum 2 G.H. no difference
Pennisetum glaucum 2 G.C. no difference
Cucurbita pepo 3 G.C. no difference


Compari son c/ Reference
d.w. Dumpert and Boscher (1982)
d.w. Ambler et al. (1978)
d.w. Biggs and Kossuth (1978)
d.w. Biggs and Kossuth (1978)
a/   G.H. = greenhouse; G.C. = growth chamber;  F = field

b/   If data presented, sensitive means UV-B radiation reduced d.w. (see footnote c) by at least 10% over  control plants.
     Tolerant indicates that UV-B resulted in less than 10% reduction in growth.  In some cases, growth of tolerant
     plants was even stimulated by UV-B radiation.
c_/   d.w. = total plant dry weight;  p.s.  = net photosynthesis.

-------
                                    11-10
completely attenuate the damaging effects of UV radiation.  Some studies
suggest that despite large increases in flavonoid concentration, metabolic
processes, such as photosynthesis, are still affected (Sisson 1981, Teramura
et al.  1984, Mirecki and Teramura 1984).  Also, total leaf flavonoid
concentrations alone do not account for the range of responses observed in
species sensitivity.  For example, total leaf flavonoid levels found in UV-B
irradiated soybeans were less than those found in cucumber, yet cucumber is
much more sensitive to UV radiation (Murali and Teramura 1986a,c).  Therefore,
species sensitivity to UV-B radiation is probably the product of a number of
UV-protective mechanisms acting in concert within the plant.  We currently
need more specific information concerning plant adaptations to UV-B radiation
before we can further refine our estimates of the ability of natural plant
protective mechanisms to compensate for the projected increases in solar UV-B
radiation.

ISSUES ASSOCIATED WITH THE EXTRAPOLATION OF DATA FROM
CONTROLLED ENVIRONMENTS TO THE FIELD

    Plant responses in growth chambers or greenhouses may neither
quantitatively nor qualitatively resemble responses that would occur in the
field because the conditions found in controlled environments are unlike those
found in nature.  Plants grown in growth chambers are more sensitive to a
given UV dose than are field-grown plants (Caldwell 1981, Bennett 1981,
Teramura 1982, Mirecki and Teramura 1984), because in artificial environments
a single factor is generally manipulated, while all other factors are kept
constant or are optimized for growth.  Plants outdoors commonly experience
simultaneous, multiple stresses, including water or nutrient limitation.  In
addition to these differences in physical factors, artificial environments
lack biotic factors such as the interactions between other plants, insects,
diseases, etc.

    However, enormous complexities are inherent in field studies.  Daily
changes in the environment are superimposed upon longer-scale seasonal and
annual changes.  Both temporal and spatial variability often result in
inconsistencies in plant responses between one year and the next  (for examples
see Biggs et al. 1984; Gold and Caldwell 1983; Teramura 1981b; Lydon,
Teramura, and Summers 1986) .

    Therefore, one useful approach in understanding the effects of UV-B
radiation on plants under more realistic conditions has been the  study of the
interactions between UV radiation and other commonly experienced, plant
stresses.  These studies revealed that exposure to enhanced levels of UV
radiation may affect the susceptibility of some plants to water stress, and
alter the sensitivity to UV radiation by inducing flavonoid biosynthesis
(Teramura, Tevini, and Iwanzik 1983; Tevini, Thoma, and Iwanzik 1983; Teramura
et al. 1984a; Teramura, Forseth, and Lydon 1984).  The results of field
experiments  (Murali and Teramura 1986c) revealed no additional deleterious
effects of UV-B radiation when radiation was combined with water  stress.  It
was hypothesized that changes in leaf anatomy, increased flavonoid production,
and reduced  growth induced by water stress masked the UV effects.

    Plant productivity (accumulation of dry matter) became more sensitive to
UV-B radiation as total mineral supply decreased  (Bogenrieder and Doute 1982).
Murali and Teramura  (1985a,b) found that plant sensitivity  to UV-B radiation

-------
                                    11-11
decreased as the phosphorus level decreased.   This suggests that the greatest
impact of UV-B enhancement might appear in well-fertilized (agricxiltural)
regions,  rather than in areas of low fertility.

    In most growth chambers,  visible irradiances range from 10% to 40% of
average midday irradiances.  This is a cause for concern, since many of the
deleterious effects of UV radiation may be ameliorated if plants are exposed
to UV radiation in the longer wavelength range.   Therefore, growth chamber
experiments could overestimate the impact that UV radiation would have in the
field.  Despite the wide range of plant species and growth conditions, a clear
trend emerges:  plants grown in higher levels of visible radiation are less
sensitive to UV-B radiation.   A corollary to this conclusion is that plant
sensitivity to UV-B radiation is strongly influenced by the level of visible
radiation available during growth and development, and that shaded
environments maximize this sensitivity.

    Lydon, Teramura, and Summers (1986) and Teramura and Murali (1986) have
specifically examined the differences in UV-B radiation response between
field- and greenhouse-grown soybean.  Six soybean cultivars were grown in an
unshaded greenhouse and in the field under a similar UV enhancement regime.
Cultivar sensitivity was ranked according to a combined plant response, which
included changes in total plant dry weight, in leaf area, and in plant height.
The relative ranking for UV-B sensitivity in the greenhouse was quite similar
to that found in the field.  The major difference was that UV-B radiation
produced a substantially larger (2- to 10-fold)  effect on greenhouse plants
compared with field-grown plants.  Also, in specific instances, different
conclusions could be drawn from the individual data sets.  For example,
according to greenhouse data, the James cultivar was sensitive, and the York,
cultivar resistant, to UV-B.   In the field, however, these cultivars
demonstrated the opposite response.  Therefore,  if controlled
environment-to-field extrapolations are necessary, they must be done with the
utmost caution.  At best, general trends may be implied, but specific or
quantitative extrapolations do not yet seem feasible.

    Despite a great range of experimental growth conditions and UV doses, most
cultivars show large individual variation in response to UV-B radiation
(Exhibit 11-1).  Therefore, the potential for ameliorating the impacts of
projected increases in solar UV radiation may be present in our current crop
germ plasm by selecting for UV tolerance.  However, we have little
understanding of the mechanisms responsible for these cultivar differences and
the relationship of genetic factors that provide UV tolerance to other
desirable plant characteristics.  If tolerance to UV-B radiation were linked
(in a genetic sense) to some undesirable characteristics, crop breeding may
not ameliorate the impacts of UV-B damage.

UNCERTAINTIES IN OUR CURRENT KNOWLEDGE OF UV-B EFFECTS ON
TERRESTRIAL ECOSYSTEMS AND PLANT GROWTH FORMS

    Most of our knowledge of the biological effects of increasing solar UV
radiation stems from research focused upon agricultural crops.  Plants from
only four of ten ecosystems,  representing only about 27% of global net primary
productivity (NPP), have been examined (Exhibit 11-2).  In temperate forests
and temperate grasslands, only very limited preliminary data are available.
Also, of 25 categories of major plant growth forms, the effectiveness of UV-B

-------
                                     11-12
                                 EXHIBIT 11-2



                  Survey of UV Studies by Major Terrestrial

                    Plant Ecosystems (after Whittaker 1975)

Ecosystem
Tropical forest
Temperate forest
Savanna
Boreal forest
Agricultural
Woodland and scrubland
Temperate grassland
Swamp and marsh
Desert and semidesert
Tundra and alpine
Global NPP
(109 ton/yr)
49.4
14.9
13.5
9.6
9.1
6.0
5.4
4.0
1.7
1.1
Total Area
(106 km2)
24.5
12.0
15.0
12.0
14.0
8.5
9.0
2.0
42.0
8.0
Included in -/
UV Study
no
yes
no
no
yes
no
yes
no
no
yes
a/
    Only studies examining some aspect of growth.

-------
                                     11-13
radiation has only been examined in eight (32%) .  Only very limited data exist
for trees, vines, small woody shrubs, epiphytes, or lower vascular plants.  Thus
the experimental data base represents only 19 of the 314 plant families in the
world (Cronquest 1981).  Of those families tested, only seven include
representative species in which harvestable yield was examined.  Therefore, we
have very little information from which to calculate the potential impacts of
increasing levels of solar UV radiation on global terrestrial productivity of
ecosystem dynamics.

    Ecosystem composition and function are dependent upon the influence of
numerous biotic and abiotic factors.  Within populations, plants may possess
mechanisms that protect them from UV-B, but the ability of higher plants to
select for those mechanisms within the timeframe of expected changes in UV-B
radiation is unclear as to the effect of that selection.  Differential
sensitivity could result in subtle shifts in species interactions such as
competition.  These shifts, as well as changes  in abiotic factors (soil
nutrients, climate, etc.), could produce significant changes in ecosystem
composition.

UNCERTAINTIES WITH THE ABILITY TO EXTRAPOLATE KNOWLEDGE TO
HIGHER AMBIENT C02 ENVIRONMENT AND OTHER ATMOSPHERIC POLLUTANTS

    Global atmospheric carbon dioxide (C02) concentration has been gradually
increasing from 205 ppm some 20,000 years ago to approximately 340 ppm today
(Keelings 1978, Neftel et al. 1982).  It is anticipated that sometime between
2075 and 2100, the atmospheric C02 concentration will reach 600 ppm (Gates
1983).

    At present we have no direct experimental evidence on the effects of UV-B
radiation under increased levels of atmospheric C02.   However, in general, it is
believed that increased C02 increases photosynthesis and water use efficiency in
most plants (both of which are reduced by UV-B  radiation),  especially in C3
plants.   Exhibit 11-3 summarizes some of that research.

    Recently,  however, some confounding evidence has been found for perennials,
in which high levels of C02 led to a decrease in biomass production (Fried
1986).  Furthermore, most C02 work has been done in growth chambers, in which
conditions differ from those that would occur in the field.  The net effect of
an interaction between UV-B radiation and increased levels of C02 cannot,  at
this time, be predicted.   Whether the deleterious effects of UV-B will be
proportionally higher or lower in a higher C02 world is uncertain.  Similarly,
the combined impacts of higher C02 and UV-B on  inter- and intraspecific
competitive balance is impossible to ascertain.  Other changes in the
environment, such as those due to global warming and climatic changes,  may also
play important roles in determining effects, but cannot be assessed.

    Further uncertainty stems from the absence of knowledge on the interactions
between increased UV-B levels and that of other pollutants in the lower
atmosphere.   Primarily anthropogenic in origin, the major pollutants include
ozone, sulfur oxides,  and nitrogen oxides.   According to the National

-------
                                     11-14
                                 EXHIBIT 11-3

                   Summary of UV-B and CO2  Effects on Plants
                          (Lemon 1983; Teramura 1983)
Plant Characteristic
   Enhanced UV-B
   Doubling of C02
Photosynthes is
Leaf conductance
Decreases in many C3
and C4 plants
No effect in many plants
Water use efficiency   Decreases in most plants
Dry matter produc-
tion and yield
Leaf area


Specific leaf weight

Crop maturity

Flowering
Interspecific
differences

Intraspecific
differences

Drought stress
Decreases in many plants
Decreases in many plants


Increases in many plants

No effect

May inhibit or stimulate
flowering in some plants

Species may vary in
degree of response

Response varies among
cultivars

Plants become less
sensitive to UV-B but
not tolerant to drought
In C3 plants increases
up to 100% but in C4
plant only a small
increase

Decreases both in C3 and
C4 plants

Increases in both C3
and C4 plants

In C3 plants almost
doubles but in C4 plants
only a small increase

Increases more in C3
than in C4 plants

Increases

Accelerated

Flowers produced earlier
Major differences occur
between C3 and C4 plants

Response may vary among
cultivars

Plants become more
drought tolerant

-------
                                    11-15
Crop Loss Assessment Network, farm crop losses in the United States in 1981
due to air pollution amounted to between $1 billion and $2 billion.  Many
studies have demonstrated that, in combination, the deleterious effects of air
pollutants were additive and, in a few cases, multiplicative (Reinert, Heagle,
and Heck 1975).

RISKS TO CROP YIELD RESULTING FROM AN INCREASE IN SOLAR
UV-B RADIATION

    During the past 10 years, only nine field studies have examined the
effects of UV-B radiation on crop yield (Exhibits 11-4 and 11-5).  These nine
studies included 22 crop species.  Corn (Zea mays) was included in six of the
nine studies, soybean (Glycine max) in four, and tomato (Lycopersicon
esculentum),  bean (Phaseolus vulgaris),  and potato (Solanum tuberosum) each
were found in three of the studies.  In about half of these species, UV-B
reduced yield.

    Ambler, Rowland, and Maher (1978) grew eight species of crops in a field
in Beltsville, Maryland (30°N).  Plants were grown under a lamp irradiation
system maintained 1.6m above them in a linear arrangement of unfiltered
Westinghouse BZS-CLG and FS-40 sunlamps.  A two-dimensional gradient was
established:   one parallel to the lamps and another at right angles.  Only
broccoli showed a significant UV effect, although the authors suggested that
sorghum and corn also were affected.   The use of unfiltered lamps, which emit
both UV-B and UV-C, make interpretation of the results difficult, since UV-C
radiation produces qualitatively and quantitatively different effects from
UV-B (Nachtway 1975).

    Bartholic, Halsey, and Garrard (1975) conducted field exclusion studies in
Gainesville,  Florida (29° 36'N), in which ambient UV-B was filtered through
plastic films of Mylar Type S or polyethylene.  Since plants were grown under
panels covered with these films, only the direct beam component of ambient UV
was removed (approximately one-half the total ambient UV present).  Compared
with uncovered plots, total yield decreased in beans and corn growing under
both Mylar and polyethylene.  Tomato matured significantly earlier under
Mylar, resulting in an apparent decrease in fruit weight in plants growing in
uncovered plots.   There may have been a shading problem associated with the
panel framework,  or a difference may have been maintained in leaf temperatures
or soil moisture as a direct result of the exclosures themselves.

    Becwar, Moore, and Burke (1982) also conducted an exclusion study located
at 3000 m in the Colorado Rocky Mountains (39° ll'N).  Mylar, Aclar, and
cellulose acetate filters were used.   The only UV effect reported was a
decrease in wheat height (between 8% and 19% depending upon plant age),  with
no corresponding change in total plant dry weight.  A second study was
conducted using filtered FS-40 sunlamps to supplement ambient levels of solar
UV-B.  No significant effect on crop yield was produced despite a calculated
52% UV enhancement compared with sea level irradiance,  simulating the dose
plants received at sea level as control to make actual comparisons.

    Biggs and Kossuth (1978) reported yields for seven crops grown in beds
filled with a synthetic soil mix in Gainesville, Florida (29° 36').  UV-B

-------
                                                        EXHIBIT 11-4

                       Suranary of Field Studies Examining the Effects of HV-B Radiation on Crop Yields
                                      (Values represent percent changes from controls)
Ambler
et al.
(1978)
(a)
Bartholic
et al.
(1975)
(b)
Becwar
et al.
(1982)
(c)
Biggs and
Kossuth
(1978)
(d)
Biggs
et al.
(1984)
(d)
Eisenstark
et al.
(1985)
(d)

Esser
(1980)
(e)
Hart
et al.
(1975)
(a)
Cucurbita maxima
Cucurbita pepo
Phaseolus vulgaris
Triticum aestivum
Zea mays
Spinacia oleracea
Sorghum bicolor
Capsicum annum
Glycine max
Cynodon dactylon
Beta vulgaris
Brassica oleracea
  var. capitata
Brassica oleracea
  var. botrytis
Lycopersicon
  esculentum
Nicotiana tobaccum
Raphanus sativus
Pennisetum glaucum
Solanum tuberosum
Brassica juncea
Vigna unguiculata
Oryza sativa
Arachis hypogaea
-24 to
  -45%
                                  -14  to  -90
          +12 to +15
          +29 to +39
                                               -5
                                                0
                                                                       +53 to  -75
                                                         -79  to  -87                    0
                                                                       +11 to  -56
                                                                       +19 to -49
          -5 to  -26
                                 -11 to -39
                                                                       -2 to -41
                                 -9 to -43
                                 -18 to -38
(a)  Unfiltered Westinghouse BZS-CLG and FS-40  sunlamps
(b)  Ambient UV filtered with Mylar Type S  or poleyethylene
(c)  Ambient UV filtered with cellulose acetate, Aclar, or Mylar
(d)  Westinghouse FS-40 sunlamps filtered with  cellulose acetate or Mylar
(e)  Unfiltered Philips TL 40/12 sunlamps and lamps filtered with Schott WG 305  (2 and 3 mm) filters

-------
                                                              EXHIBIT 11-5
                                           Details of Fisld Study by Teramnxa (1981-1985) a/


1981 1982
CULTIVAR b/ -16 -25 -16 -25
Bay -10 NE d/ +6 -8
Essex -25 e/ NE -12 -23
James -14 e/ NE -22 -25
Williams +22 e/ NE +9 +14
York -12 NE -25 e/ -8
Forrest NE NE -14 +27
SIMULATED OZONE CHANGE
(percent)
1983 c/ 1984 c/ 1985
-16 -25 -16 -25 -16 -25

+39 e/ +6 e/ +14 e/ -7 +6 -20 e/

+13 £/ -11 e/ +15 e/ +10 e/ +18 e/ +4 e/


a/   Values reported are percent changes in seed yields for given simulated ozone changes (-16% and -25%)  compared to controls (yields
     measured on a dry weight basis).  Irradiance by either Mylar (control) or cellulose acetate (UV-B supplemented) filtered FS-40.

b/   After initially screening 23 cultivars (16 of which were sensitive to UV-B;  -10% yield reduction) in  the greenhouse,  these six were
     chosen for field experimentation.  The six represent the full range of UV sensitivity found in the greenhouse,  including very
     sensitive and very tolerant cultivars.  Beginning in 1983 only Essex (very sensitive) and Williams (very tolerant) were planted  in
     the field to increase the experimental sample size to 200.   As shown in the table,  yields for Essex were generally reduced at the
     higher UV level (ozone change of -25%) but were mixed (relative to controls) at the other UV level.   Yields for Williams were
     generally enhanced by increased UV.  Of note is that Essex is currently replacing other older cultivars (including Williams)  and is
     becoming one of the most widely planted soybeans in the U.S.   In a UV-enhanced environment,  Essex will be deleteriously affected.
     Therefore, superior cultivars being developed today by crop breeders may not be suitable for the future should the UV environment
     change.

£/   Drought year.  Low yields for both controls and experimental (i.e.,  dosed) plants.

d/   NE = Not evaluated.

e/   Significantly different at p - 0.05 level.

-------
                                    11-18
radiation was supplemented with a linear arrangement of six FS-40 sunlamps at
a 12° angle from horizontal,  producing a gradient in UV-B irradiances.
Control plants were grown without lamps above them and were adjacent to those
receiving the highest UV dose.   Although some "significant" UV effects  were
reported, no indication of statistical tests nor descriptive statistics were
given.  The greatest reductions in yield were often found in plants receiving
the lowest UV irradiance, and the largest UV effects were found in plants
growing adjacent to ones showing little or no effect.  Although each plant
along the gradient received a uniquely different UV dose, plants and
treatments were pooled for analysis.   These manipulations undoubtedly added a
great deal of experimental variability and resulted in interpretation
difficulty.

    In another field experiment, Biggs et al. (1984) conducted a two-year
evaluation of crop yield in rice, wheat, corn, and soybean.  Plants were again
grown in beds filled with synthetic soils, but UV irradiation was supplied by
FS-40 sunlamps filtered with either Mylar or 3-, 5-, or 10-ml cellulose
acetate providing 0, 32%, 23% or 16% UV enhancements.  The only significant
yield reduction reported was for wheat (5% reduction) and only for one  of the
two experimental years.   The data for rice, corn, and soybean were highly
variable, and therefore no statistical differences could be detected.

    Eisenstark et al. (1985) grew corn in large pots for three growing seasons
in Columbia, Missouri (38° 57'N).  Two levels of UV radiation were supplied by
cellulose acetate-filtered FS-40 sunlamps simulating a 7% and 21% ozone
depletion.  Controls included lamps filtered with Mylar and no overhead lamps.
Plants were found to be particularly susceptible to UV-induced effects  during
tassel development (Eisenstark and Perrot 1985).  Plants irradiated as
seedlings produced total grain yield reductions of 23% and 32%, respectively,
when compared with Mylar control plants.  Even larger differences were found
when these plants were compared with those grown without overhead lamps; yield
was reduced by 80% in plants filtered by Mylar and 87% in those grown under
cellulose acetate.  The authors suggested that this large lamp effect could be
due to additional UV-A emitted from the sunlamps, but this UV-A supplement was
very small relative to UV-A present in midday solar radiation  (Caldwell et al.
1986) .  The implications of this study for field crops (as opposed to potted
plants) is doubtful, however, since the pots allowed root structures to reach
higher temperatures than they would in fields.

    Esser  (1980) conducted field studies in Frankfurt, Federal Republic of
Germany (50°N), on bean, cabbage, spinach, and potato.  He used linearly
arranged Philips TL 40/12 sunlamps suspended 3 m above the plants and produced
four UV enhancements using a combination of reflectors, Schott WG305 filters,
and unfiltered lamps.  Significant yield reductions were found only under
unfiltered lamps in the four crops tested.  Yields for spinach, cabbage, and
bean  increased under filtered lamps.

    Hart et al. (1975) grew plants under linearly arranged unfiltered FS-40
sunlamps in Beltsville, Maryland (39°N).  The absence of experimental data and
details on this study limit its usefulness.  Also, the presence of UV-C
radiation  from these unfiltered lamps greatly decreases the validity of these
data.

-------
                                    11-19
    Teramura has grown soybeans for five seasons in Beltsville, Maryland
 (39°N) (Exhibit 11-5).  During the first two years, six soybean cultivars were
 grown and two cultivars were grown in subsequent years under filtered FS-40
 sunlamps oriented perpendicularly to the soybean rows.  This arrangement
 avoids the large variation in UV irradiance along the length of the bulb,
 which is a major problem with linearly arranged (end-to-end) lamps.  Control
 plants were grown under Mylar-filtered lamps.  Field experiments simulated 0,
 16%, and 25% ozone depletions.  Seed yield was consistently reduced at a
 simulated 25% ozone depletion in cultivar Essex.  Yield reductions of up to
 20% were reported in three of the five years.  The "UV-tolerant" cultivar
 Williams was less affected but did show an 11% reduction in yield in 1983.
 Also in 1983, a substantially higher relative yield (6% increase) was found in
 cultivar Essex, which was opposite to that observed in the previous two years.
 One critical caveat that must be included with this observation is that 1983
 was an extremely dry year.  Actual seed weights of control plants during 1983
 were only 20%-30% of those harvested in 1981 and 1982.  Therefore, it is
 questionable whether these 1983 data are representative of true (normal) field
 trends.  This graphically illustrates the importance of multi-year studies and
 emphasizes the need to monitor other environmental variables in addition to UV
 radiation.  If the results on Essex in drought years was repeated, it would
 indicate that higher UV radiation actually helped plants in severe drought
 years, reducing the crop loss by a few percent.

    Rowe and Adams (1987) reviewed the information developed by Teramura to
 determine the done-response relationship between UV-B radiation levels and
 soybean yield.   To calculate a representative relationship from the Teramura
 data, Rowe and Adams limited the data only to those points reported by
 Teramura as statistically significant at the 0.05 level.  This restriction led
 to the exclusion of data from the "drought" years.   With this approach they
 estimated a yield decline of approximately 0.3 percent for each one percent
 decrease in stratospheric ozone.

    Soybean yield is influenced by other microclimatic factors as well as
 total UV-B dose.  Yield appears to be strongly influenced by the number of
 days of precipitation and the number of days during which the maximum
 temperature exceeds 35°C.  In general,  yield decreases as the number of hot
 days increases and increases as the number of days with precipitation
 approaches 25 in the growing season.   However, further increases in the number
 of days with precipitation decreases yield.  The relative importance of UV-B
 dose is a function of the cultivar and other prevailing microclimatic factors.
 For example,  in Essex,  total seed yield can be predicted,  within 95%
 confidence intervals,  by including total UV-B doses,  number of precipitation
 events, and the number of days where air temperature exceeds 35°C in a
 regression model.

    Despite the broad range of experimental protocols and dosimetry used by
various investigators,  it appears that increases in solar UV radiation could
potentially have a deleterious impact upon global crop yields.   Even
discounting the data from the three unfiltered lamp studies (Ambler,  Rowland,
and Maher 1978, Esser 1980,  Hart et al.  1975), there are still more instances
of significant reductions in yield than reports of no effect.   While
uncertainty exists,  all of these studies were imperfect validations,  suffering
from one problem or another in experimental design or dosimetry,  and only
three studies (Biggs et al.  1984,  Eisenstark et al.  1984,  Teramura 1981-1985)

-------
                                    11-20
include multi-year observations in which longer-scale environmental
variability has been taken into consideration and only mid-latitude regions
have been analyzed.  It is reasonably likely that increases in solar UV will
decrease global crop yields.

RISKS TO YIELD DUE TO A DECREASE IN QUALITY

    Biggs and Kossuth (1978) reported that the number of abnormally shaped
tomato fruit decreased between 11% and 41% under enhanced levels of UV-B
radiation in the field.  They also reported a 6%-23% reduction in the number
of culls (rots, cracks, sunscald,  etc.) in plants receiving a moderate UV-B
dose, while those receiving a low dose produced a 10%-34% increase in culls.
Mean potato weight of grade A large potatoes showed an increase of between 3%
and 13% under enhanced levels of UV-B radiation, but no other grade of
potatoes increased in weight.  In neither case was a clear, linear
relationship between UV dose and plant response demonstrated.

    In a field study conducted in Beltsville, Maryland, Ambler, Rowland, and
Maher (1978) reported that the sugar content in sugar beets significantly
increased between 17% and 21% with increasing UV-B irradiance.  However, the
authors observed these significant increases only in plants located directly
under the lamp irradiation system and not in plants that received a comparable
UV-B dose at some distance away from the light source.   Possibly some other
uncontrolled factor (perhaps shading) was inadvertently introduced into the
experiment.

    Teramura (1982-1985) examined the effects of UV-B radiation on seed
protein and lipid concentrations of soybean grown in the field at Beltsville,
Maryland (Exhibit 11-6).  Overall, the effects of UV radiation were relatively
small, but some were nonetheless significant.  Seed protein concentrations in
cultivar Essex declined under elevated levels of UV-B radiation in two of the
four years (up to a 5% reduction.   Seed lipid concentrations in cultivar
Forrest were reduced by 3%-5% in 1982, the only year in which this
characteristic was measured in Forrest.  Williams showed a slight reduction in
seed lipid content, with a maximum reduction of 10% reported in 1984.

    The extent of UV-mediated alterations in yield quality cannot presently be
estimated with any degree of confidence due to the paucity of experimental
data at hand.  Furthermore, conclusions drawn from a single growing season may
be unreliable because of the annual variation in responses reported.
Nonetheless, the consistent reductions in yield quality in soybean suggest
that the risk in some crops may be quite high.

RISKS TO YIELD DUE TO POSSIBLE INCREASES IN DISEASE OR PEST ATTACK

    Esser (1980) reported a significant decrease in the number of aphids per
plant but no significant difference in the spidermite population with
increased UV-B radiation.  Although the results suggest that UV-B radiation
can have potentially beneficial effects on pest control, these conclusions
must be judged with caution since the observation period was less than two
weeks.

-------
                                                    EXHIBIT 11-6
                                    Svmnary of Changes in Yield Quality in Soybean
                                       Between the 1982 and 1985 Growing Seasons
                                                  (Teranura 1982-1985)
Crop
Glvcine max
cv Bay

Essex

James

Williams

Forrest

York

Yield
Character

% protein
X lipid
X protein
X lipid
% protein
X lipid
X protein
X lipid
X protein
X lipid
X protein
X lipid
1982 Signifi- 1983 Signifi- 1984 Signifi- 1985 Signifi-
X Change canceX Change canceX Change canceX Change cance

0 to -5
-1.4 to +3
-3 to -5
-2 to +.5
+.2 to +3
-1 to -2
0 to -.2
+5 to +8
+.5 to +2
-3 to -5
-2 to -3
0 to +3

ns*
ns
P=0.05 +1 to -5 P=0.05 +1 to 0 ns -0.7 to 0 ns
ns +2 to -2 ns -2 to 0 ns 0 ns
ns
ns
ns + .4 to -.2 ns +3 to +5 P=0.05 0.5 to -1 P=0.05
P=0.05 -1 to +6 ns -5 to -10 P=0.05 0 ns
ns
F=0.05
ns
ns
ns= Not significant.

-------
                                    11-22
    Growth chamber and in vitro studies of fungal pathogens showed that
hyaline spores are more sensitive to UV-B radiation than pigmented spores
(Cams, Grahm, and Ravitz 1978).  Owens and Krizek (1980), however, found that
survival of a pigmented spore (Cladosporium cucumerinum) decreased with UV-B
radiation due to a delay in germ tube emergence.   Other studies (Esser 1980,
Biggs et al. 1984) indicate that there is no clear relationship between spore
coloration and UV-B radiation.

    Three leaf fungal pathogens showed a significant decrease in disease
severity with UV-B radiation in a growth chamber study (Esser 1980),  whereas a
field study by Biggs et al. (1984) showed no significant difference in disease
severity on leaves and seeds under increased UV-B radiation.  Biggs et al.
(1984) and Biggs (1985) found that in a leaf rust-sensitive cultivar of wheat,
disease severity increased with UV-B radiation.   In a rust-resistant cultivar,
however, there were no differences in disease severity.

    Semeniuk and Goth (1980) found significant UV-B mediated reductions in
potato virus infection on Chenopodium quinoa.   At irradiances over 86 m
UV-B    no infection occurred.   In this study, virus extract was exposed to
UV-B radiation immediately upon its application over the leaf surface.
Intuitively, viruses should be  highly susceptible to UV-B radiation,  since
they consist of nucleic acids encased in proteins, both of which have high UV
absorption properties.  However, virus sensitivity must be viewed in light of
screening offered by the host tissues.

    One of the plant defense mechanisms that inhibits final development is the
production of a class of chemicals known as phytoalexins (Bell 1981).
Phytoalexin production in plants can be induced artificially through
mechanical injury, high temperature, application of fungicides and
antibiotics, and by UV radiation (Bridge and Klarman 1973, Reilly 1975).  UV-B
radiation can induce isoflavonoid phytoalexin synthesis (Bakker et al. 1983),
but excess production can be toxic due to free radical formation (Beggs et al.
1984).

    At present, it is difficult to forecast the consequences of enhanced UV-B
radiation in terms of pest and  disease damage.  From the limited information
available, it appears that in some cases, UV-B radiation might decrease
disease severity,  while in others it might aggravate disease.  The effects
vary with pathogen, plant species, and cultivar.   Further studies are
obviously needed to develop a better understanding of the consequences of
increased levels of UV-B radiation on pests and plant diseases.
RISKS TO YIELD DUE TO COMPETITION WITH OTHER PLANTS

    Many plants have been shown to exhibit a wide range of sensitivity to
enhanced UV-B radiation (Teramura 1983),  which could lead to changes in
competitive ability within plant communities through differential UV-B
resistance (Caldwell 1977).   Solar UV exclusion studies (cited in Gold and
Caldwell 1983) using natural competing species show large differences in
response to present levels of UV radiation.   This suggests that current levels
of UV radiation may be partly responsible for interspecific competition among
various native plant species.

-------
                                    11-23
    Gold and Caldwell (1983) studied intraspecific competition in wheat
(Triticum aestivum L.),  wild oats (Avena fatua L.), and goat grass (Aegilops
cylindrica Host) at various planting densities in Logan, Utah, under ambient
and UV-B levels simulating a 16% ozone reduction.  No significant differences
in shoot biomass production were observed.  Therefore, enhanced levels of UV-B
radiation may be of little consequence in terms of intraspecific plant
competition.  Preliminary results (Caldwell, personal communication; 1985,
unpublished manuscript), however, indicate that interspecific competition may
be affected.  UV-B radiation appeared to alter growth by the inhibition of
internode elongation in wild oat, but not in wheat, thus favoring wheat in
competition for available light.

    Fox and Caldwell (1978) and Gold and Caldwell (1983) examined the effects
of UV-B radiation on the competitive interaction of field-grown competing
pairs from three plant associations:  agricultural crops and associated weeds,
montane forage crops, and disturbed weedy associates.  To measure competitive
ability, relative crowding coefficients (RCCs) were used.  RCCs based upon
total above-ground biomass indicate that there were significant shifts in the
competitive balance of some agricultural crop and weed associations.  For
example, pigweed (Amaranthus) was more competitive under ambient UV
conditions, and alfalfa (Medicago) exhibited the competitive advantage under
enhanced UV-B radiation.  Under a 40% ozone reduction, wheat (Triticum) had a
competitive advantage over wild oats (Avena) and goat grass (Aegilops), but no
significant differences were observed with a 16% reduction.  In the previous
year, the competitive balance was found to be in favor of Avena.  It was
suggested that this difference was due to a late planting which resulted in
large water and temperature stresses during seedling development (Gold and
Caldwell 1983).  Similarly, the weedy Geum had the competitive advantage over
Poa grass under ambient conditions, but the balance was shifted under enhanced
UV-B radiation.  Preliminary field data suggest that the competitive ability
of wheat, based on seed biomass, increased relative to wild oats under
enhanced UV-B radiation.

    These results demonstrate that enhanced levels of UV-B radiation can alter
the competitive interactions of some species.  Total harvestable yield, as
well as its quality, is altered by the presence of weeds (Bell and Nalewaja
1968, McWhorter and Patterson 1980).  Because there are a large number of
weeds typically associated with various crop plants, increasing levels of UV-B
radiation could potentially have serious consequences if weeds gain a
competitive advantage over crops.  Even very subtle differences in sensitivity
could result in large changes in species composition over time and possibly
affect ecosystem function.  At this time,  however, it is impossible to
determine the extent of the risk.

RISK TO YIELD DUE TO CHANGES IN POLLINATION AND FLOWERING

    In plants in which pollination and subsequent fertilization take place
during the day with flowers fully open, reproductive tissues would seemingly
receive an appreciable UV-B dose.  However, ovules enclosed in the ovary are
well-protected against UV-B radiation.  Flint and Caldwell (1983) have shown
that the anther wall filters out over 98% of the incident UV-B radiation in
six plant species.   Therefore, before another dehiscence, pollen is also well
protected.

-------
                                    11-24
    Southworth (1969) found UV-absorbing compounds with a maximum absorbance
in the UV-B range in the wall of pollen.  Additional studies on the UV
absorption profiles of stigmatic surfaces and exudates of many species show
one or more peaks in the UV-B region (Martin 1970, Martin and Brewbaker 1971).
Therefore, under natural conditions, the effectiveness of UV-B may be minimal
because of UV-absorbing pigments in the anther and pollen walls and in
stigmatic surface exudate.

    Results of numerous studies that were conducted during the early part of
this century on the effects of solar UV radiation through window glass filters
indicated an inhibition of flowering (see reviews by Popp and Brown 1936 and
Caldwell 1971).  However, these experiments were generally executed with
insufficient sample sizes and failed to isolate UV irradiation as the single
contributing factor causing the difference in flowering.  For instance,
Caldwell (1968) has shown that leaf temperatures increase sufficiently under
window glass to alter the induction of flowering (see Zeevaart 1976 for a
discussion of the physiology of flowering).

    Results of the Kasperbauer and Loomis (1965) study of Melilotus and
Caldwell's (1968) study of Trifolium dasyphyllum show an increase in flowering
with the exclusion of solar UV radiation.  A growth chamber study by Klein,
Edsall, and Gentile  (1965) which incorporated primarily UV-A radiation, also
shows an increase in the number of flowers produced in marigold by the
exclusion of UV radiation.  Similarly, greenhouse trials on beans and peas
also show an inhibition of flowering and decrease in flower number due to UV-B
radiation (Biggs and Basiouny 1975).  However, field studies in which
unfiltered sunlamps  (which emit both UV-B and UV-C) were used produced no
significant effects on the flowers of marigold and tomato plants, nor on their
number or the date on which flowering occurred; neither were significant
effects produced in maize tasseling or in sorghum heading (Hart et al. 1975).
Biggs and Kossuth (1978), on the other hand, found an increase in flower
number but a decrease in  the flowering duration in potatoes, and a decrease in
flower number in tomatoes at peak flowering in UV-B irradiated plants.
Apparently, some changes  in flowering suggest a correlation with UV-B
radiation, but whether such changes would lead to an appreciable effect in
harvestable yield has not been fully investigated.

-------
                                    11-25
REFERENCES
Ambler, J.E., R.A. Rowland, and N.K. Maher, 1978.  Response of selected
vegetable and agronomic crops to increased UV-B irradiation under field
conditions.  UV-B Biological and Climatic Effects Research (BACER),  Final
Report EPA-IAG-D6-0168, EPA, Washington, D.C.

Bakker, J., F.J. Grommers,  L. Smits, A. Fuchs,  and F.W. de Vries, 1983.
Photoactvation of isoflavonoid phytoalexins:   Involvement of free radicals.
Photochem. Photobiol. 38:323-329.

Bartholic, J.F., L.H. Halsey, and L.A. Garrard, 1975.  Field trials  with
filters to test for effects of UV radiation on agricultural productivity.  In
Climatic Impact Assessment Program (CIAP),  Monograph 5 (Nachtwey, D.S., M.M.
Caldwell, and R.H. Biggs, eds.),  pp. 61-71.  U.S. Department of
Transportation, Report No.  DOT-TST-75-55, National Tech.  Infor. Serv.,
Springfield, Virginia.

Becwar, M.R., F.D. Moore III, andM.J. Burke, 1982.  Effects of deletion and
enhancement of ultraviolet-B (280-315 run) radiation on plants grown at 3000 m
elevation.  J. Amer. Soc. Hort. Sci. 107:771-779.

Beggs, C.J., U. Schneider-Ziebert, and E. Wellmann, 1986.  UV-B radiation and
adaptive mechanisms in plants.   In Stratospheric Ozone Reduction, Solar
Ultraviolet Radiation and Plant Life (Worrest,  R.C,, ed.),  Springer-Verlag.
ISBN 13875-7.

Bell, A.A., 1981.  Biochemical mechanisms of disease resistance.   Ann.  Rev.
Plant Physiol. 32:21-81.

Bell, A.R., and J.D. Nalewaja,  1968.  Competition of wild oats in wheat and
barley.  Weed Sci. 16:505-508.

Bennett, J.H., 1981.  Photosynthesis and gas diffusion in leaves  of  selected
crop plants exposed to ultraviolet-B radiation.  J. Environ.  Qual. 10:271-275.

Biggs, R.H., and F.M. Basiouny, 1975.  Plant growth responses to  elevated UV-B
irradiation under growth chamber, greenhouse, and solarium conditions.   In
Impacts of Climatic Change on the Biosphere,  CIAP Monograph 5 (Nachtway,  D.S.,
M.M. Caldwell, and R.H. Biggs,  eds.), pp. 4-197-4-248.  U.S.  Department of
Transportation, Report No.  DOT-TST-75-55, National Tech.  Info. Serv.,
Springfield, Virginia.

Biggs, R.H., and S.V. Kossuth,  1978.  Effects of ultraviolet-B radiation
enhancement under field conditions on potatoes, tomatoes, corn, rice, southern
peas, peanuts, squash, mustard and radish.   UV-B Biological and Climatic
Effects Research (BACER), Final Report, EPA,  Washington,  D.C.

Biggs, R.H., S.V. Kossuth,  and A.H.  Teramura, 1981.  Response of  19  cultivars
of soybeans to ultraviolet-B irradiance.  Physiol. Plant. 53:19-26.

-------
                                    11-26
Biggs, R.H., P.G.  Webb,  L.A.  Garrard,  T.R.  Sinclair,  and S.H.  West,  1984.   The
effects of enhanced ultraviolet-B on rice,  wheat,  corn,  soybean,  citrus,  and
duckweed.  U.  S.  Evironmental Protection Agency Year 3 Interim Report No.
808075-03, EPA,  Washington,  B.C.

Bogenrieder, A.,  and Y.  Doute,  1982.  The effects of UV on photosynthesis and
growth in dependence of mineral nutrition (Lactuca sativa L. and Rumex alpinus
L.).  In Biological Effects  of UV-B Radiation (Bauer, H.,  M.M. Caldwell,  M.
Tevini, and R.C.  Worrest, eds.),  Gesellschaft fur Strahlen-und Umweltforschung
mbH Muchen, pp.  164-169.

Bogenrieder, A.,  and R.  Klein,  1978.  Die abhangigkeit der UV-empfindlichkeit
von der lichtqualitat bel der aufzucht (Lactuca sativa L.).  Angew.  Botanik
52:283-293.

Bridge, M.A.,  and W.L. Klarman, 1973.   Soybean phytoalexin hydroxyphaseollin,
induced by ultraviolet irradiation.  Phytopathol. 63:606-608.

Caldwell, M.M.,  1968.  Solar ultraviolet radiation as an ecological factor for
alpine plants.  Ecol. Monogr. 38:243-268.

Caldwell, M.M.,  1971.  Solar UV irradiation and the growth and development of.
higher plants.  Photophysiol. VI (Giese, A.C., ed.),  pp. 131-177.  Academic
Press, New York.

Caldwell, M.M.,  1977.  The effects of solar UV-B (280-315 nm) on higher
plants:  Implications of stratospheric ozone reduction.  I_n Research in
Photobiology  (Castellani, A., ed.), pp. 597-607.  Plenum Publ. Corp., New
York.

Caldwell, M.M.,  1981.  Plant response to solar ultraviolet radiation.  In
Encyclopedia  of Plant Physiology New Series, Volume 12A.  Physiological Plant
Ecology  I  (Lange, O.L., P.S. Nobel, C.B. Osmond, and H. Ziegler, eds.), pp.
169-197, Springer-Verlag, Berlin, Germany.

Caldwell, M.M.,  L.B. Camp, C.W. Warner, and S.D. Flint, 1986.  Action spectra
and their key role in assessing biological consequences of solar UV-B
radiation change.  In Stratospheric Ozone Reduction, Solar Ultraviolet
Radiation and Plant  Life  (Worrest, R.C., ed.), Springer-Verlag. ISBN 13875-7.

Caldwell, M.M.,  W.G. Gold, G. Harris, and C.W. Ashurst, 1983.  A modulated
lamp  system for solar UV-B (280-320 nm) :  Supplementation studies in the
field.   Photochem. Photobiol. 37:479-485.

Caldwell, M.M.,  R. Robberecht, and  S.D. Flint, 1983.  Internal filters:
Prospects  for UV-allimation  in higher plants.  Physiol. Plant. 58:445-450.

Cams, H.R.,  J.H. Grahm,  and S.J. Ravitz, 1978.  Effects of UV-B radiation on
selected leaf pathogenic  fungi and  on disease severity.  EPA-IAG-D6-0168.
EAGER Program, EPA,  Washington, D.C.

Cronquist,  A., 1981.  An integrated system of classification  of  flowering
plants.   Columbia University Press, New York, p. 1262.

-------
                                    11-27
Dumpert, K.,  and J. Boscher, 1982.  Response of different crop and vegetable
cultivars to UV-B irradiance:  Preliminary results.  In Biological Effects of
UV-B Radiation (Bauer, H.,  M.M. Caldwell, M. Tevini, and R.C. Worrest, eds.),
pp. 102-107.   Gesellschaft fur Strahlen-und Umweltforschung mbH, Muchen.  ISBN
0721-1694.

Eisenstark, A., and G.H. Perrot 1985.  Enhanced UV-B irradiation (290-320 run)
of corn (Zea mays L.), II.   Effects of various stages of tassel formation
(greenhouse-plants), unpublished manuscript.

Eisenstark, A., G.H. Perrot, G. Ulmer, and C.D. Miles, 1985.  Enhanced UV-B
irradiation (290-320 run) of corn (Zea mays L.), I.  Effects on growth and
yield, unpublished manuscript.

Esser, G., 1980.   Einfluss einer nach schadstoffimission vermehrten
einstrahlung von UV-B-licht auf kulturpflanzen, 2. Versuchsjahr.  Bericht
Battelle Institut, E.V. Frankfurt.  BF-R-63, 984-1.

Flint, S.D.,  and M.M. Caldwell, 1983.  Influence of floral optical properties
on the ultraviolet radiation environment of pollen.  Amer. J. Bot.
70:1416-1419.

Flint, S.D.,  P.W. Jordan, and M.M. Caldwell, 1985.  Plant protective response
to enhanced UV-B radiation under field conditions:  Leaf optical properties
and photosynthesis.  Photochem. Photobiol. 41(l):95-99.

Fox, F.M., and M.M. Caldwell, 1978.  Competitive interaction in plant
populations exposed to supplementary ultraviolet-B radiation.  Oecologia
36:173-190.

Fried, J.S.,  K.A. Surano, P.F. Daley, J.H. Shinn,  and P. Anderson, 1986.
Biomass production and nutrient responses of ponderosa pine to long-term
elevated C02 concentrations.  North American Forest Biological Workshop.
Lawrence Livermore National Laboratory, UCRL-94281, reprint.

Garrard, L.A.,  T.K. Van, and S.H. West, 1976.  Plant response to middle
ultraviolet (UV-B) radiation:  Carbohydrate levels and chloroplast reactions.
Soil and Crop Sci. Soc. Florida Proc. 36:184-188.

Gates, D.M.,  1983.  An overview.  In C02 and Plants (Lemon, E.R.,  ed.), pp.
7-20.  Westview Press Inc., Boulder, Colorado.

Gold, W.G., and M.M. Caldwell, 1983.  The effects  of ultraviolet-B radiation
on plant competition in terrestrial ecosystems.  Physiol. Plant. 58:435-444.

Hart, R.H., G.E.  Carlson, H.H. Klueter, and H.R. Cams, 1975.  Response of
economically valuable species to ultraviolet radiation.  In Climatic Impacts
Assessment Program (CIAP),  Monograph 5 (Nachtway,  D.S., M.M. Caldwell, and
R.H. Biggs, eds.), pp. 263-275.  U.S. Department of Transportation, Report No.
DOT-TST-75-55,  National Tech. Info. Serv., Springfield, Virginia.

Hashimoto, T.,  and M. Tajima, 1980.  Effects of ultraviolet irradiation on
growth and pigmentation in seedlings.  Plant Cell  Physiol. 21:1559-1571.

-------
                                    11-28
Rowland, G.P., R.W. Hart, and M.L. Yette, 1975.  Repair of DNA strand breaks
after gamma-irradiation of protoplasts isolated from cultured wild carrot
cells.  Mutat. Res. 27:81-87.

Kasperbauer, M.J., and W.E. Loomis, 1965.  Inhibition of flowering by natural
daylight on an inbred strain of Mellotus.  Crop Sci. 5:193-194.

Keelings, C.D., 1978.  Atmospheric carbon dioxide in the 19th century.  Sci.
202:1109.

Klein, R.M., P.C.  Edsall, andA.C. Gentile, 1965.  Effects of near ultraviolet
and green radiations on plant growth.  Plant Physiol. 40:903-906.

Krizek, D.T., 1978.  Differential sensitivity of two cultivars of cucumber
(Cucumis sativus L.) to increased UV-B irradiance:  I.  Dose-response studies.
Final Report EPA-IAG-D6-0168, USDA/EPA BACER Prog.,  EPA, Washington, D.C.

Lemon, E.R., 1983.  C02 and Plants.  The Response of Plants to Raising Levels
of Atmospheric Carbon Dioxide.  Westview Press Inc., Boulder, Colorado.

Lydon, J., A.H. Teramura, and E.G. Summers, 1986.  Effects of ultraviolet-B
radiation on the growth and productivity of field grown soybeans.  In
Stratospheric Ozone Reduction, Solar Ultraviolet Radiation and Plant Life
(Worrest, R.C., ed.). Springer-Verlag.

Martin, F.W., 1970. The ultraviolet absorption profile of stigmatic extracts.
New Phytol. 69:425-430.

Martin, F.W., and J.L. Brewbaker, 1971.  The nature of the stigmatic exudate
and its role in pollen germination.  In Pollen:  Development and Physiology
(Heslop-Harrison,  ed.), pp. 262-266, Appleton-Century-Crofts, New York.

McWhorter, C.G., and D.T. Patterson, 1980.  Ecological factors affecting weed
competition in soybeans.  In World Soybean Res. Conference II:  Proceedings
(Corgin, F.T., ed.), pp. 371-392.  Westview Press, Boulder, Colorado. ISBN
0-89158-158-899-X.

Mirecki, R.M., and A.H. Teramura, 1984.  Effects of ultraviolet-B irradiance
on soybean.  V. The dependence of plant sensitivity on the photosynthetic
photon flux density during and after leaf expansion.  Plant Physiol.
74:475-480.

Murali, N.S., and A.H. Teramura, 1985a.  Effect of UV-B irradiance on soybean.
VI.  Influence of phosphorus nutrition on growth and flavonoid content.
Physiol. Plant. 63:413-416.

Murali, N.S., and A.H. Teramura, 1985b.  Effect of UV-B irradiance on soybean.
VII.  Biomass and concentration and uptake of nutrient at varying P supply.
J. Plant Nutrition  8(2) : 177-192.

Murali, N.S., and A.H. Teramura, 1986a.  Intraspecific difference in Cucumis
sativa L. sensitivity to ultraviolet-B radiation.  Physiol. Plant.  In press.

-------
                                    11-29
Murali, N.S., and A.H. Teramura, 1986b.   Effect of supplemental ultraviolet-B
radiation on the growth and physiology of field-grown soybean.  Env. Exp. Bot.
In press.

Murali, N.S., and A.H. Teramura, 1986c.   Effectiveness of UV-B radiation on
the growth and physiology of field-grown soybean modified by water stress.
Photochem. Photobiol. In press.

Murali, N.S., A.H. Teramura, and S.K. Randall, 1986.  A comparative study of
ultraviolet-B sensitivity between two soybean cultivars. Env. Exp. Bot. In
review.

Nachtway, D.S., 1975.  Linking photobiological studies at 254 nm with UV-B.
In Climatic Impact Assessment Program (CIAP),  Monograph 5 (Nachtway, D.S.,
M.M. Caldwell and R.H. Biggs, eds.),  pp. 3-50 to 3-84.  U.S. Department of
Transportation, Report No. DOT-TST-75-55,  National Tech. Info. Serv.,
Springfield, Virginia.

National Academy of Sciences (NAS), 1979.   Stratospheric ozone depletion by
halocarbons:  Chemistry and transport.  National Academy Press, Washington,
B.C.

Neftel, A., H. Oeschger, J. Schwander, B.  Stauffer, and R. Zumbrunn.  1982.
Ice core sample measurements give atmospheric C02 content during the past
40,000 years.  Nature 295:220-223.

Owens, O.V.H., and D.T. Krizek, 1980.  Multiple effects of UV radiation
(265-330 nm) on fungal spore emergence.   Photochem. Photobiol. 32:41-49.

Popp, H.W., and F. Brown, 1936.  The effects of ultra-violet radiation upon
seed plants.  In Biological Effects of Radiation (Duggar, B.M., ed.),  Vol. 2,
pp. 853-887, McGraw-Hill, New York.

Reilly, J.J., 1975.  The role of thymine dimers in the induction of
phytoalexin, hydroxyphaseollin, in ultraviolet irradiated soybean suspension
culture.  Ph.D. Thesis, University of Maryland, College Park, Maryland.

Reinert, R.A., A.S. Heagle, and W.W.  Heck, 1975.  Plant responses to pollutant
combination.  In Response of Plants to Air Pollution  (Mudd, J.B., and T.T.
Kozlowski,  eds.), pp. 159-178.  Academic Press, New York. ISBN 0-12-509450-7.

Robberecht, R., and M.M. Caldwell,  1978.  Leaf epidermal transmittance of
ultraviolet radiation and its implication for plant sensitivity to
ultraviolet-radiation induced injury.  Oecologia 32:277-287.

Rowe, Robert D. and Richard Adams.   Analysis of Economic Impacts of Lower Crop
Yields Due  to Stratospheric Ozone (Draft Report).  Prepared for the U.S.
Environmental Protection Agency, August 26, 1987.

Rupert, C.S., 1984.  Cellular repair and assessment of UV-B radiation damage.
In Stratospheric Ozone Reduction, Solar Ultraviolet Radiation and Plant Life
(Worrest, R.C., ed.) Springer-Verlag.  ISBN 13875-7.

-------
                                    11-30
Semeniuk, P., and R.W. Goth, 1980.   Effect of ultraviolet irradiation on local
lesion development of potato virus  S on Chenopodium quinoa cv.  Valdivia
leaves.  Env. Exp. Bot. 20:95-98.

Sisson, W.B., 1981.  Photosynthesis, growth,  and ultraviolet irradiance
absorbance of Cucurbita pepo L.  leaves exposed to ultraviolet-B radiation
(280-315 run).  Plant Physiol. 67:120-124.

Sisson, W.B., and M.M. Caldwell,  1976.  Photosynthesis, dark respiration, and
growth of Rumex patientia L. exposed to ultraviolet irradiance  (280 to 315
nanometers) simulating a reduced atmospheric ozone column.  Plant Physiol.
58:563-568.

Southworth, D.,  1969.  Ultraviolet  absorption spectra of pollen and spore
walls.  Grana Palynologia 9(1-3):1-15.

Soyfer, V.N., 1983.  Influence of physiological conditions on DNA repair and
mutagenesis in higher plants.  Physiol. Plant. 58:373-380.

Steinmueller, D.,  and M. Tevini,  1985.  Action of ultraviolet radiation (UV-B)
upon cuticular waxes in some crop plants.   Planta 164:557-564.

Strain, B.R, and F.A. Bazzaz, 1983.  Terrestrial plant communities.  In C02
and Plants  (Lemon, E.R., ed.), pp.  177-222.  Westview Press Inc., Boulder,
Colorado.

Tanada, T., and S.B. Hendricks,  1953.  Photoreversal of ultraviolet effects in
soybeans.  Amer. J. Bot. 40:634-637.

Teramura, A.H.,  1981a.  Differences in the photosynthetic response to UV-B
radiation between mature and immature leaves.  Plant Physiol. 67 (Suppl):  93
(Abstract).

Teramura, A.M.,  1981b.  Cultivar differences in the effects of enhanced UV-B
irradiation.  Annual Report for EPA.  EPA, Washington, D.C.

Teramura, A.H.,  1982.  The amelioration of UV-B effects on productivity by
visible radiation.  In The Role of Solar Ultraviolet Radiation in Marine
Ecosystems  (Calkins, J., ed.), pp.  367-382.  Plenum Publ. Corp. New York.
ISBN 0-306-40909-7.

Teramura, A.H.,  1983.  Effects of ultraviolet-B radiation on the growth and
yield of crop plants.  Physiol.  Plant. 58:415-427.

Teramura, A.H.,  R.H. Biggs, and S.  Kossuth, 1980.  Effects of ultraviolet-B
irradiances  on  soybean.  II.  Interaction between ultraviolet-B and
photosynthetically active radiation on net photosynthesis, dark respiration,
and transpiration.  Plant Physiol.  65:483-488.

Teramura, A.H.,  I.N. Forseth, and J.  Lydon, 1984.  Effects of ultraviolet-B
radiation on plants during mild water stress.  IV.  The insensitivity of
soybean internal water relations to ultraviolet-B radiation.  Physiol. Plant
62:384-389.

-------
                                    11-31
Teramura, A.H., and N.S. Murali, 1986.  Intraspecific differences in growth
and yield of soybean exposed to ultraviolet-B radiation under greenhouse and
field conditions.  Env. Exp. Bot. In press.

Teramura, A.H., M.C. Perry, J. Lydon, M.S. Mclntosh, and E.G. Summers, 1984.
Effects of ultraviolet-B radiation on plants during mild water stress.  III.
Effects on photosynthesis recovery and growth in soybean.  Physiol. Plant.
60:484-4-92.

Teramura, A.H., M. Tevini, and W. Iwanzik, 1983.  Effects of ultraviolet-B
irradiation on plants during mild water stress.  I.  Effects on diurnal
stomatal resistance.  Physiol. Plant. 57:175-180.

Tevini, M., W. Iwanzik, and U. Thoma, 1982.  The effects of UV-B irradiation
on higher plants.  In Nato Conference Series, Series IV (Chalkins, J., ed.),
pp. 435-448.  Plenum Press, New York and London.

Tevini, M., and W. Iwanzik, 1982.  The effects of UV-B irradiation on higher
plants.  In The Role of Solar Ultraviolet Radiation on Marine Ecosystems
(Chalkins, J., ed.), pp. 581-615.  Plenum Pub. Corp., New York.

Tevini, M., and W. Iwanzik, 1983.  Inhibition of photosynthetic activity by
UV-B radiation in radish seedlings.  Physiol. Plants 58:395-400.

Tevini, M., W. Iwanzik, and A.H. Teramura, 1983.  Effects on UV-B radiation on
plants during mild water stress.  II.  Effects on growth, protein and
flavonoid content.  Z. Pflanzenphysiol 110:459-467.

Tevini, M., W. Iwanzik, and U. Thoma, 1981.  Some effects of enhanced UV-B
irradiation on the growth and composition of plants.  Planta 153:388-394.

Tevini, M., U. Thoma, and W. Iwanzik, 1983.  Effects of enhanced UV-B
radiation on germination, seedling growth, leaf anatomy, and pigments of some
crops.  Z. Pflanzenphysiol. 109:435-448.

Van, T.K., L.A. Garrard, and S.H. West,  1976.  Effects of UV-B radiation on
net photosynthesis of some crop plants.   Crop Sci.  16:715-718.

Vu, C.V., L.H. Allen, and L.A. Garrard,  1978.  Effects of supplemental
ultraviolet radiation (UV-B) on growth of some agronomic crop plants.  Soil
and Crop Sci.  Soc. Florida Proc. 38:59-63.

Wellmann, F.,  1982.   Phenylpropanoid pigment synthesis and growth reduction as
adaptive reactions to increased UV-B radiation.   In Biological Effects on UV-B
Radiation.  (Bauer,  H., M.M. Caldwell,  M.  Tevini, andR.C.  Worrest,  eds.),  pp.
145-149.   Gesellschaft fur Strahlen-und Umweltforschung mbH,  Muchen.

Whittaker, R.H.,  1975.  Communities and Ecosystems.   MacMillan Publishing Co.,
Inc.,  New York.

Zeevaart, J.A.D.,  1976.  Physiology of flower formation.  Ann.  Rev.  Plant
Physiol.  27:321-348.

-------
                                  CHAPTER 12

                       AN ASSESSMENT OF THE EFFECTS OF
                 ULTRAVIOLET-B RADIATION ON AQUATIC ORGANISMS
SUMMARY

    Various experiments have demonstrated that UV-B radiation causes damage to
fish larvae and juveniles, shrimp larvae, crab larvae, copepods, and plants
essential to the aquatic food web.  These damaging effects include decreased
fecundity, growth, survival, and other reduced functions in these organisms.
In natural marine plant communities a change in species composition rather
than a decrease in net production is the probable result of enhanced UV-B
exposure.  The change in community composition may introduce instabilities to
ecosystems and would likely have an influence on higher trophic levels.  A
decrease in column ozone could diminish the near-surface season of
invertebrate zooplankton populations.   Whether the population could endure a
significant shortening of the surface season is unknown.

    The direct effect of UV-B radiation on food-fish larvae closely parallels
the effect on invertebrate zooplankton.  Information is required on seasonal
abundances and vertical distributions of fish larvae, vertical mixing, and
penetration of UV-B radiation into appropriate water columns before effects of
incident or increased levels of exposure to UV-B radiation can be predicted.
However, in one study involving anchovy larvae, a 20% increase in incident
UV-B radiation (which would accompany about a 9% decrease in the atmospheric
ozone column) would result in the death of all of the larvae within a 10-meter
mixed layer in April and August after 15 days.  It was calculated that about
8% of the annual larval population throughout the entire water column would be
directly killed by a 9% decrease in column ozone.

    Effects induced by solar UV-B radiation have been measured to depths of
more than 20 m in clear waters and more than 5 m in unclear water.  The
euphotic zone (i.e, those depths with levels of light sufficient for positive
net photosynthesis) is frequently taken as the water column that reaches down
to the depth at which photosynthetically active radiation is reduced 99%.  In
marine ecosystems, UV-B radiation penetrates approximately the upper 10% of
the marine euphotic zone before it is reduced to 1% of its surface irradiance.
Penetration of UV-B radiation into natural waters is a key variable in
assessing the potential impact of this radiation on any aquatic ecosystem.

-------
                                    12-2
FINDINGS

1.  INITIAL EXPERIMENTS SHOW THAT REDUCTIONS IN STRATOSPHERIC  OZONE. WHICH
   INCREASES SOLAR ULTRAVIOLET RADIATION.  HAVE THE POTENTIAL  TO  HARM AQUATIC
   LIFE.  DIFFICULTIES IN EXPERIMENTAL DESIGNS AND THE LIMITED SCOPE OF  THE
   STUDIES PREVENT THE QUANTIFICATION OF RISKS.

   la.   Increases in energy in the 290-320 nm wavelengths  that would occur if
        the ozone layer were depleted could harm aquatic life.

   Ib.   Various experiments have shown that UV-B radiation damages fish  larvae
        and juveniles, shrimp larvae, crab larvae, copepods,  and plants
        essential to the marine food web.

   Ic.     Up to some threshold level of exposure, most zooplankton show  no
          effect due to increased exposure to UV-B radiation. However,
          exposure above the dose threshold elicits significant  and
          irreversible physiological and behavioral effects.

   Id.     While the exact limits of tolerance and current exposure have  not
          been precisely determined, estimates of these two properties for a
          variety of aquatic organisms show them to be essentially equal.

   le.     The equality of tolerance and exposure suggests that solar UV-B
          radiation is currently an important limiting ecological factor,  and
          the sunlight-exposed organisms sacrifice potential  resources to
          avoid increased UV-B exposure.  Thus, even small increases of UV-B
          exposure would be likely to further injure species  currently under
          UV-B stress.

    If.    A decrease in column ozone is reasonably likely to  diminish  the  time
          that zooplankton can survive or breed at or near the surface of
          waters they inhabit.  For some zooplankton, the time they spend  at
          or near the surface is critical for breeding.  Whether the
          population could endure a significant shortening of surface  time is
          unknown.

    Ig.    Sublethal exposure of copepods produces a reduction in fecundity.

    Ih.    Of the animals tested, no zooplankton possess a sensory mechanism
          for directly detecting UV-B radiation; therefore,  it would be
          unlikely that they would actively avoid enhanced levels of exposure
          resulting from a reduction in column ozone.

    li.    Exposure of a community to UV-B stress  in controlled experiments has
          resulted in a decrease in species diversity, and therefore a
          possible reduction in ecosystem resilience and flexibility.

    Ij.    One experiment predicted an 8 percent  annual loss  of the larval
          anchovy population from a 9 percent reduction in column ozone in a
          marine system with a 10-meter mixed layer.

-------
                                    12-3
2.   IN COMMON WITH ALL OTHER LIVING ORGANISMS.  THE  AQUATIC  BIOTA COPE WITH
    SOLAR UV-B RADIATION BY AVOIDANCE.  SHIELDING. AND  REPAIR MECHANISMS.
    UNCERTAINTY EXISTS AS TO THE EXTENT TO WHICH SUCH  MITIGATION MECHANISMS
    WOULD OCCUR.

3.   DETERMINATION OF UV-B EXPOSURE IN AQUATIC SYSTEMS  IS COMPLEX BECAUSE  OF
    THE VARIABLE ATTENUATION OF UV-B RADIATION IN THE  WATER COLUMN.

    3a.   Because aquatic organisms are small and do not usually have  fixed
          locations,  it is very difficult to obtain accurate data needed  to
          model the systems and verify  results.   Current understanding of the
          life cycle of organisms is very limited.

4.   ABOUT ONE HALF OF THE WORLD'S PROTEIN IS DERIVED FROM MARINE SPECIES.   IN
    MANY THIRD WORLD COUNTRIES. THIS PERCENTAGE IS  LARGER.   RESEARCH IS NEEDED
    TO IMPROVE OUR UNDERSTANDING OF HOW OZONE DEPLETION COULD INFLUENCE THESE
    SYSTEMS.

    4a.   A comprehensive analysis of sublethal and lethal  effects of solar  UV
          on littoral, benthos, and planktonic ecosystems is needed.

    4b.   A model of energy flow analysis leading to protein production where
          solar input is augmented by increased ultraviolet radiation would be
          required to better evaluate potential effects. Marine organisms
          responses to projected increases in UV must  be considered in the
          context of the oceans as a dynamic moving fluid.

    4c.   Better documentation of the effects of present levels of ultraviolet
          light on marine organisms is  needed.

    4d.   Intensive research is needed to identify  biochemical indices that
          reflect UV stress in marine organisms.

-------
                                     12-4
INTRODUCTION

    As explained in Chapter 2, it is ozone,  though in very small
concentrations in the atmosphere, that is almost entirely responsible for the
absorption of solar ultraviolet (UV) radiation in the 290-320 nm wave band
(Exhibit 12-1).

    If Exhibit 12-1 were modified to account for a 10% ozone layer reduction,
given the scale, one would scarcely notice the difference.  The additional UV
radiation reaching the earth's surface as a result of a 10% ozone reduction
would amount to less than a 0.05% increase in the total sunlight energy.   At
the very margin of the diagram, however, the percentage increases would be
much larger because the base is so small, and it is in those wavelengths that
biological responses to the sun's ultraviolet radiation appear substantially
greater.

    Based on models of depletion, a 10% ozone reduction at 45°N latitude would
result in a 28% increase for biologically effective radiation for DNA damage
(Setlow 1974).  The generalized plant response would increase by 21% (plant
damage) (Caldwell 1968).  Exhibit 12-2 shows the relationship between ozone
depletion and action spectra.  Because of the relatively high increases in
UV-B at the lower ends of this part of the spectrum and its relative
biological effect, ozone depletion constitutes a threat to aquatic life.

BACKGROUND ON MARINE ORGANISMS AND SOLAR ULTRAVIOLET RADIATION

    Organisms in fresh water or the oceans are either swimmers  (nekton),
bottom-dwellers (benthos),  or drifters (plankton).  Plankton can be either
plants (phytoplankton) or animals (zooplankton).  An important zooplankton
group, the ichthyoplankton, is comprised of the drifting  eggs and larvae of
many species of fish.

    Phytoplankton provide essentially all of the chemical energy required by
the marine food web through the photosynthetic process (primary production).
Zooplankton are consumers (grazers).  Zooplankton are usually less than about
5 cm in diameter or length, most often less than 1-2 mm.  Nearly all groups  of
aquatic animals, at least for some phase in their life history, are considered
zooplankton -- for some species, in the egg and/or larval stage, for others,
throughout their life cycle.  Zooplankton are critical components in typical
aquatic feed webs (nutrient pathways) that lead to larger animals, including
those comprising the fisheries  (finfish and shellfish), and  therefore, affect
man himself.

    In marine ecosystems, UV-B  radiation penetrates  approximately the  upper  10
meters of the marine euphotic zone before it  is reduced to 1% of  its surface
irradiance.  Penetration of UV-B radiation into natural waters  is a key
variable in assessing the potential impact of this radiation on any aquatic
ecosystem (Exhibit 12-3).

    Photosynthesis requires light, but  adequate sunlight  seldom penetrates  to
the bottom of natural waters.   Thus, aquatic  plants  are largely confined to  a
relatively thin layer at the  surface of  the water, which  is  termed  the

-------
                                   12-5
                              EXHIBIT 12-1
         Solar Irradiance Outside the Earth's Atmosphere and at the
            Surface of the Earth for a Solar Zenith Angle of 60°
 0
 200
400      600
800     1000     1200
WAVELENGTH   (nm)
1400
1600
1800
Adapted from Forsythe 1954.  Regions of significant  absorption by atmospheric
ozone,  oxygen, water vapor, and carbon dioxide are shown.

-------
                             12-6
                         EXHIBIT 12-2

      Relationship between Ozone Depletion and Biological .
           Effectiveness of Increased UV-B Radiation
       (Based on model by Green, Cross, and Smith,  1980)
                                         Increase in Biological
Ozone
Decrease
10%
20%
30%
40%
Increase in
290-320 run
8%
17%
27%
38%
UV Radiation
290-360 run
1.1%
2.4%
3.7%
5.2%
Effectiveness
DNA
28%
67%
125%
213%
Plant
21%
49%
85%
132%
a/
    Action spectra are referenced to 300 nm - 1.00.

-------
                                  12-7
                             EXHIBIT 12-3

            Solar Spectral Irradiance at the Surface of the Ocean
                            and at Four Depths
   600r

UJ
O 500
<
g 400
cc
oc
~ 300
UJ
r- 200
100
                                     PERCENTAGE OF ENERGY
                                     2    4  6 810  20  4060100,
0.2   0.4   0.6
                       0.8    1.0   1.2    1.4   1.6
                            WAVELENGTH
                                               1.8   2.0   2.2   2.4
  Adapted from Sverdrup, Johnson, and Fleming 1942.  Insert illustrates
  percentage of energy transmitted at depth in oceanic waters.

-------
                                     12-8


euphotic zone, the surface layer where plants create more chemical energy by
photosynthesis than they use for their own metabolism.  This zone is where
UV-B can penetrate.  Zooplankton are found at all depths, depending on species
and season, but are most abundant in the sunlit, upper 100 meters where the
bulk of their food, including phytoplankton, is found.  Many species normally
live very close to the surface, even in daylight, while others occupy the
near-surface layer during only part of their life cycle.  The near-surface
layer is a very important zone in the interactions of the physical/chemical/
biological components of aquatic systems.

    Zooplankton have apparently evolved mechanisms and behavior by which they
have adjusted to current levels of UV radiation (Damkaer 1982),  but they may
not be able to adjust to relatively rapid increases in total UV exposure.  If
there are changes in abundance of Zooplankton species, those changes would
have an impact far beyond any direct effects because of the critical role
zooplankton play in energy transfer within the ecosystem.

Limitations of Pioneer Investigations

    As early as 1925, scientists were aware of damaging effects from the
ultraviolet component of sunlight on aquatic organisms (Huntsman 1925; Klugh
1929,  1930; Harvey 1930; ZoBell and McEwen 1935; Giese 1938; Bell and Hoar
1950;  Dunbar 1959; Marinaro and Bernard 1966).  These reports are, for the
most part, of historical interest only; they cannot be strictly related to
present investigations because of a lack of precise or absolute measurements
of UV dose-rates and doses.  A lack of proper instrumentation for UV
irradiance measurements in aquatic research, which included laboratory
studies, persisted into at least the late 1960's.

Limitations of Recent Investigations

    There are only a few reports that were based on experimental methods
appropriate to the problem of enhanced solar UV-B radiation.  This meager body
of knowledge does not compare well to the enormous literature available on,
for example, chemical pollution.  Even for chemical pollution there are still
great uncertainties in predicting ecosystem responses.  It must be recognized
that attempts to predict ecosystem responses to enhanced UV-B radiation are in
the earliest stages.  A number of reports are difficult to apply because the
experiments used extraordinarily high UV-B dose-rates and/or doses.
Frequently, investigators increased radiation in damaging wavelengths far
beyond the amount that can be expected from ozone depletion.

   Another problem in predicting UV effects is that we are not yet very close
to being able to extrapolate the effects on one species to those on another.
Nor can we say with much assurance how the same species might react to UV
radiation in a different environment.  There are simply too few data of an
acceptable quality to make anything other than a general assessment of risk.
All that follows must start off with these caveats.

-------
                                    12-9



EFFECTS OF UV-B RADIATION ON PHYTOPLANKTON

Direct Inpact on Prinary Productivity

    The amount of UV-B radiation reaching the ocean's surface has long been
suspected as a factor influencing primary production.  Research shows
convincingly that ultraviolet radiation,  at levels currently incident at the
surface of natural bodies of water,  has an influence on phytoplankton
productivity (Nielsen 1964;  Jitts et al.  1976; Hobson and Hartley 1983).  It
has been calculated that, near the surface of the ocean, a 25% reduction in
ozone concentration would cause a decrease in primary productivity by about
35% (Smith et al. 1980).  The estimated reduction in production for the whole
euphotic zone would be about 10%.  Because these calculations were based on
attenuation lengths (i.e., the product of depth in the water column and the
diffuse attenuation coefficient of water), waters of various turbidities and
absorption characteristics can be compared.

    If one assumes that present phytoplankton populations sense and control
their average vertical position in such a way as to limit UV-B exposure to a
tolerable level, then a 10% increase in solar UV-B radiation would necessitate
a downward movement of the average position, thereby reducing the average UV-B
exposure by 10%.  There would be a corresponding reduction of light for
photosynthesis.  The loss of photosynthetically active radiation in many
locations might not be significant.   However, in some very productive areas,
especially high latitude ocean areas,  photosynthetically active radiation is
the primary limiting factor for marine productivity  (Russell-Hunter 1970).
The loss of photosynthetically active radiation from optical measurements has
been estimated to be in the range of 2.5%-5% for a 10% increase in UV-B
radiation (Calkins 1982a).  If the photosynthetic base of aquatic ecosystems
were perturbed, one would expect ramifications to extend up through the food
web through predator-prey relationships.

    Experiments with marine diatoms have shown significant reductions in
biomass, protein, and chlorophyll at UV-B  irradiances equivalent to ozone
reductions that would occur for a 5%-15% ozone depletion.  Laboratory studies
on chain-forming diatoms and other phytoplankton show that growth increased
when UV-B radiation was filtered out of the incident solar radiation,
indicating that existing levels of UV-B radiation depress productivity
(Thomson, Worrest, and Van Dyke 1980; Worrest 1982).

Effects Because of Changes in Motility

    An additional effect is that UV-B radiation may  endanger  the survival of
microorganisms (Euglena, slime mold, some  blue-green algae) by decreasing
their motility and by inhibiting phototactic and photophobic  responses  (Haeder
1986).   The inability of a population to move into favorable  environments
could result in damage that may impair development.

-------
                                    12-10
Effects on Photosynthesis

    Jones and Kok (1966) observed a direct inhibitory effect  of UV  radiation
on photosynthesis in chloroplasts.   This form of inhibition is a  short-term
response and shows a different UV-B wavelength dependence  from most other
forms of UV-B injury.  Smith et al.  (1980) found that the  short-term
depression of photosynthesis by natural solar UV radiation measured in the
ocean resembles the direct photoinhibition described by Jones and Kok.   If
direct photoinhibition were the only pathway for reduction of photosynthesis
by marine phytoplankton, the effect of ozone reduction would  be minor;
however, Smith and Baker (1980) point out that there could be long-term
effects that might show a different action spectrum and be of more
consequence.

    In Hawaii long-term photoinhibition of the growth of six  algal  species
under natural sunlight has been measured (Jokiel and York 1984).  Two strains
could not grow at all at the levels of irradiance for full natural  surface
sunlight.  Of the species tested, those collected from tropical  surface water
showed the greatest adaptive power,  but it is reasonable to conclude that
resistance to solar UV radiation is achieved through expenditure  of resources
that can be better applied to other needs in less exposed species.

    Worrest et al. (1981) found a short-term (4-5 h) depression  of
photosynthesis in seven strains of algae upon exposure to enhanced  levels  of
UV radiation.  There was a large difference in strain sensitivity.   The
observed depression appeared to linearly correlate with UV-B  dose using the
DNA (Setlow 1974) action spectrum.  This is in contrast to the  results of
Smith et al.  (1980) who found the Jones and Kok weighting function  to be
appropriate.

Multi-Day Effects

    Measurements of photosynthesis that span only a single day could
underestimate the overall action of solar UV exposure by failing to account
for effects on the next day's population level.  The effects  of UV-B radiation
could result in delaying growth or direct killing.  Thus,  subsequent
population could fall below the numbers that an unexposed population would
attain.  Prolonged delays (about two days) in growth of the survivors have
been observed in two strains of the diatom Thalassiosira,  which had been
irradiated with simulated solar UV-B radiation at doses below lethal levels.
If unicellular organisms are in a rapid growth phase, a growth delay equalling
the time of one growth cycle produces the same effect on the subsequent
population as would be produced by a 50% killing  (without growth delay).

Shifts in Community Conposition and Possible Implications

    Shifts in species diversity and community composition of phytoplankton
communities have been observed in a simulated marine ecosystem exposed  to a
UV-B equivalent to ozone decreases of 15%  (Worrest  et al. 1981).   Natural
communities appear to show changes in species composition rather than  a
decreasing net production.  Experiments  in simulated marine  ecosystems  seem to
indicate that higher UV-B radiation will  not decrease biomass and  chlorophyll
accumulation (Worrest et al. 1981).  However, a  change  in community
composition might result in a more unstable  ecosystem and might have  an impact

-------
                                    12-11
on higher trophic levels (Kelly 1986).  For example, one effect of enhanced
levels of UV-B radiation would be to alter the size distribution of the
component producers in a marine ecosystem.  Increasing or decreasing the size
of the representative primary producers could alter the energy allotment that
grazers use for finding and consuming food.  This could reduce the feeding
efficiency of the consumer, which would have a variety of effects.  In
addition, the food quality of the producers is altered by exposure to UV-B
radiation.  It has been demonstrated that the protein content, dry weight, and
pigment concentration are all depressed or the composition altered by enhanced
levels of UV-B radiation (Doehler 1984, 1985).  Consequently, a UV-induced
shift in the distribution and abundance of different kinds of aquatic plant
organisms could have important ramifications throughout the ecosystem.

EFFECTS ON INVERTEBRATE ZOOPLANKTON

Background

    There have been several studies on the effects of UV-B radiation on a
variety of zooplankton.  Investigators have reported that the effect of
increased UV-B radiation on some marine zooplankton (e.g., copepods, shrimp
larvae, crab larvae) is increased organism mortality and decreased fecundity
in the survivors.  Regardless of the species investigated, the potential
exists for a disruption of the food web, which would indirectly affect fish
populations that are important as food.  Of course, studies that consider
economically important zooplankton species, such as larvae of certain shrimp,
crab, and fish, are important because damage to these species would directly
reduce the amount of food for human consumption.

Methodological Issues

    How well photorepair mechanisms might mitigate damage is an important
issue.  It appears that up to some level  (daily-dose threshold), photorepair
will mitigate damage.  However, it also appears that near-surface exposure
levels that could accompany a significant reduction in column ozone would
exceed threshold levels.

    The issue of thresholds is important.  It appears that regardless of the
cellular-level responses to UV irradiation, it is usually noted that up to
some level of UV exposure, the organism appears to suffer no harm.
Extrapolating survival curves for different UV exposure levels would  indicate
that organisms can cope with the damage produced by some levels of additional
UV radiation without negative effects.  At greater doses, where the survival
curve becomes steep, it appears that the repair systems may become inactivated
by the radiation, or that the damage to the general tissues has gone beyond
the capacity of the repair systems.

    A study by Damkaer et al. (1980) illustrates the importance of threshold
effect.  The authors found no alterations in activity, growth rate,
developmental rate, and survival in a number of shrimp and crab larvae exposed
to low levels of artificial UV-B radiation.  However, when a certain
combination of UV dose-rate and total dose was exceeded, negative effects
became quite sharply noticeable.

-------
                                    12-12
Lethal Effects of Enhanced DV-B Radiation

    Damkaer et al.  (1980) have done a number of studies in which enhanced UV-B
radiation produced irreversible damage and death.   UV thresholds,  based on the
lowest level of irradiance, were statistically significant (exceeding the 95%
confidence limits in five replicates).  Irreversible effects occurred compared
to the controls.  Damkaer et al. (1980) have compared the effective UV
thresholds with the maximum effective solar daily  dose rates at an
experimental site in Washington state (Exhibit 12-4).  The results indicated
that the maximum solar levels are currently exceeded for all tested organisms.
Obviously, these animals can survive threshold values for maximum solar UV
levels.  To be effective, UV thresholds probably must be exceeded during
several consecutive days.  A more realistic approach would be to compare the
threshold levels with median solar UV levels (Exhibit 12-5).  The threshold
for all groups would then appear to be above the present median solar incident
UV levels, at least until late in the time span of natural occurrence near the
surface.  The data seem to suggest that UV levels  have exerted a considerable
influence in the long-term adjustments of these populations to those specific
seasons.  Of course, many other physical as well as biological factors operate
concurrently, and the organism's life cycle is a compromise with the total
environment.  Late in the surface season, it is possible that natural UV
levels exert a detrimental effect, particularly on the shrimp larvae, which
have a lower total dose tolerance than crab larvae.

    Damkaer et al.  (1980) compared the estimated effective UV daily dose,
simulating various ozone reductions at the experimental site with the UV
survival thresholds (Exhibit 12-5).  Obviously, the probability that UV
thresholds would be exceeded increases with diminishing ozone.  With a 40%
ozone reduction, there appears only to be a short  "window" of safety at the
beginning of each group's surface season.  Even a  20% ozone reduction
significantly shortens the season.  Whether or not the populations could
endure with a drastically reduced time of near-surface occurrence is not
known.  Success of any year-class depends on the timing of a great number of
other events besides UV level.  Under current stress conditions, early larvae
may do well one year, whereas only late larvae may survive in a subsequent
year.  However, an additional stress like enhanced levels of UV-B radiation is
not likely to be beneficial.  Combined with other  stresses that make early
year survival impossible, it could devastate populations.

    That larval stage development appeared to be unaffected below survival
threshold UV levels is an important observation, since the survival  of the
unharmed larvae is critical to their proper and timely development.  Beyond
the threshold levels, however, development and survival rapidly declined.  The
apparent UV thresholds are near-current  incident surface UV levels.
Nevertheless, the implications of these experiments  for zooplankton  in nature
remains in some doubt because of lack of information on the subsurface UV
spectrum, the effects of vertical mixing, unknown  in situ behavior of
irradiated zooplankton, and unknown sublethal and pre-lethal  effects.

-------
                        12-13
                    EXHIBIT 12-4

Lethal Effects on Shrinp Larvae for Various Combinations of
             UV-B Dose-Rate and Total Dose
                             Feb-Apr Meon Daily Dose (DMA)
                                               Ozone
HUU
*i
in
Jo 30°
i
LO
00
CM
»— •
< 200
Q
r—i
CM
j= 100
"— j
LU
CO
O
Q
r— p ourruce i m neuuciion
.

C


-

O
—

0 0
_ 0 °
A a 0%
B b 16%
C c 40%

• •

•
_
0
•
• •
*
.^ •**
ON* A
- ° R C
o A B ^
Q i n 1
	 1 I 1 III ... U .11 .1.1,1
            20         40

       DAILY DOSE
                                          60         80

                                             (285-315nm)
                                            100
.002
                               .004
                                    006
                           -2
.008
                DOSE  RATE  [Wm~]DNA (285-315 nm)


Open circles indicate maximum observed combinations for survival; closed
circles indicate minimum  observed combinations for mortality.  Open vertical
bars indicate mean February-April surface solar UV-B daily dose-rates  at
various atmospheric ozone concentrations; closed vertical bars indicate mean
daily dose-rates at 1 m (adapted from Damkaer et al. 1980).

-------
                                         12-14
                                    EXHIBIT 12-5


                   Estimated Effective UV-B Solar Daily Dose  at Various
             Atmospheric Ozone Concentrations Based on 4-Year Mean of Medians
                           Manchester,  Washington, 1977-1980
CXI

 '
 j=
   400r


< 35°
z
? 300

   250
      200
 (S)

 §   150
       50
                   SUN  + SKY
Q. Shrimp larva,
    euphausid  larva

b. Adult euphausid,
    crab zoea
C. Crab megalopa
                                              Ozone
                                            reduction
40%
25%
16%
7.5%
 0%

4-year mean of medians
Manchester, Wa.
1977-1980
                    Mar   Apr  May June July  Aug  Sept  Oct   Nov
                                  MONTH
       Dose increments are from Gerstl, Zardeki, and Wiser 1986.  Also shown are
       approximate thresholds of UV-B  daily dose for principal  experimental
       zooplankton groups, in natural  seasonal position (adapted  from Damkaer et al.
       1981).

-------
                                    12-15


    The degree to which other zooplankton  groups would be affected by these
levels of UV-B irradiation cannot be predicted,  although in crustaceans,  the
same basic tissues would likely be involved,  and therefore,  the response  would
be qualitatively similar.   Species do vary substantially, however, in terms  of
their sensitivity to UV-B radiation.  If the  accumulated level of UV-B
radiation is the most important factor in  damage, then increased UV radiation
through ozone depletion may significantly  affect even species that are
moderately resistant to short exposures of high UV intensities.

    Copepods are the most important zooplankton from the perspective of
ecosystems.  Karanas,  Van Dyke, and Worrest (1979) reported on the survival  of
developmental stages of the estuarine copepod Acartia clausii exposed to  daily
UV-B doses approaching natural (Oregon coast) daily doses.  Although the  total
doses in this experiment were relevant, the dose-rates exceeded natural
dose-rates by about a factor of two.

    It is known that organisms are protected to some degree by their own
pigmentation, and this suggests that organisms may have evolved pigmentation
as a consequence of the damaging effects of solar radiation.  Damkaer (1982)
reviewed several possibilities for this in regard to zooplankton.  Siebeck
(1978) found a direct relationship between pigmentation in cladocerans of
mountain and lowland lakes and their apparent natural UV exposure.  In field
and laboratory experiments, the more-pigmented cladocerans tolerated the
greater exposures.  Siebeck also noted a marked photorepair potential.

    Hairston (1980) has drawn attention to the relationship between vertical
distribution and pigmentation in some freshwater copepods.  Typically, the
most pigmented forms are closest to the surface.  Because the pressure of
predation should be against the conspicuous pigmented forms, Hairston
concluded that "photodamage may be a more important selective  force than
previously supposed."

    These kinds of copepods do have a UV-tolerance related to  their
pigmentation, as has been verified  in laboratory experiments.  Ringelberg,
Keyser, and Flik  (1984) studied the UV-B tolerance of a  red-pigmented and a
translucent variety of the same copepod species  from separate, though nearby,
lakes in France.  The red copepod tolerated higher artificial  UV-B doses.  The
experimental doses were obtained by exposure to  a single  dose-rate for varying
amounts of time.  Sharp mortality increases began in the  translucent  form at
about half the total dose as for the pigmented  form.

    An interesting aspect of these  experiments  is that  in all  cases where
death occurs, the dose-rate was equivalent to that determined  to  be  the
threshold dose-rate for shrimp  (Damkaer, Dey, and Heron 1981).  The  total
doses tolerated by the two copepod  color-forms  bracket  the  total  dose
threshold of shrimp larvae.  The tolerance limit appears  to  be the  same  in
vulnerable zooplankton, suggesting  perhaps a universal  tolerance  limit in
near-surface organisms of the  structure and  size of shrimp  larvae and
copepods.  It would be useful  to conduct additional experiments with  such
copepod color-forms and include exposures to lower and  higher  dose-rates at
more natural periodicities.  Also,  Ringelberg et al. observed  no  photorepair,
but indicated that the intensity of visible  radiation  in their experiments was
extremely low (~  0.4 W m  ).

-------
                                    12-16
    Thomson (personal communication) has demonstrated that current levels  of
UV-B radiation are of significance in the developmental life of several
species of shellfish.  For some species a 10% decrease in column ozone could
lead to as much as an 18% increase in the number of abnormal larvae produced.
Exhibit 12-6 is a compilation by Thomson of data denoting the estimated
biologically effective UV-B doses leading to significant effects in major
zooplankton groups.

Sublethal Effects of UV-B

    There are many examples of sublethal stresses (of all kinds) that can
affect natural populations.  If stresses alter the reproductive capacity of
organisms, effects can be significant and rapid.  Karanas, Worrest, and  Van
Dyke (1981) reported a possible mechanism through which enhanced solar UV-B
radiation could have a large impact on copepod populations without direct
killing.  The estuarine copepod Acartia clausii was irradiated with a UV
radiation dose that was comparable to mid-summer surface irradiance at 45°N.
After five days, the surviving copepods were mated.  If either (but not  both)
parent had been irradiated, the egg production and subsequent hatch of the
female was significantly reduced.  If both parents had been irradiated,  the
production of viable larvae was halved.  The subsequent survival of the
hatched larvae from irradiated crosses did not appear to be significantly
affected.

    Even though Karanas, Worrest, and Van Dyke (1981) applied a natural-level
surface dose-rate of artificial UV-B radiation, this dose-rate was still
rather high and probably represents an extreme situation.  To obtain
reasonable total doses, the copepods were exposed for only 1-2 hours.  Real
dose-rates, especially at depth, would be considerably less.  These copepods
are found at the surface, as well as some distance below, but their
distribution on the required scale of 1 m or less is not well known.  Acartia
clausii are taken in a wide variety of water types, but generally this is  a
coastal/estuarine species.  Effective UV-B irradiation may not extend beyond a
depth of a few meters into the usual habitat of Acartia claussi.  In view of
the apparently profound sublethal, reproductive response  to UV-B radiation, it
would be of great interest to have additional multi-generation observations.

Possible Photorepair of DV-B Damage: Implications for use of Experimental Data

    The photochemistry and photobiology of UV-B-induced damage, particularly
nucleic acid disruption, as well as the several molecular repair mechanisms
that are consistently found in plants and animals, have been reviewed by
Caldwell and Nachtwey (1975).  Among the repair mechanisms,  the quantitative
capacity of the dark-repair systems is quite different from  those  systems that
require light.  Under typical experimental conditions, it is presumed that the
dark-repair mechanisms could function freely.   Dark-repair systems  appear to
offset a very high proportion of the damage  caused by relatively  low doses of
irradiation.  However, photorepair mechanisms  often appear capable  of
repairing only  so much damage, even over a wide UV-B dose range  (Murphy 1975).
Consequently, the restorative effects of photoreactivation may be  regarded as
particularly critical in situations where organisms are  subjected to UV-B
irradiation near and above tolerance limits.   There has  been an implication in

-------
                                12-17
                            EXHIBIT 12-6
         Estimated Biologically Effective UV-B Doses Leading
      to Significant Effects in Major Marine Zooplankton Groups
Zooplankton Group
Shrimp/Euphausiid larvae
Euphausiid adults
Copepod larvae
Copepod postlarvae
Copepod adults
Crab larvae
Crab postlarvae
Ancho vy/Macke re 1
Oyster/Mussel larvae
Dose Rate
W m (DNA eff)^7
6.0 x 10"2
9.9 x 10
3.4 x 10']
1.6 x 10'
3.4 x 10
9.9 x 10~2
2.8 x 10
6.0 x 10"2
1.2 x 10"1
Total Dose
J m (DNA eff)
2.6 x 103
6.5 x 103
1.4 x 103
2.7 x 103
3.0 x 103
6.5 x 103
6.0 x 104
2.5 x 104
2.2 x 103
Time For
Effect
4 days
6 days
1.0 hour
4 . 5 hours
2 . 5 hours
6 days
20 days
12 days
5 . 0 hours

Compilation by Thomson.

-------
                                    12-18
some review papers (e.g.,  NAS 1984)  that UV-B damage will be  overcome by  the
maximally functioning photorepair mechanism.   Up to some  level  (daily dose
threshold) this is probably true, but beyond that threshold there  will  be some
residual, irreversible damage.

   Experiments with artificial sources of UV-B radiation  usually are conducted
in the laboratory under light regimes using more than an  order  of  magnitude
less visible light than found in nature.  Near UV and/or  visible light  is
needed for photoreactivation, and the accuracy of some observations may have
been influenced by less-than-maximum photorepair in the laboratory.
Experiments with shrimp larvae and adult euphausiids (Damkaer and Dey 1983),
comparing survival at various UV-B doses and dose-rates combined with
different levels of near-UV and visible irradiance, suggest that photorepair
occurs in these organisms.  The dramatic sensitivity of both shrimp  larvae  and
adult euphausiids when deprived of photoreactivating light clearly
demonstrates the relative importance of photorepair under UV stress.  However,
this apparent photorepair reaches maximum levels at relatively low levels of
visible irradiance.  Consequently, it is doubtful that photorepair and  related
survival would have been greater if experimental levels had been equal  to
solar levels of irradiance.

   A similar relationship has been reported between survival in UV-damaged
anchovy larvae and level of irradiance of photoreactivating light needed for
repair (Kaupp and Hunter 1981).  The level needed for maximum photorepair is
far less than ambient solar irradiation.  The photoreactivation experiments on
shrimp larvae and euphausiids suggest that the previously derived
dose/dose-rate thresholds are not altered with additional photoreactivating
light.  Clearly, a significant lack of photoreactivating light, which normally
would be concurrent with solar UV-B irradiance, would certainly depress
threshold levels.  Past experiments have not approached the intensity or dose
of local solar photoreactivating light.  However, for UV-B exposures below
threshold levels, maximum photorepair potential has been reached at the  low
levels (2-9 W m" ) of photoreactivating light.  Nearly doubling the dose-rate
and dose of photoreactivating light did not  increase the survival of
euphausiids.  With anchovy larvae, Kaupp and Hunter  (1981) reported that the
photorepair mechanisms appeared  to be "fully activated" at a photoreactivating
light (320-500 nm) irradiance of 7 W m   and a daily dose of 77.5 kJ m
available over four days.  At  greater irradiance levels,  up to  levels
equivalent to solar photoreactivating light  at 45°N  latitude, Kaupp and  Hunter
observed no further increase in  anchovy survival.  There  is extraordinary
close agreement between the  relatively  low dose-rates  and doses necessary for
apparent maximum photoreactivation in anchovy  larvae  and  euphausiids.  Thus,
most of  the experimental  data  cannot be eliminated  on  the basis of failing to
account  for photorepair.

   Another important point is  that from these  observations,  it  is clear  that
in nature there would always be  sufficient photoreactivating light to  provide
maximum  photorepair potentials  in shrimp  larvae  and euphausiids (and also  in
anchovy  larvae).   But it  must  be emphasized  that even maximum  photorepair may
not be sufficient  to reverse all potential  injuries  caused by  UV-B radiation.

-------
                                    12-19
Possible Vertical Distribution: Implications for Use of Experimental Data

   The actual solar radiation to which surface-living zooplankton are exposed
in nature is very important in assessing risks to higher ambient UV-B.  Given
the large effects UV radiation had on zooplankton and the apparent existence
of threshold levels, understanding UV-B in the natural environment is critical
to assessing risks.  To assess the potential dangers to zooplankton
populations from increasing solar UV-B irradiation requires consideration of
the subsurface irradiance levels.  The data for the shrimp larvae, Exhibit
12-4, show that the mean of the observed daily UV-B surface doses (current
ozone thickness) between mid-February and mid-April is within the
experimentally determined safe zone.  Daily doses calculated for a 1-m depth
are therefore even farther within the safe zone, so that presumably the shrimp
larvae could tolerate these mean dose-rates for long periods.  With a 16% or
40% ozone reduction, the total dose thresholds would be exceeded at the
surface only if the mean dose-rates are attained for two or three consecutive
days (two sunny days).  Even then, however, the calculated mean UV-B levels at
1 m would remain in the safe zone.  Understanding the implications of higher
UV-B requires knowing where the shrimp larvae spend their time.

   The actual vertical distributions of the shrimp in nature are not known.
In the laboratory, young larvae are attracted to light and definitely prefer
the near surface.  The larvae do not appear to avoid high UV-B irradiance
(Damkaer and Dey 1982),  It is not likely they can sense this radiation in
sufficient time to prevent lethal damage.  Damkaer and Dey assumed that the
younger shrimp larvae are constantly moving toward the surface, and that
during the day the majority of the population is in the upper meter.

   Vertical motion makes it difficult to determine the probability that
particular shrimp larvae would receive lethal exposures of UV-B radiation;
Exhibit 12-7 shows an approximate estimate.  The probability is obviously
time-dependent.  At the surface, shrimp larvae are unlikely to encounter
lethal daily or bi-daily UV-B exposures until the end of the surface season
(mid-April).   This is also about the time that the mean noon dose-rate exceeds
the dose-rate threshold.  Before then, the shrimp larvae would be exposed
daily to a one-third total dose limit for three consecutive days.  This level
is in the safe exposure zone because the dose-rate would be low (Exhibit
12-7).   At 1 m, the mean dose-rates in mid-April are still less than the
shrimp's threshold dose-rates, yet the surface probabilities for lethal
exposure may be sufficiently high at that time to limit the seasonal
occurrence of the larvae.  With a 16% ozone reduction, the current mid-April
UV-B conditions would be reached near the end of March.  If UV-B radiation is
a contributing factor in regulating the seasonal occurrence of shrimp larvae,
ozone reductions may act to shorten the surface time available to the larvae.
The seasonal relationships in the UV-B tolerance of different zooplankton
groups (Exhibit 12-7) suggest that UV irradiance is an important environmental
factor.   Hunter, Kaupp, and Taylor (1982) postulate that the surface
occurrence of eggs and larvae of a number of clupeid fishes are already
limited by UV-B.

-------
                                    12-20
                                EXHIBIT 12-7


            Percentage of Total Dose Limit Likely to be Reached on Any

           Particular Day; Lethal Doses Accumulated  Only After Dose-Rate

                             Threshold is Exceeded
'£
 -3


 CO
UJ
00
o
0
<
o
LJ
o
cc
LU
Q.
100


 90


 80


 70


 60


 50


 40


 30


 20 h


 10
         2   3  4

          Feb
234

Mar
1234
   Apr

WEEKS
234

May
234

June
   Circles are 4-year weekly means of surface  daily UV-B doses; x's are  these

   doses calculated at 1 m (adapted from Damkaer et al.  1981).

-------
                                    12-21
    Knowing  that shrimp larvae are limited by UV-B already, and could be
 further limited, does not provide us with an assessment of damage.  Data on
 population dynamics and recruitment of shrimp stocks are insufficient to
 estimate the impacts of shortened surface seasons.  It is possible that
 ordinarily the greatest number of successful recruits are normally derived
 from early-season  larvae and that increased UV-B irradiance would then have no
 appreciable  direct effect on shrimp population stability.  It is also possible
 that late-season larvae are essential to maintaining local populations.  In
 fact, it is  possible that the situation varies from year to year with other
 environmental stresses, so that having a long season is critical to sustaining
 the population in  the long term.  Thus, ozone reductions may lead to no
 change, to a loss  in productivity, or to extinction of these shrimp in
 particular locations.

    It is clear that increases in UV radiation through depletion of ozone
 would create an additional stress to the seasonal near-surface stages of these
 species.  It is quite possible that natural UV levels have already had a
 selective role in  the seasonal adaptation of these species.  The question
 remains:  how much of their surface-living time could these species forego and
 still maintain survival-level populations?

 Avoidance and Its  Implications In Assessing Risk

    Investigations of the biological effects of UV-B radiation on aquatic
 organisms have included a variety of approaches, ranging from histological
 analyses to  simple survival studies.  Most of this work has been concerned
 with the direct effects of natural or enhanced UV-B radiation.  The assumption
 implicit in  some of the conclusions is that the organisms that live in the
 near-surface layer will be passive recipients of ambient UV-B radiation,
 regardless of its  irradiance level.

    Because  some of these near-surface organisms appear to be living near
 their UV tolerance limits (Damkaer et al. 1980) it is possible that they have
 evolved a sensitivity to fluctuations in UV-B radiation.  In addition to
 photorepair  mechanisms, they may possess an ability to avoid harmful levels of
 UV-B radiation by  appropriate defensive behavior.  In moderately productive
 ocean waters, the  DNA-weighted irradiance at 1 m is only about 40% incident on
 the surface  (Smith and Baker 1979).  Under these conditions, the capacity to
 sense harmful UV-B radiation and to avoid it by simply swimming or sinking a
 meter below  the surface would provide considerable protection.

    However,  even  if sensory mechanisms exist to warn an organism of dangerous
 UV-B radiation, the particular stimulus that evokes the avoidance response may
 not be effective under a new spectral distribution caused by ozone depletion.
A reduction  in the ozone layer would not increase UV-A and visible radiation.
 If the avoidance response is linked to high levels of visible irradiance, for
 example,  the avoidance response would not be activated by increased depletion.
 It seems likely that only a sensory system that directly detects irradiance in
 the UV-B range and does so before irreparable damage occurs will afford
 surface-living organisms a safety potential.

    Experiments by Damkaer and Dey (1982) suggest that shrimp possess no organ
 to sense UV-B.  They observed differences in behavior between irradiated and
non-irradiated shrimp larvae (Pandalus platyceros) or copepods (Epilabidocera

-------
                                    12-22
longipedata).   The irradiated specimens maintained their near-surface
positions in visible light (as did the controls)  until the time of decreased
activity and death at about four days.   Adult euphausiids, on the other hand,
seemed to avoid this same level of visible light  (with or without UV-B
radiation) but were attracted to low levels of visible light with or without
extremely high levels of UV-B radiation (Damkaer  and Dey 1983).   These
investigators also found that the zoea stage of the shore crab (Hemigrapsus
nudus) is extremely attracted to strong visible light.  No difference  in
response-time (seconds) was noted between control larvae and larvae receiving
lethal UV-B doses until decreased activity and death of the irradiated larvae
at about 10 days.  There was no apparent reluctance of the crab larvae to swim
toward and to hold themselves within a zone producing a lethal dose of UV-B
radiation.

    These experiments suggested that there were no differences in behavior
between UV-B irradiated organisms and the control organisms until lethal doses
of radiation reduced their activity.  This is not considered to be an active
response to UV radiation, and it would have no value in avoidance.  The
apparent inability to perceive potential danger from UV-B exposure occurred at
dose-rates well above established laboratory thresholds and at exposures to
total doses that, in most experiments,  surpassed lethal total dose thresholds
for similar animals (Damkaer et al. 1980).  If the animals tested in these
studies pose a behavioral mechanism for protection from dangerous levels of
irradiation, it seems unlikely that it could be based on the direct sensing of
UV-B radiation.  The zooplankton tested seemed to be attracted to wavelengths
of radiation longer than those in the UV-B range, and the additional exposure
to high levels of UV-B irradiance did not alter their short-term behavior.
Within the limits examined, these animals generally positioned themselves as
near the light source as possible.  That organisms continue to seek out a
strong light source even while doomed from past UV-B exposures demonstrates
not only the strength of this photo-positive response but also, probably,
their inability to independently discriminate between safe and dangerous
levels of UV-B irradiance.  Since reductions in atmospheric ozone would allow
more incident UV-B (and create no change in visible irradiation), the
currently evolved response of some organisms to future solar irradiation is
unlikely to be optimal.

Effects of UV-B on Biological Diversity and Subsequent Impact on  Food Web

    Almost all life ultimately depends on plants, and in  the ocean most plant
life is planktonic (phytoplankton).  To function, plants  must receive
sunlight, and plants have evolved  to require (and tolerate) more  or less
sunlight, depending on the species.  For plants, their balance with current
levels of solar radiation is therefore critical.  Ultimately, animal  life
exists because it obtains fixed chemical energy  from  plants.  The interaction
of life forms in obtaining energy makes them interdependent,  forming  food
chains, or, more frequently, complex food webs.  Animals  feeding  directly on
plants also have evolved receptor  mechanisms for solar  light.  These
adaptations typically  involved movements and secondary  responses  to visually
feeding predators.

    Even  if UV-B caused an adverse  effect  in only  a  few marine organisms,
others are  likely to be affected.   There have  been  several  studies  indicating
that reduced UV  irradiation  leads  to increased productivity in phytoplankton.

-------
                                    12-23
Only a very few experiments have subjected phytoplankton to reasonable
enhancements of UV radiation.   Some of these experiments (e.g.,  that by
Thomson, Worrest, and Van Dyke 1980) suggest that phytoplankton have
resistance to above-ambient UV levels.   Also,  Worrest (1982) cites a study by
Wolniakowski in which growth rates of a flagellated phytoplankton, initially
depressed by exposure to enhanced UV radiation,  returned to the base rate
after a few days.  Finally, in field observations,  Hobson and Hartley (1983)
found that summer UV levels did not depress phytoplankton photosynthesis.
This is not unexpected since it may have resulted from a combination of
seasonal succession of UV-resistant species and adaptation to increased UV
radiation.

    The most likely direct negative effect of higher UV radiation on
phytoplankton would be a change in community composition (Worrest et al.
1981); primary production may not suffer any decrease.  The species
composition that follows is likely to have differences in size and/or
nutritional value.  The impact of this kind of change, at the lowest level of
the food web, could be marked and severe.  One quite likely effect is a loss
of diversity.  Diversity is associated with stability in ecosystems, allowing
alternate routes or choices within food webs.   In nature, stresses are often
best measured not by changes in productivity or population size, but by
changes in species diversity.   With loss of species, an ecosystem loses some
of its natural resiliency and flexibility.  Other "shocks" to the system
become potentially more damaging.  Unfortunately, our knowledge base is
insufficient to make any definitive assessments.  All that can be done is to
argue by analogy.  Experience has shown that a loss in diversity leads to
instability in ecosystems, which can cascade to create significant change.

EFFECTS ON ICHTHYOPLANKTON (FISHERIES)

    As explained with invertebrate zooplankton, regardless of the
cellular-level responses to UV irradiation, it is usually noted that up  to
some UV level there is no apparent effect on ichthyoplankton.  At greater
doses either the repair systems themselves may become inactivated by the
radiation, or the damage to the general tissues is beyond the capacity of the
repair systems.  To be effective, these threshold levels probably must be
exceeded during several consecutive days.  The apparent UV thresholds are near
current-incident UV levels.  In one experiment the thresholds for all groups
appeared to be above the present median solar incident UV levels  at the  test
location, at least until late in the time span of natural occurrence near the
surface.  An ozone depletion of 20% would significantly shorten this season.
The effects of this shortening on annual productivity is not known.  Whether
or not the population could endure with a drastically reduced time of
near-surface occurrence is not known.  Success of any year-class  depends on
the timing of a great number of other events in addition to levels of exposure
to UV-B radiation.

    Hunter, Kaupp, and Taylor (1982) exposed anchovy eggs and larvae and
mackerel larvae to high doses of UV-B radiation in small closed containers.
They reported that the dose that killed one-half of the test organisms  (LDcQ)
is clearly not an acceptable criterion for biological effect.  They recognized
that "it seems unlikely that any of the damaged survivors,  regardless of

-------
                                    12-24


dosage, would be able to feed successfully."  Hunter,  Kaupp,  and Taylor (1982)
recognized that "none of the larvae surviving LD,-n doses at age 4-days would
survive to age 12-days."

    Apart from the fact that laboratory experiments can never duplicate
natural responses, the subsequent experiments on UV effects by Hunter, Kaupp,
and Taylor (1981, 1982) involved some additional shortcomings that must be
kept in mind while evaluating the results.   First, it is often difficult to
maintain test organisms in good condition.   For some common animals,  it is not
yet possible to keep them alive in the laboratory, much less experiment
meaningfully with them.  For most of the other animals the laboratory
conditions themselves exert a powerful negative influence, so that a true and
isolated measurement of UV stress is not attainable.  This can be seen in the
high 12-day mortalities (average 61%) in the controls for the anchovy larvae.
These larvae were kept in small closed containers, a poor but perhaps
necessary approximation of the open ocean.   There can be no doubt that these
larvae were under significant stress even without UV radiation.  Hunter et al.
refer to wide fluctuations in temperature in the solar-light experiments;
oxygen was probably also a limiting factor.   Perhaps the use of a flow-through
seawater system or another species of fish larvae would eliminate this serious
problem.  As in some agricultural field tests (Teramura 1986), actual field
tolerance of UV radiation is greater than that determined in the laboratory.

    The second confounding factor in the experimental design of Hunter, Kaupp,
and Taylor is that the tests apparently ended at Day 12.  Hunter, Kaupp, and
Taylor (1982) reported that at age 20-days,  anchovy larvae begin daily
vertical migrations.  This larval movement into deeper water during the day,
and into shallow water at night, would coincidentally avoid significant UV-B
exposure.  After 20-days survival, then, the gauntlet of UV exposure would
have been successfully negotiated.  A more meaningful experimental criterion
would have been survival to at least age 20-days.  However, from the  reported
mortality rates, perhaps even the control larvae would not have survived this
long.

    Related to the test endpoint is the third major problem in the
interpretation of these experiments.  When experiments are conducted  to an
appropriate developmental endpoint (e.g., to the age when  larvae avoid  the
daytime near-surface layer), it is usually apparent that below a certain
irradiation level the organisms are not seriously  affected.  On the other
hand, beyond that irradiation level there seems to be fairly general  and
irreversible damage.  In experiments with planktonic invertebrates  (Damkaer et
al. 1980; Damkaer, Dey, and Heron 1981; Damkaer and Dey  1983), it can be  seen
that a very great range of UV-B dose-rates and doses could lead to  the  same
LD   • this depends solely on the time chosen for  the observation.   In every
case, however, at the developmental endpoint, the  groups  that had exhibited an
LDc« were essentially all dead.

    Hunter, Kaupp, and Taylor  (1982) stated  that  none of the  larvae  surviving
the reported LDsn daily dose to age 4-days would  have survived to age 12-days.
The LD,-0 criterion,  then, was clearly inappropriate.  Hunter,  Kaupp,  and
Taylor admit that their LE>   daily dose  to  age  12-days  may also have
underestimated the effect or UV on survival.  In  their  1981  study  they
predicted the effects on natural populations by assuming that  all  organisms
receiving less that  this LD   value will survive,  but  this is  not  an

-------
                                    12-25
appropriate use of the LDcfv-  To the extent that organisms receiving just
under LDcn doses will probably all die,  and that LDcr. values here suggest only
that a lower dose might also lead to death, the lethal daily dose estimated by
Hunter, Kaupp and Taylor could be an underestimate of UV effects.  However, in
a reinterpretation of the data considering threshold dose-rate and dose
effects (Damkaer, personal communication), the critical daily dose can be
estimated.

    Hunter, Kaupp, and Taylor (1981) noted that the lengths of anchovy larvae
in all UV treatments, even with the lowest dose, were significantly less than
the length of control larvae.  Length, then, could be a sensitive criterion
for UV-B effects, and this, and other criteria, should be explored.  A
judgment would be necessary as to what diminished length the individual larvae
could tolerate before there is a significant effect on survival.

    Hunter, Kaupp, and Taylor (1982) realized that information is required on
seasonal abundance and vertical distributions of anchovy larvae, vertical
mixing, and penetration of UV-B radiation into anchovy-populated seawater
before effects of current and increased levels of UV-B radiation can be
predicted.  There are fairly good data on seasonal abundances and vertical
distributions of anchovy larvae, as well as good calculations on average UV-B
penetration into known water types.  To detect a mortality due to UV-B
radiation, against the high natural mortality, would be a difficult problem.
But this is not the same as predicting the effect due to this factor alone.
For this, mortality rates from other causes are not required because we are
concerned only with how UV-B radiation would act on those that survive the
other perils of their existence.

    Hunter, Kaupp, and Taylor assumed that all larvae at depths where UV-B
exceeds the LD^rv daily dose will die.  They also assumed a static, unmixed
vertical distribution.  Finally, these authors provided data that indicated
that anchovy larvae off southern California are typically centered,in
moderately productive ocean waters that have about 0.5 mg CLl-a m   (surface
chlorophyll concentration).  Baker, Smith, and Green  (1981) calculated surface
DNA-effective UV-B irradiance levels expected for this area, and Smith and
Baker (1979) calculated the penetration of UV-B radiation into moderately
productive ocean water.

    While Hunter, Kaupp, and Taylor calculated that 13.4% of the annual larval
anchovy population would die (presumably after 12-day exposures during certain
critical times) in a static situation, an analysis by Damkaer (personal
communication) indicates that somewhat more would die (14.7%) and in shorter
times (4 to 10 days), if there were no vertical mixing.  However, static
situations are not probable because of physical vertical mixing of water
(through the action of wind, density currents, and tides) and the active and
passive movement of fish larvae.  If some vertical mixing did not offer relief
from the effects of UV-B radiation, it is likely that there would be no
anchovy larvae above a few meters depth; this is clearly not the case.
Probably the surface mixed layer extends to at least  25 m from December to
February (Sverdrup, Johnson, and Fleming 1942).  A mixed layer of at least
10 m can be assumed for the rest of the year.

    For the mixed situation in February, the larvae would spend  less than  one
hour each day in the surface 1-m layer,  resulting in  exposures where the

-------
                                    12-26
threshold doses would never be reached in the larval stage.   In the February
model, with mixing to 25 m, there would have to be a six-fold increase in
incident UV-B radiation before threshold doses would be reached.   For the
months March-October, with mixing to 10 m, the threshold doses are also not
reached within the 20 days available before the larvae become vertically
migrating.  Only in April and August, with the highest surface irradiances,
are these doses approached within this available time (about 22 days).

    For March-October, with 10-m mixing, a 10% increase in incident UV-B
radiation would still not lead to threshold doses in less than 20 days.  With
a 20% increase in incident UV-B radiation (a result of about a 9% decrease in
the atmospheric ozone column) the depth of the threshold dose-rate is
increased.  With the 20% increase in incident irradiance, in the
dose/dose-rate threshold model, all of the anchovy larvae within the 10-m
mixed layer in April and August would be killed, the threshold doses being
achieved after 15 days.

    Apparently, at all months about 36% of the larval anchovy population is
above 10 m (Hunter, Kaupp, and Taylor 1981).  The greatest numbers of anchovy
larvae are found in April  (20% of annual population), so that 7.2% of the
annual population would be eliminated with a 9% ozone decrease.  Only 1.9% of
the annual population is present in August, so the loss of those above 10 m
would amount to 0.7% of the annual population.  The total predicted loss,
then, due to a 9% decrease in total ozone column, would be about 8% of the
larval anchovy population  (Exhibit 12-8).

    In the static model of Hunter, Kaupp, and Taylor, 13.4% of the larval
anchovy population dies through the current effects of solar UV-B radiation.
They predict only an additional 4.2% decrease in population if UV-B  levels
increase 60+%.  Presumably, in that model, a 20% increase in UV-B irradiance
would lead to only a 1.4%  additional loss, so the total  loss under a  20%
increase would be 14.8%.   In contrast,  the dose/dose-rate threshold  and
vertical mixing model presented by Damkaer would predict no current  solar UV-B
damage, but an 8% total larval population loss at a  20%  increase  in  UV-B
irradiance (10% depletion) (Exhibit  12-8) .  Because  of complex interactions
between mixing-depths, vertical distribution of larval anchovy,  seasonal
changes in solar irradiance and the penetration of UV  into  seawater,  and
seasonal change in anchovy abundance,  there  is not a linear relationship
between mixing-depth and predicted annual loss  of anchovy larvae.  However,
within the range of values represented in Exhibit 12-8,  there  is  a maximum
predicted loss at current  UV levels with  a  15-m mixed  layer,  and a maximum
predicted loss at a  60% increase  in UV-B  irradiance  with a  10-m  mixed, layer.

    Perhaps changes  in  lower levels  of the  food web,  if  they  occur with higher
UV-B, are likely to be more  important  than  the  direct  effects  upon the
fisheries.  Changes  in  the composition of primary production  of  organic
biomass could  alter  the mortality experienced by  larval  fish,  and perhaps
there would be a synergistic effect  on mortality.   Some  fish  die from direct
exposure, some die from lack of  food,  and some  die  from  the combination of a
reduction in food and the  weakening  from exposure  (or are weakened and become
outcompeted by other  fauna for limited food reserves).

-------
do
pi
to
o
3
PJ
rt
p)

o
It)
rt
(D
P>


€
T)
§
O
i-l
vO
00
ro
                    PREDICTED LOSS OF

                 ANNUAL POPULATION  (%)
                                                         H
                                                         W
H
ro

oo
                                                             ISJ
                                                             i
                                                             to

-------
                                    12-28
    The impact on marine fisheries as a food supply to humans would be
significant if the species of phyto- and zooplankton that adapted to enhanced
UV-B radiation were of different nutritional value (i.e., if they altered
growth and fecundity of the consumers or different accessibility to humans).
If the indirect impact of suppression upon consumers were linear, a 5%
reduction of primary production would result in a 5% reduction in fish
production. (Exhibit 12-2).  A question still under investigation is whether
the trophodynamic relationships might be nonlinear.  For example, there may  be
an amplification factor involved that results in a relatively greater impact
at higher trophic levels than at the primary-producer level.

    Levels of UV-B irradiance range latitudinally, with the highest exposures
in the tropics.   The current difference between the extremes of exposures is
about 3- to 6-fold, but biota are presently adapted to the levels that are
normally experienced at their current locations.  An effect of ozone depletion
of ecological significance is that a 10% decrease in the ozone concentration
would produce a 25% increase of biologically effective ultraviolet radiation
(DNA, Plant; Exhibit 12-2), which would correspond to migrating over 30°
toward the equator.  Continued investigations concerning the range of natural
ecological uncertainties,  which are much larger than the uncertainties in the
particular photobiological effects, will be required to permit assessment of
the possible consequences for the many complex ecological interactions, as
well as for the productivity of fisheries.

CONCLUSIONS

    At a minimum, evidence exists that both small and large changes in the
stratospheric ozone layer would affect aquatic ecosystems.  It will be very
difficult to determine the biological effects of small changes in UV-B
irradiance at the ocean surface because aquatic ecosystems have a huge
physical and biological "noise" level.  Storms, clouds, and global currents
make dramatic changes in the status of life in the ocean, which cannot at
present be predicted.  One should not, however, conclude that changes that are
lost in the "noise" of the system are automatically insignificant simply
because we cannot measure and define exactly what is taking place.  Changes
that occur because of small systemic events, such as higher UV-B, could
accumulate in time to produce much more significant change  in aquatic
organisms that appear very sensitive to current UV-B radiation.  The
possibility exists that changes outside the historical range of UV-B could
have implications far greater than we are currently able to predict with
confidence.   We ought not to mistake our ignorance about what will happen if
UV-B increases for a conclusion that all is well.  At this  time we can place
almost no limit on what will happen.  An increase in UV-B might have small
effects, or it might be much more significant.

-------
                                    12-29
REFERENCES

Baker K.S. (1979).  Calculation of biologically effective UV-B radiation in
aquatic systems.  SIO Ref 79-1, Scripps Institution of Oceanography.   46 p.

Baker K.S., Smith, R.C.,  Green, A.E.S. (1981).  Middle ultraviolet radiation
reaching the ocean surface.  Photochem Photobiol 32:367-374.

Banse K. (1982).  Cell volumes, maximal growth rates of unicellular algae and
ciliates, and the role of ciliates in the marine pelagial.  Limnol Oceanogr
27:1059-1071.

Bell G.M., Hoar, W.S. (1950).  Some effects of ultraviolet radiation on
sockeye salmon eggs and alevins.  Can J Res 28:35-43.

Caldwell, M.M. (1968).  Solar ultraviolet radiation as an ecological factor
for alpine plants.  Ecol Monogr 38:243-268.

Caldwell, M.M. (1971).  Solar UV irradiation and the growth and development of
higher plants.  Photophysiology 6:131-177.

Caldwell, M.M.,  Nachtwey, D.S. (1975).  Introduction and overview.  In:
Nachtwey, D.S.,  Caldwell, M.M., Biggs, R.H. (eds.)  Impacts of climatic change
on the biosphere, part I - ultraviolet radiation effects, pp  (1)3-(1)30, US
Dept Transport,  CIAP Monogr 5, NTIS, Springfield, Virginia.

Caldwell, M.M.,  Robberecht, R., Billings, W.D.  (1980).  A steep latitudinal
gradient of solar ultraviolet-B radiation in the arctic-alpine life zone.
Ecology 61:600-611.

Calkins, J. (1982a).  Modeling light loss versus UV-B increase for organisms
which control their vertical position in the water column.  In: Calkins, J.
(ed.) The role of solar ultraviolet radiation  in marine ecosystems, pp
539-542, Plenum, New York.

Calkins, J. (1982b).  Some considerations on  the ecological and evolutionary
effects of solar UV.  In: Calkins, J. (ed.)  The role of  solar ultraviolet
radiation in marine ecosystems, pp 685-689, Plenum, New York.

Cheng, L., Douek, M., Goring,  D.  (1978).  UV  absorption by gerrid cuticles.
Limnol Oceanogr 23:554-556.

Cote, B., Platt, T.  (1983).  Day-to-day variation  in  the  spring-summer
photosynthetic parameters of coastal marine phytoplankton.  Limnol Oceanogr
28:320-344.

Damkaer, D.M. (1982).  Possible influence of  solar UV radiation in the
evolution of marine zooplankton.   In: Calkins,  J.  (ed.)   The  role of  solar
ultraviolet radiation  in marine ecosystems, pp 701-706,  Plenum, New York.

Damkaer, D.M., Dey, D.B., and  Heron,  G.A.  (1981).  Dose/Dose-Rate Responses  of
Shrimp Larvae to UV-B Radiation.  Ecologia  (BRL)  48:178-182.

-------
                                    12-30


Damkaer, D.M., Dey, D.B. (1982).   Short-term responses of some planktonic
Crustacea exposed to enhanced UV-B radiation.   In:  Calkins,  J. (ed.)   The  role
of solar ultraviolet radiation in marine ecosystems,  pp 417-427,  Plenum, New
York.

Damkaer, D.M., Dey, D.B. (1983).   UV damage and photoreactivation potentials
of larval shrimp, Pandalus platyceros.  and adult euphausiids,  Thvsanoessa
raschii.  Oecologia 48:178-182.

Damkaer, D.M., Dey, D.B.,  Heron,  G.A.,  Prentice, E.F.  (1980).   Effects of  UV-B
radiation on near-surface zooplankton of Puget Sound.   Oecologia 44:149-158.

Doehler, G. 1984.  Effect of UV-B radiation on biomass production,
pigmentation and protein content of marine diatoms.  Z Naturf 39:634-638.

Doehler, G. 1985.  Effect of UV-B radiation (290-320 run) on the nitrogen
metabolism of several marine diatoms.  J Plant Physiol 118:391-400.

Dunbar, C.E. (1959).  Sunburn in fingerling rainbow-trout.  Progr Fish Cult
21:74.

Forsythe, W.E. (1954).  Smithsonian Physical Tables.   Smithsonian Institution
Press, Washington, D.C.  pp 721-723.

Gerstl, S.A.W., Zardecki,  A., Wiser, H.L. (1986).   A new UV-B Handbook, Vol.
1.  In: Worrest,  R.C., Caldwell,  M.M. (eds.)  Stratospheric Ozone Reduction,
Solar Ultraviolet Radiation and Plant Life, pp 87-111, Springer-Verlag, New
York, Heidelberg, Berlin.

Giese, A.C. (1938).  The effects of ultra-violet radiations of various
wave-lengths upon cleavage of sea urchin eggs.  Biol Bull 75:238-247.

Giese, A.C. (1976).  Living with Our Sun's Ultraviolet Rays.  Plenum Press,
New York. 185 pp.

Green, A.E.S., Cross, K.R.,  Smith, L.A. (1980).  Improved analytic
characterization of ultraviolet skylight.  Photochem Photobiol 31:59-65.

Haeder, D.P. (1986).  The effect of enhanced UV-B radiation on motile
microorganisms.  In: Worrest, R.C., Caldwell, M.M.  (eds.)  Stratospheric Ozone
Reduction, Solar Ultraviolet Radiation and Plant Life, pp 223-233,
Springer-Verlag,  New York, Heidelberg,  Berlin.

Hairston, N. (1980)  The vertical distribution of diaptomid copepods  in
relation to body pigmentation.  In: Kerfoot, W.  (ed.)  Evolution  and  Ecology
of Zooplankton Communities,  pp 98-110,  Univ Press New England, Hanover, New
Hampshire.

Harvey, J.M. (1930).  The action of light on Calanus finmarchicus (Gunner)  as
determined by its effect on the heart rate.  Contrib Can  Biol 5:85-92.

Hobson, L., Hartley, F. (1983).  Ultraviolet irradiance and primary production
in a Vancouver Island fjord, British Columbia, Canada.  J Plankton Res
5:325-331.

-------
                                    12-31
Hunter, J.R., Kaupp, S.E., Taylor, J.H. (1981).  Effects of solar and
artificial ultraviolet-B radiation on larval northern anchovy, Engraulis
mordax.  Photochem Photobiol 34:477-486.

Hunter, J.R., Kaupp, S.E., Taylor, J.H. (1982).  Assessment of effects of UV
radiation on marine fish larvae.  In: Calkins, J. (ed.)  The role of solar
ultraviolet radiation in marine ecosystems, pp 459-497, Plenum, New York.

Huntsman, A.G.  (1925).  Limiting factors for marine animals.  I. The lethal
effect of sunlight.  Contrib Can Biol 2:83-88.

Jitts, H.R., Morel, A., and Saijo, Y. (1976).  The relation of oceanic primary
production to available photosynthetic irridantsm Aust J Mar Freshwater Res
27:441-454.

Jokiel, P.L., York, R.H., Jr. (1984).  Importance of ultraviolet radiation in
photoinhibition of microbial growth.  Limnol Oceanog R 29:192-199.

Jones, L.W., Kok, B. (1966).  Photoinhibition of chloroplast reactions.  I.
Kinetics and action spectra.  Plant Physiol 41:1037-1043.

Karanas, J.J., Van Dyke, H., Worrest, R.C. (1979).  Midultraviolet (UV-B)
sensitivity of Acartia clausii Giesbrecht (Copepoda).   Limnol Oceanogr
24:1104-1116.

Karanas, J.J., Worrest, R.C., Van Dyke, H. (1981).  Impact of UV-B radiation
on the fecundity of the copepod Acartia clausii.  Mar Biol 65:125-133.

Kaupp, S.E., Hunter, J.R. (1981).  Photorepair in larval anchovy, Engraulis
mordax.  Photochem Photobiol 33:253-256.

Kelly, J.R. (1986).  How might enhanced levels of solar UV-B radiation affect
marine ecosystems?  In: Proceedings of EPA/UNEP International Conference on
Health and Environmental Effects of Ozone Modification and Climate Change,
July, 1986.

Klugh, A.B. (1929).  The effect of the ultra-violet component of sunlight on
certain marine organisms.  Can J Res 1:100-109.

Klugh, A.B. (1930).  The effect of ultra-violet component of the sun's
radiation upon some aquatic organisms.  Can J Res 2:312-317.

Lasker, R.  (1975).  Field criteria for survival of anchovy larvae:  the
relation between inshore chlorophyll maximum layers and successful first
feeding.  Fishery Bull 73:453-462.

Marinaro, J.Y., Bernard, M. (1966).  Contribution a 1'etude des oeufs et
larves pelagiques de poissons mediterraneens.  I. Note preliminarie sur
1'influence lethale du rayonnement solaire sur les oeufs.  Pelagos 6:49-55.

Murphy, T.  (1975).  Effects of UV radiation no nucleic acids.   In:  Nachtwey,
D.S., Caldwell, M.M.,  Biggs, R.H. (eds.)  Impacts of climatic change on  the
biosphere,  part I - ultraviolet radiation effects, pp  (3)21-(3)44, US Dept
Transport,  CIAP Monogr 5, NTIS, Springfield, Virginia.

-------
                                    12-32
Nachtwey, D.S.,  Rundel,  R.D.  (1982).   Ozone change:   biological effects.   In:
Bower, F.A.,  Ward,  R.B.  (eds.)   Stratospheric Ozone  and Man,  Vol 2,  pp.
81-121, CRC Press,  West Palm Beach.

NAS.  National Academy of Science (1984).   Causes and Effects of Changes  in
Stratospheric Ozone:   Update 1983.   Natl Acad Sci, Washington,  D.C.   254  pp.

Nielsen^ E.S. (1964).   On a complication in Maine productivity work due to the
influence of ultraviolet light,  J Cons Cons Int Explor Mer 20:130-135.

Pommeranz, T. (1974).   Resistance of plaice eggs to  mechanical stress and
light.  In:  Blaxter,  J.H.S.  (ed.)   The Early Life History of Fish,  pp
397-416, Springer-Verlag, New York,  Heidelberg, Berlin.

Ringelberg, J.,  Keyser,  A., Flik, B.  (1984).  The mortality effect of
ultraviolet radiation in a translucent and in a red  morph of Acanthodiaptomus
denticornis (Crustacea,  Copepoda) and its possible ecological relevance.
Hydrobiologia 112:217-222.

Russell-Hunter,  W.D.  (1970).   Aquatic Productivity.   MacMillan.

Ryther, J.H.  (1969).   Photosynthesis and fish production in the sea.  Science
166:72-76.

Setlow, R.B.  (1974).   The wavelengths in sunlight effective in producing skin
cancer:  a theoretical analysis.  Proc Natl Acad Sci (USA) 71:3363-3366.

Siebeck, 0, (1978).  Ultraviolet tolerance of planktonic crustaceans.  Verh
Intern Verein Limnol 20:2469-2473.

Smith, R.C.,  Baker, K.S.  (1978).  Optical classification of natural waters.
Limnol Oceanogr 23:260-267.

Smith, R.C.,  Baker, K.S.  (1979).  Penetration of UV-B and biologically
effective dose-rates in natural waters.  Photochem Photobiol 29:311-323.
Smith, R.C., Baker, K.S. (1980).  Stratospheric ozone, middle ultraviolet
radiation and carbon-14 measurements of marine productivity.  Science
208:592-593.

Smith, R.C., Baker, K.S., Holm-Hanson, 0., Olson, R.  (1980).  Photoinhibition
of photosynthesis in natural waters.  Photochem Photobiol 31:585-592.

Sverdrup, H.U.,  Johnson, M.W., Fleming, R.H.  (1942).  The Oceans, Their
Physics, Chemistry and General Biology.  Prentice-Hall, New York.  1087 p.

Teramura, A.H. (1986).  Interaction between UV-B radiation and other stresses
in plants.  In:   Worrest, R.C., Caldwell, M.M. (eds.)  Stratospheric Ozone
Reduction, Solar Ultraviolet Radiation and Plant Life, pp 327-343,
Springer-Verlag, New York, Heidelberg, Berlin.

-------
                                    12-33
Thomson, B.E., Worrest, R.C.,  Van Dyke, H. (1980).   The growth response of an
estuarine diatom (Melosira nummuloides cDillw.!  Ag.) to UV-B (290-320 nm)
radiation.  Estuaries 3:69-72.

Worrest, R.C. (1982).  Review of literature concerning the impact of UV-B
radiation upon marine organisms.  In:  Calkins,  J.  (ed.)  The role of solar
ultraviolet radiation in marine ecosystems, pp 429-457, Plenum, New York.

Worrest, R.C. (1983).  Depletion of stratospheric ozone:  impact of UV-B
radiation upon nonhuman organisms.  In:  Covello, V., et al. (eds.)  The
Analysis of Actual Versus Perceived Risks, pp 303-315, Plenum Press, New York.

Worrest, R.C., Wolniakowski, K.U., Scott, J.D.,  Brooker, D.L., Thomson, B.E.,
Van Dyke, H. (1981).  Sensitivity of marine phytoplankton to UV-B radiation:
impact upon a model ecosystem.  Photochem Photobiol 33:223-227.

ZoBell, C.E., McEwen, G.F. (1935).  The lethal action of sunlight upon
bacteria in seawater.  Biol Bull 68:93-106.

-------
                                   CHAPTER 13

                          EFFECTS OF UV-B ON POLYMERS
SUMMARY

    Several polymers (and largely the impurities present in the polymers)
readily absorb the ultraviolet (UV) radiation.  One effect of increased UV-B
radiation, resulting from possible future depletion of the ozone layer in the
stratosphere, will be accelerated degradation of the polymeric materials.
Energies associated with the UV radiation are large enough to initiate reactions
in polymers, which lead to their degradation by affecting their mechanical and
optical properties and thus reducing their service life.  Several methods are
used to stabilize the polymers to maintain a useful service life in their
various important applications.

    While much is known about the physical processes of degradation, little
research has focused to date on the potential costs and responses to such
degradation.  Initial studies suggest that one likely response to increased UV-B
induced damage is to increase the amount of light stabilizer in the polymers.
Because of the lack of relevant data, only approximate estimation methods are
available to determine the extent of light-induced damage and the degree of
stabilization required to minimize it.   The effect of increased UV-B radiation
is manifested in increased costs of production, including raw material costs,
energy costs, and maintenance costs.

-------
                                     13-2


FINDINGS

1.  INCREASED UV-B RADIATION WILL ACCELERATE THE DEGRADATION OF POLYMERS.

    la.  Several commercial polymers (e.g.,  polyethylene,  polypropylene,
         poly(vinylchloride)),  although theoretically UV transparent,  contain
         chromophore impurities that absorb  light in the UV-B region of the
         spectrum.  Other polymers (e.g.,  polycarbonate) have structural
         features in their molecules that  result in strong UV-B light
         absorption.

    Ib.  Several polymers have  important outdoor applications (e.g., used in
         siding and window glazing in the  building industry, in film and
         containers in packaging,  in housewares and toys,  and in paints and
         protective coatings).   Such polymers are likely to be exposed to
         significant amounts of UV-B radiation.  Other polymers are stored
         outside before use and could deteriorate during these periods.

    Ic.  Absorption of UV-B radiation in polymers causes photo-induced reactions
         and alters important mechanical,  physical,  or optical properties of the
         polymers (e.g.,  yellowing,  brittleness) and thus degrades (i.e.,
         reduces the useful life of) the polymers.

2.   INCREASED USE OF UV-STABILIZERS FOR PROTECTION OF POLYMERS AGAINST UV
     RADIATION WOULD HAVE NEGATIVE EFFECTS.

     2a.   Increased amounts of stabilizers  might adversely affect the
           processing and use properties of  some polymers (e.g., hardness,
           thermal conductivity, flow characteristics).  For example,  increased
           amounts of titanium  dioxide in  poly(vinylchloride) might affect its
           processing properties,  increasing its costs of production.

     2b.   Changes in the amount of stabilizer (and other additives) would
           increase costs of products.   Alternatively, manufacturers could
           develop new formulations to avoid or minimize impurities in
           production.

     2c.   The addition of stabilizers to  polymers may be limited by practical
           problems of material characteristics or manufacture.  However,  other
           responses, may be possible to limit damage.

3.   INCREASED UV-B RADIATION DUE TO OZONE DEPLETION COULD HAVE ADVERSE ECONOMIC
     EFFECTS.

     3a.   Changes in polymer processing properties can result in more equipment
           shutdowns, higher maintenance costs, and increased utility costs.

     3b.   Increased operating costs and material costs (e.g., for stabilizers,
           lubricants, and other additives)  would have an adverse economic
           impact on the polymer/plastic and related industries.

     3c.   In a case study using preliminary data and methods, and a given
           scenario of ozone depletion (26%  depletion by 2075), undiscounted

-------
                                     13-3
           cumulative (1984-2075) economic damage for poly(vinylchloride) is
           estimated at $4.7 billion (USA only).   Due to the lack of data,
           possible damage to other polymers has  not been assessed.

4.   POTENTIAL DAMAGES TO POLYMERS RELATED TO OZONE DEPLETION AND CLIMATE CHANGE
     ARE DIFFICULT TO ESTIMATE.

     4a. "  Due to lack of relevant experimental data, only approximate
           estimation methods are available to determine the potential extent of
           light-induced damage to polymers and other materials.

     4b.   Depending upon the chemical nature of a polymer, the components of
           the compound, and the weathering factors, both temperature and
           humidity tend to increase the rate of degradation.

     4c.   Research on dose-response relationships for polymers could increase
           our ability to project the effects of ozone depletion.

     4d.   Actual action spectra need to be developed for different polymers.

     4e.   The feasibility of different mitigation measures needs to be
           experimentally determined.

     4f.   The synergistic effects of increased humidity and temperature need to
           be considered.

-------
                                     13-4
PHOTODEGRADATION OF POLYMERS

Structure of Polymers

    Macromolecules in the polymers are composed of several monomeric units
joined by chemical bonds to each other.   The monomeric units contain chemical
bonds that are either in the main chain of the macromolecule or connect various
atoms or side groups to it.  Side groups,  if present,  contain additional
chemical bonds.   All of these bonds may be reaction sites in polymer
degradation, and various energy sources may be effective in supplying the energy
necessary to break the bonds.  The dissociation energies of chemical bonds in
common polymers range from about 65 kcal/mole (C-C1) to 108 kcal/mole (C-F) with
carbon-carbon bonds in the middle (75-85 kcal/mole).  The most important types
of energy that cause polymer degradation are heat, mechanical energy, and
radiation.  Thermal and mechanical degradation of polymers may occur during
thermomechanical processing.  The most common form of radiant energy that causes
degradation is that of the UV component of sunlight; the energy of a photon with
wavelength  A = 300 nm is about 95 kcal/mole, which is higher than most bond
dissociation energies in polymers, and therefore capable of breaking these
bonds.

    Wavelengths of UV radiation at which various commercial polymers have
maximum sensitivity and their corresponding photon energies are given in Exhibit
13-1.

                                  EXHIBIT  13-1

             Wavelengths of UV Radiation and Polymers With Maximum
                 Sensitivity and Corresponding Photon Energies

Polymer
Styrene-acrylonitrile copolymer
Polycarbonate
Polyethylene
Polystyrene
Polyvinyl chloride
Polyester
Vinyl chloride -vinyl acetate copolymer
Polypropylene
Wavelength
(nm)
290,325
295,345
300
318
320
325
327,364
370
Energy
kcal/mole
99,88
97,83
96
90
89
88
87,79
77
        Source:  Kelen, T., Polymer Degradation. Van Nostrand Reinhold,
                 New York, 1983.

-------
                                      13-5
Light Absorption

    A prerequisite for light-induced damage in materials is their ability to
absorb the radiation (this observation is sometimes referred to as the
Grotthus-Draper Law of Photochemistry) (Andrady 1986).  Very few materials
absorb light at all wavelengths across the spectrum.  Pure samples of
polyethylene, polypropylene, and polyvinyl chloride, for instance, are
theoretically UV transparent and are not expected to absorb light in the range
290-315 nm, the UV-B region of the spectrum.  However, completely "pure"
polymers do not exist.   During synthesis, processing, and storage, various
amounts of carbonyl and hydroperoxy groups accumulate in the commercial
polymers.  These chromophores readily absorb UV light in the 290-315 nm
wavelengths.  Initiation of degradation consists mainly of the decomposition of
these chromophores.

Photophysical and Photochemical Processes

    Degradation of the polymers involves a series of photophysical and
photochemical processes.   The physical processes involved in photodegradation
include absorption of light by the material, electronic excitation of the
molecules, and deactivation energy transfer to some acceptor.   When the
lifetime of the excited state is sufficiently long, the species can participate
in various chemical transformations.  This is especially true when the
photoinduced degradation reactions are carried out in the presence of oxygen
(see equations 1-4).


                      light
        Polymer (RH)  	> polymer excited state (RH*)                   (1)

            degradation
        RH* 	> polymer radicals or broken (scission)
                          chain end (R- )                                   (2)

        R- + air 	> polymer peroxy radicals (R02•)                    (3)

                  light
        R02- + RH 	> ROOH + R- (further polymer degradation)          (4)


    As mentioned earlier,  hydroperoxy and carbonyl groups are the two most
common impurities present in the polymers.   In the case of hydroperoxy groups
(ROOH), the energy of UV light is sufficient to cause both of the following
decompositions (Kelen 1983):

                     light
                ROOH 	> RO- + -H  (-42 kcal/mole)                     (5)

                     light
                ROOH  	> R-   + -OH (-70 kcal/mole)                     (6)

-------
                                      13-6
Dissociation of the 0-H bond is less facile:

                     light
                ROOH - > ROD- + -H  (-90 kcal/mole)                     (7)


    Because of the low bond dissociation energy  (42 kcal/mole) , decomposition
according to reaction (5) is predominant in photo -oxidation of polymers
containing hydroperoxy groups .

    The carbonyl group absorbs UV light readily  and hence is easily excited.
The excited carbonyl group decomposes via Norrish reactions of Types  I,  II, and
III (Kelen 1983).  The Norrish-I reaction (8) is a radical cleavage of the  bond
between the carbonyl group and the oc-carbon atom (oc-scission) .

                             0                   0
                        -CHL-C-CH - - >  -CH9-C- +  -CH2                  (8)
                                              I
                                              CH2'   +  CO

    The Norrish- II reaction (9) is a nonradical scission  that  occurs  through
the formation of a six-membered cyclic intermediate to form  an olefin and an
alcohol or ketone.

                    0-     -H                   OH
                  /      \      light     /
                 -C          CH-   - > -c        +        CH-           (9)

                     CHj -CH2                   ctiz          -CH2

    The Norrish-III reaction  (10) is also a  nonradical chain scission and leads
to the formation of an olefin and an aldehyde.
                        O CH           0     CH9
                        III            II    II
                        -C-CH- 	>  -CH + CH-                             (10)

    In the case of PVC, the photolytic scission  takes  place  as  follows:
                          light
                 -CH  -CH-  	>  -CH  -CH-  +  -Cl                             (11)
                  I
                  Cl

-------
                                      13-7
POLYMERS IN OUTDOOR USES AND POTENTIAL DEGRADATION

    A variety of polymeric materials are used in applications where the
material is routinely exposed to solar radiation.   These include mainly the use
of plastics in the building industry and in paints/coatings.   In other
applications such as packaging,  outdoor furnishings,  housewares, and toys,
exposure occurs from extended to more limited,  possibly intermittent,  periods
of time^.

    According to recent estimates, of the 45,506 million pounds of U.S. plastic
sales in 1985, the building industry above accounted for 9,428 million pounds,
21% of the total consumption.  Packaging applications were the largest market
consuming 10,846 million pounds (Horst et al. 1986).

    However, the bulk of the consumption is represented by a limited number of
different plastics.  These include the so-called "commodity" plastics and a few
other types selected on the basis of their unique physical-chemical properties.
Exhibit 13-2 gives a breakdown of the types of plastics used in each
application along with the likelihood of exposure to sunlight in normal use.
Some plastics/products, not experiencing direct sunlight, can still be exposed
to UV light since the UV light is easily reflected and scattered by terrain and
buildings, resulting in significant irradiation levels at surfaces illuminated
only by diffused or reflected light (Andrady 1986).  Thus, vulnerability to
degradation is a function of a polymer's normal lifetime and the percent of
time it is exposed to direct or indirect sunlight.

Light-Induced Damage

    Exposure to light and several physical-chemical processes (described
earlier) occurring simultaneously result in deterioration of polymeric
materials.

    Properties most likely to be affected by the degradation of the polymer
include both mechanical and optical properties.   Mechanical properties such as
tensile strength, elongation, modulus, and impact strength can all be affected
by UV degradation of polymeric materials.  Optical properties that have been
observed to change upon exposure to UV radiation include transparency, color,
chalking or cracking of surfaces, and yellowing.  These effects are most
noticeable on the surface of plastic materials because of the limited
penetration of UV light through the materials.   Thus, the material may show
cosmetic changes in appearance long before any degradation in its bulk
mechanical property is noted.

    The mechanical weakening of plastic material occurs because of a reduction
in the molecular weight of the polymer.  Essentially all of the mechanical
properties of polymers are related to the molecular weight of the material.
Thus, for every bond in the polymer molecule that breaks because of UV
radiation, the average molecular weight will be reduced, and the corresponding
mechanical properties will be reduced in some proportion to this molecular
weight reduction (Andrady 1986) .

-------
                                     13-8
                                 EXHIBIT 13-2

         Plastics Used in Applications Where Exposure of the Material
                         to Sunlight Might be Expected
Type of Plastic

Polyvinyl chloride
(rigid)

Polyvinyl chloride
(plasticized)

Unsaturated
Polyester
Polycarbonate
Acrylic

Polyethylene
Polypropylene
Application
Building Industry
siding
door/window
other
conduit
Irrigation pipe
other pipe
roofing
liners
wire-cable
weatherstrip
garden hose
glazing
panels/siding
pipe
glazing
fixtures
glazing
Packaging
film
containers
film
containers
Usage
1000 metric
tons

410
284
33
493
247
1912
22
25
410
38
48
41
122
237
90
9
84

696
2990
375
325
Exposure

high
high
high
moderate
high
low
high
moderate
NA*
moderate
high
high
high
low
high
low
high

NA *
NA
NA
NA
Thermoplastic
Polyester
containers
                         635
NA

-------
                                     13-9
                                 EXHIBIT 13-2
                                  (Continued)
Type of Plastic

Polyethylene
Polystyrene
Polypropylene
Source: Based on 1985
Usage
Application 1000 metric
tons
Housewares and Toys
954
378
176
estimates as reported in Mo_dern_Plastics ,
Exposure

NA
NA
NA

         (1986) 63(1) :  69.

*NA indicates exposure  information is not available.   However, based on the
application,  it is expected to be low to moderate.

-------
                                     13-10
    Discoloration of polymeric materials is also related to bond breaking,  but
in these cases, bonds that are broken may have little effect on mechanical
properties.  The best example of a polymer discoloring by yellowing is in
transparent flexible polyvinyl chloride (PVC) window materials used for rear
windows of convertible automobiles.  The PVC darkens from weathering because of
impurities in the resin initiating a dehydrochlorination reaction as outlined
below:

                                            H
                                light       |
                     -CH2-GH- 	> -C=CH- + HG1                    (12)
                          |    impurities
                          Cl
The color is caused by the multiple double bonds (-9 repeat units in sequence)
created between the carbon atoms, resulting in a polymer that absorbs the light
in the visible region of the spectrum and thus appears yellow.   Often this
discoloration is evident long before mechanical properties of the polymer
material begin to change.  This, again, is because mechanical properties are
related to the molecular weight of the polymer more than to the chemical
structure.

    Other changes in polymeric materials from UV exposure, similar to yellowing,
also occur mostly at the surface.  Such changes as surface cracking, chalking or
crazing (i.e., when a clear surface becomes translucent because of formation of
microcracks) are all evident in polyolefins, polycarbonates, vinyl, acrylic, and
styrenic polymers when exposed to UV radiation.  These surface effects
drastically reduce the life expectancy of materials that require maximum
transmission of visible light, such as Plexiglas and greenhouse windows, solar
energy collectors, and streetlamp globes.

    Several different modes of damage occur concurrently.  However, depending on
the application of the polymer, one or more of such properties may be important.
Exhibit 13-3 gives the critical mode of damage and one or more secondary modes
of damage identified for selected applications of polymers.  The critical mode
of damage is defined as the degree of damage that renders the material
"unacceptable" at the shortest duration of exposure, thus reducing the polymer's
service life.

Plastic Compounds

    Polymer resins are rarely, if ever, used by themselves in any application.
To optimize the properties demanded by a given application and to enable the
product to be easily fabricated, a number of other "ingredients" are added to
the polymer to obtain a "compound."  The nature of the non-polymer ingredients
in a compound and the. ratio in which they are mixed will depend on the
processing technique, the grade of polymer, and the desired properties in the
end product.

-------
                                    13-11
                                EXHIBIT 13-3

               Modes of Damage Experienced by Polymers Used in
                            Outdoor Applications
       "Polymer
    Application
   Damage
 Polyvinyl chloride
siding,  window frames
                             pipes



                             roofing materials

                             automobile upholstery

 Unsaturated polyethylene    outdoor surfaces
 Polyethylene/polypropylene  irrigation pipe,  outdoor
                               furniture
                             synthetic turf,  stadium
                               seats,  packaging
 Polycarbonates
glazing material
[+ yellowing]
+ chalking
- impact properties
- tensile properties
+ surface distortion

[- burst pressure]
- impact properties
+ yellowing

[+ brittleness]

+ discoloration

[ + surface erosion]
+ discoloration
- strength

[+ brittleness]
- tensile properties
- electrical properties
[+ yellowing]
+ loss of transparency
+  = increase
   = decrease
[ ]  = brackets indicate critical mode of damage

Source:  Andrady,  A.,  "Analysis of Technical Issues Related to the
         Effect of UV-B on Polymers," Research Triangle Institute,
         Research Triangle Park, N.C., March 1986.

-------
                                     13-12
    An example of a rigid PVC siding compound is given in Exhibit 13-4.

    The fabrication of the siding involves feeding the melted compound through a
slit die, a process where the polymer is heated and sheared.   Therefore  a heat
stabilizer is usually used in the compound.   The lubricant prevents sticking of
melt to metal parts in the processing machinery.  A suitable  rubber is added as
an impact modifier to reduce the brittleness of PVC resin.  The processing aid
is added to improve the melt processing characteristics while the chalk  is added
to cheapen the compound.  Titanium dioxide is the light stabilizer.  As
explained in the previous sections,  several  polymers when exposed to sunlight
(even under current UV conditions),  would degrade, some of them rapidly.
Therefore, stabilizers such as titanium dioxide are used to reduce the damage
from exposure to light.  Any depletion in the stratospheric ozone with potential
increase in the UV-B radiation would accelerate the degradation process.

Light Stabilizers in Polymer Products

    Light stability in polymeric materials is achieved by the inclusion of an
appropriate "light stabilizer" in a polymer  compound.  Light stabilizers are
substances that absorb the light in the UV-B range of the spectrum very
efficiently.  When incorporated into a transparent polymer matrix, the
stabilizer molecules compete with the polymer molecules for the available light.
Light absorbed by the UV-absorber molecules  is harmlessly dissipated and does
not lead to any damage of the polymer matrix.  For best results the light
absorber should be soluble in polymer, absorb light in the UV-B range of the
spectrum, and have a very high molar extinction coefficient (so that it  is
effective at relatively low concentrations).

    Three different methods are used to achieve light stability.  The first
method involves use of UV-absorbing compounds such as 2-hydroxy-benzophenones,
2-hydroxyphenylbenzotriazoles, 2-hydroxyphenyl-S-triazines and derivatives of
phenyl salicylates (Heller 1969).  These compounds absorb and dissipate  the
light without degrading the polymer.

    A second approach is to prevent light from reaching the bulk of the  polymer
(light shielding) by opacification of the matrix.  Inclusion of a light
absorbing (or light reflecting) insoluble pigment in the matrix prevents the
light from penetrating the bulk-of the polymer.  However, a fraction of surface
layer is unprotected and is therefore affected by light.  Titanium dioxide
absorbs UV-B radiation very effectively and is used in PVC compositions  as a
light screener.  Other pigments used as light screeners include zinc oxide,
magnesium oxide, lead carbonate, and barium sulfate.  One of the most effective
light screeners is carbon black.  Used in appropriate concentrations and
adequately dispersed in the matrix,  the small-sized carbon black particles
impart excellent light stability in polyethylene formulations.  Exhibit 13-5
gives the UV screening effectiveness of several selected pigments at
concentrations of 2% in polypropylene films of  10-ml thickness  (Uzelmier 1970).
The screening power is given as the ratio of the weathering stability (lifetime)
of the pigmented to unpigmented materials (Andrady 1986) .  Higher

-------
                  13-13
              EXHIBIT 13-4

     FVC Siding Compound Composition
   Component
Concentration (pph)*
     PVC resin
     Heat stabilizer
     Lubricant
     Titanium dioxide
     Impact modifier
     Calcium carbonate
     Processing aid
         100
         1.5
           1
          12
           6
           3
           1
* pph = parts per hundred parts resin

Source:  Andrady, A., "Analysis of Technical
         Issues Related to the Effect of UV-B on
         Polymers," Research Triangle Institute,
         Research Triangle Park,  N.C., March 1986.

-------
                                     13-14
ratio indicates longer stability of the polymer.   For example, a polymer that is
compatible with carbon black would have a lifetime 12 times longer when
pigmented with carbon black and exposed to sunlight than without carbon black.
On the other hand,  a polymer that is compatible with cadmium yellow would have a
lifetime twice as long when pigmented with cadmium yellow (at the same
concentration as carbon black in the previous polymer) than without cadmium
yellow.

    The UV absorbance at specific wavelengths is also given in Exhibit 13-5.
This is an indication of how well the pigment is able to absorb the light.
Higher absorbance for titanium dioxide (0.92 at 300 nm) than for iron oxide
(0.29 at 300 nm) means that at the same pigment concentration (2%), titanium
dioxide is able to absorb more light than iron oxide.

    The third approach to achieve light stability does not prevent or control
the interaction of light with the polymer.  Instead, the light stabilizer
additive either seeks to deactivate or quench the excited state chromophores
before the degradation process occurs, or acts as a free radical scavenger,
which mops up the damaging free radicals as soon as they are formed.  Complex
chelates of nickel, for example, have been used as light stabilizers in
polyolefins and are believed to act as quenchers and peroxide decomposers.


                                  EXHIBIT 13-5

                UV Screening Effectiveness of Selected Pigments
                                                      Absorbance
           Pigment              Screening Power    300nm  325nm  350nm
Titanium Dioxide
Iron Oxide
Cadmium Yellow
Chromium Oxide
Carbon Black
(Channel Black)
2
3
2
2
12

.25
.25
.0
.75
.0

0.
0.
0.
0.
1.

92
29
36
05
56

0
0
0
0
1

.97
.30
.36
.06
.56

1
0
0
0
1

.01
.32
.36
.07
.56

            Chemical interactions limit the use of some of the above pigments in
            several polymers.  For instance, iron oxide can be effectively used
            with polypropylene but not with PVC, as it catalyzes the photo-
            oxidation of the latter polymer (King 1968).


    In recent years the use of radical scavenger molecules, particularly the
very effective Hindered Amine Light Stabilizers (HALS), has gained wide
industrial acceptance.  Exhibit 13-6 gives the current U.S. consumption of light
stabilizers in various types of plastics.

-------
                             13-15



                          EXHIBIT 13-6

       Domestic Consumption of Light Stabilizers,  1984-85



                                Domestic Consumption (1000 Ib)

                                        1984     1985
 Plastics (UV Stabilizers)*

 Polyolefins
 Polycarbonates
 PVC and related copolymers
 Acrylic
 Polyester

 Pigment/Colorant
4290
 352
 171
 165
 165
4600
 400
 180
 170
 171
Titanium Dioxide
Carbon Black
Iron Oxide
Cadmium Compounds
Chromium Compounds
71
80
9
5
5
272
85
10
5
5
*Figures for plastics do not include light shielders such as
 titanium dioxide
 Source:  Andrady, A., "Analysis of Technical Issues Related
          to the Effect of UV-B on Polymers," Research Triangle
          Institute, Research Triangle Park,  North Carolina,
          March 1986.

-------
                                     13-16
DAMAGE FUNCTIONS AND RESPONSE TO DAMAGE

    Depletion of the stratospheric ozone would change the spectral distribution
as well as the solar radiation pattern, with a potential increase in the UV-B
radiation striking the earth's surface.  The magnitude of such changes on the
weathering of plastics and potential damage that might occur are difficult to
assess due to lack of experimental data.  However, various theoretical
approaches are considered to estimate the damage.

    There are three possible responses to potential damage from increased UV-B
radiation:

        1.  Add greater amount of light stabilizers to prevent further
            degradation of polymers.

        2.  Develop new or improved performance stabilizers to absorb the
            additional radiation.

        3.  Switch from polymers to alternative materials (e.g. ceramics, metal,
            etc.)

Several attempts have been made to estimate the increased damage to materials
from potential increase in UV-B radiation.  Because of the complexity of the
analysis, each approach has focused on only the first response.  Each approach
analyzes how much additional light stabilizer would be required for a given
ozone depletion (function of UV light intensity) scenario, to maintain the
current level of useful life for various polymers.  Thus, each approach has
attempted to relate intensity of the UV light and concentration of the light
stabilizer in the polymer.

Transparent Polymers (GIAP Approach)

    In a research study funded by the U.S. Department of Transportation
(Climatic Impact Assessment Program (CIAP)), Schultz, Gordon, and Hawkins (1974)
were the first to examine the potential adverse effects of the ozone depletion
on materials.  Although their concern was depletion of ozone from SST flights
and not from discharge of chlorofluorocarbons,  their approach appears on the
basis of first principles to be correct.

    The CIAP approach employed the Beer-Lambert law to calculate approximate
increases in UV stabilizer required to offset the effects of increased UV light
(Andrady 1986).   According to this law, for a homogeneous material with no
scattering losses at the surface or in the bulk of the polymer, the
monochromatic light absorption by the system is described as follows:

    The fraction of incident monochromatic light energy absorbed is proportional
to the concentration of the chromophores and to the path length of the light
within the substrate.

-------
                                     13-17
                    Log It/IQ = -ec£                                        (13)

    I  = intensity of the transmitted light

    I  = intensity of the incident light

     e = molar extinction coefficient (proportionality constant)

     c = molar concentration of the light absorbing moiety

     H  = thickness of the film

    An increase in the incident light intensity will result in a higher light
flux in the polymer; therefore, to maintain the same level of protection, the
concentration of the UV absorber has to be increased.  If an increase in the
light intensity by a factor x required a corresponding increase in the UV
absorber concentration by a factor y, the Beer-Lambert Law for the new system
gives
                          Log(It/I0-X) = -eyc^                              (14)

    The two above equations can be combined to give

                            y - Log(x)/H + 1                                (15)

         where H = ecJi, is a constant.


This simple equation predicts the variation of the factor increase in the UV
absorber concentration, y, with the factor increase in the intensity of light,
x.  An empirical observation made is that an increase in the UV light intensity
by a factor of 1.5 generally requires an increase in the UV absorber
concentration by a factor of 2.0.

    Based on this empirical observation and the damage function (equation 15),
Schultz, Gordon, and Hawkins developed the following relationship:

                        y = 1 + (Iog10x)/(log1()1.5)                         (16)


    Using the above relationship and using a brash extrapolation of the world
market, Schultz, Gordon, and Hawkins developed increased stabilizer requirements
for various levels of ozone depletion.  A market period of 50 years (1970-2020)
was considered in the CIAP study to estimate the future market of plastics.  The
results are summarized in Exhibit 13-7.

Battelle Study

    The Battelle study (Hattery, McGinniss,  and Taussig 1985) was a follow-up
to the CIAP study.  The basic assumptions remained the same,  but all economic
market data used for analysis was updated using more current information.

-------
                     13-18
                 EXHIBIT  13-7

         Increased Stabilization Market
                  (1970-2020)
 Ozone Depletion        By 1990	By 2020
    (Percent)         (millions of 1974 dollars)*
50
20
10
5
1,040
370
172
84
5,040
1,800
840
408
* Values at zero discount rate.

Source:  Schultz,  A., D. Gordon, and W. Hawkins,
         "Economic and Social Measures of Biologic
         and Climatic Change,"  CIAP Monograph 6,
         September 1974.

-------
                                     13-19
Also, a different ozone depletion scenario was used in the Battelle study.
Exhibit 13-8 below shows the ozone depletion estimates used in the study.
                                  EXHIBIT  13-8

                           Ozone Depletion Estimates
                          Year       Ozone Depletion
                                       (percent)
1985
1995
2005
2015
2025
2035
2045
2055
2065
2075
0.0
0.15
0.62
1.50
2.84
4.66
7.08
10.41
15.44
26.08
                         Source:  EPA estimates based on emission scenario
                                  developed by EPA and modeled by Lawrence
                                  Livermore Laboratory.


    For the projection of the polymer demand to the year 2075, Battelle used a
simplified assumption that the growth rate ratio of UV-stabilized plastics to
GDP (Gross Domestic Product) for the United States will remain constant at 1.7
percent into the 21st century.

    Historical U.S. market share of the world plastic market and projection of
the U.S. plastic market were used to project the world plastic market through
the year 2075.

    In the Battelle study, Cutchis's formula was used to estimate the effect of
ozone depletion on the average UV light intensity (Cutchis 1974):
                                log x = 0.00954D                           (17)


where x is UV light intensity factor and D is the percentage of ozone
depletion.

    To describe the relationship between light intensity (x) and the stabilizer
concentration (y),  Battelle used a mathematical formula developed by Shultz,
Gorden, and Hawkins (1974):

-------
                                     13-20
                                log x = 0.176 (y-1)                        (18)
    The combination of the Cutchis and Shultz formuli,  gave the following
relationship:
                                y-1 = 0.0542D                              (19)
    This combined formula was used in the economic impact model.

    Exhibit 13-9 gives cumulative added cost for the United States and the
world in the presence of ozone depletion.
                                 EXHIBIT  13-9

                             Cumulative Added Cost

Year

1985
2015
2045
2075

Ozone Depletion
(percent)
0.0
1.50
7.08
26.08
U.S.
(1983 $

0.0
0.13
2.3
19.4
World
billion)*

0.0
0.60
14.0
120.0
              * Values at zero discount rate.

              Source: Hattery, G.,  V. McGinniss, and P. Taussig "Costs
                      Associated with Increased Ultraviolet Degradation of
                      Polymers." BATTELLE Columbus Laboratories, Columbus,
                      Ohio, April 1985.

-------
                                     13-21
Extension to Filled Polymers (Andrady/Shultz)

    While the CIAP approach is sound and simple to use, its applicability is
limited to the transparent, homogeneous polymers.  With the exception of window
glazing, polymers used outdoors are generally opaque and are often stabilized
using light screeners .   Therefore a parallel model based on light
stabilization, applicable to opaque materials (filled polymer) ,  was developed
by the Research Triangle Institute (Andrady 1986) .

    A filled polymer can be described by a homogeneous resin matrix in which
essentially spherical filler particles are randomly dispersed.  The filler can
be a pigment with high light absorptivity, so that all the light reaching the
filler particle is absorbed.  For simplicity, the random distribution can be
described as spheres of radius r, in two dimensions, within a series of layers
of thickness 2r (see Exhibit 13-10).   If the pigment volume fraction is V, the
number of spherical pigment particles within a lamina volume element of yy2r
cm3 is N = 2ry2V/(4jr/3)r3.

    The fraction of light obliterated in passing through the laminar is the
cumulative cross -sectional area of the spheres per unit laminar area, Njrr^/y2 ,
which is equal to 3V/2 .   The surface intensity,  lo, can then be related to 1^,
intensity at depth H ,  as
                        I = I0(l-3V/2)                                     (20)

    For light intensity increased by a factor x and a corresponding increase in
the light stabilizer by a factor y, the above equation can be modified as
                             I - x-I0(l-3Vy/2)                             (21)
    Using the assumption of equal residual fractional intensity at a
characteristic depth of 2 units, it can be shown that
                       Inx =  JL •  ln((l-3V/2)/(l-3Vy/2))                   (22)
                              2r

    This equation has the same predictive feature as the CIAP equation but
requires a knowledge of S. and r for its application.

    An empirical observation relating x to y is needed in this case as well;
however, such experimental data are not presently available for the polymers of
interest.  No economic damage estimates have been developed using this
approach.

-------
                                     13-22
Comprehensive Approach (Andrady 1986)

    While the two approaches described previously can be used for clear or
filled polymers to determine the amount of additional stabilizer requirements
                                 EXHIBIT 13-10

         Diagrammatic Representation of  the Effect  of Pigment/Fillers
               as Light Shield.  (Monodisperse Spherical Filler)
                                    Light
                                                     Polymer macrix
                                                     2r
                           r
corresponding to an increase in the light intensity, neither approach takes
into account the activation spectra for degradation.

    The extent (and even the nature) of damage suffered by a polymer on
exposure to sunlight is markedly wavelength dependent.  More appropriate
methodology to obtain the damage function should include (1) source spectral
distribution, (2) activation spectra, and (3) the functional relationship
between the ozone concentration in the stratosphere and the spectral
irradiance.

    Atmospheric attenuation of sunlight is a result of several processes.  The
following are the three main processes:  (1) Rayleigh scattering -- scattering
of light due to gas molecules in the atmosphere; (2) aerosol scattering or Mie
scattering caused by fine particles and fluid droplets suspended in the
atmosphere;  and (3) absorption by ozone.  These factors should be considered in
calculating the spectral distribution and intensity of light at a given
location.

    An activation spectra should be available for each specific polymer to
assess the damage.  A normalized activation spectra then can be used as a
continuous weighting function describing the spectral sensitivity of a specific
damage process in a given polymer.

    The total damage function (index), D, for a polymer exposed to sunlight can
be presented as

-------
                                     13-23
                                 -I
F(A)H(A)dA                            (23)
where   F(A) = adjusted normalized activation spectrum (ANAS)
        H(A) = spectral distribution of sunlight for natural weathering
        A.. ,  A- = wavelength range of the ANAS

    It is assumed that the effect of each wavelength in the range is additive,
with no synergism or mutual inhibition.

    If H(A)  represents an actual measured terrestrial spectrum,  then D will be
a measure of the damage under current ozone layer conditions, arbitrarily
defined as being deteriorated to zero extent.

    The above approach of using the actual measured spectrum of the solar
irradiance as the source of spectral distribution limits the usefulness of the
analysis to specific localities and possibly to certain times of the year (the
time when H(A) is a fair representation of the solar spectrum) .   This approach
avoids various scattering coefficients and their effects.   If we assume that
the depletion of the ozone layer has the sole effect on the spectral
distribution (with other effects being small) ,  the damage  index can be modified
(D*) ,  which will relate to damage under reduced ozone level conditions.

    This requires a knowledge of the absorption coefficient of ozone.  Reliable
data have been published for the absorption coefficient, A(A) (Griggs 1968):

                A(A) = exp (38.54 - 0.1237A)                               (24)

    The effect of reduced ozone concentration on the terrestrial solar
irradiance at a given wavelength can be expressed as
                         ',0)  = I0exp(-A(A)c'sece-T(A)sec0)                 (25)

    where

        A(A)  = ozone absorption coefficient,  equation (20)

        c'    = ozone concentration at reduced level

        sec©  = approximation term for path length of light  through the  air
                 mass

        r    = Rayleigh scattering thickness

        Q    =• solar zenith angle at the  location

-------
                                     13-24
    If we neglect the Rayleigh scattering, the factor increase in the UV
radiation due to ozone loss can be given by
                              = exp(A(A)c(l-c'/c)sec9)                     (26)


where (l-c'/c) is the fractional decrease (f) in the ozone concentration.

    The damage index under the conditions of ozone layer deterioration can be
expressed as
                     D(f,6) =    I F(A)H(A)exp(A(A)-c-f-sec0)dA             (27)
                              V

    The seasonal variation of ozone concentrations, the seasonal variation of
the solar zenith angle and the diurnal variation of the sun angle (time angle)
should be considered while estimating the damage index for different time
periods (i.e., daily, annual, seasonal, etc.)  For instance, the daily damage
index at a given level of ozone depletion is given by
                                   II
                    D(f,8)24 - 2 Z I I F(A)H(A)A(A)dAdt                     (28)


where A(A) is the factor increase defined in equation (26).  The value of ozone
concentration c is no longer a constant but is time dependent c(t).   The term Z
is a constant used to convert the time angle to seconds.

Illustration of Comprehensive Approach for Polyvinyl chloride:  A Case Study

    Using the solar spectrum as observed at Miami, Florida, on March 23, 1985,
at the solar noon, spectral distribution H(A) was calculated for various
wavelengths.  An adjusted normalized action spectrum ,  F(A),  was developed
using the Miami sunlight spectrum.  The integral in equation (28) was evaluated
numerically using a computer program.  Damage function and the factor increase
in the damage were calculated for several solar zenith angles at different
fractional ozone depletion.  The results are shown in Exhibit 13-11.

    As might be expected, the factor increase in damage is higher at larger
solar zenith angles and hence will be greater for higher latitudes.   However,
the solar irradiation is also lower at these latitudes.

    The value of the damage index is very sensitive to  (a) the action spectra
used in its calculation, (b) the zenith angle of the sun at the location in
question, and (c) the thickness of the polymer used.

-------
                                   13-25
                               EXHIBIT 13-11

   Relative Damage Indices for Yellowing of PVC Under Miami (March 22nd)
       Conditions, at Different Extents of Ozone Layer Deterioration
                             Relative Damage Index at Different Solar
                          	Zenith Angles*	
Fractional Loss of Ozone            0°          30°           60°
0.0
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
1.00
1.01
1.02
1.03
1.04
1.05
1.06
1.08
1.09
1.11
1.13
1.00
1.01
1.02
1.03
1.05
1.06
1.08
1.09
1.11
1.14
1.17
1.00
1.02
1.04
1.06
1.09
1.13
1.18
1.25
1.36
1.55
1.93
* Relative damage index = D/D  where D  is the damage index calculated for
  current levels of ozone; solar irradiance values used were for a surface
  facing south at 45°.

Source:  Andrady, A., "Analysis of Technical Issues Related to the Effect of
         UV-B on Polymers," Research Triangle Institute, Research Triangle
         Park, North Carolina,  March 1986.

-------
                                     13-26
    The thickness of the polymer is an important factor.  In the case of a film
of the material, a fraction of the incident light is transmitted through it.
Often, the transmitted light is deficient in UV radiation.  In the case of a
thick polymer, the light is totally absorbed but the absorption is limited to a
thin surface layer of the polymer.  Thus in those modes of damage involving the
surface integrity, thick samples undergo as much damage as the thin films.
This is true, for instance, where "yellowing" is the critical mode of damage.

Mathtech Study

    This is the most recent study in which the economic damage from potential
depletion of ozone was estimated.  The study used the comprehensive approach
developed by the Research Triangle Institute (Andrady 1986).

    According to the comprehensive approach, the total damage D is given by


                         D =    AH(A) A (A) F(A) dA          (equation (23))


    The integral might be evaluated numerically with A(A) - 1 to represent
current baseline conditions.  However, the lack of activation spectrum for
polymers limits the usefulness of the approach.  The only action spectrum
available in literature relates to transparent PVC films and to type of damage
other than yellowing.  However, a very approximate estimate might be made using
the activation spectra for polyene formation based on Reinisch, Gloria, and
Andros 1970.  Yellowing in PVC is a direct result of the generation of long
polyene sequences.

    Based on the data, the damage (D) can be estimated as

                                Ln(D) = a + g(S)                            (29)


        where  a, g = constants
                  S = stabilizer concentration
    For reduced ozone level, this equation can be modified further to the
following form:
                             S'    1         D'
                             	  (—   -In —) + 1                         (30)
                             S     g(S)       D
    where   D' = factor increase in damage
            D

-------
                                     13-27
            S'  = factor increase in titanium dioxide stabilizer
            S

            S',D'  = stabilizer concentration and damage at reduced ozone level


    This equation allows the determination of the required concentration of the
stabilizer of increased extent of damage.  Applying the data available for
yellowing (damage) of rigid PVC versus various stabilizer concentrations to the
above equation,  ranges of factor increase in damage and stabilizer
concentration are calculated for PVC as illustrated in Exhibit 13-12.  The same
approach can be extended to other polymers and stabilizers.

    To calculate the economic damage,  Mathtech has maintained a hypothesis
similar to the other studies, which is that producers of the PVC products will
change the amount of stabilizer (titanium dioxide) in the resin compound in
order to maintain quality and lifetime characteristics of their products.
However, the earlier studies (Schultz, Gordon, and Hawkins 1974; Hattery,
McGinniss, and Taussig 1985) did not consider how changes in compound
formulation would affect production costs of PVC--which can occur in several
ways.  First, resource costs will be increased as more titanium dioxide is used
per pound of plastic produced.  Second, the change in formulation will lead to
more frequent replacements of screws and barrels since titanium dioxide is
abrasive in nature and its higher concentration will increase wear and tear of
the machinery.   Third, the change in formulation will lead to increased energy
requirements for PVC processing equipment since the melt viscosity of the
extrudate increases with titanium dioxide increase and will require higher
energy to operate the screw.  Also increased viscosity may require more
lubricant and processing for successful extrusion.

    In order to estimate the cumulative damage, demand for PVC was estimated
using historical data and related to construction activity since PVC is used
heavily in several outdoor construction activities.  The supply side of the
market for PVC products was determined through a model plant analysis.  The
information from demand and supply analyses is used to determine aggregate
price indices for PVC products and subsequent economic damage.

    The ozone depletion estimates were the same as those shown in Exhibit 13-8.
The total PVC damage associated with ozone depletion is given in Exhibit 13-13.

    The results are made for the following set of circumstances:  1) All
estimates are reported in millions of 1984 dollars; 2) the factor increase in
titanium dioxide concentration is computed for a zenith angle of 60°; and 3)
polymer producers do not respond to the hypothesized decreases in stratospheric
ozone until 10 years after the impact is observed.

    As indicated so far, because of the lack of data, the Mathtech analysis is
limited to PVC only.  In order to provide broader coverage of potential
damages, it is appropriate to consider other polymers that may be adversely
affected by increased UV-B radiation.

-------
                                  13-28
                              EXHIBIT 13-12

      Estimated Ranges of Factor Increase  in Damage and the Factor
         Increase in Stabilizer Needed to Counter the Change for
                   Yellowing of Rigid PVC Compositions
                            Zenith Angle 30° &          Zenith Angle 60°
                            Spring ~ Noontime         Fall ~ Noontime
Percent Loss of Ozone D'/D^
0-5
5-10
10-20
20-30
- 1.
1.01-
1.02-
1.03-
01
1.02
1.05
1.08
1
1
1
1
S'/SC
.01-1
.01-1
.03-1
.03-1
.02
.05
.11
.18
1
1
1
1
D'/D
.01-1
.03-1
.04-1
.07-1

.02
.04
.09
.18

1
1
1
1
S'
.01
.03
.05
.08
/s
-1.05
-1.09
-1.20
-1.38
a  Zenith angles selected to reflect North American locations.

D  D'/D = factor increase in damage

c  S'/S = factor increase in titanium dioxide stabilizer
Source:   Horst, R.,  K. Brown,  R.  Black, and M.  Kianka,  "The
          Economic Impact of Increased UV-B Radiation on
          Polymer Materials: A Case Study of Rigid PVC,"
          Mathtech, Inc., Princeton, New Jersey, June 1986.

-------
                                     13-29
                                 EXHIBIT 13-13
                        FVC Damage with Ozone Depletion
                          (Millions of 1984 dollars)
                                                   Discount Rate
                                               0%	2%	5%	10%
Low
Middle
High
2,440
4,716
9,158
603
1,137
2,159
97
174
315
10
17
27
    Source:  Horst,  R.,  K.  Brown,  R.  Black, and M.  Kianka, "The Economic
            Impact of Increased UV-B Radiation on Polymer Materials:   A Case
            Study of Rigid PVC," Mathtech, Inc., Princeton,  New Jersey, June
            1986.
    The following equations, developed by Mathtech, Inc., provide future demand
projections for each polymer:
                          Q = A (1 - exp (-kT))
(31)
For small values of parameter k, the above equation can be linearized as
follows:
                          QT = a+b (T-TQ) + c(T-TQ)2                       (32)

    where       Q = consumption per person in year T

               Tn = year for which the projection is made

          a, b, c - parameters to be estimated.

    Projections of future demands for selected polymers and selected years are
provided in Exhibit 13-14.  Damage to these .polymers is difficult to estimate
without any activation spectra.  However, the magnitude of the damage is
expected to be similar to that for PVC.

EFFECT OF TEMPERATURE AND HUMIDITY ON PHOTODEGRADATION

    A necessary element in the discussion of damage functions missing so far is
the possible role of temperature and humidity.  Increasing the temperature
almost always results in an increase in the rate of a chemical reaction.  This
is certainly true of the chemical reactions responsible for photodegradation.
The extent of such an increase depends upon the energy of activation for the
particular reaction.  The activation energy for photodegradation of

-------
                     13-30
                 EXHIBIT 13-14

Projections of Future Demand for Selected Years
           (Thousands of Metric Tons)
Year
1980
2000
2025
2050
2075
Acrylics
75.0
148.5
179.4
186.7
190.0
Polyester
644.0
1130.6
1400.0
1472.8
1504.8
    Source:   Horst,  R. ,  K.  Brown,  R.  Black,  and
              M.  Kianka,  "The Economic Impact of
              Increased UV-B Radiation on Polymer
              Materials:  A Case Study of Rigid PVC,"
              Mathtech,  Inc., Princeton, New Jersey,
              June 1986.

-------
                                     13-31


polypropylene in air, for instance, is 77-80 kcal/mole.  That for PVC (based on
polyene formation) determined by Reinisch, Gloria, and Androes (1970) was 18
kcal/mole.  These values allow the estimation of approximate factor increase in
the weathering reaction due to increased temperature.

    The general effect of humidity or water is to increase the rate of
degradation of the polymer.  It is generally believed to cause slight
plasticization (or softening) of the polymer matrix.  This is hardly likely to
be the only mechanism.  Several experimental studies strongly suggest
water/pigment interactions.  The effect of irradiation of titanium dioxide in
the presence of water was shown to generate hydrogen peroxide (Hoffmann and
Savacz 1971).  The deleterious effect of peroxides on polymers is well known.
Thus,  depending upon the chemical nature of polymer, the components of the
compound, and the weathering factors, both temperature and humidity tend to
increase the rate of degradation.  Generalized quantitative data are not
available on the polymers of interest to provide potential damage estimates.

FUTURE RESEARCH

    As evident in the discussion of this chapter, there are no experimental
data available on several polymers of interest to provide a good estimate of
damage in polymers under increased light intensity.  The EPA is currently
funding a research study to evaluate the available scientific and technical
information and conduct experimental as well as field studies to evaluate
effects of increased UV radiation on polymers.

    Polyvinyl chloride (PVC) polymers are selected for the experimental study,
since PVC is the most widely used thermoplastic for outdoor applications.  Two
main categories of PVC formulations, (a) rigid PVCs such as those used in
siding and window frames and (b) plasticized PVCs such as those used in
flexible roofing membranes and cable coatings, will be studied.  Typical
formulations and formulations with varying amounts of titanium dioxide will be
evaluated in a simulated solar radiation using a Xenon lamp source.

    The field research will study the question of UV-induced material damage
from a global point of view.  The types of UV radiation (in terms of spectral
quality) that North Americans would receive at 10%-20% ozone depletion is
typical of the solar flux received at some equatorial locations at current
levels of ozone.  Studying the material behavior at these locations would be
invaluable in estimating the effects of increased UV light at northern
latitudes.

    The yellowing of PVC will be the main criterion used to assess damage,
although other tests may be employed as necessary.  Data obtained from this
research will be used to assess damage to PVCs and to estimate additional costs
to stabilize PVC in the event of partial ozone depletion.

-------
                                     13-32
REFERENCES

Andrady, A.,  "Analysis of Technical Issues Related to the Effect of UV-B
    on Polymers," Research Triangle Institute, Research Triangle Park, North
    Carolina, March 1986.

Cutchins, 1974  (supply missing info)


Griggs, M.,  "Absorption Coefficients of Ozone in the Ultraviolet and Visible
Regions," Journal of Chemical Physics. Vol. 49(2), pp. 857, 1968.

Mattery, G.,  V. McGinniss, and P. Taussig "Costs Associated with Increased
    Ultraviolet Degradation of Polymers," BATTELLE Columbus Laboratories,
    Columbus, Ohio, April 1985.

Heller, H.,  "Protection of Polymers Against Light Irradiation," European
    Polymer Journal.  Supplementary, 105, 1969.

Horst, R., K. Brown,  R. Black, and M. Kianka, "The Economic Impact of Increased
    UV-B Radiation on Polymer Materials:  A Case Study of Rigid PVC," Mathtech,
    Inc., Princeton,  New Jersey, June 1986.

Hoffmann, E.  and A. Saracz,  "Weathering of Paint Films.  III. Influence of
    Wavelength of Radiation and Temperature on the Chalking of Latex Paints,"
    J. Oil Color Chem. Association. Vol. 54, p. 450, 1971.

Kelen, T., Polymer Degradation. Van Nostrand Reinhold, New York, 1983.

King, A., "Ultraviolet Light: Its Effects on Plastics," Plastics and Polymers.
    195,  1968.

Reinisch, R., R. Gloria, and G. Androes, Photochemistry of Macromolecules.
    Plenum Press, New York,  1970.

Schultz, A.,  D. Gordon, and W. Hawkins, "Economic and Social Measures of
    Biologic and Climatic Change,"  CIAP Monograph 6, September 1974.

Uzelmier, C.,  "Effects of Pigments on the Heat and Light Stability of
    Polypropylene," Society of Plastic Engineers. Vol 26, p. 69, 1970.

-------
                                  CHAPTER 14

               POTENTIAL EFFECTS OF  STRATOSPHERIC OZONE DEPLETION
                            ON TROPOSPHERIC OZONE
SUMMARY

    Tropospheric or ground-based ozone,  03(T), is an air pollutant formed near
the earth's surface as a result of photochemical reactions involving ultraviolet
radiation, hydrocarbons, nitrogen oxides, oxygen, and sunlight.   At high
concentrations,  often found during warmer months, tropospheric ozone can
adversely affect human health, agricultural crops, forests, and materials.  To
protect public health, the U.S. Environmental Protection Agency (EPA) has
established a primary standard of 0.12 ppm (one-hour average), which is not to
be exceeded more than one day per year.   In 1979, EPA also determined that a
secondary welfare standard, more stringent than the primary standard, was
unnecessary for the protection of vegetation.  Currently, EPA is reviewing
available scientific and technical information to determine whether these
standards are adequate to protect health and welfare.

    This chapter reviews preliminary scientific information that suggests that
increases in UV-B radiation may affect the rate of tropospheric ozone formation
in urban areas.   The results from these analyses suggest that increased
ultraviolet radiation increases the rate of ground-based ozone production and
acid rain precursors.  Moreover, global warming, associated with ozone-depleting
substances, may enhance these reactions.  Further analyses of additional cities
are being conducted.  If these analyses confirm these results, it would appear
likely that in the future more cities and regions would violate the ambient air
standards and that more restrictive measures to control hydrocarbons and
nitrogen oxides may be required in order to comply with current standards.

-------
                                     14-2
FINDINGS
    RESULTS FROM ONE MODELING STUDY AND ONE CHAMBER STUDY SUGGEST THAT INCREASED
    ULTRAVIOLET RADIATION FROM OZONE DEPLETION MAY INCREASE THE RATE OF
    TROPOSPHERIC OZONE FORMATION.

    la.   According to these studies, increases in UV-B associated with ozone
          depletion would increase the quantity of ground-based ozone associated
          with various hydrocarbon and nitrogen oxides emission levels.   Results
          for individual cities vary, depending on the city's location and on
          the exact nature of the  pollution.

    Ib.   According to these studies, global warming would enhance the effects
          of increased UV-B radiation on the formation of ground-based ozone.

    Ic.   According to these studies, ground-based ozone would form closer to
          urban centers.  This would cause larger populations in some cities to
          be exposed to peak values.

    Id.   More research is needed to verify and expand the results of these
          initial studies.

    PRELIMINARY RESULTS FROM ONE STUDY ALSO SUGGEST THAT LARGE INCREASES IN
    HYDROGEN PEROXIDE WOULD RESULT FROM INCREASED UV-B RADIATION.

    2a.   If hydrogen peroxide increases as predicted in this study, the
          oxidizing capability potential of the atmosphere, including the
          formation of acid rain,  would be influenced.

    2b.   More research, especially a chamber study, is needed to verify this
          effect.

    INCREASES IN GROUND-BASED OZONE WOULD ADVERSELY AFFECT PUBLIC HEALTH AND
    WELFARE.

    3a.   If UV-B increases enhanced ozone production, more U.S. cities would be
          unable to meet health-based ground-level ozone standards, and
          background ozone would increase.

    3b.   Crops, ecosystems, and materials would be adversely affected by
          increased ground-level ozone.

-------
                                      14-3
INTRODUCTION

    Tropospheric ozone (alternatively called ground-based ozone), 03(T), is an
air pollutant formed in the ambient air as a result of a series of complex
photochemical reactions involving ultraviolet radiation, hydrocarbons, and
nitrogen oxides emitted from mobile and stationary sources, atmospheric oxygen,
and sunlight.  At ambient concentrations, often measured during warmer months,
03 can adversely affect human health, agricultural crops, forests, ecosystems,
and materials.   Interactions of 03 with nitrogen oxides and sulfur oxides may
also contribute to the formation of acidic precipitation.  Typical short-term
(1-hour) 03 levels range from 0.01 ppm in some isolated rural areas to as high
as 0.35 ppm in one of the nation's most heavily populated metropolitan areas.
Daily daylight seasonal averages in some rural areas have been reported to be
0.06 ppm and higher.

    Using 1982-1984 data -- the most recent available -- 216 metropolitan
statistical areas (MSAs) out of 319 (68%) have enough 03 air quality data to
ascertain attainment status.  Of these 216 MSAs, 119 (55%) exceed the current
EPA health-based 03 standard of 0.12 ppm more than once per year.  Thus, more
than one-half of the MSAs with sufficient data violate the standard.
Approximately 115 million people (over one-half of the total U.S. population)
live in areas that exceed the standard.  However, this does not mean that
everyone in these areas is exposed to 03 at concentrations that exceed the
standard (EPA 1986).

    Of MSAs with sufficient data, about 14% (34) have a characteristic highest
concentrations (CHC)  of 0.16 ppm 03 or higher and over 4%  (11) of MSAs have a
CHC above 0.20 ppm 03.  There is no clear temporal trend in 03 concentration
levels in MSAs around the country, although in 1982 -- and to a lesser extent,
in 1981 -- levels were generally lower than in other years during the 1979-1984
time period.

    Generally,  the third quarter (July-September) and seasonal (April-September)
averages of the 8:00 a.m. - 4:00 p.m. daily daylight values are both in the
range of 0.043-0.050 ppm 03.  Statistically significant, but modest,
relationships exist between peak and longer-term mean indices of 03 air quality
in urban areas.

    Recently, EPA (1986) reviewed scientific and technical information on the
known and potential health effects of ozone.  The information includes
respiratory tract absorption and deposition of ozone, studies of mechanisms of
03 toxicity,  effects of exposure to 03 reported in controlled human exposure,
field, epidemiological, and animal toxicology studies, as well as air quality
information.   The results of that review suggest that:

        (1)  The mechanisms by which inhaled 03 may pose health risks
             involve (a) penetration into and absorption of 03 in
             various regions of the respiratory tract,  (b) pulmonary
             response resulting from chemical interactions of 03 along
             the respiratory tract, and  (c) extrapulmonary effects
             caused indirectly by effects of 03 in the lungs.

        (2)  The risks of adverse effects associated with  absorption
             of 03 in the tracheobronchial and alveolar regions of the

-------
                                     14-4
             respiratory tract are much greater than for absorption in
             the extrathoracic region (head).   Increased exercise
             levels are generally associated with higher ventilation
             rates and increased oronasal or oral (mouth) breathing.
             Thus, maximum penetration and exposure of sensitive lung
             tissue occurs when heavily exercising individuals are
             exposed to 03.

        (3)   Factors that affect susceptibility to 03 exposure are
             activity level and environmental stress (e.g.,  humidity,
             high temperature).

        (4)   The major subgroups of the population at greatest risk to
             the effects of 03 include:  (a)  individuals with pre-
             existing respiratory disease (e.g.,  asthmatics  and
             persons with chronic obstructive lung disease or
             allergies), (b) "responders," who are otherwise healthy
             individuals, both adults and children, but experience
             significantly greater group mean lung function  response
             to 03 exposure, and (c)  any individual exercising heavily
             during exposure to 03.

        (5)   The major health concerns associated with exposure to 03,
             in approximate order of strength of the data base
             include:

             (a)    alterations in pulmonary function;

             (b)    symptomatic effects;

             (c)    effects on work performance;

             (d)    aggravation of pre-existing respiratory  disease;

             (e)    morphological effects (lung structure damage);

             (f)    altered host defense systems (i.e., increased
                    susceptibility to respiratory infection); and

             (g)    extrapulmonary effects (e.g., effects on blood
                    enzymes, central nervous system, liver,  endocrine
                    system).

        (6)   The most useful exposure-response information is from
             controlled human exposure and field studies that provide
             a quantitative relationship between alterations in
             pulmonary function and 03 exposure concentrations.

The EPA (1986) document came to the following conclusions on vegetation impacts:

        (1)   The mechanisms by which 03 may injure plants and plant
             communities include (a)  absorption of 03 into leaf
             through stomata followed by diffusion through the cell
             wall and membrane, (b) alteration of cell structure and

-------
                                      14-5
             function, as well as critical plant processes, resulting
             from the chemical interaction of 03 with cellular
             components, and (c) occurrence of secondary effects
             including reduced growth and yield and altered carbon
             allocation;

        (2)  The magnitude of the 03-induced effects depends upon the
             physical and chemical environment of the plant, as well  ®
             as various biological factors (including genetic
             potential, the developmental age of the plant and
             interaction with plant pests);

        (3)  Effects of 03 on vegetation and ecosystems have been
             demonstrated to occur from both short-term and long-term
             exposures.  Although there are a limited number of
             studies in which short-term (1-2 hour) exposures have
             resulted in growth and yield reduction, there is a
             growing body of evidence that repeated peaks above a
             given level are important in eliciting plant response;

        (4)  Concerning long-term exposures,  the bulk of the evidence
             indicates that growth and yield losses occur in several
             plant species exposed to seasonal concentrations of 03,
             typically characterized as the daily daylight mean over
             the growing season.  In addition, evidence indicates that
             forests experience cumulative stress as a result of
             chronic exposure to 03.  Exhibit 14-1 summarizes the
             range of 03 levels and exposure times required to induce
             5% and 20% foliar injury.  Exhibit 14-2 (EPA 1986)
             provides a more complete survey; and

        (5) Damage to materials is another effect of 03.  There
            appears to be no threshold level below which material
            damage will not occur; the slight acceleration of the
            aging processes of materials occurs at the level of the
            proposed standard.  The materials known to be most
            susceptible to ozone attack are elastomers, textile fibers
            and dyes, and certain types of paint.

POTENTIAL EFFECTS OF ULTRAVIOLET RADIATION AND INCREASED TEMPERATURES
ON GROUND-BASED OZONE

    Recently, Whitten and Gery (1986) conducted a preliminary investigation into
the potential changes in urban ozone levels because of an increase in solar
ultraviolet (UV) radiation resulting from reductions in the stratospheric ozone
layer.  Estimates were also made of the effects on urban ozone chemistry
resulting from a general warming of the lower atmosphere.  The focus of this
study was the effect of reductions in stratospheric ozone by as much as 30
percent of (1) peak ozone concentrations in urban areas and (2) the attendant
control requirements predicted by a modified version of the Empirical Kinetics
Modeling Approach (EKMA).   The model used in this study was the simple
trajectory model used in the EKMA.  This model was developed by the EPA and is
widely used to assess the effectiveness of emission control scenarios to abate
urban ozone.

-------
                                14-6
                            EXHIBIT 14-1

     Ozone  Concentrations for Short-Term Exposures That Produce
                5% or 20% Injury to Vegetation Growth
                     Under Sensitive Conditions^
                            Ozone Concentrations (ppm)
             	that may Produce 5% (20%)  Injury	
 Exposure                                                   Less
Time, Hours  Sensitive Plants   Intermediate Plants   Sensitive Plants
0.

1.

2.

4.

8.
5

0

0

0

0
0
(0
0
(0
0
(0
0
(0
0
.35
.45
.15
.20
.09
.12
.04
.10
.02
to 0
to 0
to 0
to 0
to 0
to 0
to 0
to 0
to 0
.50
.60)
.25
.35)
.15
.25)
.09
.15)
.04
0
(0
0
(0
0
(0
0
(0
0
.55 to
.65 to
.25 to
.35 to
.15 to
.25 to
.10 to
.15 to
.07 to
0.70
0.85)
0.40
0.55)
0.25
0.35)
0.15
0.30)
0.12
>o

>o

>o

>o

>o
.70

.40

.30

.25

.20
(0.85)

(0.55)

(0.40)

(0.35)

(0.30)
1 The concentrations in parentheses are for the 20% injury level.

Source:  EPA (1986), p. X-4.

-------
                                  14-7
                              EXHIBIT 14-2

Ozone Concentrations at Which Significant Yield Losses Have Been Doted for
 a Variety of Plant Species Exposed Ibder Various Experimental Conditions
Plant Species
Alfalfa
Alfalfa
Pasture grass
Ladino clover
Soybean
Sweet corn
Sweet corn
Wheat
Radish
Beet
Potato

Pepper
Cotton
Carnation
Coleus
Begonia
Exposure Duration
7
2
4
6
6
6
3
4
3
2
3

3
6
24
2
it
hr/day ,
hr/day ,
hr/day,
hr/day,
hr/day ,
hr/day ,
hr/day ,
hr/day ,
hr
hr/day ,
hr/day,
120 days
hr/day,
hr/day,
hr/day,
hr
hr/day,
70 days
21 day
5 days/wk, 5 wk
5 days
133 days
64 days
3 days/wk, 8 wk
7 day

38 days
every 2 wk,

3 days/wk, 11 wk
2 days/wk, 13 wk
12 days

once every 6 days
Yield Reduction, 0
% of Control
51,
16,
20,
20,
55,
45,
13,
30,
33,
40,
25,

19,
62,
74,
20,
55,
top dry wt
top dry wt
top dry wt
shoot dry wt
seed wt/plant
seed wt/plant
ear fresh wt
seed yield
root dry wt
storage root dry wt
tuber wt

fruit dry wt
fiber dry wt
no. of flower buds
flower no.
flower wt
o Concentration
(ppm)
0
0
0
0
0
0
0
0
0
0
0

0
0
.10
.10
.09
.10
.10
.10
.20
.20
.25
.20
.20

.12
.25
0.05-0.09
0
0
.20
.25
Reference
Neely et al. , 1977
Hoffman et al. , 1975
Horsman et al. , 1980
Blum et al., 1982
Heagle et al., 1974
Beagle et al., 1972
Oshima, 1973
Shannon and Mulchi, 1974
Adedipe and Ormrod, 1974
Ogata and Maas, 1973
Pell et al., 1980

Bennett et al. , 1979
Oshima et al., 1979
Feder and Campbell, 1968
Adedipe et al. , 1972
Reinert and Nelson, 1979
for a total of 4 times
Ponderosa pine
Western white pine
Loblolly pine
Pitch pine
Poplar
Hybrid poplar
Hybrid poplar
Red maple
American sycamore
Sweet gum
White ash
Green ash
Willow oak
Sugar maple
6
6
6
6
hr/day,
hr/day ,
hr/day ,
hr/day,
12 hr/day.
12
8
8
6
6
6
6
6
6
hr/day,
hr/day,
hr/day,
hr/day,
hr/day ,
hr/day,
hr/day,
hr/day,
hr/day,
126 days
126 days
28 days
28 days
5 mo
102 days
5 day/wk , 6 wk
6 wk
28 days
28 days
28 days
28 days
28 days
28 days
21,
9,
18,
13,
stem dry wt
stem dry wt
height growth
height growth
+1333, leaf abscission
58,
50,
37,
9,
29,
17,
24,
19,
12,
height growth
shoot dry wt
height growth
height growth
height growth
total dry weight
height growth
height growth
height growth
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.10
.10
.05
.10
.041
.15
.15
.25
.05
.10
.15
.10
.15
.15
Wilhour and Neely, 1977
Wilhour and Neely, 1977
Wilhour and Neely, 1977
Wilhour and Neely, 1977
Wilhour and Neely, 1977
Patton, 1981
Patton, 1981
Dochinger and Townsend, 1979
Kress and Skelly, 1982
Kress and Skelly, 1982
Kress and Skelly, 1982
Kress and Skelly, 1982
Kress and Skelly, 1982
Kress and Skelly, 1982

-------
                                     14-8
    Whitten began by investigating how the ozone formation processes in the
lower troposphere would be affected by increased UV transmission through the
atmosphere.  From preliminary work,  it became evident that the photolysis
channel of formaldehyde that leads to radical products is selectively enhanced
at a higher UV flux.  Formaldehyde emissions are products of incomplete
combustion, and formaldehyde is a major intermediate oxidation product from
virtually all organic molecules.   Radicals from formaldehyde photolysis provide
the main source of radicals needed to drive the chain reactions that generate
photochemical ground-based ozone (Whitten 1983); therefore, an increase in the
rate of photolysis will have a bearing on the design of future control
strategies.

    The photolysis of ozone to electronically-excited oxygen atoms is thought to
be the second most important source of the radicals that drive smog formation.
However, the role of ozone photolysis is different from that of formaldehyde
because ozone is the principal ingredient of ground-based ozone.   At low levels
of oxidation potential, the excited oxygen atoms tend to accelerate
photochemical reactions, making the atmospheric chemistry more efficient in
generating ozone from minimal precursor emissions.  However, at the higher or
more severe ozone levels, the excess radicals can partially suppress the ozone
peak, making the precursors seem less efficient in generating ozone.  Other
photolysis rates can also be affected, but their contribution to smog formation
is less important.

    A specific increase in formaldehyde photolysis to radical products is
difficult  to determine accurately for a number of reasons.  Atmospheric
photolysis rates have often been calculated in 10-nm wavelength intervals.
However, the accurate calculation of formaldehyde photolysis requires fine
spectral resolution for both the formaldehyde absorption cross-section and the
surface solar flux  (related to the ozone absorption cross-section).  Only in
this manner can a comparison of the respective fine structures be performed.
Near-ground solar flux data of high resolution and known stratospheric ozone
abundance have not been readily available.  Hence, Whitten and Gery (1986)
estimated  surface solar fluxes using low-resolution information provided by the
National Aeronautics and Space Administration.

    Using  the spectral data, Whitten and Gery (1986) estimated formaldehyde and
ozone photolysis rates as a function of zenith angle and ultraviolet flux
changes due to stratospheric ozone depletion.  The estimated photolysis rates
for formaldehyde, coupled with the information compiled for surface ozone
photolysis, provide the inputs needed to calculate diurnal photolysis rates for
projected  future ozone column densities.  This information is used to evaluate
the future impact of stratospheric ozone changes on near-surface ground-based
ozone formation.

    The preliminary results of the possible effects of increased ultraviolet
radiation  on urban  ground-based ozone are based on simulation of atmospheric
conditions for three urban cities:

        (1)  Nashville  -- because it is nearly in compliance with the
             0.12 ppm  federal ozone standard.

-------
                                      14-9
        (2)  Philadelphia  --to represent cities that require moderate
             control (30%-50% reduction in organic precursors) to
             achieve the 0.12 ppm standard.

        (3)  Los Angeles -- because of the severity of the exceedance
             of the ozone  standard in that region.

Present-day estimates assume a total ozone column of 300 Dobson Units (DU) and
future predictions assume  a total ozone column of 250 and 200 DU.  A 33%
decrease in total ozone column would increase ozone photolysis by approximately
a factor of two, and increase formaldehyde photolysis to radical products by
nearly 20%.  Both of these increases are approximations because of uncertainties
regarding the spectral fine structure and other factors.

    The effects of these photolysis rate increases are given in Exhibit 14-3 for
background temperatures of 298 and 302 Kelvin (K).  For all three cities, the
model predicts higher ozone concentrations resulting either from increases in
temperature or decreases in the Dobson number.  The magnitude of the increases
in ozone is apparently a function of local hydrocarbon-to-nitrogen-oxide ratio,
reactivity, meteorology, and emission distribution.  The linearity of the
response to stratospheric  ozone depletion is not general.  For the Los Angeles
case, the increases in ozone are moderate and quite linear with decreasing
Dobson number.  For the Philadelphia case, ozone is predicted to increase
progressively as the Dobson number declines; that is, only a modest increase in
ozone is seen for a decline in Dobson number from 300 to 250 units, but a more
dramatic increase in ozone is seen when the Dobson number declines from 250 to
200 units.  For the Nashville case, the increases are all very dramatic and the
linearity exists only if temperature does not increase.  At the "current"
temperature of 298K, the simulated urban ozone tends to increase linearly with
Dobson number decrease; but at the warmer 302K temperature, the simulated ozone
increases more for the first 50-unit Dobson change than for the second 50-unit
change, down to the 200-unit limit of the studies.  Preliminary trajectory model
results suggest that peak ground-based ozone levels are reached earlier in the
day, which would expose larger populations to peak values.   Hydrogen peroxide
increased from approximately 1.7 ppb to 3.3 ppb in Los Angeles and from 0.19 ppb
to approximately 3.0 ppb in Philadelphia for a 100-unit change in Dobson.  These
later changes suggest significant potential impacts for ozone depletion on acid
rain formation.

    The effects of a 4K increase in temperature are most pronounced in the Los
Angeles and Nashville simulations where 14 to 23 ppb increases in simulated
ozone peaks are predicted.   The effects of a combined temperature increase and
reduced Dobson number appear to be additive, if not synergistic,  especially for
the Nashville simulations.   When a 4K temperature increase is combined with only
a 50-unit Dobson reduction, the simulated ozone peak jumps from a base value of
0.13 ppm to over 0.18 ppm ozone.   See Exhibit 14-4.

CONCLUSIONS AND FUTURE RESEARCH DIRECTIONS

    The preliminary study  (Whitten and Gery 1986) was limited in several ways
because the primary focus was on the atmospheric chemistry of ozone formation
(e.g.,  potential changes in acid rain were not addressed).   (Although the
effects of local warming and increased ultraviolet radiation were addressed in
the preliminary study,  increases in air pollution such as ground-based ozone and

-------
                              14-10
                          EXHIBIT 14-3

          Ozone Concentrations Predicted for Changes in
         Dobson Number and Temperature  for Three Cities
                              (ppra)
                              Ozone Concentrations
City
Los Angeles—/
Philadelphia
Nashville

300 DU
0.288
0.112
0.130
298K
250 DU
0.301
0.127
0.161

200 DU
0.315
0.149
0.195

300 DU
0.306
0.122
0.146
302K
250 DU
0.318
0.134
0.184

200 DU
0.331
0.159
0.215
—/ For Los Angeles, values under 298K used actual hourly tempera-
tures, and values under 302K are from simulations using those
temperatures increased by 4K.

-------
                                      14-8
    Whitten began by investigating how the ozone formation processes in the
lower troposphere would be affected by increased UV transmission through the
atmosphere.  From preliminary work, it became evident that the photolysis
channel of formaldehyde that leads to radical products is selectively enhanced
at a higher UV flux.  Formaldehyde emissions are products of incomplete
combustion, and formaldehyde is a major intermediate oxidation product from
virtually all organic molecules.  Radicals from formaldehyde photolysis provide
the main source of radicals needed to drive the chain reactions that generate
photochemical smog (Whitten 1983); therefore, an increase in the rate of
photolysis will have a bearing on the design of future control strategies.

    The photolysis of ozone to electronically-excited oxygen atoms is thought to
be the second most important source of the radicals that drive smog formation.
However, the role of ozone photolysis is different from that of formaldehyde
because ozone is the principal ingredient of photochemical smog.  At low levels
of oxidation potential, the excited oxygen atoms tend to accelerate smog
reactions, making the atmospheric chemistry more efficient in generating ozone
from minimal precursor emissions.  However, at the higher or more severe ozone
levels, the excess radicals can partially suppress the ozone peak, making the
precursors seem less efficient in generating ozone.  Other photolysis rates can
also be affected, but their contribution to smog formation is less important.

    A specific increase in formaldehyde photolysis to radical products is
difficult to determine accurately for a number of reasons.  Atmospheric
photolysis rates have often been calculated in 10-nm wavelength intervals.
However, the accurate calculation of formaldehyde photolysis requires fine
spectral resolution for both the formaldehyde absorption cross-section and the
surface solar flux (related to the ozone absorption cross-section).  Only in
this manner can a comparison of the respective fine structures be performed.
Near-ground solar flux data of high resolution and known stratospheric ozone
abundance have not been readily available.  Hence,  Whitten and Gery (1986)
estimated surface solar fluxes using low-resolution information provided by the
National Aeronautics and Space Administration.

    Using the spectral data,  Whitten and Gery (1986) estimated formaldehyde and
ozone photolysis rates as a function of zenith angle and ultraviolet flux
changes due to stratospheric ozone depletion.  The estimated photolysis rates
for formaldehyde, coupled with the information compiled for surface ozone
photolysis, provide the inputs needed to calculate diurnal photolysis rates for
projected future ozone column densities.   This information is used to evaluate
the future impact of stratospheric ozone changes on near-surface smog formation.

    The preliminary results of the possible effects of increased ultraviolet
radiation on urban smog are based on simulation of atmospheric conditions for
three urban cities:

        (1)  Nashville -- because it is nearly in compliance with the
             0.12 ppm federal ozone standard.

        (2)  Philadelphia --to represent cities that require moderate
             control (30%-50% reduction in organic  precursors)  to
             achieve the 0.12 ppm standard.

-------
                                     14-9
        (3)  Los Angeles -- because of the severity of the exceedance
             of the ozone standard in that region.

Present-day estimates assume a total ozone column of 300 Dobson Units (DU) and
future predictions assume a total ozone column of 250 and 200 DU.   A 33%
decrease in total ozone column would increase ozone photolysis by approximately
a factor of two, and increase formaldehyde photolysis to radical products by
nearly 20%.  Both of these increases are approximations because of uncertainties
regarding the spectral fine structure and other factors.

    The effects of these photolysis rate increases are given in Exhibit 14-3 for
background temperatures of 298 and 302 Kelvin (K).   For all three cities, the
model predicts higher ozone concentrations resulting either from increases in
temperature or decreases in the Dobson number.  The magnitude of the increases
in ozone is apparently a function of local hydrocarbon-to-nitrogen-oxide ratio,
reactivity, meteorology, and emission distribution.  The linearity of the
response to stratospheric ozone depletion is not general.  For the Los Angeles
case, the  increases in ozone are moderate and quite linear with decreasing
Dobson number.  For the Philadelphia case, ozone is predicted to increase
progressively as the Dobson number declines; that is, only a modest increase in
ozone is seen for a decline in Dobson number from 300 to 250 units, but a more
dramatic increase in ozone is seen when the Dobson number declines from 250 to
200 units.  For the Nashville case, the increases are all very dramatic and the
linearity  exists only if temperature does not increase.  At the "current"
temperature of 298K, the simulated urban ozone tends to increase linearly with
Dobson number decrease; but at the warmer 302K temperature, the simulated ozone
increases more for the first 50-unit Dobson change than for the second 50-unit
change, down to the 200-unit limit of the studies.  Preliminary trajectory model
results suggest that peak smog levels are reached earlier in the day, which
would expose larger populations to peak values.  Hydrogen peroxide increased
from approximately 1.7 ppb to 3.3 ppb in Los Angeles and from 0.19 ppb to
approximately 3.0 ppb in Philadelphia for a 100-unit change in Dobson.  These
later changes suggest significant potential impacts for ozone depletion on acid
rain formation.

    The effects of a 4K increase in temperature are most pronounced in the Los
Angeles and Nashville simulations where 14 to 23 ppb increases in simulated
ozone peaks are predicted.  The effects of a combined temperature increase and
reduced Dobson number appear to be additive, if not synergistic, especially for
the Nashville simulations.  When a 4K temperature increase is combined with only
a 50-unit  Dobson reduction, the simulated ozone peak jumps from a base value of
0.13 ppm  to over 0.18 ppm ozone.  See Exhibit 14-4.

CONCLUSIONS AND FUTURE RESEARCH DIRECTIONS

    The preliminary study  (Whitten and Gery 1986) was limited in several ways
because the primary focus was on the atmospheric chemistry of ozone formation
(e.g., potential changes  in acid rain were not addressed).   (Although the
effects of local warming  and  increased ultraviolet radiation were addressed in
the preliminary study,  increases in air pollution  such  as  smog and acid  rain can
only be addressed  through more extensive modeling using more complex atmospheric
models).   However,  the  simple, moving box model  remains applicable in the more
extensive  modeling.

-------
                              14-10
                           EXHIBIT 14-3

          Ozone Concentrations Predicted for Changes in
          Dobson Number and Temperature  for Three  Cities
     City
                              Ozone Concentrations
                          298K
                                  302K
300 DU  250 DU  200 DU   300 DU  250 DU  200 DU
 Los Angeles—/
 Philadelphia
 Nashville
0.288   0.301   0.315
0.112   0.127   0.149
0.130   0.161   0.195
0.306   0.318   0.331
0.122   0.134   0.159
0.146   0.184   0.215
—/ For Los Angeles, values under 298K used actual hourly tempera-
tures, and values under 302K are from simulations using those
temperatures increased by 4K.

-------
                               14-11
                            EXHIBIT 14-4


                Global Warming Would Exacerbate Effects of
                Depletion on Ground-Based Ozone in Nashville
17% Depletion
and
4°C Temperature
Rise
                          41.5%




17.7%
-; 23.8%";
.•••/ ,• * ,- .• / ..•• •
/.'vv'Xv/ v
^ -/I ;'••;'•
•' , . • ', • ' / •'•' ,,
A
>


                        Increase in
                      Ground-Based
                          Ozone
                                         4°C Temperature
                                       > Rise
                                        > 17% Depletion
                                         Only
  Source:  Adapted from Whitten and Gery (1986).

-------
                                     14-12
acid rain can only be addressed through more extensive modeling using more
complex atmospheric models).  However, the simple, moving box model remains
applicable in the more extensive modeling.

    Although the initial study involved the use of only one atmospheric chemical
mechanism, the Carbon Bond Mechanism  (CBM), the key reactions are virtually
identical in all currently-accepted atmospheric mechanisms.  Hence, it can be
expected that other mechanisms would  show the same or similar responses to the
global effects.  Thus, the use of the CBM is not considered a serious limitation
of the current work.  However, the increased radiation effects studied could not
utilize high resolution spectral data even though the preliminary study seemed
to show a need to use such data.

    Thus far, atmospheric modeling has addressed only three trajectories, one
each in three cities.  Three cities does not constitute an adequate sample size
for characterizing the potential range of conditions affected by these global
effects.  Although the model is sensitive to initial and boundary conditions,
these conditions were not varied, even though the hypothesized global effects
would affect such conditions.  The data for initial and boundary conditions were
primarily composed of air from upwind cities or air recirculated from the day
before.  Emissions were not varied to account for local warming, even though
many emissions increase with temperature.  Climate and other meteorological
changes would affect important factors such as mixing height and wind flow
patterns.  These possible changes were also not addressed in the preliminary
study.

    A very important finding of the preliminary study is the increased rate of
smog formation due to increased ultraviolet radiation.  Even if the peak ground-
based ozone concentration is not affected, the increased rate of formation will
move the high concentrations of ground-based ozone closer to the precursor
emissions where the population density is invariably higher.  Hence, the number
of people exposed to high ground-based ozone concentrations would increase.
This exposure effect is not addressed in the preliminary study.   A grid type
model is more appropriate for such exposure studies.

    In summary, the following limitations of the preliminary study are noted:

             o    Effects on pollutants other than urban ozone
                  (tropospheric ozone) were not addressed;
                  acid rain may be affected by global effects.

             o    The chemical effects were calculated using
                  low resolution spectral data;  high
                  resolution data appear to be needed.

             o    Since only three cities were studied,  the
                  range of conditions most affected by global
                  effects could not be elucidated.

             o    Multi-day effects such as  inter-  or
                  intracity carryover of pollutants were not
                  addressed.

-------
                                     14-13
             o    Temperature effects on emission rates or
                  changes in wind patterns,  mixing heights,  or
                  frequency of episodic meteorological
                  conditions were not addressed.

             o    Possible increases in population exposure
                  due to increased smog levels and increased
                  rates of ground-based ozone formation were
                  not addressed.

    Current studies attempt to address these limitations by  (1)  increasing the
number of trajectory cases, (2) using high resolution spectral data,  (3)
studying the range of conditions that can possibly affect carryover of
pollutants, (4) varying the emissions with temperature, (5)  studying climate
change effects, and (6) calculating population exposure changes.   The use of an
acid rain model and a grid model will be incorporated in the on-going
investigations.

-------
                                     14-14-
REFERENCES
Bass AM, LC Glasgow,  C Miller,  JP Jesson, and DL Filken.  1980.  Planet. Space
Sci. 28:675.

Bahe FC, WN Marx,  U Schurath, and EP Roth.  1983.   "Determination of the
Absolute Photolysis Rate of Ozone by Sunlight, 03 + r\v —> O^D) + Q^ ^Ag) ,  at
Ground Level."  Institut fur Physikalische Chemie der Universitat Bonn.   Bonn,
W. Germany.

EPA 1986.  "Review of the National Ambient Air Quality Standards for Ozone:
Preliminary Assessment of Scientific and Technical Information," Office of Air
Quality Planning and Standards Staff Paper.  March 1986.

Whitten GZ  1983.   "The chemistry of smog formation:   A review of current
knowledge."  Environ. International 9:447-463.

Whitten GZ, KR Styles, and MW Gery.  1986.  "Assessment of Existing Tropospheric
UV-Radiation Data and the Effect of its Increase on Ozone Formation in the
Troposphere."  Monthly Technical Narrative No. 2 to U.S. Department of Interior,
National Park Service, Washington, D.C.

Whitten GZ, and MW Gery.  1986. "Effects of Increased UV Radiation on Urban
Ozone," Presented at EPA Workshop on Global Atmospheric Change and EPA Planning.
Edited by Jeffries, H.,  EPA Report # 600/9-86016 July 1986.

-------
                                      CHAPTER 15

                        CAUSES AND EFFECTS OF SEA LEVEL RISE
SUMMARY

    One of the most widely examined impacts of the projected global warming is
the possible rise in sea level.  Researchers have identified at least four
mechanisms that might cause a significant rise:  the warming and resulting
expansion of the upper layers of the ocean, the melting of alpine glaciers, the
melting of polar ice sheets in Greenland and Antarctica, and the disintegration
of these land-based ice sheets.  Estimates of the rise through the year 2100, in
the absence of efforts to limit the greenhouse warming, range from 50 cm to over
2 m.  Even the most conservative estimate implies a substantial acceleration
over the 10-15 cm rise of the last century.

    A rise in sea level in that range would permanently innundate wetlands and
lowlands, accelerate coastal erosion, exacerbate coastal flooding, and increase
the salinity of estuaries and aquifers.  Although wetlands have kept pace with
sea level rise in the last several thousand years, a 1- to 2-m rise would
destroy a majority of U.S. coastal marshes and swamps.  River deltas, such as
those of the Mississippi, Ganges, and Nile rivers, appear to be particularly
vulnerable.

    Along the open coast, beach erosion could reach 1 to 2 m for every 1-cm rise
in sea level, in addition to whatever erosion might be caused by other factors.
Because buildings are generally found within 50 m of the shore, even the 30-cm
rise projected for the next 40 years could threaten coastal property and the
recreational use of beaches,  unless additional remedial measures are
implemented.

    Sea level rise would also increase the vulnerability of coastal areas to
flooding from storm surges and rainwater.  In the area of Charleston, S.C., for
example, the area now flooded once every 100 years would be flooded every 10
years if sea level rises 1.6m.  Protecting against increased flooding would
require improvement or construction of levees, seawalls, and drainage
facilities.

    Higher water levels would also increase the salinity of estuaries and
aquifers.  For example, Philadelphia's drinking water intake on the Delaware
River would be threatened by a 73-cm rise, as would adjacent aquifers in New
Jersey that are recharged by the (currently) fresh part of the river.
Construction of additional reservoirs might be necessary to offset salinity
increases.

    Few studies have estimated the economic significance of future sea level
rise.   One study suggests that the impacts of a 0.9- to 2.4-m rise by 2075 could
be as great as 17% to 35% of total economic activity in the Charleston,  South
Carolina, area and 5% to 16%  of the activity around Galveston,  Texas.  No one
has yet estimated the potential nationwide cost of defending shorelines and
other resources from the projected rise in sea level.

-------
                                     15-2
FINDINGS

1.   THE PROJECTED GLOBAL WARMING WOULD ACCELERATE THE CURRENT RATE OF SEA LEVEL
    RISE BY EXPANDING THE DENSITY OF OCEAN WATER. MELTING ALPINE GLACIERS.  AND
    EVENTUALLY INCREASING THE RATE AT WHICH POLAR ICE SHEETS MELT OR DISCHARGE
    ICE INTO THE OCEANS.

2.   GLOBAL AVERAGE SEA LEVEL APPEARS TO HAVE RISEN 10 TO 15 CM OVER THE LAST
    CENTURY.

    2a.  Studies of the possible contribution of thermal expansion and alpine
         meltwater to sea level rise, based on the 0.6°C warming of the past
         century, indicate that these two sources are insufficient to explain
         the estimated sea level rise that has occurred during this period.
         Consequently, some other source, such as melting of the polar ice caps,
         must be considered a possibility.

3.   ESTIMATES OF THE RISE IN SEA LEVEL THAT COULD TAKE PLACE IF MEASURES TO
    LIMIT THE GLOBAL WARMING ARE NOT UNDERTAKEN RANGE FROM 10 TO 20 CM BY THE
    YEAR 2025. AND 50 TO 200 CM BY 2100.

    3a.  According to published studies, thermal expansion of the oceans alone
         would increase sea level rise between about 30 cm and 100 cm by 2100,
         depending on the realized temperature change.  This is the most certain
         contribution.

    3b.  Melting of alpine glaciers and possibly of ice on Greenland could each
         contribute 10 to 30 cm through 2100, depending on the scenario.  This
         contribution also has a high degree of likelihood.

    3c.  The contribution of Antarctic deglaciation is more difficult to
         project.  It has been estimated at between 0 and 100 cm; however, the
         possibilities cannot be ruled out that  (1) increased snowfall could
         increase the size of the Antarctic ice sheet and thereby partially
         offset part of the sea level rise from other sources; or (2) meltwater
         and enhanced calving of the ice sheet could increase the contribution
         of Antarctic deglaciation to as much as 2 m.  The Antarctic
         contribution to sea level rise may be more sensitive to time delays
         after certain threshold conditions are reached than to the magnitude of
         total warming.

4.  OVER THE MUCH LONGER TERM (THE NEXT FEW CENTURIES) DISINTEGRATION OF THE
    WEST ANTARCTIC ICE SHEET MIGHT RAISE SEA LEVEL BY 6 METERS.

    4a.  If a disintegration takes place, glaciologists generally believe  that
         such a complete disintegration of the west Antarctic ice sheet would
         take at least  300 years, and probably at least 500 years.

    4b.    A global warming might result  in sufficient thinning of  the Ross and
           Filcher-Ronne Ice Shelves in  the next century  to make the process of
           disintegration irreversible.

-------
                                     15-3
5.   LOCAL TRENDS IN SUBSIDENCE AND EMERGENCE MUST BE ADDED OR SUBTRACTED TO
     GLOBAL RISK ESTIMATES IN ORDER TO ESTIMATE RELATIVE SEA LEVEL RISE AT
     PARTICULAR LOCATIONS.

     5a.   Most of the Atlantic and Gulf Coasts of the United States--as well as
           the Southern Pacific coast--are subsiding 10-20 cm per century.

     5b.   Louisiana is subsiding 1 m per century, while parts of Alaska are
           emerging 10-150 cm per century.

     5c.   Due to subsidence already occurring in areas such as Bangladesh,
           Bangkok, and the Nile delta,  these areas are extremely vulnerable to
           sea level rise.

6.   A SUBSTANTIAL RISE IN SEA LEVEL WOULD PERMANENTLY INUNDATE WETLANDS AND
     LOWLANDS. ACCELERATE COASTAL EROSION. EXACERBATE COASTAL FLOODING. AND
     INCREASE THE SALINITY OF ESTUARIES  AND AQUIFERS.

     6a.   Louisiana is the state most vulnerable to a rise in sea level.
           Important impacts would also  occur in Florida, Maryland,  Delaware,
           New Jersey, and in the coastal regions of other states.

     6b.   A rise in sea level of 1 to 2 m by the year 2100 could destroy 50
           percent to 80 percent of U.S. coastal wetlands.

     6c.   Limited studies predict that  increased salinity from sea level rise
           would convert cypress swamps  to open water and threaten drinking
           water supplies in areas such  as Louisiana, Philadelphia,  and New
           Jersey.  Other areas, such as Southern Florida, may also be
           vulnerable but have not been  investigated.

     6d.   Studies of Bangladesh and the Nile River Delta indicate that these
           river deltas, which are already subsiding, would be greatly affected
           by rising sea level, experiencing significant economic and
           environmental losses.

7.   EROSION PROJECTED IN VARIOUS STUDIES TO RESULT FROM ACCELERATED SEA LEVEL
     RISE COULD THREATEN U.S.  RECREATIONAL BEACHES

     la.   Case studies of beaches in New Jersey, Maryland,  California, South
           Carolina,  and Florida have concluded that a 30-cm rise in sea level
           would result in beaches eroding 20-60 m or more.   Major beach
           preservation efforts would be required if recreational beaches are to
           be maintained.

8.   ACCELERATED SEA LEVEL RISE WOULD INCREASE THE DAMAGES FROM FLOODING IN
     COASTAL AREAS

     8a.   Flood damages would increase  because higher water levels  would
           provide a higher base for storm surges.

     8b.   Erosion would increase the vulnerability to storm waves,  and
           decreased natural and artificial drainage would increase  flooding
           during rainstorms.

-------
                                     15-4
9.    ESTIMATES OF DAMAGE FROM SEA LEVEL RISE MUST CONSIDER POSSIBLE MITIGATION
     BY HUMAN RESPONSES

     9a.   The adverse impacts of sea level rise could be ameliorated through
           anticipatory land use planning and structural design changes.

     9b.   In a case study of two cities, Charleston,  South Carolina, and
           Galveston, Texas, accelerated anticipatory planning was estimated to
           reduce net damages by 20 to 60 percent.

10.  RELATED IMPACTS OF A GLOBAL WARMING WOULD ALSO AFFECT IMPACTS OF SEA LEVEL
     RISE

     lOa.  Increased droughts might amplify the salinity impacts of sea level
           rise.

     lOb.  Increased hurricanes and increased rainfall in coastal areas could
           amplify flooding from sea level rise.

     lOc.  Warmer temperatures might impair peat formation of salt marshes and
           would enable mangrove swamps to take over areas that are presently
           salt marsh.

     lOd.  Decreased northeasters might reduce damage.

11.  RESEARCH OPPORTUNITIES EXIST TO IMPROVE SEA LEVEL RISE ESTIMATES AND
     IMPACTS

     lla.  The most critical areas of research for reducing the variation in
           estimates of future sea level rise are ice melting and runoff in
           Antarctica and Greenland and ice discharge.

     lib.  Research in glacial discharge in Antarctica should focus not just on
           West Antarctica, but on Pine Island and East Antarctica.

     lie.  An improved program of tidal gauge stations, especially in the
           southern hemisphere, and satellite altimetry should be used to
           measure sea level rise and the mass balance of ice sheets.

-------
                                      15-5
CAUSES OF SEA LEVEL RISE

    The worldwide average sea level depends primarily on (1) the shape and size
of ocean basins, (2) the amount of water in the oceans, and (3) the average
density of seawater.  Only the latter two factors are influenced by climate.
Subsidence and emergence caused by natural factors such as isostatic and
tectonic adjustments of the land surface, as well as human-induced factors such
as oil and water extraction, can cause trends in relative sea level at
particular locations to differ from trends in global sea level.

Past Trends in Sea Level

    Hays and Pitman (1973) analyzed fossil records and concluded that over the
last 100 million years, changes in mid-ocean ridge systems have caused the sea
level to rise and fall over 300 m.  However, Clark, Farrell, and Peltier (1978)
have pointed out that these changes have accounted for sea level changes of less
than 1 mm per century.  No published study has indicated that this determinant
of sea level is likely to have a significant impact in the next century.

    The impact of climate on sea level has been more pronounced.  During ice
ages the glaciation of substantial portions of the northern hemisphere has
removed enough water from the oceans to lower sea level 100 m below the present
level during the last (18,000 years ago) and earlier ice ages (Donn, Farrand,
and Ewing 1962; Kennett 1982; Oldale 1985).

    Although the glaciers that once covered much of the northern hemisphere have
retreated, the world's remaining ice cover contains enough water to raise sea
level over 75 m (Hollin and Barry 1979).  As Exhibit 15-1 shows, Hollin and
Barry (1979) and Flint (1971) estimate that existing alpine glaciers contain
enough water to raise sea level 30 cm or 60 cm, respectively.  The Greenland and
West Antarctic Ice Sheets each contain enough water to raise sea level about 8
m, and East Antarctica has enough ice to raise sea level over 60 m.

    There is no evidence that either the Greenland or East Antarctic Ice Sheets
have completely disintegrated in the last two million years.  However, it is
generally recognized that sea level was about 7 m higher than today during the
last interglacial (Moore 1982; Mercer 1968; Hollin 1962), which was l°-2° warmer
than today.  Because the West Antarctic Ice Sheet is marine-based and thought by
some to be vulnerable to climatic warming, attention has focused on this source
for the higher sea level.  Mercer (1968) found that lake sediments and other
evidence suggested that summer temperatures in Antarctica have been 7°-10°C
higher than today at some point in the last two million years,  probably during
the last interglacial 125,000 years ago, and that such temperatures could have
caused a disintegration of the West Antarctic Ice Sheet.  However, others are
not certain that marine-based glaciers are more vulnerable to climate change
than land-based glaciers,  Robin (1985) suggests that the higher sea level
during the last interglacial period may have resulted from changes in the East
Antarctic Ice Sheet.

    Tidal gauges have been available to measure the change in relative sea level
at particular locations over the last century.    Studies combining these
measurements to estimate global trends have concluded that sea level has risen
10-15 cm during the last century  (Barnett 1984; Gornitz, Lebedeff, and Hansen
1982; Fairbridge and Krebs 1962;  Hicks, Debaugh, and Hickman 1983 for the U.S.

-------
                                      15-6
                                 EXHIBIT  15-1

                            Snow and Ice Components
                                                      Ice
                                         Area        Volume
                                       (106 km2)    (106 km3)
                                                              Sea Level
                                                             Equivalent a/
                                                                 (m)
Land ice:
  East Antarctica b/
  West Antarctica c/
  Greenland
  Small ice caps and mountain
    glaciers d/

Permafrost (excluding Antarctica)
  Continuous

  Discontinuous
Sea ice:
     Arctic e/
       Late February
       Late August

     Antarctic f/
       September
       February
Land snow cover g/
  N. Hemisphere
     Early February
     Late August
  S.
Hemisphere
Late July
Early May
                                     9.86
                                     2.34
                                     1.7
                                     0.54
                                     7.6

                                    17.3
                                         14.0
                                          7.0
                                         18.4
                                          3.6
                                    46.3
                                     3.7
                                          0.85
                                          0.07
25.92
 3.40
 3.0
 0.12
 0.03
  to
 0.7
 0.05
 0.02
 0.06
 0.01
 0.002
64.8
 8.5
 7.6
 0.3
 0.6
 0.08
  to
 0.17
a/  400,000 km  of ice is equivalent to 1 m global sea level.
b/  Grounded ice sheet,  excluding peripheral, floating ice shelves (which do not
    affect sea level).  The shelves have a total area of 1.62 x 10° knr and a
    volume of 0.79 x 106 km3 (Dewey and Heim 1981).
c/  Including the Antarctic Peninsula.
d/  Flint (1971); Hollin and Barry (1979)
e/  Excluding the Sea of Okhotsk, the Baltic Sea, and the Gulf of St. Lawrence
    (Walsh and Johnson 1979).  Maximum ice extents in these areas are 0.7
                                            r\
    million, 0.4 million, and 0.2 million knr,  respectively.
f/  Actual ice area excluding open water (Zwally, Parkinson, and Comiso 1983).
    Ice extent ranges between 4 million and 20 million km^.
£/  Snow cover includes  land ice but excludes snow-covered sea ice (Dewey and
    Heim 1981).
Source:  Meier et al. 1985, p. 242.  Modified from Hollin and Barry 1979.

-------
                                      15-7
Coast);  Exhibit 15-2 shows the sea level curve estimated by Gornitz,  Lebedeff,
and Hansen.  Barnett (1984) found that the rate of sea level rise over the last
50 years had been about 2.0 mm/year, while in the previous 50 years there had
been little change; however, the acceleration in the rate of sea level rise was
not statistically significant.  Emery and Aubrey (1985 and 1986) have accounted
for estimated land surface movements in their analyses of tidal gauge records in
Northern Europe and western North America and have found an acceleration in the
rate of sea level rise over the last century.   Braatz and Aubrey (1987) have
found that the rate of relative sea level rise on the east coast of North
America accelerated after 1934.

    Several researchers have investigated possible causes of current trends in
sea level.  Barnett (1984) and Gornitz, Lebedeff, and Hansen (1982) estimate
that thermal expansion of the upper layers of the oceans resulting from the
observed global warming of 0.4°C in the last century could be responsible for a
rise of 0.4-0.5 mm/year.  Recent analysis suggests that the warming has been
0.6°C, which would account for thermal expansion of 0.6-0.7 mm/year (Jones,
Wigley, and Wright 1986).  Roemmich and Wunsch (1984) examined temperature and
salinity measurements at Bermuda and concluded that the 4°C isotherm had
migrated 100 m downward, and that the resulting expansion of ocean water could
be responsible for some or all of the observed rise in relative sea level.
Roemmich (1985) showed that the warming trend 700 m below the surface was
statistically significant.  However, Barnett (1984) found no significant trend
based on an examination of the upper layers of the ocean.  Nevertheless, Braatz
and Aubrey (n.d.) note that long-term steric changes in the ocean are not
confined to the upper layers of the oceans, implying that the Barnett analysis
does not necessarily contradict the Roemmich and Wunsch conclusion.  Because
one-dimensional models that cannot capture the complexities of the ocean have
been used to estimate the past expansion of ocean water, these authors all
caution that their results should be viewed as first approximations.

    Meier (1984) estimates that retreat of alpine glaciers and small ice caps
could currently contribute between 0.2 and 0.72 mm/year to sea level.   The
National Academy of Sciences Polar Research Board (Meier et al. 1985)  concluded
that existing information is insufficient to determine whether the impacts of
Greenland and Antarctica are positive or zero.  Thus, thermal expansion and
alpine melting would explain a rise of 0.8-1.4 mm/yr, compared with the observed
rise of 1.0-1.5 mm/yr.   Although the estimated global warming of the last
century appears at least partly responsible for the last century's rise in sea
level, no study has demonstrated that global warming might be responsible for an
acceleration in the rate of sea level rise.

Impact of Future Global Warming on Sea Level

    Concern about a substantial rise in sea level as a result of the projected
global warming stemmed originally from Mercer (1968), who suggested that the
Ross and Filchner-Ronne ice shelves might disintegrate, causing a deglaciation
of the West Antarctic Ice Sheet and a resulting 6- to 7-m rise in sea level,
possibly over a period as short as 40 years (see also Mercer 1978) .

    Subsequent investigations have concluded that such a rapid rise is unlikely.
Hughes (1983) and Bentley (1983) estimated that such a disintegration would take

-------
                                      15-8



                                  EXHIBIT 15-2


                     Worldwide Sea Level  in the Last Century
            10
Sea  Level
   (cm)
            -5
             1880
  I
1920
1960
                                       Year
  Sources:   Gornitz, Lebedeff, and Hansen  1982.

-------
                                     15-9
at least 200 or 500 years,  respectively.   Other researchers have estimated that
this process would take considerably longer (Fastook 1985;  Lingle 1985).

    Researchers have turned their attention to the magnitude of sea level rise
that might occur in the next century.  The best-understood factors are the
thermal expansion of ocean water and the  melting of alpine glaciers.   In the
National Academy of Sciences report Changing Climate.  Revelle (1983)  used the
model of Cess and Goldenberg (1981) to estimate temperature increases at various
depths and latitudes resulting from a 4.2°C warming by the years 2050-2060
(Exhibit 15-3).  Although his assumed time constant of 33 years probably
resulted in a conservatively low estimate, he estimated that thermal  expansion
would result in an expansion of the upper ocean sufficient to raise sea level 30
cm.

    Using a model of the oceans developed by Lacis et al. (1981), Hoffman,
Wells, and Titus (1986) examined a variety of possible scenarios of future
emissions of greenhouse gases and global  warming.   They estimated that a warming
of between 1° and 2.6°C could result in a thermal expansion contribution to sea
level of between 12 and 26 cm by the year 2050.  They also estimated that a
global warming of 2.3°-7.0°C by 2100 would result in a thermal expansion of
28-83 cm.

    Revelle (1983) suggested that although he could not estimate the  future
contribution of alpine glaciers to sea level rise, a contribution of  12 cm
through 2080 would be reasonable.  Meier  (1984) used glacier balance  and volume
change data for 25 glaciers for which the available record exceeded 50 years to
estimate the relationship between historical temperature increases and the
resulting negative mass balances of the glaciers.   He estimated that  a 28-mm
rise had resulted from a warming of 0.5°C and concluded that a 1.5°-4.5°C
warming would result in a rise of 8-25 cm in the next century.  Using these
results, Meier et al.  (1985) concluded that the contribution of glaciers and
small ice caps to a rise in sea level through 2100 is likely to be 10-30 cm.
They noted that the gradual depletion of  remaining ice cover might reduce the
contribution of sea level rise somewhat.   However, the contribution might also
be greater because the historical rise took place over a 60-year period, while
the forecast period is over 100 years. Using Meier's estimated relationship
between global warming and the alpine contribution, Hoffman, Wells, and Titus
(1986) estimated alpine contributions through the year 2100 at 12-38  cm for a
global warming of 2.3°-7.0°C.

    The first published estimate of the contribution of Greenland meltwater to
future sea level rise was Revelle's (1983) estimate of 12 cm through  the year
2080.  Using estimates by Ambach (1980 and 1982) that the equilibrium line
(between snowfall accumulation and melting) rises 100 m for each 0.6°C rise in
air temperature, Revelle concluded that the projected 6°C warming would be
likely to raise the equilibrium line 1000 m.  He estimated that such  a change in
the equilibrium line would result in a 12-cm contribution to sea level rise for
the next century.

    Meier et al. (1985) noted that the large ablation area makes Greenland a
"significant potential contributor of meltwater to the ocean if climatic warming
causes an increase in the rate of ablation and an upward shift of the

-------
                                     15-10


                                  EXHIBIT  15-3

              Temperature  Increase At Various Depths  and Latitudes
                                   LATITUDE
               40°N      20°N
20°S     4(?S     6cPs      SOPS
u


200



6
^
0.
UJ
Q


1OOO

7
s.a

"
I
I
4.0
I
J
___

1.8


rs~
J4.a




2.8

3

1.3

I
— sJ
4.2




2.3



1.3


3.5
2.7




1.6



0.5


_2_
1.8




0.9



0.3


2.5
2.0




1. 1



0.4

1
L^n
2.7




1.6



0.5
-

5
3.9




2.2



0.7


4.5
3.5




k



0.6


4
3.4




2.3



1.0



3




2



1.


I
.4




'I



0
|


3.S
2.9




_2.oJ



0.9


3 2
2.5 1.7




1.7 1.1



0.8 0.5

Computed average increase in ocean temperatures at particular depths and
latitudes for a doubling of atmospheric carbon dioxide and probable increase in
other greenhouse gases by the year 2080, based on Flohn's (1982) prognosis.
Solid lines are isopleths for temperature changes.  Note that the top 100 m
warms 2°-7°, while at depths of 800-1000 m the warming is mostly less than 1°.
Source:  Revelle (1983).

-------
                                     15-11
equilibrium line."   They found that a 1000-m rise in the equilibrium line would
result in a contribution of 30 cm through the year 2100.  However, because
Ambach (1985) found the relationship between the equilibrium line and
temperature to be 77 m per degree (C),  the panel concluded that a 500-m shift in
the equilibrium line would be more likely.  Based on the assumption that
Greenland will warm 6.5°C by the year 2050 and that temperatures will remain
constant thereafter, they estimated that such a change would contribute about 10
cm to sea level through 2100, but also noted that "for an extreme but highly
unlikely case, with the equilibrium line raised 1000 m, the total rise would be
26 centimeters."  Bindschadler (1985) treated the two cases as equally
plausible.  However, his analysis was conducted before the results of Ambach
(1985) were known and he has since indicated agreement with the findings of
Meier et al.   (1985).4

    Available estimates of the Greenland contribution assume that all meltwater
flows into the oceans and that the ice dynamics of the glaciers do not change.
Meier et al.   (1985) suggested that much of the water would refreeze and
therefore decrease the contribution to sea level rise.  Although a change in ice
dynamics might imply additional deglaciation and eventually increase the rate of
sea level rise, they concluded that such changes were unlikely to occur in the
next century.

    The potential impact of a global warming on Antarctica in the next century
is the least certain of all the factors by which a global warming might
contribute to sea level rise.  Meltwater from East Antarctica might make a
significant contribution by the year 2100.  This source could be significant
since global warming is expected to be amplified in polar regions, and large
parts of East Antarctica are far enough from the poles to possibly warm up such
that meltwater runoff would be significant.   Unfortunately, no one has estimated
the likely contribution.   Several studies have examined "deglaciation," which
also includes the contribution of ice sliding into the oceans.   Bentley (1983)
examined the processes by which a deglaciation of West Antarctica might occur.
The first step in the process would be accelerated melting of the undersides of
the Ross and Filchner-Ronne ice shelves caused by warmer water circulating
underneath them.  The thinning of these ice shelves could cause them to become
unpinned and cause their grounding lines to retreat.   Revelle (1983) concluded
that the available literature suggests that the ice shelves might disappear in
100 years; the Antarctic ice streams would then flow directly into the oceans,
without the back pressure of the ice shelves.   Hughes (1983)  and Bentley (1983)
estimate that a complete disintegration of the West Antarctic Ice Sheet could
take an additional 200 and 500 years, respectively.

    Although a West Antarctic deglaciation would occur over a period of
centuries, it is possible that an irreversible deglaciation could commence
before 2050.   Thomas,  Sanderson,  and Rose (1979) suggested that if the ice
shelves thinned more than about 1 m per year,  the ice would move into the sea at
a sufficient speed such that even a cooling back to the temperatures of today
would not be sufficient to result in a reformation of the ice shelf.

    To estimate the likely Antarctic contribution for the next  century,  Thomas
(1985) developed four scenarios of the  impact of a 3°C global warming by the
year 2050:

-------
                                     15-12
         (1)  For a shelf melting rate of 1 m/year with seaward ice
              fronts remaining at present locations--implies a rise of
              28 cm by the year 2100.

         (2)  For a shelf melting rate of 1 m/year with ice fronts
              calving back to a line linking the areas where the shelf
              is grounded, during the  2050s--implies a rise of 1.6 m by
              2100.

         (3)  For a scenario similar to case 1 but with a melt rate of
              3 m/year--implies a rise of 1 m by 2100.

         (4)  For a scenario similar to case 2 but with a melt rate of
              3 m/year--implies a rise of 2.2 meters by 2100.

    Thomas concluded that the 28-cm rise implied by case 1 would be most likely
to occur.  He also stated that even if enhanced calving did occur, it would be
likely to occur after 2050, "suggesting that associated sea-level rise would
probably be closer to the 1 m of case  3 than the 2.2 m of case 4."

    Meier et al. (1985) evaluated the  Thomas study and papers by Lingle (1985)
and Fastook (1985).  Although Lingle estimated that the contribution to sea
level rise of West Antarctica through  2100 would be 3 to 5 cm, he did not
evaluate East Antarctica; Fastook made no estimate for the year 2100.  Thus, the
panel concluded that "imposing reasonable limits" on the Thomas model yields a
range of 20 to 80 cm by the year 2100  for the Antarctic contribution.  However,
they also noted several factors that would reduce the amount of ice discharged
into the sea:  removal of the warmest  ice from the ice shelves, retreat of
grounding lines, and increased lateral shear stress.  They also concluded that
increased precipitation over Antarctica might increase the size of the polar ice
sheets there.  Thus, they concluded that the contribution from Antarctica could
cause a rise in sea level up to 1 m or a drop of 10 cm; a rise of between 0 and
30 cm is most likely.

    Exhibit 15-4 summarizes the various estimates of future global sea level
rise for specific years.   Revelle (1983) estimated that the rise was likely to
be 70 cm, ignoring the impact of a global warming on Antarctica.  He also noted
that the latter contribution was likely to be 1 to 2 m per century after 2050,
but declined to add that to his estimate.  Meier et al. (1985) projected that
the contribution of glaciers would be sufficient to raise sea level 20 to 160
cm, with a rise of "several tenths of a meter" most likely.  Thus, if one
extrapolates from  the Revelle (1983) estimate of thermal expansion through the
year 2100, the Meier et al. report predicts a rise of between 50 and 200 cm.
Using a range of estimates for future concentrations of greenhouse gases, the
climate's sensitivity to such increases, oceanic heat uptake, and the behavior
of glaciers, Hoffman, Wells, and Titus (1986) estimated that the rise would be
10 to 21 cm by 2025 and 57 to 368 cm by 2100.

Future Trends in Local Sea Level

    Although most  attention has focused on projections of global sea level,
impacts on particular areas would depend on local relative sea level.  As
Exhibit 15-5 shows, tidal gauge measurements suggest that relative sea level
rises ten to twenty centimeters per century more rapidly along much of the U.S.

-------
                                     15-13
                                  EXHIBIT 15-4

                       Estimates  of Future Sea Level Rise
                                  (centimeters)
Year 2100 by Cause (Year 2085 for Revelle 1983):
Thermal Alpine
Expansion Glaciers Greenland Antarctica
Revelle (1983)
Hoffman et al. (1983)
Meier et al . (1985) c/
Thomas (1985)
Hoffman et al . (1986)
30 12 12 2
28-115 b/ b/ b/
10-30 10-30 -10-+100
0-200
28-83 12-37 6-27 12-220
Total
70
56-345
50-200
57-368
Total Rise in Specific Years: d/

Hoffman et al . (1983)
Low
Mid- range low
Mid- range high
High
Hoffman et al . 1986
Low
High
2000 2025 2050 2075 2085
4.8 13 23 38
8.8 26 53 91
13.2 39 79 137
17.1 55 117 212

3.5 10 20 36 44
5.5 21 55 191 258
2100
56.0
144.4
216.6
345

57
368
       a/ Revelle attributes 16 cm to other factors.

       b/ Hoffman et al.  (1983) assumed that the glacial contribution would be
   one to two times the contribution of thermal expansion.

       c_/ The Meier et al.  (1985) estimate includes extrapolation of thermal
   expansion from Revelle (1983) .

       d/ Only Hoffman et al. reports made year-by-year projections for the
   next century.

   Sources:  Hoffman et al.   (1986); Meier et al. (1985);
             Hoffman et al.   (1983): Revelle (1983); Thomas (1985).

-------
                                     15-14


                                 EXHIBIT  15-5

                             Local Sea Level Rise
Location
Portland, Maine
Boston, Massachusetts
Newport, Rhode Island
New London, Connecticut
New York, New York
Sandy Hook, New Jersey
Atlantic City, New Jersey
Philadelphia, Pennsylvania
Baltimore , Maryland
Annapolis, Maryland
Hampton Roads , Virginia
Charleston, South Carolina
Fernandina, Florida
Miami Beach, Florida
Cedar Key, Florida
Pensacola, Florida
Eugene Island, Louisiana
Calves ton, Texas
San Diego, California
Los Angeles, California
San Francisco, California
Astoria, Oregon
Seattle , Washington
Juneau, Alaska
Sitka, Alaska
Worldwide
Historical
Relative Sea
Level Trend
(mm/yr)
2
2
2
2
2
4
4
2
3
3
4
3
1
2
2
2
10
6
1
0
1
-0
1
-12
-2
1
.3
.3
.6
.2
.8
.2
.0
.6
.2
.7
.3
.4
.7
.3
.0
.4
.0
.3
.9
.6
.2
.5
.9
.9
.4
.2
Historical
Subsidence
Rate
(mm/yr)
1
1
1
1
1
3
2
1
2
2
3
2
0
1
0
1
8
5
0
-0
0
-1
0
-14
-3
0
.1
.1
.4
.0
.6
.0
.8
.4
.0
.5
.1
.2
.5
.1
.8
.2
.8
.1
.7
.6

.7
.7
.1
.6

c/
Local Rise for
One -Meter Global
Rise by 2100
(cm)
112
112
116
111
118
134
132
116
123
128
135
125
105
112
109
113
201
158
108
93
100
80
108
-62
58
100
.6
.6
.1
.5
.4
.5
.2
.1
.0
.8
.7
.3
.8
.6
.2
.8
.2
.7
.0
.1
.0
.5
.5
.2
.6
.0
Source:
Based on assumed global rise of 12 cm/century.   From Hicks, Debaugh,
and Hickman (1983).

Relative sea level trend minus consensus estimate of 12 cm/century for
global sea level rise.

One meter plus extrapolation of historical subsidence rate.

Hicks, Debaugh, and Hickman 1983.

-------
                                     15-15
coast than the worldwide average (Hicks, Debaugh, and Hickman 1983).  Important
exceptions include Louisiana, which is subsiding close to 1 m per century, and
Alaska, which is emerging 10 to 100 cm per century.

    Local subsidence and emergence are caused by a variety of factors.  Rebound
from the retreat of glaciers after the last ice age has resulted in the uplift
of northern Canada, New England, and parts of Scandinavia; emergence in Alaska
is due more to tectonic adjustments.  The uplift in polar latitudes has
resulted in subsidence in other areas, notably the U.S. Atlantic and Gulf
coasts.  Groundwater pumping has caused rapid subsidence around Houston, Texas;
Taipei, Taiwan; and Bangkok, Thailand, among other areas (Leatherman 1984).
River deltas and other newly created lands subside as the unconsolidated
materials compact.  Although subsidence and emergence trends may change in the
future, particularly where anthropogenic causes are curtailed, no one has
linked these causes to future climate change in the next century.

    However, the removal of ice from Greenland and Antarctica would deform the
ocean floor.  Clark and Lingle  (1977) have calculated the impact of a uniform 1
m contribution from West Antarctica.  They concluded that relative sea level at
Hawaii would rise 125 cm, and that along much of the U.S. Atlantic and Gulf
Coasts, the rise would be 15 cm.  On the other hand, sea level would drop at
Cape Horn by close to 10 cm, and the rise along the southern half of the
Argentine and Chilean coasts would be less than 75 cm.

    Other contributors to local sea level that might change as a result of a
global warming include currents, winds,  and freshwater flow into estuaries.
None of these impacts, however, has been estimated.

EFFECTS OF SEA LEVEL RISE

    A rise in sea level of 1 or 2 m would permanently inundate wetlands and
lowlands, accelerate coastal erosion, exacerbate coastal flooding, threaten
coastal structures, and increase the salinity of estuaries and aquifers.
Substantial research has been done on the implications of sea level rise for
coastal erosion and wetlands and relatively little work has been done in the
other areas.

    The remaining sections of this chapter summarize research on the impacts of
sea level rise.  The level of damage would depend in large measure on when and
how people respond.  Much of the research has been conducted to assist agencies
that might have to respond to the impacts of sea level rise, and has attempted
to estimate impacts for three alternative responses: (1) no countermeasures are
taken; (2) measures are taken in response to sea level rise as it occurs; and
(3) sea level rise is planned for in advance of its occurrence.  We follow this
convention and examine potential impacts, possible responses, and the relative
merits of anticipating a rise in sea level.

Submergence of Coastal Wetlands

    The most direct impact of a rise in sea level is the inundation of areas
that had been just above the water level before the sea rose.  Coastal wetlands
are generally found at elevations below the highest tide of the year and above
mean sea level.  Thus, wetlands account for most of the land less than 1 m above
sea level.

-------
                                     15-16
    Because a common means of estimating past sea level rise has been the
analysis of marsh peats,  the impact of sea level rise on wetlands is fairly well
understood.  For the last several thousand years, marshes have generally kept
pace with the rate of sea level rise through sedimentation and peat formation
(Emery and Uchupi 1972;  Redfield 1972 and 1967; Davis 1985).  As sea level rose,
new wetlands formed inland while the seaward boundary was maintained.  Because
the wetland area has expanded,  Titus, Henderson, and Teal (1984) hypothesized
that one would expect a concave marsh profile (i.e.,  there is more marsh area
than area found immediately above the marsh).  Thus,  if sea level rose more
rapidly than the marsh's ability to keep pace,  there would be a net loss of
wetlands.  Moreover, a complete loss might occur if protection of developed
areas prevented the inland formation of new wetlands (see Exhibit 15-6).

    Kana et al. (1986 and 1987) surveyed marsh transects in Charleston, South
Carolina, and in two sites near Long Beach Island, New Jersey, to evaluate the
concavity of wetland profiles and the vulnerability of wetlands to a rise in sea
level.  Their data showed that in the Charleston area,  all of the marsh was
between 30 and 110 cm above current sea level,  an elevation range of 80 cm (see
Exhibit 15-7).  The area with a similar elevation range just above the marsh was
only 20% as large.  Thus, a rise in sea level exceeding vertical marsh accretion
by 80 cm would result in an 80% loss of wetlands.  In the New Jersey sites, the
marsh was also found within an elevation range of 80 cm; a rise in sea level 80
cm in excess of marsh accretion would result in 67%-90% losses.

    The future ability of marshes to accrete vertically is uncertain.  Based on
field studies by Ward and Domeracki (1978), Hatton,  DeLaune, and Patrick (1983),
Meyerson (1972), Stearns and MacCreary (1957),  Kana,  Baca, and Williams (1986)
concluded that current vertical accretion rates are approximately 4-6 mm/year in
the two case study areas, greater than the current rate of sea level rise but
less than the rates of rise projected for the next century.  If current
accretion trends continue, rises of 87 and 160 cm by 2075 would imply 50% and
80% losses of wetlands,  respectively, in the Charleston area.  Kana et al.
(1987) also estimated 80% losses in the New Jersey sites for a 160-cm rise
through 2075.  However,  because the high marsh dominates in that area, they
concluded that the principal impact of an 87-cm rise by 2075 would be the
conversion of high to low marsh.

    In both cases, the losses of marsh could be greater if inland areas are
developed and protected with bulkheads or levees.  Because there is a buffer
zone between developed areas and the marsh in South Carolina, protecting
development from a 160-cm rise would increase the loss from 80% to 90%.  Without
the buffer, the loss would be close to 100%.

    The marshes and swamps of Louisiana, which account for 40% of the coastal
wetlands in the United States  (excluding Alaska), would be particularly
vulnerable to an accelerated rise in sea level.  The wetlands  in this state are
mostly less than 1 m above sea level, and are generally subsiding approximately
1 m per century as  its deltaic sediments compact  (Boesch 1982).  Until  the last
century, the wetlands were able to keep pace with this rate of relative sea
level rise because of the sediment conveyed  to  the wetlands by the Mississippi
River.

-------
                                                 15-17


                                             EXHIBIT 15-6

                                Evolution of Harsh as  Sea Level Rises
               5000 Years Ago
                  Today
                                   •$•• L«v«4
                                              Sedimentation ana
                                              Peat Formation
                                                                                          Current
                                                                                          S«a Level
                                            Future
Substantial Wetland Loss Where There is Vacant Upland
Complete  Wetland Loss Where House is Prorecrec?
       in Response to Rise m Sea  Level
                                        Future
                                       Sea
                                      " Current
                                      Sea Level
                                Future
                               . Sea Level
                                Current
                                Sea Level
          Coastal  marshes have kept pace with  the slow rate of  sea  level rise that has
          characterized the last several thousand years.  Thus,  the area of marsh has
          expanded over time as new lands were inundated.  If in the future, sea level
          rises  faster than the ability of  the marsh to keep pace,  the marsh area will
          contract.   Construction of bulkheads to protect economic  development may prevent
          new marsh from forming and result in a total loss of  marsh in some areas.

          Source:  Titus 1986.

-------
                                         15-18


                                      EXHIBIT 15-7

                        Composite Transect -- Charleston, S.C.
HIGHLAND 47%
                                COMPOSITE TRANSECT-
                                  CHARLESTON. S.C.
                                                                  WATER }7\
                                                                        	 PCAK VE1RIY TIOC

                                                                        	 SPRING MIGM WATER
                                                                        	 MCAN HIGH WATER
                                                                        	 MEAP MICH W
                                2OOO          3000
                              TYPICAL DISTANCE (FT)
4OOO
             5OOO
  Composite wetlands transect for Charleston area.  Area above the  marsh is much
  steeper than  the marsh.

  Source:  Kana,  Baca, and  William 1986.

-------
                                     15-19
    Human activities, however, have largely disabled the natural processes by
which coastal Louisiana might keep pace with sea level rise.  Dams, navigation
channels, canals, and flood protection levees have interrupted the flow of
sediment, freshwater, and nutrients to the wetlands.   As a result, over 100
square kilometers of wetlands convert to open water every year (Gagliano,
Meyer-Arendt, and Wicker 1981).  A substantial rise in sea level would further
accelerate the process of wetland loss in Louisiana (see Exhibit 15-8).

    Throughout the world, people have dammed, leveed,  and channelized major
rivers, curtailing the amount of sediment that reaches river deltas.  Even at
today's rate of sea level rise, substantial amounts of land are converting to
open water in Egypt and Mexico (Milliman and Meade 1983).  Other deltas, such as
the Ganges in Bangladesh and India, are currently expanding seaward.  These
areas would require increased sediment, however, to keep pace with an
accelerated rise in sea level.  Additional projects to divert the natural flow
of river water would increase the vulnerability of these areas to a rise in sea
level.

    Broadhus et al.  (1986) examined the possible impacts of future sea level
rise on Egypt and Bangladesh, the inhabited areas of which are in the deltas of
the Nile and Ganges Rivers.  They estimated that 50-  and 200-cm rises by 2100
would flood 0.35% and 0.7%, respectively, of Egypt's land area.  However,
because the nation's population is concentrated in the low-lying areas,  16% and
21% of the people currently reside in the areas that would be lost.  Broadhus et
al. (1986) also estimate that Bangladesh could lose 12% and 28% of its land,
which currently houses 9% and 27% of its population (see Exhibit 15-9).

    To develop an understanding of the potential nationwide impact of sea level
rise on coastal wetlands in the United States,  Park,  Armentano, and Cloonan
(1986)  used topographic maps to characterize wetland elevations at 52 sites
comprising 4800 square kilometers (1.2 million acres)  of wetlands, over 17% of
all U.S. coastal wetlands.  Using published vertical accretion rates, they
estimated the impact of 1.4 m and 2.1 m rises in sea level through the year 2100
for each of the sites.  Weighing their results  according to the coastal  wetland
inventory by Alexander, Broutman, and Field (1986),  Titus (1987)  estimated a
loss of wetlands between 47% and 82%,  which could be  reduced to between 31% and
70% if new wetlands are not prevented from forming inland.

Inundation

    Although coastal wetlands are found at the  lowest  elevations,  inundation of
lowlands could also be important in some areas,  particularly if sea level rises
at least 1 m.  Unfortunately, the convention of 10-ft  contours in the mapping of
most coastal areas has prevented a general assessment  of land loss.
Nevertheless, a few case studies have been conducted.

    Kana et al.  (1984) used data from aerial photographs to assess elevations in
the area around Charleston.  They concluded that 160-  and 230-cm rises would
result  in 30% and 46% losses, respectively,  of  the area's dry land.  Leatherman
(1984)  estimated that such rises would result in 9%  and 12% losses,
respectively, of the land in the area of Galveston and Texas City,  Texas,
assuming that the elaborate network of seawalls  and levees was maintained.

-------
                                        15-20


                                    EXHIBIT  15-8

                        Louisiana Shoreline  in the Year  2030
N
                 50
         miles
                                  SOURCE:  COASTAL ENVIRONMENTS, INCORPORATED
                                          COASTAL ZONE WETLANDS
PREDICTED LOUISIANA COASTLINE
IN 50 YEARS AT PRESENT LAND LOSS RATES
                                       BATON ROUGE
                                                                      NEW ORLEANS
        Gulf of Mexico
                                                 HOUMA
                                                              LOOP facility
  The solid line shows the predicted shoreline for 2030 if current  trends
  continue.  The entire shaded  area  could be lost by 2085 if sea  level  rises 1 m.
  However,  a 20-cra rise by 2020 would result in the shore retreating to this point
  by 2020 or sooner.
  Source:   Coastal Environments,  Incorporated.

-------
                                       15-21


                                   EXHIBIT 15-9

                    Distribution of Population in  Bangladesh
                                                  BANGLADESH
                                          DISTRIBUTION OF POPULATION
                                                    1971'
                                             Each dot r»pr«t»m> 2 000 p*rtont
The high scenario  represents a 2- to  2.5-m rise with sedimentation  disrupted by
human activities.   The low scenario represents a 50-cm rise and natural
sedimentation maintained.
Source:  Broadhus  et al.  (1986).

-------
                                     15-22
    The only nationwide assessment of the inundation from projected sea level
rise was conducted by Schneider and Chen (1980).   Unfortunately,  the smallest
rise in sea level they considered was a 4.5-m (15-ft) rise,  in part because
smaller contours are not generally available in topographic  maps.   Nevertheless,
their findings suggest which coastal states would be most vulnerable:   Louisiana
(which would lose 28% of its land and 51% of its  assets);  Florida (24% and 52%);
Delaware (16% and 18%); Washington, D.C. (15% and 15%);  Maryland (12%  and 5%);
and New Jersey (10% and 9%).

    As with wetland loss, the responses to inundation broadly fall into the
categories of retreat and holding back the sea.   Levees  are  used extensively in
the Netherlands and New Orleans to prevent areas  that are below sea level from
being flooded and could be similarly constructed around other major cities.  In
lightly developed areas, however,  the cost of a levee might  be greater than the
value of the property being protected.  Moreover, even where levees prove to be
cost-effective, the environmental implications of replacing  natural shorelines
with manmade structures would need to be considered.

Coastal Erosion

    Sea level rise can also result in the loss of land above sea level through
erosion.  Bruun (1962) showed that the erosion resulting from a rise in sea
level would depend upon the average slope of the  entire beach profile  extending
from the dunes out to the point where the water is too deep  for waves  to have a
significant impact on the bottom,  generally a depth of about 10 m (see Exhibit
15-10):
                                 r = s * p/(h + d)

where r = shoreline retreat;  s = sea level rise;  p = horizontal length of the
beach profile from dune to the offshore limit of significant wave action; h =
height of the dune crest; and d = depth of the offshore limit of the beach
profile.  By comparison, inundation depends only on the slope immediately above
the original sea level.  Because beach profiles are generally flatter than the
portion of the beach just above sea level, the "Bruun Rule"  generally implies
that the erosion from a rise in sea level is several times greater than the
amount of land directly inundated.

    Processes other than sea level rise also contribute to erosion.  These
include storms, structures, currents, and alongshore transport.  Because sea
level has risen slowly in recent centuries, verification of  the Bruun Rule on
the open coast has been difficult; it has been difficult to  distinguish erosion
due to sea level rise from erosion caused by other processes.  However, water
levels along the Great Lakes can fluctuate over 1 m in a decade.   Hands  (1976,
1979, and 1980) and Weishar and Wood (1983) have demonstrated that the Bruun
Rule generally predicts the erosion resulting from rises in water levels there.

    The Bruun Rule has been applied to project erosion due to sea level rise
for several areas.  Bruun  (1962) found that a 1-cm rise in sea level would
generally result in a 1-m shoreline retreat, but that the retreat could be as
great as 10 m along some parts of the Florida coast.  Everts (1985) and Kyper
and Sorensen (1985) however, found that along the coasts of Ocean City,
Maryland, and Sandy Hook, New Jersey, respectively, the shoreline retreat
implied by the Bruun Rule would be only about 75 cm.  Kana et al. (1984) found
that along the coast of South Carolina, the retreat could be 2 m.  The U.S.

-------
                                       15-23



                                   EXHIBIT 15-10

                                  The Bruun Rule
         Initnl
         Condition
                                                                    J
         Inundation When
         S»« Ltv«l Rlsct
        Subs*Qu*nt
        Erosion Ou« to
        S»« L»v«! Rla«
A rise in  sea level of s  causes immediate  inundation.   However,  it  would
eventually require the offshore bottom to  rise  by s.   The necessary s and A'
would be supplied from the  upper part of the beach A.   Total shoreline retreat
r is equal to s * p/(h +  d).

-------
                                     15-24
Army Corps of Engineers (1979)  indicated that along the coast of San Francisco,
where waves are generally larger than those found along the Atlantic Coast,  the
shore might retreat 2-4 m for a 1-cm rise in sea level.

    Dean and Maurmeyer (1983) generalized the "Bruun Rule" approach to consider
the "overwash" of barrier islands.   Coastal geologists generally believe that
coastal barriers can maintain themselves in the face of slowly rising sea level
through the landward transport of sand,  which washes over the island during
storms, building the island upward and landward.   Because this formulation of
the Bruun Rule extends the beach profile horizontally to include the entire
islands as well as the active surf zone, it always predicts greater erosion
than the Bruun Rule.  However,  the formulation may not be applicable to
developed barrier islands, where the common practice of public officials is to
bulldoze sand back onto the beach after a major storm.

    The potential erosion from a rise in sea level could be particularly
important to recreational beach resorts, which include some of the nation's
most economically valuable and intensely used land.  Although nationwide
statistics are not available, information on particular locations is
instructive.  Every weekend in the summer, approximately 250,000 people visit
Ocean City, Maryland, to sunbathe on its 15-km (9-mile) beach.   Along most of
the Ocean City shoreline, the beach is less than 15 m wide during high tide--a
width that is typical of the most extensively used beach resorts.  Relatively
few of the most intensely developed resorts along the Atlantic Coast have
beaches wider than about 30 m at high tide.  Thus, the rise in relative sea
level that is projected to occur in the next 40 to 50 years (30 cm) could erode
most recreational beaches in developed areas, unless additional erosion
response measures are taken.

Flooding and Storm Damage

    A rise in sea level could increase flooding and storm damages in coastal
areas for three reasons:  Erosion caused by sea level rise would increase the
vulnerability of communities; because of higher water levels, storm surges
would have a higher base on which to build; and higher water levels would
decrease natural and artificial drainage.

    The impact of erosion on vulnerability to storms is generally a major
consideration in proposed projects to control erosion, most of which have
historically been funded through the U.S. Army Corps of Engineers.  The impact
of sea level rise, however, has not generally been considered separately from
other causes of erosion.

    The impact of higher base water levels on flooding has been investigated
for the areas around Charleston, South Carolina, and Galveston, Texas (Barth
and Titus 1984).  Kana et al. (1984) found that around Charleston, the area
within the 10-year flood plain would increase from 33% in 1980, to 48%, 62%,
and 74% for rises in sea level of 88, 160, and 230 cm, respectively, and that
the area within the 100-year flood plain would increase from 63% to 76%, 84%,
and 90%, respectively, for the three scenarios.   Gibbs (1984) estimated that a
rise of about 90 cm would double the average annual flood damages in the
Charleston area (but that flood losses would not increase substantially for
higher rises in sea level because shoreline retreat would result in a large
part of the community being completely abandoned).

-------
                                     15-25
    Leatherman (1984) conducted a similar analysis of Galveston Island, Texas.
He estimated that the area within the 100-year flood plain would increase from
58% to 94% for an 88-cm rise in sea level,   and that for a rise greater than 1
m, the Galveston seawall would be overtopped during a 100-year storm.  Gibbs
estimated that the damage from a 100-year storm would be tripled for a rise of
about 90 cm.10

    A wide variety of shore-protection measures would be available for
communities to protect themselves from increased storm surge and wave damage
due to sea level rise (Sorensen, Weisman, and Lennon 1984).   Many of the
measures used to address erosion and inundation, including construction of
seawalls, breakwaters, and levees,  and rebuilding beaches, also provide
protection against storms.  In Galveston, which is already protected on the
ocean side by the seawall, Gibbs hypothesized that it might be necessary to
completely encircle the developed areas with a levee to prevent flooding from
the bay side; upgrading the existing seawall might also be necessary.

    Kyper and Sorensen (1985 and 1987) examined the implications of sea level
rise for the design of coastal protection works at Sea Bright, New Jersey, a
coastal community that is currently protected by a seawall and has no beach.
Because the seawall is vulnerable to even a 10-year storm, the Corps of
Engineers and the State of New Jersey have been considering a possible upgrade.
Kyper and Sorensen estimated that the cost of upgrading the seawall for current
conditions would be $3.5-6.0 million per kilometer of shoreline, noting that if
designed properly, the seawall would be useful throughout the next century.
However, they estimated that a rise in relative sea level of 30-40 cm would be
likely to result in serious damage to the seawall during a major storm because
of higher water levels and the increased wave heights resulting from the
erosion of submerged sand in front of the seawall.  To upgrade the seawall to
withstand a 1-m rise in relative sea level would cost $5.7-9.0 million per
kilometer (50% more).  They concluded that policymakers would have to weigh the
tradeoff between the cost of designing the wall to withstand projected sea
level rise and the cost of subsequent repairs and a second overhaul.

    In addition to community-wide engineering approaches, measures can also be
taken by individual property owners to prevent damages from increased flooding.
In 1968, the U.S. Congress created the National Flood Insurance Program to
encourage communities to avoid risky construction in flood-prone areas.  In
return for requiring new construction to be elevated above expected flood
levels, the federal government provides flood insurance, which is not available
from the private sector.  If sea level rises, flood risks will increase.  In
response, local ordinances will automatically require new construction to be
further elevated, and insurance rates on existing properties will rise unless
those properties are further elevated.  As currently organized, the National
Flood Insurance Program would react to sea level rise as it occurred.  Various
measures that would ensure the continued viability of the program if the sea
j_evel were to rise substantially have been proposed, including warning
policyholders that rates may increase in the future if sea level rises, denying
coverage to new construction in areas that are expected to be lost to erosion
within the next 30 years, and setting premiums according to the average risk
expected over the lifetime of the mortgage (Howard, Pilkey,  and Kaufman 1985;
Titus 1984).

-------
                                     15-26
    Case studies in Charleston,  South Carolina, and Fort Walton Beach, Florida,
have examined the implications of sea level rise for rainwater flooding and the
design of coastal drainage systems.  Waddell and Blaylock (n.d.) estimated that
a 25-year rainstorm (with no storm surge) would result in no damage for the Gap
Creek watershed in Fort Walton Beach.  However, a rise in sea level of 30-45 cm
would result in damages of $1.1-1.3 million in this community of 4,000
residents during a 25-year storm.  An upgrade costing $550,000, however, would
prevent such damages.

    LaRoche and Webb (n.d.) had previously developed the master drainage plan
for Charleston, South Carolina,  and later evaluated the implications of sea
level rise for the Grove Street watershed in that community.  They estimated
that the cost of upgrading the system for current conditions would be $4.8
million and the cost of upgrading the system for a 30-cm rise would be $5.1
million.  If the system is designed for current conditions and sea level rises,
the system would be deficient and the city would face a retrofit cost of $2.4
million.  Thus, for the additional $300,000 necessary to upgrade for a 30-cm
rise, the city could ensure that it would not have to spend an additional $2.4
million later.  Noting that the decision whether to design now for a rise in
sea level depends on the probability that sea level would rise, they concluded
that a 3% real social discount rate would imply that designing for sea level
rise is worthwhile if the probability of a 30-cm rise by 2025 is greater than
30%.  At a discount rate of 10%, they concluded, designing for future
conditions is not worthwhile.

Increased Salinity in Estuaries and Aquifers

    Although most researchers and the general public have focused on the
increased flooding and shoreline retreat associated with a rise in sea level,
the inland penetration of salt water could be  important in some areas.

    A rise in  sea level increases the salinity of an estuary by altering the
balance between freshwater and saltwater forces.  The salinity of an estuary
represents the outcome of (1) the tendency for the ocean salt water to
completely mix with  the estuarine water and  (2) the tendency of fresh water
flowing into  the estuary to dilute the saline water and push it back toward the
ocean.  During droughts, the salt water penetrates upstream; during the rainy
season, low salinity levels prevail.  A rise in sea level has an impact similar
to decreasing  the freshwater inflow.  By widening and deepening the estuary,
sea level rise increases the ability of salt water to penetrate upstream.

    The implications of sea level rise for increased salinity have only been
examined in detail for Louisiana and the Delaware Estuary.  In Louisiana,
saltwater intrusion  is currently resulting in  the conversion of cypress
swamps--which  cannot tolerate salt water--to openwater  lakes,  as well as the
conversion of  fresh  marsh and intermediate marsh to marsh with higher salinity
levels.  In response to these trends, numerous projects have been proposed  to
divert  fresh  water from the Mississippi River  to these  wetlands (Roberts et al.
1983; U.S. Army  Corps  of Engineers 1982; van Beek et al. 1982).  Although the
cause of the  saltwater intrusion has been primarily the dredging of canals  and
the  sealing off  of river tributaries that once provided the wetlands with fresh
water,  relative  sea  level  rise  is  gradually  increasing  the  saltwater  force  in
Louisiana's wetlands;  a further  rise in  sea  level would accelerate  this
process.  Haydl  (1984) also concluded that  subsidence may be  resulting  in

-------
                                     15-27
 increased salinity of drinking and irrigation water supplies in some parts of
 Louisiana.

    The impact of current sea level trends on salinity has been considered in
 the long-range plan of the Delaware River Basin Commission since 1981  (DRBC
 1981) .  The drought of the 1960s resulted in salinity levels that almost
 contaminated the water supply of Philadelphia and surrounding areas.   Hull and
 Tortoriello (1979) found that the 13-cm rise projected between 1965 and 2000
 would result in the "salt front" migrating 2-4 km farther upstream during a
 similar drought.  They found that a moderately sized reservoir (57 million
 cubic meters) to augment river flows would be needed to offset the resulting
 salinity increases.

    Hull, Thatcher,  and Tortoriello (1986) examined the potential impacts of an
 accelerated rise in sea level due to the greenhouse warming.  They estimated
 that 73-cm and 250-cm rises would result in the salt front migrating an
 additional 15 and 40 kilometers, respectively, during a repeat of the  1960s
 drought.  They also found that the health-based, 50-ppm sodium standard
 (equivalent to 73 ppm chloride) adopted by New Jersey would be exceeded 15% and
 50% of the time, respectively, and that the EPA drinking water 250-ppm chloride
 standard would be exceeded over 35% of the time in the high scenario (see
 Exhibit 15-11).

    Lennon, Wisniewski,  and Yoshioka (1986) examined the implications  of
 increased estuarine salinity for the Potomac-Raritan-Magothy aquifer system,
 which is recharged by the (currently fresh) Delaware River and serves  the New
 Jersey suburbs of Philadelphia.  During the 1960s drought, river water with
 chloride concentrations as high as 150 ppm recharged these aquifers.   Lennon et
 al. estimated that a repeat of the 1960s drought with a 73-cm rise in  sea level
 would result in river water with chloride concentrations as high as 350 ppm
 recharging the aquifer,  and that during the worst month of the drought, over
 one-half of the water recharging the aquifer system would have chloride
 concentrations greater than 250 ppm.   With a 250-cm rise, 98% of the recharge
 during the worst month of the drought would exceed 250 ppm chloride, and 75% of
 the recharge would be greater than 1000 ppm.

    Hull and Titus (1986) examined the options by which various agencies might
 respond to increased salinity in the Delaware estuary.   They concluded that
 planned but unscheduled reservoirs would be more than enough to offset the
 salinity increase from a 30-cm rise in sea level,  although those reservoirs had
 originally been intended to meet increased consumption.   They noted that
 construction of the  reservoirs would not be necessary until the rise became
 more imminent.   However,  they also suggested that,  given the uncertainties, it
 might be advisable today to identify additional reservoir sites to ensure that
 future generations retain the option of building additional reservoirs if
 necessary.

    A rise in sea level  could increase salinities  in other areas,  although the
 importance of those  impacts has not been investigated.   Kana et al.  (1984) and
 Leatherman (1984)  made preliminary inquiries  into  the potential impacts on
 coastal aquifers around Charleston and Galveston,  respectively.   However,  they
 concluded that in-depth  assessments were not  worthwhile  because the aquifers
 around Charleston are already salt-contaminated because  of overpumping, and
pumping of ground water  has been prohibited in the  Galveston area due to the

-------
                ..aft
f

-------
                                     15-29
resulting land subsidence.  The potential impacts on Florida's Everglades and
the shallow aquifers around Miami might be significant, but they have not been
investigated.

Economic Significance of Sea Level Rise

    No one has estimated a dollar value of the likely impacts of sea level rise
for the nation or any coastal state.  Schneider and Chen (1980) estimated the
economic impact of what was once (but is no longer) thought to be a plausible
scenario:  rises of 4.6-7.6 m (15-25 ft) occurring with little or no warning
during the early part of the twenty-first century.  They estimated that this
scenario would result in real property losses of $100-150 billion, representing
6.2% to 8.4% of all real property in the nation.

    The only comprehensive attempt to place a dollar value on the impacts of
sea level rise for particular communities was the study by Gibbs (1984) of the
Charleston and Galveston areas.   In addition to considering more realistic
scenarios ranging from 0.9 to 2.4 m through 2075, Gibbs made several
improvements on the approach of Schneider and Chen:  Gibbs' approach (1)
differentiated between the economic impact if preventive measures are taken in
anticipation of sea level rise and the economic impact if such measures are not
taken (with and without anticipation; see Exhibit 15-12), (2) modeled how
investment decisions might respond to floods and erosion, and (3) explicitly
considered community-wide strategies to limit losses, including shore
protection and abandonment.

    In the Charleston study, Gibbs evaluated the implications of (1) efforts to
avoid development of some vacant suburban areas likely to be flooded in the
future,  (2) a partial abandonment,  and (3) elevating the existing seawalls
protecting Charleston to provide additional protection.  For a rise of 28-64 cm
through 2025, Gibbs estimated that present value of the cumulative impact would
be $280-1065 million (5% to 19%  of economic activity in the area for the
period), which could be reduced to $160-420 million if sea level rise was
anticipated.  Most of this impact would result from a 10% to 100% increase in
expected storm damages,  although Gibbs also estimated losses to erosion at
$7-35 million.   For the period between 1980 and 2075, Gibbs estimated that the
economic impacts would be $1250-2510 million (17% to 35%) and could be reduced
to $440-1100 million through anticipatory measures.  If accurate, these
estimates suggest that Charleston would be extremely vulnerable to a rise in
sea level.   However, these estimates were necessarily based on a number of
simplifying assumptions regarding storm damage and current elevation of
structures.

    Gibbs performed a similar analysis of the Galveston area and concluded that
the economic impacts of sea level rise through 2025 would amount to $115-360
million (l.l%-3.6%) if a rise in sea level is not anticipated,  and $90-140
million if a rise is anticipated.   The impacts through 2075 were estimated at
$555-1840 million (4.9%-16%) if  not anticipated,  and $310-730 million if a rise
is anticipated.   The assumptions Gibbs used suggest that a 100-year storm today
would result in damages equal to the value of about 7% of the property in the
area,  and that average annual storm damages are equal to $6 million,  slightly
less than 0.2% of the value of property in the study area.

-------
                                     15-30
                                     EXHIBIT 15-12
                Estimates of Flood Damages for Charleston and Galveston
                            Resulting from Sea Level Rise3/
              Global           Charleston Impact b/           Galveston Impact b/
           Sea Level Rise      With        Without            With        Without
Scenario    by 2075 (cm)     Anticipation Anticipation     Anticipation  Anticipation
Trend
Low
Medium
High
11.4
75.2
146.8
219.2
c/
440
730
1100
c/
1250
1910
2510
c/
310
415
730
£/
555
965
1890
a/ The damage estimates consist of a low estimate and a high estimate.   The low
   estimate is derived by linearly interpolating the "with anticipation"
   estimates for Charleston and Galveston according to the amount of global sea
   level rise.  The high estimates are derived by linearly interpolating the
   "without anticipation" estimates.

b/  Through 2075 in millions of 1980 dollars discounted at 3%.

c/  All estimates are relative to the trend scenario.


Source:  Gibbs 1984.

-------
                                      15-31
    Other studies can be used  to  gain an understanding of the economic
 significance of particular classes of impacts.  As discussed above, a 30-cm
 rise  in sea level would erode  most recreational beaches back to the first row
 of houses.  The studies cited  in  the section on erosion indicate that the
 typical beach profile extends  out about 1000 m, which suggests that 300,000
 cubic yards of sand per kilometer of shoreline are required to raise the beach
 profile 30 cm.  If sand costs  are typically $3-10/cubic meter, the beach
 rebuilding costs of a 30-cm rise  in sea level would be $1-3 million per
 kilometer.  If the United States  has a few thousand kilometers of recreational
 beaches, it would cost billions and perhaps tens of billions of dollars to
 rebuild beaches if the sea rises  30 cm.  This estimate considers only the
 beaches themselves; raising people's lots to prevent property inundation would
 further increase the costs.

    The U.S. Army Corps of Engineers (1971) estimated that in 1971, 25,000 km
 of shoreline (exclusive of Alaska, Great Lakes, and Hawaii) were eroding, of
 which 17% were "critically eroding," and would require engineering solutions.
 If 17% of all shorelines require  erosion control, that would imply protection
 of close to 10,000 km of shoreline.  Sorensen, Weisman, and Lennon (1984)
 describes dozens of engineering options for preventing erosion, the least
 expensive of which costs $300,000 per kilometer, which would imply a cost of at
 least $3 billion for protecting all shorelines.

 Relationship to Other Impacts  of  the Greenhouse Warming

    The impacts of sea level rise on coastal areas, as well as their
 importance, is likely to depend in part on other impacts of the greenhouse
 warming.  Although future sea  level is uncertain, there is a general consensus
 that a global warming would cause sea level to rise;  by contrast,  the direction
 of most hydrological and storm event changes at particular locations is
 unknown.

    Warmer temperatures might change the ability of wetlands to keep pace with
 sea level rise.   Mangrove swamps  would replace salt marshes in some areas that
 are currently too cold for mangroves.   No one has assessed the impact of this
 vegetation change on vertical accretion.   Peat formation is generally greater
 in New England marshes than at lower latitudes.   Warmer temperatures might
 reduce that rate of vertical accretion for these wetlands.

    Changing climate could alter  the frequency and tracks of storms.   Because
hurricane formation requires water temperatures of 27°C or higher (Wendland
 1977), a global warming might result in a longer hurricane season and in the
 formation of hurricanes at higher latitudes.   Besides increasing the amount of
 storm damage,  increased frequency of severe storms would tend to flatten the
 typical beach profile,  causing substantial shoreline  retreat unless additional
 sand was placed on the beach.   A  decreased frequency  of severe winter storms
might have the opposite impact at higher latitudes.

    Changes in precipitation could also affect the impacts of sea level rise.
Because warmer temperatures would intensify the hydrologic cycle,  it is
generally recognized that a global warming would increase rainfall  worldwide.
If rainfall in maritime environments increases,  rainwater flooding  could be
increased because of decreased drainage and increased precipitation.   The
impact of sea level rise on saltwater  intrusion could be offset by  decreased

-------
                                     15-32
drought frequency or exacerbated by increased drought frequency (Rind and
Lebedeff 1984).

CONCLUSION

    Researchers  have identified a wide variety of impacts that could result
from a 50- to 200-cm rise in sea level.  These impacts include both economic
losses resulting from land loss, flooding,  shore protection,  and increased
costs for drinking water, as well as environmental costs such as the disruption
of coastal ecosystems resulting from wetland loss and increased salinity.
Although there are substantial uncertainties regarding the magnitude of both
the causes and effects of future sea level rise, even conservative estimates
suggest that important impacts are likely to result.

    A high priority for future research should be the use of existing case
studies to derive estimates of the nationwide and worldwide magnitude of the
various impacts  that have been identified.   Although preliminary estimates of
the potential loss of U.S. coastal wetlands exist, the implications of those
losses have not  been assessed.  For most other impacts,  no nationwide estimates
are available.  That type of information will be important for evaluating the
benefits of possible actions to mitigate the greenhouse warming and resulting
rise in sea level.

-------
                                     15-33
NOTES
1.  Mark Meier, University of Colorado,  Boulder (personal communication).

2.  The only true measure of absolute sea level would be the rise of sea level
    relative to the center of the earth.  Unfortunately, no such measurements
    are yet available.  Therefore, researchers have had to combine tidal gauge
    measurements of relative sea level trends at various locations,  filter out
    known movements of the land surface, and take weighted averages  to arrive
    at estimates of global sea level trends.

3.  This result was reported in the North America study.  The data also show
    this to be true in the Northern Europe study, but the result was not
    reported.  David Aubrey, Woods Hole Oceanographic Institute, Woods Hole,
    Massachusetts (personal communication).

4.  Robert Bindschadler,  Goddard Space Flight Center, Greenbelt, Maryland
    (personal communication).

5.  James Hansen, Goddard Institute for Space Studies, New York (personal
    communication).

6.  Sandy Coyman, Department of Planning and Community Development,  Town of
    Ocean City, Maryland (personal communication).

7.  Results of the Charleston and Galveston case studies have been reported in
    J.G. Titus and Michael C.  Earth, 1984, "An Overview of the Causes and
    Effects of Sea Level Rise," Appendix, In Greenhouse Effect and Sea Level
    Rise:  A Challenge for This Generation.  New York:  Van Nostrand Reinhold,
    1984.

8.  See Note 7.

9.  See Note 7.

10. See Note 7.

-------
                                     15-34
REFERENCES
Alexander, C.E., M.A. Broutman, and D.W. Field,  1986.  An Inventory of
    Coastal Wetlands of the USA.  Rockville, MD.:   National Oceanic and
    Atmospheric Administration (NOAA),  National Ocean Service (NOS).

Ambach, W.  1985.  "Climatic Shift of the Equilibrium Line--Kuhn's Concept
    Applied to the Greenland Ice Cap."   Annals of Glaciology 6:76-78.

Ambach, W.  1980 (translated by G.P. Weidhaas).  "Increased C02 Concentration
    in the Atmosphere and Climate Change:  Potential Effects on the Greenland
    Ice Sheet."  Wetter und Leben 32:135-142, Vienna.  (Available as Lawrence
    Livermore National Laboratory Report UCRL-TRANS-11767,  April 1982.)

Associated Press.  1985.  "Doubled Erosion Seen for Ocean City." The
    Washington Post. November 14th, (Maryland Section).

Barnett, T.P.  1984.  "The Estimation of 'Global'  Sea Level Change:  A
    Problem of Uniqueness."  Journal of Geophysical Research  89(C5):7980-7988.

Earth. M.C. and J.G. Titus (eds.)  1984.  Greenhouse Effect and Sea Level
    Rise:  A Challenge for this Generation.  New York:  Van Nostrand Reinhold.

Bentley, C.R.  1983.  "West Antarctic Ice Sheet:  Diagnosis and Prognosis."
    In Proceedings:  Carbon Dioxide Research Conference:   Carbon Dioxide.
    Science, and Consensus. Conference  820970.  Washington, D.C.:  Department
    of Energy.

Bindschadler, R.A.   1985.  "Contribution of the Greenland Ice Cap to Changing
    Sea Level:  Present and Future."  In Meier et al. 1985.  op. cit.

Boesch, D.F. (ed.)   1982.  Proceedings  of the Conference of Coastal Erosion
    and Wetland Modification in Louisiana:   Causes. Consequences, and Options.
    FWS-OBS-82/59.   Slidell,  Louisiana:   National Coastal Ecosystems Team, U.S.
    Fish and Wildlife Service.

Braatz, B.V. and D.G. Aubrey.  1987.  "Recent Relative Sea-Level Movement
    Along the Eastern Coastline of North America."  Journal of Sedimentary
    Petrology, in press.

Broadhus, J., et al.  1986.  "Sea Level Rise and the Damming of Rivers."  In
    Titus, J.G. (ed.)  Effects  of Changes in Stratospheric  Ozone and Global
    Climate.  Washington, D.C.:  United Nations Environment Programme and U.S.
    Environmental Protection Agency.

Bruun, P. 1962.  "Sea Level Rise as a Cause of Shore Erosion."  Journal of
    Waterways and Harbors Division (ASCE) 1:116-130.

Cess,  R.D., and S.D. Goldenberg.  1981.   "The Effect of Ocean Heat Capacity
    Upon Global Warming Due to  Increasing Atmospheric Carbon Dioxide."  Journal
    of Geophysical Research  86:498-502.

-------
                                     15-35
Clark,  J.A.,  and C.S.  Lingle.   1977.   "Future Sea-Level Changes Due to West
    Antarctic Ice Sheet Fluctuations."   Nature 269(5625):206-209.

Clark,  J.A.,  W.E. Farrell,  and W.R.  Peltier.   1978.   "Global Changes in
    Postglacial Sea Level:   A Numerical Calculation."  Quarternary Research
    9:265-87.

Davis,  R.A.   1985.   Coastal Sedimentary Environments.  New York:  Springer
    Verlag.

Dean, R.G.,  and E.M.  Maurmeyer.  1983.   "Models for Beach Profile Response."
    In Handbook of Coastal  Processes and Erosion.   Fort Belvoir:  U.S. Army
    Corps of Engineers, Coastal Engineering Research Center.

Delaware River Basin Commission (DRBC).   1981.  The Delaware River Basin
    Comprehensive (Level B) Study:   Final Report and Environmental Impact
    Statement.  West Trenton,  New Jersey:  Delaware River Basin Commission.

Dewey,  K.F.,  and R.J.  Heim.  1981.   Satellite Observations of Variations in
    Northern Hemisphere Seasonal Snow Cover.   NOAA Technical Report, NESS 87,
    Washington, D.C.

Donn, W.L.,  W.R. Farrand, and M. Ewing.   1962.  "Pleistocene Ice Volumes and
    Sea-Level Lowering."  Journal of Geology 70:206-214.

Edmonson, J., and R.  Jones.  1985.   Terreboone Parish Barrier Island and Marsh
    Management Program.  Houma, Louisiana:  Terrebonne Parish Council.

Emery,  K.O.,  and D.G.  Aubrey.   1986.   "Relative Sea Level Change from Tide
    Gauge Records of Western North America."  Journal of Geophysical Research
    91(14):914-953.

Emery,  K.O.,  and D.G.  Aubrey.   1985.   "Glacial Rebound and Relative Sea Levels
    in Europe From Tide-Gauge Records."  Tectonophysics 120:239-255.

Emery,  K.O.,  and E. Uchupi.  1972.   Western North Atlantic Ocean Memoir 17.
    Tulsa, Oklahoma:   American Association of Petroleum Geologists.

Everts, C.H.   1985.  "Effect of Sea Level Rise and Net Sand Volume Change on
    Shoreline Position at Ocean City, Maryland."  In Potential Impacts of Sea
    Level Rise on the Beach at Ocean City. Maryland.  Washington, D.C.:  U.S.
    Environmental Protection Agency.

Fairbridge,  R.W., and W.S.  Krebs, Jr.  1962.  "Sea Level and the Southern
    Oscillation."  Geophysical Journal 6:532-545.

Fastook, J.L.  1985.   "Ice Shelves and Ice Streams:  Three Modeling
    Experiments."  In Meier et al.  1985, op. cit.

Flint,  R.F.   1971.  Glacial and Quarternary Geology.  New York:  John Wiley
    and Sons.

-------
                                     15-36
Flohn, H.  1982.  "Climate change and An Ice-Free Arctic Ocean".   In Carbon
    Dioxide Review:   1982. W.C. Clark, ed.  New York:  Oxford University Press,
    pp.145-199.

Gagliano, S.M., K.J. Meyer-Arendt, and K.M. Wicker.  1981.  "Land Loss in the
    Mississippi Deltaic Plain."  In Transactions of the 31st Annual Meeting of
    the Gulf Coast Association of Geological Societies.  Corpus Christi, Texas,
    p. 293-300.

Gibbs, M.  "Economic Analysis of Sea Level Rise:  Methods and Results."  In
    Earth and Titus (eds.) 1984, CJD. cit.

Gornitz, V., S. Lebedeff, and J. Hansen.  1982.  "Global Sea Level Trend in
    the Past Century."  Science 215:1611-1614.

Hands, E.G.  1980.  Prediction of Shore Retreat and Nearshore Profile
    Adjustments to Rising Water Levels on the Great Lakes. Tech.  Paper 870-7.
    Fort Belvoir:   U.S. Army Corps of Engineers.

Hands, E.G.  1979.  Changes in Rates of Offshore Retreat:  Lake Michigan.
    Fort Belvoir:   U.S. Army Corps of Engineers Coastal Engineering Research
    Center Publication TP79-4.

Hands, E.G.  1976.  Observations of Barred Coastal Profiles Under Influence
    of Rising Water Levels:  Eastern Lake Michigan. CERC Publication TR76-1.
    Fort Belvoir:   U.S. Army Corps of Engineers.

Hatton, R.S., R.D. DeLaune, and W.H. Patrick. 1983.  "Sedimentation,
    Accretion and Subsidence in Marshes of Barataria Basin, Louisiana."
    Limnology and Oceanography 28:494-502.

Haydl, N.C.  1984.  Louisiana Coastal Area, Louisiana:   Water Supply.
    Initial Evaluation Study.  New Orleans:  U.S. Army Corps of Engineers.

Hays, J.D., and W.C. Pitman III.  1973.  "Lithospheric Plate Motion, Sea Level
    Changes, and Climatic and Ecological Consequences."  Nature 246:18-22.

Hicks, S.D., H.A.  Debaugh, and L.H.  Hickman.  1983.  Sea Level Variations for
    the United States 1855-1980.  Rockville, Maryland:   U.S. Department of
    Commerce, NOAA-NOS.

Hoffman, J.S.,  D.  Keyes,  and J.G. Titus.  1983.  Projecting Future Sea Level
    Rise.  Washington, D.C.:  Government Printing Office.

Hoffman, J.S.,  J.  Wells,  and J.G. Titus.  1986.  "Future Global Warming and
    Sea Level Rise."  In Bruun, P. (ed.), Iceland Symposium '85.   Reykjavik:
    National Energy Authority.

Hollin, J.T., 1962.   "On the Glacial History of Antarctica."  Journal of
    Glaciology 4:273-295.

-------
                                     15-37
Hollin, J.T.,  and R.G.  Barry.   1979.   "Empirical and Theoretical Evidence
    Concerning the Response of the Earth's Ice and Snow Cover to a Global
    Temperature Increase."  Environment International 2:437-444.

Howard, J.D.,  O.H. Pilkey, and A.  Kaufman.  1985.   "Strategy for Beach
    Preservation Proposed."  Geotimes 30:12:15-19.

Hughes, T. 1983.  "The  Stability of the West Antarctic Ice Sheet:  What Has
    Happened and What Will Happen."  In Proceedings:  Carbon Dioxide Research
    Conference:  Carbon Dioxide. Science,  and Consensus.  Conference 820970.
    Washington, D.C.:  Department of Energy.

Hull, C.H.J. and J.G. Titus (eds).   1986.   "Responses to Salinity
    Increases."  In Greenhouse Effect.  Sea Level Rise, and Salinity in the
    Delaware Estuary.  Washington,  D.C.:  Environmental Protection Agency and
    Delaware River Basin Commission.

Hull, C.H.J.,  M.L. Thatcher, andR.C. Tortoriello.  1986.  "Salinity in the
    Delaware Estuary."   In Hull and Titus (eds.) 1986, op. cit.

Hull, C.H.J. and R.C. Tortoriello.   1979.   Sea Level Trend and Salinity in
    the Delaware Estuary.   West Trenton, N.J.:  Delaware River Basin
    Commission.

Jones, P.O., T.M.L. Wigley, and P.B.  Wright.  1986.  "Global Temperature
    Variations between 1861 and 1984."   Nature 322:430-434.

Kana, T.W., J.  Michel,  M.O. Hayes,  and J.R.  Jensen.  1984.  "The Physical
    Impact of Sea Level Rise in the Area of Charleston, South Carolina."  In
    Barth and Titus  (eds.) 1984, op.  cit.

Kana, T.W., B.J. Baca,  and M.L. Williams.   1986.  Potential Impacts of Sea
    Level Rise on Wetlands Around Charleston. South Carolina.  Washington,
    D.C.:  U.S. Environmental Protection Agency.

Kana, T.W., W.C. Eiser, B.J. Baca, and M.L.  Williams.  1987  "Potential
    Impacts of Sea Level Rise on Wetlands Around Little Egg Harbor, New
    Jersey."  In Greenhouse Effect. Sea Level Rise and Coastal Wetlands.
    Washington, D.C.:  U.S. Environmental Protection Agency.

Kennett, J.  1982.  Marine Geology.  Englewood Cliffs, New Jersey:
    Prentice-Hall.

Kuo, C.  n.d.   Sea Level Rise. Climate Change and Coastal Drainage
    Systems.  Washington,   D.C.:  U.S. Environmental Protection Agency,  (in
    press)

Kyper, T., and R. Sorensen.  1985.   "Potential Impacts of Selected Sea Level
    Rise Scenarios on  the  Beach and Coastal Works at Sea Bright, New Jersey."
    In O.T. Magoon et  al.   (eds.), Coastal Zone  '85.  New York:  American
    Society of Civil Engineers.

-------
                                     15-38
Kyper, T. and R. Sorensen.  1987.   Potential Impacts of Sea Level Rise on the
    Beach and Coastal Structures at Sea Bright.  New Jersey.   Washington, D.C.:
    U.S. Environmental Protection Agency and U.S.  Array Corps of Engineers (in
    press).

Lacis, A., J.E. Hansen, P. Lee, T. Mitchell, and S. Lebedeff.  1981.
    "Greenhouse Effect of Trace Gases,  1970-1980."  Geophysical Research
    Letters 8:(10):1035-1038.

LaRoche, T.B.,  and M.K. Webb.   n.d.  "Impact of Sea Level Rise on Stormwater
    Drainage Systems in the Charleston, South Carolina, Area."  In Kuo, C.
    (ed.), Sea Level Rise. Climate Change and Coastal Drainage Systems.
    Washington, D.C.:  U.S. Environmental Protection Agency (in press).

Leatherman,  S.P.  1984.  "Coastal Geomorphic Responses to Sea Level Rise:
    Galveston Bay, Texas."  In Earth and Titus (eds.) 1984, op. cit.

Lennon, G.P., G.M. Wisniewski,  and G.A. Yoshioka.   1986.  "Impact of
    Increased River Salinity On New Jersey Aquifers."  In Hull and Titus (eds.)
    1986, op. cit.

Lingle, C.S.  1985.   "A Model of a Polar Ice Stream and Future Sea-Level Rise
    Due to Possible Drastic Retreat of the West Antarctic Ice Sheet."  In Meier
    et al. 1985, op. cit.

Meier, M.F.   1984.  "Contribution of Small Glaciers to Global Sea Level."
    Science 226:1418-1421.

Meier, M.F., et al.   1985.  Glaciers. Ice Sheets,  and Sea Level:  Effect of a
C02-Induced Climatic Change.  Washington, D.C.:   National Academy Press.

Mercer, J.H.  1978.   "West_Antarctic Ice Sheet and C02 Greenhouse Effect:
    A Threat of Disaster?"  Nature 271:321-325.

Mercer, J.H.  1968.   "Antarctic Ice and Sangamon Sea Level."  Geological
    Society of America Bulletin 79:471.

Meyerson, A.L.   1987.  "Pollen and Paleosalinity Analyses from a Holocene
    Tidal Marsh Sequence."  Marine Geology 12:335-357.

Milliman, J.D., and R.H. Meade.  1983.   Worldwide Delivery of River Sediment
    to the Oceans.  Journal of Geology 91(1):1-21.

Moore, W.S.   1982.  Late Pleistocene Sea-Level History.  In Uranium Series
    Disequilibrium:   Applications to Environmental Problems. M. Ivanovich and
    R.S. Harmon (eds.). Clarendon, Oxford, p. 41.

Oldale, R.  1985.   "Late Quarternary Sea Level History of New England:  A
    Review of Published Sea Level Data."  Northeastern Geology 7:192-200.

-------
                                     15-39
Park, R.A., T.R. Armentano,  and C.L. Cloonan.  1986.  "Predicting the Effects
    of Sea Level Rise on Coastal Wetlands."  In J.G. Titus (ed.), 1986.  Effects
    of Changes in Stratospheric Ozone and Global Climate.   Washington,  D.C.:
    U.S. Environmental Protection Agency and United Nations Environment
    Programme.

Redfield, A.C.   1972.  "Development of a New England Salt Marsh."  Ecological
    Monograph 42:201-237.

Redfield, A.C.   1967.  "Postglacial Change in Sea Level in the Western North
    Atlantic Ocean."  Science 157:687-692.

Revelle, R.  1983.  "Probable Future Changes in Sea Level Resulting from
    Increased Atmospheric Carbon Dioxide."  In Changing Climate.   Washington,
    D.C.:  National Academy Press.

Rind, D., and S. Lebedeff.   1984.  Potential Climate Impacts of Increasing
    Atmospheric C02 with Emphasis on Water Availability and Hydrology in the
    United States.  Washington, D.C.:  Government Printing Office.

Roberts, D., D. Davis, K.L.  Meyer-Arendt,  and K.M. Wicker.  1983.
    Recommendations for Freshwater Diversion to Barataria Basin.  Louisiana.
    Baton Rouge:  Coastal Management Section, Louisiana Department of Natural
    Resources.

Robin, G.  1985.  "Contrast in Core:  Changes in Climate or Ice Volumes?"
    Nature 316:578-589.

Roemmich, D.  1985.  "Sea Level and Thermal Variability of the Ocean."  In
    Meier et al.,  Glaciers.  Ice Sheets, and Sea Level.  Washington, D.C.:
    National Academy Press.

Roemmich, D., and C. Wunsch.  1984.  "Apparent Changes in the Climatic State
    of the Deep North Atlantic Ocean."  Nature 307:447-450.

Schneider, S.H., andR.S.  Chen.  1980.  "Carbon Dioxide Flooding:  Physical
    Factors and Climatic Impact."  Annual Review of Energy 5:107-140.

Sorensen, R.M., R.N. Weisman, and G.P. Lennon.  1984.  "Control of Erosion,
    Inundation, and Salinity Intrusion."  In Earth and Titus (eds.) 1984,  op.
    cit.

Stearns, L.A.,  and B. MacCreary.  1957.   "The Case of the Vanishing Brick
    Dust."  Mosquito News 17:303-304.

Thomas, R.H.  1985.  "Responses of the Polar Ice Sheets to Climatic
    Warming."  In Meier et al.  1985, op.  cit.

Thomas, R.H., T.J.O. Sanderson, and K.E.  Rose.  1979.  "Effect of Climatic
    Warming on the West Antarctic Ice Sheet."  Nature 227:355-358.

-------
                                     15-40
Titus, J.G.  1984.  "Planning for Sea Level Rise Before and After a
    Coastal Disaster."  In Earth and Titus (eds.) 1984, op. cit.

Titus, J.G.  1985.  "Sea Level Rise and the Maryland Coast."  In
    Potential Impacts of Sea Level Rise on the Beach at Ocean City. Maryland.
    Washington, B.C.:  U.S. Environmental Protection Agency.

Titus, J.G.  n.d.  "Greenhouse Effect, Sea Level Rise, and Society's
    Response."  In  Devoy, R.J.,  Sea Surface Studies. Cork, Ireland:
    University College.

Titus, J.G. 1986.  "Greenhouse Effect, Sea Level Rise, and Coastal Zone
    Management."  In Coastal Zone Management Journal. Volume 14, Number 3.

Titus, J.G.  1987.  "Greenhouse Effect, Sea Level Rise, and Wetland Loss."
    In Greenhouse Effect.  Sea Level Rise, and Coastal Wetlands.  Washington,
    D.C."  U.S. Environmental Protection Agency.

Titus, J.G., T. Henderson, and J.M. Teal.  1984.  "Sea Level Rise and
    Wetlands Loss in the United States."  National Wetlands Newsletter 6:4.

U.S. Army Corps of Engineers.  1982.  Louisiana Coastal Area.  Louisiana:
    Freshwater Diversion to Barataria and Breton Sound Basins.   New Orleans:
    U.S. Army Corps of Engineers.

U.S. Army Corps of Engineers.  1980.  Feasibility Report and Final
    Environmental Impact Statement:  Atlantic Coast of Maryland and Assateague
    Island, Virginia.  Baltimore:  U.S. Army Corps of Engineers.

U.S. Army Corps of Engineers.  1979.  Ocean Beach Study:  Feasibility
    Report.  San Francisco:  U.S. Army Corps of Engineers.

U.S. Army Corps of Engineers, 1977. Shore Protection Manual.  Fort Belvoir,
    Virginia:  Coastal Engineering Research Center.

U.S. Army Corps of Engineers.  1971.  National Shoreline Study.  Washington,
    D.C.:  U.S. Army Corps of Engineers.

Van Beek, J.L., D. Roberts, D. Davis,  D. Sabins, and S.M.  Gagliano.  1982.
    Recommendations for Freshwater Diversion to Louisiana Estuaries East of
    the Mississippi River.  Baton Rouge:  Coastal Management Section,
    Louisiana Department of Natural Resources.

Waddell, J.O., and R.A. Blaylock.  n.d.  "Impact of Sea Level  Rise on Gap
    Creek Watershed in the Fort Walton Beach, Florida Area."  In Kuo, C.Y.
    (ed.),  op. cit.

Walsh, J.E., and C.M. Johnson.  1979.   "Interannual Atmospheric Variability
    and Associated Fluctuations in Arctic Sea-Ice Extent."   Journal of
    Geophysical Research 84:6915-6928.

Ward,  L.G., and D.D.  Domeracki.   1978.  "The Stratographic  Significance of
    Back-Barrier Tidal Channel Migration."  Geological Society of America
    10(4):201.

-------
                                     15-41
Weishar, L.L., and W.L. Wood.  1983.  "An Evaluation of Offshore and Beach
    Changes on a Tideless Coast."  Journal of Sedimentary Petrology
    53(3):847-858.

Wendland, W.M.  1977.  "Tropical Storm Frequencies Related to Sea Surface
    Temperatures."  Journal of Applied Meteorology 16:480.

Zwally, H.J., C.L. Parkinson, andJ.C. Comiso.  1983.  "Variability of
    Antarctic Sea Ice and Changes in Carbon Dioxide."  Science 220:1005-1012.

-------
                                   CHAPTER 16

                  POTENTIAL EFFECTS OF FUTURE CLIMATE CHANGES
                 ON FORESTS AND VEGETATION, AGRICULTURE, WATER
                          RESOURCES, AND HUMAN HEALTH
SUMMARY

    The greenhouse effect resulting from increased levels of C02,
chlorofluorocarbons, methane, N20, and other trace gases in the atmosphere has
been recognized by the scientific community for several decades as a potential
cause of future climate change.  In the last few years, estimates  of the rate of
change of these gases in the atmosphere has heightened concern about global
warming and associated climate and environmental change.  Chapter 6 -- Global
Warming--presents a review of recent chemical and physical evidence supporting
the greenhouse phenomenon.  From this evidence it is generally concluded that in
the relatively short period of time of the next 50-100 years the earth's climate
will undergo significant changes.  These include potential increases in
temperatures, and changes in precipitation, humidity, windfields,  ocean
currents,  and in the frequency of extreme events such as hurricanes.
Furthermore, these climate parameters will induce still other shifts in sea
levels, ice margins, the hydrologic cycle, air pollution episodes  and other
phenomena.

    Recently (1986) the World Meterological Organization (WHO), the United
Nations Environment Program (UNEP),  and the International Council  of Scientific
Unions (ICSU) summarized current scientific data on global climate change.
These findings are presented in Exhibit 16-1.  Similar findings have been
reported by NAS (1979 and 1982).

                                  EXHIBIT 16-1

             Summary of Findings from the WMO/ONEP/ICSU Conference
            on  Global  Climate Held in Villach, Austria,  October 1985


    o   Many important economic and social decisions are being made
        today on long-term projects --  major water resource management
        activities such as irrigation and hydro-power, drought relief,
        agricultural land use,  structural designs and coastal
        engineering projects, and energy planning -- all based on the
        assumption that past climatic data, without modification,  are a
        reliable guide to future climate conditions.  This is no longer
        a valid assumption,  since the increasing concentrations of
        greenhouse gases are expected to cause a significant warming of
        the global climate in the next  century.   It is a matter of
        urgency to refine estimates of  future climate conditions to
        improve these decisions.

    o   The amounts of some trace gases in the troposhere,  notably
        carbon dioxide (C02), nitrous oxide (N20),  methane (CH4),  ozone
        (03), and chlorofluorocarbons (CFCs), are increasing.   These

-------
                              16-2
                          EXHIBIT 16-1
                          (continued)
gases are essentially transparent to incoming short-wave solar
radiation, but absorb and emit long-wave radiation and are thus
able to influence the earth's climate.

The role of greenhouse gases other than C02 in changing the
climate is already about as important as that of C02.  If
present trends continue, the combined concentrations of
atmospheric C02 and other greenhouse gases would be
(radiatively) equivalent to a doubling of C02 from
pre-industrial levels, possibly as early as the 2030's.

The most advanced experiments with general circulation models of
the climatic system show increases of the global mean
equilibrium surface temperature of between 1.5°C and 4.5°C for a
doubling of the atmospheric C02 concentration or the equivalent.
Because of the complexity of the climatic system and the
imperfections of the models, particularly with respect to
ocean-atmosphere interactions and clouds, values outside of this
range cannot be excluded.  The realization of such changes will
be slowed by the inertia of the oceans; the delay in reaching
the mean equilibrium temperatures corresponding to doubled
greenhouse gas concentrations is expected to be a matter of
decades.

While other factors such as aerosol concentrations, changes in
solar energy input, and changes in vegetation may also influence
climate, the increased amounts of greenhouse gases are likely to
be the most important cause of climate change over the next
century.

Regional scale changes in climate have not yet been modeled with
confidence.  However, regional differences from the global
averages show that warming may be greater in high latitudes
during  late autumn and winter than in the tropics; annual mean
runoff may increase in high latitudes, and summer dryness may
become more frequent over the continents at middle latitudes in
the Northern Hemisphere.  In tropical regions, temperature
increases are expected to be smaller than the average  global
rise, but the effects on ecosystems and humans could have far-
reaching consequences.  Potential evapotranspiration probably
will increase throughout the tropics, whereas in moist, tropical
regions, convective rainfall could increase.

Based on the observed changes since the beginning of this
century, it is estimated that global warming of 1.5°-4.5°C would
lead to a sea-level rise of 20-140 cm.  A sea-level rise in the
upper portion of this range would have major direct effects on
coastal areas and estuaries.  A significant melting of the West
Antarctic ice sheet leading to a much larger rise in

-------
                                     16-3
                                 EXHIBIT 16-1
                                  (continued)
        sea level,  although possible at some future date,  is not
        expected during the next century.

    o   Based on analyses of observational data,  the estimated increase
        in global mean temperature of between 0.3°C and 0.7°C during the
        last 100 years is consistent with the projected temperature
        increase attributable to the observed increase in C02 and other
        greenhouse gases, although it cannot be ascribed in a
        scientifically rigorous manner to these factors alone.

    o   Based on evidence of the effects of past climate changes, there
        is little doubt that a future change in climate of the magnitude
        obtained from climate models could have a profound effect on
        global ecosystems, agriculture, water resources, and sea ice.


    As noted in the WMO/UNEP/ICSU report,  the projected changes in climate will
have important impacts on all aspects of society.  Agriculture, forests, human
health, water resources, wildlife, energy planning, and recreation are among the
sectors likely to be affected.  Moreover,  all these sectors are likely to be
affected simultaneously throughout the world, but to different extents.  Today
we know a great deal from paleoclimatic records about how past shifts in climate
affected the growth and development of forest systems, the location of lakes,
and the development of agriculture.  But the changes that occurred to these
systems in the past took 18 to 20 thousand years to unfold as the earth warmed
approximately 4°-5°C.  During that time, forest composition shifted, some lake
systems were lost and new ones were formed.  Most important, the changes took
place during a period when the earth's population was small and civilizations
were in formative stages.

    Today modern society is much more complex, but still vulnerable to climatic
changes.  Our industrial society relies on a sustained climate to replenish
natural resources as a source of raw materials, for transport of goods, and for
the food we eat.  We assume that the climate that supports our society, while
variable and difficult to predict over short periods, will not shift
appreciably.  Indeed, most decisions made by farmers, forest managers, state and
local water management officials, utility executives and government policy
makers are based on the assumption that climate will be constant.  However, if
current predictions from global climate models prove to be correct, some
increase in global temperatures may be inevitable simply because of the presence
of trace gases that have already been emitted to the atmosphere.

    Our current understanding of the effects of global climate change on the
environment is incomplete.  Moreover, several features of the greenhouse
phenomenon make it unique, different from other environmental issues and
difficult to analyze.  Among these features are the following:

        o   The effects will not take place immediately, but over many
            decades.

-------
                                      16-4
        o   The effects will be virtually irreversible over several
            centuries.

        o   All nations of the world will experience the effects at
            the same time.

        o   There is no historical analog for the amount of global
            warming likely to occur in the relatively short period of
            the next 50-100 years.

    Scientists have only begun to analyze the potential impacts from global
warming and changes in other climate variables.   Insights are available from
historical data and from the application of predictive models.   In most cases,
however, our understanding of the consequences,  both beneficial and detrimental,
is in a formative stage.  Historical analogs provide qualitative information
about likely effects, but they cannot predict the future because the anticipated
rate and increase in temperature are beyond the range of previous warm periods.
Predictive models of both the climate system and potential effects often do not
include complete parameterizations of important system variables.  Thus, more
advanced global climate models capable of providing regional predictions are
needed, and more comprehensive and sophisticated analyses of environmental
effects are necessary to understand fully the implications throughout the world.

    Recognizing these limitations, the following section summarizes what is
known about climate impacts on the environment with emphasis on forests,
agriculture, water resources, and human health.   Perhaps of equal importance are
potential impacts that have not been reported or analyzed, including potential
impacts on ports, electricity demand and supply, population shifts, hurricane
frequency, air pollution emissions, wildlife management, and the U.S.'
competitive position in the world.  These potential impact areas and many others
represent the challenges to be investigated as the science supporting
predictions of global climate change improves.

-------
                                      16-5
FINDINGS

    The following findings are summarized from Appendix B of the Risk
Assessment, which provides a comprehensive review of potential impacts of global
climate change.

1.  CLIMATE CHANGE HAS HAD A SIGNIFICANT IMPACT ON FORESTS IN THE PAST.  IF
    CURRENT PREDICTIONS PROVE ACCURATE. THERE IS A POTENTIAL FOR DRAMATIC SHIFTS
    IN FORESTS AND VEGETATION OVER THE NEXT 100 YEARS.

    la.   Climate models predict that a global warming of approximately 1.5°C to
          4.5°C will be induced by a doubling of atmospheric C02 and other trace
          gases during the next 50 to 100 years.  The period 18,000 to 0 years
          B.P. is the only general analog for a global climate change of this
          magnitude.  The geological record from this glacial to inter-glacial
          interval provides a basis for qualitatively understanding how
          vegetation may change in response to large climatic change.

    Ib.   The paleovegetational record shows that climatic change as large as
          that expected to occur in response to C02 doubling is likely to induce
          significant changes in the composition and patterns of the world's
          biomes.  Changes of 2°C to 4°C have been significant enough to alter
          the composition of biomes, and to cause new biomes to appear and
          others to disappear.  At 18,000 B.P., the vegetation in eastern North
          America was quite distinct from that of the present day.  The cold,
          dry climate of that time seems to have precluded the widespread growth
          of birch, hemlock, beech, alder, hornbeam, ash, elm, and chestnut, all
          of which are fairly abundant in present-day deciduous forest.
          Southern pines were limited to grow with oak and hickory in Florida.

    Ic.   Available paleoecological and paleoclimatological records do not
          provide an analog for the high rate of climate change and
          unprecedented global warming predicted to occur over the next century.
          Previous changes in vegetation have been associated with climates that
          were nearly 5°C to 7°C cooler and took thousands of years to evolve
          rather than decades, the time during which such changes are now
          predicted to occur.  Insufficient temporal resolution (e.g., via
          radiocarbon dates) limits our ability to analyze the decadal-scale
          rates of change that occurred prior to the present millennium.

    Id.   Limited experiments conducted with dynamic vegetation models for North
          America suggest that decreases in net biomass may occur and that
          significant changes in species composition are likely.  Experiments
          with one model suggest that eastern North American biomass may be
          reduced by 11 megagrams per hectare (10% of live biomass) given the
          equivalent of a doubled C02 environment.  Plant taxa will respond
          individualistically rather than as whole communities to regional
          changes in climate variables.  At this time such analyses must be
          treated as only suggestive of the kinds of change that could occur.
          Many critical processes are simplified or omitted and the actual
          situation could be worse or better.

-------
                                     16-6
     le.   Future forest management decisions  in major timber-growing regions are
          likely to be affected by changes in natural growing conditions.   For
          example,  one study suggests that loblolly pine populations are likely
          to move north and northeast into Pennsylvania and New Jersey,  while
          its range shrinks in the west.   The total geographic range of  the
          species may increase,  but a net loss in productivity may result
          because of shifts to less accessible and less productive sites.   While
          the extent of such changes is unclear,  adjustments will be needed in
          forest technology,  resource allocation,  planning,  tree breeding
          programs, and decision-making to maintain and increase productivity.

    If.    Dynamic vegetation models based on  theoretical descriptions of all
          factors that could influence plant  growth must be improved and/or
          developed for all major kinds of vegetation.   In order to make more
          accurate future predictions,  these  models must be validated using the
          geological record and empirical ecological response surfaces.   In
          particular,  the geological record can be used to test the ability of
          vegetation models to simulate vegetation that grew under climate
          conditions unlike any of the modern day conditions.

    Ig.    Dynamic vegetation models should incorporate direct effects of
          atmospheric C02 increases on plant  growth and other air pollution
          effects.   Improved estimates of future regional climates are also
          required in order to make accurate  predictions of future vegetation
          changes.

2.   LIMITED ASSESSMENTS SUGGEST THAT IMPORTANT CHANGES IN AGRICULTURE AND FARM
    PRODUCTIVITY ARE LIKELY THROUGHOUT THE WORLD IF CLIMATE CHANGE OCCURS AS
    PREDICTED.  ESTIMATES OF IMPACTS ON SPECIFIC REGIONS ARE DIFFICULT TO MAKE
    BECAUSE REGIONAL PROJECTIONS OF CHANGE CANNOT BE RELIABLY MADE.  CURRENT
    CLIMATIC KNOWLEDGE IS ONLY SUFFICIENT TO  SUPPORT VULNERABILITY STUDIES FOR
    ALTERNATIVE SCENARIOS.

    2a.    Climate has had a significant impact on farm productivity and
          geographical distribution of crops.  Examples include the 1983
          drought,  which contributed to a nearly 30 percent reduction in corn
          yields in the U.S.; the persistent  Great Plains drought between
          1932-1937, which contributed to nearly 200,000 farm bankruptcies; and
          the climate shift of the Little Ice Age (1500-1800), which led to the
          abandonment of agricultural settlements in Scotland and Norway.

    2b.    World agriculture is likely to undergo significant shifts if
          trace-gas-induced climate warming in the range of 1.5°C to 4.5°C
          occurs over the next 50 to 100 years.  Climatic effects on agriculture
          will extend from local to regional  and international levels.  However,
          modern agriculture is very dynamic  and is constantly responding to
          changes in production, marketing, and government programs.

    2c.    The main effects likely to occur at the field level will be physical
          impacts of changes in thermal regimes,  water conditions, and pest
          infestations.  High temperatures have caused direct damage to  crops
          such as wheat and corn; moisture stress, often associated with
          elevated temperatures, is harmful to corn, soybean, and wheat  during

-------
                                     16-7
          flowering and grain fill;  and increased pests are associated with
          higher,  more favorable temperatures.

    2d.    Even relatively small increases in the mean temperature can increase
          the probability of harmful effects in some regions.   Analysis of
          historical data has shown that an increase of 1.7°C  (3°F)  in mean
          temperature changes by about a factor of three the likelihood of a
          five-consecutive-day maximum temperature event of at least 35°C (95°F)
          occurring in a city like Des Moines.   In regions where crops are grown
          close to their maximum tolerance limits, extreme temperature events
          may have significant harmful effects  on crop growth  and yield.

    2e.    Limited experiments using climate scenarios and agricultural
          productivity models have demonstrated the sensitivity of agricultural
          systems to climate change.  Future farm yields are likely to be
          affected by climate because of changes in the length of the growing
          season,  heating units, extreme winter temperatures,  precipitation,  and
          evaporative demand.  In addition, field evaluations  show that total
          productivity is a function of the drought tolerance  of the land and
          the moisture reserve, the availability of land, the  ability of farmers
          to shift to different crops, and other factors.

    2f.    The transition costs associated with  adjusting to global climatic
          change are not easily calculated, but are likely to  be very large.
          Accommodating to climate change may require shifting to new lands and
          crops, creating support services and  industries, improving and
          relocating irrigation systems, developing new soil management and pest
          control programs,  and breeding and introducing new heat- or
          drought-tolerant species.   The consequences of these decisions on the
          total quantity, quality, and cost of  food are difficult to predict.

    2g.    Current projections of the effects of climate change on agriculture
          are limited because of uncertainties  in predicting local temperature
          and precipitation patterns using global climate models, and because of
          the need for improved research studies using controlled atmospheres,
          statistical regression models, dynamic crop models and integrated
          modeling approaches.

3.   WATER RESOURCE SYSTEMS HAVE UNDERGONE IMPORTANT CHANGES AS THE EARTH'S
    CLIMATE HAS SHIFTED IN THE PAST.  CURRENT ANALYSES SUGGEST AN INTENSIFIED
    HYDROLOGIC CYCLE. IF CLIMATE CHANGE OCCURS  AS PREDICTED.

    3a.    There is evidence that climate change since the last ice age (18,000
          years B.P.) has significantly altered the location of lakes --
          although the extent of present day lakes is broadly  comparable with
          18,000 years B.P.   For example, there is evidence indicating the
          existence of many tropical lakes and  swamps in the Sahara,  Arabian,
          and Thor Deserts around 9,000 to 8,000 years B.P.

    3b.    The inextricable linkages between the water cycle and climate ensure
          that potential future climate change  will significantly alter
          hydrologic processes throughout the world.  All natural hydrologic
          processes--precipitation,  infiltration, storage and  movement of soil

-------
                                     16-8
          moisture,  surface  and  subsurface  runoff,  recharge  of  groundwater,  and
          evapotranspiration--will be  affected if climate  changes.

    3c.    As  a result  of  changes in key  hydrologic variables such  as
          precipitation,  evaporation,  soil  moisture,  and runoff, climate  change
          is  expected  to  have  significant effects on  water availability.  Early
          hydrologic impact  studies provide evidence  that  relatively  small
          changes  in precipitation and evaporation patterns  might  result  in
          significant,  perhaps critical, changes  in water  availability.   For
          many aspects of water  resources,  including  human consumption,
          agricultural water supply, flooding and drought  management,
          groundwater  use and recharge,  and reservoir design and operation,
          these hydrologic changes will  have serious  implications.

    3d.    Despite  significant differences among climate change  scenarios, a
          consistent finding among hydrologic impact  studies is the prediction
          of  a reduction  in  summer soil  moisture  and  changes in the timing and
          magnitude  of runoff.  Winter runoff is  expected  to increase and summer
          runoff to  decrease.  These results appear to be  robust across  a range
          of  climate change  scenarios.

    3e.    Future directions  for  research and analyses suggest that improved
          estimates  of climate variables are needed from large-scale  climate
          models;  innovative techniques  are needed for regional assessments;
          increased  numbers  of assessments  are necessary to  broaden our
          knowledge  of effects on different users; and increased analyses of the
          impacts  of changes in  water  resources on the economy and society are
          necessary.

4.   MORBIDITY AND.. MORTALITY iRATES ARE ASSOCIATED WITH _WEATHER EXTREMES  IN OUR
     SOCIETY (chapter  16).

     4a.   Weather has a  profound effect on human health and well  being.   It has
           been demonstrated that weather  is associated with changes  in birth
           rates,  outbreaks  of pneumonia,  influenza,  and bronchitis,  and related
           to other  morbidity effects, and  is linked  to pollen concentrations
           and high  pollution levels.

     4b.   Large increases in mortality  have occurred during previous heat and
           cold waves.  It is estimated  that 1,327 fatalities occurred in the
           United States  as  a result  of  the 1980  heat wave,  and Missouri alone
           accounted for  over 25 percent of that  total.

     4c.   Hot weather extremes  appear to  have a  more substantial  impact on
           mortality than cold wave  episodes.

     4d.   Threshold temperatures,  which represent maximum and minimum
           temperatures associated with increases in total mortality, have been
           determined for various cities.   These  threshold temperatures vary
           regionally; for example,  the  threshold temperature for winter
           mortality in mild  southern cities such as  Atlanta is 0°C and for more
           northerly cities such as  Philadelphia, threshold temperature is  -5°C.

-------
                                16-9
4e.   If future global warming induced by increased concentrations of trace
      gases does occur, it has the potential to affect human mortality
      significantly.   In one study,  total summertime mortality in New York
      City was estimated to increase by over 3,200 deaths per year for a
      7°F trace-gas-induced warming without acclimatization.   If New
      Yorkers fully acclimatize,  the number of additional deaths is
      estimated to be no different than today.  It is hypothesized that if
      climate warming occurs,  some additional deaths are likely to occur
      because economic conditions and the basic infrastructure of the city
      will prohibit full acclimatization even if behavior changes.

-------
                                     16-10
REFERENCES

National Academy of Sciences (1979), Carbon Dioxide and Climate:   A Scientific
Assessment. National Academy Press, Washington, B.C.

National Academy of Sciences (1982), Carbon Dioxide and Climate:   A Second
Assessment. National Academy Press, Washington, D.C.

World Meterological Organization, United Nations Environment Programme,  and
International Council of Scientific Unions (1986), Report of the International
Conference on the Assessment of the Role of Carbon Dioxide and of Other
Greenhouse Gases in Climate Variations and Associated Impacts. Villach Austria,
9-15 October 1985, WMO No. 661.

-------
                                   CHAPTER 17

               MODELS FOR INTEGRATING THE ANALYSES OF HEALTH AND
             ENVIRONMENTAL RISKS ASSOCIATED WITH  OZONE MODIFICATION
SUMMARY

    Analyses of the potential risks due to emissions of ozone-modifying
substances and greenhouse gases must assess a variety of effects on human health
and the environment.  To assess risks numerous factors must be examined
together,  including: economic and population growth; chemical use and emissions;
ozone modification; changes in ultraviolet (UV) radiation flux;  climate change;
and effects on human health and the environment.

    To evaluate risks the Integrating Analysis Model has been developed to
provide a framework for systematically evaluating the risks and costs associated
with alternative assumptions about the ozone modification issue and policies for
limiting the emissions of ozone-modifying substances.  This chapter describes
the manner in which the Integrating Analysis Model (the model) draws from the
analyses described in earlier chapters.

    The model begins with assumptions about the potential future use of
ozone-modifying substances and concentrations of other trace gases, including
chlorofluorocarbons, methyl chloroform,  carbon tetrachloride,  Halons
(bromine-containing compounds), carbon dioxide, methane, and nitrous oxide.
This potential future use and emissions may be modified by control strategies.
The resulting global emissions of the substances are computed, followed by the
atmospheric science module, which assesses the impacts of these emissions on
ozone.  The risks of ozone modification for human health and the environment are
evaluated using specified dose-response relationships where available for each
of the areas of interest.

    Using this overall framework, the model can reflect a wide range of
assumptions and the joint implications of all the assumptions can be identified.
Of note is that the model is not a substitute for choosing the assumptions,
relationships, and data to describe the system being modeled.   It merely
reflects these assumptions and data in a consistent framework.
     •*- When used for performing policy analysis, the model does not make
decisions,  for example, about which, if any, control strategies are preferred or
about which,  if any, ozone depletion estimates are correct.   Decisions about the
need to protect stratospheric ozone or the best way to do so must be based not
only on model results,  but also on considerations and judgments either not
reflected in the model  or reflected implicitly in the assumptions provided as
inputs to the model.

-------
                                     17-2
INTRODUCTION

    Analyses of the potential risks due to emissions of ozone-modifying
substances and greenhouse gases must assess a variety of effects on human health
and the environment.   As described earlier in this document, numerous factors
influence ozone modification and its effects.  To assess risks,  the following
must be characterized, and their joint implications (including their
uncertainties) examined together:  economic and population growth;  chemical use
and emissions; ozone modification; changes in ultraviolet (UV) radiation flux;
climate change; and effects on human health and the environment.

    Analyzing and combining these factors is a complex task.  To make this task
tractable, two models have been developed:

        1.  Integrating Analysis Model.  The primary role of the
            Integrating Analysis Model is to provide a framework for
            systematically evaluating the risks and costs associated
            with alternative assumptions about the ozone modification
            issue and policies for limiting the emissions of
            ozone-modifying substances.

        2.  Transient Temperature and Sea Level Rise Model.   The
            primary role of this model is to assess the implications
            of emissions of greenhouse gases in terms of increases in
            global average temperature and sea level over time.

Together, these two models provide a comprehensive framework for evaluating the
potential risks of emissions of ozone-modifying substances and greenhouse gases.
This chapter describes the manner in which the Integrating Analysis Model (the
model) draws from the analyses described in earlier chapters.   The Transient
Temperature and Sea Level Rise Model is described in Appendix A of Chapter 6.

    This  chapter first describes the framework that the model provides for the
analysis.  Then, an overview of the model's analytic procedure is presented.
The third section of  this chapter presents the major limitations of the model.
The chapter concludes with a series of appendices that describe the model's
design and operation  in detail.

THE MODEL AS A FRAMEWORK

    The objective of  the model is to provide a framework within which the
implications of alternative assumptions and policies can be  identified.  The
framework reflects the order and structure of the analysis,   and the user may
define the assumptions and data to be used in the model for  any given run.
Exhibit 17-1 displays the major parts or  "modules" of the model.  The model
begins with assumptions about  the potential  future use of ozone-modifying
substances and concentrations  of other trace gases, including chlorofluoro-
carbons,  methyl chloroform, carbon tetrachloride, Halons  (bromine-containing
     ^ The Integrating Analysis Model was originally  described  in Gibbs  and
Hoffman  (1986).

-------
                                  EXHIBIT 17-1
                               MODULAR STRUCTURE
Potential Future
Use of Ozone-Depleting
Substances and
Concentrations of
Other Trace Gases
                             Global
                             Emissions
Atmospheric
Science

Health and
Environmental
Effects
dular Structure
      Control
      Strategy

-------
                                      17-4
compounds), carbon dioxide, methane, and nitrous oxide.  This potential future
use and emissions may be modified by control strategies.   The resulting global
emissions of the substances are computed, followed by the atmospheric science
module.  which assesses the impacts of these emissions on ozone.    The risks of
ozone modification for human health and the environment are evaluated using
specified dose-response relationships where available for each of the areas of
interest.

    Using this overall framework, the model can reflect a wide range of
assumptions within each of the modules.  When the model is run,  the joint
implications of all the assumptions are identified.  Of note is  that the model
is not a substitute for choosing the assumptions, relationships,  and data to
describe the system being modeled.  It merely reflects these assumptions and
data in a consistent framework.  In its simplest terms, the model is a
calculator that is needed to assist in computing multiple estimates of possible
values.   Decisions about best assumptions or the need to protect stratospheric
ozone or the best way to do so may be made using the model and its results, but
must also be based on considerations and judgments either not reflected in the
model or reflected implicitly in the assumptions provided as inputs to the
model.

ANALYSIS PROCEDURE

    A model "run" is comprised of:

        o   inputs provided by the user that define the data and
            assumptions to use in the analysis;

        o   algorithms that are carried out according to the user's
            inputs and instructions contained within the model;  and

        o   outputs prepared by the model that describe the results of
            the algorithms that are carried out.

The model is "deterministic," meaning that a unique output or result is produced
by a specified input; there are no random (i.e., stochastic) elements within the
model.

    To initiate a model run, the user must specify inputs for each module he
wishes to use.  (The user may choose to run only a portion of the model.)
Exhibit 17-2 provides an overview of the input choices for each of the following
modules:

        o   Potential Use.  The user may choose  one of the six
            scenarios of future use that have been incorporated into
            the model.  The compounds included are:  chlorofluoro-
            carbons  (CFC-11; CFC-12; HCFC-22; CFC-113); carbon-
            tetrachloride  (CCL4); methyl chloroform (CH3CCL3);
            Halon-1211; and Halon-1301.  Each compound can be
     3 These emission rates are also used  in the Transient Temperature and Sea
Level Rise Model to evaluate climate change effects.

-------
                                                                 17-5

                                                           EXHIBIT  17-2
                                                 Major Model  Input:  Choices
        MODULE
                                   CHOICES
                                                                      OMffiHTS
                                                                                                          SOURCES
 Potential Future Use
 of Ozone-Modifying
 Substances
Future Use:

— Scenarios based on studies of
   historical trends  and  projec-
   tions of economic  growth  and
   population growth

— Any user-supplied  estimates of
   future use can be  incorporated
Future Use:

— Scenarios  reflect a wide range
   of potential  future use.
Based on:   UNEP (1986).
Control Strategies
Strategies can be defined  in
terms of limits to production in
various ways,  including:

— limits in kilograms;

— limits in kilograms per
   capita;

— bans of specified uses;

— requirements to implement
   specified technologies; and

— fees or taxes.
                                                          Data on  control options are based
                                                          on U.S.  data.
                                   Estimates of the  costs and effec-
                                   tiveness  of controls are based on
                                   data developed  in EPA (1987).
Global Emissions
                      Several estimates of the  rate of
                      release after use have  been
                      developed based on analyses of
                      historical data.   Emissions
                      reflect the effectiveness of the
                      control strategy  in limiting
                      emissions.   The user may  choose
                      several scenarios of effective-    j
                      ness, which reflects the  level of  [
                      reductions  expected to  occur as    |
                      the result  of the control strate-  |
                      gy identified.                     I
                                   Data on  the  effectiveness of con-
                                   trol options are based on U.S.
                                   data.  Information on other parts
                                   of the world may be entered by
                                   the user.
                                   The rates of emissions following
                                   use are based on data presented
                                   in Quinn (1986).

-------
                                                               17-6

                                                 EXHIBIT 17-2  (continued)

                                                MAJOR MODEL INPUT  CHOICES
        MODULE
                                  CHOICES
                                                                    COMMENTS
                                                                                                       SOURCES
Atmospheric Science
Ozone Modification:

— Parameterized relationship
   between global emissions  and
   global ozone depletion.

— The user may specify his  own
   estimate of ozone depletion
   for use in the model.   If this
   choice is made, the global
   emissions estimate is not
   used.
Ozone Modification:  Global
estimates based on analysis of
one-dimensional atmospheric
model.
Based on Connell  (1986).
Health and Environ-
mental Effects
Ranges of  relationships with UV
radiation  are developed for each
effect.  The user may specify for
each effect low, medium, or high
estimates.
Ranges based on published esti-
mates.
Based on analyses by:  Scotto,
Fears,  and Fraumeni  (1981),
Scotto and Fears (in press),
Hiller,  Sperduto, and Ederer
(1983),  Horst  (1986), Andrady
(1986),  and  Pitcher  (in press).

-------
                                      17-7
            analyzed for up to six end uses, in 10 regions of the world.  The
            six scenarios reflect a wide range of potential future use, and are
            based on historical trends and projections.   Alternatively, the
            user may specify his own scenario of future use.

            Control Strategy.  The user may specify a variety of controls on
            future use, such as kilogram limits, use limits, technology
            requirements, or fees.  The potential costs and benefits of
            controls are not evaluated in this document because the objective
            of this risk assessment is to evaluate risks in the absence of
            regulation.

            Global Emissions.  Emissions of the chlorofluorocarbons, CCL4,
            CH3CCL3, Halon-1211 and Halon-1301 are computed from the production
            and use estimates based on release rates that reflect the fact that
            these chemicals are stored in certain products for various periods
            of time.  The effects of control strategies are reflected in the
            release estimates, and the user may specify alternative release
            rates as desired.  Scenarios of atmospheric concentrations are also
            specified for carbon dioxide (C02),  methane (CH4) and nitrous oxide
            (N20).

            Atmospheric Science.  A parameterized numerical fit to a
            one-dimensional (1-D) atmospheric model is used to compute global
            ozone depletion as a function of global emissions of
            chlorofluorocarbons, CC14, CH3CC13,  and Halons,  and concentrations
            of C02, N20 and CH4.5  The uncertainty in the estimates of global
            depletion may be modeled.   Alternatively, the user may bypass this
            algorithm and use a different relationship or independent estimates
            of global ozone depletion.

            Health and Environmental Effects.  The association between ozone
            depletion, ultraviolet (UV) radiation, and effects may be specified
            by the user.  Ranges of dose-response relationships may be
            identified for melanoma and nonmelanoma skin cancers,  and
            cataracts.  Population projections (by race, age, and sex) are
            incorporated into the model for purposes of computing future
            incidence of human health effects.  Alternative population
            estimates may be identified by the model user.  The effects on PVC
            polymers are also reported.  Potential impacts on aquatic
            organisms, plants, ground-based ozone, and coastal areas are also
            identified based on existing case study data.   Estimates of effects
            are made for the U.S. only.
     ^ These six scenarios are described in Chapter 18.

       See Connell (1986) for a description of this method.

       The range of uncertainty identified in Stolarski and Douglass (1985) is
currently incorporated into the model.  These uncertainties reflect the results
of a Monte Carlo analysis in which kinetic rate coefficients and cross sections
are varied according to the JPL 85-37 panel.   Consequently, not all uncertain-
ties are included.  The user may provide alternative uncertainty estimates.

-------
    Based on these inputs,  the model performs calculations on a year-by-year
basis.  For example,  at the beginning of year t,  the model first computes the
potential use of each of the ozone-modifying substances in each of its
applications in each portion of the world being analyzed.   This estimate of use
reflects the user's choice  of scenario (for the first module) as well as the
effects of any controls initiated according to the user's  instructions prior to
and including year t.   Based on this estimate of use in year t, and estimates
for all years prior to year t, the global emissions in year t are computed for
each potential ozone-depleting substance.

    The atmospheric science module computes an estimate of global ozone
depletion for year t based  on the emission estimates for year t, atmospheric
conditions due to emissions prior to year t, and user-specified assumptions
regarding the concentrations of carbon dioxide, methane, and nitrous oxide.
Global ozone depletion is computed and then translated into associated changes
in UV flux, using three action spectra for weighting the different UV
wavelengths:  DNA damage; Robertson-Berger Meter (RB-Meter),  and erythema.  The
changes in UV flux in year  t (and prior years) are used in dose-response
relationships to estimate increased risks of melanoma and nonmelanoma skin
cancers and cataracts.  Other impacts, such as polymer degradation, are also
computed.

    After the computations  for year t are performed, the model repeats the
computations for year t+1.   This annual series of computations is consequently
repeated for each year of the model run.  After all the years have been
analyzed, the following summary outputs are produced:

        o   future production of ozone-depleting substances;

        o   future emissions of ozone-depleting substances;

        o   future global ozone depletion;

        o   future ozone depletion by latitude;

        o   future increases in UV flux; and

        o   estimates of effects, including:

                melanoma skin cancers (incidence and mortality);
                nonmelanoma skin cancers (incidence and mortality);
                cataracts (incidence);
                costs of damages to polymers  (PVC used outdoors); and
                potential impacts on ground-based ozone, plants,
                aquatic organisms, and coastal areas.

The effects estimates are currently based primarily on data that describe the
U.S.  Consequently, only U.S. effects estimates are currently produced.
       Ranges of assumptions for these gases are described in Chapter 4.

-------
                                          17-9
 MODEL LIMITATIONS

     In using  any model,  it  is  important  to  understand its  limitations.   The  model  is
 being developed in a  neutral fashion.  Its  flexible  design allows  users  to evaluate
 wide ranges of inputs and assumptions.   The model  cannot  incorporate, however,  numerous
 factors that  are as yet  undefined or poorly understood.   Consequently, the model
 excludes a variety of potential  factors  and analyses that  may  be important for  making
 decisions regarding stratospheric ozone  protection.

     The most  important exclusions from the  model are a variety of  effects of ozone
 modification.  Many effects that increased  UV radiation will have  on human health  and
 the  environment are not  well quantified.  For example, only limited data are available
 that describe the  effects of UV  radiation on plants  and aquatic organisms.   These  data
 cannot necessarily be extrapolated to all crops and  ecosystems to  evaluate the  net
 effects of ozone modification  and global warming;  actual damages could be higher or
 lower.   The potential for effects is substantial,  but current  data do not allow
 reliable quantitative estimates  at this  time.  Similarly,  the  effects of UV  radiation
 on ground oxidants are not well  quantified  at this time for most cities  or rural areas
 (see Whitten  (1986) for  case studies of  effects on oxidants).

     Similarly, too little knowledge currently exists to allow  immune suppression
 effects of increased  UV  radiation to be  quantified.   These  effects may be particularly
 important because  they may affect all individuals  in the world in  important  ways that
 are  not currently  anticipated.

     Exhibit 17-3 summarizes the  effects  for which  reliable  projections are not  yet
 possible.  The omission  of these known effects would be a  clear bias reflected  in  the
 estimates of  the risks associated with ozone  depletion.  In addition, there  may be
 other links among  UV  radiation,  global warming, and  human health and the environment
 that have not yet  been identified.  The potential  bias of  omitting the results  of  these
 latter areas due to lack of information on  these effects is unknown.

     The model could be used to address issues  related to risk  management and policy
 analysis.   However,  this document does not  address  these  issues because it  is  focused
 on risk assessment.
       The model could be used to address a variety of risk management issues,
including:  (1) the feasibility of alternative control strategies; and (2) the number
of countries likely to participate in international agreements to limit the use and
emissions of ozone-depleting substances.  It does not have the capability, however, to
evaluate potential trade of ozone-modifying substances among regions of the world.

    Decision makers must use their own judgments regarding these factors.   The model
may be used to test the implications of these judgments (e.g., the implications of only
half the countries participating in a control strategy may be identified by specifying
the expected participation as a model input).  However, the model is not meant to be
used in place of these types of judgments, or any others,  that are required of decision
makers.

    Similarly,  extrapolation of case study results to a wider scope is unlikely to
yield reliable estimates of damages:   they may be Loo high or too low.  Nonetheless,  it
is a question of judgment: whether examination of such output adds information to the
decision-making process.

-------
                                     17-10


                                 EXHIBIT  17-3

                            Effects Not Quantified
              Effect
               Comments
Impact on plants
Impact on aquatic organisms
National impact on urban and rural
ground-based ozone formation (smog)
Increased incidence of cutaneous
diseases as a result of immune
suppression

Impacts of climate change on human
health, water resources,  forestry,
and agriculture

Impacts of sea level rise on wetlands

Impacts on polymers other than PVC
Preliminary experiment data indicate
risks of decreased yields.   See Chapter
11.

Laboratory data indicate increase UV
radiation may adversely affect
survivability and breeding.  See
Chapter 12.

Case studies of three cities indicate
potential increases in smog formation,
and formation earlier in the day.  See
Chapter 14.

See Chapter 9.
See Chapter 16.



See Chapter 15.

See Chapter 13.

-------
                                     17-11
REFERENCES
Andrady, A. (1986), "Analysis of Technical Issues Related to the Effect of UV-B
    on Polymers," prepared for the U.S. Environmental Protection Agency,
    Washington, B.C.

Camm, F. et al. (1986), "Social Cost of Technical Control Option to Reduce
    Emissions of Potential Ozone Depleters in the United States: An Update,"
    The RAND Corporation,  prepared for the U.S.  Environmental Protection
    Agency, Washington, D.C.

Connell, Peter S. (1986),  "A Parameterized Numerical Fit to Total Column Ozone
    Changes Calculated by the LLNL 1-D Model of the Troposphere and
    Stratosphere," Lawrence Livermore National Laboratory, Livermore,
    California.

EPA (1987).  Draft Regulatory Impact Analysis:  Protection of Stratospheric
Ozone.  Office of Air and Radiation, U.S.  Environmental Protection Agency,
Washington, D.C.

Gibbs,  M.J. and J.H. Hoffman (1986), "STRATO:  A Model for Analyzing the
    Implications of Control Strategies for Protecting Stratospheric Ozone,"
    presented at the EPA Workshop on CFCs,  March 1986.

Hiller, R., R. Sperduto,  and F. Ederer (1983), "Epidemiologic Associations
    with Cataract in the 1971-1972 National Health and Nutrition Examination
    Survey," American Journal of Epidemiology. Vol. 118, No. 2, pp. 239-248.

Horst,  R.L. (1986), "The Economic Impacts of Increased UV-B Radiation on
    Polymer Materials:  A-Case Study of Rigid PVC," prepared for the U.S.
    Environmental Protection Agency, Washington, D.C.

Isaksen, I.S.A. (1986), "Ozone Perturbations Studies in a Two-Dimensional Model
    with Temperature Feedback in the Stratosphere Included," presented at UNEP
    Workshop on the Control of Chlorofluorocarbons, Leesburg, Virginia,
    September 1986.

Pitcher, H.  (1986), "Melanoma Death Rates and Ultraviolet Radiation in the
    United States," U.S.  Environmental Protection Agency, Washington,  D.C.

Quinn,  T.H., et al. (1986),  "Projected Use,  Emissions, and Banks of Potential
    Ozone-depleting Substances," The RAND Corporation, prepared for the U.S.
    Environmental Protection Agency, Washington, D.C.

Scotto, J., T.R. Fears, and J.F. Fraumeni,  Jr. (1981), "Incidence of
    Nonmelanoma Skin Cancer in the United States," U.S.  Department of Health
    and Human Services, (NIH) 82-2433,  Bethesda, Maryland.

Scotto, J., and T. Fears  (in press), "The Association of Solar Ultraviolet
    Radiation and Skin Melanoma Among Caucasians in the United States," Cancer
    Investigations.

-------
                                     17-12
Stolarski, R.S., and A.R. Douglas.  (1985), ''Sensitivity of an Atmospheric
    Photochemistry Model to Chlorine Perturbations Including Consideration of
    Uncertainty Propagation," NASA/Goddard Space Flight Center, Greenbelt,
    Maryland, and Applied Research Corporation, Landover, Maryland.

UNEP (1986), "Summary Paper for Topic #2:  Projections of Future Demand,"
    presented at the UNEP Workshop on the Control of Chlorofluorocarbons, Rome,
    Italy, May 1986.

-------
                                   APPENDIX A

                          MODEL DESIGN AND MODEL FLOW
INTRODUCTION

    This appendix describes the design and flow of the CFG Policy Model (the
model).  The model is deterministic -- it computes a series of outputs from data
supplied by the model user and from data that reside in files associated with
the model.  A model "run" is comprised of (1) a set of instructions (i.e.,
inputs) supplied by the model user; (2) a set of computations performed within
the model; and (3) a set of output tables that present the results of the
computations performed by the model.   The model is designed with a modular
structure, with each "module" performing a designated function.  The modules are
independent entities that communicate with each other.

    The modular structure facilitates the maintenance and development of the
model by:

        o   allowing desired modifications to the model to be
            localized to individual modules;

        o   allowing communication among modules to be identified
            easily (communication among modules is particularly
            important for modeling feedbacks  from one portion of the
            model to another portion);  and

        o   minimizing memory requirements and execution time.

    The remainder of this appendix is divided into the following sections:

        o   Level of Aggregation;
        o   Model Flow;
        o   Specifying a Model Run; and
        o   Limitations.

LEVEL OF AGGREGATION

    The smallest unit of analysis  in the model includes compound,  region,  end
use,  and time,  as follows:

        o   Compound - -  Up to ten  compounds may be analyzed
            separately:

            --   CFC-11;
            --   CFC-12;
            --   HCFC-22;
            --   CFC-113;
            --   CFC-114;

-------
                                      A-2
            --  CFC-115;
                Carbon tetrachloride (CGI,);
            --  Methyl chloroform (CH CCLT) ;
            --  Halon-1211;  and
            --  Halon-1301.

            Data are currently not analyzed for CFC-114 and CFC-115.

        o   Region -- The world is divided into 10 separate regions
            (the regions  can be analyzed individually or in groups):

            --  USA;
                Canada;
                Western Europe;
                Japan, Australia,  and New Zealand;
            --  USSR and East Bloc;
                Centrally Planned Asia;
            --  Middle East;
                Africa;
                Latin America; and
                South and East Asia.

        o   End Use -- Each compound may be identified with up to six
            applications  or "end uses."   Each end use can be analyzed
            separately.   End uses currently incorporated into the
            model include:

                aerosol propellant;
                flexible  foam;
                rigid polyurethane foam;
                rigid nonurethane foam;
                refrigeration;
                solvent;
                fire extinguisher; and
                miscellaneous.

        o   Time --a year is the smallest unit of time in this model.
            The current range is 1931-2165.

Analysis is performed at these levels of information.  Major revisions to the
model are required to modify this level of disaggregation.

MODEL FLOW

    The model is divided into two main parts:  (1) the Analysis Program reads
input data, performs computations, and creates output files; (2) the Report
Writer Program reads output files from the Analysis Program, and creates
formatted reports as specified by the user.  This section discusses each part of
the model in turn.

-------
                                      A-3
Analysis Program

    The Analysis Program begins by reading an input file that describes the
configuration of the model run desired by the user, including:

        o   which years will be analyzed;

        o   the portions of the model (i.e., modules) that will be
            used in the analysis;

        o   the data files that will be used as inputs;

        o   the years for which output values should be computed; and

        o   the discount rates used for estimating the present value
            of costs and benefits.

Based on these user-supplied data, a series of variables are initialized and the
model run specified by the user then proceeds.

    The flow of a model run can be described in terms of the following five
primary parts or "modules" of the Analysis Program:

        1.  Production Scenarios Module;
        2.  Policy Alternatives Module;
        3.  Emissions Module;
        4.  Atmospheric Science Module;  and
        5.  Effects Module.

As shown in Exhibit A-l, these five modules are all executed within a "year
loop," i.e., all the information from all the modules is computed for one year,
then the model moves on to the next year.  This structure has several benefits,
including:

        o   the results of year "t" can be used to indicate what would
            happen in year "t+1" (in other words,  feedbacks between
            modules and across years can be modeled);

        o   the information required to be stored is limited to the
            current year being analyzed.

Appendices B through F describe each of the modules in turn.  Briefly, each
module performs the following computations:

        o   Appendix B:   Production Scenarios Module - - This module
            computes a scenario of production over time for each
            compound.   The scenario for each compound is disaggregated
            by region and end use.

        o   Appendix C:   Policy Alternatives Module -- This module
            identifies the implications of alternative control
            policies specified by the user.   The module's inputs
            include policy specifications (from the user),  production
            scenarios  (from the previous module),  and candidate

-------
                                           A-4




                                       EXHIBIT A-l


                                Flow of Analysis Program
Year
Loop
                    Input
                    Files
                  Production
                  Scenarios
                   Module
                    Policy
                 Alternatives
                   Module
                                 "*]
                        I
                        I	
Emissions
 Module
                        I
                        I	
                 Atmospheric
                   Science
                   Module
                        I	
                                  -H
                    Effects
                    Module
                        I	
Summary
File
^ ^^


Report
Writer
Program


Output
File
- ^^

-------
                                      A-5
            technical  options  for controlling the  use  of potential
            ozone-depleting substances  (identified in  an input  file).
            The final  results  of this module  are  (1) the incurred
            costs  of control attributable  to  the policy;  and (2)  the
            reductions in production realized as a consequence  of
            control.

        o   Appendix D:   Emissions Module  - -  This  module calculates
            global emissions of each compound by applying a release
            algorithm  to the production scenarios  (which have been
            modified by the Policy Alternatives Module).   The release
            algorithm  specifies the time it  takes  (after production)
            for the compound to be released  for each end use of each
            compound.   The algorithm reflects the  fact that the
            chemicals  are stored in products  for many  years.

        o   Appendix E:   Atmospheric Science  Module -- This module
            uses the global emissions  estimates from the Emissions
            Module and estimates of atmospheric concentrations  of
            carbon dioxide (C02), methane  (CH4) and nitrous oxide
            (N20)  to calculate global  and  latitudinal  ozone depletion.
            Ozone depletion is calculated  using parameterized
            equations  that represent a one-dimensional model.
            Uncertainty is also modeled based on  investigations of
            uncertainty propagation through  an atmospheric model.

        o   Appendix F:   Effects Module -- This module projects the
            effects that ozone depletion may have  on human health and
            the environment.  Human health effects are focused  on
            melanoma and nonmelanoma skin  cancers  and  cataracts.
            Other effects include degradation of  materials, soybean
            yield impacts based on the Essex and  Williams soybean
            cultivars, and ground-based ozone levels in three cities
            (which are used to estimate a  'national'  average).

    The results of each module are written to a "Summary File."  This Summary
File is used by the Report Writer Program to create detailed tables  as specified
by the user.

Report Writer Program

    The Report Writer  Program reads in the data stored in the Summary File and
creates reports for the model run as specified by the  user.  The reports for a
given model run may show results disaggregated by region, compound,  end use, and
time.  Results across  runs may also be displayed  and compared.   Of note is that
once a Summary File is created by the  Analysis Program,  the Report Writer
Program can be run several times to generate reports at various levels of
aggregation without re-running the Analysis Program.

SPECIFYING A MODEL RUN

    To specify a model run the user fills  out a set of data files that describe
the assumptions and data to be used.  The names of these data files, and their
purposes, are listed in Exhibit A-2.  Each is described in turn.

-------
                                      A-6


                                  EXHIBIT A-2

                        Files Required to Specify a Run
        File Name
                 Purpose
Run File
To identify modules included in the run
and data files to be used, specify
years for which outputs are desired,
and define the discount rates used to
calculate costs and benefits.
Production Scenarios File
To choose a scenario of future
production for each compound and of
future population and economic growth.
Policy Alternatives File
To define the control strategy to be
evaluated and the scenario of available
technical controls.
Atmospheric Science File
To choose a method for estimating ozone
depletion and specify a range of
uncertainty surrounding the method.
Also to specify scenarios of
concentrations of other trace gases.
Effects File
To specify the dose-response
coefficients for each human health
effect, the action spectrum to be used
for each human health effect, and the
economic value to be placed on each
human health effect.  Also, to specify
the values used to estimate polymer
degradation.
Report Writer File
To identify runs for which reports are
to be written and the reports desired.

-------
                                      A-7
        o   The Run File identifies the overall model flow, including
            which modules are to be executed.  This file includes:

                the range of years for the model run;

                the names of modules to be executed;

                the names of the data files to be used;

                the years for which summary results will be retained;
                and

                the discount rates for cost/benefit analysis.

        o   The Production Scenarios File specifies the scenario of
            future production of potential ozone-depleting compounds
            to use in the analysis.  The user may choose one of a set
            of defined scenarios, or may create his own.  The user
            also chooses (or creates) a scenario of population and
            economic growth.

        o   The Policy Alternatives File defines the control strategy
            (if any) that the user is evaluating.  In addition to
            defining the level of controls or the policy objective,
            the user must choose among alternative scenarios that
            define the technical options available for reducing the
            use of potential ozone-depleting compounds.

        o   The Atmospheric Science File allows the user to choose
            among alternative methods for computing ozone depletion.
            The user may specify a range of uncertainty surrounding
            the ozone depletion estimate.  The assumptions regarding
            the concentrations of other trace gases (carbon dioxide,
            methane, and nitrous oxide) are also specified in this
            file.

        o   The Effects File identifies the assumptions used to (1)
            estimate effects (in their natural units); and (2)
            evaluate the effects in monetary terms.   The user chooses
            (or supplies) dose-response coefficients for each effect.
            The value (in dollars) of each effect may also be
            specified.

        o   The Report Writer File identifies the runs for which
            reports are to be written and the reports desired.

Of note is that once these files have been specified, and the model is run, the
entire model need not be re-run to perform additional analyses.  For example, to
analyze an alternative set of effects assumptions,  only the Effects File needs
to be filled out,  and only the Effects Module needs to be executed.  The results
from the previous  modules can be stored for future use.  This flexibility in the
modular design of the model helps to facilitate numerous comparison runs and
reduces the time and expense of the analysis.

-------
LIMITATIONS

    Although the model was implemented in a modular fashion to allow for
flexibility in development and use, it has several important limitations,

        o   Size.   The size of the model was kept manageable by
            limiting the units of analysis to 10 regions,  10
            compounds, and six end uses per compound.   More regions
            (e.g.,  for analyses of individual countries) cannot be
            accommodated easily.   The number of compounds and end uses
            per compound are also fixed.

        o   Feedback between compounds.  Each compound is analyzed
            independently.  The control of one compound (e.g., CFC-12)
            does not influence the use of other compounds.   (This
            aspect of the model is under revision so that it will be
            possible to consider, for example, how HCFC-22  and methyl
            chloroform emissions increase as they substitute for CFC-
            11, 12, 113 etc.)   These feedbacks may be analyzed by
            specifying alternative production scenarios that reflect
            the expected substitutions.

        o   Feedback between years.  Each year is analyzed
            independently, and feedbacks are handled by making
            previous years' data available through special  data
            structures.  To minimize memory requirements,  special data
            structures are established only for those data needed to
            model feedbacks.  Consequently, the addition of new
            feedbacks between years would likely require modifications
            to the data structure.

-------
                                   APPENDIX B

             SCENARIOS OF CHEMICAL PRODUCTION, POPULATION, AND GNP
INTRODUCTION

    This portion of the model generates scenarios of (1) historical and future
production of potential ozone-depleting substances, (2) population, and (3)
economic activity (i.e., gross national product, or GNP).  Ten potential
ozone-depleting compounds may be specified for analysis:  CFC-11; CFC-12;
HCFC-22; CFC-113; CFC-114; CFC-115; Carbon Tetrachloride (CC14);  Methyl
Chloroform (CH3CC13);  Halon-1211; and Halon-1301.  The production of each
compound is disaggregated by up to 10 regions of the world, and up to six
different end uses or applications.  The population and GNP scenarios may also
be disaggregated by up to 10 regions of the world.

    The remainder of this appendix is organized as follows:

        o   Production Scenarios describes the standard procedure for
            specifying scenarios of production (historical and
            future);

        o   Population and GNP Scenarios describes the specification
            of these scenarios;

        o   User-Modified Scenarios describes how the user may
            override portions of the standard procedures for
            specifying the production, population, and GNP scenarios;
            and

        o   Limitations presents the limitations of this portion of
            the model.

PRODUCTION SCENARIOS

    Production scenarios may cover any period from 1931 to 2100.   Historical
data are used through 1984,  and a variety of scenarios are specified to reflect
potential future production from 1985 onward.  To run the model,  the user may
(1) choose an existing scenario included in the Production Scenario File,-'- (2)
create a new scenario in the Production Scenario File,  (3) modify an existing
scenario using the procedure outlined below in the section on user-modified
scenarios, or (4) omit the specification of production from the model by
specifying emissions directly (see Appendix D) or ozone depletion directly (see
Appendix E).   This section describes the format of the existing scenarios
specified in the Production Scenarios File.   These scenarios are  defined in
three parts:   global production;  regional shares; and end use shares.  Each part
is described in turn.
       The six scenarios included in the Production Scenario File are described
in Chapter 18.

-------
                                      B-2
Global Production

    World production for each of the 10 potential ozone-depleting compounds is
reported in units of millions of kilograms per year.   Exhibit B-l shows an
example production scenario for all eight compounds currently analyzed.
Intermediate values between the years reported are calculated by the model
(using either linear or exponential interpolation, as specified by the user).
Production is assumed to be constant following 2050.

Regional Shares

    Regional shares are used to divide global production into the following 10
regions:

        o   U.S.;
        o   Canada (CAN);
        o   Western Europe (WEUR);
        o   Japan, Australia, and New Zealand (JANZ);
        o   USSR and East Bloc (EUSSR);
        o   Centrally Planned Asia (ASCEN);
        o   Middle East (HIDE);
        o   Africa (AFR);
        o   Latin America (LA); and
        o   South and East Asia (SEASIA).

Exhibit B-2 shows a scenario of regional shares for CFC-11.  The regional shares
are multiplied by the global production to produce regional production
estimates.  For example, to compute 2000 U.S. CFC-11 production, the model
multiplies CFC-11 global production for that year, 543.0 million kg, by the
regional share for the U.S. for that year, 0.212, to calculate 115.1 million kg
of production.  For each year, the shares add to one, indicating that all global
production is accounted for when disaggregating across regions.  Also notice
that the regional shares may change over time (e.g.,  in this scenario the U.S.
share of world CFC-11 production shrinks from 73.7 percent [0.737] in 1958 to
19.9 percent  [0.199] by 1980).

End Use Shares

    An "end use share" is the fraction of regional production that is allocated
to a particular application or end use.  Each compound may have up to six
different end uses.  End use shares may vary by the 10 regions of the world and
may change over time.  The model currently includes the following end uses:

        o   aerosol propellant;
        o   flexible urethane foam;
        o   rigid polyurethane foam;
        o   rigid nonurethane foam;
        o   refrigeration;
        o   solvent;
        o   fire extinguisher; and
        o   miscellaneous.

-------
                                      B-3
                                  EXHIBIT B-l

                      Example Global  Production Scenario
                            (millions of kilograms)
      CFC-11   CFC-12  HCFC-22  CFC-113  CC14   CH3CC13  Halon-1211  Halon-1301

1985    375.0    475.0  111.0    163.0    71.2    545.0     10.8        10.8

2000    543.0    688.0  161.0    283.0   103.0    848.0     19.7        19.7

2025  1,006.9  1,275.6  298.1    525.0   191.0  1,534.6     36.5        32.3

2050  1,867.0  2,365.0  552.0    974.0   354.2  2,777.0     67.8        53.1

2075  1,867.0  2,365.0  552.0    974.0   354.2  2,777.0     67.8        53.1

-------
                                      B-4







                                  EXHIBIT B-2




                        Regional Use Shares  for CFG-11









Year  U.S.    CAN   WEUR   JANZ   EUSSR  ASCEN  HIDE     AFR    LA    SEASIA  TOTAL




Historical Data




1931  1.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000   1.000




1958  0.737  0.023  0.165  0.075  0.000  0.000  0.000  0.000  0.000  0.000   1.000




1965  0.577  0.037  0.263  0.120  0.001  0.000  0.000  0.000  0.000  0.000   1.000




1970  0.410  0.052  0.364  0.166  0.003  0.001  0.001  0.001  0.001  0.001   1.000




1975  0.333  0.057  0.406  0.185  0.007  0.003  0.002  0.002  0.003  0.002   1.000




1980  0.199  0.067  0.470  0.215  0.019  0.007  0.004  0.004  0,009  0.006   1.000







Scenario Of Future Regional Shares




1985  0.211  0.035  0.251  0.114  0.149  0.057  0.034  0.034  0.069  0.046   1.000




2000  0.212  0.033  0.231  0.118  0.133  0.066  0.037  0.037  0.081  0.052   1.000




2025  0.212  0.030  0.214  0.116  0.113  0.069  0.046  0.042  0.097  0.061   1.000




2050  0.212  0.028  0.197  0.115  0.093  0.072  0.054  0.046  0.113  0.070   1.000




2100  0.212  0.028  0.197  0.115  0.093  0.072  0.054  0.046  0.113  0.070   1.000

-------
                                      B-5
Exhibit B-3 shows an example of end use shares for CFG-11 in the U.S.   All end
use shares for a year add to one (i.e., all production is assigned to some type
of end use).  In the module, end use shares are multiplied by estimated regional
production to estimate the amount of production going into each end use.  For
example, U.S. CFC-11 production, used to make rigid polyurethane foam,  would be
calculated for the year 2000 as follows:

       543.0             0.212            0.588      =    67.7 million kg
    (2000 CFC-11  x  (2000 CFC-11   x  (2000 CFC-11     (2000 CFC-11 U.S.
       Global        U.S. Regional      U.S. Rigid      Rigid Polyurethane
    Production)         Share)         Polyurethane         Foam Use)
                                       Foam Share)

    Using these three factors (global production,  regional shares, and end use
shares), the historical and potential future production of each of the 10
compounds can be specified in detail.  For ease of use,  the model incorporates a
range of scenarios that reflect a large range of potential future growth in
production (80 percent reduction in use to 5 percent annual growth).   The user
may choose one of these scenarios, create a new scenario, or modify an existing
scenario as described below in the section entitled User-Modified Scenarios.

POPULATION AND GNP SCENARIOS
                                                                       o
    A series of population and GNP scenarios are provided in the model.    The
user has the option of choosing one of these scenarios,  or creating his own
scenario.  Existing scenarios reflect a wide range of projected growth,  ranging
from 0.5 percent to 1.1 percent growth per year (population from 1985 to 2075)
and 1.0 percent to 3.6 percent growth per year for GNP (1985 to 2075).

    The population scenario is specified for each of 10 regions in terms of a
base year value (1975) and a list of growth indices.  Each index value is
multiplied by the base year value to produce a population projection for a given
year for a region.  Exhibit B-4 presents the scenario for the U.S. population.
To compute the U.S. population projection for 2050, the base year population for
the U.S. (213,925,000) is multiplied by the index for 2050 (1.447) which results
in a U.S. population projection of 309,488,000 in 2050.   This method allows for
the rates of growth of population to vary by region, and to change over time.

    Regional GNP scenarios are specified in a similar manner.  Exhibit B-5 shows
the scenario for U.S. GNP.  The base GNP in 1975 is in 1975 U.S. dollars and the
indices correspond to projected years.  For example, the model would calculate
projected GNP for the U.S. in 2050 by multiplying the base year GNP,  $1,519,890
million, by the corresponding index for the U.S. in the year 2050, 4.899.  This
computation results in a projected 2050 GNP for the U.S.  of $7,445,941 million
(in 1975 dollars).  GNP for other regions is also calculated in 1975 U.S.
dollars .
     o
     ^ These scenarios can be varied by the user if desired.

-------
                                B-6

                           EXHIBIT B-3

               U.S. End Use  Shares  for CFG-11
                 Flexible     Rigid         Rigid
       Aerosol     Urethane   Folyurethane  Nonurethane
Year   Propellant    Foam       Foam         Foam
HISTORICAL DATA



SCENARIO OF FUTURE
END USE SHARES


1931
1958
1970
1980
1985

2000
2050
0.
0,
0.
0.
0.

0,
0.
.000
.869
,542
.132
.055

.038
.011
0
0
0
0
0

0
6
.000 0.000 0.000 1.000
.045 0.046 0.003 0.011
.157 0.162 0.009 0.037
.284 0.371 0.020 0.064
.290 0.490 0.026 0.061

.275 0.588 0.031 0.053
.283 0.604 0.031 0.055
0.000 1.000
0.026 1.000
0.093 1.000
0.129 1.000
0.078 1.000

0.015 1.000
0.016 1.000

-------
                   B-7


               EXHIBIT B-4

        U.S.  Population  Scenario
                              Population
Year     Population Index     (thousands)

1975          1.000             213,925
1985          1.115             238,631
2000          1.253             267,955
2025          1.409             301,394
2050          1.447             309,488
2075          1.452             310,639
               EXHIBIT B-5

            U.S. GNP Scenario
                                  GNP
                              (millions of
Year        GNP Index         (1975 U.S.S)

1975          1.000            1,519,890
1985          1.348            2,048,812
2000          2.040            3,100,576
2025          3.242            4,927,483
2050          4.899            7,445,941
2075          7.830           11,900,739

-------
                                      B-8
USER-MODIFIED SCENARIOS

    The user has the option of overriding portions of the scenarios described in
the previous sections.   This capability expands flexibility by allowing modified
scenarios to be analyzed without creating complete scenario data files for each.
This option is especially useful for performing sensitivity analysis.   There are
six options for modifying the scenarios:

        o   Population -- The user may specify population projections
            over time for individual regions.  The specified
            production(s) override(s) the appropriate portion(s) of
            the standard scenario without changing the projections for
            other regions.

        o   GNP -- The user may specify GNP projections over time for
            individual regions, thereby overriding the projections
            from the standard scenario without changing the
            projections for other regions.

        o   Production by End Use -- The user may specify production
            (by end use) over time for individual regions.
            Projections for end uses and regions not explicitly
            specified are drawn from one of the standard production
            scenarios.

        o   Production as a Function of Population - - The user may
            specify the production of each compound as a function of
            population for individual regions.

        o   Production as a Function of GNP --  The user may specify
            the production of each compound in each region as a
            function of GNP.

        o   Production as a Function of GNP per Capita - - The user may
            specify the production of each compound in each region as
            a function of GNP per capita.  GNP per capita is
            calculated within the model from the GNP and population
            scenarios for each region.

    For each user-modified scenario, the user must specify in a table  the type
of projection (population,  GNP, etc.), the applicable region, the compound (if
appropriate), and the end use (if appropriate).  For example, to specify 2.0
percent annual growth in the production of CFC-11 in the U.S. from 1985 to 2050,
and zero growth thereafter, the user would fill out the table shown in Exhibit
B-6.  The values in this table would override the values for CFC-11 in the U.S.
reported in the standard scenarios.

    Another example of a user-specified projection type would be to have U.S.
CFC-11 production vary with U.S. population.  Exhibit B-7 shows an example of
how to specify this production versus population information.  The model
utilizes this information by (1) determining the level of population for that
region, and (2) identifying (by interpolation)  the associated production for
that population.  Because the information specifying the user-modified scenarios

-------
                      B-9
                  EXHIBIT B-6

User-Modified Scenario Specifying a Growth Rate
  of 2  Percent Annually for  CFG-11  in the  U.S.
       Year     Millions of kg of CFG-11

       1985                   78
       2000                  105
       2025                  172
       2050                  283
       2075                  283
                  EXHIBIT B-7

     A User-Modified Scenario:  Production
          as a Function of Population
    Population                Production
Millions of People     Millions of kg of CFC-11

       200                         50
       250                        100
       300                        200

-------
                                     B-10
is entered in the form of tables,  the user is free to adopt any arbitrary
function or "shape" for his data.   This capability enhances the ability to use
the model for sensitivity analysis.

LIMITATIONS

    The primary limitations of this scenarios module are described below.

        o   Range of Standard Scenarios.  A set of standard (i.e.,
            pre-defined) scenarios are included in the module, and the
            user may choose among them.  To perform an analysis with
            scenarios outside the range reflected by the standard
            scenarios, the user must define his own scenario or modify
            a standard scenario.  The limitation of this approach is
            that the scenarios are comprised of a significant amount
            of data, thereby making the establishment of an additional
            scenario somewhat cumbersome.  This limitation is
            alleviated to some extent by allowing the user to make
            selective modification to standard scenarios to allow
            sensitivity analysis.

        o   Sub-regional Scenarios.  Scenarios for areas other than
            the 10 regions in the model cannot be specified.

        o   Consistency.  The user is free to choose or create
            individual production, population, and GNP scenarios.  In
            fact, the future production of the different compounds was
            generally projected under various assumptions regarding
            population and GNP growth.  Therefore, the user should be
            careful to choose scenarios that are consistent with each
            other.  The model does not check for consistency among the
            scenarios chosen by the user (e.g., the user may choose a
            high production scenario and a low population scenario
            simultaneously).

-------
                                     B-ll
REFERENCES
Gibbs, Michael J.  (1986),  "Scenarios of CFG Use: 1985 to 2075," prepared for
        the U.S. Environmental Protection Agency, Washington,  B.C.

-------
                                   APPENDIX C

                       EVALUATION OF POLICY ALTERNATIVES
    This module evaluates control policy alternatives.  If a policy specified by
the model user requires reductions in the production of potential ozone -
depleting substances, the model simulates these reductions by choosing from a
list of possible technical control options.  Each technical control is defined
by its cost and the percentage reduction that it can achieve.

    Because this risk assessment is focused on evaluating risks in the absence
of regulation, this module is not employed in the analysis at this time.
Therefore, this module is not described here.  For discussion of control policy
alternatives, the interested reader is referred to Regulatory Impact Analysis:
Protection of Stratospheric Ozone. U.S. Environmental Protection Agency.

-------
                                  APPENDIX D

               EMISSIONS OF POTENTIAL OZONE-DEPLETING COMPOUNDS
INTRODUCTION

    Emissions are estimated on a global basis for each compound analyzed.
Two approaches are available for modeling emissions:

        o   Internal estimates -- Emissions are modeled based on the
            production scenarios described in Appendix B and modified
            through the implementation of policies described in
            Appendix C.  A release algorithm is used to describe how
            fast the compounds are released into the atmosphere after
            production.  This algorithm is based on "release tables"
            that are specific for each compound and end use.  Each
            release table shows the annual fraction of the chemical
            that is released over time.

        o   Exogenous specification -- Emission scenarios may be
            defined directly by the user.

INTERNAL ESTIMATES OF EMISSIONS

    The model calculates emissions from release tables and estimates of
production developed by the Production Scenario Module (Appendix B) and modified
by the Policy Alternatives Module (Appendix C).  This section first defines a
release table, then describes how releases are estimated from one year's
production, and, finally, presents how emissions are calculated from production
over a series of years.

Release Tables

    Compounds are released into the atmosphere at different rates following
production, depending on their uses.  Some uses (such as aerosol propellants)
release compounds into the atmosphere very soon after production.  However, some
uses (such as refrigeration and rigid foam production) store compounds for many
years, releasing them gradually over time.  This storage is called a "bank."

    A release table represents the fraction of production emitted over time for
a particular compound and end use.  Release tables have been defined for each of
the end uses in the model based on Quinn (1986), and the user may specify
alternative release tables as desired.  Exhibit D-l presents the release tables
for CFC-11 in its end uses.  A total of 20 years is shown, and the fraction of
the production from year 1 released into the atmosphere in the following years
is listed.  For example, during the year of production of refrigeration uses,
19.0 percent of the compound is released.  By the end of the second year, a
cotal of 27.1 percent of the production from year 1 is released.  Releases
continue through the fourth year, after which all of the production from year 1
is released.  Annual release rates may be calculated by subtracting each year's
cumulative fraction from the previous year's cumulative fraction.  For example,
8.1 percent (0.271 minus 0.190) is released in year 2, 7.3 percent (0.344 minus
0.271) is released in year 3.

-------
                                                           D-2

                                                      EXHIBIT  D-l

                                            Release Tables  for CFG-11
                                             Rigid               Rigid
     Aerosol Propellant     Flexible Foam      Polyurethane Foam   Nonurethane Foam       Refrigeration       Miscellaneous
Cumulative Annual Cumulative
Fraction Fraction Fraction
Year Released Released Released
1 1.000 1.000 1.000
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Annual Cumulative
Fraction Fraction
Released Released
1.000 0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
.145
.190
.235
.280
.325
.370
.415
.460
.505
.550
.595
.640
.685
.730
.775
.820
865
.910
955
.000
Annual Cumulative Annual Cumulative Annual Cumulative Annual
Fraction Fraction Fraction Fraction Fraction Fraction Fraction
Released Released Released Released Released Released Released
0.
0
0.
0
0.
0
0
0,
0
0.
o
o
0
0
0
0
0
0
0
0
,145 0.600 0.600 0.190 0.190 1.000 1.000
,045 1.000 0.400 0.271 0.081
,045 0.344 0.073
.045 1.000 0.656
,045
,045
.045
,045
.045
,045
.045
.045
.045
.045
.045
.045
.045
.045
.045
.045
Source:  Quinn,  T.,  et al. (1980), "Projected Use, Emissions, and Bank  of Potential Ozone Depleting Substances," The
        RAND Corporation, prepared for the U.S. Environmental Protection Agency,  Washington,  D.C.

-------
                                      D-3
    The release tables vary in length (some end uses store their compounds for
many years).  Also, the release table need not reflect complete emissions.  If
the cumulative release fraction in the table reaches 1.0, then all of the
compound that is produced is eventually released into the atmosphere.  If some
of the compound is expected to be retained forever within a product, or
destroyed,  the release table would have as its highest value a number less than
1.0.  The difference between the reported value and 1.0 would be the portion of
the compound that is never emitted.

Estimating Emissions From One Year's Production

    Exhibit D-2 shows the emissions from the hypothetical production of 100
million kg in 1985.  The cumulative fraction of production released over time is
presented in column 2.  The annual fraction of production released is shown in
column 3.  The model calculates annual emissions by multiplying the annual
fraction by the amount of production in the initial year (100 million kg in this
example).  The projected emissions per year are shown in column 4.  In this
example the total emissions sum to the initial production amount, indicating
that all of the production is eventually emitted into the atmosphere.

Estimating Emissions From Production Over a Series of Years

    Exhibit D-3 presents how the model calculates emissions from production over
a series of years.  Columns 1 and 2 show years and production in that year,
respectively.  In this example, production grows by 10 million kg per year, and,
for simplicity, emissions from production before 1985 are omitted (emissions
from past production are not omitted in the model).   Columns 3 through 8 show
emissions over time due to the production shown in column 2, each row showing
how releases of its production are spread over the course of years.

    The first row shows the emissions from 1985 production.  The second row
shows emissions from the 1986 production, and so on.  The total emissions for a
year are calculated by summing emissions values in the columns to compute a
total.  For example, emissions in 1985 are 19.0, emissions in 1986 are 29.0,
emissions in 1987 are 39.0, and so on.  By 1990, a total of 441.1 million kg has
been emitted, and a total of 750 million kg has been produced.  Therefore, 750 -
441.1, or 308.9 million kg, remains to be emitted after 1990.

    This method of estimating emissions reflects the delays between the
production and release of many potential ozone-depleting substances.  The
implications of the method are that releases in a given year are caused not only
by production in that year, but also by production and use in previous years.

EXOGENOUS SPECIFICATION OF EMISSIONS

    The user has the option of directly specifying the level of emissions over
time.  This is accomplished by filling out an emissions table; a sample table is
presented in Exhibit D-4.  This table shows emissions estimates for eight
chemical compounds over a time period of 65 years.  The module can interpolate
intermediate emissions values by either linear or exponential interpolation.
These specified emissions override the estimates of emissions that the model
would have generated from production data as discussed in the previous section.
This ability to specify emissions directly to the model allows the flexibility

-------
                    D-4
                EXHIBIT D-2

       Emissions from a Hypothetical
100 Million Kilograms of Production in 1985
I 2
Cumulative
Year Fraction Released
1985 0.190
1986 0.271
1987 0.344
1988 1.000
TOTAL
3
Annual
Fraction Released
0
0
0
0
1
.190
.081
.073
.656
.000
4
Annual Emissions
(millions of kg)
19.0
8.1
7.3
65.6
100.0

-------
                               D-5
                           EXHIBIT D-3

         Emissions  from Production Over a Series of Years
                     (millions of kilograms)
Year
1985
1986
1987
1988
1989
1990
Production
100
110
120
130
140
150
1985 1986 1987
19.0 8.1 7.3
20.9 8.9
22.8
--
. .
	 	 	
1988
65.
8.
9.
24.

_ _

6
0
7
7


1989
--
72.
8.
10.
26.
_ _


2
8
5
6

1990
--
--
78.
9.
11.
28.



7
5
3
5
TOTAL EMISSIONS
  PER YEAR            19.0    29.0    39.0    108.0   118.1   128.0

-------
                                      D-6
                                  EXHIBIT D-4

                Sample Table of Exogenously Specified Emissions
                            (millions of kilograms)
Year  CFC-11  CFC-12 HCFC-22  CFC-113  CC14  CH3CC13  Halon-1211  Halon-1301

1985     298    438      81     138     71     500         3           3

2000     462    668     151     240    103     721         9           9

2025     897  1,237     279     446    191   1,303        25          22

2050   1,650  2,281     516     828    355   2,356        48          38

-------
                                      D-7
to evaluate the ozone depletion and other effects that may be associated with
given emission levels,  without having to specify a policy that would result in
such emissions.

LIMITATIONS

    The primary limitation of this module is that the emission tables are not
designed to change over time.  Therefore, changes in technology that influence
emissions rates of existing end use products cannot be modeled.

-------
                                      D-8
REFERENCES
Quinn, T.,  et al.  (1986), "Projected Use, Emissions, and Bank of Potential
    Ozone-Depleting Substances," The RAND Corporation, prepared for the U.S.
    Environmental Protection Agency, Washington, D.C.

-------
                                  APPENDIX E

                          ATMOSPHERIC  SCIENCE MODULE
INTRODUCTION

    This portion of the model computes the level of global average ozone
depletion as a function of the emissions of the 10 potential ozone-depleting
substances.  The emissions are computed in the previous module (presented in
Appendix D).  The user has three options for estimating ozone depletion:

        o   Use a relationship between emissions and ozone depletion developed
            to reflect the results of a one-dimensional (1-D) atmospheric model;

        o   Specify his own relationship; and

        o   Specify the level of ozone depletion directly.

Each of these options is described in turn.  Then, the method used to reflect
uncertainty is presented.  Finally, the limitations of the atmospheric science
module are presented.

PARAMETERIZED RELATIONSHIP

    The module incorporates a series of equations that describe the change in
total column ozone as a function of:

        o   emissions of potential ozone-depleting substances; and

        o   abundances of three trace gases (C02 in ppm, N20 in ppb,
            and CH4 in ppm).

This relationship was developed by Connell (1986), and it computes global total
column ozone change (in percent) relative to 1985.

    To use this relationship, the user must specify scenarios of trace gas
abundances to use in conjunction with the emissions of potential ozone-depleting
substances.  A range of scenarios has been developed from which the user may
choose, or the user may specify his own scenarios.  The reference case set of
assumptions used in the development of the relationship is shown in Exhibit E-l.
Exhibit E-2 compares the results of the LLNL 1-D model to the relationship
developed by Connell for the base case of emissions of ozone-depleting
substances.

    The following six steps are performed to compute global ozone depletion in
year t using the parameterization:

    1.  Compute the change in the chlorine concentration in year t relative to
        1985 as:

-------
                         E-2
                     EXHIBIT E-l

        Trace Gas Assumptions Used To Develop
         the Ozone-Depletion Relationship-'
         Carbon Dioxide     Nitrous Oxide     Methane
Year
1985
1995
2005
2015
2025
2035
2045
2055
2065
2075
(ppm)
350.
361.
377.
399.
422.
456.
490.
531.
578.
625.
2
3
2
6
0
4
8
4
2
0
(ppb) (ppm)
303
309
315
321
328
334
341
348
355
362
.1
.2
.5
.8
.3
.9
.7
.6
.6
.8
1
1
2
2
2
2
2
2
3
3
.756
.926
.096
.266
.436
.606
.776
.946
.116
.286
   a/
      This base case set of assumptions was one of several
used to develop the relationship between emissions of
ozone-depleting substances and ozone depletion.

Source:  Connell,  Peter S. (1986),  "A Parameterized Numerical
         Fit to Total Column Ozone  Changes Calculated
         by the LLNL 1-D Model of the Troposphere and
         Stratosphere," Lawrence Livermore National
         Laboratory, Livermore, California.

-------
                        E-3
                    EXHIBIT E-2
    Comparison of Total Column Ozone Depletion
        Results from the 1-D Model and the
           Parameterized Numerical Fit3'
          Ozone Depletion (percent)
Year
1985
1995
2005
2015
2025
2035
2045
2055
2065
2075
Numerical Fit
0
-0
-0
-1
-2
-4
-6
-9
-14
-21
.0
.18
.74
.64
.86
.48
.63
.60
.0
.6
1-D Model
0
-0
-0
-1
-2
-4
-6
-9
-12
-19
.0
.18
.61
.42
.65
.30
.41
.10
.9
.9
Difference
0
0
0
0
0
0
0
0
1
1
.00
.00
.13
.22
.21
.18
.22
.50
.10
.70
   a/
      This base case was one of many cases used to develop
the relationship between emissions of ozone-depleting
substances and ozone depletion.  This result does not
necessarily reflect the base case estimates from this
model.

Source:  Connell, Peter S. (1986), "A Parameterized Numerical
         Fit to Total Column Ozone Changes Calculated
         by the LLNL 1-D Model of the Troposphere and
         Stratosphere," Lawrence Livermore National
         Laboratory, Livermore, California.

-------
                                  E-4
        dCl(t) - dCl(o) + S [CF(j Remissions(i,j)
                            * exp(-(t-i)/lifetime(j))
                            * (l-exp(-(-t-i)/mixing time))]

where:  dCl(t) equals the change in chlorine concentrations in year t;

    dCl(o) equals the change in chlorine concentration in year t due
           to emissions prior to 1985;

    CF equals a conversion factor for each compound j;

    emissions are the annual emissions for compound j for years 1985
           to year t;

    lifetime equals the average atmospheric lifetime for each compound j;
           and

    mixing time equals the time it takes for source emissions to become
           well mixed.

2.  Compute the changes in C02,  N20 and CH4 for year t as follows:

        dC02 = C02(t)/C02(1985);
        dN20 = N20(t)/N20(1985); and
        dCH4 = CH4(t)/CH4(1985),

where the values for C02,  N20 and CH4 over time are model inputs as
discussed above.

3.  Compute ozone depletion due to CFCs as:

        d03(t) = 14.58[asinh(.332(21.05(H-(dN20-l)/(dCH4)2)
                            - dCl(t)*exp(-.15(dCH4-l))))
                            - asinh(.332*21.05(l+(dN20-l)/(dCH4)2))]

where d03(t) equals the ozone depletion in year t (in percent).

4.  Compute ozone depletion (in percent) from C02,  N20, and CH4 as follows:

        d03(t) - 3.61n(dC02);
        d03(t) = 3.75(dCH4-l); and
        d03(t) = -7.0(dN20-l).

5.  Compute ozone depletion (in percent) from Halons as follows:

        d03(t) = -0.0302[concentration of Halon-1301]
                 -0.0618[concentration of Halon 1211]

6.  Compute total ozone depletion (in percent) as the sum of the individual
    estimates computed in Steps 3, 4, and 5.

-------
                                      E-5
These equations are used to represent the potential for ozone depletion as a
function of emissions and concentrations of the various substances analyzed.
The details of this method are described in Connell (1986).

USER-SPECIFIED RELATIONSHIP

    The user has the option of specifying his own relationship between emissions
and ozone depletion.  For each substance, the user specifies a "conversion
factor" and a "lifetime."  These values are used in the following equation to
compute an atmospheric abundance for each compound over time:

                                       T   ,  .       , .  -(T-t)/lifetime1
    Abundance (T) = Conversion Factor  S   [emissions (t)e                ].
                                     t=t
The abundances are summed across the compounds to create a total abundance.  The
amount of ozone depletion associated with each level of abundance is then
specified by the user in a table.

    This formulation of the user-specified ozone-depletion relationship is
similar to the parameterized numerical fit developed by Connell (1986).  It can
produce abundances of the different compounds that exhibit behavior similar to
the results produced by 1-D models.  The value of this option is that the user
may identify, in tabular form, his best understanding of the manner in which
ozone depletion will vary with the abundances of the compounds in the
atmosphere.

USER-SPECIFIED OZONE DEPLETION

    The user has the option of bypassing the computation of ozone depletion
based on emissions of potential ozone-depleting substances.  Instead, the user
may provide estimates of global ozone depletion directly to the model.  This
option allows the user to evaluate the impacts of such a level of depletion
(possibly computed via other means) without identifying a level of emissions or
a control policy that would produce such depletion.

    To specify levels of ozone depletion, the user fills out the table shown in
Exhibit E-3.  For each year, a level of total column ozone depletion  (or
increase) is identified.  The model interpolates between the values supplied in
the table.

UNCERTAINTY

    The user has the option of recognizing the uncertainty inherent in current
models of ozone depletion by specifying a range within which the "true" level of
ozone depletion is expected to fall.  This range is specified using
multiplicative scaling factors, as shown in Exhibit E-4.  The scaling factors
are multiplied by the estimate of ozone depletion produced within the

-------
                         E-6


                     EXHIBIT E-3

HYPOTHETICAL TABLE OF USER-SPECIFIED OZONE DEPLETION*
                               Total Column
                          Global Ozone Depletion
Year
1985
1990
2000
2025
2050
2075
2100
(percent")
0.00
0.01
0.10
1.00
5.00
10.00
20.00
    * Values are illustrative only.

-------
                     E-7


                    EXHIBIT E-4

      EXAMPLE OZONE-DEPLETION SCALING FACTORS*
 Ozone Depletion   Scaling Factors
    (percent)      Low	High

    <15.0          0.4        2.0

    >20.0          0.5        1.8
*These factors represent the 10th and 90th fractile
estimates from a lognormal distribution of
uncertainty developed from values reported in:
Stolarski, R.S., and A.R. Douglass (1986),
Sensitivity of an Atmospheric Photochemistry Model to
Chlorine Perturbation Including Consideration of
Uncertainty Propagation. Draft Report to the U.S.
Environmental Protection Agency, Washington, D.C.

-------
                                      E-8


model.   These multiplications produce a range of ozone depletion values for
each year; the end points of the range (as well as the middle value estimated
within the model) are passed to the effects module for evaluation.

    As shown in Exhibit E-4, the scaling factors may change at different levels
of ozone depletion.   The values shown in the exhibit are based on a Monte Carlo
uncertainty analysis performed by Stolarski and Douglass (1985).   Each Monte
Carlo iteration draws from a distribution of estimated reaction rates and cross
sections to generate predictions of atmospheric abundances.  These predicted
abundances can be compared to actual measurements in the atmosphere; those
iterations that do not compare favorably with actual measurements can be
screened out.   The values reported in Exhibit E-4 are based on those iterations
that compared favorably with actual measurements and were not screened out.

    Because the ozone-depletion models are not considered reliable at high
levels of ozone depletion, the user may also specify a "cap" on ozone depletion.
If the computed level of ozone depletion (after scaling) exceeds the user
specified cap, then the ozone depletion used to evaluate effects (in the next
module) is reduced to the level of the cap.

LIMITATIONS

    The Atmospheric Science Module is a simplified representation of a complex
set of atmospheric reactions.  In using the module,  the user should become aware
of the range of emissions and trace gas values for which the parameterized
relationship is considered valid.  Outside this range, the equations do not
necessarily reflect current estimates from 1-D models of the atmosphere.  The
range of production scenarios discussed in Appendix B to this chapter and
Chapter 18 are generally well within the range for which these equations are
valid.  In those few instances when they are not within the range, the allowable
values are capped to avoid extrapolating beyond the bounds of the analyses from
which the equations were developed.
     o
       The scaling factors are not used when ozone depletion is specified
directly by the user.

-------
                                      E-9
REFERENCES

Connell, Peter S. (1986), "A Parameterized Numerical Fit to Total Column Ozone
    Changes Calculated by the LLNL 1-D Model of the Troposphere and
    Stratosphere," Lawrence Livermore National Laboratory, Livermore,
    California.

Stolarski, R.S., and A.R. Douglass (1985), "Sensitivity of an Atmospheric
    Photochemistry Model to Chlorine Perturbation Including Consideration of
    Uncertainty Propagation," Draft Report to the U.S. Environmental Protection
    Agency, Washington, D.C.

-------
                                  APPENDIX F

              HEALTH AND ENVIRONMENTAL IMPACTS OF OZONE DEPLETION
INTRODUCTION

    This appendix presents the methods used to model the health and
environmental effects of ozone depletion in the CFC Policy Model.   This module
incorporates analyses performed by a variety of researchers,  including analyses
of:

        o   the relationship between global ozone depletion and
            latitudinal-dependent ozone depletion (Isaksen 1986);

        o   the relationship between ozone depletion and ultraviolet
            radiation (UV);^

        o   the relationship between UV and the incidence (and
            mortality) of nonmelanoma skin cancers,

        o   the relationship between UV and the incidence (and
            mortality) of melanoma skin cancers,

        o   the relationship between UV and the prevalence (and
            incidence) of senile cataracts,  and

        o   the relationship between ozone depletion and the costs of
            modifying polymers to prevent degradation of polymer
            appearance and/or performance due to increased UV.

    These analyses pertain primarily to the U.S., consequently, this module
focuses on the risks in the U.S. associated with ozone depletion.   The module is
designed to allow risk estimates to be performed for other portions of the
world if:  (1) the model user supplies data needed to quantify effects in these
other regions; or (2) the model user assumes that the data for the U.S apply to
other portions of the world.  The descriptions below relate exclusively to the
analysis of impacts in the U.S.
     1 See Serafino and Frederick (1986) for a description of the model used to
estimate the relationship between ozone depletion and UV flux.

     2 See Chapter 7 for a discussion of the relationship between nonmelanoma
skin cancer and UV.

     3 See Chapter 8 for a discussion of the relationship between melanoma skin
cancer and UV.

     ^ See Chapter 10 for a discussion of the relationship between senile
cataracts and UV.

     -* See Chapter 13 for a discussion of the relationship between polymer
degradation and UV.

-------
                                      F-2
    This appendix is divided into the following sections:

        o   Changes in UV Flux presents the method used to estimate
            changes in the flux of UV for given levels of ozone
            depletion;

        o   Skin Cancers and Cataracts presents the methods used to
            model melanoma and nonmelanoma skin cancers and cataracts;

        o   Materials presents the methods used to model the costs
            associated with the degradation of polymers; and

        o   Limitations summarizes the major limitations of the methods
            used, including the risks believed to be associated with
            increases in ambient levels of UV that are not modeled at
            this time in the CFC Policy Model.

CHANGES IN UV FLDX

    This portion of the module evaluates the expected changes in UV flux
associated with changes in total column ozone.  The estimates of changes in UV
flux vary by latitude because the change in UV flux (measured in percent) for a
given change in total column ozone (also measured in percent) varies as a
function of the initial total abundance of ozone (i.e., the level of ozone
abundance in the absence of ozone depletion).

    The relationship between ozone change and UV flux will also vary for
different measures of UV energy.  Three measures of UV, also referred to as
action spectra, are currently incorporated into the model:  Robertson-Berger
Meter (RB-Meter); DNA Action Spectrum; and Erythema Action Spectrum.  Each
measure has a unique weighting of the wavelengths in the UV spectrum.  Because
ozone selectively screens out radiation at given wavelengths, the different
weighting schemes will produce different estimates of increased radiation levels
when total column ozone abundance changes.

    The UV Model developed by Serafino and Frederick (1986) was used to evaluate
the expected changes in UV flux for given levels of total column ozone change.
Pitcher (1986) performed a validation of the UV model, comparing monthly model
estimates to monthly field instrumentation readings for three cities:  El Paso,
San Francisco, and Minneapolis.  Although the model was found to overestimate UV
flux in the months of September through March, and underestimate UV flux in May,
June, and July, the annual estimates are generally within 10 percent of the
RB-meter readings in the field  (differences are probably attributable to errors
in input data about clouds from real world situations, in which that data base
is weak).

    Changes in UV flux  (as measured by the three action spectra) as a function
of changes in ozone abundance were evaluated using the UV model for 24 cities
across the U.S.  These cities are divided into three regions of the U.S.:
north, middle, and south (see Exhibit F-l).  These three regions are used to
evaluate potential changes in the incidence and mortality due to skin cancers.

-------
                                   F-3
                               EXHIBIT F-l

                Cities  Used to  Evaluate  Changes  in UV Flux
                    for the Three Regions of the U.S.
REGION 1:  NORTH             REGION 2:  MIDDLE           REGION 3:  SOUTH
  New York                   Chicago                     Los Angeles
  Detroit                    Philadelphia                San Diego
  Milwaukee                  Baltimore                   Houston
  Boston                     San Francisco               Dallas/Fort Worth
  Seattle                    Washington                  Phoenix
  Minneapolis                Denver                      New Orleans
  Portland                   Salt Lake City              Miami
  Buffalo                    Kansas City                 Atlanta

-------
                                      F-4
The states included in each region are shown in Exhibit F-2.  The population
weighted average latitudes for the three regions are also shown.

    Exhibit F-3 shows the percent change in UV (measured with the DNA action
spectrum) in the three regions of the U.S. associated with a range of percent
changes in total column ozone (from a 10 percent increase in ozone (listed as a
-10 percent ozone depletion) to a 30 percent ozone depletion).  As expected, the
change in UV for a given level of ozone change is not constant across the three
regions of the U.S., although they are very similar.  Statistical analysis
indicates that the estimates for the three regions are statistically different
from each other.

    Similar tables have been developed for the RB-Meter and the Erythema action
spectra.   The values for the Erythema Action Spectrum are approximately 80 to 85
percent of the DNA Action Spectrum values; the RB-Meter values are 35 to 45
percent.   The user may define additional action spectra and prepare additional
tables that relate changes in UV flux to changes in ozone depletion if he wishes
to evaluate effects as a function of different action spectra.

SKIN CANCERS AND CATARACTS

    The objective of this portion of the module is to evaluate the potential
changes in risks of melanoma and nonmelanoma skin cancers and cataracts
associated with changes in total column ozone abundance over time.  The purpose
of the evaluation is to provide information for assessing the implications of
alternative emission rates of potential ozone-modifying substances.  The inputs
to this portion of the module include estimates of changes in UV flux associated
with changes in total column ozone abundance.  The output of this portion of the
module includes estimates of the increased number of cases and deaths associated
with skin cancers over time and an estimate of the increased number of cases of
cataracts.

    This analysis is organized into four parts as follows:

        1.   characterize the population at risk;

        2.   estimate the baseline incidence of the conditions being
            evaluated;

        3.   apply dose-response relationships to estimate the change in
            the risk of the conditions as a function of changes in ozone
            abundance over time;  and

        4.   summarize the results.

    Data have been developed primarily for the three regions of the U.S. for
purposes of evaluating these skin cancer and cataract risks (see above for the
definition of the three regions in the U.S.).  The risks of these cancers in
other portions of the world can be evaluated by the model user if:  (1) the user
supplies data to characterize the populations at risk in the other regions,  the
baseline incidence in the other regions, and the dose-response relationships

-------
                                   F-5
                               EXHIBIT F-2

             States Included  in the Three  Regions of  the U.S.
REGION 1:   NORTH
REGION 2:  MIDDLE
REGION 3:  SOUTH
  Alaska
  Connecticut
  Idaho
  Maine
  Massachusetts
  Michigan
  Minnesota
  Montana
  New Hampshire
  New York
  North Dakota
  Oregon
  Rhode Island
  South Dakota
  Vermont
  Washington
  Wisconsin
  Latitude = 43.3 N
California (N) a/
Colorado
Delaware
District of Columbia
Illinois
Indiana
Iowa
Kansas
Kentucky
Maryland
Missouri
Nebraska
Nevada
New Jersey
North Carolina
Ohio
Oklahoma
Pennsylvania
Tennessee
Utah
Virginia
West Virginia
Wyoming

Latitude = 39.1 N
Alabama
Arizona
Arkansas
California (S) a/
Florida
Georgia
Hawaii
Louisiana
Mississippi
New Mexico
South Carolina
Texas
Latitude = 31.8 N
a/  California is divided in half,  one half included in the Middle Region,
    and one half included in the South Region.

Source:  Latitude estimates based on population centroids for each state
         from the 1980 U.S. census,  Master Area Reference File #2, Geography
         Section, U.S. Bureau of Census,  Department of Commerce.

-------
                      F-6
                  EXHIBIT F-3

Percent Change in UV as a Function of Change in
     Ozone Abundance  for Three U.S. Regions
             (DNA Action Spectrum)
CHANGE IN UV (%)
OZONE DEPLETION (%) North
-10
-5
-2
0
2
5
10
20
30
-17.3
-9.3
-3.8
0.0
4.2
10.8
22.9
53.8
96.0
Middle
-17
-9
-3
0
4
10
22
53
94
.2
.1
.8
.0
.3
.6
.8
.2
.8
South
-16
-8
-3
0
4
10
22
51
90
.7
.9
.8
.0
.2
.5
.2
.0
.4
Source:  Based on analyses using the UV Model developed
         by Serafino and Frederick (1986).

-------
                                      F-7
that apply to the other regions;  or (2) the user assumes that the information
provided for the U.S.  applies to the other regions.   Each of the four steps
discussed below is presented for the U.S.

Step 1 -- Characterize the Population at Risk

    The incidence of skin cancers and cataracts has been found to vary by race,
age, sex, and the ambient level of UV in the area of residence of the
individual.^  The population in the U.S. is characterized by age, sex, and race
using data from the U.S.  Bureau of the Census.'  Each of the three regions in
the U.S. are characterized separately.

    For each region, the current and expected age distribution of the population
is defined in a series of tables as shown in Exhibit F-4.  These age
distributions were obtained by taking age distributions for the United States
and assuming the age distribution for each region (in this case the north
region) were identical.

    The age distribution of the U.S. population is expected to change over time
due to changing survival and birth rates.   These changes in the distribution
through 2080 are included in the characterization of the population, as shown in
Exhibit F-4.  Changes beyond 2080 have not been estimated, and consequently, the
age distribution is assumed to remain fixed beyond 2080.  If the U.S.
population becomes older (younger) on average after 2080, this fixed population
distribution underestimates (overestimates) the population of older individuals,
and may consequently underestimate (overestimate) risks of ozone depletion.  The
user may relax this assumption by inserting alternative data for years beyond
2080 as appropriate.

    The regional distribution of the U.S.  population is assumed to remain
unchanged over time.  Recent trends do not indicate a definite shift from one
region to another.  If the U.S. population distribution were to shift toward the
south (north),  the risks of ozone depletion would be underestimated
(overestimated) by this assumption.  The user may alter this assumption by
inserting alternative data on regional distribution patterns.

    The portions of the population in each region that are male and female,
black and white are also defined.  These portions could be changed over time
within the model to reflect changing demographic patterns; however, in this
analysis these population variables were assumed to be equal across regions and
over time.

    From these data the fraction of the U.S. population is characterized by
age, race,  and sex for each of the three regions of the country.  The total
U.S. population over time is provided as an input to the Scenarios Module (see
       See Chapter 7 for a discussion of nonmelanoma skin cancer, Chapter 8 for
a discussion of melanoma skin cancer, and Chapter 10 for a discussion of
cataracts.
       The Census Data were obtained directly from the U.S. Bureau of the
Census.

-------
                                      F-8
                                  EXHIBIT F-4

                    Age Distribution of the U.S. Population
                         Over Time in the North Region
                  (fraction of total population  in the region)
AGE GROUP
YEAR
1985
2000
2025
2050
2080
0-14
.2074
.1990
.1752
.1692
.1665
15-24
.1619
.1303
.1168
.1151
.1146
25-34
.1739
.1329
.1240
.1214
.1186
35-44
.1360
.1633
.1310
.1235
.1219
45-54
.0961
.1422
.1162
.1190
.1196
55-64
.0973
.0925
.1282
.1229
.1174
65-74
.0748
.0700
.1166
.1005
.1040
75-84
.0404
.0498
.0651
.0723
.0757
85+
.0122
.0200
.0269
.0561
.0617
FRACTION OF
TOTAL U.S. POP.*
.2486
.2486
.2486
.2486
.2486
* The fraction of the total U.S.  population represents the portion of the U.S.
population that is estimated to reside in this region (in this case the North
region).  This fraction does not change over time,  thereby reflecting stable
regional distribution patterns.

Source:  Age distribution from:  U. S. Bureau of Census, Current Population
         Reports, series P-25, No. 952, Projections of the Population of the
         United States, by Age. Sex and Race: 1983 to 2080. U.S. Government
         Printing Office, Washington, D.C., 1984.

-------
Appendix B),  and is used as the basis of estimating the number of people in
each age, race, and sex grouping.

Step 2 -- Baseline Incidence

    To evaluate the potential future risk of skin cancers in the U.S., the
baseline risk (i.e., the risk in the absence of changes in the level of total
column ozone) is required.  Incidence of melanoma and nonmelanoma skin cancers
has been found to vary by latitude.  Therefore, baseline incidence values have
been developed for each of the three U.S. regions separately.  The age-sex
specific baseline rates used for melanoma and nonmelanoma are described
separately.  The baseline data for cataracts follow.

    Nonmelanoma Skin Cancer

    There are two major types of nonmelanoma skin cancers:  basal cell and
squamous cell (Scotto, Fears, and Fraumeni 1981; and see Chapter 7).  Because
the two types may respond differently to changes in UV, their baseline
incidences are estimated separately (Scotto, Fears, and Fraumeni 1981).   Data
describing the age-sex specific incidence for whites are available for 10
locations from Scotto, Fears, and Fraumeni (1981).  The incidence rates by age
a-nd sex were estimated for each of the three U.S. regions by taking the
population-weighted averages of the rates for the cities reported in Scotto,
Fears, and Fraumeni (1981) that are in each of the regions.   The locations in
each region are:

         North:  Seattle, Minneapolis-St. Paul, Detroit;
         Middle:  Utah, San Francisco, Iowa; and
         South:  Atlanta, New Orleans, New Mexico, Dallas-Fort Worth.

Exhibit F-5 displays an example of the age-specific incidence rates for basal
and squamous cell skin cancers in the North region of the U.S.  These data
apply only to whites.

    The baseline incidence rates for blacks and members of other races are not
well characterized.  The rates are believed to be approximately one order of
magnitude or more below the rates for whites, and are consequently not expected
to play a major role in evaluating risks of skin cancers in the U.S. (they may
play a more important role in the risks in other countries).   For purposes of
this analysis, risks for blacks and members of other races are not evaluated.

    The mortality associated with each of these cancers is believed to be quite
low.  Approximately 1.0 percent of all cases of nonmelanoma skin cancer result
in death (Scotto, Fears, and Fraumeni 1981).  The rates for basal and squamous
cell cancer differ, and have been estimated from data in Scotto, Fears and
Fraumeni to be:  basal cell -- 0.3 percent;  and squamous cell -- 3.75 percent.
These percentages are multiplied by the baseline incidence rates to estimate
the baseline rates of death associated with each cancer type over time.

-------
                                                  F-10
                                              EXHIBIT F-5

                         Baseline  Incidence for Nonmelanoma Skin Cancers
                       (North Region of  the U.S. --  White  Population  Only)
                                          (Rate per  100,000)
                             15    15-24
                                                           AGE GROUP
                                            25-34
                                                    35-44
                                                              45-54
                                                                      55-64     65-74
                                                                                        75-84
                                                                                                 85+
NONMELANOMA SKIN CANCER


  Basal Cell Skin Cancer
Male
Female
Squamous Cell Skin Cancer
Male
Female
0.1 2.9 22.1 91.1 259.2 465.8 761.0 1162.8 1311.3
0.5 5.6 22.2 91.0 201.8 287.4 465.9 638.2 754.1

0.2 0.3 1.6 7.4 32.6 87.4 147.4 349.7 431.8
0.0 0.1 1.4 4.1 10.5 27.5 54.8 112.5 167.7
Source:  Derived from Scotto, Fears, and Fraumeni (1981).

-------
                                      F-ll
    The model user may also specify that the baseline incidence and mortality
ratio may change over time (increase or decrease).  The purpose of this
specification is to allow the user to include the implications of potential
changes in the baseline incidence of these skin cancers that are not related to
changes in UV flux.

    Melanoma Skin Cancer

    There are four principal classes of melanoma skin cancers:  Hutchinson's
melanotic freckle, superficial spreading melanoma, nodular melanoma, and
unclassifiable melanoma (Elder et al.  1980).  For purposes of modeling the
baseline incidence of these cancers,  two groups are defined that are based on
the location of the cancer on the body:

         o   Group 1: Face, Head, Neck, and Upper extremities; and
         o   Group 2: Trunk and Lower Extremities.

These groups have been defined by Scotto and Fears (in press), and incidence
data have been collected for seven U.S. cities that divide the occurrence of
melanoma skin cancers into these groups.  The division of these cancers into
these groups appears to be warranted because Scotto and Fears have identified
different correlations with UV for each of the two groups.

    The incidence data for these groups of melanoma skin cancer were developed
for the three regions of the U.S. by taking population-weighted averages of the
incidence data reported for the following seven locations in the three
regions:

         o   North:  Seattle, Detroit;
         o   Middle:  Utah, San Francisco, Iowa; and
         o   South:  Atlanta, New Mexico.

The incidence data were taken from the National Cancer Institute, Cancer
Incidence and Mortality in the United States:  SEER Report (1984), and the data
apply to whites only.  Because the SEER data do not divide the incidence by the
two groups of melanoma skin cancer defined above, the fractions of age-specific
incidence associated with each type of cancer were taken from Scotto and Fears
(in press) .

    Exhibit F-6 displays sample baseline incidence data for melanoma skin
cancer by sex and age for whites in the North U.S. region.  As shown in the
exhibit,  the incidence rates for the two types of melanoma were estimated
separately.

    Mortality data were taken from the SEER Report (1984) for whites only.
Because the SEER data do not divide the mortality rates by the two groups of
     8 Although the National Cancer Institute SEER Report (1984) reports data
for ten locations, incidence data for only seven locations were used.  These
seven locations correspond to the locations for which Scotto and Fears (in
press) reported body site percentages (face, head, neck, and upper extremities
vs. trunk and lower extremities) for melanoma incidence.  The three locations
that were not included in this analysis are:  Connecticut, Hawaii, and Puerto Rico.

-------
                                     F-12
                                  EXHIBIT F-6

                 Baseline  Incidence for Melanoma  Skin Cancers
              (North Region of the U.S. -- White Population Only)
                               (Rate per 100,000)
                                               Aee Group
Melanoma Skin Cancer  10-14  15-24  25-34  35-44  45-54  55-64  65-74  75-84  85+
Face. Head, Neck and
Upper Extremities
    Males              0.0    0.8    2.4    5.3    5.5    8.4   10.4    9.2    9.1
    Females            0.1    0.9    3.2    4.2    5.3    5.4    4.5    4.6    4.9
Trunk and Lower
Extremities
Males
Females

0
0

.0
.1

1.2
1.4

3.4
5.3

7,
6,

.5
.9

7.9
8.6

12.0
8.9

14.9
7.4

13.2
7.4

13.0
8.0
Source:  Derived from Scotto and Fears (in press) and National Cancer Institute
         SEER Report (1984).

-------
                                      F-13
melanoma skin cancer defined above,  mortality is modeled without dividing the
data into the two cancer types.  Exhibit F-7 displays the mortality rates
currently used in the model.

    The incidence and mortality associated with melanoma skin cancers in
nonwhite populations in the U.S. are less well defined.  The rates for
nonwhites are much lower than the rates for whites,  and are consequently not
modeled.

    The model user may also specify that the baseline incidence rates may
change over time (increase or decrease).  The purpose of this specification is
to allow the user to include the implications of potential changes in the
baseline incidence of melanoma skin cancer that are not related to changes in
UV flux.

    Senile Cataracts

    Epidemiological studies have identified a correlation between the
prevalence of various types of cataracts in humans and the flux of sunlight or
ultraviolet radiation reaching the earth's surface (Hiller, Giacometti and Yuen
1977; Zigman, Datiler, and Torczynski 1979; Taylor 1980; Hollows and Moran
1981).  Hiller, Sperduto and Ederer (1983) developed a multivariate logistic
risk function that describes the correlation found between the prevalence of
senile cataracts and the flux of UV-B and other risk factors.  The results of
this study were used to indicate the magnitude of change in the prevalence of
senile cataracts that could be associated with changes in UV-B flux due to
ozone depletion.

    The study by Hiller, Sperduto, and Ederer included HANES data on a total of
2,225 persons between the ages of 45 and 74 years who had resided in the
state where the HANES examination took place for at least one-half of their
lives.  Of these 2,225 people, 413 (18.6 percent) were placed in the cataract
or aphakia outcome category.

    UV-B flux data were developed by NOAA for the 35 HANES locations used in
the study based on a statistical analysis of UV-B data collected at 10
locations using Robertson-Berger meters.  The statistical analysis incorporates
season, latitude, elevation,  weather (clouds), and haze.  Subsequent validation
of the estimates at six locations indicated that the differences between the
estimated and observed mean daily flux average about seven percent.

    These data on UV-B and outcome (i.e., cataract)  were used by Hiller,
Sperduto, and Ederer in conjunction with demographic and medical history data
to estimate the following multivariate logistic risk function:
                    P =
                        1 +
where P is the probability (or risk) of having a cataract, and X. are risk
factors.  In addition to UV-B, the following risk factors were analyzed:  age;
race; sex; education; diabetes; systolic blood pressure; and residence  (urban,
rural).

-------
                                     F-14
                                  EXHIBIT F-7

                   Mortality Rates for Melanoma Skin Cancers
                            (White Population Only)
                               (Rate per  100,000)
                                      Age Group
         10-14   15-24   25-34   35-44   45-54   55-64   65-74   75-84   85+
Male      0.0

Female    0.0
0.5     1.0     2.9     4.6     5.5     8.0

0.0     0.8     1.6     2.8     2.9     3.3
                                                                  7.7    10.3

                                                                  5.3    5.5
Source:   Derived from Scotto and Fears (in press) and National Cancer
         Institute SEER Report (1984).

-------
                                      F-15
    Exhibit F-8 displays the standardized regression coefficients estimated for
each of the risk factors.  Positive coefficients indicate factors that are
correlated with increased risk; negative coefficients indicate factors that are
correlated with decreased risk.  The coefficients presented in the exhibit are
"standardized," meaning that they represent the expected change in the logit of
P  (equal to In (P/l-P)) for a one standard deviation change in the risk factor.
Standardization of the coefficients allows the relative importance of the risk
factors to be identified by the relative size of the standardized coefficients.

    As shown in Exhibit F-8, three risk functions were estimated:  (1)
univariate (outcomes as a function of the risk factor); (2) bivariate (outcome
as a function of the risk factor and age); and (3) outcome as a function of all
the risk factors simultaneously.  For all three formulations, UV-B is
statistically significant, and positively correlated with the increased risk of
being in the cataract outcome category.

    Using the multivariate risk function coefficients, and the mean values for
all the risk factors other than UV-B, the change in the prevalence of cataract
for each 1.0 percent change in UV-B is estimated to be approximately 0.5
percent.  This relationship holds for changes in UV-B as large as minus 20
percent to plus 30 percent.  Outside of this range, reductions in UV-B are
associated with less of a reduction in cataract prevalence, and increases in
UV-B are associated with larger increases.

    Of note is that this estimated relationship between UV-B and cataract
prevalence varies with age; UV-B has a larger effect on prevalence (on a
percentage basis) among younger individuals.  Exhibit F-9 displays the percent
increase in cataract prevalence due to increases in UV-B for persons of
different ages.   As shown in the exhibit, the percentage increase in prevalence
due to changes in UV-B are estimated to be larger for 50 year olds than for 70
ye ir olds.

    Although the effect of UV-B on prevalence is estimated to be larger at
younger ages (on a percentage basis) using the multivariate risk function, the
prevalence of senile cataracts is known to increase substantially with age.
Leske and Sperduto (1983) report the prevalence of senile cataracts in both
sexes found in the Framingham Eye Study to be as follows:   52 to 64 years old --
4.5 percent;  65  to 74 years old -- 18.0 percent;  75 to 85 years old -- 45.9
percent.   These  prevalence estimates use the same definition of cataracts as
used by Hiller,  Sperduto, and Ederer.   Because cataracts are more prevalent in
older individuals,  increases in the actual number of cases of cataracts would
likely be larger for older individuals,  even though the percentage increase in
risk has been estimated to be larger for younger individuals.

Step 3 -- Apply Dose-Response Relationships

    The purpose  of applying the dose-response relationships is to estimate how
the risks of melanoma and nonmelanoma skin cancers and cataracts may change as
the flux of UV reaching the earth's surface changes due to changes in total
column ozone  abundance.   Two factors are required to apply dose-response
relationships:   (1)  a measure of exposure;  and (2)  a dose-response equation.
Each is described in turn.

-------
                                     F-16


                                  EXHIBIT F-8

               STANDARDIZED REGRESSION COEFFICIENTS FOR CATARACTS
Risk Factor a/ Mean
Age
Race
Sex
Education
Diabetes
Systolic Blood Pressure
UV-B
Residence
61.
1.
1.
2.
0.
146
3.
1.
b/
97
22
52
87
08
.3
59
37
Univariate c_/ Bivariate d/ Multivariate e/
1.
0.
0.
-0,
0.
0.
0,
0,
,22
,18
,07
,43
,25
,33
.19
.18
£/
f / 0.20
0.08
f/ -0.25 f/
f/ 0.23 f/
£/ 0.15 g/
f/ 0.20 f/
f/ 0.21 f/
1
0
0
-0
0
0
0
0
.20
.13
.08
.14
.21
.08
.13
.19
f/
£/

£/
f/

&/
f/
     a/ Values for categorical risk factors:  race:  1 = white, 2 = black; sex:
1 = male, 2 = female; education 1 = <5 grades, 2=5-8 grades, 3 = 9-11 grades,
4 = 12 grades, 5 = college; diabetes:  0 = absent, 1 = present; residence:  1 =
urban, 2 = rural.

     b/ Mean value for the risk factor in the 2,225 persons in the study.

     c/ Each risk factor analyzed separately.

     d/ Each risk factor analyzed with age only.

     e/ All risk factors analyzed simultaneously.

     f/ p(two sided) <0.005.

     £/ p(two sided) <0.05.

     Source:  Hiller, R., R. Sperduto, and F. Ederer, "Epidemiologic
Associations with Cataract in the 1971-1972 National Health and Nutrition
Examination Survey," American Journal of Epidemiology, Vol. 118, No. 2, pp.
239-249, 1983.

-------
                                          F-17
                                       EXHIBIT F-9


                    ESTIMATED REIATIONSHIP BETWEEN RISK OF CATARACT
                                     AND UV-B FLUX
Percent
Increase
in
Cataract
Prevalence
                                                                            50

                                                                            60
= 70
                                10        15       20       25


                              Percent Increase in UV-B Flux
    Increased UV-B flux  (measured with  an RB-meter)  is associated with increased
    prevalence of cataract.  The percent  change  in prevalence varies by age.


          Source:  Developed from data  presented in R.  Killer,  R. Sperduto, and
    F. Ederer, "Epidemiologic Associations with  Cataract in the 1971-1972 National
    Health and Nutrition Examination  Survey,"  American Journal of Epidemiology. Vol,
    118, No. 2, pp. 239-249, 1983.

-------
                                     F-18
    The measure of exposure used to model skin cancers and cataracts is related
to the level of ambient UV flux estimated to reach the earth's surface in the
three regions of the U.S.   To use the dose-response relationships incorporated
into the model, the change in UV flux (in percent) associated with changes in
ozone abundance is used -- the change in the absolute level of UV energy
reaching the earth's surface (in units such as millijoules per meter squared) is
not used.  (The method used to estimate changes in UV flux (in percent) along
several scales of UV measurement was described above.)

    Given these estimates  of percent changes in UV over time for given locations
in the U.S.  (i.e., the three regions), the change in exposure for individuals
estimated to live in those areas is estimated on an age-specific basis.  In year
T of the analysis, the change in UV flux for a person who is 40 years old is a
function of the change in UV estimated from years T-40 to year T.  Changes in UV
for years prior to 1985 are assumed to be zero.

    The current understanding for nonmelanoma skin cancer is that the cumulative
exposure received during a person's life is the appropriate measure to use for
modeling the dose-response relationship.  Insufficient evidence exists to define
the appropriate measure of exposure for purposes of modeling cataract
prevalence.   Similarly, a major question remains regarding the most appropriate
measure of exposure for modeling the incidence and mortality of melanoma skin
cancer.  Recent analyses of melanoma incidence by Scotto and Fears (in press)
have used annual RB-meter values as estimates of exposure.  Pitcher (1986) used
peak values to model mortality.  Although some evidence indicates that peak or
intermittent exposures to UV may best explain melanoma incidence patterns (see
Chapter 8),  significant research remains to be done on this issue.  Of note is
that the UV Model indicates that changes in UV flux (in percent) as a function
of changes in ozone levels are about the same for the annual values and peak
values for a given action spectrum.  However, the choice of peak UV flux versus
annual UV flux (or flux measured over some other time period) influences the
estimate of the relationship between mortality risks and UV significantly (by as
much as 60 percent).  Therefore, the choice of annual or peak UV flux as an
exposure estimate for purposes of modeling risk influences the risk estimates
substantially.

    A related issue regarding modeling exposure for purposes of estimating
melanoma risks is the relative weight to put on exposures (annual or peak)
experienced throughout a person's life.  Migration studies indicate that
exposure before the age of 15 may be most important and should be given more
weight than subsequent exposures (see Chapter 8).  These relative weights have
yet to be quantified, however.  To allow alternative assumptions regarding the
relative importance of exposure during a person's life to be modeled, relative
weights can be assigned to different ages during a person's life for purposes of
estimating changes in exposure.  Exhibit F-10 presents a simplified version of
the tables that the user can use to define these weights.

    As shown in the exhibit, for a person who is currently 40 years old, the
relative weight to give to the exposure that a person of that age has received
during his life is:  2.0 times the exposure received from ages zero to 10, 2.0
times the exposure received from ages 10 to 20; 1.0 times the exposure received
from ages 20 to 30; and 1.0 times the exposure received from ages 30 to 40.
With these assumptions, the childhood exposures of individuals that are
currently 40 years old are weighted twice as heavily as subsequent exposures.

-------
                             F-19
                         EXHIBIT F-10

             Sample Table for Specifying Relative
       Weights for Exposure During a Person's Lifetime
Current Age
(years)
10
20
30
40
50
60

0
2
2
2
2
2
2

-10
.0
.0
.0
.0
.0
.0

10

2
2
2
2
2
Age
-20
-
.0
.0
.0
.0
.0
During Exposure (years)
20-30 30-40 40-50
-
-
1.0
1.0 1.0
1.0 1.0 1.0
1.0 1.0 1.0

50-60
-
-
-
-
-
1.0
All values are illustrative.

-------
                                       F-20


The assumptions used in the model for both melanoma skin cancers and cataracts
are equal weights across all ages.  The user may modify these assumptions as
desired to evaluate alternative formulations.

    The dose-response relationships currently included in the model are called
"power functions" and are generally of the form:

                    In(incidence) = a + b * ln(E),

    where:

        incidence = the age-specific incidence for a given race and sex;

        a = a constant that varies by age, race, and sex;

        b = a constant that varies by age, race and sex; and

        E = measure of lifetime exposure to ultraviolet radiation.

Using this equation, the fractional change in incidence as a function of the
fractional change in UV can be expressed as:

    fractional change in incidence = (fractional change in exposure +1)"3 -1

The value of this formulation is that:

        o   the change in incidence is expressed as a fractional change
            that can be multiplied by the baseline incidence to compute
            the increased age-specific incidence; and

        o   the change in exposure can be expressed as a fractional
            change, thereby avoiding potential difficulties in
            specifying the absolute levels of changes in exposure to UV
            (see Serafino and Frederick (1986) for a description of
            these potential difficulties).

    To specify the dose-response relationship, the user must choose a measure of
UV (e.g., RB-Meter, DNA, or Erythema action spectrum) and provide estimates of b
in the above equations.  Scotto, Fears, and Fraumeni (1981) present a range of
estimates of b for white males and females for use with the RB-Meter action
spectrum."  Their analysis indicates that these coefficients do not vary by age,
indicating that the incidence at each age is correlated to changes in exposure
to UV in the same manner.
       The coefficients for nonmelanoma were derived from the regression
coefficients presented in Scotto, Fears, and Fraumeni (1981, p. 10).  The values
reported were for an exponential formulation of the risk model.  To translate
the coefficients into the power formulation, the coefficients were multiplied by
135 UV units, the ambient UV exposure level at which the power function and
exponential function models predict equal incidence results.

-------
                                     F-21
    As described in Chapter 7, an analysis of the data in Scotto, Fears, and
Fraumeni (1981) was performed using the DNA action spectrum.  The estimated
coefficients did not differ significantly from the coefficients reported by
Scotto, Fears and Fraumeni for the RB-Meter action spectrum.  This result is
expected because at current ozone levels the RB-Meter and DNA UV flux values are
highly correlated.  Exhibit F-ll displays the estimated coefficients for the
incidence of basal and squamous cell nonmelanoma skin cancer used with the DNA
action spectrum.  The risk estimates in Chapter 18 are based on the DNA Action
Spectrum, with a sensitivity analysis to evaluate the implications of using the
Erythema Action Spectrum in its place.

    The coefficients for the dose-response relationships for each of the types
of melanoma were derived from data presented in Scotto and Fears (in press). u
These coefficients vary by type of melanoma and sex.  As with the coefficients
for nonmelanoma, the values do not vary by age, and are for whites only.  These
values are also presented in Exhibit F-ll.  These coefficients are assumed to be
applicable for the DNA action spectrum.

    A dose-response relationship for nonmelanoma mortality has not been
estimated.  For purposes of modeling, the relationships for nonmelanoma
incidence (basal and squamous) are assumed to apply.

    A separate analysis of melanoma mortality has been performed by Pitcher
(1986).  The coefficients estimated as the result of that investigation are used
to model the potential increase in mortality due to melanoma skin cancer.  The
coefficients estimated using peak UV flux values (weighted using the DNA action
spectrum) as the measure of exposure are reported in Exhibit F-12.  Analysis is
continuing on alternative measures of exposure.  Preliminary results indicate
that using annual UV flux as the exposure measure would result in coefficients
that are as much as 60 percent smaller.
     lO The information reported in Scotto (in press) reported the percent
increase in the incidence in each of the different types of melanoma for a 10
percent increase in UV.  To translate these data into coefficients in the power
formulation of the risk model, the power function was rearranged to solve for
the coefficient (b) as follows:

                              Change in incidence
                       ln(l + - )
                                   incidence
                   b=  - ' -
                                 Change in UV
                                 - )
                                      UV

For example, an 8 percent increase in incidence resulting from a 10 percent
increase in UV yields the following coefficient:  ln(l . 08)/ln(l . 1) = 0.807.

-------
                             F-22
                        EXHIBIT  F-ll

                Coefficients Relating Percent
   Change in UV to Percent Change in Skin Cancer Incidence
    (For vise with the DNA Action Spectrum -- Whites only)
                                  Low
          Middle
           High
NONMELANOMA SKIN CANCER
  Squamous
    Male
    Female

  Basal
    Male
    Female
1.42
1.47
0.932
0.316
2.03
2.22
1.29
0.739
2.64
2.98
1.65
1.16
MELANOMA SKIN CANCER
  Face.  Head and Neck
    Male                          0.661     0.846      1.029
    Female                        0.798     1.019      1.236

  Trunk and Lower Extremities
    Male                          0.421     0.651      0.875
    Female                        0.341     0.522      0.700
a/ Middle value minus one standard error.

b/ Middle value plus one standard error.

Sources:  Melanoma coefficients derived from Scotto and Fears
          (in press).  Nonmelanoma coefficients presented in
          Chapter 7.

-------
                      F-23
                  EXHIBIT F-12

  Coefficients Relating Percent Change in UV to
      Percent Change in Melanoma Mortality
(For Use with DMA Action Spectrum --  Whites Only)
                    a/                    b/
                 Low       Middle     High
     Males       0.78       0.85       0.92

     Females     0.50       0.58       0.66
     a/ Middle estimate minus one standard
        error.

     b/ Middle estimate plus one standard
        error.

     Source:   Pitcher (1986).

-------
                                      F-24
    Coefficients for dose-response relationships for blacks and members of other
races have not been developed.  Skin cancer risks for blacks and members of
other races are not modeled.

    The correlation between UV-B and cataract prevalence estimated by Killer,
Sperduto and Ederer (1983) is used to model the potential changes in the
incidence of cataracts as a function of changes in UV-exposure   These
prevalence values were used to compute incidence estimates (again, in the
absence of ozone depletion) under the assumption that mortality and cataract
prevalence are not related, and that the people in each age group (e.g., 55 to
64 year olds) are distributed uniformly across the ages.  Based on these
assumptions, the incidence rates assuming zero ozone depletion are as follows:
55 to 64 years old -- 450 per 100,000; 65 to 74 years old -- 1,350 per 100,000;
75 to 85 years old -- 2,750 per 100,000.  Because (in the absence of ozone
depletion) prevalence of cataracts in those 85+ years old is assumed to be equal
to the prevalence in those 75 to 85 years old, there is no incidence among those
over 85.

    To evaluate the impact of ozone depletion on cataract incidence, a "power
model" was used to relate increases in prevalence to changes in lifetime UV
radiation exposure.  The use of the power model is justified because the percent
change in prevalence divided by the percent change in UV radiation is fairly
constant over a wide range of changes in UV.  Exhibit F-13 displays the
age-weighted coefficients used in the analysis.  The age-weighting was performed
because the dose-response coefficient varies by age.

    Of note is that these coefficients were developed using Robertson-Berger
(R-B) meter measures of annual UV radiation flux as the surrogate measure of
exposure, and that the coefficients were then used with the DNA-damage action
spectrum and erythema action spectrum estimates of changes in UV flux due to
ozone depletion.  Use of R-B meter-derived coefficients for the dose-response
relationship coupled with estimates of changes in DNA-damage-weighted and
erythema-weighted UV radiation flux in response to ozone depletion may result in
overestimating the increase in cataract incidence in response to ozone
depletion.

    There are various important limitations in the use of these estimates and
data.  The correlation between UV-B and cataracts reported by Hiller,
Sperduto, and Ederer does not prove a causal connection -- other  (unknown)
factors could be playing a role.  These (unknown) factors would have to be
correlated with UV-B flux.  Also, the study does not have estimates of
individual lifetime UV-B exposure, thereby limiting the strength of the evidence
for the association between UV-B exposure and cataracts.  Additionally, the
sample population analyzed may not be representative of the entire U.S.
population.  Finally, the outcome category used in the study does not
differentiate between different types of cataracts, some of which may be more
strongly related than others to UV-B exposure.

    Confidence in the estimates developed here are strengthened by several
considerations.  The correlation between UV-B flux and sunlight flux is high,
and a correlation between sunlight and cataracts has also been found in
Australia  (Taylor 1980) and in China  (Mao and Hu 1982).  An association between
UV-B exposure and cataracts has also been demonstrated in laboratory animals.

-------
                  F-25
              EXHIBIT F-13

DOSE-RESPONSE COEFFICIENTS -- CATARACTS




Low a/          Middle          High b/


0.127           0.225            0.296


a/ Middle minus one standard error.

b/ Middle plus one standard error.

Source:  Derived from data presented in:
         Hiller, Sperduto,  and Ederer,
         "Epidemiologic Associations with
         Cataract in 1971-1972 National
         Health and Nutrition Examination
         Survey, "American Journal of
         'Epidemiology. Vol. 118,  No. 2,
         pp. 239-249, 1983.

-------
                                      F-26


Therefore, although considerable investigation remains to be performed,
indications are that the association between UV-B and cataracts is a reasonable
basis for evaluating potential impacts due to increased UV-B flux associated
with ozone depletion.

Step 4 -- Summarize the Results

    The dose-response relationships are applied to each of the age groups of the
population over time using estimated changes in UV exposure as outlined above.
The results of the method include:

         o    changes in the total  population incidence rates
              (e.g., per 100,000 people) for a standard population
              age, sex, and race distribution;

         o    the total number of cases of each type of cancer
              over time, and the change in the number of cases
              from the number that  would exist in the absence of
              changes in UV flux associated with changes on total
              column ozone;  and

         o    the total number of deaths due to each type of
              cancer over time, and the change in the number of
              deaths from the number that would exist in the
              absence of changes in UV flux associated with
              changes on total column ozone.

    These results are reported for  the U.S. as a whole (the calculations are
performed separately for each o-f the three U.S. regions).  The increased number
of cases and deaths are reported separately for those individuals that are alive
today and for those individuals who are born in the future.

DEGRADATION OF POLYMERS

    The purpose of this portion of  the module is to evaluate the potential
damages and costs associated with polymer degradation that may be caused by
increases in the flux of ultraviolet radiation (UV) reaching the earth's
surface.  As described in Chapter 13, UV can degrade the physical and appearance
characteristics of polymers.  This  module evaluates the economic impacts that
increased UV flux may have on polymer performance based on studies by Andrady
(1986) and Horst (1986).

    There are several types of polymers that may be adversely affected by
increases in UV, including:   polyvinylchloride (PVC);  acrylics; polycarbonate;
polypropylene;  and polyester (Andrady 1986, p. 22).  To date, data for
assessing the potential future markets and potential damages due to UV have only
been developed for the portion of the PVC market used in construction (siding,
window profiles, rainwater systems, and pipe and conduit).  This PVC market
accounts for about 26 percent of all polymers subject to exposure to UV
(measured by production volume), and consequently, the estimates from this
module are underestimates of the potential risks of polymer degradation due to
ozone depletion.

-------
                                      F-27
    Recognizing that the current data only account for a portion of the polymer
market, the module is designed to evaluate up to five different types of
polymers using the method outlined for PVC.  As data for other polymers become
available, it will be integrated into the module as appropriate.

    The effects of UV on polymers are evaluated in three steps:

         1.   Project the size of the market for polymers that may be
              subject to degradation due to UV.

         2.   Assess the damage that increases in UV (due to ozone
              depletion) may have on the polymers.

         3.   Assess the costs of the damage to the polymers.

Each step is described in turn for evaluating PVC in construction applications.

Step 1 -- Market for Polymers

    The future market for PVC polymers in construction will be influenced by
construction activity and the costs and performance of substitute materials.
Horst  (1986) developed several approaches for projecting the expected production
and use of PVC through 2075.   The approach used here reflects expected market
saturation and is of the following form:
                             A ( 1 - exp (-k(T - TQ)))
    where:
         Q = quantity per person per year in pounds;
         A = a constant;
         k = a constant;
         T = the current year of the projection; and
         T  = a base year identified for the polymer type.

    The constants and base year computed for PVC by Horst (1986, p. 4A-3) using
statistical analyses of historical data are as follows:  A = 63.0308; k =
0.01265; and T  = 1966.  These values are used along with the U.S. population
scenario (see Appendix B) to compute a middle estimate of the size of the future
market for PVC in pounds.  By using the population scenario from the scenarios
module, the estimates of the PVC market will be consistent with the values used
to develop the scenarios of production of CFCs and the other trace gases and the
evaluation of human health effects.

    Because there is uncertainty in the expected future market for PVC, the
module also produces low and high estimates of future demand.  The low and high
estimates are based on the statistical uncertainty in the estimates of A and k.
By approximating the future demand, Q, with the first three terms of its
expansion series, Horst  (1986) showed that the variance of Q could be estimated
from the covariance matrix of A and k.  Specifically,

-------
                                      F-28


                         Var (Q) = Var(Ak)  (T-T )2.

                                   - Covar(Ak,Ak2) (T-T )3
                                                       o

                                              (T-T
The model computes the variance of demand in each year.  The low estimate is set
at one standard deviation below the middle estimate; the high estimate is set at
one standard deviation above.   These low and high values (along with the middle
value described above) are used in step 3 to evaluate the costs associated with
the degradation of these polymers.

Step 2 -- Polymer Damage

    A damage index that indicates the level of degradation of polymer
characteristics was developed by Andrady (1986).  When there is no degradation,
the index value equals 1.0.  Degradation is indicated by increasing values of
the index, such as 1.1.  This index is used here to describe the potential
damage to polymers due to increased UV flux associated with ozone depletion.

    Andrady (1986) also developed an estimate of how the damage index may vary
as a function of ozone depletion.  Due to uncertainties in the relationship
between ozone depletion and UV, and UV and polymer characteristics, the
relationship between ozone depletion and the damage index is represented as a
range.  These ranges are displayed in Exhibit F-14.

    To evaluate the damage index in each year, the ozone depletion estimate for
that year is used get a range of values in the damage index table.  The ozone
depletion estimate for the U.S. is taken as the estimate for the middle U.S.
region.  The result of this method is a range of damage index values for each
year (low, middle, and high).

Step 3 -- Assess Damage Costs

    The costs of polymer degradation depend on the manner in which the
characteristics of the polymers degrade, and the steps that are taken in
response to the degradation.  Horst (1986) developed an approach for evaluating
the costs associated with the production of new PVC each year.  These costs
represent the implications of changing the formulation of PVC during manufacture
in order to maintain its characteristics in light of increased UV exposure.
These costs do not include the potential damages to PVC already in place.
Additionally, these estimates do not reflect the losses that may be associated
with polymer degradation in the absence of changes in the polymer formulation.
Therefore, these estimates are an underestimate.

    Horst (1986) identified the costs of changing the formulation of PVC as a
function of the increased amount of stabilizer that needs to be added to the
polymer to maintain its characteristics.  A 25 percent increase in stabilizer
was estimated to lead to a 1.86 percent increase in the price of PVC, from its
current cost of $0.604 per pound  (Horst 1986, p. 5-16 and 6-8).  The increased
amount of stabilizer required as a function of ozone depletion was identified by
Andrady (1986) and is presented in Exhibit F-14.

-------
                               F-29
                            EXHIBIT F-14

              Damage Index and Increase in Stabilizer
                   For Ranges of Ozone Depletion
Ozone Depletion
(percent)
0-5
5-10
10-15
15-20
Damage Index Stabilizer Increase (%)
Low Middle High Low Middle High
1.01 1.015 1.02 1.0 3.0 5.0
1.01 1.025 1.04 1.0 5.0 9.0
1.02 1.055 1.09 3.0 11.5 20.0
1.03 1.105 1.18 3.0 20.5 38.0
Source:  Derived from Horst (1986),  p,  6-10.

-------
                                      F-30
    To estimate the costs of increasing the amount of stabilizer in response to
ozone depletion, Horst recommends using a ten year lag between the increase in
stabilizer required and the increase in the price of PVC.  This lag reflects the
time needed for the industry to respond to changing environmental conditions and
the fact that the current stabilizer concentrations include a margin of safety.
This ten year lag is used in the module, although the model user may override it
with an alternative assumption.

    To compute the costs in year T, the following computations are performed:

        o   compute the increased amount of stabilizer required, based
            on the ozone depletion in year T - 10;

        o   compute the increase in price by interpolating between no
            increase in price,  and a 1.86 percent increase in price for
            a 25 percent increase in stabilizer;

        o   given the increase in price, compute the cost as follows:

            C(T) = (D(T)/PQb)/(l+b) * [P^1+b) . p^1^]

            where:

                C(T) = cost in year T;

                D(T) = demand in year T;

                P    = price of PVC in the absence of changes in the
                       formulation of the polymer;

                b    = price elasticity of demand for PVC;

                PI    = the new price for PVC given the change in the
                       formulation of the polymer.

    This equation represents the annual loss in consumer surplus associated with
the estimated changes in price.  The price elasticity of demand estimated by
Horst (1986) is -1.956.  This value represents (in part) the estimated
availability of appropriate substitutes for PVC.

    This equation can be evaluated three times each year to generate low,
medium, and high estimates of the costs.  The low estimate uses the low demand
value (from step 1) and the estimate of the change in price associated with the
low estimate of the increase in stabilizer associated with ozone depletion.  The
middle and high estimates use the middle and high values for these components,
respectively.  If there is an increase in ozone abundance, the damage index is
set to 1.0 and there are no estimated costs.  The maximum price increase (1.86
percent) estimated by Horst is associated with a 25 percent increase in
stabilizer.  Price increases are capped at this value (1.86 percent) even when
an increase in stabilizer exceeding 25 percent may be indicated as being
required.  The present value of the costs over time can be evaluated by
discounting these annual costs at user-supplied discount rates.

-------
                                      F-31
LIMITATIONS

    The Effects Module draws together a diverse set of analyses and applies
these analyses to estimates of global ozone depletion (from the Atmospheric
Science Module) to evaluate potential risks to human health.  The analyses
require that a chain of events from ozone depletion to health outcomes be
modeled.  There are limitations in each link of this chain, creating
uncertainties in the estimates of the outcomes.

    The estimates of changes in UV as a function of changes in the ozone level
are based on a model.   The model is in its early stages of field validation, and
additional analyses remain to be performed.

    The dose-response relationships for skin cancers and cataracts all depend on
epidemiologic studies of humans in the U.S.  Because the studies are based on
humans in the U.S., there are no extrapolation problems (e.g., from animals to
humans, or from one population of humans to the U.S. population).  However, the
epidemiologic studies are generally limited by their measures of environmental
exposure to UV, as opposed to individual lifetime exposures.  In addition to the
limitations of the estimates of the dose-response relationship themselves, the
characterization of baseline evidence (in the absence of ozone depletion) and
baseline population characteristics are subject to uncertainty.

    In addition to uncertainties and limitations in the estimates produced by
the module, the module is limited because there are a variety of important
effects that are not evaluated.  These omitted effects include:

        o   polymers other than PVC;

        o   impacts on plants;

        o   impacts on aquatic organisms;

        o   impacts on urban smog formation;

        o   potential immune suppression in humans; and

        o   possible other diseases.

The omission of these effects biases  the estimates downward.  In Chapter 18,
qualitative estimates of some of these effects are presented based on
interpolations and extrapolations from case studies and research in early
stages.  Data used to prepare these estimates include:

        o   Impacts on soybeans.  For depletion of 25 percent, a 7.5
            percent reduction in yield is used (average of
            statistically-significant results reported in Chapter 11 for
            Essex and Williams cultivars).   For depletion between zero
            and 25 percent,  a value is interpolated between zero and 7.5
            percent.  Values are not  extrapolated outside this range.

-------
                                     F-32
            Impacts on Ground-Based Ozone (Smog).   Estimates of
            increased smog levels are interpolated from case study
            estimates (Chapter 14) as follows:
                Los Angeles:
                Philadelphia:
                Nashville:
16.6% depletion yields a 4.5 percent
increase;
33.3% depletion yields a 9.4 percent
increase;

16.6% depletion yields a 13.4
percent increase;
33.3% depletion yields a 33.0 percent
increase;

16.6% depletion yields a 23.8 percent
increase;  and
33.3% depletion yields a 50.0 percent
increase.
            Impacts on Aquatic Organisms.   Estimates for reduced anchovy
            populations are interpolated from laboratory tank
            experiments (described in Chapter 12)  as follows:
                    Percent Change
                       in UV-B
        Percent Annual
       Population Death
                          0%
                         10%
                         20%
                         30%
                         40%
                         50%
                         60%
           0%
           0%
           0% to 4.8%
           0% to 11.5%
         2.5% to 18.0%
         6.0% to 23.0%
        11.0% to 25.0%
Because these data are based on case studies and research in early stages,  all
the estimates of these effects are very preliminary and should be viewed with
caution.

-------
                                     F-33
REFERENCES


Andrady,  A.  (1986),  "Analysis of Technical Issues Related to the Effect of UV-B
    on Polymers," prepared for the U.S.  Environmental Protection Agency,
    Washington,  D.C.

Cole, C.A.,  P.D.  Forbes,  and R.E. Davies (1986),  "An Action Spectrum for UV
    Photocarcinogenesis," Photochemistry and Photobiology.  pp.  275-284.

Connell,  P.S.  (1986),  "A Parameterized Numerical Fit to Total Column Ozone
    Changes Calculated by the LLNL I-D Model of the Troposphere and
    Stratosphere," Lawrence Liverpool National Laboratory,  Livermore,
    California.

Elder, D.E., Jacoby P.M., Tuthill, R.J.  and Clark W.H. Jr.  (1980), "The
    Classification of Malignant Melanoma," American Journal of Dermatopathology.
    Vol.  2,  pp.  315-320.

Killer, R.,  R. Sperduto and F. Ederer (1981), "Epidemiologic Association with
    Cataract in the 1971-1972 National Health and Nutrition Examination Survey,"
    American Journal of Epidemiology. Vol. 118, No. 2, pp.  239-298.

Horst, R.L.  (1986),  "The Economic Impacts of Increased UV-B Radiation on
    Polymer Materials:  A Case Study of Rigid PVC," prepared for the U.S.
    Environmental Protection Agency, Washington,  D.C.

Isaksen,  I.S.A.  (1986), "Ozone, Perturbations Studies in a Two Dimensional
    Model with Temperature Feedbacks in the Stratosphere Included," presented at
    UNEP Workshop on the Control of Chlorofluorocarbons,  Leesburg, Virginia,
    September 1986.

Leske, C.L., and R.D.  Sperduto (1983), "The Epidemiology of Senile Cataracts:
    A Review," American Journal of Epidemiology.  Vol. 118,  No.  2, pp. 152-165.

Pitcher,  H.  (1986),  "Melanoma Death Rates and Ultraviolet Radiation in the
    United States 1950-1979," U.S. Environmental Protection Agency, Washington,
    D.C.   This paper has been revised and submitted for publication as: Pitcher,
    H. (1987), "Examination of the Empirical Relationships Between Melanoma
    Death Rates in the United States 1950-1979 and Satellite-based Estimates of
    Exposure to Ultraviolet Radiation."

Scotto, Fears, and Fraumeni (1981), "Incidence of Nonmelanoma Skin Cancer in
    the United States," U.S. Department of Health and Human Services, (NIH)
    82-2433, Bethesda, Maryland.

Scotto J. and T.  Fears (in press), "The Association of Solar Ultraviolet
    Radiation and Skin Melanoma Among Caucasians in the United States,"
    Cancer Investigation.

Serafino, G. and J.  Frederick (1986), "Global Modeling of the Ultraviolet
    Solar Flux Incident on the Biosphere," prepared for the U.S. Environmental
    Protection Agency, Washington, D.C.

-------
                                     F-34
U.S. Bureau of Census. (1984), Projections of the Population of the United
    States by Age. Sex. Race: 1983 to 2080.  Current Population Reports,  series
    P-25, No. 952, Washington, D.C.

-------
                                  CHAPTER 18

                    HUMAN HEALTH AND ENVIRONMENTAL EFFECTS
SUMMARY

    Future damage to human health and the environment depends on the magnitude
of future emissions of ozone depleters and of greenhouse gases into the
atmosphere.  This chapter presents a range of quantitative estimates of the
potential risks to human health and the environment arising from various
scenarios of production and emissions of ozone-modifying substances.

    The risk estimates draw on the data and methods described in earlier
chapters.  These chapters describe:  (1) potential future emissions and
concentrations of ozone-modifying substances; (2)  the effects that these
substances may have on stratospheric ozone and global climate; (3) the potential
damage to human health and the environment arising from ozone modification and
climate change; and (4) a comprehensive modeling framework for estimating these
risks.

    The major types of effects for which quantitative estimates are provided
include: ozone depletion; additional cases and deaths related to skin cancers;
additional cases of senile cataract; damage to polyvinyl chloride materials;
rise in global equilibrium temperature; sea level rise; cost of sea level rise;
reduction in soybean yield; increases in ground-based ozone; and effects on the
survivability of a selected aquatic organism.  Certain important risks discussed
in earlier chapters are not evaluated because of data limitations and the
absence of relevant scientific information.  In most instances, the estimates
cover the U.S. only.

    Estimates of the potential risks due to emissions of ozone-modifying
substances are necessarily uncertain for two reasons.  First, estimating future
risks requires projections of key factors, such as population and economic
growth.  Second, the many relationships that define how emissions result in
risks are themselves uncertain.

    To investigate the significance of key assumptions and data used to model
risks,  the chapter presents a range of risk estimates by varying the most
important assumptions one at a time.  The range of results obtained by varying
each assumption is used to determine which factors used to model risks are most
important.

-------
                                     18-2
FINDINGS

1.  MODIFICATION OF THE TRACE GAS COMPOSITION OF THE ATMOSPHERE CAN BE EXPECTED
    TO ALTER COLUMN OZONE ABUNDANCE.

    la.  The range of global average  total column ozone change projected for the
         year 2075 based on a parameterized representation of a one-dimensional
         model could vary from as high as over 50 percent depletion, for a case
         where global use of chlorine and bromine bearing substances grows at an
         average annual rate of 2.8 percent from 1985 to 2100 (5.0 percent per
         year from 1985 to 2050,  followed by no growth through 2100),  to
         increased abundance of ozone of approximately 3 percent,  for a case
         where global use of chlorine and bromine bearing substances declines to
         20 percent of its 1985 value by 2010.  The six "what if"  scenarios
         examined include:

           o   80% Reduction:  Use of chlorine and bromine bearing substances
               declines to 20 percent of its 1985 value by 2010, and remains
               constant thereafter, yielding approximately 3.0 percent increased
               ozone abundance by 2075;

           °   No Growth:  no growth  in use of chlorine and bromine-bearing
               substances from 1985 to 2100, yielding approximately 0.3 percent
               increased ozone abundance by 2075;

           o   1.... 2% Growth:   1.2 percent growth from 1985 to 2050, followed by
               no growth, yielding approximately 4.5 percent depletion by 2075;

           o   2.5% Growth:   2.5 percent growth from 1985 to 2050, followed by
               no growth, yielding approximately 25 percent depletion by 2075;

           o   3.8% Growth:   3.8 percent growth from 1985 to 2050, followed by
               no growth, yielding over 50 percent depletion by 2075;

           o   5.0% Growth:   5.0 percent growth from 1985 to 2050, followed by
               no growth, yielding over 50 percent depletion by 2075.

           The trace gas concentration assumptions used in these six cases are:
           C02 -- NAS 50th percentile; CH4 -- 0.017 ppm per year (approximately
           1 percent of current CH4 concentration); and N20 -- 0.20 percent per
           year.

     Ib.   Current data are not sufficient for distinguishing whether CH4
           concentrations are likely to increase in a linear manner (e.g, at
           0.017 ppm per year, or approximately 1 percent of current
           concentrations) or in a compound manner (e.g., at 1 percent per year,
           compounded annually).  The sensitivity of the ozone change estimates
           in 2075 was evaluated for the following six assumptions regarding
           future CH4 concentrations:

           o   Scenario A:  compound annual growth of 1 percent from 1985 to
               2010, followed by constant concentrations at 2.23 ppm; and

-------
                                     18-3
           o   Scenario B:   linear growth at 0.01275  ppm per year (75 percent of
               the 0.017 ppm growth);

           o   Scenario C:   linear growth of 0.017  ppm per year (approximately 1
               percent of current concentrations);

           o   Scenario D:   linear growth at 0.02125  ppm per year (125 percent
               of the 0.017 ppm growth);

           o   Scenario E:   compound annual growth  of 1 percent;

           o   Scenario F:   compound annual growth  of 1 percent from 1985 to
               2020,  growing to 1.5 percent compound  annual growth by 2050 and
               thereafter.

           For the 2.5% Growth scenario,  the estimate of ozone depletion by 2075
           ranges from about 14 percent (Scenario F)  to 30 percent (Scenario A)
           across these six CH4 assumptions evaluated.   The difference between
           the 1 percent linear (0.017  ppm per year)  and 1 percent compounded
           assumptions (Scenarios C and E) is approximately 6 percent depletion.
           This sensitivity of the ozone depletion  estimates to the assumption
           about linear versus compound growth of CH4 concentrations is much
           larger than the sensitivity  to the range of assumptions examined
           regarding future C02 concentrations (from  the 25th to the 75th
           percentile NAS estimates)  and regarding  future N20 concentrations
           (from 0.15 percent annual compound growth  to 0.25 percent annual
           compound growth).

2.    TWO-DIMENSIONAL (2-D)  MODELS PREDICT GREATER AVERAGE GLOBAL DEPLETION THAN
     ONE-DIMENSIONAL (1-D)  MODELS.  2-D MODELS ALSO PREDICT THAT OZONE DEPLETION
     WILL EXCEED THE GLOBAL AVERAGE AT  HIGH LATITUDES AND BE LESS THAN THE
     GLOBAL AVERAGE AT THE EQUATOR.

     2a.   For a case of 3 percent annual growth in emissions of CFCs, no
           emissions of Halons,  and increases in trace gases of:  C02 --
           approximately 0.6 percent per year; CH4  -- 1 percent per year; and
           N20 -- 0.25 percent per year,  a 2-D model  estimates approximately 5.4
           percent global average depletion by 2030.   For the same scenario of
           emissions and trace gas concentrations,  the parameterized
           representation of a 1-D model estimates  only 3.0 percent depletion by
           2030.

     2b.   For this same case of emissions and trace  gas concentrations, the 2-D
           model estimates of ozone depletion in 2030 at high latitudes are
           approximately: 60°N -- 8.7 percent; and  50°N -- 7.0 percent.

3.    ESTIMATES OF ATMOSPHERIC MODIFICATION. SKIN CANCER CASES AND DEATHS.
     CATARACT CASES.  MATERIALS DAMAGE.  GLOBAL TEMPERATURE. AND SEA LEVEL DEPEND
     ON THE RATE AT WHICH OZONE-DEPLETING GASES GROW. ATMOSPHERIC RESPONSE. DOSE
     RESPONSE. AND WHETHER GREENHOUSE GASES THAT COUNTER OZONE DEPLETION GROW
     INDEFINITELY.  THE ASSUMPTIONS BEHIND QUANTITATIVE PROJECTIONS MUST BE
     NOTED CAREFULLY.

-------
                                     18-4
     3a.    The  models  used in  this  risk assessment  assume  that Antarctic  ozone
           depletion has  no global  implications  and that global  trends  do not
           invalidate  estimates  of  current  models.

     3b.    Except  as noted,  projected effects  assume that:  greenhouse gases  grow
           at historical  rates indefinitely; current one-dimensional models
           accurately  project  depletion;  production of ozone depleters  does  not
           grow after  2050;  ozone depletion is limited to  50 percent; the action
           spectrum causing skin cancers is DNA;  and the temperature sensitivity
           of the  earth to doubled  C02 is 3°C.

     3c.    In 2100,  projections  of  ozone depletion  range from over 50 percent
           for  the 5%  Growth scenario (ozone depletion is  constrained at  50
           percent in  this analysis)  to 47  percent  for the 2.5%  Growth  scenario
           to an increase in column ozone abundance of nearly 5  percent for  the
           80%  Reduction  scenario.

     3d.    For  cohorts born before  2075,  the number of nonmelanoma skin cancers
           projected ranges from a  261.5 million increase  for the 5% Growth
           scenario to a  115 million increase  for the 2.5% Growth scenario to  a
           reduction of 6.5 million skin cancers for the scenario of 80%
           Reduction in all ozone depleters.

     3e.    For  cohorts born before  2075,  total melanoma cases ranges from a  1.3
           million case increase for the 5% Growth  scenario to  a 609,000
           increase for the 2.5% Growth scenario to 54,000 fewer cases  for the
           scenario of an 80%  Reduction in all ozone depleters.

     3f.    For  cohorts born before  2075,  total mortality from melanoma  and
           nonmelanoma ranges  from  a 5.6 million increase  for the 5% Growth
           scenario to a  2.4 million increase  for the 2.5% Growth scenario to
           115,000 fewer  cases for  the scenario  of  80% Reduction in all ozone
           depleters.

     3g.    For  cohorts born before  2075,  the increase in total  cataract cases
           ranges from 26 million  for the 5% Growth scenario to 15.1 million for
           the  2.5% Growth scenario to 9,500 for the scenario of 80% Reduction
           in ozone depleters.

     3h.    The  rise in global  temperature by 2075 ranges from 11.6°C  in the  5%
           Growth scenario to  5.6°C in the 2.5%  Growth scenario to 4°C  in the
           scenario of 80% Reduction in all ozone depleters.

     3i.    Impacts are also projected for other  areas such as sea level rise,
           ground-based ozone, materials, aquatics, and soybean yield.

4.   QUANTITATIVE ESTIMATES OF RISKS VARY WITH ASSUMPTIONS ABOUT FUTURE
     EMISSIONS  OF GREENHOUSE GASES  THAT WILL CONTRIBUTE TO GLOBAL WARMING.

     4a.    Model projections that  extrapolate historical growth rates  of
           greenhouse  gases, which tend to counter ozone depletion, into  the
           indefinite  future assume certain policy decisions from future
           decisionmakers; alternative assumptions are possible.

     4b.    If future decisionmakers limit the concentrations of C02,  N20, and
           CH4 to prevent global warming from exceeding 2°C  (±50%) in 2075,  they

-------
                                     18-5
           would by necessity have to limit growth of ozone depleters to the No
           Growth case;  for other cases increases in ozone depleters would be
           too large to  achieve that objective.

     4c.    Ozone depletion associated with the No Growth or 1.2% Growth
           scenarios increases nearly 3 to 5 percent if global warming is
           limited to 3°C (+50%);  skin cancer deaths would increase 43 percent
           for people alive today.

     4d.    Estimates of  methane emissions are inherently uncertain even without
           consideration of future policy decisions and could affect
           quantitative  risk estimates.

5.    QUANTITATIVE ESTIMATES OF RISK VARY WITH UNCERTAINTY ABOUT DOSE-RESPONSE
     COEFFICIENTS.  ACTION SPECTRUM.  LIMITS OF OZONE DEPLETION. AND
     RESPONSIVENESS OF MODELS TO ATMOSPHERIC DEPLETION.

     5a.    For people alive today and born before 2075,  additional skin cancer
           cases would be reduced 45 percent if  one assumes the lower dose-
           response coefficients that are one standard error below the best
           estimate and  66 percent higher if one assumes the higher coefficients
           that are one  standard error above the best estimate.

     5b.    For people alive today and born before 2075,  additional skin cancer
           cases would be reduced 11 percent if  the Erythema action spectrum,
           rather than the DNA action spectrum,  were used to measure health
           effects.

     5c.    Limiting projected depletion to 50 percent from what the
           parameterized 1-D model would project reduces projected deaths for
           later cohorts.   For people born from  2030 to 2074,  limiting depletion
           to 50 percent reduces deaths by 13 percent for the 2.5% Growth
           scenario and  66 percent for the 5% Growth scenario.

     5d.    For people alive today and born before 2075,  skin cancer cases would
           be reduced 62 percent in the 2.5% Growth scenario if the atmosphere
           were less sensitive to potential ozone depleters (using the 10th
           percentile),  and increased 54 percent if the atmosphere were more
           sensitive (using the 90th percentile).

6.    WHILE NATIONAL QUANTITATIVE ESTIMATES OF AQUATIC.  CROP.  GROUND-BASED OZONE.
     AND  SEA LEVEL RISE  DAMAGE CANNOT BE MADE AT THIS TIME. CASE STUDY RESULTS
     INDICATE THAT SIGNIFICANT INCREASES IN GROUND-BASED OZONE. LOSS OF AQUATIC
     LIFE. SEA LEVEL RISE DAMAGE.  AND LOSS OF CROP YIELD ARE POSSIBLE.

-------
                                      18-6
INTRODUCTION

    The previous chapters of this document have described:  (1) best estimates
of the future production, emission, and concentrations of ozone-modifying
substances; (2) the likely effects that these emissions and concentrations may
have on stratospheric ozone and global climate; (3) the potential risks to human
health and the environment arising from ozone-modification and global climate
change; and (4) a comprehensive modeling framework for estimating these risks.
This chapter integrates the data and methods presented in earlier chapters to
develop quantitative estimates of risks for human health and the environment due
to ozone modification and climate change.

    Quantitative risk estimates are made for six different "what if" scenarios
of emissions of ozone-depleting substances.  The scenarios have been designed to
cover a larger range of emissions than would be likely to occur either in the
absence of regulations or with very stringent regulations, thereby providing a
comprehensive assessment of risks of possible trajectories of future emissions.
In general, the scenarios of ozone-depleting substances are run through models
in conjunction with scenarios of greenhouse gas emissions that assume that
historical rates of growth for these trace species continue indefinitely.
Sensitivity analysis is performed, however, to assess the potential consequences
of limits to greenhouse gas emissions taken in order to reduce global warming.
With the exceptions of global sea level rise, global ozone depletion, and change
in global equilibrium temperature, all effects are estimated for the U.S. only.-*-

    The following ozone-depleting substances are considered in estimating
risks:^

        o   CFC-11, CFC-12, HCFC-22, and CFC-113;
        o   methyl chloroform (CH3CC13);
        o   carbon tetrachloride (CCL4); and
        o   Halon-1211 and 1301.

The following trace gases that add ozone or counter ozone depletion are
considered:^

        o   carbon dioxide (C02);
        o   methane (CH4); and
        o   nitrous oxide (N20).
    Exhibit 18-1 displays the types of effects that are estimated in this
     1 The estimates presented in this chapter are for the entire U.S.  As
described in Chapter 17, the U.S. estimates combine analyses of three geographic
regions (by latitude):  north, middle, and south.

     ^ See Chapter 3 for a description of the available estimates of future
production and emissions.

     3 See Chapter 4 for a discussion of the future concentrations of these
greenhouse gases.  These greenhouse gases also counter ozone depletion in  high
chlorine cases.  CC>2 is thought to increase ozone by cooling the stratosphere;
CH4 adds ozone to the troposphere and counters chlorine-induced depletion in the
stratosphere; nitrous oxide increases depletion  in the stratosphere if chlorine
levels are high.

-------
                                      18-7
chapter along with the chapter that describes the data and methods used to
derive the effects.  The effects are divided into two sets:  (1) effects that
can be estimated quantitatively for the U.S.; and (2) effects estimated on the
basis of case studies and research in early stages.   The first set is estimated
using detailed epidemiologic,  economic, demographic, and other scientific
information within the modeling framework described in Chapter 17.  Effects
estimated in this first set are representative of risks to the entire U.S., and
include:

        o   global average ozone depletion;

        o   nonmelanoma skin cancer (cases and deaths);

        o   melanoma skin cancer (cases and deaths);

        o   senile cataracts (cases);

        o   damage to polyvinyl chloride (PVC) materials;

        o   increased global average temperature;  and

        o   global sea level rise.

The second set of estimates is based either on research in early stages or on a
small number of case studies,  the results of which are more difficult to
generalize to the entire U.S.   Effects estimates included in the second set
should be considered preliminary and are included here to indicate the potential
order of magnitude of these effects.  The second set of estimates includes the
following types of effects:

        o   costs of sea level rise;

        o   reductions in soybean yields (for two cultivars; 2 out
            of 3 cultivars appear sensitive);

        o   increases in ground-based ozone  (smog);  and

        o   effects on the survivability of a selected aquatic
            organism (the Northern Anchovy population).

    Because of data limitations and the absence of relevant scientific
information, certain important effects discussed in earlier chapters were not
evaluated.  The major effects  not analyzed in this chapter are:

        o   overall impacts on agricultural and natural plants
            (Chapter 11);

        o   impacts of climate change  on human health, water
            resources,  forestry, and wetlands (Chapter 16);

        o   increased incidence of cutaneous diseases from immune
            suppression (Chapter 9);

-------
                             18-8
                         EXHIBIT 18-1

   Types of Human Health and Environmental Effects Estimated
  TYPE OF EFFECT
(Reference Chapter)
          UNITS
  Effects Estimated Quantitatively for the U.S.
  Ozone Depletion
  (Chapter 5)

  Nonmelanoma  Skin Cancer
  (Chapter 7)
  Melanoma Skin Cancer
  (Chapter 8)
  Senile Cataracts
  (Chapter 10)
  Damage to Polyvinyl Chloride
  Materials (Chapter 13)

  Rise in Equilibrium Temperature
  from 1985 to 2075 (Chapter 6)

  Sea Level Rise from 1985 to 2075
  (Chapter 15)
Percentage Depletion from
1985 to 2075

Additional Cases and Deaths
for People Alive Today; People
Born 1986-2029; and People
Born 2030-2074.

Additional Cases and Deaths
for People Alive Today; People
Born 1986-2029; and People
Born 2030-2074.

Additional Cases for People
Alive Today; People Born
1986-2029; and People Born
2030-2074.

Present Value Costs in
1985 U.S. Dollars

Degrees Centigrade
Centimeters

-------
                             18-9
                         EXHIBIT 18-1
                          (Continued)

   Types of Human Health and Environmental Effects Estimated
  TYPE OF EFFECT                                UNITS
(Reference Chapter)


  Effects Based on Case Studies and Research in Early Stages:

  Reduction in Soybean Yield          Percent in 2075
  (Chapter 11)

  Increase in Fatality Rate           Percent in 2075
  of Northern Anchovy Population
  (Chapter 12)

  Increase in Ground-Based Ozone      Percent in 2075
  (Chapter 14)

  Cost of Sea Level  Rise in           Present Value Costs
  Charleston,  SC and Galveston,        in 1985 U.S.  Dollars
  TX (Chapter 15)

-------
                                     18-10
        o   impacts of sea level rise on wetlands (Chapter 15);  and

        o   impacts on materials other than polyvinyl chloride
            (Chapter 13).

    Estimates of human health and environmental effects are necessarily
uncertain for two reasons.  Effects are estimated into the future, requiring
projections of such factors as population growth; economic growth; and
production, use, and emissions of ozone-modifying substances and greenhouse
gases.  Furthermore the linkages from emissions to risks are themselves
uncertain.  For example, the models and analyses used to evaluate ozone
modification and global warming continue to be developed and may undergo change,

    In order to understand the significance of uncertainties in the numerous
factors and assumptions used to model the chain of events that starts with
emissions and results in risks, following the presentation of the quantitative
risk estimates for six scenarios, the implications of various uncertainties
systematically explored so that a quantitative assessment of their relative
importance can be made.

    The remainder of this chapter is organized as follows:

        o   Methods for Estimating Health and Environmental Risks
            summarizes the major steps used to estimate risks.

        o   Description of Range of Production. Emissions, and
            Concentrations Scenarios For Evaluating Risks describes
            the key scenarios discussed throughout this chapter.

        o   Sensitivity of Health and Environmental Effects to
            Differences in Emissions of Ozone Depleters assesses the
            sensitivity of potential risks to assumptions about the
            growth in production and emissions of ozone-depleting
            gases, assuming greenhouse gases that counter ozone
            depletion grow at historical rates.

        o   Sensitivity of Results To Alternative Atmospheric
            Assumptions explores the relationships of potential
            risks to assumptions about: (1) the ozone-depletion
            model used; (2) sensitivity of global warming to
            concentrations of greenhouse gases; (3) future
            concentrations of greenhouse gases (4) ozone level
            response to emissions; and (5) the maximum level of
            ozone depletion allowed in modeling.

        o   Sensitivity of Effects to Uncertainty in Dose Response
            identifies how the risk estimates are influenced by
            uncertainty in estimates of:   (1) dose-response
            coefficients; and  (2) choice of action spectrum (i.e.,
            the weighting of wavelengths in the increased flux of UV
            that would be associated with various levels of ozone
            depletion).

-------
                                     18-11
        o   Relative Importance of Key Uncertainties summarizes the
            results of previous sections and compares the
            significance of the different uncertainties analyzed in
            this chapter.


METHODS FOR ESTIMATING HEALTH AND ENVIRONMENTAL RISKS

    A complete description of these methods is given in Chapter 17.  Briefly,
emissions of ozone-modifying substances and greenhouse gases lead to effects on
human health and the environment by means of the following principal mechanisms:

        1.  Production leads to emissions of ozone depleters and
            greenhouse gases, which in turn result in a time stream
            of ozone depletion.

        2.  Ozone depletion produces increases in ultraviolet
            radiation reaching the earth's surface.

        3.  Ultraviolet radiation increases have health impacts,
            including melanoma skin cancer, nonmelanoma skin cancer,
            and cataracts.

        4.  Changes in concentrations of ozone depleters and trace
            gases that counter ozone depletion directly cause global
            warming.

        5.  Changes in global temperature and ultraviolet radiation
            have other environmental impacts on sea level, aquatic
            organisms, agricultural yields, energy demand, and other
            concerns.

    As discussed in chapters 7, 8, 9 and 17, to quantify risks for human health
and the environment, each of these mechanisms was evaluated using mathematical
relationships.  For example, estimates of additional cases of melanoma skin
cancers rely on mathematical relationships between:  (1) emissions and ozone
depletion, (2) ozone depletion and UV, and (3) UV and cancer incidence.  The
sources and derivation of the mathematical relationships are described in
earlier chapters.

    The mathematical relationships were integrated in the modeling framework
described in Chapter 17.  The framework provides a method of estimating expected
risks as  well as analyzing the joint implications of key assumptions and
uncertainties in each mathematical relationship.  The primary period of analysis
used in the models  is 1985 to 2075.  Emissions of ozone-modifying substances and
concentrations of greenhouse gases are analyzed during  this period.  Risks to
human health are evaluated for individuals born before  2075, in three discrete
cohorts (i.e., groups):

        o  those individuals alive today  (i.e., 1985);

        o  those individuals born between now and 2029; and

-------
                                     18-12
        o   those individuals born between 2030 and 2074
            (the end of the analysis period).

Risk estimates for these individuals are based on simulated lifetime exposures
to increased UV.  For example, a 50 year old person who is alive today is
assumed to experience no increase in UV prior to 1986.   During his remaining
life, however, this 50 year old may experience increased exposure to UV as total
column ozone depletes.  Exposure histories for each population cohort are
analyzed separately as a basis for estimating risks.

    All the risk estimates presented below assume constant baseline incidence
and mortality rates; recent secular trends of increasing incidence are not
extrapolated into the future.

    Of note is that some of the individuals in the second and third cohorts
(born after 1985) will live beyond the end of the primary period of analysis
(2075).  To estimate the lifetime risk for these individuals, the time horizon
was extended to 2165 (i.e., a 90-year lifetime was assumed), holding emissions
constant after 2075.  Risk estimates for the second two population cohorts are
affected by the assumption that emissions are constant after 2075.  In most of
the scenarios investigated, depletion, UV radiation, and temperature are still
increasing in 2075.

    Environmental risks (e.g., global warming, sea level rise, and PVC damages)
are reported for 2075.  Risk estimates based on preliminary research and case
studies (aquatics, food, and ground-based ozone) are also reported for 2075.

    It is important to note that the models used to estimate risks do not
include feedbacks among factors-.  For example, the economic production used to
generate an emissions growth rate (2.5 percent for CFG emissions, for example)
is assumed not to be affected by the rate of global warming  (5.6°C for the 2.5%
Growth scenario).  To the extent that there are feedbacks from effects to
subsequent economic growth, that is, to the extent that effects may reduce
economic growth, the earth will be somewhat self-limiting in the growth of CFCs.
As now conceived, the risk models implicitly assume that the damage from effects
is not large enough to alter the economic growth of society.


DESCRIPTION OF RANGE OF PRODUCTION, EMISSIONS, AND CONCENTRATIONS SCENARIOS FOR
EVALUATING RISKS

    For purposes of evaluating the risks of emitting ozone-modifying substances
and greenhouse gases, scenarios of future production, emissions, and
concentrations of ozone depleters are required.  The purpose of the scenarios
chosen here is to cover the complete range of possible futures--that is, the
largest rate of  growth in the absence of regulation to the  greatest reduction  in
growth in the case of unanticipated technological breakthroughs or stringent
regulation.  It  is unnecessary to decide on a single reference  case for
evaluating risks  -- all quantitative estimates will be given in terms of
additional (i.e., incremental) effects.
     ^ This assumption of no prior increase  in UVB  is based  on  the  conclusion  in
 Chapter  5  that  it  is premature for this risk assessment  to conclude that  global
 ozone has  decreased.  If it turns out later  that  global  ozone has already begun
 to  decrease,  then  the risk estimates presented herein will be too low.

-------
                                     18-13
    As described above, the key substances of concern are the major ozone-
depleting substances:  CFCs; chlorocarbons;  and Halons (see Chapter 3).  In
addition, for the first set of runs we will assume that the greenhouse gases
that counter ozone depletion -- carbon dioxide (C02),  nitrous oxide (N20), and
methane (CH4) -- grow at historical rates into the definite future.

    In order to evaluate the risks of stratospheric changes due to these
substances, the scenarios of emissions and concentrations must extend over very
long periods of time because the impacts of the compounds are long-lived.  The
lifetimes of the key CFCs are on the order of 100 or more years.  Therefore, to
assess properly the implications of current and future emissions of these
substances, long-term scenarios are required.

    Scenarios for Ozone-modifying Substances

    Although Chapter 3 summarized the best available estimates from a variety of
authors of future production and emissions of the CFCs, chlorocarbons, and
Halons, use of those estimates is unnecessary at this juncture.  As discussed in
the chapter, the global production and emissions of these compounds is expected
to increase in the near term (e.g., through the year 2000) in the absence of
regulatory intervention due to factors such as economic growth, population
growth, the development of new applications for the compounds, and increases in
the saturation of existing applications.

    For purposes of assessing the full range of potential risks, however, we
will expand the production emission, and concentration scenarios far beyond the
best estimates made in Chapter 3, using six "what if" scenarios of global
production.  Through out this chapter, these six scenarios of future production
and emissions will be referred to as:

        o   80% Reduction Scenario--80% reduction in ozone-depleting
            substances by 2010, constant production thereafter.

        o   No Growth Scenario--0.0% annual growth through 2050, constant
            production thereafter.

        o   1.2% Growth Scenario--1.2% annual growth through 2050, constant
            production thereafter.

        o   2.5% Growth Scenario--2.5% annual growth through 2050, constant
            production thereafter.

        o   3.8% Growth Scenario--3.8% annual growth through 2050, constant
            production thereafter.

        o   5.0% Growth Scenario--5.0% annual growth through 2050, constant
            production thereafter.

    For each of these scenarios, CFC-11, CFC-12, HCFC-22, and carbon
tetrachloride (CCL4) all grow at the specified rates.  The following compounds
grow at slightly different rates:

-------
                                     18-14
    o   CFC-113 is assumed to grow at 1.5 times the growth rates in the
        1.2%, 2.5%, 3.8%, and 5.0% scenarios from 1985 to 2000.

    o   Methyl chloroform is assumed to grow slightly slower in the
        1.2%, 3.8% and 5.0% scenarios.

    o   Halons are assumed to grow as projected by Camm and Hammitt
        (1986).

    The production and emissions for each scenario result in simulated changes
in atmospheric chlorine concentrations (Clx),  which drive stratospheric ozone
depletion.  To assist readers in translating these scenarios tc various "real
world" situations, Exhibit 18-2 shows, for two of the scenarios, alternative
ways in which emissions could grow in the real world and achieve essentially the
same Clx levels.

    In Exhibit 18-2a the simulated Clx concentration in the No Growth case is
shown to increase to 5.0 ppbv by 2075.  Approximately equal future Clx values
could be achieved if the underlying growth rate is assumed to be the same as
shown in the 2.5% Growth scenario, but global limitations are placed on the
growth of a subset of the chemicals, with 10 percent of the world's producing
countries not complying.  Of course, the real growth rate could be 1.2 percent
or 3.8 percent, which would require lesser or greater stringency and coverage to
achieve equivalent Clx concentrations.

    Exhibit 18-2b provides another example of how various scenarios are
equivalent.  The 1.2% Growth scenario in the risk assessment leads to a Clx
concentration of 9.9 ppbv by 2075.  This would be similar to the 3.8% Growth
scenario if global limitations led to a 50 percent reduction.  That is, if world
growth actually were 3.8 percent, to achieve 1.2 percent would require a 50
percent cutback for CFC-11, -12, -113, and CC14 in nations with 90 percent of
the world's CFG production.

    The scenarios in Exhibit 18-2a and 2b demonstrate that there are a variety
of ways in which the Clx levels associated with individual scenarios can be
achieved.  A global policy of reductions in use for a subset of the compounds,
assuming less than 100 percent participation,  could result in the same Clx
values as the assumed No Growth scenario (which assumes 100 percent compliance
across all the compounds).

    Given these examples, therefore, it is clear that the six "what if"
scenarios used here to assess risks reflect a great diversity of potential
future outcomes.  Furthermore, interpreting real world options does not require
these scenarios to be exactly as simulated vis a vis specific substances since
the same Clx growth paths can be achieved by many difficult scenarios.

    Scenarios of future production and emissions of other CFCs, including
CFC-13, CFC-14, CFC-21, CFC-23, CFC-114, CFC-115, CFC-142b, and CFC-152a are not
presented here because these chemicals are only produced in limited quantities.

    Representatives from DuPont, Allied, and Imperial Chemical Industries have
reported that they had investigated chemical substitutes beginning a decade ago.
HCFC-123 and FC-134a were identified as attractive candidates for replacing

-------
                                      18-15


                                 EXHIBIT  18-2A

       Real World Equivalent to the No Growth Scenario in Risk Assessment
Risk Assessment No Growth Scenario
Equivalent Real World Case
Chemical Coverage:

o CFCs:  11, 12, 113, 22
o Methyl chloroform
o Carbon tetrachloride
o Halons:  1211, 1301
o Other ozone depleters assumed
  to have zero emissions
                 a/
Chemical Coverage

o CFCs:  11, 12, 113
o Carbon tetrachloride
Stringency:

o Freeze at 1985 levels



Compliance:

o 100 percent of global production
Stringency:

o 70 percent reduction from 1985
  levels by 1990


Compliance:

o Nations representing 90 percent
  of global production
Resulting Increases in Clx
  Concentrations (ppbv)
1985
2000
2025
2050
2075
0.0
1.3
2.9
4.1
5.0
Resulting Increases in Clx
  Concentrations (ppbv)
1985
2000
2025
2050
2075
0.0
1.2
2.2
3.7
5.1
a/  Chemicals not covered grow as follows:  HCFC-22:  2.5%/yr through 2050;
    methyl chloroform:  2.5%/yr through 2050; Halon-1211:  2.9%/yr through 2050-
    Halon-1301:  2.5%/yr.

b/  Non-compliers are assumed to grow at 2.5%/yr through 2050 for CFC-11 and
    CFC-12, and 2.8%/yr for CFC-113.

-------
                                     18-16


                                 EXHIBIT 18-2B

     Real World  Equivalent  to the 1.2% Growth  Scenario  in Risk Assessment
Risk Assessment No Growth Scenario
Equivalent Real World Case
Chemical Coverage:

o CFCs:  11, 12, 113, 22
o Methyl chloroform
o Carbon tetrachloride
o Halons:  1211, 1301
o Other ozone depleters assumed
  to have zero emissions
                 a/
Chemical Coverage

o CFCs:  11, 12, 113
o Carbon tetrachloride
Stringency:

o Limited to 1.2% Growth



Compliance:

o 100 percent of global production
Stringency:

o 50 percent reduction from 1985
  levels by 1990


Compliance:

o Nations representing 90 percent
  of global production
Resulting Increases in Clx
  Concentrations (ppby)
1985
2000
2025
2050
2075
0.0
1.5
4.0
6.9
9.9
Resulting Increases in Clx
  Concentrations (ppbv)
1985
2000
2025
2050
2075
0.0
1.4
3.3
6.6
10.3
a/  Chemicals not covered grow as follows:  HCFC-22:  3.8%/yr through 2050;
    methyl chloroform:  3.4%/yr through 2050; Halon-1211:  4.2%/yr through 2050;
    Halon-1301:  3.8%/yr.

b/  Non-compliers are assumed to grow at 3.8%/yr through 2050 for CFC-11 and
    CFG-12, and 4.3%/yr for CFC-113.

-------
                                     18-17
CFC-11 and CFC-12.  HCFC-123 has a hydrogen atom and would probably have a short
atmospheric lifetime.  FC-134a has no chlorine and thus no depletion potential.
HCFC-123 and FC-134a are not now commercially produced.  At the current time it
is believed that producers will not produce these chemicals unless environmental
controls lead to increases in the prices of existing CFCs,  unless there is an
improvement in the technology of producing these chemicals.

    Greenhouse Gas Scenarios

    Chapter 4 summarizes the data available on the potential future levels of
emissions and concentrations of three key greenhouse gases:  C02,  CH4, and N20.
All three gases also influence stratospheric ozone.  Because stratospheric ozone
is most sensitive to the uncertainty surrounding the potential future levels of
CH4 concentrations, many CH4 scenarios are examined, including:

        o   linear growth of 0.017 ppm/yr (about 1.0 percent of
            current concentrations);

        o   compounded annual growth of 1.0 percent per year;

        o   linear growth of 0.01275 ppm/yr (75 percent of the 0.017
            ppm/yr growth);

        o   linear growth of 0.02125 ppm/yr (125 percent of the
            0.017 ppm/yr growth);

        o   compounded annual growth of 1.0 percent per year from
            1985 to 2010, followed by no growth thereafter; and

        o   compounded annual growth of 1.0 percent per year from
            1985 to 2020, growing to 1.5 percent per year by 2050
            and thereafter.

These six scenarios reflect the potential consequences of the diverse set of
factors affecting CH4 concentrations discussed in Chapter 4.

    The levels of C02 and N20 concentrations,  although important for evaluating
stratospheric ozone,  have less of an impact than the CH4 concentrations.
Initial assumptions for these two gases are:

        o   C02:   50th percentile estimate reported by NAS  (1984),
            which implies an annual rate of growth in concentrations
            of about 0.7 percent over the period of the analysis;
            and

        o   N20:   compounded annual growth of 0.2 percent per year.

    Because C02,  N20 and CH4 are radiatively active gases that may contribute to
future global warming,  governments may take action to limit future increases in
their concentrations if they become concerned about global  warming.   The
potential influence of such government actions could be significant over the
time period used to assess risks of stratospheric modification.  Rather than
assume that future decision makers choose to never control  these gases to limit
warming,  a range of scenarios of greenhouse gas growth is presented in order to
explore the eventuality of limiting global warming to various levels.

-------
                                     18-18
SENSITIVITY OF HEALTH AND ENVIRONMENTAL EFFECTS TO DIFFERENCES IN EMISSIONS OF
OZONE DEPLETERS5

    Using the six scenarios presented in the previous section, this section
identifies the potential risks to assumptions regarding the future production,
use, and emissions of ozone-modifying substances.

    Exhibit 18-3 displays estimates of ozone depletion for the six scenarios.
The 80% Reduction scenario shows a slight decline in ozone levels until 2010,
after which ozone levels increase.  The 2.5%, 3.8%, and 5.0% Growth scenarios
result in greatly accelerated ozone depletion.  Ozone depletion reaches nearly
50 percent on or before 2100 in these scenarios.

    For purposes of evaluating potential health effects, the maximum ozone
depletion was (arbitrarily) set to 50 percent.  Depletion exceeding 50 percent
was considered to be outside the range for which the atmospheric models used
here was tested.  This arbitrary cutoff affected ozone depletion estimates in
the 2.5%, 3.8%, and 5.0% Growth scenarios only.  Once 50 percent depletion is
reached, depletion is held constant for the remainder of the analysis period.
(The implications of relaxing this assumption are explored below.)

    Exhibits 18-4 to 18-8 present estimates of the health risks across the
emissions scenarios for the three population cohorts.  All of the estimates
reflect human health impacts due to ozone depletion:

        o   Exhibit 18-4 shows the additional cases of nonmelanoma by
            type of nonmelanoma (basal or squamous).

        o   Exhibit 18-5 shows the additional mortality from nonmelanoma
            by type of nonmelanoma.

        o   Exhibit 18-6 shows the additional cases of melanoma.
      ^ Many of  the estimates of ozone depletion and health risks presented in
 this  section  show increases relative to  the  estimates  in  the previous draft of
 this  document.  Modification to the analyses  that resulted in  increased
 estimates  of  risks include:  (1)  inclusion of the expected increases in the
 average  age of  the U.S. population through the year 2080, rather than assuming
 the population  profile  does not age as expected (as had been assumed for
 purposes of simplification in  the draft);  (2)  use of 0.017 ppm/yr. growth rate
 for methane and 0.20  percent/yr for nitrous  oxide concentrations, as opposed to
 earlier  larger  estimates;  (3)  use of estimates of global  ozone depletion for
 purposes of estimating  risks rather than relying on a  single 2-D model result;
 and  (4)  inclusion of  simulated changes in chlorine concentrations through 2165
 for estimating  ozone  depletion through 2165  (previous  analysis held ozone
 depletion  constant after  2100, implying  constant atmospheric concentrations
 after 2100).

-------
                                     18-19



                                  EXHIBIT 18-3


              Global Average Ozone Depletion:   Emission Scenarios
c
0

o
Q.
4)
Q


-------
                                     18-20
Assumptions:
                                 EXHIBIT 18-4

                 Additional Cases of Nonmelanoma Skin Cancer
                            By type of Nonmelanoma
                                 (Whites only)
Scenario
5.0%
3.8%
2.5%
1.2%
Growth
Basal
Squamous
Total
Growth
Basal
Squamous
Total
Growth
Basal
Squamous
Total
Growth
Basal
Squamous
Total
No Growth
Basal
Squamous
Total
80% Reduction
Basal
Squamous
Total
People
Alive Today
6,882,800
5.431.800
12,314,600
3,481,500
2.587.200
6,068,700
1,289,300
846.500
2,135,800
623,700
390.800
1,014,500
284,200
169.900
454,100
-74,400
-57.300
-131,700
People Born
1986-2029
46,429,300
42.922.400
89,351,700
34,272,500
30.330.000
64,602,500
11,998,800
9.091.300
21,090,100
2,301,600
1.411.900
3,713,500
288,000
150.500
438,500
-1,164,800
-698.100
-1,862,900
People Born
2030-2074
78,600,200
81.244.300
159,844,500
74,802,800
76.314.500
151,117,300
47,744,000
44.078.600
91,822,600
4,191,800
2.564.500
6,756,300
-572,900
-354.400
-927,300
-2,853,400
-1.622.200
-4,475,600
Total
131,912,300
129.598.500
261,510,800
112,556,800
109.231.700
221,788,500
61,032,100
54.016.400
115,048,500
7,117,100
4.367.200
11,484,300
-700
-34.000
-34,700
-4,092,600
-2.377.600
-6,470,200
     Current 1-D models accurately reflect global depletion; Antarctic ozone
     hole has no impact on global ozone levels.
     Greenhouse gases that counter depletion grow at historical rates.
     Growth rates for ozone depleters are for global emissions; it is assumed
     that emissions do not increase after 2050.
     Ozone depletion limited to 50 percent.

-------
                                     18-21
                                  EXHIBIT 18-5

               Additional Mortality From Nonmelanoma Skin Cancer
                            By Type of Nonmelanoma
                                 (Whites only)
Scenario
People
Alive Today
People Born
1986-2029
People Born
2030-2074
Total
5.0% Growth
     Basal
     Squamous
     Total

3.8% Growth
     Basal
     Squamous
     Total

2.5% Growth
     Basal
     Squamous
     Total

1.2% Growth
     Basal
     Squamous
     Total

No Growth
     Basal
     Squamous
     Total

80% Reduction
     Basal
     Squamous
     Total

Assumptions:
 21,300
203.700
225,000
 10,800
 97.000
107,800
  4,000
 31.700
 35,700
  1,900
 14.700
 16,600
   -100
 -2.100
 -2,200
  143,900
1.609.600
1,753,500
  106,200
1.137.400
1,243,600
   37,200
  340.900
  378,100
    7,100
   52.900
   60,000
   -3,500
  •26.100
  -29,600
  243,700
3.046.700
3,290,400
  231,900
2.861.800
3,093,700
  148,000
1.652.900
1,800,900
   13,000
   96.200
  109,200
                                -1,700
                               -13.200
                               -14,900
   -8,700
  -60.700
  -69,400
  408,900
4.860.000
5,268,900
  348,900
4.096.200
4,445,100
  189,200
2.025.500
2,214,700
   22,000
  163.800
  185,800
                                       100
                                    -1.200
                                    -1,100
  -12,300
  -88.900
 •101,200
     Current 1-D models accurately reflect global depletion; Antarctic ozone
     hole has no impact on global ozone levels.
     Greenhouse gases that counter depletion grow at historical rates.
     Growth rates for ozone depleters are for global emissions; it is assumed
     that emissions do not increase after 2050.
     Ozone depletion limited to 50 percent.

-------
                                     18-22
                                 EXHIBIT  18-6

              Additional Cases of Melanoma  Skin Cancer By Cohort
                                 (Whites only)
Scenario
5.0% Growth
3.8% Growth
2.5% Growth
1.2% Growth
No Growth
80% Reduction
People
Alive Today
62,500
33,700
13,900
7,100
3,500
-300
People Born
1986-2029
445,900
324,400
124,400
29,800
5,200
-14,300
People Born
2030-2074
771,500
734,300
470,400
56,800
-6,400
-39,800
Total
1,279,900
1,092,400
608,700
93,700
2,300
-54,400
Assumptions:
     Current 1-D models accurately reflect global depletion; Antarctic ozone
     hole has no impact on global ozone levels.

     Greenhouse gases that counter depletion grow at historical rates.

     Growth rates for ozone depleters are for global emissions; it is assumed
     that emissions do not increase after 2050.
     Ozone depletion limited to 50 percent.

-------
                                     18-23
        o   Exhibit 18-7 shows the additional mortality from melanoma.

        o   Exhibit 18-8 shows the additional senile cataract cases.

The exhibits provide a range of risk estimates and illustrate how risk estimates
respond to assumptions about emissions.

    The 80% Reduction and No Growth scenarios result in the fewest additional
cumulative cases and deaths related to nonmelanoma skin cancer, melanoma skin
cancer, and senile cataract.  The reduction in risk results from lower ozone
depletion (and UV radiation) in these scenarios.  For example, in the 80%
Reduction scenario in Exhibit 18-4, it is estimated that people alive today
would actually experience fewer cases of basal cell skin cancer than in the
absence of ozone depletion due to an increase in ozone abundance.

    The 3.8% and 5.0% Growth scenarios produce the largest increases in cases
and deaths.  The large increases in cases and deaths in the higher emission
scenarios can lead to estimates of total cases that exceed the size of the
population.  For example, in Exhibit 18-4 the total number of nonmelanoma cases
in the 3.8% and 5.0% Growth scenarios among people born from 2030 to 2074
exceeds the total size of the population.  This level of increase in risk is
possible since many people often develop two or more cases of nonmelanoma in
their lives, while other people may not develop nonmelanoma at all.  In general,
however, the risk estimates for the higher emission scenarios should be treated
with caution because the dose-response coefficients used to derive risk
estimates were estimated over a limited range of UV radiation levels.  The
dose-response parameters may not be valid for the extreme changes in UV
radiation expected in these scenarios (it is possible they overestimate or
underestimate total changes).

    The risk estimates shown in Exhibit 18-4 to 18-8 are incremental to the
risks that can be expected in the absence of depletion.  If current levels of
skin cancer and cataracts are extrapolated into the future (without changing
their baseline incidence and mortality rates), there would be 160 million
nonmelanoma skin cancer cases for the three cohorts, 1.77 million deaths from
nonmelanoma, 4.2 million melanoma cases, 1.2 million melanoma deaths, and 182
million cataract cases.

    Exhibit 18-9 displays estimates of environmental risks for those effects
that could be estimated quantitatively.  Exhibit 18-10 displays estimates for
environmental effects that were estimated based on other case studies or
research in its early stages (these effects are more difficult to quantify).   As
with the health risks, the range of environmental effects is very large.  Of
note is that the global warming estimates for 2075 range from 4.0°C in the 80%
Reduction scenario to 11.6°C in the 5.0% Growth scenario.  Exhibit 18-11 shows
the greenhouse equilibrium temperature increase over time for the six scenarios.

SENSITIVITY OF RESULTS TO ALTERNATIVE ATMOSPHERIC ASSUMPTIONS

    This section identifies the sensitivity of potential risks to assumptions
regarding:  (1) the type of model used to characterize ozone depletion; (2) the
sensitivity of the global climate to greenhouse gas forcings; (3) the
concentration of greenhouse gas emissions; (4) ozone level response to

-------
                                     18-24
                                 EXHIBIT  18-7

           Additional Mortality From Melanoma  Skin Cancer By Cohort
                                 (Whites only)
Scenario
5.0% Growth
3.8% Growth
2.5% Growth
1.2% Growth
No Growth
80% Reduction
People
Alive Today
17,800
9,500
3,800
1,900
900
-100
People Born
1986-2029
107,900
81,900
32,500
7,300
1,100
-3,600
People Born
2030-2074
170,300
163,500
111,900
13,400
-1,700
-9,600
Total
296,000
254,900
148,200
22,600
300
-13,300
Assumptions:
     Current 1-D models accurately reflect global depletion; Antarctic ozone
     hole has no impact on global ozone levels.

     Greenhouse gases that counter depletion grow at historical rates.

     Growth rates for ozone depleters are for global emissions; it is assumed
     that emissions do not increase after 2050.

     Ozone depletion limited to 50 percent.

-------
                                       18-25


                                    EXHIBIT 18-8

                     Additional Senile Cataract Cases By Cohort
Scenario
5.0% Growth
3.8% Growth
2.5% Growth
1.2% Growth
No Growth
80% Reduction
People
Alive Today
3,238,700
1,829,400
613,900
267,400
102,800
9,500
People Born
1986-2029
10,361,300
8,928,100
4,311,900
752,000
49,100
0
People Born
2030-2074
12,405,300
12,333,500
10,214,500
1,157,600
0
0
Total
26,005,300
23,091,000
15,140,300
217,700
151,900
9,500
Assumptions:
       Current 1-D models accurately reflect global depletion; Antarctic ozone hole
       has no impact on global ozone levels.

       Greenhouse gases that counter depletion grow at historical rates.

       Growth rates for ozone depleters are for global emissions; it is assumed that
       emissions do not increase after 2050.

       Ozone depletion limited to 50 percent.

-------
                                     18-26


                                 EXHIBIT  18-9

          Environmental Effects Estimated Quantitatively for the U.S.
Scenario


5.0% Growth

3.8% Growth

2.5% Growth

1.2% Growth

No Growth

80% Reduction
Materials Damage5
   (106 1985 $)
                                            Rise In
                                          Equilibrium
                                         Temperature by
                                             2075^
                                                      Sea Level
                                                     Rise by 2075
                                                         (cm)
971
829
606
305
272
131
11.6
7.6
5.6
4.7
4.3
4.0
134.3
108.8
95.1
88.6
85.3
82.2
b/
Present value, discounted over 1985-2075 using a real discount rate of 3
percent.

Assuming a climate sensitivity of 3°C for doubled CC^.  A range of climate
sensitivity from 1.5 to 4.5°C could be used based on NAS (1979).  Recent
climate model developments indicate that 4.0°C may be the most likely
estimate.  Estimates of equilibrium warming that would occur if one chose
1.5°C and 4.5°C can be made by multiplying the values shown by 50 percent
and 150 percent, respectively.
Assumptions:
     Current 1-D models accurately reflect global depletion; Antarctic ozone
     hole has no impact on global ozone levels.

     Greenhouse gases that counter depletion grow at historical rates.

     Growth rates for ozone depleters are for global emissions; it is assumed
     that emissions do not increase after 2050.
     Ozone depletion limited to 50 percent.

-------
                                     18-27
                                 EXHIBIT 18-10

                Environmental Effects Based On Case Studies and
                           Research In Early Stages
               Cost of Sea Level   Reduction for  Increase In
                                                             Loss of
                                                             Northern
                                                             Anchovy
Scenario
5.0% Growth
3.8% Growth
2.5% Growth
1.2% Growth
No Growth
80% Reduction
          Rise In Charleston  Soybean, Yield  Ground-Based   . Population
           and Galveston3/    by 2075^      Ozone by 2075^ By 2075s/
             (106 1985 $)
1
1
1
1
1
1
,403-
,221-
,123-
,076-
,053-
,031-
3
3
2
2
2
2
,507
,013
,748
,622
,558
,498
>7.5
>7.5
>7.5
1.4
e/
e/
>9
>9
7
1


.4
.4
.0
.2


- >50.0
- >50.0
-36.9
- 6.5
e/
e/
>11 0 -
>11.0 -
>11.0 -
0.0
e/
e/
>25.
>25.
>25.
- 3.3


0
0
0



b/
e/
Lowest estimate based on anticipation of sea level rise; highest estimate
without anticipation.  Present value discounted using a real discount rate
of 3 percent.

Based on statistically significant results developed by Teramura (1986,
1987) for two soybean cultivars -- Essex and Williams --as reported in
Rowe, et al. (1987).  Maximum value reported is 7.5 percent.

Lowest estimate is for Los Angeles, California; (9.4 percent is maximum
value reported) highest estimate is for Nashville, Tennessee (50.0 percent
is maximum value reported).

Lowest estimate assumes 15-meter vertical mixing of the top ocean layer (11
percent is maximum value reported); highest estimate assumes 10-meter
vertical mixing if ozone depletion is above 11.2 percent, no vertical
mixing if below 11.2 percent (25 percent is maximum value reported).

Impacts were not evaluated because ozone levels are estimated to increase,
not decrease, by 2075 in these scenarios.
Assumptions:
     Current 1-D models accurately reflect global depletion; Antarctic ozone
     hole has no impact on global ozone levels.
     Greenhouse gases that counter depletion grow at historical rates.
     Growth rates for ozone depleters are for global emissions; it is assumed
     that emissions do not increase after 2050.
     Ozone depletion limited to 50 percent.

-------
                                     18-28
                                 EXHIBIT 18-11
         Equilibriun Temperature Change for the Six Emission Scenarios
                    Assuming 3.0°C Warming for Doubled C02*
    12.0

i-
                                                                         5.0% Growth
                                                                         3.8% Growth
                                                                         2.5% Growth

                                                                         1.2% Growth
                                                                         No Growth
                                                                         80% Reduction
      1985
             1995
                    2005
                           2015
                                  2025
                                         2035
                                                 2045
                                                        2055
                                                               2065
                                                                      2075
     * Computed assuming that  the climate  sensitivity to a doubling of carbon
dioxide is 3°C.  This assumption is  in  the middle  of the NAS  range of 1.5°C to
4.5°C (see Chapter 6).  Note that the actual warming that may be realized will
lag by several decades or more.  To  compute the  equilibrium warming associated
with high or low NAS estimates multiply the y  axis "temperature change" by 1.5
or 0.5.

     Growth levels refer to global estimates of  production of all ozone
depleters.

-------
                                     18-29
emissions; and (5) the maximum level of ozone depletion simulated to occur.  For
purposes of comparison, the results of these sensitivity analyses are shown for
the 2.5% Growth scenario only (unless indicated otherwise).  Information is
presented for one scenario to avoid complicating the presentation of results
unnecessarily.  The selection of the 2.5% Growth scenario case should not be
interpreted as other than a presentational device;  comparison could have been
made with any of the six scenarios to demonstrate the impact of the alternative
assumptions.

Comparison with Results from a 2-Dimensional Atmospheric Model

     A variety of models have been developed to assess potential changes in the
stratosphere due to emissions of CFCs and other ozone-modifying substances (see
Chapter 5).  The various scenario risk estimates rely on a parameterized
representation of a one-dimensional model of the atmosphere (see Chapter 17).
For comparison purposes, the results of a time-dependent two-dimensional model
developed by Isaksen (1986) are presented here.  Two-dimensional models are
believed to be preferable for risk assessment (see Chapter 5).

     The results of Isaksen's analysis are currently only available beyond 2030.
Therefore, a complete comparison of risks is not possible at this time (as
additional data become available, a full comparison of risks will be performed).

    Exhibit 18-12 displays estimates of ozone depletion from the one- and
two-dimensional atmospheric models for two scenarios of emissions.  The first
scenario is a 3.0 percent growth rate of CFCs-11, 12, and 22, methyl chloroform,
and carbon tetrachloride.  This scenario is not directly comparable to any of
the scenarios described above.  The exhibit shows that the two-dimensional model
results in larger estimates of ozone depletion than does the parameterized
representation of the one-dimensional model for this scenario.

    The second scenario shown in the exhibit is a 0.0 percent per year growth
scenario.  Comparing the one-and two-dimensional model estimates, the results
from Isaksen's two-dimensional model show more depletion.  In general,
two-dimensional models appear to produce larger estimates of ozone depletion
than do one-dimensional models (see Chapter 5).

Alternative Assumptions about: Sensitivity of Atmosphere to Greenhouse Gases

    The 2.5% Growth scenario (as well as the other five scenarios) assumes that
the sensitivity of the global climate system is such that a doubling of
atmospheric C02 would result in a 3°C increase in the average global
temperature.   This climate sensitivity is the middle of a range published by the
NAS in 1983 of 1.5°C to 4.5°C.  Exhibit 18-13 displays estimates of
temperature-sensitive effects for the limits of this NAS range.  As expected,
the warming and sea level rise estimates are lower for the 1.5°C sensitivity and
higher for the 4.5°C sensitivity.

-------
                                           18-30
                                      EXHIBIT 18-12



                             Global Average Ozone Depletion:

              Comparison to Results with a 2-Dimensional Atmospheric Model
Q
a.
UJ
a
0
N

0
       -5 -
       -6
                                                             2-0  Constant Emissions  Growth
                                      1-D Constant Emissions Growth
                                               2-D 3 Percent  Emissions Growth
                                                    L-D 3  Percent

                                                   Emissions Growth
         1985
1995
                2005
                                                         2015
                                                2025

-------
                                                             EXHIBIT 18-13

                                                      Cliaate and Other Effects:
                               Sensitivity to Relationship Between Climate Change and 002 Concentrations
                               (Figures in Parentheses are Percentage Changes frcn 2.5Z Growth Scenario)
Type of Effect
Effects Estimated Quantitatively
for the U.S.
Rise in Equilibrium
Temperature by 207S
Sea Level Rise by 2075
Effects Based on Case Studies
and Research In Early Stages
DOUBLED C02 CLIMATE SENSITIVITY
1.5°C 3°C 4°C 4.5°C UNITS
2.7 5.6 7.6 8.7 Degrees Centigrade
(-52) (36) (55)
76 95 104 107 Centimeters
(-20) (9) (13)
                                                                                                                                                     00
Cost of Sea Level Rise in
Charleston and Galveston a/
                                             977-2,188
1,123-2,748
1,177-2,897  1,199-2,955
Present Value
(millions of 1985 dollars)
a/ Lowest estimate with anticipation of sea level rise;  highest estimate without anticipation.

-------
                                     18-32


    Also shown in the exhibit are estimates for a 4°C sensitivity.  Recent model
runs from the major general circulation models in the U.S. produced a warming of
4°C for a doubling of C02 (see Chapter 6).

Alternative Assumptions about Emissions of Trace Gases

    Uncertainty exists about potential future changes in atmospheric
concentrations of the greenhouse gases:  carbon dioxide (C02);  methane (CH4);
and nitrous oxide (N20).   This uncertainty arises from at least two sources.
First, to limit the potential for future global warming, governments may agree
to take steps to limit growth in the concentrations of these gases.  However,
the timing and stringencies of such potential steps are not known.  Second,
future concentrations of methane are influenced by numerous natural as well as
anthropogenic factors that have not been well characterized to date (see Chapter
4).  Therefore, the rate of future increases in methane concentrations is
particularly uncertain even in the absence of potential government regulations
to limit global warming.

    Through their roles in chemical reactions in the stratosphere, trace gases
affect both total column ozone and the potential for future global warming.
Exhibit 18-14 summarizes the implications of changes in concentrations of the
gases for ozone depletion and global warming.  In general, each greenhouse  gas
increases global temperature but counters ozone depletion.

    Any risk assessment requires an assumption about future global warming  and
whether governments take actions to limit warming.  The "standard assumption"
generally has been that warming is never limited.  To examine the sensitivity of
the risk estimates to this issue, the implications of reducing the growth in
concentrations of the trace gases to limit warming was examined for all six
scenarios for two different limits: (1) 2°C by 2075 and (2) 3°C by 2075.  In
some scenarios, it may not be possible to limit warming to these levels by
reducing the growth in concentrations of trace gases because even if no growth
in these gases occurs, the limit will still be exceeded due to emissions from
other CFCs.  In fact, the only scenarios for which a limit of 2°C can be
examined are the 80% Reduction and No Growth scenarios, since the 2°C limit
cannot be achieved because of CFG growth in the other scenarios even if there is
no growth in the concentrations of other trace gases; with a limit of 3°C,  only
the 80% Reduction, No Growth, 1.2% Growth, and 2.5% Growth scenarios can be
estimated.

    Exhibits 18-15 to 18-19 compare the results from the 80% Reduction, No
Growth, 1.2% Growth and 2.5% Growth scenarios without limiting warming to the
situation in which global warming is limited to 2°C and 3°C for those scenarios.
As shown in Exhibit 18-15, the increase in ozone depletion can be significant;
e.g., by 2050 in the 2.5% Growth scenario with a 3°C limit the difference is
about 4 percent depletion.  Exhibits 18-16 to 18-18 show the increased health
risks expected for the three population cohorts.  In general, the latter cohorts
are affected more.  For example, in the 2.5% Growth scenario, the people alive
today may experience an increase in health risks on the order of 40 percent
relative to the No Limits value.  People born from 2030 to 2074 experience  an
increase in health risks up to 90 percent.

-------
                                18-33
                            EXHIBIT 18-14

      Summary of Effects of Greenhouse Gases on Ozone Depletion
                  and Global  Equilibrium Temperature
   GREENHOUSE GAS         STRATOSPHERIC OZONE       GLOBAL TEMPERATURE
Methane (CH4)           Counters Depletion              Increases

Nitrous Oxide (N20)     Counters Depletion              Increases
                        in High Chlorine Cases

Carbon Dioxide (C02)    Adds Ozone                      Increases

-------
                                      18-34
                                  EXHIBIT 18-15

              Global Average Ozone Depletion:   Scenarios of Limits
                            to Future Global Warming
                                                               80% Reduction
                                                               3°C
                                                               2°C
                                       2045
                                              2065
Assumptions:
                                                              No Growth
                                                             t, 3°C
                                                              2°C
     Current 1-D models  accurately reflect global depletion; Antarctic ozone
     hole has no impact  on global ozone levels.

     Greenhouse gases other than ozone depleters are limited sufficiently so
     that total equilibrium warming does not exceed 2°C or 3°C in 2075.

     A doubling of C02 would ultimately warm the earth 3°C.  (Thus, if the
     sensitivity were 4°C,  the  two limits would be 4°C and 2.67°C,
     respectively.)

-------
                                      18-35
                                  EXHIBIT 18-15
                                   (Continued)

              Global Average Ozone Depletion:  Scenarios  of Limits
                            to Future Global Warning
Assumptions:
           S   -200 -
           O   -300 -
                                                                1.2% Growth
                                                              ... 3«C
                                                                2.5% Growth
     Current 1-D models accurately reflect  global  depletion;  Antarctic ozone
     hole has no impact on global ozone  levels.

     Greenhouse gases other than ozone depleters are  limited sufficiently so
     that total equilibrium warming does not  exceed 2°C or 3°C in 2075.

     A doubling of C02 would ultimately warm  the earth 3°C.   (Thus,  if the
     sensitivity were 4°C, the two limits would be 4°C and 2.67°C,
     respectively.)

-------
                                                                        EXHIBIT 18-16

                                              Hunan Health Effects:  Scenarios of Limits to Future Global Harming
                                          Additional emulative Cases and Deaths Over Lifetimes of People Alive Today
                                              (Figures in Parentheses are Percentage Changes from Ho Limit Case)
Nonmelanoma Skin Tumors
Additional Basal Cases
Additional Squamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
a/ Percentage changes are not
Assumptions:

2°C

166,800
a/
96,100
a/
4,100
a/

2,200
a/
500
a/

51,400
(441)
computed

80% Reduction
3"C

47,400
a/
20 , 100
a/
900
a/

900
a/
200
a/

17,900
(88)
from negative


No Limit 2°C

-74,400 607,000
(114)
-57,300 379,300
(123)
-2,200 16,100
(121)

-300 6,900
(97)
-100 1,800
(100)

9,500 257,100
(150)
to positive values.

No Growth
3°C No Limit

471,400 284,200
(66)
291,100 169,900
(71)
12,400 7,300
(70)

5,500 3,500
(57)
1,500 900
(67)

192,800 102,800
(88)

1.2%
3CC

883,500
(42)
561,700
(44)
23,800
(43)

9,800
(38)
2,700
(42)

391,400
(46)

Growth 2.5%
No Limit 3'C
623,700 1,925,600
(42)
390,800 1,224,000
(45)
16,600 51,600
(44)

7,100 19,200
(38)
1,900 5,300
(39)

267,400 878,700
(43)

Growth
No Limit
1,289,300
846,500
35,700

13,900
3,800

613,900

                                                                                                                                                            00
Temperature limits shown assumed that a doubling of C02 ultimately warms  earth  3°C; numbers  could be  larger or smaller  if doubling warms  earth  1.5°C or  4.5°C.

Assumes 1-D models results are accurate; Antarctic ozone hole has no relevance  to  global  ozone.

Emissions growth estimates refer to 1985 to 2050;  held constant  thereafter.

-------
                                                                       DCHTBIT  18-17

                                             Hman Health Effects:  Scenarios of Units to Future Global Harming
                                       Additional Cnmolative Cases and Deaths Over Lifetimes of People Bom 1986-2029
                                             (Figures in Parentheses are Percentage Changes from Ho Limit Case)
Honmelanoma Skin Tumors
Additional Basal Cases
Additional Squamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
a/ Percentage changes are
Assumptions:

2°C

-6,300
(99)
-15,500
(98)
-500
(98)

800
a/
100
«/

6,000
not computed

SOX Reduction
3°C

-570,500 -
(51)
-350,600
(50)
-14,800
(50)

-6.500
(54)
-1,700
(53)

0
from negative


No Limit 2'C

1,164,800 1,949,300
(577)
-698,100 1,175,900
(681)
-29,600 50,100
(671)

-14,300 26,100
(402)
-3,600 6,300
(473)

0 604,100
(1,130)
to positive values.

No Growth
3°C No Limit

1,260,700 288,000
(338)
746,500 150,500
(396)
31,900 6,500
(391)

17,500 5,200
(237)
4,200 1,100
(282)

371,300 49,100
(656)

1.2% Growth
3°C No Limit

3,840,200 2,301,600
(67)
2,421,500 1,411,900
(72)
102,700 60,000
(71)

48,100 29,800
(61)
11,900 7,300
(63)

1,277,700 752,000
(70)

2.5X
3°C

19,746,400
(65)
16,527,400
(82)
681,000
(80)

191,200
(54)
49,800
(53)

6,476,200
(50)

Growth
No Limit
11,998,800
9,091,300
378,100

124,400
32,500

4,311,900

Temperature limits shown assumed that a doubling  of C02 ultimately warms  earth  3°C; numbers  could be larger  or  smaller  if doubling warms  earth  1.5 C or 4.5°C.

Assumes 1-D models results are accurate;  Antarctic  ozone hole has no  relevance  to  global ozone.

Emissions growth estimates refer to 1985 to 2050; held constant  thereafter.

-------
                                                                            EXHIBIT  18-18

                                                 Human Health Effects:  Scenarios of Limits to Future Global Warming
                                            Additional emulative Cases and Deaths Orer Lifetimes of People Born 2030-2074
                                                  (Figures in Parentheses are Percentage Changes from Ho Limit Case)
 Nonmelanoma Skin Tumors

 Additional Basal Cases


 Additional Squamous Cases


 Additional Deaths


 Melanoma Skin Tumors

 Additional Cases


 Additional Deaths


 Senile Cataract

Additional Cases



 a/ No percentage estimated.


Assumptions:

- Temperature limits shown assumed that a doubling of C02 ultimately warms  earth  3°C; numbers  could be  larger  or  smaller  if doubling warms  earth 1.5°C or 4.5°C.

- Assumes 1-D models results are accurate;  Antarctic  ozone hole has no relevance  to  global  ozone.

- Emissions  growth estimates refer to 1985  to 2050; held constant thereafter.
80%
2°C
-514,300 -
(82)
-305,800
(81)
-12,900
(81)
-6,600
(83)
-1,600
(83)
0
(0)
Reduction
3"C No Limit
1,637,500 -2,853,400
(43)
-947,700 -1,622,200
(42)
-40,400 -69,400
(42)
-22,400 -39,800
(44)
-5,400 -9,600
(44)
0 0
(0)

2"C
2,949,200
(515)
1,772,400
(500)
75,600
(507)
41,000
(641)
9,600
(565)
799,400
a/
No Growth
3°C No Limit
1,498,100 -572,900
(261)
876,600 -354,400
(247)
37,500 -14,900
(252)
21,800 -6,400
(341)
5,000 -1,700
(294)
379,900 0
a/
1.2% Growth
3°C No Limit
8,243,500 4,191,800
(97)
5,326,400 2,564,500
(108)
225,300 109,200
(106)
106,100 56,800
(87)
25,300 13,400
(89)
2,297,300 1,157,600
(98)
2.5%
3°C
76,861,100
(61)
84,308,400
(91)
3,399,900
(89)
685,600
(46)
159,600
(43)
13,901,600
(36)
Growth
No Limit
47,744,000

44,078,600

1,800,900

470,400

111,900

10,214,500

00

-------
    Type of Effect

Effects Estimated
Quantitatively for the U.S.

Materials Damage a/
Rise in Equilibrium
Temperature by 2075

Sea Level Rise by 2075

Effects Based on Case
Studies and Research in
Early Stages
                                                                            EXHIBIT 18-19

                                         Materials, Climate,  and Other Effects:   Scenarios  of Limits  to Future Global Harming
                                                  (Figures  in Parentheses are Percentage Changes fron Mo  Limit Case)
                                   80% Reduction
                                2'C
228
(74)

2.0
67
         3°C   Mo Limit
163
(24)

3.0
         75
                131
                                              4.0
                 82
                                                               No Growth
301
(11)

2.0
                            67
272s7  272
(0)
                                                                  3.0
                                    75
                                           4.3
                                            85
                                              1.2% Growth
                                                    No Limit
397
(30)

3.0
                                                       75
                                                                                             305
                                                               4.7
                                                                89
                                                                                                         2.5Z Growth
                                             3°C
739
(22)

3.0
                                                                          75
                                                                                Ho Limit
                                                                                                               606
                                                                                 5.6
                                                                                  95
Present Value
(millions of 1985 dollars)

Degrees Centigrade
                                                                                                Centimeters
Cost of Sea Level Rise in 913-
Charleston and Galveston b/ 2,180
Reduction in Soybean Seed f/
Yield c/
Increase in Ground-Based f/
Ozone d/
Loss of Northern Anchovy f/
Population £/
971- 1,031-
2,335 2,498
f/ f/

£/ £/

f/ f/

913-
2
1

0

0

,180
.0

.9-4.

.0-2.

971-
2,
0.

6 0.

4 0.

335
5

5-2.

0-1.

1,053-
2,558
f/

.6 f/

.3 f/

971-
2,335
2.3

2.1-11.1

0-5.9

1
2
1

1

0

,076-
,622
.4

.2-6.5

.0-3.3

971-
2,335
>7.5

>9.4->50.

>11.0->2i

1,123-
2,748
>7.5

,0 7.0-36.9

5.0 >11.0->:

                                                                                                                              Present Value
                                                                                                                              (millions of 1985 dollars)

                                                                                                                              Percent in Year 2075
                                                                                                                              Percent in Year 2075
                                                                                                                                                                 00
                                                                                                                                                                 i

                                                                                                                                                                 vo
a/ Discounted over 1985-2075 using a real discount rate of 3 percent.
b/ Lowest estimate with anticipation of sea level rise; highest estimate without anticipation.
cj Essex and Williams cultivars; statistically-significant results only.
d/ Lowest estimate is for Los Angeles, California;  highest estimate is for Nashville,  Tennessee.
el Lowest estimate 15-meter vertical mixing of the top ocean layer; highest estimate 10-meter vertical mixing if ozone depletion is greater than 11.2 percent, while
  ozone depletion less than 11.2 percent assumes no vertical mixing.
f/ Impacts were not evaluated because ozone levels are estimated to increase,  not decrease,  by  2075 in these scenarios.
£/ Value of materials damage for this case does not differ from the No Limit case because the difference in ozone depletion between the two cases is too small, i.e.,
   the damage function from which the damage estimates are computed is a step function that  defines materials damage for various ranges of ozone depletion.

As sumptions:
- Temperature limits shown assumed that a doubling of C02 ultimately warms earth 3°C;  numbers could be larger or smaller if doubling warms earth 1.5°C or 4.5°C.
- Assumes 1-D models results are accurate; Antarctic ozone hole has no relevance to global ozone.
- Emissions growth estimates refer to 1985 to 2050;  held constant thereafter.

-------
                                     18-40
    Exhibit 18-19 displays environmental risks.  As shown in the exhibit,
warming and its impacts are reduced from 5.6°C in the 2.5% Growth scenario to
3°C.  As a result, sea level rise impacts are also reduced.  The ozone-sensitive
risks are increased, however.   If global warming were held to less than 3°C by
2075 the risks from ozone depletion would be even greater.

    A second sensitivity case is examined in which several alternative scenarios
of methane concentrations are examined.  To evaluate the sensitivity of the
results to the methane growth rates,  five methane scenarios were examined in
addition to the assumption of 0.017 ppm/yr used above.  Other trace gas
assumptions were not changed.   The six methane scenarios examined are:

        o   Scenario A - - compound annual growth of 1 percent from
            1985 to 2010, followed by constant concentrations at
            2.23 ppm;

        o   Scenario B -- linear growth at 0.01275 ppm per year (75
            percent of the 0.017 ppm/yr growth);

        o   Scenario C -- linear growth at 0.017 ppm per year;

        o   Scenario D -- linear growth at 0.02125 ppm per year (125
            percent of the 0.017 ppm/yr growth);

        o   Scenario E -- compound annual growth of 1 percent; and

        o   Scenario F -- compound annual growth of 1 percent from 1985
            to 2020, growing to 1.5 percent compound annual growth by
            2050 and thereafter.

    Exhibit 18-20 shows that higher concentrations of atmospheric methane would
reduce ozone depletion, while lower concentrations result in more depletion.
Contrasting the highest methane scenario (Scenario F) to the assumption used
above (0.017 ppm/yr - Scenario C), ozone depletion is roughly equal until 2025,
when the contributions to ozone from increased methane concentrations begin to
dampen depletion.  The influence of increased methane concentrations becomes
more pronounced after 2025, and depletion reaches an inflection point in 2085,
after which ozone levels rise.  The lower methane scenarios (i.e., Scenarios A
and B) cause the ozone depletion estimates to increase compared to Scenario C.

    The implications of the resulting differences in ozone depletion for the
risk estimates from the six methane scenarios are shown in Exhibits 18-21
through 18-24.  Of note in Exhibit 18-24 is that the lower methane scenarios
reduce the estimate of global warming, while the higher scenarios increase it.

Effects at High and Low Sensitivities of Ozone Depletion to Emissions

    A previous section analyzed the sensitivity of risk estimates to assumptions
about the levels of emissions of ozone-modifying substances.  This section
explores another source of uncertainty in the risk estimates:  the relationship
between emissions (and their resulting chlorine concentrations) and ozone
depletion.  This uncertainty was quantified using Monte Carlo modeling
techniques to reflect the implications of the convolution of the  individual

-------
                                       18-41


                                   EXHIBIT 18-20

                Global Average Ozone Depletion:   Methane Scenarios
c
o
'•M
.!£
a

O
O
    -10.0 -
-20.0 -
                                                                         j' Scenario F
    -30.0 -
    -40.0 -
                                                                       Scenario E
    -50.0
        1985
                   2005
                              2025
                                          2045
                                                     2065
                                                                2085
                                                                            Scenario  D
                                                                            Scenario  C
                                                                         Scenario  B
  Scenario A  -   Compound annual growth of 1 percent from 1985 to 2010,  followed
                •  by constant concentrations at 2.23 ppm.

  Scenario B  =   Linear growth at 0.01275 ppm per year.

  Scenario C  -   Linear growth at 0.017 ppm per year.

  Scenario D  -   Linear growth at 0.02125 ppm per year.

  Scenario E  =   Compound annual growth of 1 percent per year.

  Scenario F  =   Compound annual growth of 1 percent from 1985 to 2020,  growing
                  to 1.5 percent compound annual growth by 2050 and thereafter.

  Assumptions:

      Current 1-D  models accurately reflect global depletion; Antarctic ozone hole
      has no impact on global ozone levels.

      Greenhouse gases that counter depletion (other than CH4) grow at historical
      rates.

      Growth rates for ozone depleters are for global emissions; it is assumed
      that emissions do not increase after 2050.
      Ozone depletion limited to 50 percent.

-------
                                                  EXHIBIT 18-21

                                     Hunan Health Effects:   Methane Scenarios
                   Additional emulative Cases and Deaths Over Lifetimes of People Alive Today
                         (Figures in Parentheses are Percentage Changes from Scenario O*
METHANE SCENARIOS
HEALTH EFFECT Scenario A Scenario B
Nonmelanoma Skin Tumors
Additional Basal Cases 1,434,100 1,455,300

Additional Squamous
Cases
Additional Deaths

Melanoma Skin Tumors
Additional Cases

Additional Deaths

Senile Cataract
Additional Cases

(11) (13)
953,200 956,100
(13) (13)
40,100 40,400
(12) (13)

15,200 15,700
(9) (13)
4,200 4,300
(11) (13)

698,300 685,300
(14) (12)
* All estimates from 2. 52 Growth Scenario.
Scenario A = Compound annual growth of 1 percent from
ppm.
Scenario B = Linear growth at
Scenario C = Linear growth at
Scenario D = Linear growth at
Scenario E = Compound annual
Scenario F = Compound annual

0.01275 ppm per year.
0.017 ppm per year.
0.02125 ppm per year.
growth rate of 1 percent
growth of 1 percent from
Scenario C Scenario D Scenario E Scenario F
1,289,300 1,122,200
(-13)
846,500 736,600
(-13)
35,700 31,100
(-13)

13,900 12,000
(-14)
3,800 3,300
(-13)

613,900 541,500
(-12)
1,104,200 1,
(-14)
719,100
(-15)
30,400
(-15)

12,000
(-14)
3,300
(-13)

521,400
(-15)
1985 to 2010, followed by constant concentrations at




per year.
1985 to 2020, growing to 1.5





percent compound annual
071,700
(-17)
692,200
(-18)
29,300
(-18)

11,700
(-16)
3,200
(-16)

494,700
(-19)
2.23





growth
             by 2050 and thereafter.

Assumptions:

- Current 1-D models accurately reflect global depletion;  Antarctic ozone hole has no impact on global ozone
  levels.
- Greenhouse gases (other than CH4) that counter depletion grow at historical rates.
- Growth rates for ozone depleters are for global emissions;  it is assumed that emissions do not increase after
  2050.
- Ozone depletion limited to 50 percent.
                                                                                                                                             00
                                                                                                                                             i-
                                                                                                                                             to

-------
                                                  EXHIBIT 18-22
                                          Health Effects:  Methane Scenarios
                  Additional Cmlative Cases and Deaths Over Lifetimes of People Born 1986-2029
                         (Figures in Parentheses are Percentage Changes from Scenario O*
METHANE SCENARIOS
HEALTH EFFECT
Scenario A
Scenario B
Scenario C
Scenario D
Scenario £ Scenario F
Nonmelanoma Skin Tumors
Additional
Additional
Cases
Additional
Melanoma Si
Additional
Additional
Basal Cases
(21)
Squamous
(25)
Deaths
(24)
tin Tumors
Cases
(18)
Deaths
(18)
14,520,400
(11)
11,352,200
(12)
470,700
(12)
147,000
(10)
38,400
(10)
13,277,100
10,210,300
424,100
136,500
35,600
11,998,800
(-11)
9,091,300
(-13)
378,100
(-13)
124,400
(-11)
32,500
(-10)
10,619,500
(-29)
7,896,700
(-34)
329,000
(-33)
111,300
(-25)
29,100
(-25)
8,539,000
(-48)
6,021,700
(-55)
252,300
(-55)
93,800
(-41)
24,300
(-43)
6,186,600
4,075,100
172,000
72,900
18,500
Senile Cataract
Additional
Cases
(17)
5,044,100
(9)
4,678,800
4,311,900
(-10)
3,887,600
(-28)
3,085,700
(-51)
2,098,100
* All estimates for 2. 51 Growth Scenario.
Scenario A
Scenario B
Scenario C
Scenario D
Scenario E
Scenario F
= Compound annual growth of 1 percent from
ppm.
= Linear growth at 0.01275 ppm per year.
= Linear growth at 0.017 ppm per year.
= Linear growth at 0.02125 ppm per year.
= Compound annual growth rate of 1 percent
= Compound annual growth of 1 percent from
1985 to 2010,
per year.
1985 to 2020,
followed by constant concentrations at 2.23
growing to 1.5 percent compound annual growth
             by 2050 and thereafter.

Assumptions:

- Current 1-D models accurately reflect global depletion;  Antarctic ozone hole has no impact on global ozone
  levels.
- Greenhouse gases (other than CH4) that counter depletion grow at historical rates.
- Growth rates for ozone depleters are for global emissions; it is assumed that emissions do not increase after
  2050.
- Ozone depletion limited to 50 percent.
                                                                                                                                              oo

-------
                                                  EXHIBIT 18-23

                                     Hunan Health Effects:   Methane Scenarios
                  Additional Cumulative Cases and Deaths Over Lifetimes of People Born 2030-2074
                         (Figures in Parentheses are Percentage Changes from Scenario C)*
METHANE SCENARIOS
HEALTH EFFECT
Scenario A
Scenario B Scenario C
Scenario D
Scenario E
Scenario F
Nonmelanoma Skin Tumors
Additional Basal Cases 53,384,400
(12)
Additional Squamous
Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
* All estimates for 2
50,458,200
(14)
2,057,700
(14)

523,600
(11)
123,000
(10)

10,785,600
(6)
.5% Growth Scenario
50,593,800 47,744,000
(6)
47,282,900 44,078,600
(7)
1,929,900 1,800,900
(7)

497,100 470,400
(6)
117,500 111,900
(5)

10,512,300 10,214,500
(3)

Scenario A = Compound annual growth of 1 percent from 1985 to 2010,
ppm.
Scenario B = Linear growth at 0.01275 ppm per year.
Scenario C = Linear growth at 0.017 ppm per year.
Scenario D = Linear growth at 0.02125 ppm per year.
Scenario E ~ Compound annual growth of 1 percent per year.
Scenario F = Compound annual growth of 1 percent from 1985 to 2020,
by 2050 and thereafter.
43,994,600
(-8)
39,902,600
(-9)
1,632,700
(-9)

436,400
(-7)
104,500
(-7)

9,779,800
(-4)

28,059,300
(-41)
21,938,900
(-50)
909,700
(-49)

311,200
(-34)
74,100
(-34)

6,673,500
(-35)

10,538,800
(-78)
6,596,600
(-85)
280,100
(-84)

147,300
(-69)
33,100
(-70)

2,145,900
(-79)

followed by constant concentrations at 2.23
growing to 1.5 percent compound annual growth
Assumptions:

-  Current 1-D models accurately reflect global depletion;  Antarctic ozone hole has no impact on global ozone
   levels.
-  Greenhouse gases (other than CH4) that counter depletion grow at historical rates.
   Growth rates for ozone depleters are for global emissions;  it is assumed that emissions do not increase after
   2050.
-  Ozone depletion limited to 50 percent.

-------
                                                                           EXHIBIT 18-24

                                                      Materials,  Climate,  and Other Effects:   Methane Scenarios
                                                  (Figures in Parentheses  are Percentage Changes from Scenario C
METHANE SCENARIOS
TYPE OF EFFECT
Effects Estimated Quantitatively
for the U.S.
Materials Damage a/
Rise in Equilibrium
Temperature by 2075 b/
Sea Level Rise by 2075
Effects Based on Case Studies
and Research in Early Stages
Scenario A
662
(9)
5.4
(-4)
93
(-2)
Scenario B Scenario C
646 606
(7)
5.6 5.6
(0)
94 95
(-1)
Scenario D
570
(-6)
5.7
(2)
96
(1)
Scenario E
553
(-9)
5.8
(4)
96
(1)
Scenario F
532
(-12)
5.9
(5)
97
(2)
Units
Present Value
(millions of 1985 dollars)
Degrees Centigrade
Centimeters
    Cost of Sea Level Rise in
    Charleston and Galveston £

    Reduction in Soybean
    Seed Yield d/

    Increase in Ground-Based
    Ozone e/

    Loss of Northern Anchovy
    Population f/
                                   1100-2684
>7.5
8.3-44.1
                                                1106-2703
             >7.5
             7.6-40.2
                                                             1123-2748
                          >7.5
                          7.0-36.9
                                                                          1120-2742
                                       6.8
                                       6.3-33.4
                                                                                       1120-2742
                                                    5.7
                                                    5.2-27.7
>11.0->25.0  >11.0->25.0  >11.0->25.0  >11.0->25.0  6.3-17.4
                                                                                                    1128-2761
                                                                 4.3
                                                                 3.8-20.3
                                                                 1.4-12.3
                                                                                     Present Value (millions of 1985
                                                                                     dollars)
                                                                                     Percent in Year 2075
                                                                                     Percent in Year 2075
                                                                                                                                                                00
a/  Discounted over 1985-2075 using a real discount rate of 3 percent.
b/  Estimates of the rise in equilibrium temperature neglect increases  and decreases in tropospheric that would occur with deviations from the 2.5% Growth case of
    methane growth (1%) and changes in stratospheric water vapor.
£/  Lowest estimate with anticipation of sea level rise; highest estimate without anticipation.
d/  Essex and Williams cultivars; statistically significant results only.
e/  Lowest estimate is for Los Angeles, California; highest estimate is for Nashville,  Tennessee.
f/  Lowest estimate 15-meter vertical mixing of top ocean layer; highest estimate 10-meter vertical mixing if ozone depletion is greater than 11.2 percent and no
    mixing if ozone depletion is less than 11.2 percent.
g/  All estimates for 2.5% Growth Scenario.

Assumptions:

    Current 1-D models accurately reflect global depletion; Antarctic ozone hole has no impact on global ozone levels.
    Greenhouse gases (other than CH4) that counter depletion grow at historical rates.
    Growth rates for ozone depletion are for global emissions; it is assumed that emissions do not increase after 2050.
    Ozone depletion limited to 50 percent.
See text for definition of methane cases.

-------
                                        18-46
                                    EXHIBIT 18-25

                    Global Average Ozone Depletion:   Sensitivity
                to Relationship Between Ozone Depletion and Emissions
            (Based on 10th and  90th percentiles  of Monte  Carlo Analysis
                        of Rate Constants  and Cross-Sections
CL
a)
Q
0)
c
o
0
o
o
      0.0
    -10.0 -
    -20.0 -
    -300 -
    -40 0 -
    -50.0
        1985
                   2005
                              2025
      High
(90th Percentile)
	1	1	1— -~r	1	1	
      for
                                  Low
                                  (10th
                                   Percentile)

                                   for 2.5%
                                   growth
                                                                           'best'
                                                                           1-D estimate
                                                                           for
                                                                           2.5% Growth
                                                    2065
                                             2.5%  growth
                       2085
  Assumptions:

       Scenario  is 2.5% Growth scenario.

       Global depletion varies from best 1-D estimates.

       Greenhouse gases that counter depletion grow at historical rates.

       Growth rates for ozone depleters are for global emissions; it is assumed
       that  emissions do not increase after 2050.
       Ozone  depletion limited to 50 percent.

-------
                                     18-47
ranges of uncertainty that surround estimates of rate constants for atmospheric
reactions.   The range analyzed here is the tenth and ninetieth percentile
estimates.

    The effect of varying the sensitivity of ozone depletion to emissions is
illustrated in Exhibit 18-25.  The exhibit shows that the tenth percentile
estimate of ozone depletion is about one half of the 2.5% Growth scenario value.
The ninetieth percentile estimate is about twice the 2.5% Growth scenario value
and reaches 50 percent depletion in approximately 2078.

    The implications for health risks are displayed in Exhibits 18-26 to 18-28
for each of the three population cohorts.  In general, increased health risks
would be about 50 to 65 percent lower than the 2.5% Growth in the tenth
percentile case,  and roughly double the increased risks in the ninetieth
percentile case for the first two cohorts and about 30 to 45 percent higher for
people born from 2030 to 2074.  The differences are not as high for the third
cohort because ozone depletion was arbitrarily limited to 50 percent; the 50
percent depletion level is reached just after 2100 in the 2.5% Growth scenario
and about 2078 in the ninetieth percentile case, implying that for the period
during which the third cohort is alive, the amount of ozone depletion between
the 2.5% Growth scenario and the High (90th percentile)  case is very similar.

Alternative Maximum Ozone Depletion Assumption

    Throughout this Risk Assessment, the maximum level of ozone depletion
allowed was 50 percent.  This arbitrary limit was applied to avoid extrapolating
beyond the bounds of the atmospheric model upon which the parameterized
quantitative risk estimates have been based.  In instances where the 50 percent
limit is reached, this assumption underestimates the risks posed by ozone
depletion.   To evaluate the potential impact of this assumption, the sensitivity
of the risk estimates was analyzed assuming a maximum level of ozone depletion
of 95 percent.6

    The health risks estimated with the 95 percent ozone depletion limit are
shown in Exhibits 18-29 to 18-31 for the 2.5% and 5.0% Growth scenarios.  These
health risk estimates tend to increase significantly (1) as the rate of growth
in emissions increases, which increases the rate of ozone depletion and hence,
the likelihood that the 50% limit will become binding; and (2) for the later
cohorts, since ozone depletion is more likely to reach the 50% limit as time
goes on.  For example, for all cohorts a 95 percent limit has a higher
percentage impact on the risk estimates in the 5.0% Growth scenario compared to
the 2.5% Growth scenario.  This impact is also greater on people born later,
since in the 2.5% Growth scenario there is no increase in risk for people alive
today since 50% depletion is not reached in their lifetimes (see Exhibit 18-29),
while people born from 2030 to 2074 will have about a 5 to 10 percent increase
in risk (see Exhibit 18-31).
       This level of depletion is projected using the parameterized model.  Two
questions arise: (1) would the 1-D model project the same depletion level for
the same emissions scenario, and (2) would real world depletion occur the way
the model projects.  We do not know the answer to these questions, so caution
must be used in interpreting these results.

-------
                                     18-48
                                 EXHIBIT 18-26

                             Human Health Effects:
       Sensitivity to Relationship Between Ozone Depletion and Emissions
  Additional Cumulative Gases and Deaths Over Lifetimes of People Alive Today
       (Figures in Parentheses are Percentage Changes from Current Model)
         HEALTH EFFECT
                                  SENSITIVITY OF OZONE DEPLETION TO EMISSIONS
                                            2
                                                                             I/
  Senile Cataract

    Additional Cases
 Lo
Current Model
High"
Nonmelanoma Skin Tumors
Additional Basal Cases
Additional Squamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths

503,300
(-61)
324,400
(-62)
13,800
(-61)

5,500
(-60)
1,500
(-61)

1,289,300 2,772,600
(115)
846,500 1,891,900
(123)
35,700 79,500
(123)

13,900 29,000
(109)
3,800 8,000
(111)
240,500     613,900   1,307,100
 (-61)                  (113)
\/ Emissions of ozone depleters are from the 2.5% Growth scenario.
2/ 10% percentile of Monte Carlo runs of rate constants and cross-sections.
3/ 90% percentile of Monte Carlo runs of rate constants and cross-sections.

Assumptions

    Current 1-D models accurately reflect global depletion; Antarctic ozone hole
    has no impact on global ozone levels.
    Greenhouse gases that counter depletion grow at historical rates.
    Growth rates for ozone depleters are for global emissions; it is assumed
    that emissions do not increase after 2050.
    Ozone depletion limited to 50 percent.

-------
                                     18-49
                                  EXHIBIT 18-27

                              Hunan Health Effects:
        Sensitivity to Relationship Between Ozone Depletion and Emissions
 Additional  Cumulative Cases  and Deaths  Over Lifetimes  of People  Born 1986-2029
       (Figures  in Parentheses are  Percentage Changes  from Current Model)-'
                                  SENSITIVITY OF OZONE DEPLETION TO EMISSIONS
HEALTH EFFECT
Nonmelanoma Skin Tumors
Additional Basal Cases
Additional Squamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Low*7

4,665,600
(-61)
3,185,400
(-65)
134,000
(-65)
51,300
(-59)
13,600
(-58)
Current Model High^

11,998,800 23,670,800
(97)
9,091,300 19,638,900
(116)
378,100 809,900
(114)
124,400 234,500
(89)
32,500 59,800
(84)
  Senile Cataract

    Additional Cases
1,884,300     4,311,900     7,046,300
  (-56)                       (63)
  i/  Emissions of ozone depleters are from the 2.5% Growth scenario.
  2/  10% percentile of Monte Carlo runs of rate constants and cross-sections.
  3_/  90% percentile of Monte Carlo runs of rate constants and cross-sections.

Assumptions

    Current 1-D models accurately reflect global depletion; Antarctic ozone hole
    has no impact on global ozone levels.
    Greenhouse gases that counter depletion grow at historical rates.
--  Growth rates for ozone depleters are for global emissions; it is assumed
    that emissions do not increase after 2050.
--  Ozone depletion limited to 50 percent.

-------
                                     18-50
                                  EXHIBIT 18-28

                              Hunan Health Effects:
        Sensitivity to Relationship Between Ozone Depletion and Emissions
 Additional Cumulative Cases and Deaths Over Lifetimes of People Born 2030-2074
      (Figures in Parentheses are Percentage Changes from Current Model)-7
         HEALTH EFFECT
SENSITIVITY OF OZONE DEPLETION TO EMISSIONS

     Low-/     Current Model     High"/
  Nonmelanoma Skin Tumors

    Additional Basal Cases


    Additional Squamous Cases


    Additional Deaths


  Melanoma Skin Tumors

    Additional Cases


    Additional Deaths


  Senile Cataract

    Additional Cases
 19,949,600     47,744,000    64,862,800
    (-58)                       (36)

 14,947,400     44,078,600    63,901,200
    (-66)                       (45)

    622,300      1,800,900     2,597,400
    (-65)                       (44)
    221,900
    (-53)

     54,400
    (-51)
  5,444,800    10,214,500     11,717,700
    (-47)                        (15)
470,400
111,900
637,200
(35)
145 , 300
(30)
  !/  Emissions of ozone depleters are from the 2.5% Growth scenario.
  2./  10% percentile of Monte Carlo runs of rate constants and cross-sections.
  3>/  90% percentile of Monte Carlo runs of rate constants and cross-sections.

Assumptions

    Current 1-D models accurately reflect global depletion; Antarctic ozone hole
    has no impact on global ozone levels.
    Greenhouse gases that counter depletion grow at historical rates.
    Growth rates for ozone depleters are for global emissions; it is assumed
    that emissions do not increase after 2050.
--  Ozone depletion limited to 50 percent.

-------
                                                  EXHIBIT 18-29

                              Hunan Health Effects:   Mar<«.» Depletion of 95 Percent;
                   Additional Cuaulative Cases and Deaths Over Lifetimes of People Alive Today
                      (Figures in Parentheses are Percentage Changes from 501 Limit Scenario)
HEALTH EFFECT
Nonmelanoma Skin Tumors
Additional Basal Cases
Additional Squamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
EMISSIONS SCENARIOS
2.5X Growth 5,
SOX Limit 95X Limit 50% Limit

1,289,300 1,289,300 6,882,800
(0)
816,500 846,500 5,431,800
(0)
35,700 35,700 225,000
(0)

13,900 13,900 62,500
(0)
3,800 3,800 17,800
(0)

613,900 613,900 3,238,700
(0)

.OX Growth
95X Limit

8,780,400
(28)
7,624,500
(40)
313,100
(39)

75,300
(20)
21,500
(21)

4,094,600
(26)
Assumptions:

- Current 1-D models accurately reflect global depletion; Antarctic ozone hole has no impact on global ozone
  levels.
- Greenhouse gases that counter depletion grow at historical rates.
- Growth rates for ozone depleters are for global emissions; it is assumed that emissions do not increase after
  2050.
- Ozone depletion limited to 50 percent.

-------
                                                 EXHIBIT 18-30

                             Hunan Health Effects:   Minrin«»n Depletion of 95 Percent
                 Additional Cumulative Cases and Deaths Over Lifetimes of People Born 1986-2029
                     (Figures  in Parentheses are Percentage Changes from 50Z Limit Scenario)
HEALTH EFFECT
Nonmelanoma Skin Tumors
Additional Basal Cases
Additional Scfuamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
EMISSIONS
2.5% Growth
50% Limit 95% Limit

11,998,800 12,011,300
(1)
9,091,300 9,107,600
(2)
378,100 378,700
(2)

124,400 124,500
(0.1)
32,500 32,500
(0)

4,311,900 4,315,600
(0.1)
SCENARIOS
5.0% Growth
50% Limit 95% Limit

46,429,300 81,851,800
(76)
42,922,400 97,750,000
(128)
1,753,500 3,919,300
(224)

445,900 653,000
(46)
107,900 158,700
(47)

10,361,300 15,627,400
(51)
Assumptions:
     Current 1-D models accurately reflect global depletion;  Antarctic ozone hole has no impact on global ozone
     levels.
     Greenhouse gases that counter depletion grow at historical rates.
     Growth rates for ozone depleters are for global emissions;  it is assumed that emissions do not increase
     after 2050.
     Ozone depletion limited to 50 percent.
                                                                                                                                                       oo

-------
                                                 EXHIBIT 18-31

                             Hunan Health Effects:   M«rrtmn Depletion of 95 Percent
                 Additional Cumulative Cases and Deaths Over Lifetimes of People Bom 2030-2074
                     (Figures in Parentheses are Percentage Changes froH 501 LiMit Scenario)
EMISSIONS SCENARIOS
HEALTH EFFECT
Nonmelanoma Skin Tumors
Additional Basal Cases
Additional Squamous Cases
Additional Deaths
Melanoma Skin Tumors
Additional Cases
Additional Deaths
Senile Cataract
Additional Cases
2.5Z Growth
50Z Limit 95X Limit

47,744,000 50,452,800
(6)
44,078,600 48,085,800
(9)
1,800,900 1,959,600
(9)

470,400 485,400
(3)
111,900 115,900
(4)

10,214,500 10,818,100
(6)
5. OX Growth
50X Limit 95X Limit

78,600,200 167,287,400
(113)
81,244,300 246,526,800
(203)
3,290,400 9,763,400
(297)

771,500 1,293,100
(68)
170,300 278,700
(64)

12,405,300 19,740,300
(59)
Assumptions:
     Current 1-D models accurately reflect global depletion; Antarctic ozone hole has no impact on global ozone
     levels.
     Greenhouse gases that counter depletion grow at historical rates.
     Growth rates for ozone depleters are for global emissions; it is assumed that emissions do not increase
     after 2050.
     Ozone depletion limited to 50 percent.
                                                                                                                                                       oo
                                                                                                                                                        i
                                                                                                                                                       LO

-------
                                     18-54
SENSITIVITY OF EFFECTS TO UNCERTAINTY IN DOSE RESPONSE

    This section analyzes the sensitivity of risk estimates to uncertainties in
two areas:  (1) estimates of dose-response between UV radiation and health
effects, and (2) the choice of action spectrum.

    The statistical uncertainty in the dose-response relationships reported here
does not reflect uncertainties in the underlying specifications of the
dose-response relationships.  For example, as described in Chapter 8, there
remains a major question regarding the appropriate measure of exposure to use
for modeling melanoma skin cancer risks.  The uncertainty estimates reported
below reflect the implications of the statistical variation reported using one
particular measure of exposure.  Using alternative exposure measures would
increase the range of risk estimates.

Effects at High and Low Estimates of UV Radiation Impact on Human Health

    The dose-response relationships between increases in UV radiation and
associated health risks were obtained using a variety of statistical techniques
(see Chapters 7, 8, 10 and 17).  In these studies, the responsiveness of health
effects to changes in UV radiation were characterized by dose-response
parameters estimated using mathematical functions fitted statistically to
detailed epidemiologic data.  The results of the statistical analysis included
estimates of the standard errors of dose-response parameters.

    The standard errors of the dose-response parameters were used to develop
high and low estimates of health risks.  Low risk estimates were calculated by
using the 2.5% Growth scenario dose-response coefficients minus one standard
error; high risk estimates use the 2.5% Growth scenario use coefficients plus
one standard error.  All other assumptions are held fixed.

    Health risks estimated using the low, middle and high coefficients are
presented in Exhibits 18-32 through 18-34 for the 2.5% Growth scenario.  The low
dose-response coefficients produced an approximately 10 to 50 percent reduction
in health effects, while the high dose-response coefficients produced about a 10
to 85 percent increase.  The magnitude of the differences with the 2.5% Growth
scenario tend to increase as time goes on.

Choice of Action Spectrum

    This section analyzes the sensitivity of risk estimates to the action
spectrum that is specified to determine the relationship between ozone abundance
and UV flux.  In the 2.5% Growth scenario, the DNA-damage action spectrum
developed by Setlow was used to derive biologically effective UV radiation
fluxes for changes in ozone levels.  The sensitivity of the risk estimates to
the erythema action spectrum was investigated.  Health risks estimated using the
DNA action spectrum (which was used in the 2.5% Growth scenario) and the
erythema action spectrum are presented in Exhibits 18-35 to 18-37 for each of
the population cohorts.  On average, health risks using the erythema action
spectrum are about 10-15 percent lower than risks estimated with the DNA-action
spectrum due to the lower levels of weighted UV flux derived with the erythema
action spectrum.

-------
                                   18-55
                               EXHIBIT 18-32

     Human Health  Effects:   Sensitivity to Dose-Response Relationship
Additional Cumulative Cases and Deaths Over Lifetimes of People Alive Today
 (Figures in Parentheses are Percentage Changes from 2.5% Growth Scenario)
             HEALTH EFFECT
SENSITIVITY OF EFFECT TO UV DOSE
  Low      Standard     High
          Coefficient
      Nonmelanoma Skin Tumors

        Additional Basal Cases


        Additional Squamous Cases


        Additional Deaths


      Melanoma Skin Tumors

        Additional Cases


        Additional Deaths


      Senile Cataract

        Additional Cases
831,400
 (-36)

570,900
 (-33)

 24,000
 (-33)
 10,000
 (-28)

  3,400
 (-11)
345,800
 (-44)
1,289,300


  846,500


   35,700
   13,900
    3,800
  613,900
1,757,300
   (36)

1,134,100
   (34)

   47,900
   (34)
   17,700
   (27)

    4,200
   (11)
  810,100
   (32)
      Assumptions:

      - Current 1-D models accurately reflect global depletion; Antarctic
        ozone hole has no impact on global ozone levels.

      - Greenhouse gases that counter depletion grow at historical rates.

      - Growth rates for ozone depleters are for global emissions; it is
        assumed that emissions do not increase after 2050.
      - Ozone depletion limited to 50 percent.

-------
                                  18-56
                                EXHIBIT  18-33

       Human Health Effects:  Sensitivity to Dose-Response Relationship
Additional Cumulative Cases and Deaths Over Lifetimes of People Born 1986-2029
   (Figures  in Parentheses  are Percentage Changes  from 2.5% Growth Scenario)
          HEALTH  EFFECT
  SENSITIVITY OF EFFECT TO UV DOSE
   Low       Standard         High
 	Coefficient	
   Nonmelanoma  Skin Tumors

    Additional Basal  Cases


    Additional Squamous  Cases


    Additional Deaths


   Melanoma Skin Tumors

    Additional Cases


    Additional Deaths


   Senile Cataract

    Additional Cases
7,360,900
  (-39)

5,476,200
  (-40)

  228,200
  (-40)
2,387,400
  (-45)
11,998,800


 9,091,300


   378,100
 4,311,900
17,296,900
   (44)

13,771,600
   (51)

   570,000
   (51)
88,100
(-29)
29,200
(-10)
124,400
32,500
161,600
(30)
35,800
(10)
 5,761,300
   (34)
   Assumptions:

   -  Current 1-D models accurately reflect global depletion;  Antarctic ozone
     hole has no impact on global ozone levels.

   -  Greenhouse  gases that counter depletion grow at historical rates.

   -  Growth rates for ozone depleters are for global emissions; it is assumed
     that emissions do not increase after 2050.
   - Ozone depletion limited to 50 percent.

-------
                                    18-57
                                 EXHIBIT 18-34

       Human Health Effects:  Sensitivity to Dose-Response Relationship
Additional Cumulative Gases and Deaths Over Lifetimes of People Born 2030-2074
   (Figures in Parentheses are Percentage Changes fron 2.5% Growth Scenario)
                                       SENSITIVITY OF EFFECT TO UV DOSE
          HEALTH EFFECT
   Low
 Standard         High
Coefficient	
   Nonmelanoma Skin Tumors

     Additional Basal  Cases


     Additional Squamous Cases


     Additional Deaths


   Melanoma Skin Tumors

     Additional Cases


     Additional Deaths


   Senile Cataract

     Additional Cases
26,911,600
  (-44)

22,076,300
  (-50)

   911,300
  (-49)
   324,400
  (-31)

    99,600
  (-11)
 5,529,300
  (-46)
47,744,000


44,078,600


 1,800,900
   470,400


   111,900
10,214,500
75,759,400
   (59)

81,760,300
   (85)

 3,300,900
   (83)
   628,600
   (34)

   124,200
   (11)
13,877,300
   (36)
   Assumptions:

   -  Current 1-D models accurately reflect global depletion;  Antarctic ozone
     hole has no impact on global ozone levels.

   -  Greenhouse  gases that counter depletion grow at historical rates.

   -  Growth rates for ozone depleters are for global emissions; it is assumed
     that emissions do not increase after 2050.
   -  Ozone depletion limited to 50 percent.

-------
                                     18-58
                                 EXHIBIT 18-35
                             Human Health Effects:
    Sensitivity to Relationship Between Ozone Depletion and Action Spectrum
  Additional Cumulative Cases and Deaths Over Lifetimes of People Alive Today
   (Figures in Parentheses are Percentage Changes from 2.5% Growth Scenario)
                            SENSITIVITY OF OZONE DEPLETION TO ACTION SPECTRUM
         HEALTH EFFECT             Erythema                DNA
	(2.5% Growth Scenario')   (2.5% Growth Scenario)

Nonmelanoma Skin Tumors

    Additional Basal Cases           1,170,300          1,289,300
                                        (-9)

    Additional Squamous Cases          763,300            846,500
                                        (-10)

    Additional Deaths                   32,200             35,700
                                        (-10)

Melanoma Skin Tumors

    Additional Cases                    11,500             13,900
                                        (-17)

    Additional Deaths                    3,500              3,800
                                         (-8)

Senile Cataract

    Additional Cases                   509,800            613,900
                                        (-17)


Assumptions:

    Current 1-D models accurately reflect global depletion; Antarctic ozone hole
    has no impact on global ozone levels.

    Greenhouse gases that counter depletion grow at historical rates.

    Growth rates for ozone depleters are for global emissions; it is assumed
    that emissions do not increase after 2050.

    Ozone depletion limited to 50 percent.

-------
                                      18-59
                                 EXHIBIT 18-36
                             Human Health Effects:
    Sensitivity to Relationship Between Ozone Depletion and Action Spectrum
 Additional Cumulative Cases and Deaths Over Lifetimes of People Born 1986-2029
    (Figures in Parentheses are Percentage Changes from  2.5% Growth Scenario)
                            SENSITIVITY OF OZONE DEPLETION TO ACTION SPECTRUM
         HEALTH EFFECT                 Erythema                DNA
	(2.5% Growth Scenario)    (2.5% Growth Scenario)

Nonmelanoma Skin Tumors

    Additional Basal Cases           10,800,800            11,998,800
                                        (-10)

    Additional Squamous Cases         8,004,100             9,091,300
                                        (-12)

    Additional Deaths                   333,700               378,100
                                        (-12)

Melanoma Skin Tumors

    Additional Cases                    102,800               124,400
                                        (-17)

    Additional Deaths                    29,600                32,500
                                         (-9)

Senile Cataract

    Additional Cases                  3,583,400             4,311,900
                                        (-17)


Assumptions:

    Current 1-D models accurately reflect global depletion; Antarctic ozone hole
    has no impact on global ozone levels.

    Greenhouse gases that counter depletion grow at historical rates.

    Growth rates for ozone depleters are for global emissions; it is assumed
    that emissions do not increase after 2050.

    Ozone depletion limited to 50 percent.

-------
                                     18-60
                                 EXHIBIT 18-37

                             Human Health Effects:
    Sensitivity to Relationship Between Ozone Depletion and Action Spectrum
Additional Cumulative Cases and Deaths Over Lifetimes of People Born 2030-2074
   (Figures in Parentheses are Percentage Changes from 2.5% Growth Scenario)
                            SENSITIVITY OF OZONE DEPLETION TO ACTION SPECTRUM
         HEALTH EFFECT                 Erythema                DNA
	(2.5% Growth Scenario)   (2.5% Growth Scenario)

Nonmelanoma Skin Tumors

    Additional Basal Cases           43,173,200            47,744,000
                                        (-10)

    Additional Squamous Cases        38,368,300            44,078,600
                                        (-13)

    Additional Deaths                 1,572,600             1,800,900
                                        (-13)

Melanoma Skin Tumors

    Additional Cases                    390,500               470,400
                                        (-17)

    Additional Deaths                   103,200               111,900
                                         (-8)

Senile Cataract

    Additional Cases                  8,631,500            10,214,500
                                        (-15)


Assumptions:

    Current 1-D models accurately reflect global depletion; Antarctic ozone hole
    has no impact on global ozone levels.

    Greenhouse gases that counter depletion grow at historical rates.

    Growth rates for ozone depleters are for global emissions; it is assumed
    that emissions do not increase after 2050.

    Ozone depletion limited to 50 percent.

-------
                                     18-61
RELATIVE IMPORTANCE OF KEY UNCERTAINTIES

    As expected, quantitative estimates of risks to human health and the
environment are sensitive to the assumptions, methods, and data used to model
risks.  It appears that the 1-D parameterization underestimates ozone depletion
compared to 2-D models.

    The largest range of risk estimates was produced by varying assumptions
about future production, use, and emissions of ozone-modifying substances.
Estimates of ozone depletion were particularly sensitive to these assumptions.
The 80% Reduction scenario leads to a proj ected increase in ozone levels; the
highest case scenario (i.e., the 5.0% Growth scenario) results in extremely
rapid depletion.  In addition, the 3.8% Growth and 5.0% Growth scenarios caused
ozone depletion estimates by 2100 to reach an amount believed to be outside the
valid range of the atmospheric models used to analyze changes in total column
ozone (in the 2.5% Growth scenario, this amount of depletion was reached just
after 2100).

    Because estimates of human health effects are driven by projections of ozone
depletion, health risks were similarly affected by this range of emissions
assumptions.  Increased risks to people alive today were found to vary by as
much as 500 percent.  The sensitivity of health risks to emissions is greater in
future generations; individuals yet to be borne were simulated to experience
greater changes in exposure to UV radiation over a larger fraction of their
lives.  Environmental effects were also shown to be most sensitive to emissions
assumptions; however, compared to health effects, the range of results was
narrower.

    The health risk estimates were also sensitive to variations in the
relationship between ozone depletion and emissions.  Changing the responsiveness
of ozone depletion to emissions in the 2.5% Growth scenario resulted in up to 65
percent fewer additional cases of skin cancers and senile cataract among people
alive today (low sensitivity) and up to 120 percent more additional cases for
these individuals (high sensitivity).

    Assumptions about potential actions taken to limit global warming were also
found to be important for estimating health risks.  Actions aimed at limiting
future warming to 2°C or 3°C would reduce future potential concentrations of
greenhouse gases.  Because these gases also counter ozone depletion, an increase
in ozone depletion and associated UV exposure would accompany such actions.  For
example, in the 2.5% Growth scenario where global warming is limited to 3°C,
people alive today would experience about 45 percent more cases of skin cancer
and senile cataract compared to the number of cases estimated to occur in the
absence of actions to limit warming.  The sensitivity of health risks to
assumptions about potential limits to global warming is greater for future
generations; in the 2.5% Growth scenario up to 90 percent more cases of skin
cancer are estimated for people born during 2030 through 2074 if global warming
is limited to 3°C.

    Health risk estimates also were sensitive to assumptions about future
concentrations of methane,  but to a lesser extent.  The analysis suggests that
for the 2.5% Growth scenario people alive today will contract about 10 to 15
percent more cases of skin cancer and senile cataract if lowest methane growth
assumptions are used.  Note that lowering methane concentrations increases ozone
depletion and hence increases UV radiation damaging to human health.

-------
                                     18-62
    Health risks were also sensitive to the uncertainty of the parameters used
to characterize dose-response.  However, low and high estimates of cases and
deaths associated with low and high dose-response coefficients (plus and minus
one standard error) fell well within the ranges of risk estimates produced by
varying assumptions about emissions and other factors.  Nonetheless, health
risks were found to decrease by as much as 50 percent for low dose-response
coefficients and increase up to 85 percent for high dose-response coefficients.

    The choice of action spectrum could also affect the health risk estimates,
although the variation between the DNA action spectrum and the Erythema action
spectrum is unlikely to exceed 10 to 15 percent (the Erythema action spectrum
has lower impacts).

    Health risks were also sensitive to the assumption that ozone depletion
could not exceed 50 percent.  For example, in the 2.5% Growth scenario, risk
estimates could increase up to 10 percent if ozone depletion exceeds 50 percent;
for higher emission scenarios, the increase in risk could be significantly
greater.

    Exhibit 18-38 shows the comparative impact of these uncertainties, using
mortality among people born before 2075 as an indicator of the sensitivity of
the risk estimates to each factor.  Clearly, for current models, the emission of
ozone depleting substances is the most sensitive variable, although future
greenhouse policies, dose-response coefficients, and model response all are
about half as important in explaining differences in estimates of mortality as
production estimates.

SUMMARY

    The risk of stratospheric modification on human health and the environment
depends on the level of future emissions of ozone depleters.  Lower levels of
emissions always reduces risks, with smaller increments of risk reduction
achieved for greater percentage reductions in emissions.  The benefits of lower
emissions of ozone depleters  could substantially increase if reductions are made
in greenhouse gases  (evaluated within this Risk Assessment) and if the Antarctic
ozone hole is found  to have global implications  (assumed herein not to be
relevant).  While uncertainly in many factors makes risk estimates
quantitatively variable, the  range of uncertainty does not alter the  fundamental
conclusion that risks are significant.

-------
Mortality
(millions)
                                              18-63


                                          EXHIBIT 18-38

           Comparative Sensitivity of Mortality From Skin Cancer to Various Factors"
6.0
5.0
4.0
3.0
2.0
1.0
0.0
—
5.6
;

4.7

SENSITIVITY TO EMISSIONS GROWTH
2.4
0.2
<-0.01 -0.1

          5% Growth     3.8% Growth    2.5% Growth    1.2% Growth    No Growth    80% Reduction
         Growth rates refer to global production of chlorine and bromine bearing
         compounds.
         a/  Estimate of additional mortality from skin cancer for all people born before
             2075.

         Assumptions:

             Current 1-D models accurately reflect global depletion; Antarctic  ozone hole
             has no impact on global ozone levels.

             Greenhouse gases that counter depletion grow at historically-extrapolated
             rates.

             Growth rates for ozone depleters are for global emissions;  it  is assumed
             that emissions do not increase after 2050.

             Ozone  depletion limited to 50 percent.

-------
Mortality
(millions)
                                               18-64
                                          EXHIBIT 18-38
                                            (Continued)
            Comparative Sensitivity of Mortality From Skin Cancer to Various Factors"/
    5.0


    4.0


    3.0


    2.0


    1.0


    0.0
      4.3
  2.4
                  2.4
              1.3
SENSITIVITY TO VARIOUS ASSUMPTIONS
                        -,  (All estimates use 2.5% Growth)
                     //A                             3.7
                                  2.4
                                      2.1
                                                 2.4
                                              0.8
                                                                 2.4
                                           2.5
             No Llmit/3C Limit Low/Middle/High
                Greenhouse        Dose-
                               Response
                               Coefficient
Gas Limits
     DNA/Erythema 10% / 50% / 90%
       Action          Model
      Spectrum        Response
50%Limit/95%Limit
    Limiting
   Depletion
     to 50%
         a/  Estimate of additional mortality from  skin  cancer  for all people born before
             2075.

         Assumptions:

             Current 1-D models accurately reflect  global  depletion;  Antarctic ozone hole
             has no impact on global ozone levels.

         --  Greenhouse gases that counter depletion grow  at  historically-extrapolated
             rates (except in the scenario of greenhouse gas  limits).

             Growth rates for ozone depleters are for  global  emissions;  it is assumed
             that emissions do not increase after 2050.

             Ozone depletion limited to 50 percent  (except in the  sensitivity scenario in
             which the limit is increased to 95 percent).

-------
                                      18-65
REFERENCES


Canan, F. and J.K. Hammitt  (1986).   "An Analytic  Method for Constructing
    Scenarios From a Subjective  Joint Probability Distribution,"  The RAND
    Corporation, Santa Monica, California.

Hansen, J., A. Lacis, D. Rind, G.  Russell,  P.  Stone,  I.  Fung,  R. Ruedy, and
    J. Lerner (1984), "Climate Sensitivity:  Analysis of Feedback Mechanisms,"
    in Hansen, J.E., and T.  Takahashi,  (eds.), Climate Processes and Climate
    Sensitivity. Geophysical Monograph 29,  Maurice Ewing Volume 5, American
    Geophysical Union, Washington,  DC.

Isaksen, I.S.A., and F. Stordal, (1986),  "Ozone  Perturbations  by Enhanced
    Levels of CFCs, N20, and CH4:   A Two-Dimensional  Diabatic  Circulation Study
    Including Uncertainty  Estimates," Journal  of Geophysical Research. 91(D4):
    5249-5263.

Manabe, S., and R.T. Wetherald  (1986),  "Reduction in  Summer Soil Wetness
    Induced by an Increase in Atmospheric Carbon Dioxide," Science. 232:
    626-628.

Ramanathan, V., R.J. Cicerone, H.B. Singh,  and J.T. Kiehl (1985), "Trace Gases
    and Their Potential Role in  Climate  Change," National Center for Atmospheric
    Research, NCAR/0304/84-9, Boulder,  Colorado.

Rowe, R.D. , and R.M. Adams (1987.),  Analysis of Economic Impacts Of Lower Crop
    Yields Due To Stratospheric  Ozone.  Draft Report,  prepared  for U.S.
    Environmental Protection Agency, No.  68-01-7033.

Washington, W.M., and G.A. Meehl (1984),  "Seasonal Cycle Experiment on the
    Climate Sensitivity Due to  a Doubling of C02 With an Atmospheric General
    Circulation Model Coupled to a Simple Mixed-Layer Ocean Model," Journal of
    Geophysical Research.  89(D6):  9475-9503.
. 5. GOVERNMENT PRINTING OFF ICE:1938-516-002:80035

-------